Figure 1a shows the schematic diagram of the fabricated MoS2 photo-TFT with the conventional bottom gate structure. The number of MoS2 layers in the as-fabricated phototransistor was confirmed from the AFM height profiles, as depicted in Fig. 1b. As shown in the figure, the height of the mono-layer MoS2 on the substrate is around 0.7 nm or more, which is slightly higher than the theoretical thickness of 6.15 Å due to the absorber on the MoS2 surface. Figure 1c presents the Raman spectrum of MoS2 at different thicknesses. The \( {\text{E}}^{ 1}_{{ 2 {\text{g}}}} \) mode close to 383 cm−1 and the A1g mode close to 408 cm−1 are observed from the mono-layer to the bulk MoS2. As shown in Fig. 1d, with an increase in the number of MoS2 layers, the frequency of the \( {\text{E}}^{ 1}_{{ 2 {\text{g}}}} \) peak decreases whereas that of the A1g peak increases. An increase in the number of MoS2 layers resulted in a decrease in the Van der Waals force [19] between adjacent layers, causing a red shift of the \( {\text{E}}^{ 1}_{{ 2 {\text{g}}}} \) peak. Moreover, the Van der Waals force at each MoS2 layer suppresses the vibration as the number of layers is increased. This produces a higher force constant [20], resulting in a blue shift of the A1g modes.
Figure 2a presents the IDS–VGS characteristics of the MoS2 phototransistors with different layer thicknesses under dark and illuminated conditions. An increase in the dark-state ON current as well as the photocurrent in the illuminated state was observed with an increase in the MoS2 thickness from the monolayer to the bulk. Figure 2b shows the photocurrent (IPhoto) to dark current (IDark) ratio in the off state and the drain current in the on state as a function of the layer thickness. The increase in the drain current with the layer thickness can be explained by the increased carrier concentration. The effects of the layer thickness on the persistent photoconductivity (PPC) of the fabricated MoS2 phototransistors are shown in Fig. 2c. The PPC measurements were carried out by exposing the MoS2 TFTs to light pulses at a wavelength of 400 nm with a fixed intensity (5 mW/cm2). Figure 2d shows the decay time with the maximum photocurrent for the different layers obtained from Fig. 2c. Here, the decay time represents the time required for the photocurrent to decrease from the maximum level to one-fifth of its maximum value. It can be seen that the decay time and the magnitude of the maximum photocurrent of the phototransistors increase with an increase in the layer thickness.
The optical property of the as fabricated MoS2 phototransistors were significantly enhanced by adapting see-through transparent electrodes instead of the traditional global bottom gate or patterned local bottom gate structures and were compared with those of conventional opaque electrodes and transparent IZO electrodes. Figure 3a–c show the IDS–VGS characteristics of MoS2 phototransistors with various metal electrodes. This measurement was carried out under both dark and light conditions using a focused laser with different wavelengths at steps of 100 nm.
From Fig. 3c, it can be seen that the use of IZO transparent metals limits the optical performance of the MoS2 phototransistor due to the high Schottky barrier resulting from the Fermi-level pinning effect caused by its high workfunction (~ 5 eV). Furthermore, the sheet resistance of the IZO metal electrode obtained from four-point probe measurements was as high as 105 Ω/square. On the other hand, the sheet resistance of the see-through metal electrode was 8 Ω/square, much lower than that of IZO, which is not significantly different from the value of 1.4 Ω/square, which is the sheet resistance of a conventional Ti/Au metal electrode. The see-through metal was chosen not only for its sheet resistance properties but also for its transmittance capabilities. This see-through metal electrode shows transmittance of 70% under visible light of 532 nm and allows incident light to reach the entire channel area below the source and drain electrodes.
The optical properties of the external quantum efficiency (EQE), responsiveness (R) and collected carrier density (ncoll) of the phototransistors were extracted and calculated from Eqs. (1), (2) and (3) as a function of the wavelength depending on the different contact electrodes [21, 22]. These results are shown in Fig. 4.
$$ {\text{EQE}} = \frac{{I_{DS} /q}}{{P_{i} /h\upsilon }} $$
(1)
$$ {\text{R}} = \frac{{J_{total} - J_{dark} }}{{P_{i} }} $$
(2)
$$ \eta_{coll} = \frac{{I_{DS} }}{{q\mu_{FE} \left( {W/L} \right)t_{s} V_{DS} }}. $$
(3)
Here Pi is the power density in the illumination state, hν is the incident photon energy, JTotal is the current density in the illumination state, JDark is the current density in the dark state, μFE is the field-effect mobility of each device and tS is the MoS2 layer thickness.
As shown in Fig. 4, MoS2 phototransistors with see-through metal electrodes exhibit significantly improved optical properties as compared to the thick opaque or IZO transparent metal in the visible region. This results from the enhancement in the photocurrent due to the penetration of incident light to the entire active region below the transparent electrode, as described above.
To identify the photocurrent mechanism, the photocurrent of the multi-layer MoS2 phototransistor with the see-through metal electrode was measured by locally illuminating the MoS2 channel at different positions (inset of Fig. 5b). A beam with a wavelength of 532 nm at an intensity level of 0.99 μW was used for this purpose. As indicated by the IDS–VGS characteristics presented in Fig. 5b, the photocurrent of the MoS2 TFT is highest when the beam is located at the source position (A), after which it decreases along the channel (B, C, D), and is lowest at the drain (E). This can be explained by the barrier height variation (BHV), i.e., ΔφB, between the source and the channel due to the incident light. The BHV in this case is mainly caused by the electrostatic force induced at the junction between the metal and the semiconductor. It can be expressed by the following equation.
$$ I_{DS} = I_{DS0} \exp \left( {\alpha \frac{{q\Delta \varPhi_{B} }}{kT}} \right). $$
(4)
Here IDS0 is a reference current value without variation of the barrier height, α is a constant and kT is the thermal energy at room temperature.
Figure 5b shows the BHV depending on the beam position obtained from Eq. (4). As expected, the BHV at the source position has the largest value. Due to the increased BHV, a greater amount of electron injection (ninj) occurs from the source and causes an increase in the photocurrent [23]. From Eqs. (3) and (4), ninj can be deduced as follows:
$$ I_{DS} = I_{DS0} \exp \left( {\alpha \frac{{q\Delta \varPhi_{B} }}{kT}} \right). $$
(5)
Here n0 is a constant indicating the reference carrier density.