Study of the Electrical and Structural Characteristics of ... - CiteSeerX

Journal of The Electrochemical Society, 151 共4兲 G223-G226 共2004兲

G223

0013-4651/2004/151共4兲/G223/4/$7.00 © The Electrochemical Society, Inc.

Study of the Electrical and Structural Characteristics of AlÕPt Ohmic Contacts on n-Type ZnO Epitaxial Layer Han-Ki Kim,a,b,c,z I. Adesida,a K.-K. Kim,b S.-J. Park,b and T.-Y. Seongb,* a

Department of Electrical and Computer Engineering and Micro and Nanotechnology Laboratory, University of Illinois, Urbana, Illinois 61801, USA b Department of Materials Science and Engineering, Kwangju Institute of Science and Technology, Kwangju 500-712, Korea We have investigated Al/Pt 共20/50 nm兲 ohmic contacts on a n-type zinc oxide 共ZnO:Al兲 epitaxial layer. The samples were annealed at temperatures of 300°C and 600°C for 1 min in nitrogen ambient. Current-voltage measurements indicated that the as-deposited sample was ohmic with a specific contact resistivity of 1.25 (⫾0.05) ⫻ 10⫺5 ⍀ cm2 . However, the annealing at 300°C resulted in a significantly better ohmic behavior, with a contact resistivity of 2 (⫾0.25) ⫻ 10⫺6 ⍀ cm2 . A further increase in the annealing temperature to 600°C led to a decrease in specific contact resistivity due to extensive interfacial reactions between Al and ZnO. Both Auger electron spectroscopy and glancing angle X-ray diffraction were employed to investigate the nature of the interfacial reaction between the Al/Pt and ZnO layer with increasing annealing temperature. Possible explanation is given to describe the temperature dependence of the specific contact resistivity. © 2004 The Electrochemical Society. 关DOI: 10.1149/1.1647576兴 All rights reserved. Manuscript received June 2, 2003. Available electronically February 19, 2004.

ZnO and related oxide semiconductors, such as MgZnO and CdZnO, have acquired considerable importance as the basis for a new generation of optoelectronic devices due to their wide direct bandgap (E g ⫽ 3.37 eV), and large exciton binding energy of 60 meV at room temperature, which is much higher than that of GaN 共25 meV兲 and ZnSe 共18 meV兲.1-4 ZnO is an alternate candidate for optoelectronic applications, for use in the short wavelength range 共green, blue, UV兲.5 Although great progress has been achieved in the area of ZnO-based oxide semiconductors, many problems remain, such as the difficulty of p-ZnO growth and lack of high quality ohmic or Schottky contacts for n, p-ZnO. In particular, to realize commercialized ZnO-based optoelectronic devices, high quality ohmic and Schottky contacts are required. Several groups have reported ohmic contact schemes, however, most work is the initial stages. Akane et al. investigating In ohmic contacts to undoped n-ZnO, demonstrated that a postannealing treatment for 1 min at 300°C resulted in ohmic contact with a specific contact resistivity of 7 ⫻ 10⫺1 ⍀ cm2 . 6 Kudo et al. fabricating transparent n-ZnO/p-SrCu2 O2 junction diode, employed n⫹-ZnO and ITO as n-type and p-type electrodes, respectively.2 In our previous studies, we reported on Au, Ti/Au, and Ru ohmic contact scheme with specific contact resistivity of 10⫺3 -10⫺5 ⍀ cm2 . 7-9 Recently, an Al ohmic scheme was used for a metal-semiconductor-metal device structure and p-n homojunction diodes.4,10 Although the Al have been used as n-type ohmic contact metal, there are still lack of detailed interfacial reaction study between Al and ZnO. In addition, the detailed electrical and structural examinations of the Al/Pt ohmic schemes have not been hitherto performed. In this work, we reported on electrical and structural examination of Al/Pt multilayer on Al-doped ZnO epitaxy layer with increasing annealing temperature. Using glancing angle X-ray diffraction 共GXRD兲 and Auger electron spectroscopy 共AES兲 depth profiles analysis, we suggested possible mechanism to describe temperature dependence of Al/Pt ohmic contact on n-ZnO. Experimental An Al-doped n-ZnO epitaxial layer was grown on a 共0001兲 sapphire substrate by means of a high temperature epitaxy radio frequency sputtering system 共HTE rf sputtering system: Korea vacuum-KVS-25060兲 at 800°C using a 2 in. target containing 1 wt % Al2 O3 powder 共Pure Tech兲.11 The carrier concentration and mo-

