Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Electroluminescence and Photocurrent Generation from Atomically Sharp WSe2/MoS2 Heterojunction p–n Diodes

Electroluminescence and Photocurrent Generation from Atomically Sharp WSe2/MoS2 Heterojunction... Letter pubs.acs.org/NanoLett Terms of Use Electroluminescence and Photocurrent Generation from Atomically Sharp WSe /MoS Heterojunction p−n Diodes 2 2 † ‡ ‡ † ‡ ‡ † † Rui Cheng, Dehui Li, Hailong Zhou, Chen Wang, Anxiang Yin, Shan Jiang, Yuan Liu, Yu Chen, ,†,§ ,‡,§ Yu Huang,* and Xiangfeng Duan* Department of Materials Science and Engineering, University of California, Los Angeles, California 90095, United States Department of Chemistry and Biochemistry, University of California, Los Angeles, California 90095, United States California Nanosystems Institute, University of California, Los Angeles, California 90095, United States * Supporting Information ABSTRACT: The p−n diodes represent the most fundamental device building blocks for diverse optoelectronic functions, but are difficult to achieve in atomically thin transition metal dichalcogenides (TMDs) due to the challenges in selectively doping them into p-or n-type semiconductors. Here, we demonstrate that an atomically thin and sharp heterojunction p−n diode can be created by vertically stacking p-type monolayer tungsten diselenide (WSe ) and n-type few-layer molybdenum disulfide (MoS ). 2 2 Electrical measurements of the vertically staked WSe /MoS heterojunctions reveal excellent current rectification behavior with 2 2 an ideality factor of 1.2. Photocurrent mapping shows rapid photoresponse over the entire overlapping region with a highest external quantum efficiency up to 12%. Electroluminescence studies show prominent band edge excitonic emission and strikingly enhanced hot-electron luminescence. A systematic investigation shows distinct layer-number dependent emission characteristics and reveals important insight about the origin of hot-electron luminescence and the nature of electron−orbital interaction in TMDs. We believe that these atomically thin heterojunction p−n diodes represent an interesting system for probing the fundamental electro-optical properties in TMDs and can open up a new pathway to novel optoelectronic devices such as atomically thin photodetectors, photovoltaics, as well as spin- and valley-polarized light emitting diodes, on-chip lasers. KEYWORDS: WSe , MoS , heterojunction, electroluminescence, photocurrent, van der Waals 2 2 wo-dimensional layered materials, such as graphene, from ML-MoS has been reported in a metal-MoS Schottky 2 2 T MoS , and WSe , are emerging as an exciting material junction through a hot carrier process. Electrostatic doping 2 2 system for a new generation of atomically thin optoelectronics, has also been used to create planar p−n diodes, but usually with 1−7 8 9 including photodetectors, ultrafast lasers, polarizers, touch relatively gradual doping profile (limited by the fringe electrical 10 11 panels, and optical modulators due to their unique field) and typically relatively low optoelectronic efficiency (e.g., 12−23 electronic and optical properties. In this regard, the photon to electron conversion external quantum efficiency monolayer transition metal dichalcogenides (ML-TMDs) is 26−28 (EQE) ∼0.1−1%). particularly interesting due to their direct energy bandgap and The atomically thin geometry of these 2D materials can allow 12,13 the non-centrosymmetric lattice structure. The p−n diodes band structure modulation in a vertically stacked hetero- represent the most fundamental device building blocks for most structures to form atomically sharp heterojunctions. For optoelectronic functions, including photodiodes and light example, this strategy allows gapless graphene to be used in emitting diodes. However, it is particularly difficult to create p−n diodes in atomically thin TMDs due to the challenges in selectively doping them into p- or n-type semiconductors. Received: June 4, 2014 Contact engineering has been explored to create p−n diodes in Revised: August 15, 2014 TMD based layered materials. Electroluminescence (EL) Published: August 26, 2014 © 2014 American Chemical Society 5590 dx.doi.org/10.1021/nl502075n | Nano Lett. 2014, 14, 5590−5597 Nano Letters Letter Figure 1. Schematic illustration and band diagram of the WSe /MoS vertical heterojunction p−n diode. (a) A schematic illustration of the WSe / 2 2 2 MoS vertical heterojunction device shows that a transferred MoS flake on synthetic WSe forms a vertical heterojunction. (b) A schematic 2 2 2 illustration the cross-sectional view of the WSe /MoS vertical heterojunction device. (c) The ideal band diagram of WSe /MoS heterojunction p−n 2 2 2 2 diode under zero bias. Figure 2. Structural characterization of the WSe /MoS heterojunction p−n diode. (a) Optical microscopy image of a truncated triangular domain of 2 2 monolayer WSe with an inverted triangular bilayer region at the center. (b) The false color SEM image of the WSe /MoS vertical heterojunction 2 2 2 device, with ML-WSe highlighted by blue color, BL-WSe area by violet color, MoS by green color, and metal electrodes by golden color. The scale 2 2 2 bar is 3 μm. (c) The PL mapping of the WSe /MoS heterojunction device, with red color representing the PL from MoS and the green color 2 2 2 representing PL from WSe . (d) The PL spectra of synthetic ML- and BL-WSe and few-layer MoS flakes with the A, B exciton peaks and indirect 2 2 2 transition I peak labeled. The intensities of BL-WSe2 and FL-MoS are multiplied by 10 times for better visibility. Inset, the B exciton peak in ML- WSe . (e) The cross-sectional HRTEM image of the WSe /MoS heterojunction interface. The scale bar is 5 nm. (f) The EDS element distribution 2 2 2 profile from the bottom to the top of panel e. The black square represent the distribution profile of W-L characteristic peaks. The red line represents the fitting curve for W-L distribution profile, with a full width at half-maximum of 1.2 nm, corresponding to bilayer WSe . 29,30 31 32 field-effect tunnelling devices, barristors, inverters, and shows fast photoresponse and demonstrates that the p−n photodetectors while staked together with other 2D materials junction is created throughout the entire WSe /MoS over- 2 2 in the vertical direction. Although the nearly perfect 2D lapping area. Furthermore, prominent EL is observed with rich structure and low density of states in graphene provide spectral features that can reveal important insights about advantages in some heterostructure devices, its gapless nature electron−orbital interaction in TMD based materials. prevents the formation of a large potential barrier for charge The vertical heterojunction p−n diode is formed between separation and current rectification. On the other hand, the synthetic p-type ML-WSe and exfoliated n-type MoS flake 2 2 vertical heterojunction p−n diode formed between a TMD (Figure 1a,b). Triangular domains of ML-WSe was first material and a bulk material has recently been reported, but synthesized on 300 nm Si/SiO substrate typically with a 33,34 35 usually with no EL or very weak EL. bilayer (BL) region in the center (see Figure 2a), which were Here we report an atomically thin p−n diode based on a characterized by using optical microscope, atomic force heterojunction between synthetic p-type ML-WSe and microscopy (AFM), and Raman spectroscopy (Supporting exfoliated n-type MoS flake. The atomically thin p−n diode Information Figure S1). Mechanically exfoliated MoS flakes 2 2 exhibits well-defined current rectification behavior and can were then transferred onto synthetic WSe domains to form enable efficient photocurrent generation with an EQE up to vertically stacked heterojunctions. Electron-beam lithography 12%. Unlike the planar structures where the active area is and electron beam evaporation were used to define the contact confined to the lateral interface region, photocurrent mapping electrodes. A thin Ni/Au film (5 nm/50 nm) and Au film (50 5591 dx.doi.org/10.1021/nl502075n | Nano Lett. 2014, 14, 5590−5597 Nano Letters Letter Figure 3. Electrical characterization of the WSe /MoS heterojunction p−n diode. (a) The I −V characteristics of n-type MoS FET transistor with 2 2 ds ds 2 Ni/Au (5/50 nm) contacts. (b) The I −V characteristics of p-type WSe FET transistor with Au (50 nm) contacts. (c) Gate-tunable output ds ds 2 characteristics of the WSe /MoS heterojunction p−n diode. (d) The derivation of the p−n diode ideality factor by using a model consists of an ideal 2 2 p−n diode with a series resistor. An ideality factor of 1.2 was derived with a series resistor of 80 MΩ at 0 V gate voltage (red circle), and an ideality factor of 1.3 was derived with a series resistor of 33 MΩ at −20 V gate voltage (green triangle). nm) were used as the electrode for MoS flake and WSe peak in MoS . It is also important to note that the “B” exciton 2 2 2 domain to form Ohmic contacts with minimized contact peak can be observed in both WSe (605 nm) and MoS (620 2 2 resistance and potential barrier (Figure 1b). Figure 1c shows nm), but with the intensity typically 2−3 order magnitude the ideal band diagrams of the heterojunction p−n diode at weaker than the “A” exciton peak (Figure 2d inset). zero bias. The built-in potential and applied voltage are mainly We have further characterized the atomic structure of the supported by a depletion layer with abrupt atomic boundaries, stacked heterojunction using cross-sectional transmission and outside the boundaries, the semiconductor is assumed to electron microscope (TEM) studies. The high resolution be neutral. TEM image clearly shows the WSe /MoS heterojunction 2 2 Figure 2a shows an optical microscopy image of a synthetic with a 13-layer MoS flake on top of BL-WSe (Figure 2e). 