Detailed studies of MgZnO-based metal-semiconductor-metal (MSM) solar-blind ultraviolet photodetector with different electrode structures are performed. A two-dimensional physical model is established based on the Poisson’s equation and time-dependent continuity equations, which is verified by our experimental data of conventional electrode MSM detector. The steady-state characteristics and transient response of semicircular and triangular electrode MSM detectors are also investigated by this model. Compared with the conventional electrode, semicircular and triangular electrode devices exhibit a substantial improvement on the photocurrent. At a bias of 10 V, the steady-state saturated photocurrents for semicircular and triangular electrode devices are 14.69 nA and 24.37 nA respectively, corresponding to a 20.5% and 100% increase over the conventional electrode detector. Meanwhile, the transient peak photocurrents reach 31.38 nA and 52.09 nA respectively, both of which are notably larger than that of conventional device.
©2013 Optical Society of America
In recent years, solar-blind photodetectors (SBPDs) have drawn a great deal of interest because of their wide applications in missile warning, flame monitoring and near-Earth communication [1–3]. MgZnO, with a large tunable band-gap energy (3.3-7.8eV) [4,5], high radiation hardness  and low growth temperature , is a promising material for fabricating SBPDs. In contrast to p-n, p-i-n and Schottky structures, metal-semiconductor-metal (MSM) photodetectors [8–10] are a prior choice for photodetectors due to their lower intrinsic capacitance and suitability for monolithic integration. At present, there have been a number of experimental reports on MgZnO-based MSM photodetectors for ultraviolet (UV) applications [11–13]. Unfortunately, optical losses including reflection and electrode shadowing lead to low responsivity and external quantum efficiency. So far, some studies have been carried out to improve the performance of MSM photodetectors such as utilizing ZnO or SrTiO3 buffer layer  and fabricating a device with a heterojunction . In addition, K.-W. Ang et al. adopted novel Schottky barrier enhancement layer to reduce the dark current and achieve a good spectral response and quantum efficiency [16,17]. K. Lee et al. proposed an unactivated cap layer and transparent indium tin oxide structure to obtain a lower noise level and a higher normalized detectivity . D. Crouse et al. presented a theoretical analysis on the various electromagnetic resonance effects (Horizontal surface plasmons, Cavity modes, ‘hybrid’ modes) in MSM photodetector . After that, they studied the various optical modes including horizontal surface plasmons (HSPs), vertical surface resonances (VSRs), Wood–Rayleigh anomalies (WRs), Fabry–Perot cavity resonances (FPCRs), and diffracted modes in MSM-PD to enhance the bandwidth and responsivity [20–22]. Compared with those methods mentioned above, novel electrode structures such as semicircular and triangular electrodes may provide an alternative solution to improve the optoelectronic performance of MgZnO-based MSM detector. Up to now, new types of electrode structures have been applied to GaN and SiC-based MSM detectors [23,24]. However, a systematic investigation of different electrode structures on the steady-state characteristics and transient response of MgZnO-based MSM detectors is still not available.
In this paper, a two-dimensional numerical simulation program for MgZnO-based MSM ultraviolet (UV) detector is developed based on traditional drift-diffusion model. Poisson’s equation is solved coupling with time-dependent continuity equations for electrons and holes. Numerical simulation and experimental data of conventional electrode MSM detector were compared to demonstrate the accuracy of the model. The steady-state and transient photocurrents variation of semicircular and triangular electrode MSM detectors are discussed in detail based on the physical model. The results indicate that the triangular and semicircular electrode detectors have an outstanding superiority over conventional MSM detector, and are suitable for optoelectronic integrated circuits applications.
The paper is organized as follows. In Section 2, the structure and fabrication process of the MgZnO-based MSM photodetector are briefly described. The physical model for our simulation is introduced in Section 3. Section 4, the theoretical and experimental results of the devices are presented and discussed. Finally, our conclusion is given in Section 5.
2. Device fabrication
The basic layout of MgZnO-based MSM photodetector is illustrated in Fig. 1(a). A 0.1 µm Mg0.5Zn0.5O layer was deposited on glass substrate by Radio-frequency magnetron sputtering method. The MSM structure photodetector was fabricated on this MgZnO film according to the following procedure: A top 0.2 µm Au layer was evaporated onto the MgZnO film. Interdigitated finger electrodes were obtained through conventional photolithography, metal evaporation and lift-off techniques. The finger is 3 in width, 300in length and the spacing between the fingers is 4 µm. The optical sensitive area of the photodetector is 200×300 µm2. Current-voltage characteristics of the fabricated detector were measured by using a Keithley 6514 high resistance electrometer. The photoresponse of the photodetector was measured by using a Xe lamp and a monochromator.