bility of the annealed layers were measured at room temperature by means of Hall effect measurements with Van der Pauw geometry. The measurements showed that the carrier concentration and Hall mobility are 2 (⫾0.5) ⫻ 1018 cm⫺3 and of the order of 60 共⫾5兲 cm2/Vs, respectively. Prior to lithography, the samples were ultrasonically degreased with acetone and methanol for 1 min in each step, and then rinsed with deionized 共DI兲 water. The native contamination layer was removed by treatment with a buffered oxide etchant 共BOE兲 solution then blown dry by nitrogen gas. The Al 共20 nm兲/Pt 共50 nm兲 bilayer was subsequently deposited n-ZnO by electron beam evaporation. The circular pad was patterned by standard photolithography techniques for measurement of specific contact resistivity using a circular-transmission line model 共CTLM兲.12 The use of CTLM structure is advantageous because no etching of ZnO is required for feature isolation. The inner dot radius was 105 ␮m, and the spacing between the inner and the outer radii were 3, 4, 6, 12, 13, 16, and 21 ␮m. The actual gap spacing was measured by scanning electron microscopy and was used in the CTLM analysis. The measured total resistance, R T , between contacts for the circular model configuration can be expressed as RT ⫽

冋冉

冊冉

R1 1 Rs 1 ln ⫹ LT ⫹ 2␲ R1 ⫺ d R1 R1 ⫺ d

冊册

关1兴

where R s is the sheet resistance of the materials, R 1 the outer circular dot radius, L T the transfer length, and d the gap spacing between the contacts. The total resistance (R T) was measured for various spacings and plotted as function of ln(R1 /R1 ⫺ d).13,14 The least squares curve fitting method was used to obtain a straight line plot of R T vs. ln(R1 /R1 ⫺ d). Thus, the specific contact resistivity can be calculated by ␳ c ⫽ L T2R s Current-voltage 共I-V兲 data were measured using a parameter analyzer 共HP 4155A兲. To characterize the extent of indiffusion between Al and n-ZnO, AES was used 共PHI 670 model兲 with an electron beam of 10 keV and 0.0236 ␮A depth profiles. The interfacial reaction products were identified by GXRD, which was carried out with a Rigaku diffractometer 共D/MAX-RC兲. Atomic force microscopy 共AFM兲 was employed to characterize the surface morphology of the samples. Results and Discussion

* Electrochemical Society Active Member. c

Present address: Samsung SDI, Gyeonggi-Do 442-391, Korea. z E-mail: [email protected]

Core

Technology

Laboratory,

Suwon,

Figure 1 shows the I-V characteristics of the Al/Pt metallization scheme on ZnO with increasing annealing temperatures. They were measured with a circular transmission line method of 4 ␮m spacing. Both as-deposited and annealed samples show linear I-V character-

G224


Figure 1. Current-voltage 共I-V兲 characteristics of the Al/Pt contacts on n-ZnO:Al with increasing annealing temperature.

