Few-layer hexagonal bismuth telluride (Bi2Te3) nanoplates with high-performance UV-Vis photodetection

It is widely known that the excellent intrinsic electronic and optoelectronic advantages of bismuthene and tellurene make them attractive for applications in transistors and logic and optoelectronic devices. However, their poor optoelectronic performances, such as photocurrent density and photoresponsivity, under ambient conditions severely hinder their practical application. To satisfy the demand of high-performance optoelectronic devices and topological insulators, bismuth telluride nanoplates (Bi2Te3 NPs) with different sizes, successfully synthesized by a solvothermal approach have been, for the first time, employed to fabricate a working electrode for photoelectrochemical (PEC)-type photodetection. It is demonstrated that the as-prepared Bi2Te3 NP-based photodetectors exhibit remarkably improved photocurrent density, enhanced photoresponsivity, and faster response time and recovery time in the UV-Vis region, compared to bismuthene and tellurene-based photodetectors. Additionally, the PEC stability measurements show that Bi2Te3 NPs have a comparable long-term stability for on/off switching behaviour for the bismuthene and tellurene-based photodetectors. Therefore, it is anticipated that the present work can provide fundamental acknowledgement of the optoelectronic performance of a PEC-type Bi2Te3 NP-based photodetector, shedding light on new designs of high-performance topological insulator-based optoelectronic devices.


Introduction
The individual elements tellurium (Te) and bismuth (Bi) have been widely applied to the fabrication and applications of photodetectors over the past decade due to their simple composition, excellent nonlinear photonic performance, 1-4 and intriguing thermoelectric [5][6][7] and photoelectric properties. [8][9][10] Similar to the famous black phosphorus (BP), both Te and Bi have a layer-dependent energy band gap (E g ) 1,3,8 that can be easily tuned (Te: 0.35-1.0 eV; 1 Bi: 0-0.55 eV (ref. 3)) when the thickness of Te or Bi decreases from bulk to monolayer. Recently, Ye et al. reported that ultrathin 2D nonlayered Te nanosheets synthesized by a substrate-free solution process, displayed high on/off ratios (10 6 ), remarkable eld-effect mobility (700 cm 2 V À1 s À1 ) and comparable air-stable performance. 8 Besides, the ultrathin Te nanosheets fabricated by a simple liquid phase exfoliation (LPE) method, showed excellent photoresponse behaviors from the UV to the visible region in association with strong time and cycle stability for the on/off switching behaviors. 10 In 2017, Zhang et al. reported that ultrasmall Bi quantum dots fabricated by a LPE approach, exhibited good photoresponse performance from the UV to visible region as well as long-term photoresponse stability. 11 Nevertheless, the poor photoresponse performances of Te or Bi nanomaterials, especially photoelectrochemical (PEC) photocurrent density and photoreponsivity, still severely limit their device development for practical applications. Therefore, it is still a challenge to explore a new method to improve their photoresponse performance under ambient conditions. However, bismuth telluride (Bi 2 Te 3 ), basically known as a compound of the post-transition metal element Bi and the non-metal element Te, also exhibits a thickness-dependent E g (from 0.16 eV to 1.36 eV), 12,13 and a high structural stability. 14,15 Versatile strategies have already been employed to synthesize Bi 2 Te 3 nanomaterials, including template synthesis, 16 evaporation, 17 electrochemical deposition, 18 chemical solution process, 19,20 solvothermal approaches 21,22 and microwave-assisted methods. 23 Bi 2 Te 3 as one of the common topological insulators, features an unconventional phase of quantum matter possessing an insulating bulk state as well as a metallic surface state. 17 Such metallic surface states were experimentally evidenced to be protected by time-reversal symmetry and demonstrated to be robust against non-magnetic perturbation. 24,25 In addition, topological insulator Bi 2 Te 3 exhibits an excellent surface mobility 26 and good optoelectronic performance. 27 This, combined with the relatively narrow E g of Bi 2 Te 3 and low cost and facile synthesis of Bi 2 Te 3 nanomaterials, has drawn great interest in photodetection, [28][29][30] eld effect transistors, 26,31 spintronics, 32,33 thermoelectrics, 22,34 and lasers. [35][36][37] These advantages of Bi 2 Te 3 merit it to be qualied for the practical application in high-performance optoelectronic devices.
