SEM image and XRD analysis of the grown TiS3 microcrystals are presented in Figure S1. Before transferring the flakes, the surface of paper was investigated by SEM and EDX measurements. As can be seen in Figure S2a, the paper is formed by stacked cellulose fibers and it presents deep gaps/voids between fibers. EDX analysis also confirms the presence of carbon (C), oxygen (O), calcium (Ca), silicon (Si), aluminum (Al), iron (Fe), sodium (Na), and manganese (Mg) in the raw paper where their atomic percentages are listed in the inset of Figure S2b.
Figure 1 shows the characterization of the deposited flakes on paper substrates. The schematic illustration of the TiS3-MoS2 heterostructure and its corresponding photograph are also presented in panels (a) and (b) of Fig. 1, respectively. Figure 1c exhibits the SEM image of the MoS2 flakes on the paper. According to the figure, MoS2 flakes cover the fibers and gaps between fibers of the paper. The deposited TiS3 flakes are also presented in Fig. 1d in which the thickness of the film is measured to be ~ 10 µm. Moreover, the SEM image of the TiS3-MoS2 heterostructure is observed in Fig. 1e. Due to softness and flexibility of MoS2, it forms a smoother film than TiS3 on paper, which is visible in Fig. 1e. The EDX analysis of MoS2, TiS3, and TiS3-MoS2 samples on papers is presented in Figs. 1f-g. By rubbing MoS2 on the paper, the Mo and S elements becomes significant (Fig. 1f). The EDX analysis of TiS3 sample also proves the presence of S and Ti elements according to Fig. 1g. SEM-EDX analysis of transferred TiS3 flakes is also presented in Figure S3. In the case of TiS3-MoS2 film, dominant elements are S, Mo, and Ti (Fig. 1h). The C, O and Ca elements come from raw paper. Table S1 summarizes the element contents of all samples with their weight and atomic percentages.
Figure 2a exhibits the magnified SEM image of the TiS3-MoS2 film where the MoS2 film is deposited on the TiS3 film. However, some TiS3 microcrystals may also appear on top of the film due to the rubbing process. The Raman spectra of MoS2, TiS3, and TiS3-MoS2 samples are also presented in Fig. 2b. For MoS2, E12g and A1g peaks are found at 379 and 403 cm− 1, respectively15. In the case of TiS3, the 193, 385 and 627 cm− 1 peaks refer to Ag and Eg modes, respectively16. There is also one additional peak located at 507 cm− 1, which refers to B1g peak of TiO217. This oxide peak may have appeared during the growth of TiS3 microcrystals in the ampoule process. The Raman spectrum of TiS3-MoS2 sample contains all peaks of both structures. Figure 2c shows the mapping analysis of the TiS3-MoS2 sample where the dominant elements are separately shown in Fig. 2d. Accordingly, O, C, and Ca elements are due to the raw paper, and elements of Ti, Mo, and S originate from the TiS3 and MoS2 films. It can be seen that all elements are uniformly distributed on the paper.
To investigate the optoelectronic properties of the films, MoS2, TiS3, and TiS3-MoS2 PDs were fabricated. Details of their fabrication are provided in the experimental section. The fabrication steps of MoS2 and TiS3 PDs are shown in Figures S4 and S5, respectively. In the case of TiS3-MoS2 PD, the fabrication steps are presented in Fig. 3.
