High‐Performance Photodetector based on a 3D Dirac Semimetal Cd3As2/Tungsten Disulfide (WS2) van der Waals Heterojunction

Abstract

Recently, heterojunction photodetectors have attracted significant interest due to the multiple degrees of freedom reorganization, integrating advantages of different typed materials. Herein, a high‐performance photodetector based on a 3D Dirac semimetal Cd₃As₂/tungsten disulfide (WS₂) heterojunction is demonstrated, which is constructed by directly transferring exfoliated 2D few layer WS₂ on Cd₃As₂ nano‐belt and following by annealing treatment. The resulting Cd₃As₂/WS₂ heterojunction device presents superior performance with a high on/off ratio (≈5.3 × 10⁴) and a responsivity (R_i) of about 223.5 AW⁻¹ at 520 nm, as well as an outstanding detectivity (D*) of about 2.05 × 10¹⁴ Jones at 808 nm near‐IR waveband. However, the optimized noise equivalent power (NEP) is evaluated about 6.17 × 10⁻¹⁴ WHz^−1/2 by the noise power density spectrum. The excellent performance can be attributed to a high‐quality heterostructure interface, strong light absorption capacity, and ultralow dark current in a Cd₃As₂/WS₂ heterojunction system. This work provides a promising platform to develop a high‐performance optoelectronic device based on 3D Dirac semimetal and 2D TMDs families.

Introduction

A high-performance photodetector is always a desired pursuit in photodetection field due to the vital application in a sensing network system including optical communication, IR imaging, and the Internet of Things sensor, which of core is efficient optoelectronic functional materials that mainly constructed by some expensive III–V semiconductors in commercial applications at present.^[] Since the discovery of graphene transistors in 2009, other 2D materials such as layered transition metal sulfides (TMDs) and black phosphorene (BP) have triggered an extensive attention due to fascinating physical properties (e.g., adjustable bandgap and strong light interacts with matter), which are particularly attractive in designing an electronic and optoelectronic device.^[] As one of the prominent among abundant 2D TMDs families, tungsten disulfide (WS₂) has an outstanding optical absorption performance with adjustable bandgap (1.3–2.1 eV) that varies depending on its number of layers.^[] The feature of no dangling bonds on surface permits it to be van der Waals integrated with a variety of distinctive materials to improve device performance without considering lattice matching constraint.^[] Although a lot of WS₂-based photodetectors have constantly been investigated based on different photodetection mechanisms including photoconductive,^[] photogating,^[] photovoltaic,^[] and plasmon^[] in the past, the spectral responsivity and detection performance of those photodetectors are far from meeting actuality requirements. There are still much room to strengthen device performance parameters whether in responsivity, detectivity, or detection waveband. Among various promotion strategies mentioned earlier, constructing a van der Waals heterojunction device may be considered an efficient approach to achieve high-property photoelectric detection using complementary performance photosensitive materials.

In recent years, topological quantum materials including Dirac semimetals, Weyl semimetals, and topological insulators have aroused intense research interest in electrical and photonics due to some novel physical phenomena such as 3D quantum anomalous Hall effect.^[] In particular, topological semimetal usually has a unique topological zero gap energy band structure and super high charge mobility to achieve remarkable charge transport, which has aspiring and flourishing application prospects in IR optoelectronic devices.^[] At present, the reports on topological semimetal photodetectors are increasing gradually, bringing new vitality to the further innovation of broad spectra photodetection field.^[] As a representative of Dirac semimetal, cadmium arsenide (Cd₃As₂) is called 3D graphene to possess ultrahigh charge mobility that attributed to suppressed backscattering of high Fermi velocity 3D Dirac fermions. Besides, the existing strong light–matter interaction in Cd₃As₂ means it has excellent optical absorption properties compared to that of a monatomic layer of graphene.^[] These excellent characteristics demonstrate the important research value it has for broad spectra photodetection. For instance, Wang et al. developed the ultrafast broadband photodetectors based on 3D Dirac semimetal Cd₃As₂ nano-plate, which achieved the broadband responses from 532 nm to 10.6 μm.^[] Also, the broad spectra heterojunction photodetectors based on Cd₃As₂/organics thin film were also fabricated and investigated by Yang et al.^[] However, to the best of our knowledge, the van der Waals heterojunction photodetector combined 3D Dirac semimetal Cd₃As₂ with 2D materials has rarely involved and lack of in-depth developed so far, which is expected to effectively integrate the advantages of both to realize high-efficiency photodetection.^[]

