pubs.acs.org/Langmuir © 2009 American Chemical Society
Ultraviolet ZnO Nanorod Photosensors Y. K. Su,†,‡,§ S. M. Peng,† L. W. Ji,*,‡,^ C. Z. Wu,^ W. B. Cheng,^ and C. H. Liu^ †
Institute of Microelectronics & Department of Electrical Engineering, National Cheng Kung University, Tainan 701, Taiwan, ‡Advanced Optoelectronic Technology Center, National Cheng Kung University, Tainan 701, Taiwan, §Department of Electrical Engineering, Kun Shan University, Tainan 710, Taiwan, and ^Institute of Electro-Optical and Materials Science, National Formosa University, Yunlin 632, Taiwan Received June 16, 2009. Revised Manuscript Received September 5, 2009
This study fabricates and characterizes ultraviolet (UV) photosensors with ZnO nanorods (NRs). The NR arrays were selectively grown in the gap between interdigitated (IDT) electrodes of devices using hydrothermal solution processes and a lithography-based technique. Compared with a conventional ZnO photosensor without NRs, the proposed UV NR photosensors have much higher photoresponse in the UV region. Additionally, the photoconductive gain of an NR photosensor increased as UV illumination time increased; it varied at 34.45-5.32 102 under illumination by 18.28 mW/cm2 optical power. Consequently, the substantial photoconductive gain can be attributed to high surface-to-volume ratio of ZnO NRs. The high density of hole-trap states on NR surfaces lead to a persistent photoconductivity (PPC) state, promoting the transport of carriers through devices.
1. Introduction Notably, ZnO is a wide-direct-gap semiconductor with a high exciton binding energy of 60 meV and bandgap energy of 3.37 eV at room temperature; thus, ZnO can be regarded as a promising photonic material to replace GaN in optoelectronic applications in the ultraviolet (UV) spectral range.1-3 Such one-dimensional (1D) materials as nanowires (NWs), nanobelts (NBs), and nanorods (NRs) have attracted considerable interest in recent years.4-6 These 1D materials present the greatest challenge to semiconductor technology, making possible fascinating novel devices. Among such 1D materials, 1D ZnO nanostructures have been utilized in various functional devices over the past decade, such as light-emitting diodes (LEDs),7 nanolasers,8 photodetectors (PDs),9-11 field-effect transistors (FETs),12,13 photovoltaic devices,14 and nanogenerators.15,16 *Corresponding author: Tel þ886-5-631-5679, Fax þ886-5-632-9257, e-mail
[email protected] or
[email protected]. (1) Look, D. C.; Reynolds, D. C.; Litton, C. W.; Jone, R. L.; Eason, D. B.; Cantwell, G. Appl. Phys. Lett. 2002, 81, 1830. (2) Ponce, F. A.; Bour, D. P. Nature 1997, 386, 351. (3) Kwon, M. K.; Kim, J. Y.; Park, I. K.; Cho, C. Y.; Byeon, C. C.; Park, S. J. Adv. Mater. 2008, 20, 1253. (4) Ji, L. W.; Fang, T. H.; Meen, T. H. Phys. Lett. A 2006, 355, 118. (5) Young, S. J.; Ji, L. W.; Chang, S. J.; Su, Y. K. J. Cryst. Growth 2006, 293, 43. (6) Ji, L. W.; Young, S. J.; Fang, T. H.; Liu, C. H. Appl. Phys. Lett. 2007, 90, 033109. (7) Jeong, M. C.; Oh, B. Y.; Ham, M. H.; Lee, S. W.; Myoung, J. M. Small 2007, 3, 568. (8) Ruhle, S.; van Vugt, L. K.; Li, H. Y.; Keizer, N. A.; Kuipers, L.; Vanmaekelbergh, D. Nano Lett. 2008, 8, 119. (9) Kind, H.; Yan, H. Q.; Messer, B.; Law, M.; Yang, P. D. Adv. Mater. 2002, 14, 158. (10) Soci, C.; Zhang, A.; Xiang, B.; Dayeh, S. A.; Aplin, D. P. R.; Park, J.; Bao, X. Y.; Lo, Y. H.; Wang, D. Nano Lett. 2007, 7, 1003. (11) Jin, Y. Z.; Wang, J. P.; Sun, B. Q.; Blakesley, J. C.; Greenham, N. C. Nano Lett. 2008, 8, 1649. (12) Jin, S.; Whang, D.; McAlpine, M. C.; Friedman, R. S.; Wu, Y.; Lieber, C. M. Nano Lett. 2004, 4, 915. (13) Keem, K.; Jeong, D. Y.; Kim, S. Nano Lett. 2006, 6, 1454. (14) Law, M.; Greene, L. E.; Radenovic, A.; Kuykendall, T.; Liphardt, J.; Yang, P. J. Phys. Chem. B 2006, 110, 22652. (15) Wang, X. D.; Song, J. H.; Liu, J.; Wang, Z. L. Science 2007, 316, 102. (16) Yang, R.; Qin, Y.; Li, C.; Dai, L.; Wang, Z. L. Appl. Phys. Lett. 2009, 94, 022905.
