Home
Search
Collections
Journals
About
Contact us
My IOPscience
High-detectivity nanowire photodetectors governed by bulk photocurrent dynamics with thermally stable carbide contacts
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2013 Nanotechnology 24 495701 (http://iopscience.iop.org/0957-4484/24/49/495701) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 130.237.165.40 This content was downloaded on 08/09/2015 at 19:50
Please note that terms and conditions apply.
IOP PUBLISHING
NANOTECHNOLOGY
Nanotechnology 24 (2013) 495701 (7pp)
doi:10.1088/0957-4484/24/49/495701
High-detectivity nanowire photodetectors governed by bulk photocurrent dynamics with thermally stable carbide contacts Rujia Zou1 , Zhenyu Zhang2 , Junqing Hu1 , Liwen Sang3 , Yauso Koide4 and Meiyong Liao4 1
State Key Laboratory for Modification of Chemical Fibers and Polymer Materials, College of Materials Science and Engineering, Donghua University, Shanghai 201620, People’s Republic of China 2 Center of Super-Diamond and Advanced Films (COSDAF), Department of Physics and Materials Science, City University of Hong Kong, Hong Kong, People’s Republic of China 3 International Center for Young Scientists (ICYS), National Institute for Materials Science (NIMS), 1-1 Namiki, Tsukuba, Ibaraki 305-0044, Japan 4 Optical and Electronic Materials Unit, National Institute for Materials Science (NIMS), 1-1 Namiki, Tsukuba, Ibaraki 305-0044, Japan E-mail: [email protected] and [email protected]
Received 27 August 2013, in final form 5 October 2013 Published 14 November 2013 Online at stacks.iop.org/Nano/24/495701 Abstract Photodetectors fabricated from one-dimensional semiconductors are always dominated by the surface states due to their large surface-to-volume ratio. Therefore, the basic 5S requirements (high sensitivity, high signal-to-noise ratio, high spectral selectivity, high speed, and high stability) for practical photodetectors are difficult to satisfy. We report on high-temperature and high-detectivity solar-blind deep-ultraviolet (DUV) photodetectors based on β-Ga2 O3 nanowires, in which the photoresponse behavior is dominated by the bulk instead of the surface states. Ohmic contact to the β-Ga2 O3 nanowires was achieved by using a thermally stable tungsten carbide electrode. As a result, the DUV responsivity at 250 nm shows the highest values—4492 A W−1 at room temperature (RT) and 3000 A W−1 at 553 K (280 ◦ C)—among the DUV photodetectors. The detectivity is as high as 1.26 × 1016 cm Hz1/2 W−1 at RT, and still remains 4.1 × 1014 cm Hz1/2 W−1 at as high a temperature as 553 K. The photocurrent dynamics from the β-Ga2 O3 nanowire is discussed in terms of the bulk dominated photoresponse behavior. Other wide bandgap DUV detectors based on nanostructures could also be developed for high-temperature applications based on this work. S Online supplementary data available from stacks.iop.org/Nano/24/495701/mmedia (Some figures may appear in colour only in the online journal)
1. Introduction
nanowires [4], In2 Ge2 O7 nanobelts [5], or ZnS and ZnO nanowires [6], have been exploited for high-performance UV or DUV photodetection as they show enhanced photosensitivity due to their high surface-to-volume ratios and high crystal quality [7] compared to their bulk counterparts. Generally speaking, as a photodetector for practical applications, the 5S (high sensitivity, high signal-to-noise ratio, high spectral selectivity, high speed, and high stability) requirements
Deep-ultraviolet (DUV) photodetectors operating in the solar-blind range lower than 280 nm are promising in many applications, such as missile tracking, flame detection, biomedicine, and environmental monitoring [1, 2]. One-dimensional (1D) wide bandgap nanostructured semiconductors, such as ZnGa2 O4 nanowires [3], Zn2 GeO4 0957-4484/13/495701+07$33.00
1
c 2013 IOP Publishing Ltd Printed in the UK & the USA
Nanotechnology 24 (2013) 495701
R Zou et al
2. Experimental section
should be satisfied [8]. For DUV detection, one of the most important requirements is the ability to detect very weak signals at high temperatures, such as flame detection with high surface temperatures [2]. Nevertheless, high-temperature DUV detection based on nanostructured semiconductors faces a predicament. The main reason being that for nanostructurebased photodetectors, surface states always govern the photocurrent [9]. As a result, these photodetectors either suffer from instability or the 5S requirements are difficult to satisfy if the environments are harsh, such as elevated temperatures or different gas atmospheres. To resolve this difficulty, bulk dominated photocurrent dynamics is preferred. The traditional thin-film Si-based photodetector devices are susceptible to temperature fluctuations and degrade at high temperatures with a large increase of dark current due to the inherently narrow bandgap [10–12], resulting in a low detectivity. High-temperature photodetectors fabricated using thin-film wide bandgap semiconductors such as InGaN, AlN and diamond always involve time-consuming growth techniques and degraded photoresponse properties due to crystal defects [13, 14]. Therefore, the development of a high-performance DUV photodetector using nanostructured semiconductors with bulk dominated performance working at high temperatures is an attractive, yet challenging requirement. The monoclinic gallium oxide (β-Ga2 O3 ) semiconductor, with a wide bandgap of ∼4.9 eV in the DUV region, is one of the most promising candidates for high-performance DUV photodetectors at elevated temperatures. Individual β-Ga2 O3 nanostructures, such as nanobelts and nanowires, have been demonstrated by several reports to be potential solar-blind DUV photodetectors [15–20]. However, the photoresponse properties of these β-Ga2 O3 photodetectors are strongly affected by the surface states, as mentioned above. Therefore, novel strategies are required to improve the photoresponse properties of β-Ga2 O3 nanostructured photodetectors. In this work, we report on high-temperature and high-detectivity solar-blind DUV photodetectors based on an individual β-Ga2 O3 nanowire, in which the photoresponse behavior is dominated by the bulk instead of the surface states. We succeeded in the formation of Ohmic contact to the β-Ga2 O3 nanowire upon illumination by using a thermally stable tungsten carbide electrode, while the tungsten carbide electrode blocks contact to the β-Ga2 O3 nanowire in the dark [14]. This provides a low dark current with a high DUV sensitivity. The fabricated β-Ga2 O3 nanowire photodetector meets the 5S requirements even at elevated temperatures. The operating temperature reached as high as 553 K, which is the highest value up to now for high-temperature UV detectors. The DUV responsivity at 250 nm shows the highest values of 4492 A W−1 at room temperature (RT), and 3000 A W−1 at 553 K. The detectivity is as high as 1.26×1016 cm Hz1/2 W−1 at RT, and still remains 4.1 × 1014 cm Hz1/2 W−1 at as high a temperature as 553 K. The results reveal that photodetectors based on β-Ga2 O3 nanowires are excellent candidates for high-temperature applications.