Herein, we demonstrated a 3D/2D van der Waals heterojunction photodetector combining Dirac semimetal Cd₃As₂ with few layer WS₂ using a directly precise transfer process. Then, annealing treatment for the Cd₃As₂/WS₂ heterojunction device was carried out to improve interface and contact quality in reducing atmosphere, so that detection performance would be further optimized. The fabricated heterojunction has achieved high-performance photoelectric conversion characters from visible (405 nm) to near-IR region (808 nm) at room temperature, in which the optimal responsivity (R_i) and detectivity (D*) were evaluated about 223.5 AW⁻¹ and 2.05 × 10¹⁴ Jones, respectively. The results indicate that the high performance of Cd₃As₂/WS₂ van der Waals heterojunction may be a great potential in photoelectric detection.

Results and Discussion

Figure  presents the optical microscope (OM) and scanning electron microscopy (SEM) images of the as-prepared Cd₃As₂/WS₂ heterojunction device (the detailed device fabrication process is shown in Experimental Section), respectively. As shown in Figure , the planar Cd₃As₂/WS₂ heterojunction was constructed by that few-layered WS₂ was mechanically exfoliated from corresponding bulk materials and then precisely transferred between 3D Dirac semimetal Cd₃As₂ nano-belt and as-prepared metal electrodes (corresponding flow diagram is revealed in Figure S1, Supporting Information). The overlapping region area of Cd₃As₂ and WS₂ can be further clearly determined from SEM image in Figure , in which the photosensitive area is estimated about 5 μm². Obviously, different color brightness of WS₂ generally represents layer thickness, and the heterostructure region of WS₂ could be considered as few layer according to previous publications.^[] However, three discrete devices can be observed including Cd₃As₂ photoconductive device (port of ➀-➁), Cd₃As₂/WS₂ heterojunction device (port of ➀-➂), and WS₂ flake device (port of ➂-➃), indicating that the device has the potential to switch between different modes. To evidentiary characterize materials constituent in the heterojunction device, energy disperse spectrum (EDS) mapping and Raman spectra were carried out. Figure  displays the EDS distribution of Cd₃As₂ nano-belt located in the yellow area of Figure . Apparently, the elements of Cd and As are evenly distributed, and element percentage of both is ≈3:2, proving existence of Cd₃As₂. For EDS mapping in heterojunction region, the four elements of As, Cd, S, and W can also be observed (Figure S2, Supporting Information), which are originated from Cd₃As₂ and WS₂, respectively. Moreover, Raman spectra in Figure  convincingly confirm a successful construction of the high-quality Cd₃As₂/WS₂ heterostructure device. Specifically, individual WS₂ presents two characteristic peaks at the positions of 354 and 419 cm⁻¹, which corresponds to E¹_2g and A_1g modes caused by in-plane vibration and out-of-plane vibration, respectively,^[] whereas Raman spectrum of heterostructure area selected in blue area of Figure  has two additional distinct characteristic peaks in 195 and 250 cm⁻¹ that can be assigned to Cd₃As₂.^[] All aforementioned results indicate that the heterostructure region was composed of Cd₃As₂ and WS₂.