Langmuir 2010, 26(1), 603–606
In this work, a simple bottom-up wet chemical method, aqueous solution growth route, and photolithography technique are used to fabricate ZnO NR metal-semiconductor-metal (MSM) photosensors, which differ from existing 1D ZnO devices. A conventional ZnO film MSM photosensor was also fabricated for comparison. Such characteristics as photoresponse, timeresolved photocurrent, and the time-dependent photoconductive gain of these devices were measured under the same experimental conditions.
2. Experimental Section Before fabricating the NR device, 40 nm thick ZnO seed layers were deposited onto glass substrates using the radio-frequency (rf) magnetron sputter deposition technique. The as-grown ZnO films were annealed at 450 C for 30 min under ambient oxygen. The 120 nm thick Ag electrodes were then deposited onto the ZnO film by electron beam evaporation and served as Schottky contacts. Lithography was applied to define the interdigitated (IDT) contact pattern. The fingers of the Ag contact electrodes were 150 μm long and 10 μm wide with a 10 μm spacing. The active area of the entire device was 150 160 μm2. Photoresists were then employed to protect the electrode patterns using a lithography-based technique. The sample with photoresist-protected IDT electrodes was then immersed in a Zn(NO3)2/NH4OH aqueous solution for 4 h at 90 C. The fabricated MSM photosensors were removed from the solution, rinsed with distilled water, and dried in air. Finally, the photoresists were removed from the IDT electrode surface of the devices. Figure 1 shows the schematic structure of the proposed ZnO NR MSM photosensor. The conventional MSM photosensor with a 40 nm thick ZnO film was fabricated for comparison. Microphotoluminescence (micro-PL), X-ray diffraction (XRD), micro-Raman, and high-resolution transmission electron microscopy (HRTEM) (JEOL JEM-3010) were then utilized to characterize the optical and crystallographic properties of the asgrown ZnO NRs. Surface morphologies of the NR photosensor were characterized by field-emission scanning electron microscopy (FE-SEM) (JEOL JSM-6700F). An HP-4156C semiconductor parameter analyzer was then employed to measure current-voltage (I-V) characteristics of the proposed ZnO NR MSM photosensor. Spectral responsivity measurements were
Published on Web 11/06/2009
DOI: 10.1021/la902171j
603
Article
Su et al.
Figure 1. Schematic for the fabricated UV photosensors with ZnO NR arrays.
Figure 2. (a) XRD result of ZnO NRs on ZnO/glass substrate. The inset shows room temperature PL spectrum of ZnO NRs. (b) HRTEM image of an individual ZnO NR; inset is the corresponding SAED image. (c) Raman spectrum of the as-grown ZnO NRs. (d) Absorption spectra for ZnO films and ZnO NRs. obtained using the TRAIX 180 system with a 300 W xenon arc lamp light source and standard synchronous detection scheme.
3. Results and Discussion Figure 2a shows the PL spectrum of the ZnO NRs excited by a 325 nm He-Cd laser at room temperature; the XRD result is shown in the inset. With a strong UV emission centered at 378 nm, the ZnO NRs were preferentially oriented in the [002] direction and had good crystal quality (Figure 2a). The HRTEM image and selective area electron diffraction (SAED) pattern (Figure 2b) reveal that the ZnO NRs are structurally uniform and contain no defects such as dislocations or stacking. The lattice spacing of 0.52 nm corresponds to a d-spacing of (002) crystal planes, indicating growth of the crystalline ZnO NRs along the c-axis direction. The hexagonal wurzite ZnO belongs to the C46v space group; two formula units per primitive cell exist and all atoms occupy the sites of symmetry C3v. Group theory indicates that the single crystalline ZnO structure should have eight sets of optical phonon modes at the Γ point of the Brillouin zone (2A1 þ 2E1 þ 2B1 þ 2E2). Among these optical phonon modes, one set of A1 and E1 modes are acoustic, and the remaining six optical modes, 604 DOI: 10.1021/la902171j
A1 þ E1 þ 2B1 þ 2E2, can be observed for first-order Raman scattering. Here, the A1 and E1 modes are both Raman and infrared (IR) active, while the two E2 modes are only Raman active, and the two B1 modes are neither Raman nor IR active (silent modes).17 Additionally, the A1 and E1 modes can be split into longitudinal optical (LO) and transverse optical (TO) branches. Figure 2c shows the Raman spectrum of the as-grown ZnO NR arrays. The Raman spectrum can be found at only the E2 and A1 (LO) modes at 438 and 584 cm-1, respectively. According to Raman selection rules, only E2 and A1 (LO) modes are allowed; that is, TO (A1 and E1) modes are forbidden in this configuration. Thus, the absence of TO modes in this measurement result further confirms that the ZnO NRs are strongly oriented along the c-axis.17 Figure 2d shows the absorption spectra of the ZnO film and NRs. The NRs have higher absorption than films in the UV region, leading to higher photoresponse for such NR photosensors compared with that of conventional film photosensors. (17) Zhang, Y.; Jia, H. B.; Wang, R. M.; Chen, C. P.; Luo, X. H.; Yu, D. P.; Lee, C. J. Appl. Phys. Lett. 2003, 83, 4631.