All chemicals were purchased from the Sinopharm Chemical Reagent Co. (Shanghai, China) and were used without further purification. The β-Ga2 O3 nanowires were grown in a high-temperature alumina tube furnace, which is described in detail here. GaN powders were placed in an alumina boat in the central region of an alumina tube. A long alumina plate, which was ultrasonically cleaned in acetone and used as a substrate, was inserted downstream into the tube. The furnace was heated at a rate of 30 ◦ C min−1 to 900 ◦ C and kept at this temperature for 90 min, then further heated to 1400 ◦ C at the same rate and maintained at this temperature for 150 min. The whole process was carried out under a constant flow of Ar at a rate of 150 sccm and at the ambient pressure in the tube. After the reaction was terminated, the products were collected from the alumina substrate for the following experimental use. The products were characterized using a x-ray powder diffractometer (Rigaku Co., Japan), a scanning electron microscope (S4800), and a transmission electron microscope (JEM-2100F, equipped with an energy-dispersive x-ray spectrometer). The electrical and photoresponse measurements of as-fabricated β-Ga2 O3 nanowire photodetectors were carried out in air. Their current–voltage (I–V) characteristics were measured using an Advantest picoammeter (R8340A) and a direct current–voltage source (R6144) from room temperature to 553 K. The spectral response was obtained by using a standard lock-in detection technique from 480 to 210 nm with a 500 W xenon lamp. In situ heating experiments of as-grown material were performed inside a transmission electron microscope (TEM; JEM-2100F) using a Gatan heating holder (Gatan models: 901). In these in situ experiments, the hot by electric current is directly observed by TEM imaging and recorded using a CCD camera attached to the microscope; the current for heating and the heating rate (1 ◦ C min−1 ) can be easily controlled by dedicated software.
3. Results and discussion The β-Ga2 O3 nanowires were prepared by a high-temperature thermal reaction inside an alumina tube. After the synthesis, white wool-like products were obtained from the alumina substrate inside the tube furnace. A low-magnification scanning electron microscope (SEM) image shows that the synthesized products consist of a large quantity of long and straight nanowires with lengths up to several tens of micrometers, as shown in figure 1(a). A transmission electron microscope (TEM) image in figure 1(b) demonstrates that the wires have a uniform diameter along their whole length, and most of them have diameters of ∼300–600 nm. A high-resolution TEM (HRTEM) image of a β-Ga2 O3 wire in figure 1(c) reveals lattice fringes with d-spacings of 0.56 nm ¯ and 0.27 nm, which match those of the (002) and (111) lattice planes of the β-Ga2 O3 crystal, respectively; the wire edge is clean and abrupt on an atomic scale, and there are no amorphous layers covering the surface. By measuring the angles (61.6◦ ) of the reflections and the corresponding 2
Nanotechnology 24 (2013) 495701
R Zou et al
Figure 1. (a) SEM and (b) TEM images of as-synthesized β-Ga2 O3 nanowires. (c) HRTEM image showing a Ga2 O3 nanowire edged surface; the inset shows the ED pattern along the [101] axis of the β-Ga2 O3 crystal. (d) XRD patterns of the as-synthesized β-Ga2 O3 nanowires (i) and the β-Ga2 O3 powders from the JCPDS files (ii) (card no.: 43-1012), respectively.
illumination (49 µW cm−2 ) with a high photocurrent-to-dark current ratio of more than three orders of magnitude. The absolute photocurrent reaches 7.6 nA at an applied voltage of 10 V. The photoresponsivity of an individual β-Ga2 O3 nanowire, defined as the photocurrent generated per unit power of the incident light on the effective area, is as high as 4492 A W−1 at 10 V, corresponding to a photocurrent gain (G) of 22460. This is the largest value compared with the previous reported photodetectors based on β-Ga2 O3 nanobelts or nanowires (table S1 available at stacks.iop.org/Nano/24/ 495701/mmedia) [16–20]. It is noteworthy that the I–V characteristics become linear under DUV light illumination, as shown in figure 2(c). Therefore, the device behaves as a photoconductor upon illumination. This is markedly different from previous reports [16, 20], which showed blocking-type I–V behavior under illumination. The Ohmic behavior under illumination in this work is due to the sputter-deposited tungsten carbide electrical contacts. These Ohmic properties provide a high DUV responsivity with a high photocurrent gain. The relationship between the mobility–lifetime product µτ and the gain G for a photoconductor can be evaluated from the following equation [21, 22]:
spacing distances of the crystal planes from the electron diffraction (ED) pattern, as shown in the inset of figure 1(c), the reflections can be indexed to the [101] zone axis of the β-Ga2 O3 phase. Figure 1(d) shows comparative x-ray diffraction (XRD) patterns of the as-synthesized β-Ga2 O3 nanowires (i) and the β-Ga2 O3 powders (ii) from the Joint Committee on Powder Diffraction Standards (JCPDS) files (card no.: 43-1012), respectively. All the diffraction peaks from the products can be indexed to those of the monoclinic β-Ga2 O3 phase with lattice parameters a = 1.223 nm, b = 0.304 nm, c = 0.580 nm, and β = 103.7◦ . Thus, it is concluded that the as-synthesized nanowires exhibit the overallcrystal structure of the β-Ga2 O3 phase with a high phase purity of the products. The schematic configuration and the optical plan-view of a photodetector device