Panels (a) to (c) of Fig. 4 display I-V characteristics of the MoS2, TiS3, and TiS3-MoS2 PDs at a bias range of -10 to + 10 V in dark and under 532 nm laser illumination at different power intensities. In all three PDs, significant photocurrents were generated compared to the dark state. The incident power is normalized in terms of laser spot and PD’s channel areas. Accordingly, the effective incident power is calculated as:
Peff = Plaser×Adevice/Alaser (1)
Where Plaser is the power of the laser, Adevice is the area of the PD’s channel, and Alaser is the area of the laser spot. The diameter of the 532 nm laser spot was 2.81 mm and its area (Alaser) was measured to be 6.22 mm2. As the laser intensity increases, the photocurrent also increases for all three PDs. In detail, the drain currents were measured in the range of ~ 0–4 nA, ~ 0–8 µA and ~ 0-450 nA for MoS2, TiS3, and TiS3-MoS2 PDs at a biasing voltage of 0–10 V, respectively. Interestingly, the Ids in TiS3-MoS2 is significantly higher than that of individual MoS2, which indicates the effective role of the band offset formation in increasing the photocurrent. The I-V curves of the all PDs in response to 405, 655, and 810 nm laser illuminations are also shown in Figure S6 where all PDs respond to the laser irradiation by an increase in current. Figure 4d-i compare the photocurrent and photoresponsivity of all three PDs at different power intensities of 532 nm at an applied voltage of 10 V. According to Fig. 4d-f, with increasing effective incident power (from zero to a few mW), increase in the photocurrents are occurred. In the MoS2 PD, the photocurrent increases to about 1.6 nA. For TiS3, this value is more noticeable and a 1.0 µA increase of photocurrent is observed. In the case of TiS3-MoS2 PD, photocurrent increases up to 300 nA. The photoresponsivity of the PDs are calculated based on the following Eq. 18:
R = Iph/Peff (2)
Where Iph is the photocurrent of the PDs. In the MoS2 PD, increasing the effective power associates with decrease of the photoresponsivity. The same trends are also observed for TiS3, and TiS3-MoS2 PDs. As the incident power increases, more photocarriers are generated, which increases the recombination rate of photogenerated carriers or the possibility of being captured by the traps, leading to a decrease in R 19,20. In general, the photoresponsivities are measured in the range of 0.4–1.2 µA/W, 0.4–1.6 mA/W, and 90–170 µA/W for MoS2, TiS3, and TiS3-MoS2 PDs, respectively. Accordingly, the photoresponsivity of the last two PDs is three and two orders of magnitude greater than that of MoS2 PD.
Figure 5a-c show the measured photocurrent in terms of the different effective powers of the 532 nm laser. Accordingly, an increase in the optical powers leads to an increase in photocurrent in all PDs. Panels (d) to (f) of Fig. 5 display the time trace response of the PDs to 532 nm laser illumination at a drain voltage of 10 V at an incident power of 10 mW. Rise time (τon) is defined as the time taken to increase the current from 10–90% of its baseline under laser illumination21. The fall time (τoff) is the time required to reduce the current from 90–10% of its baseline22. For MoS2, the rise and fall times are measured to be 2.17 and 9.99 s, respectively. In the TiS3 PD, the τon and τoff of 2.4 and 17.44 s are measured and in the TiS3-MoS2 PD, τon and τoff of about 0.96 and 8.42 s are calculated that it shows faster switching performance compared with two other PDs.