2 2 WSe domain on 300 nm Si/SiO substrate. A triangular shaped Energy dispersive X-ray spectroscopy (EDS) was further used 2 2 BL-WSe domain was typically observed in the center of the to analyze the elemental distribution across the heterojunction triangular ML-WSe domain, indicating the nearly perfect interface. An EDS elemental line scan in vertical direction lattice structure of the synthetic WSe . Figure 2b shows a top- shows a rather narrow tungsten distribution profile located near view scanning electron microscopy (SEM) image of the vertical the heterojunction interface (Figure 2f), with the full width at heterojunction. The MoS ,WSe layers and the contact half-maximum ∼1.2 nm, corresponding to BL-WSe observed 2 2 2 electrodes are labeled with different artificial colors to highlight in Figure 2e. Together, these structural and PL character- the device structure. Photoluminescence (PL) mapping was izations demonstrate that the atomically sharp heterojunctions used to illustrate the stacking structure of WSe /MoS are formed by vertically stacking atomically thin WSe and 2 2 2 heterojunction (Figure 2c). The PL mapping shows distinct MoS . PL emission from WSe (red region in Figure 2c) and MoS Before testing the electrical characteristics of the hetero- 2 2 region (green region in Figure 2c), with a consistent structure junction p−n diodes, we have first characterized the electrical layout as that observed in the SEM image (Figure 2b). The transport properties of MoS and WSe to ensure Ohmic 2 2 uniform PL from WSe and MoS also indicates the excellent contacts were achieved. To this end, the MoS and WSe field 2 2 2 2 effect transistors (FETs) were fabricated on Si/SiO substrate, quality of the TMD materials. The PL spectra of WSe show a 2 2 strong layer-number dependence (Figure 2d), with the PL with Ni/Au thin film as the source-drain contacts for MoS , and intensity in ML-WSe at least 10 times stronger than that in Au thin film as the contacts for WSe , and the silicon substrate 2 2 BL-WSe . The PL spectrum in ML-WSe shows a peak at 785 as a back gate electrodes. Figure 3a and b show the I −V 2 2 ds ds nm, corresponding to the “A” exciton peak. The PL in BL- characteristics at varying back gate voltages for MoS and WSe , 2 2 WSe also exhibits the “A” exciton peak with an additional respectively. Importantly, a linear I −V relationship is clearly 2 ds ds broad peak at ∼877 nm, which is attributed to indirect band observed for both MoS and WSe , indicating Ohmic contacts 2 2 gap emission (typically label as “I” peak). These PL studies are achieved for both materials. The formation of Ohmic are consistent with previous experimental studies and contacts for both MoS and WSe is very important because the 2 2 16,37 theoretical calculations, indicating the good crystalline Schottky barrier at the contact area may severely affect the quality of the synthetic WSe . The PL spectrum from MoS electronic and optoelectronic characteristics of the vertical 2 2 flake shows a peak at 677 nm, corresponding to “A” exciton heterojunction and could induce photocurrent generation or 5592 dx.doi.org/10.1021/nl502075n | Nano Lett. 2014, 14, 5590−5597 Nano Letters Letter Figure 4. Photoresponse of the WSe /MoS heterojunction p−n diode. (a) Optical microscpope image of the WSe /MoS heterojunction. (b) False 2 2 2 2 color scanning photocurrent micrograph of the WSe /MoS heterojunction device acquired at V = 0 V and V = 0 V under irradiation 514 nm 2 2 ds BG laser (5 μW). The purple square dotted line outlines the ML-WSe and the dark purple square dotted line outlines the BL-WSe . The blue circle 2 2 dotted line outlines the MoS and the golden solid line outlines the gold electrodes. Photocurrent were observed in the entire overlapping junction area. (c) Experimental output (I −V ) characteristic of the vertical heterojunction device in the dark (black) and under illumination (wavelength: ds ds 514 nm; power, 5 μW). Inset, temporal response of the photocurrent generation under 514 nm illumination (10 μW). (d) Power-dependent EQE of the heterojunction device under 514 and 633 nm laser excitation wavelengths at V = 0 V and V = 0 V. A maximum EQE of 12% was observed. ds BG EL at the contact region. Furthermore, I −V plots at of a p−n diode with a series resistor (Figure 3d). Importantly, ds ds varying back gate voltage show that the current increases with an ideality factor of n = 1.2 was derived with a series resistance increasing positive gate voltage for MoS , indicating an n-type of 80 MΩ, at zero gate voltage, and an ideality factor of n = 1.3 semiconductor behavior. On the contrary, the current increases was derived with a series resistance of 33 MΩ,at −20 V gate with decreasing negative gate voltage in WSe FET, consistent voltage. The achievement of ideality factor close to 1 indicates with the p-type characteristics. the excellent diode behavior of our atomically sharp With Ohmic contacts formed for both MoS and WSe ,we heterojunction p−n diode. The decrease of the series resistance 2 2 continue to probe the electrical transport properties of the with increasing negative gate voltage is also consistent with our heterojunction p−n diode. Importantly, a clear current model that the p-type WSe is the limiting series resistor at high rectification behavior is observed in (I −V ) plots for the forward bias. ds ds WSe /MoS heterojunction (Figure 3c), with current only The electrical measurements indicate excellent diode 2 2 being able to pass through the device when the p-type WSe is behavior in the atomically thin vertical heterojunction. To positively biased. The observation of current rectification clearly further characterize the diode characteristics in our vertical demonstrates a p−n diode is formed within the atomically thin heterojunction, the photocurrent mapping was carried out at WSe /MoS heterojunction. The ultrathin nature of the zero bias under a confocal microscope. Figure 4a shows an 2 2 heterojunction allows gate tunability of the diode character- optical microscope image of the WSe /MoS heterojunction 2 2 istics. The diode output characteristics (I −V ) under different depicting the relative position between WSe , MoS , and the ds ds 2 2 back gate voltage show that the output current decreases with electrodes. The corresponding photocurrent mapping at zero increasing positive gate voltage, suggesting that the p-type WSe bias with a 514 nm laser excitation (5 μW) is shown in Figure is partly limiting the charge transport in the device. 4b, with the ML-WSe region outlined by purple dotted line, The I −V output characteristics of the vertical hetero- few-layer MoS outlined by blue dotted line, and the electrodes ds ds 2 junction under forward bias can be viewed as a vertical outlined by golden solid lines. The photocurrent mapping heterojunction p−n diode in series with an additional p-type shows clear photoresponse from the entire overlapping region, FET due to the side contact on WSe . In general, the indicating the formation of p−n junction across the entire heterojunction p−n diode resistance decreases exponentially WSe /MoS overlapping area. It is also interesting to note that 2 2 with increasing bias voltage, and the series p-FET resistance is the photocurrent in ML-WSe /MoS region is much stronger 2 2 nearly constant with bias voltage. Therefore, the heterojunction than that in BL-WSe /MoS region, suggesting that the direct 2 2 resistance is dominated by p−n diode at low bias and band gap plays an important role in the photocurrent 6,7 dominated by the p-type WSe FET under high forward bias. generation process. A detailed understanding of the different We have also fitted the diode characteristics and calculated the response of ML vs BL-WSe /MoS will be an interesting topic 2 2 ideality factor of our heterojunction device based on the model for future studies. No measurable photocurrent was observed 5593 dx.doi.org/10.1021/nl502075n | Nano Lett. 2014, 14, 5590−5597 Nano Letters Letter Figure 5. Electroluminescence (EL) from the WSe /MoS heterojunction p−n diode. (a) The false color EL image of the heterojunction device 2 2 under an injection current of 100 μA. The purple dashed line outlines the ML-WSe , the blue dotted line outlines the MoS and the golden solid line 2 2 outlines the gold electrodes. (b) The EL spectra of a ML-WSe /MoS heterojunction at different injection current. (c) The EL spectra of a BL- 2 2 WSe /MoS heterojunction at different injection current. (d) The EL intensity as a function of injection current for both ML- and BL-WSe /MoS 2 2 2 2 heterojunction. (e) The ideal band diagram of the WSe /MoS heterojunction under small forward bias. The conduction band in MoS is below that 2 2 2 in WSe , the valence band in WSe is below that in MoS . At small bias, holes can go cross the junction and inject into n-type region, while the 2 2 2 electrons cannot go cross the junction. (f) The ideal band diagram of the WSe /MoS heterojunction under large forward bias. The conduction band 2 2 in MoS shifts upward and is higher than that in WSe , and the valence band in WSe is below it in MoS . At large bias, both electrons and holes can 2 2 2 2 go cross the junction and inject into the other side of the heterojunction. (g) The EL spectra of a ML-WSe /MoS heterojunction at different 2 2 temperature ranging from 25 to 75 °C. (h) The EL spectra of a BL-WSe /MoS heterojunction at different temperature ranging from 25 to 75 °C. 2 2 The injection current is fixed at 250 μA. (i) The normalized intensities of A and B′ peaks in the EL spectra of both ML- and BL-WSe as a function of temperature. from the non-overlapping regions (only WSe or MoS ) or the EQE in our vertical heterojunction can reach 11% under a 514 2 2 electrical contacts, which is expected for zero bias photocurrent nm laser excitation with a power of 5 μW. Furthermore, it is because the photogenerated carries in the regions outside p−n found that the EQE decreases with increasing excitation power junction cannot be effectively separated and extracted. (Figure 4d), with a maximum EQE of 12% observed under an The output characteristics (I −V )ofthe vertical excitation power of 0.5 μW. The decreasing EQE with ds ds heterojunction with and without laser illumination (514 nm, increasing