fabricated using a Ga2 O3 nanowire are shown in figure 2(a). The β-Ga2 O3 nanowires were placed horizontally on the SiO2 /Si substrate, and the interdigital metal–semiconductor–metal (MSM) photodetectors were fabricated using a standard photolithography process with 50 nm thick tungsten carbon electrodes deposited by sputtering. The electrode gap is around 5 µm. Electrical and photoresponse measurements were carried out both in air and in a vacuum chamber (5 Pa). Current–voltage (I–V) characteristics were measured using an Advantest picoammeter (R8340A) and a dc voltage source (R6144) from room temperature (RT) to 553 K. The spectral response was obtained from 480 to 210 nm with a UV-enhanced 500 W xenon lamp. The developed photodetector shows a low dark current of an approximately pA level even at an applied voltage of 10 V, as shown in figure 2(b). The tungsten carbide electrode displays a blocking behavior in the dark. A significant increase in the current was observed under 250 nm DUV light
G = µτ V/L2 where V is the applied voltage and L is the gap between the electrodes. The product µτ was calculated to be 3.56 × 10−4 cm2 V−1 . Despite the existence of the photocurrent gain, the β-Ga2 O3 nanowire photodetector exhibits a fast response speed without exhibiting clear persistent photoconductivity (PPC). The time-dependent photoresponse was measured by periodically turning on/off the 250 nm light at different voltages, as shown in figure 2(d). As can be seen, the electrical current drops by more than two orders of magnitude 3
Nanotechnology 24 (2013) 495701
R Zou et al
Figure 2. Photoresponse properties of β-Ga2 O3 nanowire based photodetectors at RT. (a) Device geometry and practical optical image of the photodetectors. The electrode gap is around 5 µm. WC denotes tungsten carbide. (b) Dark current and photocurrent–voltage characteristics under 250 nm light illumination (49 µW cm−2 ). (c) The photocurrent–voltage characteristics showing linear Ohmic behavior. (d) Time response under 250 nm light illumination at various applied voltages. (e) Spectral response measured at 5 V. (f) Photocurrent dependence on the 250 nm light intensity, showing a linear behavior.
within 0.3 s (limitation of the measurement system) once the DUV light is mechanically turned off. High reversibility and high stability of the ‘ON’ and ‘OFF’ state was observed in the as-fabricated β-Ga2 O3 nanowire photodetector. The discrepancy between the dark current and the I–V curve reveals that there are some defects in the device. However, no clear dependence of a time response on the applied voltage was observed. Since the Ohmic behavior of tungsten carbide to the β-Ga2 O3 was observed under illumination, these interface defects must either be at a shallow energy level or their density is low. Figure 2(e) represents the spectral response of the β-Ga2 O3 nanowire photodetector measured at a bias of 5 V at RT. It can be seen that the β-Ga2 O3 nanowire device is insensitive to light with wavelengths longer than 300 nm. When the wavelength is shorter than 300 nm, the sensitivity increases gradually and reaches a maximum value at about 250 nm, but further wavelength shortening causes a drop in sensitivity. The photoresponsivity at 250 nm is more than four orders of magnitude larger than that in the solar light range, which indicates a high selectivity towards solar blindness. This value also is the highest among the nanostructured wide bandgap materials (table S1 available at stacks.iop.org/Nano/24/495701/mmedia) [3–5, 16–20]. The dominant peak at 250 nm indicates that the photocurrent mainly comes from bandgap absorption. The decreased
sensitivity at wavelengths shorter than 250 nm is attributed to enhanced absorption of high-energy photons near the surface region of the semiconducting β-Ga2 O3 material, where the recombination rate of the electron–hole pairs becomes obvious [23]. Figure 2(f) plots the photocurrent as a function of the intensity of the 250 nm DUV light at different voltages and exhibits a good linear behavior, further demonstrating the good performance of the β-Ga2 O3 nanowire photodetectors. Interest has been growing in the development of photodetectors with excellent thermal stability and reliability at high temperature. Using a hot TEM holder [24], the thermal stability of as-synthesized β-Ga2 O3 nanowires was directly observed under a TEM and recorded using a CCD camera attached to the microscope. Figure S1 (available at stacks. iop.org/Nano/24/495701/mmedia) shows consecutive TEM images of β-Ga2 O3 nanowires heated from RT to 900 ◦ C. It is found that the β-Ga2 O3 nanowires show excellent stability at temperatures as high as 900 ◦ C. To investigate the thermal stability of the β-Ga2 O3 nanowire photodetectors, I–V characteristics in the dark and under DUV light were evaluated from RT to 553 K. Figure 3(a) shows the dark I–V characteristics for various temperatures. The dark current shows only a slight increase from 0.1 to 35.9 pA as the temperature increases from 323 to 553 K. The photoresponsivity of the β-Ga2 O3 nanowire photodetector also demonstrates good thermal stability, as 4
Nanotechnology 24 (2013) 495701
R Zou et al
Figure 3. Photoresponse properties of β-Ga2 O3 nanowire based photodetectors at various elevated temperatures. (a) Dark current–voltage characteristics at various temperatures. (b) Responsivity under 250 nm light illumination for various temperatures. (c) Time response under 250 nm light illumination at various temperatures at 5 V. (e) Spectral response measured at 5 V at different temperatures.