[IMAGE OMITTED. SEE PDF]

Figure  shows the schematic diagram of the Cd₃As₂/WS₂ heterojunction photodetector on SiO₂/Si⁺⁺ substrate, and the corresponding energy band arrangement is shown in Figure . Theoretically, the Fermi level energy of 3D Dirac semimetal Cd₃As₂ is approximately about −4.5 eV, whereas the valence band (VB) and the conduction band (CB) of WS₂ are about −4.13 and −5.06 eV, respectively.^[] Hence, a directional built-in electric field would be formed at the heterostructure interface of the Cd₃As₂/WS₂ device because Fermi level could reach equilibrium again once the two touch. When the heterojunction device was exposed to light radiation, photogenerated carriers would be generated, separated, transferred, and collected by electrode under collective action of existing built-in electric field (E_in) and applied bias, and the relevant energy band illustration of Cd₃As₂/WS₂ heterojunction at different biases is presented in Figure S3, Supporting Information.

[IMAGE OMITTED. SEE PDF]

To investigate the photoelectric detection characteristic of the Cd₃As₂/WS₂ heterojunction, related electrical and optical–electrical measurements have been carefully carried out. As shown in Figure S4, Supporting Information, a series of current–voltage (I–V) characteristics under dark condition were first analyzed for individual Cd₃As₂, WS₂, and Cd₃As₂/WS₂ heterojunction before/after annealing, respectively. Obviously, the individual Cd₃As₂ nano-belt photoconductive device presented relative high dark current due to its topological semimetallic properties (shown in Figure S4a, Supporting Information). In contrast, the dark current in the Cd₃As₂/WS₂ heterojunction device (shown in Figure S4c,d, Supporting Information) was successfully suppressed after introduction of relatively low charge mobility of 2D WS₂, which may help to improve on/off ratio and detection performance of device. Remarkably, the Cd₃As₂/WS₂ heterojunction (Figure S4d, Supporting Information) has a diode-like rectification characteristic with extremely low dark current with an order magnitude of 10⁻¹¹ A even at a bias voltage of -1 V. Compared with the heterojunction before annealing (Figure S4c, Supporting Information), the dark current especially in case of applying forward bias has a significant improvement for the Cd₃As₂/WS₂ heterojunction after annealing, which may be related to enhanced crystallinity, interface quality, and optimized contact resistance between material and electrode in the heterostructure device.^[] Figure  displays the corresponding I–V curves under dark condition and an irradiation of 520 nm light source with a power density of about 855 μW mm⁻². We can see that the heterojunction device presents the prominent photocurrent response in comparison with negligible dark current. From logarithmic plots of I–V curve in the inset of Figure , high on/off ratio characteristics can be distinctly observed, in which the maximum value could be reach about 4.3 × 10⁶ at 0 V. The high response photocurrent capability may be benefited from the complementary advantages of 3D Dirac semimetal Cd₃As₂ and 2D WS₂. Figure  further shows the photocurrent response curves in time domain of the heterojunction at a bias of ±0.5 V, in which the corresponding on/off ratio is about 1.3 × 10⁴ at a forward bias of +0.5 V and 5.3 × 10⁴ at a reverse bias of −0.5 V, respectively. Relative to that of the individual Cd₃As₂ photoconductive device with a lot of noise (as displayed in Figure S5, Supporting Information), the photoelectric performance of the Cd₃As₂/WS₂ heterojunction has been optimized due to the effective synergistic effect of Cd₃As₂ and WS_2.

Based on the calculated formula of photocurrent (I_ph), responsivity (R_i), and detectivity (D*) as follows, the corresponding performance parameters of the heterojunction were further assessed.1$I_{\text{ph}}=I_{\text{photo}}–I_{\text{dark}}$2$R_{\mathrm{i}}=I_{\text{photo}}/(P_{\text{light}}\times A)$3$D^{*}=R_{\mathrm{i}}/{(A/2e{I}_{\text{dark}})}^{1/2}$where I_ph is the photocurrent, P_light is the incident light power density, and A represents the effective photosensitive area of a device.