excitation power could be attributed partly to 5 μW) show clear photovoltaic effect with an open-circuit absorption saturation in WSe and partly to the screening of the voltage of ∼0.27 V and a short-circuit current of ∼0.22 μA built-in electric field by the excited holes in the valence band of (Figure 4c). In general, the photoresponse exhibits a rapid WSe . The power dependent EQE of the same device under temporal response beyond our experimental time resolution of 633 nm excitation shows a similar trend but with generally 100 μs (Figure 4c inset), demonstrating that the photoresponse lower values than those under 514 nm excitation, which may be is originated from photocarrier generation rather than any other attributed to the spectral dependent optical absorption extrinsic effects. Based on the photocurrent response and input coefficient. It is important to note that the EQE observed in laser power, we can determine the external quantum efficiency the vertical WSe /MoS heterostructure devices is much higher 2 2 (EQE) of the photon to electron conversion. The EQE (η)is than those in lateral electrostatically doped WSe p−n 27,28 defined as the ratio of the number of carriers collected by homojunctions (0.1−3%), which may be partly attributed electrodes to the number of the incident photon, or η =(I / to more efficient charge separation resulting from an atomically ph q)/(P/hν) × 100% where I is the photocurrent, h is Planck’s sharp vertical p−n junction. In contrast, the electrostatic doping ph constant, ν is the frequency of light, q is the electron charge, would typically exhibit a spatial doping gradient, and is difficult and P is the incident light power. Our study showed that the to achieve atomically sharp junctions. 5594 dx.doi.org/10.1021/nl502075n | Nano Lett. 2014, 14, 5590−5597 Nano Letters Letter The above electrical transport and photocurrent studies HEL A′,B′ peaks, which are usually 100−1000 times weaker demonstrate excellent p−n diode characteristics in the than the A exciton peak in the PL measurements and have atomically sharp WSe /MoS heterojunction. Because p−n not been reported in EL previously. In contrast, the EL spectra 2 2 diode represents the basic device element for a light-emitting of the vertical heterojunction show that the intensities of these diode, we have further investigated the electroluminescence HEL peaks are only 3−10 times weaker than the A exciton from these heterojunction p−n diodes. Figure 5a shows an EL peak, suggesting a relative enhancement of the HEL by about 2 image acquired under a forward bias of 3 V and a forward orders of magnitude in our EL studies, which may be attributed current of ∼100 μA. The shapes of WSe , MoS , and gold to the electric field induced carrier redistribution. 2 2 electrodes were outlined in the same way as before to identify The origin of the HEL peaks in TMD materials remains the the position of the EL. In contrast to the photocurrent subject of debate and is difficult to probe due to their low 37,40−42 generation from the entire overlapping area, it is important to emission probability. The HEL peaks A′ and B′ are note that the EL is localized at the overlapping area in close generally believed to arise from the splitting of the ground and proximity to the electrodes. This can be explained by the excited states of A and B transitions due to the electron−orbital electric field distribution in the heterojunction under different interaction via either inter- or intralayer perturbation or 40,41 both. However, there is no yet clear evidence to prove bias. For photocurrent mapping at zero bias (or a small bias less than the turn on voltage), the p−n diode junction resistance which perturbation dominates the electron−orbital interaction. dominates the entire device, and therefore, photocurrent can be The emergence of intense HEL emission in our ML-WSe / seen from the entire overlapping area where there is a p−n MoS and BL-WSe /MoS heterojunction can offer a new 2 2 2 junction. For EL studies at much higher forward bias exceeding platform to probe the origin of HEL peaks and the nature of the p−n diode turn-on voltage, the resistance of the ML-WSe electron−orbital interaction in TMDs. The presence of HEL becomes an increasingly important component of the total peaks A′ and B′ in EL spectra of ML-WSe /MoS 2 2 resistance. Therefore, the most voltage drop occurs across the heterojunction (Figure 5b) indicates that intralayer perturba- heterojunction edge near the electrodes due to the large series tion plays a role in the formation of these HEL peaks. On the resistance of the ML-WSe . This is also consistent with the other hand, it is noted that the relative intensities of HEL peaks result recently reported for MoS /Si heterojunctions. (comparing with the respective A peak) in BL-WSe /MoS 2 2 2 Figure 5b and c show the EL spectra of a ML- and a BL- heterojunction (Figure 5c) are clearly much stronger than that WSe /MoS heterojunction with increasing injection current. in ML-WSe /MoS heterojunction, suggesting that interlayer 2 2 2 2 The plot of the overall EL intensity as a function of injection perturbation may also contribute to the HEL peaks (which can current shows an apparent threshold (Figure 5d), with little EL be further supported by temperature dependent studies; see below the threshold, and linear increase above threshold. The below). threshold current may be explained by the band alignment of To further probe physical mechanism governing the photon the heterojunction under different bias voltages (Figure 5e and emission process in the atomically thin p−n diode, we have also f). In general, due to different band gap and band alignment conducted the temperature dependent EL studies at 25, 50, and among the conduction band and valence band edge, the barrier 75 °C for both the ML- and BL-WSe /MoS heterojunctions 2 2 for hole transport across the junction is smaller than that for the (Figure 5g and h) and plotted the normalized peak intensities electrons. With increasing forward bias (below a certain for A and B′ peaks as a function of temperature (Figure 5i). For threshold), the holes from WSe are first injected into n-type ML-WSe /MoS heterojunction, the EL intensity of all spectral 2 2 2 MoS region, whereas few electrons can overcome the barrier peaks show a consistent decrease with increasing temperature to reach WSe (Figure 5e). Due to the nature of indirect band (Figure 5g and i), which is a common phenomenon in the LED gap in few-layer MoS , the yield of radiative recombination is devices and can be attributed to the exponential enhancement relatively low at this point. As a result, the EL intensity is very of nonradiative recombination rate with increasing temper- low when the hole injection dominates the charge transfer ature. In striking contrast, temperature dependent EL in the across the heterojunction. With further increasing bias across BL-WSe /MoS heterojunction displays highly distinct features. 2 2 the heterojunction (above electron injection threshold), the First, the A exciton peak in BL-WSe /MoS heterojunction 2 2 conduction band of MoS is shifted upward, both electrons and shows an unusual increase (instead of decrease) with increasing holes can go cross the heterojunction and are injected into p- temperature. (Figure 5h and i). This increase in A exciton type and n-type region, respectively (Figure 5f). At this point, emission may be explained by thermally decoupling neighbor- the radiative recombination in WSe dominates the EL with its ing layers via interlayer thermal expansion, which can induce a intensity increasing linearly with the injection current. It is band gap crossover from the indirect gap to the direct one with noted that the EL intensity observed in ML-WSe /MoS the increasing decoupling at higher temperature. A similar 2 2 heterojunction is much stronger than that in BL-WSe /MoS thermal decoupling effect has been observed in MoSe by PL 2 2 2 heterojunction due to the higher radiative recombination rate in studies. Second, the HEL peak B′ (and A′) shows a much direct band gap ML-WSe vs indirect bandgap BL-WSe . greater decrease with increasing temperature than that in ML 2 2 The EL spectra show rich spectral features and can be well -WSe /MoS , indicating the weakening of electron−orbital 2 2 fitted using multiple Gaussian functions (Supporting Informa- interaction with the decoupling neighboring layers. These tion Figure S2) with five main peaks, which can be assigned as temperature dependent characteristics are consistent seen in excitonic peaks A (∼792 nm) and B (∼626 nm), hot electron three devices studies and further suggest that the interlayer luminescence (HEL) peaks A′ (∼546 nm) and B′ (∼483 nm) perturbation plays an important role in electron−orbital and an indirect band gap emission peak I (∼880 nm). The A interaction in WSe , which is consistent with the observation exciton peak dominates the spectra of the EL in ML-WSe of strong interlayer excitons in TMD based heterojunc- 44,45 (Figure 5a), whereas the indirect band gap emission I is tions. significant in BL-WSe (Figure 5b). Strikingly, the EL spectra of In summary, we have fabricated WSe /MoS heterojunction 2 2 2 both ML- and BL-WSe show prominent B exciton peak and p−n diodes with atomically thin geometry and atomically sharp 5595 dx.doi.org/10.1021/nl502075n | Nano Lett. 2014, 14, 5590−5597 Nano Letters Letter interface. The scanning photocurrent measurement demon- spectra were taken by using an Acton 2300i spectrometer with strates that the p−n junction was formed over the entire 150 g/mm grating and liquid-nitrogen-cooled CCD. overlapping area with a maximum photon-to-electron con- ASSOCIATED CONTENT version EQE of 12%. The EL measurement allows for the identification of emission from different optical transitions. Hot * Supporting Information electron luminescence peaks were observed in EL spectra of The AFM and Raman characterizations of synthetic WSe WSe for the first time and used to investigate the electron− 2 domain on 300 