of the present β-Ga2 O3 nanowire detector. The detectivity is still as high as 2.7 × 1015 cm Hz1/2 W−1 at 453 K, indicating an excellent thermal stability. As a result of the increased dark current with elevated temperatures, the detectivity is reduced, but retains a high value of 4.1 × 1014 cm Hz1/2 W−1 at 553 K—still three or four orders of magnitude higher than that of normal photodetectors at RT [21, 27–29]. The fabricated β-Ga2 O3 nanowire photodetector also shows a fast response at elevated temperatures. Figure 3(c) shows the plot of the time response at a bias of 5 V at temperatures of 298, 323, 373, 423, 453, and 503 K. As can be seen, the time response is thermally stable with elevated temperatures. The response properties show only a slight PPC which is independent of the temperature. For example, at temperatures as high as 503 K, on turning off the 250 nm DUV light, the electrical current drops by nearly two orders of magnitude within the limitation time of 0.3 s for the measurement system. The possible origin of the slight PPC is the surface traps of the Ga2 O3 nanowires, which display a slight dependence on the temperature. This also suggests that the photocurrent dynamics of the β-Ga2 O3 nanowire photodetector is governed by its bulk behavior. In addition to the thermal stability in the electrical properties and response time, the β-Ga2 O3 nanowire photodetector also shows thermal stability in the high DUV/visible light discrimination ratio. Figure 3(d) is the spectral response at a bias of 5 V measured at room temperature, 423 K, and 503 K. The discrimination ratio between 250 nm and 480 nm is more than 2 × 104 at RT. This value remains higher than 104 at temperatures as high as 423 K and 2 × 103 at 503 K. This temperature-dependent property indicates good thermal stability of the β-Ga2 O3 nanowire photodetector, and such a high discrimination
shown in figure 3(b). The responsivity under 250 nm DUV light illumination decreases slightly from 832 A W−1 at RT to 667 A W−1 at 453 K and then to 535 A W−1 at 553 K at a bias of 2 V. The responsivity at 10 V is still as high as 2786 A W−1 up to 553 K. The photocurrent gain at 553 K biased at 10 V is 13 820, corresponding to an external quantum efficiency (EQE) of 1.38 × 106 %. These values at high temperatures are much higher even than those of nanostructured Ga2 O3 photodetectors at RT, as shown in table S1 (available at stacks.iop.org/Nano/24/ 495701/mmedia) [16–19]. We should mention that a small persistent photoconductivity was observed—this changes the dark current level and explains the dark current discrepancy between figures 3(a) and (c) at the same voltage. Detectivity D∗ is one of the most important indices used to characterize the performance of photodetectors. There are three contributions that limit D∗ : shot noise from the dark current, Johnson noise, and thermal noise. If the shot noise from the dark current is the dominant contribution to D∗ , the detectivity can be expressed as [25, 26]: 1/2
D∗ = R/(2qJd )1/2 = (Jp /Llight )(2qJd )
where Jp is the photocurrent, Jd is the dark current, q is the absolute value of electron charge and Llight is the intensity of the incident light. The detectivity calculated (based on the measured photocurrent, dark current, and incident light intensity) using this equation when exposed to 250 nm light with a bias of 10 V for the β-Ga2 O3 nanowire photodetector at RT is 1.26 × 1016 cm Hz1/2 W−1 . We note that this value is much higher than those of silicon, GaAs- or GaN-based thin-film photodetectors, which are usually of the order of 1010–14 cm Hz1/2 W−1 , respectively [21, 27–29]. This is mainly due to the low dark current and high DUV responsivity 5
Nanotechnology 24 (2013) 495701
R Zou et al
Figure 4. (a) Photocurrent–voltage characteristics of the β-Ga2 O3 nanowire based photodetectors at 553 K, showing an ideal Ohmic behavior. (b) Photocurrent dependence on the 250 nm light intensity at various voltages at 553 K.
Figure 5. (a) Time response under different vacuum conditions at 5 V. (b) Photocurrent dependence on the 250 nm light intensity at various voltages under atmospheric pressure and a vacuum pressure of 5 Pa at RT, respectively.
ratio guarantees realizable applications of the present photodetectors working at high temperatures. The photocurrent dependence on light intensity at high temperatures also shows similar behavior to that at RT. Figure 4(a) presents the I–V curves of the β-Ga2 O3 nanowire photodetector illuminated by 250 nm DUV light at 553 K with an intensity varying from 0.37 to 49 µW cm−2 . The Ohmic I–V behavior obviously remains without a dependence on the temperature and light intensity. It is also disclosed that the photocurrent increases with increasing light intensity, consistent with the fact that the charge carrier photo-generation efficiency is proportional to the absorbed photon flux. The photocurrent at a fixed bias, i.e., 5 and 10 V, is linearly dependent on the light intensity, as shown in figure 4(b). If we fit the photocurrent to light intensity using the power law, Ip ∼ Pθ , where θ determines the response of the photocurrent to light intensity, then θ has a value of 1, suggesting the bulk dominated photoconductive behavior of the present devices [23]. This means the DUV responsivity is constant at various light intensities, providing a simple method for DUV monitoring. This is an obvious advantage over previous reports [16–20]. As disclosed by previous reports, high-responsivity photocurrent behavior from nanostructure-based photodetectors including ZnO, ZnGa2 O4 , and β-Ga2 O3 , originated from the molecular absorption and desorption dynamics in the air [3, 6, 15]. To further distinguish between the effects from the surface and the bulk, the photoresponse under different vacuum levels was investigated. Figure 5(a) shows the time response at different vacuum levels. The photocurrent at an
applied voltage of 5 V increases only slightly from 4.6 to 5.1 nA when the vacuum level decreases from atmospheric pressure to 5 Pa, while the response speed does not obviously slow down. From this pressure dependence of the photocurrent, we suppose that the photocurrent of the present β-Ga2 O3 nanowire photodetector includes photo-generated carriers from both the bulk and the surface. Since the photocurrent increases only slightly, bulk photo-generated carriers dominate the photoresponse behavior. This conclusion is supported by the temperature-dependent photoresponse behavior shown in figure 3. The reason is as follows. If air desorption is the main mechanism for the photocurrent dynamics, the photocurrent should increase at elevated temperature due to the reduction of recombination centers at the surface. In contrast, the DUV responsivity decreases at elevated temperatures, suggesting the recombination centers in the bulk are activated. We also note that the light intensity dependence on the photocurrent at different voltages is still almost the same regardless of the vacuum conditions, as shown in figure 5(b). This again supports the photocurrent transport within the β-Ga2 O3 nanowire structure being mainly from the bulk. From the above results, the β-Ga2 O3 nanowire based device shows a high sensitivity (4492 A W−1 @ 10 V at RT), high signal current-to-dark current (more than three orders of magnitude at 10 V), high spectral selectivity (2 × 104 at RT), high speed (
Search
Collections
Journals
About
Contact us
My IOPscience
High-detectivity nanowire photodetectors governed by bulk photocurrent dynamics with thermally stable carbide contacts
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2013 Nanotechnology 24 495701 (http://iopscience.iop.org/0957-4484/24/49/495701) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 130.237.165.40 This content was downloaded on 08/09/2015 at 19:50
Please note that terms and conditions apply.