Figure  displays a histogram of photocurrent value of the heterojunction device from visible (405 nm) to near-IR waveband (808 nm) driven at ±0.5 V. Specifically, as displayed in Figure , the values of R_i are evaluated at the waveband of 405, 450, 520, 650, and 808 nm for the device, in which the value corresponds to 8.4, 53.9, 135.6, 35.8, and 1.0 A W⁻¹ at a forward bias of 0.5 V and 20.7, 99.5, 223.5, 102.6, and 1.9 A W⁻¹ at an inverse bias of −0.5 V, respectively. It can be concluded that the Cd₃As₂/WS₂ heterojunction has a better responsiveness of nearly 520 nm, and its response bandwidth can be continued to near-IR region.

[IMAGE OMITTED. SEE PDF]

As presented in Figure , dependence of photocurrent (I_p) and responsivity (R_i) on various bias voltage from −0.5 to 0.5 V at an irradiation of 520 nm were further exhibited. The photocurrent almost linearly increases with an increase in the absolute value of bias, suggesting that more carriers are collected by electrode. However, it can be clearly observed that the values of R_i have a positive linear dependence on bias, and the acquired R_i values were quantified about 28.3 (135.6) and 82.8 (223.5) A/W at a forward bias of 0.1 (0.5) V and a backward bias of −0.1 (−0.5) V, respectively. The response time usually reflects the ability of photodetector to track light signals, in which the rise (decay) time is the time taken for the photocurrent to increase from 10% to 90% (drop from 90% to 10%) in normalized curve. As shown in Figure , the rising edge and falling edge at 620 nm were assessed as 15 and 16 ms, respectively, indicating that the heterojunction photodetector has a fast response speed. To investigate the photoelectric response characteristics of near-IR band in detail, Figure  presents a variation trend of photoinduced current with a power density from 14 mW mm⁻² to 0.1 μW mm⁻² of 808 nm at −0.5 and +0.5 V, respectively, which have an approximate linear dependence. An overall increasing trend with an increase in power density can be depicted, because stronger light produces more photogenerated carriers. Moreover, the corresponding R_i and D* are also described in Figure . The maximum values are up to 3.8 A W⁻¹ and 2.05 × 10¹⁴ Jones, respectively, at a near-IR waveband of 808 nm, indicating that the heterojunction has a tremendous potential for high efficiency of photoelectric conversion.

[IMAGE OMITTED. SEE PDF]

To shed light on enhanced photocurrent response mechanism of Cd₃As₂/WS₂ heterojunction from experiment, the photocurrent characteristics of the heterojunction before/after annealing were contrasted and analyzed. As shown in Figure , the photocurrent response capacities before/after annealing treatment are presented under an irradiation at 405 and 808 nm with a forward bias of 0.5 V, respectively. Intuitively, the photocurrent is significantly enhanced after annealing, which also distinguished from the I–V curves for Cd₃As₂/WS₂ before/after annealing treatment (Figure S6a,b, Supporting Information). Apparently, the photocurrent has a significant increase with augment of an optical power density of 808 nm light source, because more photon streams were injected into Cd₃As₂/WS₂. Compared with I–V characteristic of the heterojunction before annealing in Figure S6a, Supporting Information, the photoelectric response capability especially in the case of forward bias has significantly enhanced after annealing (Figure S6b, Supporting Information). Similar situations also occur in other wavebands, which is shown in Figure . The photocurrent value has about 3–5 times enhancement after annealing at different wavebands. Concretely, the values of R_i have increased to some extent from the 5.7, 34.1, 86.1, 31.9, and 0.3 A W⁻¹ before annealing to 20.7, 99.5, 223.5, 102.6, and 1.9 A W⁻¹ after annealing at 405, 450, 520, 650, and 808 nm, respectively. Figure  presents the noise power density as a function of frequency of the Cd₃As₂/WS₂ heterojunction before and after annealing treatment, in which the noise has been obviously optimized for Cd₃As₂/WS₂ heterojunction after annealing. According to the obtained value (1.91 × 10⁻²² A² Hz) from noise power spectrum at a reverse bias of −0.5 V and the frequency of 1 Hz, the NEP could be evaluated as 6.17 × 10⁻¹⁴ WHz^−1/2.^[]