nm Si/SiO substrate and the analysis and orbital interaction in WSe . Our novel heterojunction structure 2 peak fittings of the EL spectra. This material is available free of offers an interesting platform for fundamental investigation of charge via the Internet at http://pubs.acs.org. the microscopic nature of the carrier generation, recombination and electro-optical properties of single or few-layer TMD AUTHOR INFORMATION materials, and can open up a new pathway to novel Corresponding Authors optoelectronic devices including atomically thin photodetec- *E-mail: [email protected]. tors, photovoltaics, as well as spin- or valley-polarized light *E-mail: [email protected]. emitting diodes and on-chip lasers. Notes Note added: During the finalization of this manuscript we 46,47 The authors declare no competing financial interest. became aware of two related studies. Methods. Fabrication of the Vertical Heterostructure ACKNOWLEDGMENTS Devices. To fabricate the vertical WSe /MoS heterojunction 2 2 We acknowledge the Nanoelectronics Research Facility (NRF) devices, WSe was grown using a physical vapor deposition at UCLA for technical support. X.D. acknowledges support by process on a Si/SiO (300 nm SiO ) substrate. A total of 0.2 g 2 2 ONR Young Investigator Award N00014-12-1-0745. Y.H. WSe powder (Alfa Aesar, 13084) was added into an alumina acknowledges support by the NIH Grant 1DP2OD007279. boat as precursor. The blank Si/SiO substrates (1 cm × 5 cm) were loaded into a home-built vapor deposition system in a REFERENCES horizontal tube furnace (Lindberg/Blue M) with 1 in. quartz (1) Xia, F.; Mueller, T.; Lin, Y.-m.; Valdes-Garcia, A.; Avouris, P. Nat. tube. The system was pumped down to a vacuum of 10 mTorr Nanotechnol. 2009, 4, 839−843. in 10 min, and refilled with 150 sccm of ultrahigh purity argon (2) Mueller, T.; Xia, F.; Avouris, P. Nat. Photonics. 2010, 4, 297−301. gas (Airgas, ∼ 99.9999%) then heated to desired growth (3) Liu, Y.; Cheng, R.; Liao, L.; Zhou, H.; Bai, J.; Liu, G.; Liu, L.; temperature within 30 min. After that, the growth kept at the Huang, Y.; Duan, X. Nat. Commun. 2011, 2, 579. desired temperature for 30 min, and then terminated by (4) Sun, D.; Aivazian, G.; Jones, A. M.; Ross, J. S.; Yao, W.; Cobden, shutting off the power of the furnace. The sample was naturally D.; Xu, X. Nat. Nanotechnol. 2012, 7, 114−118. cooled down to ambient temperature. The MoS flakes were (5) Lopez-Sanchez, O.; Lembke, D.; Kayci, M.; Radenovic, A.; Kis, A. then exfoliated onto the WSe flakes through a micro- 2 Nat. Nanotechnol. 2013, 8, 497−501. mechanical cleavage approach. The metal electrodes (for (6) Yu, W. J.; Liu, Y.; Zhou, H.; Yin, A.; Li, Z.; Huang, Y.; Duan, X. probe contact or wire bonding purposes) were patterned on Nat. Nanotechnol. 2013, 8, 952−958. (7) Britnell, L.; Ribeiro, R.; Eckmann, A.; Jalil, R.; Belle, B.; the Si/SiO substrates by using electron-beam lithography and Mishchenko, A.; Kim, Y.-J.; Gorbachev, R.; Georgiou, T.; Morozov, S. electron-beam deposition of Ti/Au (50/50 nm) thin film. Ni/ Science 2013, 340, 1311−1314. Au (5/50 nm) contact electrodes were then deposited to form (8) Sun, Z.; Hasan, T.; Torrisi, F.; Popa, D.; Privitera, G.; Wang, F.; the Ohmic contact to MoS and Au (50 nm) was deposited to Bonaccorso, F.; Basko, D. M.; Ferrari, A. C. ACS Nano 2010, 4, 803− form the Ohmic contact to WSe . Microscopic, Electrical, Optical, and Optoelectrical Char- (9) Bao, Q.; Zhang, H.; Wang, B.; Ni, Z.; Lim, C. H. Y. X.; Wang, Y.; acterizations. The microstructures and morphologies of the Tang, D. Y.; Loh, K. P. Nat. Photonics. 2011, 5, 411−415. nanostructures are characterized by a JEOL 6700 scanning (10) Bae, S.; Kim, H.; Lee, Y.; Xu, X.; Park, J.-S.; Zheng, Y.; electron microscope (SEM). The cross-sectional image of the Balakrishnan, J.; Lei, T.; Kim, H. R.; Song, Y. I. Nat. Nanotechnol. 2010, 5, 574−578. heterostructure device is obtained by an FEI Titan transmission (11) Liu, M.; Yin, X.; Ulin-Avila, E.; Geng, B.; Zentgraf, T.; Ju, L.; electron microscope (TEM). The DC electrical transport Wang, F.; Zhang, X. Nature 2011, 474,64−67. measurements were conducted with a Lakeshore probe station (12) Bonaccorso, F.; Sun, Z.; Hasan, T.; Ferrari, A. Nat. Photonics. (Model TTP4) and a computer-controlled analogue-to-digital 2010, 4, 611−622. converter (National Instruments model 6030E). The confocal (13) Avouris, P. Nano Lett. 2010, 10, 4285−4294. micro-PL and Raman measurements were conducted on a (14) Mak, K. F.; Lee, C.; Hone, J.; Shan, J.; Heinz, T. F. Phys. Rev. Horiba LABHR confocal Raman system with 600 g/mm Lett. 2010, 105, 136805. grating, 50× diffraction-limited objective (N.A. = 0.75), with an (15) Splendiani, A.; Sun, L.; Zhang, Y.; Li, T.; Kim, J.; Chim, C.-Y.; Ar laser (514 nm) or a He−Ne laser (633 nm) excitation. The Galli, G.; Wang, F. Nano Lett. 2010, 10, 1271−1275. scanning photocurrent measurements were conducted with the (16) Ramasubramaniam, A. Phys. Rev. B 2012, 86, 115409. (17) Cao, T.; Wang, G.; Han, W.; Ye, H.; Zhu, C.; Shi, J.; Niu, Q.; same Horiba LABHR confocal Raman system combined with Tan, P.; Wang, E.; Liu, B. Nat. Commun. 2012, 3, 887. the same electrical measurement system. The EL measurements (18) Feng, J.; Qian, X.; Huang, C.-W.; Li, J. Nat. Photonics 2012, 6, were performed on a home-built confocal PL measurement 866−872. system combining with the same electrical measurement system (19) Mak, K. F.; He, K.; Shan, J.; Heinz, T. F. Nat. Nanotechnol. with a temperature control in Ar environment. Unless 2012, 7, 494−498. mentioned in the main text, all measurements were conducted (20) Xiao, D.; Liu, G.-B.; Feng, W.; Xu, X.; Yao, W. Phys. Rev. Lett. at room temperature. The EL images were collected by a 50× 2012, 108, 196802. objective (N.A. = 0.5) and captured by a liquid-nitrogen-cooled (21) Zeng, H.; Dai, J.; Yao, W.; Xiao, D.; Cui, X. Nat. Nanotechnol. CCD camera (Princeton instruments PyLoN 400F). The 2012, 7, 490−493. 5596 dx.doi.org/10.1021/nl502075n | Nano Lett. 2014, 14, 5590−5597 Nano Letters Letter (22) Jones, A. M.; Yu, H.; Ghimire, N. J.; Wu, S.; Aivazian, G.; Ross, J. S.; Zhao, B.; Yan, J.; Mandrus, D. G.; Xiao, D. Nat. Nanotechnol. 2013, 8, 634−638. (23) Wang, Q. H.; Kalantar-Zadeh, K.; Kis, A.; Coleman, J. N.; Strano, M. S. Nat. Nanotechnol. 2012, 7, 699−712. (24) Chuang, S.; Battaglia, C.; Azcatl, A.; McDonnell, S.; Kang, J. S.; Yin, X.; Tosun, M.; Kapadia, R.; Fang, H.; Wallace, R. M.; Javey, A. Nano Lett. 2014, 14, 1337−1342. (25) Sundaram, R.; Engel, M.; Lombardo, A.; Krupke, R.; Ferrari, A.; Avouris, P.; Steiner, M. Nano Lett. 2013, 13, 1416−1421. (26) Ross, J. S.; Klement, P.; Jones, A. M.; Ghimire, N. J.; Yan, J.; Mandrus, D.; Taniguchi, T.; Watanabe, K.; Kitamura, K.; Yao, W. Nat. Nanotechnol. 2014, 9, 268−272. (27) Pospischil, A.; Furchi, M. M.; Mueller, T. Nat. Nanotechnol. 2014, 9, 257−261. (28) Baugher, B. W.; Churchill, H. O.; Yang, Y.; Jarillo-Herrero, P. Nat. Nanotechnol. 2014, 9, 262−267. (29) Britnell, L.; Gorbachev, R.; Jalil, R.; Belle, B.; Schedin, F.; Mishchenko, A.; Georgiou, T.; Katsnelson, M.; Eaves, L.; Morozov, S. Science 2012, 335, 947−950. (30) Georgiou, T.; Jalil, R.; Belle, B. D.; Britnell, L.; Gorbachev, R. V.; Morozov, S. V.; Kim, Y.-J.; Gholinia, A.; Haigh, S. J.; Makarovsky, O. Nat. Nanotechnol. 2013, 8, 100−103. (31) Yang, H.; Heo, J.; Park, S.; Song, H. J.; Seo, D. H.; Byun, K.-E.; Kim, P.; Yoo, I.; Chung, H.-J.; Kim, K. Science 2012, 336, 1140−1143. (32) Yu, W. J.; Li, Z.; Zhou, H.; Chen, Y.; Wang, Y.; Huang, Y.; Duan, X. Nat. Mater. 2013, 12, 246−252. (33) Chuang, S.; Kapadia, R.; Fang, H.; Chang, T. C.; Yen, W.-C.; Chueh, Y.-L.; Javey, A. Appl. Phys. Lett. 2013, 102, 242101. (34) Jariwala, D.; Sangwan, V. K.; Wu, C.-C.; Prabhumirashi, P. L.; Geier, M. L.; Marks, T. J.; Lauhon, L. J.; Hersam, M. C. Proc. Natl. Acad. Sci. U. S. A. 2013, 110, 18076−18080. (35) Ye, Y.; Ye, Z.; Gharghi, M.; Zhu, H.; Zhao, M.; Yin, X.; Zhang, X. 2014, arXiv:1305.4235. arXiv.org e-Print archive. http://arxiv.org/ abs/1305.4235 (accessed June 2014) (36) Liu, H.; Ye, P. D. IEEE Electron Device Lett. 2012, 33, 546−548. (37) Zhao, W.; Ghorannevis, Z.; Chu, L.; Toh, M.; Kloc, C.; Tan, P.- H.; Eda, G. ACS Nano 2012, 7, 791−797. (38) Whitaker, J. C. The electronics handbook; CRC Press: Boca Raton, FL, 1996. (39) Sze, S. M.; Ng, K. K. Physics of semiconductor devices; John Wiley & Sons: Hoboken, NJ, 2006. (40) Beal, A.; Knights, J.; Liang, W. J. Phys. C: Solid State Phys. 1972, 5, 3540. (41) Bromley, R.; Murray, R.; Yoffe, A. J. Phys. C: Solid State Phys. 1972, 5, 759. (42) Wilson, J.; Yoffe, A. Adv. Phys. 1969, 18, 193−335. (43) Tongay, S.; Zhou, J.; Ataca, C.; Lo, K.; Matthews, T. S.; Li, J.; Grossman, J. C.; Wu, J. Nano Lett. 2012, 12, 5576−5580. (44) Fang, H.; Battaglia, C.; Carraro, C.; Nemsak, S.; Ozdol, B.; Kang, J. S.; Bechtel, H. A.; Desai, S. B.; Kronast, F.; Unal, A. A.; Conti, G.; Conlon, C.; Palsson, G. K.; Martin, M. C.; Minor, A. M.; Fadley, C. S.; Yablonovitch, E.; Maboudian, R.; Javey, A. Proc. Natl. Acad. Sci. U. S. A. 2014, 111, 6198−6202. (45) Rivera, P.; Schaibley, J. R.; Jones, A. M.; Ross, J. S.; Wu, S.; Aivazian, G.; Klement, P.; Ghimire, N. J.; Yan, J.; Mandrus, D. G.; Yao, W.; Xu, X. 2014, arXiv:1403.4985. arXiv.org e-Print archive. http:// arxiv.org/abs/1403.4985 (accessed June 2014). (46) Furchi, M. M.; Pospischil, A.; Libisch, F.; Burgdörfer, J.; Mueller, T. 2014, arXiv:1403.2652. arXiv.org e-Print archive. http://arxiv.org/ abs/1403.2652 (accessed June 2014). (47) Lee, C.-H.; Lee, G.-H.; van der Zande, A. M.; Chen, W.; Li, Y.; Han, M.; Cui, X.; Arefe, G.; Nuckolls, C.; Heinz, T. F.; Guo, J.; Hone, J.; Kim, P. 2014, arXiv:1403.4985. arXiv.org e-Print archive. http:// arxiv.org/abs/1305.4235 (accessed June 2014). 5597 dx.doi.org/10.1021/nl502075n | Nano Lett. 2014, 14, 5590−5597 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nano Letters Pubmed Central

Electroluminescence and Photocurrent Generation from Atomically Sharp WSe2/MoS2 Heterojunction p–n Diodes

Download PDF
8 pages

Loading next page...