In this work, Bi 2 Te 3 nanoplates (NPs) with a rhombohedral phase in the space group D 5 3d (R3m), have been successfully synthesized by a solvothermal approach. To determine the sizedependent PEC performances of Bi 2 Te 3 NPs, different sizes of Bi 2 Te 3 NPs were readily obtained by simply tuning the reaction time. The as-synthesized Bi 2 Te 3 NPs were, for the rst time, developed as working materials to fabricate a PEC-type photodetector in various electrolytes. The PEC results demonstrate that the Bi 2 Te 3 NP-based photodetectors exhibit not only a largely improved photocurrent density and photoresponsivity, but also a comparable photoresponse stability compared to that of Bi or Te nanomaterial-based photodetectors. It is anticipated that this work can provide fundamental guidance for constructing high-performance Bi 2 Te 3 NP-based photodetectors, paving the way to new designs of topological insulator-based optoelectronic devices with excellent properties.

Synthesis of Bi 2 Te 3 hexagonal NPs
In a typical procedure, 20 mmol Bi(NO 3 ) 3 $5H 2 O and 40 mmol Na 2 TeO 3 were rst dissolved in 30 mL DI water. Then 0.03 mmol PVP was added into the solution and it was kept stirring for 30 min to form a homogeneous mixture. The mixture was transferred into a 50 mL Teon-lined autoclave and placed in an oven at 180 C. Aer a predetermined reaction time (2 h or 12 h), the reaction was stopped by quenching the system to room temperature. The Bi 2 Te 3 hexagonal NPs were obtained by centrifugation at 4000 rpm for 20 min and washed with deionized water, ethanol and acetone, each. The product was nally dried in a vacuum oven at 80 C overnight for the next use.

Characterization
The morphologies and dimensions of Bi 2 Te 3 NPs were determined by both scanning electron microscopy (SEM, Hitachi-SU8010) and transmission electron microscopy (TEM, FEI Tecnai G2 F30). High-resolution TEM (HRTEM) was performed to determine the atomic arrangements of the as-synthesized Bi 2 Te 3 NPs. Energy-dispersive X-ray spectroscopy (EDS) was carried out using an FEI Tecnai G2 F30 TEM equipped with an Oxford EDAX EDS system. Atomic force microscopy (AFM, Bruker, with 512 pixels per line) was performed aer depositing a drop of dispersion onto a silicon substrate. The X-ray diffraction (XRD) analysis was performed on an X'Pert-Pro MPD diffractometer with a Cu K-a radiation source at room temperature. Ultraviolet-visible (UV-Vis) absorption spectroscopy was performed with a spectral range of 200-1500 nm by using a UV-Vis absorbance spectrometer (Cary 60, Agilent) at room temperature.

Photoresponse activity
The PEC measurement system in Scheme S1 † was used to characterize the photoresponse behaviour of Bi 2 Te 3 NPs. A standard three-electrode system, that is, a working electrode (for example, Bi 2 Te 3 NPs deposited on ITO-coated glass, photoanode), a counter electrode (platinum wire, photocathode), and a reference electrode (Ag/AgCl electrode), was assembled in various aqueous electrolytes, including KOH (0.1 M, 0.5 M, and 1.0 M), KCl (0.5 M), and HCl (0.5 M). To ensure good adhesion between ITO-coated glass and the sample, the as-synthesized samples were rst re-dispersed in a 0.2 mg mL À1 PVDF/DMF solution, and then the dispersion was deposited onto ITOcoated glass and dried under vacuum at 80 C overnight. Electrochemical impedance spectra (EIS) were obtained in the frequency range from 1 to 10 5 Hz with an amplitude of 0.005 V. Amperometric current-time (I-t) curves were recorded at bias voltages of 0 V, 0.3 V, and 0.6 V with increasing power densities at a sampling interval of 5 s. Simulated light (300-800 nm) and lasers with different l (l ¼ 365 nm, 400 nm, 475 nm, 550 nm, 600 nm and 700 nm) were employed to irradiate the Bi 2 Te 3 NPbased photodetectors. Light power densities (P l ) of these irradiations with labels of Dark, I, II, III, IV, and VI levels gradually increased (Table S1 †). As a control experiment, a piece of naked ITO-coated glass was also irradiated by using a SL under the same conditions.