[IMAGE OMITTED. SEE PDF]

It is worth noting that the surface physical adsorbates (such as H₂O and O₂ molecules) are always prone to degradation of device performance.^[] After annealing process in a reducing atmosphere, the material interface may be closer contact and passivated, which are beneficial to reduce the adsorption of O₂ and H₂O on the surface of WS₂ and increase charge transfer efficiency.^[] Moreover, due to excellent light absorption capacity and ultrahigh charge mobility, 3D Dirac semimetal Cd₃As₂ plays important role in generating photogenerated charges and charge transport, whereas the combination of WS₂ with Cd₃As₂ further reduces dark current and improves detectability. Therefore, the device shows more excellent photodetection performance.

As shown in Table , some representative optoelectronic performance indicators including R_i, D*, response time, and On/off for our Cd₃As₂/WS₂ heterojunction and some photodetectors based on 2D materials/topological quantum materials are listed. Also, further effective efforts could be made in optimizing performance to realize a photodetector array combining 3D Dirac semimetal and 2D material to realize high-response imaging function in future.

Table 1 Comparison of photodetectors performance based on 2D materials or topological quantum material

NO.	Devices	R i [A/W]	D* [Jones]	Time [ms]	On/off	Bias [mV]	Wavelength [nm]	Ref.
1	MoTe2/WS2	0.14	NA	0.285/0.186	103	1000	825	[]
2	Cd3As2 nano-plate	5.9 × 10−3	NA	3.9 ps (intrinsic)	NA	10	633	[]
3	TaAs	0.7 × 10−3	1.68 × 108	NA	NA	0.1	438.5	[]
4	Bi2Se3 NWs	300	7.5 × 109	550/400	1.0013	150	1064	[]
5	Sb2Te3/MoS2	330	1012	0.36/0.47	NA	−1000	520	[]
6	SnTe/Bi2Se3	0.146	1.15 × 1010	6.9/19.2	6.52	−5000	1550	[]
7	Bi2Te3 film/WS2	30.7	2.3 × 1011	20/20	NA	3000	1550	[]
8	Cd3As2/WS2	223.5	2.1 × 1014	15/16	5.3 × 104	−500	520	This work

Conclusion

In summary, the high-performance photodetector based on 3D Dirac semimetal Cd₃As₂/WS₂ van der Waals heterojunction was designed and fabricated by transferring few-layer WS₂ nano-sheet to Cd₃As₂ nano-belt and following by annealing in high-purity Ar atmosphere. That process is conducive to establish intimate contact and reduce physical adsorbates (such as H₂O and O₂ molecules in air) on surface of WS₂. Consequentially, the photoelectric response performance of the heterojunction device has been significantly improved benefiting from suppressed dark current and excellent light absorption abilities. Hence, the superior performance indicators including a high responsivity of 223.5 AW⁻¹, an excellent detectivity of 2.1 × 10¹⁴ Jones, and a fast response speed of ≈15 ms were achieved. This work offers a paradigm for fabricating the van der Waals heterojunction photodetector combining 3D Dirac semimetals and 2D layered materials to develop next-generation high-performance and low-cost photodetectors.

Experimental Section

Material Growth and Preparation

The Cd₃As₂ nano-belts were grown by depositing high-purity Cd₃As₂ powers in argon atmosphere using a chemical vapor deposition equipment, which is similar to previous reports.^[] Specifically, as shown in Figure S1, Supporting Information, the precursors of Cd₃As₂ powders were first placed in heating center of a horizontal tube furnace, and dull-polished quartz substrates were placed downstream, in which Argon was a carrier gas. To remove air, the tube furnace was first pumped and flushed with Ar to complete the air washing process. At the most important stage of crystal growth, the tube furnace was heated to 760 °C in 15 min and held constant for 40 min until cooled naturally to room temperature with a constant flow of argon. Finally, the micro-nano-sized Cd₃As₂ was obtained on the quartz substrate.