 
/lp/pubmed-central/electroluminescence-and-photocurrent-generation-from-atomically-sharp-3ePlse0FcE

References (112)

Publisher
Pubmed Central
Copyright
Copyright © 2014 American Chemical Society
ISSN
1530-6984
eISSN
1530-6992
DOI
10.1021/nl502075n
Publisher site
See Article on Publisher Site

Abstract

Letter pubs.acs.org/NanoLett Terms of Use Electroluminescence and Photocurrent Generation from Atomically Sharp WSe /MoS Heterojunction p−n Diodes 2 2 † ‡ ‡ † ‡ ‡ † † Rui Cheng, Dehui Li, Hailong Zhou, Chen Wang, Anxiang Yin, Shan Jiang, Yuan Liu, Yu Chen, ,†,§ ,‡,§ Yu Huang,* and Xiangfeng Duan* Department of Materials Science and Engineering, University of California, Los Angeles, California 90095, United States Department of Chemistry and Biochemistry, University of California, Los Angeles, California 90095, United States California Nanosystems Institute, University of California, Los Angeles, California 90095, United States * Supporting Information ABSTRACT: The p−n diodes represent the most fundamental device building blocks for diverse optoelectronic functions, but are difficult to achieve in atomically thin transition metal dichalcogenides (TMDs) due to the challenges in selectively doping them into p-or n-type semiconductors. Here, we demonstrate that an atomically thin and sharp heterojunction p−n diode can be created by vertically stacking p-type monolayer tungsten diselenide (WSe ) and n-type few-layer molybdenum disulfide (MoS ). 2 2 Electrical measurements of the vertically staked WSe /MoS heterojunctions reveal excellent current rectification behavior with 2 2 an ideality factor of 1.2. Photocurrent mapping shows rapid photoresponse over the entire overlapping region with a highest external quantum efficiency up to 12%. Electroluminescence studies show prominent band edge excitonic emission and strikingly enhanced hot-electron luminescence. A systematic investigation shows distinct layer-number dependent emission characteristics and reveals important insight about the origin of hot-electron luminescence and the nature of electron−orbital interaction in TMDs. We believe that these atomically thin heterojunction p−n diodes represent an interesting system for probing the fundamental electro-optical properties in TMDs and can open up a new pathway to novel optoelectronic devices such as atomically thin photodetectors, photovoltaics, as well as spin- and valley-polarized light emitting diodes, on-chip lasers. KEYWORDS: WSe , MoS , heterojunction, electroluminescence, photocurrent, van der Waals 2 2 wo-dimensional layered materials, such as graphene, from ML-MoS has been reported in a metal-MoS Schottky 2 2 T MoS , and WSe , are emerging as an exciting material junction through a hot carrier process. Electrostatic doping 2 2 system for a new generation of atomically thin optoelectronics, has also been used to create planar p−n diodes, but usually with 1−7 8 9 including photodetectors, ultrafast lasers, polarizers, touch relatively gradual doping profile (limited by the fringe electrical 10 11 panels, and optical modulators due to their unique field) and typically relatively low optoelectronic efficiency (e.g., 12−23 electronic and optical properties. In this regard, the photon to electron conversion external quantum efficiency monolayer transition metal dichalcogenides (ML-TMDs) is 26−28 (EQE) ∼0.1−1%). particularly interesting due to their direct energy bandgap and The atomically thin geometry of these 2D materials can allow 12,13 the non-centrosymmetric lattice structure. The p−n diodes band structure modulation in a vertically stacked hetero- represent the most fundamental device building blocks for most structures to form atomically sharp heterojunctions. For optoelectronic functions, including photodiodes and light example, this strategy allows gapless graphene to be used in emitting diodes. However, it is particularly difficult to create p−n diodes in atomically thin TMDs due to the challenges in selectively doping them into p- or n-type semiconductors. Received: June 4, 2014 Contact engineering has been explored to create p−n diodes in Revised: August 15, 2014 TMD based layered materials. Electroluminescence (EL) Published: August 26, 2014 © 2014 American Chemical Society 5590 dx.doi.org/10.1021/nl502075n | Nano Lett. 2014, 14, 5590−5597 Nano Letters Letter Figure 1. Schematic illustration and band diagram of the WSe /MoS vertical heterojunction p−n diode. (a) A schematic illustration of the WSe / 2 2 2 MoS vertical heterojunction device shows that a transferred MoS flake on synthetic WSe forms a vertical heterojunction. (b) A schematic 2 2 2 illustration the cross-sectional view of the WSe /MoS vertical heterojunction device. (c) The ideal band diagram of WSe /MoS heterojunction p−n 2 2 2 2 diode under zero bias. Figure 2. Structural characterization of the WSe /MoS heterojunction p−n diode. (a) Optical microscopy image of a truncated triangular domain of 2 2 monolayer WSe with an inverted triangular bilayer region at the center. (b) The false color SEM image of the WSe /MoS vertical heterojunction 2 2 2 device, with ML-WSe highlighted by blue color, BL-WSe area by violet color, MoS by green color, and metal electrodes by golden color. The scale 2 2 2 bar is 3 μm. (c) The PL mapping of the WSe /MoS heterojunction device, with red color representing the PL from MoS and the green color 2 2 2 representing PL from WSe . (d) The PL spectra of synthetic ML- and BL-WSe and few-layer MoS flakes with the A, B exciton peaks and indirect 2 2 2 transition I peak labeled. The intensities of BL-WSe2 and FL-MoS are multiplied by 10 times for better visibility. Inset, the B exciton peak in ML- WSe . (e) The cross-sectional HRTEM image of the WSe /MoS heterojunction interface. The scale bar is 5 nm. (f) The EDS element distribution 2 2 2 profile from the bottom to the top of panel e. The black square represent the distribution profile of W-L characteristic peaks. The red line represents the fitting curve for W-L distribution profile, with a full width at half-maximum of 1.2 nm, corresponding to bilayer WSe . 29,30 31 32 field-effect tunnelling devices, barristors, inverters, and shows fast photoresponse and demonstrates that the p−n photodetectors while staked together with other 2D materials junction is created throughout the entire WSe /MoS over- 2 2 in the vertical direction. Although the nearly perfect 2D lapping area. Furthermore, prominent EL is observed with rich structure and low density of states in graphene provide spectral features that can reveal important insights about advantages in some heterostructure devices, its gapless nature electron−orbital interaction in TMD based materials. prevents the formation of a large potential barrier for charge The vertical heterojunction p−n diode is formed between separation and current rectification. On the other hand, the synthetic p-type ML-WSe and exfoliated n-type MoS flake 2 2 vertical heterojunction p−n diode formed between a TMD (Figure 1a,b). Triangular domains of ML-WSe was first material and a bulk material has recently been reported, but synthesized on 300 nm Si/SiO substrate typically with a 33,34 35 usually with no EL or very weak EL. bilayer (BL) region in the center (see Figure 2a), which were Here we report an atomically thin p−n diode based on a characterized by using optical microscope, atomic force heterojunction between synthetic p-type ML-WSe and microscopy (AFM), and Raman spectroscopy (Supporting exfoliated n-type MoS flake. The atomically thin p−n diode Information Figure S1). Mechanically exfoliated MoS flakes 2 2 exhibits well-defined current rectification behavior and can were then transferred onto synthetic WSe domains to form enable efficient photocurrent generation with an EQE up to vertically stacked heterojunctions. Electron-beam lithography 12%. Unlike the planar structures where the active area is and electron beam evaporation were used to define the contact confined to the lateral interface region, photocurrent mapping electrodes. A thin Ni/Au film (5 nm/50 nm) and Au film (50 5591 dx.doi.org/10.1021/nl502075n | Nano Lett. 2014, 14, 5590−5597 Nano Letters Letter Figure 3. Electrical characterization of the WSe /MoS heterojunction p−n diode. (a) The I −V characteristics of n-type MoS FET transistor with 2 2 ds ds 2 Ni/Au (5/50 nm) contacts. (b) The I −V characteristics of p-type WSe FET transistor with Au (50 nm) contacts. (c) Gate-tunable output ds ds 2 characteristics of the WSe /MoS heterojunction p−n diode. (d) The derivation of the p−n diode ideality factor by using a model consists of an ideal 2 2 p−n diode with a series resistor. An ideality factor of 1.2 was derived with a series resistor of 80 MΩ at 0 V gate voltage (red circle), and an ideality factor of 1.3 was derived with a series resistor of 33 MΩ at −20 V gate voltage (green triangle). nm) were used as the electrode for MoS flake and WSe peak in MoS . It is also important to note that the “B” exciton 2 2 2 domain to form Ohmic contacts with minimized contact peak can be observed in both WSe (605 nm) and MoS (620 2 2 resistance and potential barrier (Figure 1b). Figure 1c shows nm), but with the intensity typically 2−3 order magnitude the ideal band diagrams of the heterojunction p−n diode at weaker than the “A” exciton peak (Figure 2d inset). zero bias. The built-in potential and applied voltage are mainly We have further characterized the atomic structure of the supported by a depletion layer with abrupt atomic boundaries, stacked heterojunction using cross-sectional transmission and outside the boundaries, the semiconductor is assumed to electron microscope (TEM) studies. The high resolution be neutral. TEM image clearly shows the WSe /MoS heterojunction 2 2 Figure 2a shows an optical microscopy image of a synthetic with a 13-layer MoS flake on top of BL-WSe (Figure 2e). 