3. Model description
The unit cells of conventional, semicircular and triangular electrode MSM photodetectors for simulation are shown in Fig. 1(b). To understand the realistic behavior of carries and electric fields in MSM PDs, the coupled Poisson’s equation and time-dependent continuity equations are solved on a two dimensional domain [25,26], which are listed as follows:
The Shockley-Hall-Read model is used to simulate the generation-recombination process of the carriers. The optical generation rate with the consideration of the thickness of the active region can be given as27] varies strongly with the optical wavelength.
A schematic view of light reflection by semicircular and triangular electrodes is shown in Fig. 2. It is seen that a fraction of the incident light rays are reflected back to the MgZnO photosensitive area by two novel electrodes, which is equivalent to increase the effective incident light. The deep analyses of the MSM Schottky barrier photodetector in our work are based on the qualitative theoretical study of GaN-based photoconductive UV PDs . Here, we assume the incident light rays illuminating the detector are parallel. For triangular electrode, the light rays reflected to are also parallel. If the reflected ray from point just reaches point , all the rays reflected by surface of the electrode can reach , which is also applicable to surface due to the symmetry of the electrode. According to the reflection law and geometrical relationship, we can get
Similarly, for semicircular electrode, we can obtainEq. (8), which is used to describe the optical generation rate of the conventional device, the new optical generation rate can be achieved considering the contribution from the semicircular or triangular electrode reflection, which is derived asEqs. (2) and (3), the influences of novel electrode structures on the characteristics of MSM photodetector can be achieved after solving coupled Eqs. (1)–(7) along with the proper boundary conditions on a uniform two dimensional mesh. For the surface and both sides of the analyzed region, Newman conditions are applied. The Dirichlet boundary condition of potential at the Schottky contacts is expressed as28], the boundary conditions for current density at the Schottky contacts are
The finite difference method is adopted to the discretization of the Poisson’s equation, currents densities and continuity equations. The discretizated Poisson’s equation is solved using the Gaussian elimination method and then the distribution of potential is achieved. The electric field is calculated in each mesh point using the formulaEqs. (4)–(7). After that, the electron and hole concentrations at every time step are obtained from Eq. (21) and Eq. (22). The expressions are
4. Results and discussion
Figure 3 presents the measured dark characteristics of the detector. It can be seen that the dark current increases with the increasing of voltage and the detector shows a breakdown voltage up to 160V. The simulated dark current characteristics of MgZnO photodetector are in a good agreement with experimental data when a bias lower than 12V, as shown in the inset of Fig. 3. For the bias higher than 12V, the discrepancy between calculations and experimental data may be related to an idealization of numerical model with only the SRH combination considered and constant mobility adopted. Under 10V bias, the dark current of MgZnO detector is much smaller than that of ZnO detector. The lower dark current benefits from the large band gap of MgZnO . Figure 4 plots the calculated responsivity and the corresponding experimental data of MgZnO detector at 3V bias voltage, from which it can be observed they show excellent consistency in the UV range of 230-400 nm. The peak responsivity is 4.73 mA/W at 266 nm, and the cutoff wavelength is about 285 nm. The rejection ratio is over two orders, indicating a good signal-to-noise ratio of the detector. The reason for the relatively low responsivity of our device may be the smaller bias voltage of the detector and the thinner MgZnO epilayer which produces less photogenerated carriers because of insufficient photoabsorption. All in all, comparisons of the simulations and experimental results confirm the effectiveness of our present model.
To investigate the internal electric field inner the device, simulations on conventional electrode MSM photodetector were performed. Figures 5(a) and 5(b) illustrate the distributionof longitudinal and transversal components of electric field andrespectively. It is observed that and is symmetrically distributed due to uniform potential drop between the anode and cathode. Both two components of the electric field are notably larger at the corner of the two adjacent contacts and decay into the rest of the MSM structure. Because the electric field always points toward the contacts, the longitudinal component is positive near both contacts, while the transverse component changes polarity from one contact to the other. This occurrence has been found by other authors on GaAs and InP/GaInAs-based MSM detectors [30,31].