istics, indicating good ohmic contact over the whole range of voltage ⫺4 and ⫹4 V. Compared to previously reported Al-based ohmic contact results,4,10 the Al/Pt contact on ZnO layer shows markedly improved I-V characteristics. This improvement in the Al/Pt contact on the ZnO layer can be explained by the high doping concentration of ZnO, the use of Pt capping layer on Al, and the different surface treatment method. However, the annealing at 600°C resulted in degradation in a performance, as compared with sample annealed at 300°C. To obtain the specific contact resistivity, the total resistance data were plotted as linear function of vs. ln(R1 /R1 ⫺ d). From the linear fitted data, the specific contact resistivity was calculated to be about 1.25 (⫾0.05) ⫻ 10⫺5 ⍀ cm2 for an as-deposited contact, 2 (⫾0.25) ⫻ 10⫺6 ⍀ cm2 for a 300°C contact, and 8 (⫾0.3) ⫻ 10⫺5 ⍀ cm2 for a 600°C contact, respectively. The annealing dependence of I-V behavior can be explained in terms of interdiffusion between the metal and ZnO layers during the annealing as discussed below. Figure 2 show AES depth profiles of the Al/Pt contacts on ZnO with increasing annealing temperatures. For the as-deposited sample, Fig. 2a, there is no obvious evidence for dissociation between the metal layers, except for some amount of oxygen that had outdiffused into Al layer. Note that the dissociation of the ZnO at the surface region is possible in the absence of any annealing process, due to the strong reaction between Al and O in the ZnO layer.14 Similar results were previously reported for a Ti/Au ohmic contact on ZnO at room temperature.7 For the sample annealed at 300°C, Fig. 2b, large amount of oxygen outdiffused from the ZnO to the Al layer. Note that the extent of ZnO dissociation for the 300°C annealed sample was greater than that of the as-deposited one. However, there was no evidence for outdiffusion of Zn into the Al layer. In addition, this shows that Al outdiffused through the Pt layer indicating the formation of Pt-Al phases. For the sample annealed at 600°C, Fig. 2c, it shows that a significant amount of oxygen outdiffused from ZnO resulting in the formation of a thick oxide interfacial layer. The degraded I-V characteristics of the 600°C annealed sample can be explained by the presence of a thick oxide interface layer, which prevents the carrier injection into the ZnO layer. Figure 3 shows the O AES signal at the interface region between Al and ZnO as a function of sputtering time. The data show that the O 共KLL transition兲 spectra signal of both the as-deposited and annealed samples shifts toward the lower kinetic-energy side, which is indicative of an interfacial reaction between Al and O. Compared to the as-deposited and 300°C annealed samples, the 600°C annealed sample shows much more outdiffusion of O from ZnO. This large outdiffusion of oxygen indicates that the ZnO is completely covered

Figure 2. AES depth profiles of 共a兲 as-deposited, 共b兲 300°C, and 共c兲 600°C annealed Al/Pt contacts on ZnO. Note that the dissociation of ZnO at room temperature is possible due to strong interfacial reactions between Al and O atoms.

by thick insulating Al-O or Al-Zn-O phases. Therefore, the higher specific contact resistivity of 600°C annealed sample, compared to that of 300°C annealed sample can be explained by the presence of thick insulating oxide layers that cover the ZnO surface. GXRD analysis was performed on as-deposited and annealed samples to identify interfacial reaction products. GXRD plot of the as-deposited sample 共Fig. 4a兲 exhibited the characteristic diffraction peaks (2␪ ⫽ 34.47°) for ZnO 共002兲. In addition, there are extra peaks for Pt共111兲 (2␪ ⫽ 40.02°) and 共200兲 (2␪ ⫽ 46.66°). However, no peaks corresponding to Al metal layer such as Al共111兲 (2␪ ⫽ 38.472°) and 共200兲 (2␪ ⫽ 44.738°) were evident, due to overlapping with the strong Pt共111兲 and 共200兲 peaks. In addition, the interfacial reaction products shown in AES depth profile were not detected due to the limitations in the resolution in the GXRD ex-


G225

Figure 3. Oxygen AES signal at interfacial 共Al/ZnO兲 results for 共a兲 asdeposited, 共b兲 300°C, and 共c兲 600°C annealed samples.