2 2 WSe domain on 300 nm Si/SiO substrate. A triangular shaped Energy dispersive X-ray spectroscopy (EDS) was further used 2 2 BL-WSe domain was typically observed in the center of the to analyze the elemental distribution across the heterojunction triangular ML-WSe domain, indicating the nearly perfect interface. An EDS elemental line scan in vertical direction lattice structure of the synthetic WSe . Figure 2b shows a top- shows a rather narrow tungsten distribution profile located near view scanning electron microscopy (SEM) image of the vertical the heterojunction interface (Figure 2f), with the full width at heterojunction. The MoS ,WSe layers and the contact half-maximum ∼1.2 nm, corresponding to BL-WSe observed 2 2 2 electrodes are labeled with different artificial colors to highlight in Figure 2e. Together, these structural and PL character- the device structure. Photoluminescence (PL) mapping was izations demonstrate that the atomically sharp heterojunctions used to illustrate the stacking structure of WSe /MoS are formed by vertically stacking atomically thin WSe and 2 2 2 heterojunction (Figure 2c). The PL mapping shows distinct MoS . PL emission from WSe (red region in Figure 2c) and MoS Before testing the electrical characteristics of the hetero- 2 2 region (green region in Figure 2c), with a consistent structure junction p−n diodes, we have first characterized the electrical layout as that observed in the SEM image (Figure 2b). The transport properties of MoS and WSe to ensure Ohmic 2 2 uniform PL from WSe and MoS also indicates the excellent contacts were achieved. To this end, the MoS and WSe field 2 2 2 2 effect transistors (FETs) were fabricated on Si/SiO substrate, quality of the TMD materials. The PL spectra of WSe show a 2 2 strong layer-number dependence (Figure 2d), with the PL with Ni/Au thin film as the source-drain contacts for MoS , and intensity in ML-WSe at least 10 times stronger than that in Au thin film as the contacts for WSe , and the silicon substrate 2 2 BL-WSe . The PL spectrum in ML-WSe shows a peak at 785 as a back gate electrodes. Figure 3a and b show the I −V 2 2 ds ds nm, corresponding to the “A” exciton peak. The PL in BL- characteristics at varying back gate voltages for MoS and WSe , 2 2 WSe also exhibits the “A” exciton peak with an additional respectively. Importantly, a linear I −V relationship is clearly 2 ds ds broad peak at ∼877 nm, which is attributed to indirect band observed for both MoS and WSe , indicating Ohmic contacts 2 2 gap emission (typically label as “I” peak). These PL studies are achieved for both materials. The formation of Ohmic are consistent with previous experimental studies and contacts for both MoS and WSe is very important because the 2 2 16,37 theoretical calculations, indicating the good crystalline Schottky barrier at the contact area may severely affect the quality of the synthetic WSe . The PL spectrum from MoS electronic and optoelectronic characteristics of the vertical 2 2 flake shows a peak at 677 nm, corresponding to “A” exciton heterojunction and could induce photocurrent generation or 5592 dx.doi.org/10.1021/nl502075n | Nano Lett. 2014, 14, 5590−5597 Nano Letters Letter Figure 4. Photoresponse of the WSe /MoS heterojunction p−n diode. (a) Optical microscpope image of the WSe /MoS heterojunction. (b) False 2 2 2 2 color scanning photocurrent micrograph of the WSe /MoS heterojunction device acquired at V = 0 V and V = 0 V under irradiation 514 nm 2 2 ds BG laser (5 μW). The purple square dotted line outlines the ML-WSe and the dark purple square dotted line outlines the BL-WSe . The blue circle 2 2 dotted line outlines the MoS and the golden solid line outlines the gold electrodes. Photocurrent were observed in the entire overlapping junction area. (c) Experimental output (I −V ) characteristic of the vertical heterojunction device in the dark (black) and under illumination (wavelength: ds ds 514 nm; power, 5 μW). Inset, temporal response of the photocurrent generation under 514 nm illumination (10 μW). (d) Power-dependent EQE of the heterojunction device under 514 and 633 nm laser excitation wavelengths at V = 0 V and V = 0 V. A maximum EQE of 12% was observed. ds BG EL at the contact region. Furthermore, I −V plots at of a p−n diode with a series resistor (Figure 3d). Importantly, ds ds varying back gate voltage show that the current increases with an ideality factor of n = 1.2 was derived with a series resistance increasing positive gate voltage for MoS , indicating an n-type of 80 MΩ, at zero gate voltage, and an ideality factor of n = 1.3 semiconductor behavior. On the contrary, the current increases was derived with a series resistance of 33 MΩ,at −20 V gate with decreasing negative gate voltage in WSe FET, consistent voltage. The achievement of ideality factor close to 1 indicates with the p-type characteristics. the excellent diode behavior of our atomically sharp With Ohmic contacts formed for both MoS and WSe ,we heterojunction p−n diode. The decrease of the series resistance 2 2 continue to probe the electrical transport properties of the with increasing negative gate voltage is also consistent with our heterojunction p−n diode. Importantly, a clear current model that the p-type WSe is the limiting series resistor at high rectification behavior is observed in (I −V ) plots for the forward bias. ds ds WSe /MoS heterojunction (Figure 3c), with current only The electrical measurements indicate excellent diode 2 2 being able to pass through the device when the p-type WSe is behavior in the atomically thin vertical heterojunction. To positively biased. The observation of current rectification clearly further characterize the diode characteristics in our vertical demonstrates a p−n diode is formed within the atomically thin heterojunction, the photocurrent mapping was carried out at WSe /MoS heterojunction. The ultrathin nature of the zero bias under a confocal microscope. Figure 4a shows an 2 2 heterojunction allows gate tunability of the diode character- optical microscope image of the WSe /MoS heterojunction 2 2 istics. The diode output characteristics (I −V ) under different depicting the relative position between WSe , MoS , and the ds ds 2 2 back gate voltage show that the output current decreases with electrodes. The corresponding photocurrent mapping at zero increasing positive gate voltage, suggesting that the p-type WSe bias with a 514 nm laser excitation (5 μW) is shown in Figure is partly limiting the charge transport in the device. 4b, with the ML-WSe region outlined by purple dotted line, The I −V output characteristics of the vertical hetero- few-layer MoS outlined by blue dotted line, and the electrodes ds ds 2 junction under forward bias can be viewed as a vertical outlined by golden solid lines. The photocurrent mapping heterojunction p−n diode in series with an additional p-type shows clear photoresponse from the entire overlapping region, FET due to the side contact on WSe . In general, the indicating the formation of p−n junction across the entire heterojunction p−n diode resistance decreases exponentially WSe /MoS overlapping area. It is also interesting to note that 2 2 with increasing bias voltage, and the series p-FET resistance is the photocurrent in ML-WSe /MoS region is much stronger 2 2 nearly constant with bias voltage. Therefore, the heterojunction than that in BL-WSe /MoS region, suggesting that the direct 2 2 resistance is dominated by p−n diode at low bias and band gap plays an important role in the photocurrent 6,7 dominated by the p-type WSe FET under high forward bias. generation process. A detailed understanding of the different We have also fitted the diode characteristics and calculated the response of ML vs BL-WSe /MoS will be an interesting topic 2 2 ideality factor of our heterojunction device based on the model for future studies. No measurable photocurrent was observed 5593 dx.doi.org/10.1021/nl502075n | Nano Lett. 2014, 14, 5590−5597 Nano Letters Letter Figure 5. Electroluminescence (EL) from the WSe /MoS heterojunction p−n diode. (a) The false color EL image of the heterojunction device 2 2 under an injection current of 100 μA. The purple dashed line outlines the ML-WSe , the blue dotted line outlines the MoS and the golden solid line 2 2 outlines the gold electrodes. (b) The EL spectra of a ML-WSe /MoS heterojunction at different injection current. (c) The EL spectra of a BL- 2 2 WSe /MoS heterojunction at different injection current. (d) The EL intensity as a function of injection current for both ML- and BL-WSe /MoS 2 2 2 2 heterojunction. (e) The ideal band diagram of the WSe /MoS heterojunction under small forward bias. The conduction band in MoS is below that 2 2 2 in WSe , the valence band in WSe is below that in MoS . At small bias, holes can go cross the junction and inject into n-type region, while the 2 2 2 electrons cannot go cross the junction. (f) The ideal band diagram of the WSe /MoS heterojunction under large forward bias. The conduction band 2 2 in MoS shifts upward and is higher than that in WSe , and the valence band in WSe is below it in MoS . At large bias, both electrons and holes can 2 2 2 2 go cross the junction and inject into the other side of the heterojunction. (g) The EL spectra of a ML-WSe /MoS heterojunction at different 2 2 temperature ranging from 25 to 75 °C. (h) The EL spectra of a BL-WSe /MoS heterojunction at different temperature ranging from 25 to 75 °C. 2 2 The injection current is fixed at 250 μA. (i) The normalized intensities of A and B′ peaks in the EL spectra of both ML- and BL-WSe as a function of temperature. from the non-overlapping regions (only WSe or MoS ) or the EQE in our vertical heterojunction can reach 11% under a 514 2 2 electrical contacts, which is expected for zero bias photocurrent nm laser excitation with a power of 5 μW. Furthermore, it is because the photogenerated carries in the regions outside p−n found that the EQE decreases with increasing excitation power junction cannot be effectively separated and extracted. (Figure 4d), with a maximum EQE of 12% observed under an The output characteristics (I −V )ofthe vertical excitation power of 0.5 μW. The