Fig. 1. (a) Working principle of plasmon-enhanced photodetector. There are three primary processes responsible for the eventual migration of electrons to the conduction band of a semiconductor, including I, II, III. (b) Experimental and simulated reflection and transmission spectra of gold disk arrays with a diameter of 1.5 μm, a period of 2.5 μm, and a thickness of 40 nm. (c) Absorption spectra of pristine WS2 (black line), gold disk arrays (yellow line), WS2 (on the gold disk array) (blue line), and the assembly of WS2 and underlying Au disk array (red line). The blue line represents the absorption of the WS2 sheet that is on the gold disk array, while the red line represents that of the assembly of the WS2 sheet and the gold disk array under it. (d) Electric field distributions of interface between WS2 and gold disks at the wavelengths of 405 nm, 532 nm, 635 nm, 808 nm, 1060 nm, and 1310 nm. The minimum and maximum values in the color bar represent the minimum and maximum values of the respective electric fields at each wavelength.

Fig. 2. (a) Schematic diagram of the fabrication processes of the plasmonic 2D photodetector with five parts: (I) the prefabrication of plasmonic disk array, using direct writing photolithography for patterning and electron beam evaporation for gold deposition, (II) wet transfer of a monolayer h-BN, (III) dry transfer of a few-layer WS2 nanosheet, (IV) a thermal treatment of 90°C for 5 min, and (V) the fabrication of electrodes. (b) An optical micrograph of the fabricated plasmonic 2D photodetector. The white dashed line indicates the position of WS2, and the yellow dashed line portrays gold electrodes while the blue circle represents the position of the underlying plasmonic disks. The WS2 sheet is about 30 nm in thickness. (c) AFM morphology of the plasmonic disk array with a thickness of about 40 nm. (d) AFM 3D image of WS2 on the plasmonic disk array.

Fig. 3. (a) PL spectra of the WS2 photodetectors with and without plasmonic disks under 532 nm laser excitation, revealing a difference of 23 nm in peak wavelengths. (b) Spatial distribution of the PL intensities of WS2 on plasmonic disks under 532 nm laser excitation. (c) I−V characteristic curves of the WS2 photodetectors with (red curves) and without (black curves) plasmonic disks at a variety of intensities ranging from 0 to 109.7  mW/cm2 under 635 nm illumination. (d) Responsivity of the WS2 photodetectors with and without plasmonic disks as a function of light power intensity at a bias voltage of 1 V under 635 nm illumination. (e) Detectivity of the WS2 photodetectors with and without plasmonic disks as a function of light power intensity at a bias voltage of 1 V under 635 nm illumination. (f) Comparison between the responsivity and detectivity of our work and the existing 2D photodetectors working with plasmonic nanostructures.

Fig. 4. (a) I−V characteristics of the WS2 photodetectors with and without plasmonic disks in the dark and at a fixed light intensity under 405–1060 nm illumination. The dashed line represents the dark current of the two devices, while the solid line represents the photocurrents at a fixed light intensity. (b) I−V curves of the WS2 photodetectors with plasmonic disks under 1060 nm illumination with light intensity from 0 to 300.75  mW/cm2. (c) Reproducible on−off switching under 808 nm illumination (with a light intensity of 184.71  mW/cm2) at zero bias. (d) Normalized transient response of the WS2 photodetectors with and without plasmonic disks at a bias voltage of 1 V under 405–1310 nm illumination. (e) An enlarged transient photoresponse cycle at a bias voltage of 1 V under 635 nm excitation with 109  mW/cm2 intensity.

Fig. 5. (a) Top and side views of WS2. (b) Band structure of WS2 with 0.3% strain and without applied strain. (c) Connection among the conduction band (yellow line), valence band (blue line), and Fermi level (red line) of WS2 and tensile strain. (d) Raman spectra of the WS2 photodetectors with and without plasmonic disks under 532 nm laser excitation.

Fig. 6. (a)–(f) Electric field distributions of interface between WS2 and gold disks under the wavelengths of 405 nm, 532 nm, 635 nm, 808 nm, 1060 nm, and 1310 nm light illumination. The maximum value of the color bar is the maximum value of their respective electric field at each wavelength. The minimum value of the color bar is set to 0.8 (E/E0). The positions with relative electric fields <0.8 are given the same color as 0.8. The value of 0.8 corresponds to the electric fields at the surface of a blank silica substrate without other structures under plain wave incidence. Under such settings, the colors differing from the color of 0.8 indicate enhancements when compared with a blank substrate. Thus, these figures can indicate the spatial areas where electric fields are enhanced.