Results and discussion
The Bi 2 Te 3 NPs with a well-dened hexagonal shape were synthesized by the solvothermal method. In order to investigate the inuence of the size of Bi 2 Te 3 NPs on photoresponse performances, two kinds of Bi 2 Te 3 NPs with different sizes have been facilely synthesized by tuning the reaction time. For convenience, the Bi 2 Te 3 NPs reacted aer 2 h and 12 h are abbreviated as Bi 2 Te 3 NPs-1 and Bi 2 Te 3 NPs-2, respectively. Fig. 1a and b give the SEM images of the as-prepared Bi 2 Te 3 NPs-1 and Bi 2 Te 3 NPs-2, both of which exhibit a well-dened hexagonal shape, while the average lateral dimensions of Bi 2 Te 3 NPs-1 and Bi 2 Te 3 NPs-2 are 620 AE 150 nm and 730 AE 210 nm, respectively. The TEM characterization ( Fig. 1c and d) reveals that the as-prepared Bi 2 Te 3 NPs also represent a well-dened hexagonal shape. Besides, Bi 2 Te 3 NPs-2 have darker hexagons (Fig. 1d), compared to Bi 2 Te 3 NPs-1 (Fig. 1c), indicating that the longer the reaction time the thicker the Bi 2 Te 3 NPs are, which is in good agreement with the "oriented attachment" mechanism of Bi 2 Te 3 NPs. 38,39 The HRTEM image (Fig. 1e) shows a clear lattice spacing of 0.22 nm, consistent with the (1120) plane of layered Bi 2 Te 3 . 27 Sharp diffraction spots are observed in the selected area electron diffraction (SAED) pattern (Fig. 1f), and the EDS line scan analysis (Fig. 1g) reveals compositional variation in a single Bi 2 Te 3 NP, suggesting that Bi and Te are evenly distributed ( Fig. 1h and j).
The thicknesses of the as-synthesized Bi 2 Te 3 NPs-1 and Bi 2 Te 3 NPs-2 were characterized by AFM, as shown in Fig. 2a and b, respectively. It can be clearly seen in Fig. 2c and d that with the increase in the reaction time, the measured thickness of Bi 2 Te 3 NPs obviously increases from 11.1 nm to 17.2 nm, which correspond to 11 and 17 layers, respectively, given that one layer is regarded as an average quintuple layer of Te-Bi-Te-Bi-Te with a thickness of 1.0 nm. 17 XRD patterns of the assynthesized Bi 2 Te 3 NPs, as shown in Fig. 2e, can be indexed to a rhombohedral Bi 2 Te 3 structure (JCPDS Card Number 15-0863). UV-Vis absorption spectroscopy was employed to characterize the optical response of differently sized Bi 2 Te 3 NPs (Fig. 2f). Broadband absorption from 260 nm to 1500 nm, is observed for Bi 2 Te 3 NPs-1 and Bi 2 Te 3 NPs-2, which is in good agreement with previously reported results, 40,41 implying great potential for application in broadband optoelectronic devices. Besides, Tauc plots of Bi 2 Te 3 NPs-1 and Bi 2 Te 3 NPs-2 (Fig. 1g) were calculated based on the results in Fig. 1f, and sizedependent E g values of 0.83 eV (Bi 2 Te 3 NPs-1) and 1.0 eV (Bi 2 Te 3 NPs-2) were obtained, close to that of the previously reported Bi 2 Te 3 nanoparticles, 42 suggesting that the E g of Bi 2 Te 3 NPs could be easily controlled by simply tuning the reaction conditions.