Figure 6a compares the photoresponsivity of several fabricated PDs at an applied voltage of 10 V under a laser wavelength of 532 nm. It is observed that all three TiS3 PDs have the highest photoresponsivity, followed by TiS3-MoS2 and MoS2 samples, respectively. In general, the R are measured in the range of 0.67–1.56 mA/W for TiS3 PDs and 0.20–1.93 µA/W in the case of MoS2 PDs. For TiS3-MoS2 PDs, these values are measured in the range of 0.08 to 0.19 mA/W. Accordingly, MoS2 has the lowest photoresponsivity and TiS3 possess the highest photoresponsivity. However, both photoresponsivity and photocurrent of TiS3-MoS2 PDs are considerable compared to the MoS2 PD. Moreover, a similar trend is observed in the order of magnitude of all calculated R in these PDs. Figure 6b shows the photoresponsivity of all three devices in terms of different laser wavelengths at the same power of 15 mW and supply voltage of 10 V. All three PDs have a broad response within the range of 400 to 800 nm. The photoresponsivity of the TiS3 PDs are one order of magnitude greater than that of the TiS3-MoS2 PDs and two orders greater than that of the MoS2 PDs. Figure 6c compares the measured Ion/Ioff ratio and dark current of all three PDs under 532 nm laser irradiation. The error line shows the deviation of the values obtained for the same PDs. The on-off ratio (Ion/Ioff) is usually used to describe the signal-to-noise ratio23. According to Fig. 6c, the values of 1.35, 0.17, and 1.82 are obtained for MoS2, TiS3, and TiS3-MoS2 PDs, respectively. Based on them, the highest signal-to-noise ratio is related to the TiS3-MoS2 PDs. Interestingly, the dark currents are lowest in MoS2 (~ 1.2 nA) and highest in TiS3 (~ 6.0 µA). Since the light detection mechanism is based on the change of drain current in these PDs, less dark current provides better performance in light detection as will be discussed below24. Another important parameters are the rise and fall times which are presented in Fig. 6d. Accordingly, the TiS3-MoS2 PD has the fastest response time to laser radiation than the other two PDs. Figure 6e shows the energy band diagram of the TiS3-MoS2 heterostructure and its corresponding photodetection mechanism. TiS3 is an n-type semiconductor with an energy band gap of 1.1 eV and MoS2 is also an n-type semiconductor with an energy band gap in the range of 1.2–1.8 eV 25,26. Here, the energy band gap of bulk MoS2 is considered because most layers are thick, although the deposited film can contain single and few layers of MoS2. The electron affinities of MoS2 and TiS3 are around 4 and 4.7 eV, respectively, and hence the conduction band of TiS3 is located of MoS2, forming an n-n+ heterostructure27,28. In the MoS2 film under laser irradiation, electron-hole pairs are generated and the photogenerated electrons enter into the TiS3 due to its lower conduction band which prevents electrons from recombination with the holes. Moreover, the valence band of MoS2 is located higher than that of TiS3 resulting in transferring of minority holes from TiS3 into MoS2. Hence, the recombination rate of the carriers is decreased in the heterostructure. Such charge separation in the heterostructure is responsible for faster response time, and improved photoresponsivity of the heterostructure compared with individual MoS2 PDs.
Since the substrate is made of paper, the performance of the introduced PDs was thoroughly investigated under applied strain. For this purpose, a homemade setup was fabricated to apply strain to the substrates as shown in Figure S7. This setup includes a cage covered with aluminum foil to act as a Faraday cage. The sample is fixed to a motor shaft through a hook to provide upward and downward strains by moving in a clockwise or counter-clockwise direction. The strain applied to the PDs was calculated according to ε = t/2R equation, where t is the thickness of the substrate and r is the radius of bending curvature29. Figure S8 presents the applied strains in the PDs, which are performed in the range of -0.54% to + 0.54%˚. As can be seen, at strains larger than ± 0.33%, a significant bending is occurred in the substrates.
Figures 7a and 7b show the schematic of applying the upward and downward strain in the PDs under exposing to 532 nm laser illuminations, respectively. Panels (c), (d), and (e) of Fig. 7 compare the Ids-V curves of the MoS2, TiS3, and TiS3-MoS2 PDs in the flat state and under negative and positive strains, respectively. For the MoS2 sample, it is observed that the drain current in the flat state almost remains the same as under different applied strains without any significant change. The flexible and soft structure of MoS2 is the main reason for this behavior30. For TiS3, it is found that the Ids decreases after applying downward strains, but a more significant decrease is observed in the upward strains. Moreover, after upward strain, by returning the TiS3 to its flat position, current does not return to its original value of before any strain. The structure of TiS3 is more fragile than that of MoS2, and bending can break the contact between TiS3 and Ag paste, affecting the conductivity31. As a result, even after the sample has returned to the flat state, the current dramatically suffers from decline. A similar trend is observed in the TiS3-MoS2 PD, but with a more noticeable decrease after positive strains, which could be due to the separation of the two films at the interface and the reduction of the electric field in the channel. Figures 7f-h show the photocurrent characteristics of all PDs under 532 nm laser irradiation at a drain voltage of 10 V and an incident power of 15 mW. For MoS2 PD, the photocurrent is well switched under negative and positive strains of 0.33% similar to the flat state. In the case of TiS3, the photocurrent decreases in both applied strains of -0.33% and + 0.33%, but TiS3-MoS2 shows a more noticeable decrease compared to its flat state. As discussed, the brittle structure of TiS3 as well as the separation of the two films at the junction, especially at a strain of + 0.33%, play important roles in this performance deterioration of the TiS3 based PDs.