Fig. 7. Comparison of Raman spectra of the WS2 sheet before and after the thermal treatment, which shows that the treatment at 90°C for 5 min has no observable influence on the WS2 sheet.

Fig. 8. (a) Optical microscope image of pristine WS2 device. (b) AFM morphology of WS2 material in the pristine WS2 device, the thickness of which is 33 nm. (c) AFM image of WS2 material in the plasmon-enhanced 2D photodetector, the thickness of which is 30 nm. (d) SEM image of gold disk arrays, displaying a uniform diameter.

Fig. 9. (a)–(e) Mapping images of responsivity of the plasmon-enhanced 2D photodetector at different bias voltages and incident power densities under 405–1060 nm illumination. The maximum value of responsivity can reach 242 A/W.

Fig. 10. (a)–(e) Mapping images of detectivity of the plasmon-enhanced 2D photodetector at different bias voltages and incident power densities under 405–1060 nm illumination. The peak values at low power intensity and high positive bias can be achieved. This reflects the excellent ability of the plasmon-enhanced 2D photodetector for detecting low light. The maximum value of detectivity can reach 3.9×1014 Jones.

Fig. 11. (a)–(c) I−V characteristic curves of the pristine WS2 (black dashed line) and the plasmon-enhanced 2D photodetector (color line) at a variety of intensities under 405–808 nm illumination. (d)–(f) Responsivity of the pristine WS2 and the plasmon-enhanced 2D photodetector as a function of light power intensity at a bias voltage of 1 V under 405–808 nm illumination. The responsivity of the plasmon-enhanced 2D photodetector was up to 307 folds higher than that of the pristine WS2. (g)–(i) Detectivity of the pristine WS2 and the plasmon-enhanced 2D photodetector as a function of light power intensity at a bias voltage of 1 V under 405–808 nm illumination. The detectivity of the plasmon-enhanced 2D photodetector was up to 297 folds higher than that of the pristine WS2.

Fig. 12. Responsivity and detectivity of the plasmon-enhanced 2D photodetector as a function of light power intensity at a bias voltage of 1 V under 1060 nm illumination. As for the I−V characteristic curve of the plasmon-enhanced 2D photodetector under 1060 nm illumination, it has been shown in the paper and the pristine WS2 device has no response at this wavelength.

Fig. 13. Photoswitching behavior of the plasmon-enhanced 2D photodetector at a bias voltage of 1 V under 635 nm illumination with a power density of 324.84  mW/cm2. Encouragingly, the plasmon-enhanced 2D photodetector has an excellent switching behavior, demonstrating remarkable repeatability as well as stability in operation.

Fig. 14. Normalized photocurrent versus time photoswitch characteristic curves of the pristine WS2 and the plasmon-enhanced 2D photodetectors at different wavelengths of 405–1310 nm. The response time (rise time and fall time) of the plasmon-enhanced 2D photodetector is slightly faster than that of pristine WS2 under 405–808 nm illumination. The plasmon-enhanced 2D device has a significant response while the pristine WS2 device has no response at all under 1060–1310 nm illumination, confirming that the plasmon-enhanced 2D device is capable of detecting visible to near infrared light.

Fig. 15. Bandgap change of WS2 under different strains. It quantifies the bandgap reduction of 40 meV under an average 0.3% tensile strain. Narrowing the bandgap enlarges the absorption cutoff wavelength, broadening the detection spectrum for the plasmon-enhanced 2D photodetector.

Fig. 16. Relative shift of E2g1 peak of the WS2 around a plasmonic disk under 532 nm laser excitation. The plotted data are got from the spatial distribution of the E2g1 peak Raman shifts, and the values are corresponding to the differences between the E2g1 peaks’ wave numbers at each point and the minimum wave number of the E2g1 peaks in the measurement. As indicated by the distribution of the relative shift of E2g1 peak and the profile of the plasmonic disk, the induced strain is strongly correlated with the pattern of the Au disk.