IOP PUBLISHING
NANOTECHNOLOGY
Nanotechnology 24 (2013) 495701 (7pp)
doi:10.1088/0957-4484/24/49/495701
High-detectivity nanowire photodetectors governed by bulk photocurrent dynamics with thermally stable carbide contacts Rujia Zou1 , Zhenyu Zhang2 , Junqing Hu1 , Liwen Sang3 , Yauso Koide4 and Meiyong Liao4 1
State Key Laboratory for Modification of Chemical Fibers and Polymer Materials, College of Materials Science and Engineering, Donghua University, Shanghai 201620, People’s Republic of China 2 Center of Super-Diamond and Advanced Films (COSDAF), Department of Physics and Materials Science, City University of Hong Kong, Hong Kong, People’s Republic of China 3 International Center for Young Scientists (ICYS), National Institute for Materials Science (NIMS), 1-1 Namiki, Tsukuba, Ibaraki 305-0044, Japan 4 Optical and Electronic Materials Unit, National Institute for Materials Science (NIMS), 1-1 Namiki, Tsukuba, Ibaraki 305-0044, Japan E-mail: [email protected] and [email protected]
Received 27 August 2013, in final form 5 October 2013 Published 14 November 2013 Online at stacks.iop.org/Nano/24/495701 Abstract Photodetectors fabricated from one-dimensional semiconductors are always dominated by the surface states due to their large surface-to-volume ratio. Therefore, the basic 5S requirements (high sensitivity, high signal-to-noise ratio, high spectral selectivity, high speed, and high stability) for practical photodetectors are difficult to satisfy. We report on high-temperature and high-detectivity solar-blind deep-ultraviolet (DUV) photodetectors based on β-Ga2 O3 nanowires, in which the photoresponse behavior is dominated by the bulk instead of the surface states. Ohmic contact to the β-Ga2 O3 nanowires was achieved by using a thermally stable tungsten carbide electrode. As a result, the DUV responsivity at 250 nm shows the highest values—4492 A W−1 at room temperature (RT) and 3000 A W−1 at 553 K (280 ◦ C)—among the DUV photodetectors. The detectivity is as high as 1.26 × 1016 cm Hz1/2 W−1 at RT, and still remains 4.1 × 1014 cm Hz1/2 W−1 at as high a temperature as 553 K. The photocurrent dynamics from the β-Ga2 O3 nanowire is discussed in terms of the bulk dominated photoresponse behavior. Other wide bandgap DUV detectors based on nanostructures could also be developed for high-temperature applications based on this work. S Online supplementary data available from stacks.iop.org/Nano/24/495701/mmedia (Some figures may appear in colour only in the online journal)
1. Introduction
nanowires [4], In2 Ge2 O7 nanobelts [5], or ZnS and ZnO nanowires [6], have been exploited for high-performance UV or DUV photodetection as they show enhanced photosensitivity due to their high surface-to-volume ratios and high crystal quality [7] compared to their bulk counterparts. Generally speaking, as a photodetector for practical applications, the 5S (high sensitivity, high signal-to-noise ratio, high spectral selectivity, high speed, and high stability) requirements
Deep-ultraviolet (DUV) photodetectors operating in the solar-blind range lower than 280 nm are promising in many applications, such as missile tracking, flame detection, biomedicine, and environmental monitoring [1, 2]. One-dimensional (1D) wide bandgap nanostructured semiconductors, such as ZnGa2 O4 nanowires [3], Zn2 GeO4 0957-4484/13/495701+07$33.00
1
c 2013 IOP Publishing Ltd Printed in the UK & the USA
Nanotechnology 24 (2013) 495701
R Zou et al
2. Experimental section
should be satisfied [8]. For DUV detection, one of the most important requirements is the ability to detect very weak signals at high temperatures, such as flame detection with high surface temperatures [2]. Nevertheless, high-temperature DUV detection based on nanostructured semiconductors faces a predicament. The main reason being that for nanostructurebased photodetectors, surface states always govern the photocurrent [9]. As a result, these photodetectors either suffer from instability or the 5S requirements are difficult to satisfy if the environments are harsh, such as elevated temperatures or different gas atmospheres. To resolve this difficulty, bulk dominated photocurrent dynamics is preferred. The traditional thin-film Si-based photodetector devices are susceptible to temperature fluctuations and degrade at high temperatures with a large increase of dark current due to the inherently narrow bandgap [10–12], resulting in a low detectivity. High-temperature photodetectors fabricated using thin-film wide bandgap semiconductors such as InGaN, AlN and diamond always involve time-consuming growth techniques and degraded photoresponse properties due to crystal defects [13, 14]. Therefore, the development of a high-performance DUV photodetector using nanostructured semiconductors with bulk dominated performance working at high temperatures is an attractive, yet challenging requirement. The monoclinic gallium oxide (β-Ga2 O3 ) semiconductor, with a wide bandgap of ∼4.9 eV in the DUV region, is one of the most promising candidates for high-performance DUV photodetectors at elevated temperatures. Individual β-Ga2 O3 nanostructures, such as nanobelts and nanowires, have been demonstrated by several reports to be potential solar-blind DUV photodetectors [15–20]. However, the photoresponse properties of these β-Ga2 O3 photodetectors are strongly affected by the surface states, as mentioned above. Therefore, novel strategies are required to improve the photoresponse properties of β-Ga2 O3 nanostructured photodetectors. In this work, we report on high-temperature and high-detectivity solar-blind DUV photodetectors based on an individual β-Ga2 O3 nanowire, in which the photoresponse behavior is dominated by the bulk instead of the surface states. We succeeded in the formation of Ohmic contact to the β-Ga2 O3 nanowire upon illumination by using a thermally stable tungsten carbide electrode, while the tungsten carbide electrode blocks contact to the β-Ga2 O3 nanowire in the dark [14]. This provides a low dark current with a high DUV sensitivity. The fabricated β-Ga2 O3 nanowire photodetector meets the 5S requirements even at elevated temperatures. The operating temperature reached as high as 553 K, which is the highest value up to now for high-temperature UV detectors. The DUV responsivity at 250 nm shows the highest values of 4492 A W−1 at room temperature (RT), and 3000 A W−1 at 553 K. The detectivity is as high as 1.26×1016 cm Hz1/2 W−1 at RT, and still remains 4.1 × 1014 cm Hz1/2 W−1 at as high a temperature as 553 K. The results reveal that photodetectors based on β-Ga2 O3 nanowires are excellent candidates for high-temperature applications.