Numerical simulation is employed to present further insight about the physics of light-induced generation-recombination processes for charged carriers in these devices. We used the model to calculate the photocurrent of the devices in response to the incident light under various biasing conditions. Details of the simulation can be found in the supporting information section. Panels (a) to (c) of Fig. 8 depict the schematic of the devices consisting MoS2, TiS3 and heterostructure with stacked layers of both materials. Overlapped part in the heterostructure device was exposed to the incident light with the wavelength of 532 nm. Stray electromagnetic field due to unintentional exposure is also took into the account. Corresponding light-induced electric field can be seen in Fig. 8d for all devices due to laser exposure. Vertical electric field was established by applying a voltage difference between the two terminals at both ends, where contacts are located. Figure 8e-f show the electron and hole concentrations under this biasing condition in a logarithmic scale, respectively. As can be seen, the distribution of electrons and holes in the two structures of MoS2 and TiS3 is almost uniform, but in the TiS3-MoS2 device, the band offset accelerates the separation of electrons and holes and causes a non-uniform distribution of carriers at the two-layer boundary. These results indicate that the formed n-n + heterostructure can more efficiently separate the charge and confirms its superior optoelectronic performance compared with the individual MoS2 case.
At various light intensities, I-V characteristic curves has been simulated for MoS2, TiS3, and TiS3-MoS2 photodetectors. The obtained curves for drain currents are consistent with the measurements as illustrated in Figure S9a-c. Relative differences between the amplitude of photocurrent is in all devices is in total agreement with the experiments where TiS3 and MoS2 PDs have higher and lower photocurrents, respect to the hybrid structure under identical biasing and light exposure conditions.
Table 1 compares the performance of the introduced PDs with some other reported paper-based PDs. According to it, most of them have a photoresponsivity in the range of a few µA/W. In the case of MoS2, it is observed that our introduced PD has a slightly higher photoresponsivity than the others. The TiS3 PD shows a much higher photoresponsivity (in the range of mA/W), which has a significant improvement in performance compared to other reported PDs. However, it shows a larger dark current and a smaller on/off ratio than MoS2 PDs which can limit its optical detection performance. Finally, the photoresponsivity of the TiS3-MoS2 heterostructure shows a significant improvement, which indicates its superiority over other works.
Table 1
Comparison of the performance of introduced paper-based PDs with other PDs based on different 2D materials.
Paper-based PDs
|
Vds
|
λ
|
R
|
ref
|
Te—TiO2
|
0 V
|
400 nm
|
30 µA/W
|
32
|
ZnS-MoS2
|
1 V
|
554 nm
|
18 µA/W
|
33
|
MoTe2
|
1 V
|
532 nm
|
1.1 µA/W
|
34
|
MoSe2
|
1 V
|
532 nm
|
0.2 µA/W
|
WS2
|
1 V
|
532 nm
|
0.06 µA/W
|
MoS2
|
1 V
|
532 nm
|
0.02 µA/W
|
MoS2
|
20 V
|
532 nm
|
1.1 µA/W
|
14
|
MoS2
|
10 V
|
532 nm
|
1.2 µA/w
|
This work
|
TiS3
|
10 V
|
532 nm
|
1600 µA/W
|
TiS3-MoS2
|
10 V
|
532 nm
|
170 µA/W
|