amination. Concerning the 300°C annealed sample 共Fig. 5b兲, the GXRD plot shows new diffraction peaks that include AlPt3 (100) (2␪ ⫽ 23.10°), 共101兲 (2␪ ⫽ 31.75°), 共111兲 (2␪ ⫽ 40.24°), 共002兲 (2␪ ⫽ 46.85°), 共201兲 (2␪ ⫽ 53.02°), 共211兲 (2␪ ⫽ 58.5°), and AlPt2 共102兲 (2␪ ⫽ 28.09°), 共111兲 (2␪ ⫽ 29.27°). Thin interfacial reaction layer was also not detected due to the limitation in GXRD resolution. As the annealing temperature was increased to 600°C, the intensity of the AlPt3 peaks was also enhanced and the AlPt2 peaks disappeared. These results indicate that most of the Pt and Al layers were consumed to form the AlPt3 phase by the further outdiffusion of Al. In addition, as expected from AES depth profile, there were new ␣-Al2 O3 共220兲 (2␪ ⫽ 31.6°), 共311兲 (2␪ ⫽ 38.6°), 共400兲 (2␪ ⫽ 45.3°), and AlZn2 O4 共311兲 (2␪ ⫽ 36.3°) peaks, which was indicative of the formation of a thick insulating layer on ZnO. Figure 5 shows AFM images of the as-deposited and annealed Al/Pt contacts on n-ZnO. The surface morphology of the asdeposited contact is fairly smooth with a root-mean-square 共rms兲 roughness of 3.2 nm, Fig. 6a. Compared to the as-deposited Ti/Au 共5.4 nm rms roughness兲10 and the Al/Au 共6.3 nm rms roughness兲 ohmic contact scheme, the surface of Al/Pt is smoother and is very small 3.2 nm. Generally, the Au injection layer has a high surface energy and, as a result, shows a fairly rough surface and caused by the agglomeration of Au atoms. Therefore, the stable Pt injection layer represents a promising top layer, which prevents the surface degradation of an ohmic contact. However, the 300 and 600°C annealed Al/Pt contacts show degraded surfaces with rms roughness of 6.7 and 7.0 nm, respectively. Based on I-V characteristics, AES and GXRD, the annealing dependence of the electrical properties of the Al/Pt contacts on n-ZnO can be explained as follows. The AES results show that the oxygen outdiffused into the Al metal layer and results in Al-O bond in the interface region. This is consistent with the fact that the standard Gibb’s free formation energy for ␣-Al2 O3 (⌬G o298 ⫽ ⫺1492 kJ/mol) is much smaller than that for ZnO (⌬G o298 ⫽ ⫺324 kJ/mol). 15,16 Figure 6 shows a comparison of the Gibb’s

free formation energy for Al2 O3 and ZnO with increasing temperature. For ZnO, ⌬G of ⫽ ⫺377956 ⫹ 180.444T, while for ␣-Al2 O3 , ⌬G fo ⫽ ⫺1588824 ⫹ 325.584T. 17 The outdiffusion of oxygen atoms is indicative of the accumulation of oxygen vacancies in the region near the ZnO layer.18,19 Thus, the surface region of n-ZnO would be expected to contain a very high concentration of carriers due to the predominance of these oxygen vacancies and the accumulation of Al donor atoms. This highly doped n⫹-ZnO surface could result in a tunneling process between the metal and n⫹-ZnO layers as shown in Fig. 7. When the field emission process dominates the current transport at room temperature, the specific contact resistivity, ␳ c , is given as20

␳ c ⬀ exp

冉冊

冋

q␾ B 2 冑␧ sm * ⫽ exp E00 共 h/2␲ 兲

冉冊册 ␾B

冑N D

关2兴

where ␧ s is the semiconductor permittivity, m * the effective mass, h the Plank constant, and ␾ B the barrier height. The above equation shows that in the field emission process, the specific contact resistivity is strongly dependent on the doping concentration. Thus, the increase in carrier concentration near surface of the ZnO layer is responsible for the very linear I-V characteristics of 300°C annealed sample. Furthermore, interfacial reactions between Al and O produce dissociated Zn, leading to a low n-type barrier height.15,21,22 Finally, the interfacial layer can serve as a diffusion barrier for Zn atoms, as shown in the AES depth profile.9 Therefore, the improved specific contact resistivity of Al/Pt ohmic contact by annealing at 300°C can be attributed to the combined effects of the increase carrier concentration by the outdiffusion of oxygen and indiffusion of Al atoms. However, the degraded ohmic contact characteristics of the 600°C annealed sample can be attributed to the formation of a thick-insulating ␣-Al2 O3 phase at the interface and surface of Al/Pt contact, as shown in the AES and GXRD results.

Figure 4. Glancing angle X-ray diffraction plots of 共a兲 as-deposited, 共b兲 300°C, and 共c兲 600°C annealed samples.

G226


Figure 5. AFM images of 共a兲 asdeposited, 共b兲 300°C, and 共c兲 600°C annealed samples.