The few-layered WS₂ nano-sheets were prepared by the standard mechanical exfoliation process from the corresponding WS₂ bulk material (that purchased from Nanjing MuKe Nanotechnology Co., Ltd.) using polydimethylsiloxane and Stoke tape, in which the thickness of acquired WS₂ nano-sheets was determined by means of Raman spectra and sample color depth under an OM.

Device Fabrication

For the fabricated Cd₃As₂/WS₂ heterojunction device, previously acquired appropriate size 3D Dirac semimetal Cd₃As₂ nano-belt was first directly transferred to SiO₂/Si substrate with specific alignment mark. After spin-coating methyl methacrylate/polymethyl methacrylate double layer photoresist, a standard electron-beam lithography (EBL) process was conducted to pattern electrode shape, and Ag/Au (60/100 nm) was thermal evaporated as metal electrode, in which metal electrode was only connected to one end of Cd₃As₂ nano-plates for next step to transfer stripped WS₂. By a lift-off process in acetone, semifinished Cd₃As₂ nano-belt devices were obtained. Subsequently, with an aid of a 2D material transfer platform, the exfoliated few-layered WS₂ nano-sheets could be precisely transferred as channel connecting Cd₃As₂ and metal electrodes. To further improve interface contact quality, an annealing process was further taken for the Cd₃As₂/WS₂ heterojunction in high-purity Ar atmosphere.

Device Performance Characterization

The SEM images and Raman spectra of Cd₃As₂/WS₂ heterojunction were collected with a thermal field-emission scanning electron microscope (EM-8100F) equipped with the EDS analyzer and a confocal microprobe Raman spectrometer with a helium–neon laser at a wavelength of 514.5 nm, respectively. The photoelectric response performance of the Cd₃As₂/WS₂ heterojunction was measured by a Keithley 2636B source meter system (Tektronix Inc.) with various waveband commercialized small semiconductor lasers. The response rate and noise power density of the device were measured by a PDA semiconductor parameter analyzer.

Acknowledgements

This work was supported by Outstanding Youth Foundation of National Natural Science Foundation of China (No. 61922022), Foundation for Innovative Research Groups of the National Natural Science Foundation of China (Grant No. 61421002), and National Natural Science Foundation of China (No. 61875031).

Conflict of Interest

The authors declare no conflict of interest.

Data Availability Statement

Research data are not shared.

References

R. Soref, Nat. Photonics 2010, 4, 495.

E. Delli, V. Letka, P. D. Hodgson, E. Repiso, J. P. Hayton, A. P. Craig, Q. Lu, R. Beanland, A. Krier, A. R. J. Marshall, P. J. Carrington, ACS Photonics 2019, 6, 538.

S. Mauthe, Y. Baumgartner, M. Sousa, Q. Ding, M. D. Rossell, A. Schenk, L. Czornomaz, I. E. Moselund, Nat. Commun. 2020, 11, 1.

J. Wang, J. Y. Han, X. Q. Chen, X. R. Wang, InfoMat 2019, 1, 33.

T. Dong, J. Simões, Z. C. Yang, Adv. Mater. Interfaces 2020, 7, 1901657.

F. Wang, Z. X. Wang, L. Yin, R. Q. Cheng, J. J. Wang, Y. Wen, T. A. Shifa, F. M. Wang, Y. Zhang, X. Y. He, J. Zhan, Soc. Rev. 2018, 47, 629.