decreasing EQE with ds ds heterojunction with and without laser illumination (514 nm, increasing excitation power could be attributed partly to 5 μW) show clear photovoltaic effect with an open-circuit absorption saturation in WSe and partly to the screening of the voltage of ∼0.27 V and a short-circuit current of ∼0.22 μA built-in electric field by the excited holes in the valence band of (Figure 4c). In general, the photoresponse exhibits a rapid WSe . The power dependent EQE of the same device under temporal response beyond our experimental time resolution of 633 nm excitation shows a similar trend but with generally 100 μs (Figure 4c inset), demonstrating that the photoresponse lower values than those under 514 nm excitation, which may be is originated from photocarrier generation rather than any other attributed to the spectral dependent optical absorption extrinsic effects. Based on the photocurrent response and input coefficient. It is important to note that the EQE observed in laser power, we can determine the external quantum efficiency the vertical WSe /MoS heterostructure devices is much higher 2 2 (EQE) of the photon to electron conversion. The EQE (η)is than those in lateral electrostatically doped WSe p−n 27,28 defined as the ratio of the number of carriers collected by homojunctions (0.1−3%), which may be partly attributed electrodes to the number of the incident photon, or η =(I / to more efficient charge separation resulting from an atomically ph q)/(P/hν) × 100% where I is the photocurrent, h is Planck’s sharp vertical p−n junction. In contrast, the electrostatic doping ph constant, ν is the frequency of light, q is the electron charge, would typically exhibit a spatial doping gradient, and is difficult and P is the incident light power. Our study showed that the to achieve atomically sharp junctions. 5594 dx.doi.org/10.1021/nl502075n | Nano Lett. 2014, 14, 5590−5597 Nano Letters Letter The above electrical transport and photocurrent studies HEL A′,B′ peaks, which are usually 100−1000 times weaker demonstrate excellent p−n diode characteristics in the than the A exciton peak in the PL measurements and have atomically sharp WSe /MoS heterojunction. Because p−n not been reported in EL previously. In contrast, the EL spectra 2 2 diode represents the basic device element for a light-emitting of the vertical heterojunction show that the intensities of these diode, we have further investigated the electroluminescence HEL peaks are only 3−10 times weaker than the A exciton from these heterojunction p−n diodes. Figure 5a shows an EL peak, suggesting a relative enhancement of the HEL by about 2 image acquired under a forward bias of 3 V and a forward orders of magnitude in our EL studies, which may be attributed current of ∼100 μA. The shapes of WSe , MoS , and gold to the electric field induced carrier redistribution. 2 2 electrodes were outlined in the same way as before to identify The origin of the HEL peaks in TMD materials remains the the position of the EL. In contrast to the photocurrent subject of debate and is difficult to probe due to their low 37,40−42 generation from the entire overlapping area, it is important to emission probability. The HEL peaks A′ and B′ are note that the EL is localized at the overlapping area in close generally believed to arise from the splitting of the ground and proximity to the electrodes. This can be explained by the excited states of A and B transitions due to the electron−orbital electric field distribution in the heterojunction under different interaction via either inter- or intralayer perturbation or 40,41 both. However, there is no yet clear evidence to prove bias. For photocurrent mapping at zero bias (or a small bias less than the turn on voltage), the p−n diode junction resistance which perturbation dominates the electron−orbital interaction. dominates the entire device, and therefore, photocurrent can be The emergence of intense HEL emission in our ML-WSe / seen from the entire overlapping area where there is a p−n MoS and BL-WSe /MoS heterojunction can offer a new 2 2 2 junction. For EL studies at much higher forward bias exceeding platform to probe the origin of HEL peaks and the nature of the p−n diode turn-on voltage, the resistance of the ML-WSe electron−orbital interaction in TMDs. The presence of HEL becomes an increasingly important component of the total peaks A′ and B′ in EL spectra of ML-WSe /MoS 2 2 resistance. Therefore, the most voltage drop occurs across the heterojunction (Figure 5b) indicates that intralayer perturba- heterojunction edge near the electrodes due to the large series tion plays a role in the formation of these HEL peaks. On the resistance of the ML-WSe . This is also consistent with the other hand, it is noted that the relative intensities of HEL peaks result recently reported for MoS /Si heterojunctions. (comparing with the respective A peak) in BL-WSe /MoS 2 2 2 Figure 5b and c show the EL spectra of a ML- and a BL- heterojunction (Figure 5c) are clearly much stronger than that WSe /MoS heterojunction with increasing injection current. in ML-WSe /MoS heterojunction, suggesting that interlayer 2 2 2 2 The plot of the overall EL intensity as a function of injection perturbation may also contribute to the HEL peaks (which can current shows an apparent threshold (Figure 5d), with little EL be further supported by temperature dependent studies; see below the threshold, and linear increase above threshold. The below). threshold current may be explained by the band alignment of To further probe physical mechanism governing the photon the heterojunction under different bias voltages (Figure 5e and emission process in the atomically thin p−n diode, we have also f). In general, due to different band gap and band alignment conducted the temperature dependent EL studies at 25, 50, and among the conduction band and valence band edge, the barrier 75 °C for both the ML- and BL-WSe /MoS heterojunctions 2 2 for hole transport across the junction is smaller than that for the (Figure 5g and h) and plotted the normalized peak intensities electrons. With increasing forward bias (below a certain for A and B′ peaks as a function of temperature (Figure 5i). For threshold), the holes from WSe are first injected into n-type ML-WSe /MoS heterojunction, the EL intensity of all spectral 2 2 2 MoS region, whereas few electrons can overcome the barrier peaks show a consistent decrease with increasing temperature to reach WSe (Figure 5e). Due to the nature of indirect band (Figure 5g and i), which is a common phenomenon in the LED gap in few-layer MoS , the yield of radiative recombination is devices and can be attributed to the exponential enhancement relatively low at this point. As a result, the EL intensity is very of nonradiative recombination rate with increasing temper- low when the hole injection dominates the charge transfer ature. In striking contrast, temperature dependent EL in the across the heterojunction. With further increasing bias across BL-WSe /MoS heterojunction displays highly distinct features. 2 2 the heterojunction (above electron injection threshold), the First, the A exciton peak in BL-WSe /MoS heterojunction 2 2 conduction band of MoS is shifted upward, both electrons and shows an unusual increase (instead of decrease) with increasing holes can go cross the heterojunction and are injected into p- temperature. (Figure 5h and i). This increase in A exciton type and n-type region, respectively (Figure 5f). At this point, emission may be explained by thermally decoupling neighbor- the radiative recombination in WSe dominates the EL with its ing layers via interlayer thermal expansion, which can induce a intensity increasing linearly with the injection current. It is band gap crossover from the indirect gap to the direct one with noted that the EL intensity observed in ML-WSe /MoS the increasing decoupling at higher temperature. A similar 2 2 heterojunction is much stronger than that in BL-WSe /MoS thermal decoupling effect has been observed in MoSe by PL 2 2 2 heterojunction due to the higher radiative recombination rate in studies. Second, the HEL peak B′ (and A′) shows a much direct band gap ML-WSe vs indirect bandgap BL-WSe . greater decrease with increasing temperature than that in ML 2 2 The EL spectra show rich spectral features and can be well -WSe /MoS , indicating the weakening of electron−orbital 2 2 fitted using multiple Gaussian functions (Supporting Informa- interaction with the decoupling neighboring layers. These tion Figure S2) with five main peaks, which can be assigned as temperature dependent characteristics are consistent seen in excitonic peaks A (∼792 nm) and B (∼626 nm), hot electron three devices studies and further suggest that the interlayer luminescence (HEL) peaks A′ (∼546 nm) and B′ (∼483 nm) perturbation plays an important role in electron−orbital and an indirect band gap emission peak I (∼880 nm). The A interaction in WSe , which is consistent with the observation exciton peak dominates the spectra of the EL in ML-WSe of strong interlayer excitons in TMD based heterojunc- 44,45 (Figure 5a), whereas the indirect band gap emission I is tions. significant in BL-WSe (Figure 5b). Strikingly, the EL spectra of In summary, we have fabricated WSe /MoS heterojunction 2 2 2 both ML- and BL-WSe show prominent B exciton peak and p−n diodes with atomically thin geometry and atomically sharp 5595 dx.doi.org/10.1021/nl502075n | Nano Lett. 2014, 14, 5590−5597 Nano Letters Letter interface. The scanning photocurrent measurement demon- spectra were taken by using an Acton 2300i spectrometer with strates that the p−n junction was formed over the entire 150 g/mm grating and liquid-nitrogen-cooled CCD. overlapping area with a maximum photon-to-electron con- ASSOCIATED CONTENT version EQE of 12%. The EL measurement allows for the identification of emission from different optical transitions. Hot * Supporting Information electron luminescence peaks were observed in EL spectra of The AFM and Raman