Langmuir 2010, 26(1), 603–606
Su et al.
Article
Figure 3. (a) Cross-sectional SEM image in low magnification of the ZnO NR MSM photosensors; the inset shows high-magnification image of IDT contact pattern. (b) The FE-SEM image with 45 tilt angle for the ZnO NR devices.
Figure 3a shows a cross-sectional FE-SEM image of the ZnO NR photosensor device. The well-aligned ZnO NR arrays grew vertically on the ZnO/glass substrate (inset of Figure 3a). The average length and diameter of these ZnO NRs were estimated at 2.2 μm and 60-80 nm, respectively. The FE-SEM image with a 45 tilt angle (Figure 3b) shows the ZnO NR arrays selectively distributed among the IDT electrodes of the devices. Figure 4 shows the spectral response characteristics of the fabricated conventional ZnO film and ZnO NR MSM photosensors with the same Ag IDT electrodes, which were measured at applied bias of 5 V. Sharp cutoffs occurred at ∼370 nm for both devices. The ratio of UV to visible rejection can be defined as the responsivity measured at 360 nm divided by that at 450 nm. With this definition and a 5 V bias, the ratios of UV to visible rejection were 28.28 and 336.65 for film and NR photosensors, respectively. The responsivities of the fabricated conventional ZnO film and NR photosensors under illumination of 370 nm were 0.13 and 41.22 A/W, respectively. Furthermore, photoconductive gain (G) is defined as the ratio of the number of electrons collected per unit time (Nel) to the number of absorbed photons per unit time (Nph); this ratio can be derived as10,18 Nel 1:24 1 τ Iph ¼ ð1Þ G ¼ ¼R ¼ λ ðμmÞ η τtr qF Nph where R is the responsivity of a detector, λ is incident light wavelength, η is quantum efficiency, τ is hole (minority) lifetime, τtr is electron transit time, Iph is photocurrent, q is the elementary charge, and F is photon absorption rate. The quantum efficiency, η, as mentioned, is the number of electron-hole pairs generated per incident photon:19 η ¼ ðIph =qÞ=ðPopt =hυÞ
ð2Þ
where Iph is the photogenerated current by absorption of incident optical power, Popt, at a wavelength λ (corresponding (18) Jiang, D.; Zhang, J.; Lu, Y.; Liu, K.; Zhao, D.; Zhang, Z.; Shen, D.; Fan, X. Solid-State Electron. 2008, 52, 679. (19) Sze, S. M. Physics of Semiconductor Devices, 2nd ed.; Wiley: New York, 1981; Chapter 13. (20) Li, Q. H.; Gao, T.; Wang, Y. G.; Wang, T. H. Appl. Phys. Lett. 2005, 86, 123117.
Langmuir 2010, 26(1), 603–606
Figure 4. Measured spectral reponsivities of the conventional ZnO film and NR MSM photosensors at 5 V applied bias. The inset shows the dark current for film and NR photosensor.
to a photon energy hυ). The transit time of an electron is given by τtr ¼
L2 μe V
ð3Þ
where L, V, and μe are the IDT spacing, applied bias, and electron mobility, respectively. By applying 41.22 A/W and 370 nm to the expression, the gain is estimated at 1.38 102 by assuming η = 1 for simplicity. Such a computational result suggests that the NR device has internal gain. Figure 5a shows the photocurrent rise and decay by turning the continuous UV light (λ = 370 nm) on and off at an applied bias of 5 V for both photosensors. The time decay process follows a firstorder exponential relaxation function with an estimated time constant. When UV illumination was off, the carrier lifetime (decay time constant), τfilm, was roughly 2.42 s for the ZnO film photosensor, much shorter than that of the ZnO NR photosensor (τNR was roughly 15.44 s). The current baseline did not fully recover after switching off the UV light, and a persistent photoconductance was maintained, which was much higher than the DOI: 10.1021/la902171j
605
Article
Su et al.