A group of photocurrent-voltage () characteristics for three kinds of MgZnO MSM detectors under illumination at 260nm are shown in Fig. 6. The curves exhibit two distinct operating regions, referred to as the photoconductive region and saturation region respectively. It can be found that for all kinds of detectors the current increases sharply under lower bias in the photoconductive region. With the increasing of applied voltage, the photocurrent achieves saturation and remains constant. Furthermore, the finger spacing is wider, the photocurrent of conventional electrode detector at a given bias is larger. It is seen that the photocurrent increases from 12.19 nA to 17.41 nA when the finger spacing increases from 2 to 5. In contrast to the conventional electrode detector, the photocurrents of new types of electrode devices are much larger with the same spacing. The photocurrents and the corresponding responsivity are 14.69 nA, 0.317A/W and 24.37 nA, 0.515 A/W, for semicircular and triangular electrode detectors with 2 finger spacing, respectively. This is due to the fact that under the same illumination the novel electrode structures will redirect some fraction of the incident ultraviolet beams to photosensitive area to increase the photocurrent. These results indicate that the electrode structure plays an important role in the device performance.
Based on the same physical model and illumination condition, the electron, hole and total photocurrents for conventional electrode device are shown in Figs. 7(a) and 7(b) at 3V bias voltage. It can be seen that the electron current decrease approximately exponentially with the variation of time. Compared with electron current, the hole current is very small and almost remains constant. Thus the total current is a little bit larger and looks much like the electron current. In this region, the electrons with higher drift velocity are collected quite quickly whereas holes collected much slowly. It is also seen that smaller finger spacing leads to more steeply decreased current and as a result shorter fall time is obtained. Hence the electrode spacing has a significant influence on the transit time for photogenerated carriers moving to the finger.
The influences of various contact electrode structures on transient response of MgZnO MSM detectors with finger spacing of 2 and 3are illustrated in Figs. 7(c) and 7(d). In each subplot, it is found that the peak photocurrents of novel electrode detectors are larger compared with the corresponding conventional electrode device. This is due to the fact that more reflected ultraviolet beams for semicircular and triangular electrode detectors can reach the photosensitive area. It is also found that three kinds of detectors have the same fall time in their response currents. Hence, the electrode structure is not the main factor affecting the fall time of the detector.
As can be seen, the semicircular and triangular electrodes can obviously enhance the photocurrents compared to the conventional electrode. From our simulation, the triangular electrode outperforms the semicircular electrode. However, as far as the fabrication process is concerned, the latter may be simpler. For semicircular electrode, based on the manufacturing process of rectangular electrode, making the cross-sectional area of the rectangular electrode equal to that of the semicircular electrode and keeping a good control of the thickness, we increase the temperature to the melting point of the electrode metal to make it melt. Then, the shape of the electrode would become semi-cylindrical automatically under the action of surface tension. But for triangular electrode, the fabrication process is more complex. The vacuum evaporation technique with two photo-resist layers of different photosensitivity could be adopted in which the triangular electrode can be formed by the metal deposition through “eaves” edge structure with very precise control of the thickness and gap width. Based on the above considerations, accurate shape and profile control for the two novel electrodes would be the key technology for further work.
In summary, we present theoretical and experimental analysis of MgZnO solar-blind MSM photodetector. The longitudinal and transversal distribution of the electric field and the current characteristics of the detector are calculated. By direct comparing the calculations with experimental data of conventional electrode device, the validity of our model is verified. Afterwards, the steady-state photocurrents and transient response of the detectors with different electrode structures are investigated. Both steady-state and transient results confirm that the semicircular and triangular electrode photodetectors can enhance the photocurrents obviously compared with conventional electrode device. For semicircular and triangular electrode devices, the steady-state photocurrents are 14.69 nA and 24.37 nA, the transient peak currents are 31.38 nA and 52.09 nA, respectively, both of them are larger than that of conventional device. The results indicate that the semicircular and triangular electrode MSM photodetectors are proper and promising candidates for optoelectronic integrated circuits applications.
Project supported by the China Postdoctoral Science Special Foundation (Grant No. 201104 659), the China Postdoctoral Science Foundation (Grant No. 20100481322), the Foundation of State Key Lab on Integrated Service Networks (Grant No. ISN1003006), and the Fundamental Research Funds for the Central Universities (Grant No. K50511010023).