Conclusion The electrical and structural properties of the Al/Pt layer on n-ZnO, annealed at temperatures of 300 and 600°C in nitrogen ambient were investigated using glancing angle current-voltage measurement, GXRD and AES. Clearly, the electrical behavior of the Al/Pt contacts was dependent on the annealing temperature. The Al/Pt contact scheme showed the best ohmic behavior when annealed at 300°C with a specific contact resistivity of 2 (⫾0.3) ⫻ 10⫺6 ⍀ cm2 . AES depth profile measurements show the large amounts of oxygen atoms diffuse from the ZnO and accumulate oxygen vacancies, which likely act as a donor. However, an increase of annealing temperature to 600°C resulted in the degradation of electrical properties via the formation of a thick interfacial insulating layer. Therefore, the low specific contact resistivity of the Al/Pt contact scheme can be attributed to increased carrier concentration at the region of the ZnO surface by the formation of thin Al-O layers. Acknowledgments This work is supported, in part, by Brain Korea 21 project and the U.S. Air Force Office of Scientific Research 共AFOSR兲/Asian Office of Aerospace Research and Development 共AOARD, monitor: Joanne Maurice兲. The University of Illinois and the Kwangju Institute of Science and Technology assisted in meeting the publication costs of this article.

References Figure 6. Standard Gibb’s free energy for the formation of ZnO and ␣-Al2 O3 with increasing temperature.

Figure 7. Schematic band diagram of Al contact on n-ZnO after interfacial reaction at room temperature.

1. Z. K. Tang, G. K. L. Wong, P. Yu, M. Kawasaki, A. Ohtomo, H. Koinuma, and Y. Segawa, Appl. Phys. Lett., 72, 3270 共1998兲. 2. A. Kudo, H. Yanagi, K. Ueda, H. Hosono, H. Kawazoe, and Y. Yano, Appl. Phys. Lett., 75, 2851 共1999兲. 3. H. Ohta, M. Orita, M. Hirano, and H. Hosono, J. Appl. Phys., 89, 5720 共2001兲. 4. S. Liang, H. Sheng, Y. Liu, Z. Huo, and H. Shen, J. Cryst. Growth, 225, 110 共2001兲. 5. D. G. Thomas, J. Phys. Chem. Solids, 15, 86 共1960兲. 6. T. Akane, K. Sugioka, and K. Midorikawa, J. Vac. Sci. Technol. B, 18, 1406 共2000兲. 7. H.-K. Kim, S.-H. Han, W. K. Choi, and T.-Y. Seong, Appl. Phys. Lett., 77, 1647 共2000兲. 8. H.-K. Kim, S.-H. Han, W. K. Choi, and T.-Y. Seong, J. Electrochem. Soc., 148, G114 共2001兲. 9. H.-K. Kim, K.-K. Kim, S.-J. Park, T.-Y. Seong, and Y. S. Yoon, Jpn. J. Appl. Phys., Part 2, 41, L546 共2002兲. 10. G. Xiong, J. Wilkinson, B. Mischuck, S. Tuzemen, K. B. Ucer, and R. T. Williams, Appl. Phys. Lett., 80, 1195 共2002兲. 11. K.-K. Kim, J.-H. Song, H.-J. Jung, W.-K. Choi, S.-J. Park, and J.-H. Song, J. Appl. Phys., 87, 3573 共2000兲. 12. G. S. Marlow and M. B. Das, Solid-State Electron., 25, 91 共1982兲. 13. J.-K. Ho, C.-S. Jong, C. C. Chiu, C.-N. Huang, K.-K. Shin, L.-C. Chen, F.-R. Chen, and J.-J. Kai, J. Appl. Phys., 86, 4491 共1999兲. 14. H.-K. Kim, K.-K. Kim, T.-Y. Seong, and I. Adesida, Electrochem. Solid-State Lett., Submitted. 15. G. V. Samsonov, The Oxide Handbook, 2nd ed, IFI Plenum Data Company, New York 共1981兲. 16. CRC Handbook of Chemistry and Physics, 73rd ed., D. R. Lide, Editor, Chemical Rubber Corp., Boca Raton, FL 共1992兲. 17. G. Chen and C.-X. Sun, Mater. Sci. Eng., A, A244, 291 共1998兲. 18. K. Vanheusden, C. H. Seager, W. L. Warren, D. R. Tallant, and J. A. Voigt, Appl. Phys. Lett., 68, 403 共1996兲. 19. A. Poppl and G. Volkel, Phys. Status Solidi A, A125, 571 共1991兲. 20. A. Y. C. Yu, Solid-State Electron., 13, 239 共1970兲. 21. H. Sheng, S. Muthukumar, N. W. Emanetoglu, and Y. Lu, Appl. Phys. Lett., 80, 2132 共2002兲. 22. L. J. Brillson, Surf. Sci. Rep., 2, 295 共1982兲.