All chemicals were purchased from the Sinopharm Chemical Reagent Co. (Shanghai, China) and were used without further purification. The β-Ga2 O3 nanowires were grown in a high-temperature alumina tube furnace, which is described in detail here. GaN powders were placed in an alumina boat in the central region of an alumina tube. A long alumina plate, which was ultrasonically cleaned in acetone and used as a substrate, was inserted downstream into the tube. The furnace was heated at a rate of 30 ◦ C min−1 to 900 ◦ C and kept at this temperature for 90 min, then further heated to 1400 ◦ C at the same rate and maintained at this temperature for 150 min. The whole process was carried out under a constant flow of Ar at a rate of 150 sccm and at the ambient pressure in the tube. After the reaction was terminated, the products were collected from the alumina substrate for the following experimental use. The products were characterized using a x-ray powder diffractometer (Rigaku Co., Japan), a scanning electron microscope (S4800), and a transmission electron microscope (JEM-2100F, equipped with an energy-dispersive x-ray spectrometer). The electrical and photoresponse measurements of as-fabricated β-Ga2 O3 nanowire photodetectors were carried out in air. Their current–voltage (I–V) characteristics were measured using an Advantest picoammeter (R8340A) and a direct current–voltage source (R6144) from room temperature to 553 K. The spectral response was obtained by using a standard lock-in detection technique from 480 to 210 nm with a 500 W xenon lamp. In situ heating experiments of as-grown material were performed inside a transmission electron microscope (TEM; JEM-2100F) using a Gatan heating holder (Gatan models: 901). In these in situ experiments, the hot by electric current is directly observed by TEM imaging and recorded using a CCD camera attached to the microscope; the current for heating and the heating rate (1 ◦ C min−1 ) can be easily controlled by dedicated software.
3. Results and discussion The β-Ga2 O3 nanowires were prepared by a high-temperature thermal reaction inside an alumina tube. After the synthesis, white wool-like products were obtained from the alumina substrate inside the tube furnace. A low-magnification scanning electron microscope (SEM) image shows that the synthesized products consist of a large quantity of long and straight nanowires with lengths up to several tens of micrometers, as shown in figure 1(a). A transmission electron microscope (TEM) image in figure 1(b) demonstrates that the wires have a uniform diameter along their whole length, and most of them have diameters of ∼300–600 nm. A high-resolution TEM (HRTEM) image of a β-Ga2 O3 wire in figure 1(c) reveals lattice fringes with d-spacings of 0.56 nm ¯ and 0.27 nm, which match those of the (002) and (111) lattice planes of the β-Ga2 O3 crystal, respectively; the wire edge is clean and abrupt on an atomic scale, and there are no amorphous layers covering the surface. By measuring the angles (61.6◦ ) of the reflections and the corresponding 2
Nanotechnology 24 (2013) 495701
R Zou et al
Figure 1. (a) SEM and (b) TEM images of as-synthesized β-Ga2 O3 nanowires. (c) HRTEM image showing a Ga2 O3 nanowire edged surface; the inset shows the ED pattern along the [101] axis of the β-Ga2 O3 crystal. (d) XRD patterns of the as-synthesized β-Ga2 O3 nanowires (i) and the β-Ga2 O3 powders from the JCPDS files (ii) (card no.: 43-1012), respectively.
illumination (49 µW cm−2 ) with a high photocurrent-to-dark current ratio of more than three orders of magnitude. The absolute photocurrent reaches 7.6 nA at an applied voltage of 10 V. The photoresponsivity of an individual β-Ga2 O3 nanowire, defined as the photocurrent generated per unit power of the incident light on the effective area, is as high as 4492 A W−1 at 10 V, corresponding to a photocurrent gain (G) of 22460. This is the largest value compared with the previous reported photodetectors based on β-Ga2 O3 nanobelts or nanowires (table S1 available at stacks.iop.org/Nano/24/ 495701/mmedia) [16–20]. It is noteworthy that the I–V characteristics become linear under DUV light illumination, as shown in figure 2(c). Therefore, the device behaves as a photoconductor upon illumination. This is markedly different from previous reports [16, 20], which showed blocking-type I–V behavior under illumination. The Ohmic behavior under illumination in this work is due to the sputter-deposited tungsten carbide electrical contacts. These Ohmic properties provide a high DUV responsivity with a high photocurrent gain. The relationship between the mobility–lifetime product µτ and the gain G for a photoconductor can be evaluated from the following equation [21, 22]:
spacing distances of the crystal planes from the electron diffraction (ED) pattern, as shown in the inset of figure 1(c), the reflections can be indexed to the [101] zone axis of the β-Ga2 O3 phase. Figure 1(d) shows comparative x-ray diffraction (XRD) patterns of the as-synthesized β-Ga2 O3 nanowires (i) and the β-Ga2 O3 powders (ii) from the Joint Committee on Powder Diffraction Standards (JCPDS) files (card no.: 43-1012), respectively. All the diffraction peaks from the products can be indexed to those of the monoclinic β-Ga2 O3 phase with lattice parameters a = 1.223 nm, b = 0.304 nm, c = 0.580 nm, and β = 103.7◦ . Thus, it is concluded that the as-synthesized nanowires exhibit the overallcrystal structure of the β-Ga2 O3 phase with a high phase purity of the products. The schematic configuration and the optical plan-view of a photodetector device