The curve of photoconductive gain of the NR photosensor varied as UV illumination time increased (Figure 5b). The photoconductive gain increased from 34.45 to 5.32 102 at an optical power of 18.28 mW/cm2, 5 V applied bias, and 370 nm illumination. The photoconductive gain gradually saturated as UV illumination time increased (Figure 5b). By eq 1, the photoconduction gain, G, depends on photocurrent, Iph, whereas the photon absorption rate, F, can be a constant at an illumination of 370 nm. Although high-density hole-trap states exist on NR surfaces, researchers generally accept that oxygen molecules are adsorbed onto ZnO surfaces by capturing free electrons from the n-type ZnO [O2(g) þ e- f O2-(ad)], which has a low conductive depletion layer that forms near the surface. The electron-hole pairs are photogenerated [hυ f e- þ hþ], while the photon energy of illumination exceeds bandgap energy, Eg (hυ > Eg). The holes migrating to the surface along the potential gradient produced by band-bending discharge are negatively charged adsorbed oxygen ions [hþ þ O2-(ad) f O2(g)], and consequently, oxygen is desorbed from surfaces, thereby increasing the free carrier concentration and decreasing the width of the depletion layer.9-11 Restated, the defects state in the surface layer of ZnO NRs readily act as oxygen-related hole-trap states reaction [hþ þ O2-(ad) f O2(g)] under UV illumination.10,20,21 The photogenerated electron-hole pairs are separated by an applied bias (external electrical field), leading to a steady-state persistent photocurrent flow along the seed layer.
Figure 5. (a) Time-resolved photocurrent in response to turn off the UV light at 5 V applied bias. (b) Time dependence of the photoconductive gain under 18.28 mW/cm2 UV illumination (370 nm).
initial dark conductance (inset of Figure 4 and Figure 5a). This measurement demonstrates the existence of persistent photoconductivity (PPC) of the ZnO NR photosensors. Conversely, no obvious PPC existed during the operation of film photosensors. The effects of UV light on ZnO NRs are more complex than a simple band-to-band photoresponse because a persistent increase in electrical conductivity is typically induced and lasts for a long time after exposure to UV light and requires a simple posttreatment to recover the initial dark current value.9,10,21-23 However, the origin of this PPC phenomenon remains controversial; it may be point defects exhibiting metastable charge states or electron-hole separation related to the surface characteristics of metal oxides. Thus, a widely accepted description has not yet been presented.21-23 (21) Liao, Z. M.; Lu, Y.; Xu, J.; Zhang, J. M.; Yu, D. P. Appl. Phys. A: Mater. Sci. Process. 2009, 95, 363. (22) Lee, S. W.; Jeong, M. C.; Myoung, J. M.; Chae, G. S.; Chung, I. J. Appl. Phys. Lett. 2007, 90, 133115. (23) Prades, J. D.; Ramirez, F. H.; Diaz, R. J.; Manzanares, M.; Andreu, T.; Cirera, A.; Rodriguez, A. R.; Morante, J. R. Nanotechnology 2008, 19, 465501.
606 DOI: 10.1021/la902171j
4. Conclusions In summary, the MSM UV photosensors with ZnO NRs were fabricated and characterized. Compared with conventional MSM photosensors based on ZnO films, the photoresponsities with 5 V applied bias under 370 nm illumination were 41.22 and 0.13 A/W, and the ratios of UV to visible rejection were 336.65 and 28.28, respectively. The measured photoresponsivity values of NR photosensors are much larger than those of film photosensors. Time-resolved photocurrent measurements reveal that the carrier lifetime of film photosensors (τfilm of roughly 2.42 s) was shorter than that of NRs photosensors (τNR of roughly 15.44 s); thus, the longer carrier lifetime can be attributed to an obvious PPC effect in NR devices. Furthermore, photoconductive gain increased from 34.45 to 5.32 102 as illumination time increased at an optical power of 18.28 mW/cm2. Acknowledgment. Funding from the Advanced Optoelectronic Technology Center, National Cheng Kung University, under projects from the Ministry of Education and the National Science Council (NSC 96-2221-E-006-079-MY3) of Taiwan is gratefully acknowledged. This work was partially supported by TDPA “Lamp Development of White Light-Emitting Diode for Local Lighting” program and in part by National Science Council of the Republic of China (R.O.C.) in Taiwan under Contracts TDPA 97-EC-17-A-07-S1-105, NSC 97-2623-E-168-001-IT and NSC-982221-E-150-005-MY3.
Langmuir 2010, 26(1), 603–606