References and links
1. Q. Zheng, F. Huang, K. Ding, J. Huang, D. Chen, Z. Zhan, and Z. Lin, “MgZnO-based metal-semiconductor-metal solar-blind photodetectors on ZnO substrates,” Appl. Phys. Lett. 98(22), 221112 (2011). [CrossRef]
2. Q. Zheng, F. Huang, J. Huang, Q. Hu, D. Chen, and K. Ding, “High-responsivity solar-blind photodetectors based on Mg0.46Zn0.54O thin film,” IEEE Electron Device Lett. 33(7), 1033–1035 (2012). [CrossRef]
3. R. McClintock, A. Haddadi, and M. Razeghi, “Free-space optical communication using mid-infrared or solar-blind ultraviolet sources and detectors,” Proc. SPIE 8268, 826810 (2012). [CrossRef]
5. J. Narayan, A. K. Sharma, A. Kvit, C. Jin, J. Muth, and O. Holland, “Novel cubic ZnxMg1−xO epitaxial heterostructures on Si (100) substrates,” Solid State Commun. 121(1), 9–13 (2001). [CrossRef]
6. F. Auret, S. Goodman, M. Hayes, M. Legodi, H. Van Laarhoven, and D. C. Look, “Electrical characterization of 1.8 MeV proton-bombarded ZnO,” Appl. Phys. Lett. 79(19), 3074–3076 (2001). [CrossRef]
7. S. Choopun, R. D. Vispute, W. Yang, R. P. Sharma, T. Venkatesan, and H. Shen, “Realization of band gap above 5.0 eV in metastable cubic-phase MgxZn1-xO alloy films,” Appl. Phys. Lett. 80(9), 1529–1531 (2002). [CrossRef]
8. D. Crouse, M. Arend, J. Zou, and P. Keshavareddy, “Numerical modeling of electromagnetic resonance enhanced silicon metal-semiconductor-metal photodetectors,” Opt. Express 14(6), 2047–2061 (2006). [CrossRef] [PubMed]
9. A. Müller, G. Konstantinidis, M. Dragoman, D. Neculoiu, A. Kostopoulos, M. Androulidaki, M. Kayambaki, and D. Vasilache, “GaN membrane metal-semiconductor-metal ultraviolet photodetector,” Appl. Opt. 47(10), 1453–1456 (2008). [CrossRef] [PubMed]
10. M. Mikulics, M. Marso, S. Wu, A. Fox, M. Lepsa, D. Grützmacher, R. Sobolewski, and P. Kordoˇs, “Sensitivity enhancement of metal-semiconductor-metal photodetectors on low-temperature-grown GaAs using alloyed contacts,” IEEE Photon. Technol. Lett. 20(12), 1054–1056 (2008). [CrossRef]
11. Y. Hou, Z. Mei, Z. Liu, T. Zhang, and X. Du, “Mg0.55Zn0.45O solar-blind ultraviolet detector with high photoresponse performance and large internal gain,” Appl. Phys. Lett. 98(10), 103506 (2011). [CrossRef]
12. L. Wang, Z. Ju, C. Shan, J. Zheng, D. Shen, B. Yao, D. Zhao, Z. Zhang, B. Li, and J. Zhang, “MgZnO metal–semiconductor–metal structured solar-blind photodetector with fast response,” Solid State Commun. 149(45-46), 2021–2023 (2009). [CrossRef]
13. S. Han, J. Zhang, Z. Zhang, Y. Zhao, L. Wang, J. Zheng, B. Yao, D. Zhao, and D. Shen, “Mg0.58Zn0.42O Thin Films on MgO Substrates with MgO Buffer Layer,” ACS Appl. Mater. Interfaces 2(7), 1918–1921 (2010). [CrossRef]
14. W. I. Park, G. C. Yi, and H. M. Jang, “Metalorganic vapor-phase epitaxial growth and photoluminescent properties of Zn1-xMgxO thin films,” Appl. Phys. Lett. 79(13), 2022–2024 (2001). [CrossRef]
15. Y. Zhu, H. Zhou, G. Fang, and M. Li, “Middle-ultraviolet-enhanced photodetectors based on Mg0.4Zn0.6O/ZnO homojunction with a high selectivity for 300 nm around light,” Semicond. Sci. Technol. 27(6), 065003 (2012). [CrossRef]