fabricated using a Ga2 O3 nanowire are shown in figure 2(a). The β-Ga2 O3 nanowires were placed horizontally on the SiO2 /Si substrate, and the interdigital metal–semiconductor–metal (MSM) photodetectors were fabricated using a standard photolithography process with 50 nm thick tungsten carbon electrodes deposited by sputtering. The electrode gap is around 5 µm. Electrical and photoresponse measurements were carried out both in air and in a vacuum chamber (5 Pa). Current–voltage (I–V) characteristics were measured using an Advantest picoammeter (R8340A) and a dc voltage source (R6144) from room temperature (RT) to 553 K. The spectral response was obtained from 480 to 210 nm with a UV-enhanced 500 W xenon lamp. The developed photodetector shows a low dark current of an approximately pA level even at an applied voltage of 10 V, as shown in figure 2(b). The tungsten carbide electrode displays a blocking behavior in the dark. A significant increase in the current was observed under 250 nm DUV light
G = µτ V/L2 where V is the applied voltage and L is the gap between the electrodes. The product µτ was calculated to be 3.56 × 10−4 cm2 V−1 . Despite the existence of the photocurrent gain, the β-Ga2 O3 nanowire photodetector exhibits a fast response speed without exhibiting clear persistent photoconductivity (PPC). The time-dependent photoresponse was measured by periodically turning on/off the 250 nm light at different voltages, as shown in figure 2(d). As can be seen, the electrical current drops by more than two orders of magnitude 3
Nanotechnology 24 (2013) 495701
R Zou et al
Figure 2. Photoresponse properties of β-Ga2 O3 nanowire based photodetectors at RT. (a) Device geometry and practical optical image of the photodetectors. The electrode gap is around 5 µm. WC denotes tungsten carbide. (b) Dark current and photocurrent–voltage characteristics under 250 nm light illumination (49 µW cm−2 ). (c) The photocurrent–voltage characteristics showing linear Ohmic behavior. (d) Time response under 250 nm light illumination at various applied voltages. (e) Spectral response measured at 5 V. (f) Photocurrent dependence on the 250 nm light intensity, showing a linear behavior.
within 0.3 s (limitation of the measurement system) once the DUV light is mechanically turned off. High reversibility and high stability of the ‘ON’ and ‘OFF’ state was observed in the as-fabricated β-Ga2 O3 nanowire photodetector. The discrepancy between the dark current and the I–V curve reveals that there are some defects in the device. However, no clear dependence of a time response on the applied voltage was observed. Since the Ohmic behavior of tungsten carbide to the β-Ga2 O3 was observed under illumination, these interface defects must either be at a shallow energy level or their density is low. Figure 2(e) represents the spectral response of the β-Ga2 O3 nanowire photodetector measured at a bias of 5 V at RT. It can be seen that the β-Ga2 O3 nanowire device is insensitive to light with wavelengths longer than 300 nm. When the wavelength is shorter than 300 nm, the sensitivity increases gradually and reaches a maximum value at about 250 nm, but further wavelength shortening causes a drop in sensitivity. The photoresponsivity at 250 nm is more than four orders of magnitude larger than that in the solar light range, which indicates a high selectivity towards solar blindness. This value also is the highest among the nanostructured wide bandgap materials (table S1 available at stacks.iop.org/Nano/24/495701/mmedia) [3–5, 16–20]. The dominant peak at 250 nm indicates that the photocurrent mainly comes from bandgap absorption. The decreased
sensitivity at wavelengths shorter than 250 nm is attributed to enhanced absorption of high-energy photons near the surface region of the semiconducting β-Ga2 O3 material, where the recombination rate of the electron–hole pairs becomes obvious [23]. Figure 2(f) plots the photocurrent as a function of the intensity of the 250 nm DUV light at different voltages and exhibits a good linear behavior, further demonstrating the good performance of the β-Ga2 O3 nanowire photodetectors. Interest has been growing in the development of photodetectors with excellent thermal stability and reliability at high temperature. Using a hot TEM holder [24], the thermal stability of as-synthesized β-Ga2 O3 nanowires was directly observed under a TEM and recorded using a CCD camera attached to the microscope. Figure S1 (available at stacks. iop.org/Nano/24/495701/mmedia) shows consecutive TEM images of β-Ga2 O3 nanowires heated from RT to 900 ◦ C. It is found that the β-Ga2 O3 nanowires show excellent stability at temperatures as high as 900 ◦ C. To investigate the thermal stability of the β-Ga2 O3 nanowire photodetectors, I–V characteristics in the dark and under DUV light were evaluated from RT to 553 K. Figure 3(a) shows the dark I–V characteristics for various temperatures. The dark current shows only a slight increase from 0.1 to 35.9 pA as the temperature increases from 323 to 553 K. The photoresponsivity of the β-Ga2 O3 nanowire photodetector also demonstrates good thermal stability, as 4
Nanotechnology 24 (2013) 495701
R Zou et al
Figure 3. Photoresponse properties of β-Ga2 O3 nanowire based photodetectors at various elevated temperatures. (a) Dark current–voltage characteristics at various temperatures. (b) Responsivity under 250 nm light illumination for various temperatures. (c) Time response under 250 nm light illumination at various temperatures at 5 V. (e) Spectral response measured at 5 V at different temperatures.
