Ultrasensitive MoS2 photodetector by serial nano-bridge multi-heterojunction

The recent reports of various photodetectors based on molybdenum disulfide (MoS2) field effect transistors showed that it was difficult to obtain optoelectronic performances in the broad detection range [visible–infrared (IR)] applicable to various fields. Here, by forming a mono-/multi-layer nano-bridge multi-heterojunction structure (more than > 300 junctions with 25 nm intervals) through the selective layer control of multi-layer MoS2, a photodetector with ultrasensitive optoelectronic performances in a broad spectral range (photoresponsivity of 2.67 × 106 A/W at λ = 520 nm and 1.65 × 104 A/W at λ = 1064 nm) superior to the previously reported MoS2-based photodetectors could be successfully fabricated. The nano-bridge multi-heterojunction is believed to be an important device technology that can be applied to broadband light sensing, highly sensitive fluorescence imaging, ultrasensitive biomedical diagnostics, and ultrafast optoelectronic integrated circuits through the formation of a nanoscale serial multi-heterojunction, just by adding a selective layer control process.


Supplementary Note 2: X-ray photoelectron spectroscopic (XPS) analysis of MoS 2 surfaces
Supplementary Figure 2a shows no change in the XPS data of Mo 3d and S 2p between the pristine 6L-MoS 2 and the 6L-MoS 2 after 5 th cycle ALE. However, Mo 6+ was observed at ~236 eV after an O 2 plasma exposure (120 s at 6.7 mTorr and 70 sccm of O 2 and 200 W of a 13.56 MHz ICP power) on the pristine 6L-MoS 2 surface. The Mo 6+ peak appears when the MoS 2 surface is oxidized and indicates the changes of the MoS 2 intrinsic properties 1,2 . Supplementary Figure 2b shows XPS data of Cl 2p when one cycle ALE process is performed on the pristine 6L-MoS 2 . Cl peaks (2p 1/2 and 2p 3/2 ) were completely removed after Ar + -ion beam desorption (2 nd step) was performed on Cl-adsorbed MoS 2 , indicating no etch residue on the MoS 2 surface after the ALE process. Supplementary Figure 2c shows the comparison of the S/Mo ratio. The S/Mo ratios of pristine 6L-MoS 2 , after 5 th cycle ALE and after the O 2 plasma exposure were 2.023, 2.012 and 1.547, respectively, indicating that the S/Mo ratio of the MoS 2 layer is preserved even after the 5 th cycle ALE while the ratio is changed after the oxygen plasma exposure. Supplementary Figure 2d shows the change of the XPS binding energy positions of the Mo peaks (3d 3/2 and 3d 5/2 ) and S peaks (2p 1/2 and 2p 3/2 ). The XPS binding energy positions of pristine 6L-MoS 2 and that after 5 th cycle ALE were all the same, but after the O 2 plasma exposure, both Mo and S peaks were red-shifted. In general, the red shift is observed when the MoS 2 surface is oxidized and damaged 2 . Figure 2. XPS analysis of MoS 2 surface characteristics before/after the MoS 2 ALE or an O 2 plasma exposure. a, XPS data of Mo 3d and S 2p for pristine 6L-MoS 2 , after 5 th cycle ALE, and after an O 2 plasma exposure. b, XPS data of Cl 2p after Cl radical adsorption and after Ar + -ion desorption. c, Change of the S/Mo ratio for pristine 6L-MoS 2 , after 5 th cycle ALE, and after an O 2 plasma exposure. d, Change of the XPS binding energy positions of the Mo peaks (3d 3/2 and 3d 5/2 ) and S peaks (2p 1/2 and 2p 3/2 ) for pristine 6L-MoS 2 , after 5 th cycle ALE, and after O 2 plasma exposure. O 2 plasma was generated for 120 s at 6.7 mTorr/70 sccm of O 2 and 200 W of a 13.56 MHz ICP power.

Supplementary Note 3: Properties of MoS 2 layers prepared by different methods
In case of the pristine 6L-MoS 2 , an indirect bandgap of 1.394 eV was obtained by PL measurement at room temperature as shown in Supplementary Figure 3a

Supplementary Note 5: Photoresponse time extraction
The photoresponse time (rise and decay) characteristics of the photodetector were analyzed in a laser on/off cycle (20 s of laser on-state and 20 s of laser off-state), and the maximum photocurrents (I max ) were normalized as 1.0 for the accurate comparison of Type (1)~(6). The rise time and decay time were extracted from (the time between I 10% to I 90% ) and (the time between I 90% to I 10% ) of the measured I max . Therefore, for Type (1)

Supplementary Note 6: Electrical characterization
Here, the electrical properties of Type (5)-parallel nano-bridge multi-heterojunction were remarkably degraded and were similar to Type (1)-mono-layer possibly due to the series of energy barriers between the source and drain which block the carrier transport. On the other hand, in Type (6)-serial nano-bridge multi-heterojunction, it is possible to maximize the optoelectronic performances without changing electrical characteristics by showing similar electrical characteristics as Type (2)-multi-layer due to no energy barrier such as Type (5) Figure 7c) and laser on-state (Supplementary Figure 7d), the carrier transit time was decreased with increasing V d , and the effective barrier height was decreased with increasing V g . Therefore, a gradual increase in I d is confirmed and acceptable contact quality is also expected 6 . In addition, the reason why the change of  Figure 7. Characteristics of hysteresis, output curve (for laser on-/off-state), I d -V g (as function of V d ), and photoresponsivity (as function of V g ). The hysteresis comparison between a, Type (2) and b, Type (6). The output curve of Type (6) according to laser c, off-state and d, on-state. I d -V g characteristics of Type (6) as function of V d e, (0.1 V), f, (2 V), and g, (5 V). h, Photoresponsivity of Type (1)~(6) extracted as a function of V g (-30 ~ +30 V).

Supplementary Note 8: Optical and Raman mapping images
In order to observe the optoelectronic performance according to the number of parallel-type and serial-type mono-/multi-layer (6L) MoS 2 heterojunctions, 6L-MoS 2 channel was patterned with PR (width of 1 μm) at 2 μm intervals through a photolithographic process and 5 MoS 2 layers were selectively removed by 5 ALE cycles. Optical and Raman mapping images showed that the multiheterojunctions with the mono-/multi-layer MoS 2 structure were uniformly formed on the ~10 µm

Supplementary Note 11: Noise power density of Type (6) photodetector
In order to calculate the detectivity (D*), we measured the noise power density of the Type (6) serial nano-bridge multi-heterojunction photodetector (Supplementary Figure 11). The noise power density values of the Type (6) photodetector was extracted at 1 Hz. Detectivity (D*) was calculated from the D* = (AB) 1/2 /NEP = R(A) 1/2 /S n , where, R is the photoresponsivity, A is the effective area of the photodetector, and S n is the noise spectral density. The measured D* were 1.93 × 10 8 jones at V g = 0 V and 2.38 × 10 13 jones at V g = -25 V and these were compared with previously reported photodetectors (Supplementary Data 2). Figure 11. Noise spectra of Type (6) serial nano-bridge multi-heterojunction photodetector. Here, the data was measured at V d = 5 V, V g = 0 and -25 V, and fitted 1/f at two gate voltages.