16. K.-W. Ang, S. Zhu, M. Yu, G.-Q. Lo, and D.-L. Kwong, “High-performance waveguided Ge-on-SOI metal–semiconductor–metal photodetectors with novel silicon–carbon (Si: C) Schottky barrier enhancement layer,” IEEE Photon. Technol. Lett. 20(9), 754–756 (2008). [CrossRef]
17. K.-W. Ang, S.-Y. Zhu, J. Wang, K.-T. Chua, M.-B. Yu, G.-Q. Lo, and D.-L. Kwong, “Novel silicon-carbon (Si: C) Schottky barrier enhancement layer for dark-current suppression in Ge-on-SOI MSM photodetectors,” IEEE Electron Device Lett. 29(7), 704–707 (2008). [CrossRef]
18. K. Lee, P. C. Chang, S. J. Chang, C. Yu, Y. Wang, and S. Wu, “GaN MSM photodetectors with an unactivated Mg-doped GaN cap layer and sputtered ITO electrodes,” J. Electrochem. Soc. 155(6), J165–J167 (2008). [CrossRef]
19. D. T. Crouse, “Surface plasmon effects in metal-semiconductor-metal photodetectors,” Proc. SPIE 5594, 45–56 (2004). [CrossRef]
20. D. Crouse and R. Solomon, “Numerical modeling of surface plasmon enhanced silicon on insulator avalanche photodiodes,” Solid-State Electron. 49(10), 1697–1701 (2005). [CrossRef]
21. D. Crouse and P. Keshavareddy, “A method for designing electromagnetic resonance enhanced silicon-on-insulator metal–semiconductor–metal photodetectors,” J. Opt. A, Pure Appl. Opt. 8(2), 175–181 (2006). [CrossRef]
22. I. Bendoym, A. B. Golovin, and D. T. Crouse, “The light filtering and guiding properties of high finesse phase resonant compound gratings,” Opt. Express 20(20), 22830–22846 (2012). [CrossRef] [PubMed]
23. H.-M. Lü, G.-D. Chen, and J.-S. Yuan, “The relation between the electrode shape and the sensitivity of photoconductive ultraviolet detectors,” Acta Photonica Sinica 35, 1052–1055 (2006).
24. B. Chen, Y.-T. Yang, C.-C. Chai, and X.-J. Zhang, “Quantitatively exploring the effect of a triangular electrode on performance enhancement in a 4H-SiC metal-semiconductor-metal ultraviolet photodetector,” Chin. Phys. Lett. 28(6), 068501 (2011). [CrossRef]
25. S. Selberherr, Analysis and Simulation of Semiconductor Devices (Springer-Verlag, 1984.)
26. A. F. Salem, A. W. Smith, and K. F. Brennan, “Theoretical study of the effect of an AlGaAs double heterostructure on metal-semiconductor-metal photodetector performance,” IEEE Trans. Electron. Dev. 41(7), 1112–1119 (1994). [CrossRef]
27. S. Han, Z. Zhang, J. Zhang, L. Wang, J. Zheng, H. Zhao, Y. Zhang, M. Jiang, S. Wang, D. Zhao, C. X. Shan, B. Li, and D. Shen, “Photoconductive gain in solar-blind ultraviolet photodetector based on Mg0.52Zn0.48O thin film,” Appl. Phys. Lett. 99(24), 242105 (2011). [CrossRef]
28. S. M. Sze, Physics of Semiconductor Devices (Wiley, 1981.)
29. J. Chen, W. Shen, N. Chen, D. Qiu, and H. Wu, “The study of composition non-uniformity in ternary MgxZn1− xO thin films,” J. Phys. Condens. Matter 15(30), L475–L482 (2003). [CrossRef]
30. S. Averine, O. Bondarenko, and R. Sachot, “High-speed limitations of the metal-semiconductor-metal photodiode structures with submicron gap between the interdigitated contacts,” Solid-State Electron. 46(12), 2045–2051 (2002). [CrossRef]
31. G. Guarino, W. Donaldson, M. Mikulics, M. Marso, P. Kordoš, and R. Sobolewski, “Finite element simulation of metal–semiconductor–metal photodetector,” Solid-State Electron. 53(10), 1144–1148 (2009). [CrossRef]