The typical photoresponse behaviour of the Bi 2 Te 3 NP-based photodetector was evaluated using a PEC system equipped with a standard three-electrode conguration, as shown in Scheme S1. † Fig. 3a gives the patterns of the as-fabricated Bi 2 Te 3 NPs cast onto ITO-coated glass, exhibiting strong on/off switching photoresponse behaviours at an applied bias voltage of 0.6 V. For clarity, the prole of the naked ITO-coated glass was added in Fig. 3a, which displays a negligible signal as compared to Bi 2 Te 3 NPs under the same conditions, revealing that the photoresponse signal indeed comes from the Bi 2 Te 3 NPs rather than ITO. In addition, it should be noted that the type of electrolytes plays an important role in the PEC performance. As is shown in Fig. 3b, c and e, Bi 2 Te 3 NPs-2 irradiated by using several single-wavelength lasers, exhibit an excellent on/off switching photoresponse behaviour in 0.5 M KCl (Fig. 3c) and 0.5 M KOH (Fig. 3e) while they show poor performance in 0.5 M HCl (Fig. 3b) at 0.6 V, suggesting that KCl and KOH are preferred electrolytes for Bi 2 Te 3 NPs in this PEC system but HCl is not. Besides, the inuences of lasers with different l and P l on the PEC performance were investigated (Fig. 3c-f). Six lasers with specic l (l ¼ 365, 400, 475, 550, 650, and 700 nm) and P l were employed to demonstrate the laser wavelength-dependent PEC performance of Bi 2 Te 3 NPs. Similarly, the proles of naked ITO-coated glass irradiated by using a SL are also added for comparison in Fig. 3c-f. It can be observed that when the l value  Nanoscale Advances is less than 550, Bi 2 Te 3 NPs show a strong PEC signal in KCl and KOH and the signal gradually increases with the decrease of the l value ( Fig. 3c-f), which can be attributed to the relatively highe laser energy and P l (Table S1 †). However, when l $ 550 nm, a negligible photoresponse signal of Bi 2 Te 3 NPs can be observed, different from the results in absorption spectra, which is attributed to the very weak laser energy employed for PEC measurements in this work. In addition, it should be pointed out that the signal of naked ITO-coated glass irradiated by using a SL with much higher P l is obviously lower than that of Bi 2 Te 3 NPs (Fig. 3c-f) irradiated by using single-wavelength lasers with shorter l, e.g., 365 nm, while higher than that irradiated by using single-wavelength lasers with longer l, e.g., 650 nm and 700 nm, reconrming the truth of the PEC signal of Bi 2 Te 3 NPs. Furthermore, it shows the same trend when the P l gradually increases from dark to VI in both KCl and KOH, that is, the PEC signal of Bi 2 Te 3 NPs increases with the P l (Fig. 3c-f).
To quantitatively evaluate the photoresponse performance of Bi 2 Te 3 NPs, the photocurrent density (P ph ) and photoresponsivity (R ph ) can be obtained by: 43,44 P ph ¼ (I light À I dark )/S (1) where, I light and I dark are the drain current with and without light, respectively; P l and S are the light power density and effective area of the Bi 2 Te 3 NPs on ITO-coated glass, respectively. The bias voltage dependent on the PEC performance was also studied ( Fig. 3e and S1 †). The P ph of Bi 2 Te 3 NPs-2 in 0.5 M KOH gradually increases with the applied bias voltage, i.e., the P ph of Bi 2 Te 3 NPs-2 irradiated by using a 365 nm laser increases from 44.8 nA cm À2 (0 V), to 96.4 nA cm À2 (0.3 V), to 2.52 mA cm À2 (0.6 V), similar to that in 0.5 M KCl (Fig. 3c and S2 †). Photocurrent generation of the Bi 2 Te 3 NPs at 0 V means that the Bi 2 Te 3 NP-based photodetector is able to display selfpowered PEC performance in KOH, yet additional external bias voltage can strengthen the photocurrent generation, which can be ascribed to the fact that the external bias voltage across the photoelectrode can construct a potential gradient within Bi 2 Te 3 NPs and enhance the separation of photogenerated holes and electrons. 