J. S. Miao, C. Wang, Nano Res. 2020, 1.

H. S. Ra, M. H. Jeong, T. Yoon, S. Kim, Y. J. Song, J. S. Lee, Adv. Sci. 2020, 7, 2001475.

W. Jiang, T. Zheng, B. Wu, H. X. Jiao, X. D. Wang, Y. Chen, X. Y. Zhang, M. Peng, H. L. Wang, T. Lin, H. Shen, J. Ge, W. D. Hu, X. F. Xu, X. J. Meng, J. H. Chu, J. L. Wang, Light: Sci. Appl. 2020, 9, 160.

H. J. Tan, Y. Fan, Y. Q. Zhou, Q. Chen, W. S. Xu, J. H. Warner, ACS Nano 2016, 10, 7866.

L. Xie, M. Z. Liao, S. P. Wang, H. Yu, L. J. Du, J. Tang, J. Zhao, J. Zhang, P. Chen, X. B. Lu, G. L. Wang, G. B. Xie, R. Yang, D. X. Shi, G. Y. Zhang, Adv. Mater. 2017, 29, 1702522.

T. H. Tsai, Z. Y. Liang, Y. C. Lin, C. C. Wang, K. I. Lin, K. Suenaga, P. W. Chiu, ACS Nano 2020, 14, 4559.

Q. Xu, Z. Yang, D. F. Peng, J. G. Xi, P. Lin, Y. Heng, H. Liu, C. F. Pan, Nano Energy 2019, 65, 104001.

Y. Z. Xue, Y. P. Zhang, Y. Liu, H. T. Liu, J. C. Song, J. Sophia, J. Y. Liu, Z. Q. Xu, Q. Y. Xu, Z. Y. Wang, J. L. Zheng, Y. Q. Liu, S. J. Li, Q. L. Bao, ACS Nano 2016, 10, 573.

Q. S. Lv, F. G. Yan, X. Wei, K. Y. Wang, Adv. Opt. Mater. 2018, 6, 1700490.

W. K. Zhu, F. G. Yan, X. Wei, Q. S. Lv, H. Yang, K. Y. Wang, Adv. Mater. Lett. 2019, 10, 329.

G. C. Wang, L. Li, W. H. Fan, R. Y. Wang, S. S. Zhou, J. T. Lü, L. Gan, T. Y. Zhai, Adv. Funct. Mater. 2018, 28, 1800339.

P. Z. Liu, J. R. Williams, J. J. Cha, Nat. Rev. Mater. 2019, 4, 479.

C. Zhang, Y. Zhang, X. Yuan, S. H. Lu, J. L. Zhang, A. Narayan, Y. W. Liu, H. Q. Zhang, Z. L. Ni, R. Liu, E. S. Choi, A. Suslov, S. Sanvito, P. Li, H. Z. Lu, A. C. Potter, F. X. Xiu, Nature 2019, 565, 331.

A. Q. Wang, X. G. Ye, D. P. Yu, Z. M. Liao, ACS Nano 2020, 14, 3755.

J. Liu, F. N. Xia, D. Xiao, F. J. Abajo, D. Sun, Nat. Mater. 2020, 19, 830.

H. T. Chorsi, S. Y. Yue, P. P. Lyer, M. Goyal, T. Schumann, S. Stemmer, B. L. Liao, J. A. Schuller, Adv. Opt. Mater. 2020, 8, 1901192.

Q. S. Wang, C. Z. Li, S. Ge, J. G. Li, W. Lu, J. W. Lai, X. F. Liu, J. C. Ma, D. P. Yu, Z. M. Liao, D. Sun, Nano Lett. 2017, 17, 834.

M. Yang, J. Wang, Y. K. Yang, Q. Zhang, C. Ji, G. Wu, Y. Su, J. Gou, Z. Wu, K. Yuan, F. Xiu, J. Phys. Chem. Lett. 2019, 10, 3914.

Z. H. Huang, Y. D. Jiang, Q. Han, M. Yang, J. Y. Han, F. Wang, M. Luo, Q. Li, H. Zhu, X. C. Liu, J. Gou, J. Wang, Nanotechnol. 2019, 31, 064001.