of the present β-Ga2 O3 nanowire detector. The detectivity is still as high as 2.7 × 1015 cm Hz1/2 W−1 at 453 K, indicating an excellent thermal stability. As a result of the increased dark current with elevated temperatures, the detectivity is reduced, but retains a high value of 4.1 × 1014 cm Hz1/2 W−1 at 553 K—still three or four orders of magnitude higher than that of normal photodetectors at RT [21, 27–29]. The fabricated β-Ga2 O3 nanowire photodetector also shows a fast response at elevated temperatures. Figure 3(c) shows the plot of the time response at a bias of 5 V at temperatures of 298, 323, 373, 423, 453, and 503 K. As can be seen, the time response is thermally stable with elevated temperatures. The response properties show only a slight PPC which is independent of the temperature. For example, at temperatures as high as 503 K, on turning off the 250 nm DUV light, the electrical current drops by nearly two orders of magnitude within the limitation time of 0.3 s for the measurement system. The possible origin of the slight PPC is the surface traps of the Ga2 O3 nanowires, which display a slight dependence on the temperature. This also suggests that the photocurrent dynamics of the β-Ga2 O3 nanowire photodetector is governed by its bulk behavior. In addition to the thermal stability in the electrical properties and response time, the β-Ga2 O3 nanowire photodetector also shows thermal stability in the high DUV/visible light discrimination ratio. Figure 3(d) is the spectral response at a bias of 5 V measured at room temperature, 423 K, and 503 K. The discrimination ratio between 250 nm and 480 nm is more than 2 × 104 at RT. This value remains higher than 104 at temperatures as high as 423 K and 2 × 103 at 503 K. This temperature-dependent property indicates good thermal stability of the β-Ga2 O3 nanowire photodetector, and such a high discrimination
shown in figure 3(b). The responsivity under 250 nm DUV light illumination decreases slightly from 832 A W−1 at RT to 667 A W−1 at 453 K and then to 535 A W−1 at 553 K at a bias of 2 V. The responsivity at 10 V is still as high as 2786 A W−1 up to 553 K. The photocurrent gain at 553 K biased at 10 V is 13 820, corresponding to an external quantum efficiency (EQE) of 1.38 × 106 %. These values at high temperatures are much higher even than those of nanostructured Ga2 O3 photodetectors at RT, as shown in table S1 (available at stacks.iop.org/Nano/24/ 495701/mmedia) [16–19]. We should mention that a small persistent photoconductivity was observed—this changes the dark current level and explains the dark current discrepancy between figures 3(a) and (c) at the same voltage. Detectivity D∗ is one of the most important indices used to characterize the performance of photodetectors. There are three contributions that limit D∗ : shot noise from the dark current, Johnson noise, and thermal noise. If the shot noise from the dark current is the dominant contribution to D∗ , the detectivity can be expressed as [25, 26]: 1/2
D∗ = R/(2qJd )1/2 = (Jp /Llight )(2qJd )
where Jp is the photocurrent, Jd is the dark current, q is the absolute value of electron charge and Llight is the intensity of the incident light. The detectivity calculated (based on the measured photocurrent, dark current, and incident light intensity) using this equation when exposed to 250 nm light with a bias of 10 V for the β-Ga2 O3 nanowire photodetector at RT is 1.26 × 1016 cm Hz1/2 W−1 . We note that this value is much higher than those of silicon, GaAs- or GaN-based thin-film photodetectors, which are usually of the order of 1010–14 cm Hz1/2 W−1 , respectively [21, 27–29]. This is mainly due to the low dark current and high DUV responsivity 5
Nanotechnology 24 (2013) 495701
R Zou et al
Figure 4. (a) Photocurrent–voltage characteristics of the β-Ga2 O3 nanowire based photodetectors at 553 K, showing an ideal Ohmic behavior. (b) Photocurrent dependence on the 250 nm light intensity at various voltages at 553 K.
Figure 5. (a) Time response under different vacuum conditions at 5 V. (b) Photocurrent dependence on the 250 nm light intensity at various voltages under atmospheric pressure and a vacuum pressure of 5 Pa at RT, respectively.
ratio guarantees realizable applications of the present photodetectors working at high temperatures. The photocurrent dependence on light intensity at high temperatures also shows similar behavior to that at RT. Figure 4(a) presents the I–V curves of the β-Ga2 O3 nanowire photodetector illuminated by 250 nm DUV light at 553 K with an intensity varying from 0.37 to 49 µW cm−2 . The Ohmic I–V behavior obviously remains without a dependence on the temperature and light intensity. It is also disclosed that the photocurrent increases with increasing light intensity, consistent with the fact that the charge carrier photo-generation efficiency is proportional to the absorbed photon flux. The photocurrent at a fixed bias, i.e., 5 and 10 V, is linearly dependent on the light intensity, as shown in figure 4(b). If we fit the photocurrent to light intensity using the power law, Ip ∼ Pθ , where θ determines the response of the photocurrent to light intensity, then θ has a value of 1, suggesting the bulk dominated photoconductive behavior of the present devices [23]. This means the DUV responsivity is constant at various light intensities, providing a simple method for DUV monitoring. This is an obvious advantage over previous reports [16–20]. As disclosed by previous reports, high-responsivity photocurrent behavior from nanostructure-based photodetectors including ZnO, ZnGa2 O4 , and β-Ga2 O3 , originated from the molecular absorption and desorption dynamics in the air [3, 6, 15]. To further distinguish between the effects from the surface and the bulk, the photoresponse under different vacuum levels was investigated. Figure 5(a) shows the time response at different vacuum levels. The photocurrent at an
applied voltage of 5 V increases only slightly from 4.6 to 5.1 nA when the vacuum level decreases from atmospheric pressure to 5 Pa, while the response speed does not obviously slow down. From this pressure dependence of the photocurrent, we suppose that the photocurrent of the present β-Ga2 O3 nanowire photodetector includes photo-generated carriers from both the bulk and the surface. Since the photocurrent increases only slightly, bulk photo-generated carriers dominate the photoresponse behavior. This conclusion is supported by the temperature-dependent photoresponse behavior shown in figure 3. The reason is as follows. If air desorption is the main mechanism for the photocurrent dynamics, the photocurrent should increase at elevated temperature due to the reduction of recombination centers at the surface. In contrast, the DUV responsivity decreases at elevated temperatures, suggesting the recombination centers in the bulk are activated. We also note that the light intensity dependence on the photocurrent at different voltages is still almost the same regardless of the vacuum conditions, as shown in figure 5(b). This again supports the photocurrent transport within the β-Ga2 O3 nanowire structure being mainly from the bulk. From the above results, the β-Ga2 O3 nanowire based device shows a high sensitivity (4492 A W−1 @ 10 V at RT), high signal current-to-dark current (more than three orders of magnitude at 10 V), high spectral selectivity (2 × 104 at RT), high speed (