45 Therefore, we think that the photoresponse mechanism of Bi 2 Te 3 NPs is similar to that of bismuthene and tellurene: (i) formation of electron (e À )-hole (h + ) pairs by photoexcitation and (ii) photoinduced charge transportation. 10,11 Surprisingly, it should be noted that the P ph and R ph of Bi 2 Te 3 NPs can reach up to 8.68 mA cm À2 and 395 mA W À1 (Fig. 3g and h), respectively, both of which largely outperform the reported bismuthene-based or tellurene-based photodetectors, 10,11 which could be attributed to the unique property of the topological insulator, Bi 2 Te 3 . Notably, the P ph of Bi 2 Te 3 NPs in this work is also remarkably superior to those of ZnO homojunction nanowires ($0.28 nA cm À2 ) 46 and GaN nanowires ($0.45 nA cm À2 ), 47 considering the same effective area of the measured samples (2.2 cm 2 ). Moreover, it is observed ( Fig. 3g and h) that electrolyte concentration has a great effect on the PEC signal of Bi 2 Te 3 NPs, that is, the P ph and R ph of Bi 2 Te 3 NPs irradiated by using a 400 nm laser increase in the range of KOH concentration from 0.1 M to 0.5 M while both of them decrease from 0.5 M to 1.0 M, possibly ascribed to the slight electrochemical reaction at both bias voltage and high electrolyte concentration. 48 The resistance (R) at the interface between the electrolyte and electrode gradually decreases with the increase of KOH concentration,  (Fig. 3i) can be due to the different functionalities between Bi 2 Te 3 NPs and electrolytes, similar to the reported results of black phosphorus nanosheets 49 and bismuth sulde(III) nanosheets. 45 Due to the size-dependent E g of Bi 2 Te 3 NPs (Fig. 2g), the photoresponse behaviours of the Bi 2 Te 3 NPs-1 and Bi 2 Te 3 NPs-2 irradiated by using a SL and three kinds of lasers (365, 400, and 475 nm) were studied to understand the inuence of the size of Bi 2 Te 3 NPs on the PEC performance, as shown in Fig. 4. It can be seen that the PEC signals of both Bi 2 Te 3 NPs-1 and Bi 2 Te 3 NPs-2 increase with the P l (Fig. 4a) and show the same trend as those of the Bi 2 Te 3 NPs-2, i.e., the PEC signal declines as the l value increases (Fig. 4b). In addition, it is noted that the PEC signal of Bi 2 Te 3 NPs-2 is obviously stronger than that of Bi 2 Te 3 NPs-1, no matter which laser was employed ( Fig. 4a  and b), e.g., the P ph of Bi 2 Te 3 NPs-2 irradiated by using a 400 nm laser at 4.65 mW cm À2 is 2.52 mA cm À2 while that of Bi 2 Te 3 NPs-1 is only 0.729 mA cm À2 . This could be attributed to the synergistic effect of the suitable E g and the number of accessible active sites on the Bi 2 Te 3 NPs-2. Since the E g of Bi 2 Te 3 NPs inversely correlates with size, the larger Bi 2 Te 3 NPs-1 have stronger absorption under incident light. However, with the size decrease of Bi 2 Te 3 NPs, the specic surface area becomes higher and accessible active sites on the PEC performance of Bi 2 Te 3 NP-based photodetectors, similar to previously reported results. 10,44,50 Therefore, the size of nanomaterials has a great inuence on the PEC performance; Bi 2 Te 3 NPs tend to be larger, resulting in higher efficiency, which provides fundamental acknowledgement of the design and optimization of PEC-type devices.