B. Ayse, R. G. Humberto, R. B. Andres, P. Nestor, L. E. Ana, C. I. Chia, B. Wang, H. C. Vincent, L. U. Florentino, J. C. Charlier, H. Terrones, M. Terrones, Sci. Rep. 2013, 3, 1755.

P. K. Nayak, C. H. Yeh, Y. C. Chen, P. W. Chiu, ACS Appl. Mater. Interfaces 2014, 6, 16020.

Y. Zhao, S. Y. Han, J. Zhang, L. M. Tong, J Raman Spectrosc. 2020, 1.

K. Zhang, H. Y. Pan, M. H. Zhang, Z. X. Wei, M. G. Song, X. F. Wang, R. Zhan, RSC Adv. 2017, 7, 17689.

Y. Z. Xue, Y. P. Zhang, Y. Liu, H. T. Liu, J. C. Song, J. Sophia, J. Y. Liu, Z. Q. Xu, Q. Y. Xu, Z. Y. Wang, J. L. Zheng, Y. Q. Liu, S. J. Li, Q. L. Bao, ACS Nano 2016, 10, 573.

A. Islam, J. Lee, P. X. L. Feng, J. Appl. Phys. 2018, 123, 025701.

J. Garnier, R. Parize, E. Appert, O. Chaix-Pluchery, A. Kaminski-Cachopo, V. Consonni, ACS Appl. Mater. Interfaces 2015, 7, 5820.

G. H. Shin, J. G. Park, K. J. Lee, G. B. Lee, H. B. Jeon, Y. K. Choi, K. Yu, S. Y. Choi, ACS Appl. Mater. Interfaces 2019, 11, 7626.

S. Hao, S. N. Yan, Y. Wang, T. Xu, H. Zhang, X. Cong, L. F. Li, X. W. Liu, T. J. Cao, A. Y. Gao, L. L. Zhang, L. X. Jia, M. S. Long, W. D. Hu, X. M. Wang, P. H. Tan, L. T. Sun, X. Y. Cui, S. J. Liang, F. Miao, Small 2020, 16, 1905902.

C. Y. Lan, C. Li, Y. Yin, Y. Liu, Nanoscale 2015, 7, 5974.

C. Zhang, A. Narayan, S. Lu, J. L. Zhang, H. Q. Zhang, Z. L. Ni, Y. W. Liu, J. H. Park, E. Z. Zhang, W. Y. Wang, S. S. Liu, L. Cheng, L. Pi, Z. G. Sheng, S. Sanvito, F. X. Xiu, Nat. Commun. 2017, 8, 1.

E. Z. Zhang, Y. W. Liu, W. Y. Wang, C. Zhang, P. Zhou, Z. G. Chen, J. Zou, F. X. Xiu, ACS Nano 2015, 9, 8843.

S. M. Chi, Z. L. Li, Y. Xie, Y. G. Zhao, Z. Y. Wang, L. Li, H. H. Yu, G. Wang, H. M. Weng, H. J. Zhang, J. Y. Wang, Adv. Mater. 2018, 30, 1801372.

A. Sharma, B. Bhattacharyya, A. K. Srivastava, T. D. Senguttuvan, S. Husale, Sci. Rep. 2016, 6, 19138.

H. W. Liu, D. Li, C. Ma, X. H. Zhang, X. X. Sun, C. G. Zhu, B. Y. Zheng, Z. X. Zou, Z. Y. Luo, X. L. Zhu, X. Wang, A. L. Pan, Nano Energy 2019, 59, 66.

H. B. Zhang, Z. L. Song, D. Li, Y. C. Xu, J. Li, C. J. Bai, B. Y. Man, Appl. Surf. Sci. 2020, 509, 145290.

J. D. Yao, Z. Q. Zhang, G. W. Yang, J. Mater. Chem. C 2016, 4, 7831.