characterizations of synthetic WSe WSe for the first time and used to investigate the electron− 2 domain on 300 nm Si/SiO substrate and the analysis and orbital interaction in WSe . Our novel heterojunction structure 2 peak fittings of the EL spectra. This material is available free of offers an interesting platform for fundamental investigation of charge via the Internet at http://pubs.acs.org. the microscopic nature of the carrier generation, recombination and electro-optical properties of single or few-layer TMD AUTHOR INFORMATION materials, and can open up a new pathway to novel Corresponding Authors optoelectronic devices including atomically thin photodetec- *E-mail: [email protected]. tors, photovoltaics, as well as spin- or valley-polarized light *E-mail: [email protected]. emitting diodes and on-chip lasers. Notes Note added: During the finalization of this manuscript we 46,47 The authors declare no competing financial interest. became aware of two related studies. Methods. Fabrication of the Vertical Heterostructure ACKNOWLEDGMENTS Devices. To fabricate the vertical WSe /MoS heterojunction 2 2 We acknowledge the Nanoelectronics Research Facility (NRF) devices, WSe was grown using a physical vapor deposition at UCLA for technical support. X.D. acknowledges support by process on a Si/SiO (300 nm SiO ) substrate. A total of 0.2 g 2 2 ONR Young Investigator Award N00014-12-1-0745. Y.H. WSe powder (Alfa Aesar, 13084) was added into an alumina acknowledges support by the NIH Grant 1DP2OD007279. boat as precursor. The blank Si/SiO substrates (1 cm × 5 cm) were loaded into a home-built vapor deposition system in a REFERENCES horizontal tube furnace (Lindberg/Blue M) with 1 in. quartz (1) Xia, F.; Mueller, T.; Lin, Y.-m.; Valdes-Garcia, A.; Avouris, P. Nat. tube. The system was pumped down to a vacuum of 10 mTorr Nanotechnol. 2009, 4, 839−843. in 10 min, and refilled with 150 sccm of ultrahigh purity argon (2) Mueller, T.; Xia, F.; Avouris, P. Nat. Photonics. 2010, 4, 297−301. gas (Airgas, ∼ 99.9999%) then heated to desired growth (3) Liu, Y.; Cheng, R.; Liao, L.; Zhou, H.; Bai, J.; Liu, G.; Liu, L.; temperature within 30 min. After that, the growth kept at the Huang, Y.; Duan, X. Nat. Commun. 2011, 2, 579. desired temperature for 30 min, and then terminated by (4) Sun, D.; Aivazian, G.; Jones, A. M.; Ross, J. S.; Yao, W.; Cobden, shutting off the power of the furnace. The sample was naturally D.; Xu, X. Nat. Nanotechnol. 2012, 7, 114−118. cooled down to ambient temperature. The MoS flakes were (5) Lopez-Sanchez, O.; Lembke, D.; Kayci, M.; Radenovic, A.; Kis, A. then exfoliated onto the WSe flakes through a micro- 2 Nat. Nanotechnol. 2013, 8, 497−501. mechanical cleavage approach. The metal electrodes (for (6) Yu, W. J.; Liu, Y.; Zhou, H.; Yin, A.; Li, Z.; Huang, Y.; Duan, X. probe contact or wire bonding purposes) were patterned on Nat. Nanotechnol. 2013, 8, 952−958. (7) Britnell, L.; Ribeiro, R.; Eckmann, A.; Jalil, R.; Belle, B.; the Si/SiO substrates by using electron-beam lithography and Mishchenko, A.; Kim, Y.-J.; Gorbachev, R.; Georgiou, T.; Morozov, S. electron-beam deposition of Ti/Au (50/50 nm) thin film. Ni/ Science 2013, 340, 1311−1314. Au (5/50 nm) contact electrodes were then deposited to form (8) Sun, Z.; Hasan, T.; Torrisi, F.; Popa, D.; Privitera, G.; Wang, F.; the Ohmic contact to MoS and Au (50 nm) was deposited to Bonaccorso, F.; Basko, D. M.; Ferrari, A. C. ACS Nano 2010, 4, 803− form the Ohmic contact to WSe . Microscopic, Electrical, Optical, and Optoelectrical Char- (9) Bao, Q.; Zhang, H.; Wang, B.; Ni, Z.; Lim, C. H. Y. X.; Wang, Y.; acterizations. The microstructures and morphologies of the Tang, D. Y.; Loh, K. P. Nat. Photonics. 2011, 5, 411−415. nanostructures are characterized by a JEOL 6700 scanning (10) Bae, S.; Kim, H.; Lee, Y.; Xu, X.; Park, J.-S.; Zheng, Y.; electron microscope (SEM). The cross-sectional image of the Balakrishnan, J.; Lei, T.; Kim, H. R.; Song, Y. I. Nat. Nanotechnol. 2010, 5, 574−578. heterostructure device is obtained by an FEI Titan transmission (11) Liu, M.; Yin, X.; Ulin-Avila, E.; Geng, B.; Zentgraf, T.; Ju, L.; electron microscope (TEM). The DC electrical transport Wang, F.; Zhang, X. Nature 2011, 474,64−67. measurements were conducted with a Lakeshore probe station (12) Bonaccorso, F.; Sun, Z.; Hasan, T.; Ferrari, A. Nat. Photonics. (Model TTP4) and a computer-controlled analogue-to-digital 2010, 4, 611−622. converter (National Instruments model 6030E). The confocal (13) Avouris, P. Nano Lett. 2010, 10, 4285−4294. micro-PL and Raman measurements were conducted on a (14) Mak, K. F.; Lee, C.; Hone, J.; Shan, J.; Heinz, T. F. Phys. Rev. Horiba LABHR confocal Raman system with 600 g/mm Lett. 2010, 105, 136805. grating, 50× diffraction-limited objective (N.A. = 0.75), with an (15) Splendiani, A.; Sun, L.; Zhang, Y.; Li, T.; Kim, J.; Chim, C.-Y.; Ar laser (514 nm) or a He−Ne laser (633 nm) excitation. The Galli, G.; Wang, F. Nano Lett. 2010, 10, 1271−1275. scanning photocurrent measurements were conducted with the (16) Ramasubramaniam, A. Phys. Rev. B 2012, 86, 115409. (17) Cao, T.; Wang, G.; Han, W.; Ye, H.; Zhu, C.; Shi, J.; Niu, Q.; same Horiba LABHR confocal Raman system combined with Tan, P.; Wang, E.; Liu, B. Nat. Commun. 2012, 3, 887. the same electrical measurement system. The EL measurements (18) Feng, J.; Qian, X.; Huang, C.-W.; Li, J. Nat. Photonics 2012, 6, were performed on a home-built confocal PL measurement 866−872. system combining with the same electrical measurement system (19) Mak, K. F.; He, K.; Shan, J.; Heinz, T. F. Nat. Nanotechnol. with a temperature control in Ar environment. Unless 2012, 7, 494−498. mentioned in the main text, all measurements were conducted (20) Xiao, D.; Liu, G.-B.; Feng, W.; Xu, X.; Yao, W. Phys. Rev. Lett. at room temperature. The EL images were collected by a 50× 2012, 108, 196802. objective (N.A. = 0.5) and captured by a liquid-nitrogen-cooled (21) Zeng, H.; Dai, J.; Yao, W.; Xiao, D.; Cui, X. Nat. Nanotechnol. CCD camera (Princeton instruments PyLoN 400F). The 2012, 7, 490−493. 5596 dx.doi.org/10.1021/nl502075n | Nano Lett. 2014, 14, 5590−5597 Nano Letters Letter (22) Jones, A. M.; Yu, H.; Ghimire, N. J.; Wu, S.; Aivazian, G.; Ross, J. S.; Zhao, B.; Yan, J.; Mandrus, D. G.; Xiao, D. Nat. Nanotechnol. 2013, 8, 634−638. (23) Wang, Q. H.; Kalantar-Zadeh, K.; Kis, A.; Coleman, J. N.; Strano, M. S. Nat. Nanotechnol. 2012, 7, 699−712. (24) Chuang, S.; Battaglia, C.; Azcatl, A.; McDonnell, S.; Kang, J. S.; Yin, X.; Tosun, M.; Kapadia, R.; Fang, H.; Wallace, R. M.; Javey, A. Nano Lett. 2014, 14, 1337−1342. (25) Sundaram, R.; Engel, M.; Lombardo, A.; Krupke, R.; Ferrari, A.; Avouris, P.; Steiner, M. Nano Lett. 2013, 13, 1416−1421. (26) Ross, J. S.; Klement, P.; Jones, A. M.; Ghimire, N. J.; Yan, J.; Mandrus, D.; Taniguchi, T.; Watanabe, K.; Kitamura, K.; Yao, W. Nat. Nanotechnol. 2014, 9, 268−272. (27) Pospischil, A.; Furchi, M. M.; Mueller, T. Nat. Nanotechnol. 2014, 9, 257−261. (28) Baugher, B. W.; Churchill, H. O.; Yang, Y.; Jarillo-Herrero, P. Nat. Nanotechnol. 2014, 9, 262−267. (29) Britnell, L.; Gorbachev, R.; Jalil, R.; Belle, B.; Schedin, F.; Mishchenko, A.; Georgiou, T.; Katsnelson, M.; Eaves, L.; Morozov, S. Science 2012, 335, 947−950. (30) Georgiou, T.; Jalil, R.; Belle, B. D.; Britnell, L.; Gorbachev, R. V.; Morozov, S. V.; Kim, Y.-J.; Gholinia, A.; Haigh, S. J.; Makarovsky, O. Nat. Nanotechnol. 2013, 8, 100−103. (31) Yang, H.; Heo, J.; Park, S.; Song, H. J.; Seo, D. H.; Byun, K.-E.; Kim, P.; Yoo, I.; Chung, H.-J.; Kim, K. Science 2012, 336, 1140−1143. (32) Yu, W. J.; Li, Z.; Zhou, H.; Chen, Y.; Wang, Y.; Huang, Y.; Duan, X. Nat. Mater. 2013, 12, 246−252. (33) Chuang, S.; Kapadia, R.; Fang, H.; Chang, T. C.; Yen, W.-C.; Chueh, Y.-L.; Javey, A. Appl. Phys. Lett. 2013, 102, 242101. (34) Jariwala, D.; Sangwan, V. K.; Wu, C.-C.; Prabhumirashi, P. L.; Geier, M. L.; Marks, T. J.; Lauhon, L. J.; Hersam, M. C. Proc. Natl. Acad. Sci. U. S. A. 2013, 110, 18076−18080. (35) Ye, Y.; Ye, Z.; Gharghi, M.; Zhu, H.; Zhao, M.; Yin, X.; Zhang, X. 2014, arXiv:1305.4235. arXiv.org e-Print archive. http://arxiv.org/ abs/1305.4235 (accessed June 2014) (36) Liu, H.; Ye, P. D. IEEE Electron Device Lett. 2012, 33, 546−548. (37) Zhao, W.; Ghorannevis, Z.; Chu, L.; Toh, M.; Kloc, C.; Tan, P.- H.; Eda, G. ACS Nano 2012, 7, 791−797. (38) Whitaker, J. C. The electronics handbook; CRC Press: Boca Raton, FL, 1996. (39) Sze, S. M.; Ng, K. K. Physics of semiconductor devices; John Wiley & Sons: Hoboken, NJ, 2006. (40) Beal, A.; Knights, J.; Liang, W. J. Phys. C: Solid State Phys. 1972, 5, 3540. (41) Bromley, R.; Murray, R.; Yoffe, A. J. Phys. C: Solid State Phys. 1972, 5, 759. (42) Wilson, J.; Yoffe, A. Adv. Phys. 1969, 18, 193−335. (43) Tongay, S.; Zhou, J.; Ataca, C.; Lo, K.; Matthews, T. S.; Li, J.; Grossman, J. C.; Wu, J. Nano Lett. 2012, 12, 5576−5580. (44) Fang, H.; Battaglia, C.; Carraro, C.; Nemsak, S.; Ozdol, B.; Kang, J. S.; Bechtel, H. A.; Desai, S. B.; Kronast, F.; Unal, A. A.; Conti, G.; Conlon, C.; Palsson, G. K.; Martin, M. C.; Minor, A. M.; Fadley, C. S.; Yablonovitch, E.; Maboudian, R.; Javey, A. Proc. Natl. Acad. Sci. U. S. A. 2014, 111, 6198−6202. (45) Rivera, P.; Schaibley, J. R.; Jones, A. M.; Ross, J. S.; Wu, S.; Aivazian, G.; Klement, P.; Ghimire, N. J.; Yan, J.; Mandrus, D. G.; Yao, W.; Xu, X. 2014, arXiv:1403.4985. arXiv.org e-Print archive. http:// arxiv.org/abs/1403.4985 (accessed June 2014). (46) Furchi, M. M.; Pospischil, A.; Libisch, F.; Burgdörfer, J.; Mueller, T. 2014, arXiv:1403.2652. arXiv.org e-Print archive. http://arxiv.org/ abs/1403.2652 (accessed June 2014). (47) Lee, C.-H.; Lee, G.-H.; van der Zande, A. M.; Chen, W.; Li, Y.; Han, M.; Cui, X.; Arefe, G.; Nuckolls, C.; Heinz, T. F.; Guo, J.; Hone, J.; Kim, P. 2014, arXiv:1403.4985. arXiv.org e-Print archive. http:// arxiv.org/abs/1305.4235 (accessed June 2014). 5597 dx.doi.org/10.1021/nl502075n | Nano Lett. 2014, 14, 5590−5597

Journal

Nano Letters – Pubmed Central

Published: Aug 26, 2014

There are no references for this article.