In addition, the response time (t res ) and recovery time (t rec ) of the Bi 2 Te 3 NP-based photodetector were ascribed to the time interval for the rise and decay from 10% to 90% and from 90% to 10% of its peak value, respectively. 48,51 It can be observed that regardless of the size of Bi 2 Te 3 NPs and the type of electrolytes, the Bi 2 Te 3 NP-based photodetector always shows fast t res (0.001-0.09 s) as well as t rec (0.001-0.07 s) (Fig. 5), both of which are superior to those of the bismuthenebased photodetector (t res ¼ 0.2 s, t rec ¼ 0.2 s), 11 tellurene-based photodetector (t res ¼ 0.2 s, t rec ¼ 0.2 s), 10 ZnO homojunction nanowires (t res ¼ $50 s, t rec ¼ $200 s) 46 and GaN nanowires (t res ¼ 0.003 s, t rec ¼ 0.003 s). 47 This could be attributed to the unconventional phase of Bi 2 Te 3 quantum matter. This  indicates that the Bi 2 Te 3 NP-based photodetector has appealing potential in the eld of PEC-type devices. Long-term stability measurements of the photoresponse of Bi 2 Te 3 NP-based photodetectors are of great importance for practical application. Fig. 6 gives the stability proles of the Bi 2 Te 3 NP-1-based photodetector irradiated by using a SL in 0.1 M KOH at a bias voltage of 0.6 V. The 600 cycles with 5 s intervals of on/off switching were traced (Fig. 6a), and the 481-500 th cycles were chosen to evaluate the PEC stability of the Bi 2 Te 3 NP-1-based photodetector (Fig. 6b). It should be pointed out that no obvious change is observed by visual inspection of Bi 2 Te 3 NPs-1 specimens aer PEC stability measurements in 0.1 M KOH (Fig. S3 †), indicating the excellent PEC stability of the as-fabricated Bi 2 Te 3 NP-1-based photodetector. A notable on/off switching behaviour can be observed even aer one month, suggesting the long-term PEC stability of Bi 2 Te 3 NPs under ambient conditions. Furthermore, it is calculated from Fig. 6b that the P ph s of the fresh Bi 2 Te 3 NPs-1 in the 481-500 th cycles is 879 nA cm À2 in 0.1 M KOH, declining to 439 nA cm À2 aer one month. An approximate reduction of 50.1% of P ph was obtained, comparable to that of the bismuthene-based photodetector 11 and tellurene-based photodetector. 10 The decline of P ph could be ascribed to the weak electrochemical reaction at high bias voltage (0.6 V) under a SL with high P l (134 mW cm À2 ) and slight peel-off during long-running measurements, which can be efficiently solved by coating conductive polymers, such as polyaniline and polypyrrole, onto the surface of Bi 2 Te 3 NPs to remarkably lower the electrochemical reaction of Bi 2 Te 3 NPs and employ stronger binders to make Bi 2 Te 3 NPs more strongly xed on the surface of ITO-coated glass to avoid the slight peeloff during long-running measurements.

Conclusions
In summary, topological insulator Bi 2 Te 3 NPs were successfully synthesized by a solvothermal approach and the size of Bi 2 Te 3 NPs can be readily controlled by simply tuning the reaction time. The hexagonal structure of Bi 2 Te 3 NPs was well-characterized, and UV-Vis spectra revealed a broadband absorption range from 260 nm to 1500 nm. The as-synthesized Bi 2 Te 3 NPs were, for the rst time, employed as a working material in a PEC-type photodetector. The PEC result not only shows that the P ph and R ph signicantly improve but also exhibits faster t res and t rec , compared to those of bismuthene-based or tellurenebased photodetectors. It was also shown that Bi 2 Te 3 NPs-2 displayed better PEC performance, attributed to the synergistic effect of optical absorbance and the number of accessible active sites on a Bi 2 Te 3 NP. In addition, good PEC stability of the Bi 2 Te 3 NP-based photodetector was obtained in 0.1 M KOH aer one month without any protection, comparable to the bismuthene-based or tellurene-based photodetector. Because of the facile synthesis, easy size control of Bi 2 Te 3 NPs, excellent photoresponse performance, and good long-term stability of the Bi 2 Te 3 NP-based photodetector, we believe that Bi 2 Te 3 can pave a new way for the design of bismuthene or tellurene nanomaterial-based high-performance PEC-type devices with practical applicability.

Conflicts of interest
There are no conicts to declare.