Metal/nonpolar m-plane ZnO contacts with and without thin Al2O3 interlayer deposited by atomic layer deposition Hogyoung Kim, Dong Ha Kim, Sungyeon Ryu & Byung Joon Choi

Journal of Materials Science: Materials in Electronics ISSN 0957-4522 J Mater Sci: Mater Electron DOI 10.1007/s10854-017-7370-z

1 23

Your article is protected by copyright and all rights are held exclusively by Springer Science+Business Media, LLC. This e-offprint is for personal use only and shall not be selfarchived in electronic repositories. If you wish to self-archive your article, please use the accepted manuscript version for posting on your own website. You may further deposit the accepted manuscript version in any repository, provided it is only made publicly available 12 months after official publication or later and provided acknowledgement is given to the original source of publication and a link is inserted to the published article on Springer's website. The link must be accompanied by the following text: "The final publication is available at link.springer.com”.

1 23

Author's personal copy J Mater Sci: Mater Electron DOI 10.1007/s10854-017-7370-z

Metal/nonpolar m-plane ZnO contacts with and without thin Al2O3 interlayer deposited by atomic layer deposition Hogyoung Kim1,2 · Dong Ha Kim3 · Sungyeon Ryu3 · Byung Joon Choi3 

Received: 24 March 2017 / Accepted: 17 June 2017 © Springer Science+Business Media, LLC 2017

Abstract Using the temperature dependent current–voltage (I–V) measurements, the electrical properties of Au/ nonpolar m-plane ZnO Schottky diodes with an Al2O3 interlayer prepared by atomic layer deposition (ALD) was investigated. With an Al2O3 interlayer, it was found to have higher barrier heights and higher rectifying ratio. Modified Richardson plots produced effective Richardson constants of 30.0 and 37.6 Acm−2K−2 for the samples with and without Al2O3 interlayer, respectively, which are similar to the theoretical value of 32.0  Acm− 2K− 2 for n-ZnO. Scanning transmission electron microscope (STEM) results showed that the oxygen-contained layer on ZnO surface degraded the film quality of subsequently deposited Al2O3 layer. In addition, the inter-diffusion of Au and Al atoms into ZnO subsurface region also modulated the electrical properties of Au/ZnO contacts.

* Hogyoung Kim [email protected] * Byung Joon Choi [email protected] 1

Department of Visual Optics, Seoul National University of Science and Technology (Seoultech), Seoul 01811, South Korea

2

Convergence Institute of Biomedical Engineering & Biomaterials, Seoul National University of Science and Technology (Seoultech), Seoul 01811, South Korea

3

Department of Materials Science and Engineering, Seoul National University of Science and Technology (Seoultech), Seoul 01811, South Korea

1 Introduction Zinc oxide (ZnO) has attracted considerable interest for its applications in the electronic and optoelectronic devices such as thin film transistors (TFTs), light-emitting diodes (LEDs), white light sources and photovoltaic solar cells due to its wide band gap (3.4 eV at 300 K) and high exciton binding energy (60 meV) [1]. As is well known, the electric field in wurtzite structures is polarized along the c-axis due to spontaneous and piezoelectric polarizations. This polarization field can degrade the device performance of light emitting diodes (LEDs) by decreasing internal quantum efficiency of the emitting devices [2, 3]. Then, it is advantageous to obtain high quality nonpolar ZnO films because electric field along the growth direction is zero in nonpolar ZnO films. Several methods have been proposed to obtain nonpolar ZnO films. For example, a-plane ZnO films grown on r-plane sapphire substrates by metal–organic chemical vapor deposition (MOCVD) [4] and molecular beam epitaxy (MBE) [5] were reported. Cagin et al. grew m-plane ZnO films on MgO (001) substrates by MBE [6]. Single crystal nonpolar ZnO substrates can also be used to grow nonpolar ZnO films with low defect density. However, various crystalline defects present in the bulk ZnO will affect the subsequently grown ZnO film quality. Due to the presence of interface states in the metal–semiconductor (MS) junction, the discrepancy between the Schottky barrier height (SBH) and the metal work function arises in the real fabricated diodes. Modulation of electrical properties of MS contacts by inserting an interfacial layer between the metal and the semiconductor layers have been researched in several semiconductors such as Ge [7], InGaAs [8] and SiC [9]. The interfacial layer acting as a dangling bond terminator at the semiconductor surface reduces the surface charge trap density, alleviating

13

Vol.:(0123456789)

Author's personal copy J Mater Sci: Mater Electron

the Fermi level pinning (FLP) effect [10]. Krajewski et al. performed similar work on the ZnO based devices using hafnium dioxide (HfO2) as an insertion layer [11]. They observed the increased rectifying ratio with a HfO2 interlayer, which was attributed to the passivation of ZnO surface accumulation layer. Because the constant progress in the ultra-large-scale integration systems requires high dielectric constant (high-k) materials as gate materials [12], systematic investigation on the role of very thin high-k dielectric materials in MS junctions is helpful to develop the device technology. In this work, we applied very thin Al2O3 interlayer grown by atomic layer deposition (ALD) to Au/nonpolar m-plane bulk ZnO Schottky contacts and comparatively investigated the electrical properties under various temperatures.

2 Experimental A double side polished unintentionally-doped (undoped), hydrothermally-grown, m-plane (1100) bulk ZnO single crystals (thickness: 500  µm, carrier concentration: 5 × 1014  cm−3) purchased from Tokyo Denpa Ltd. were used as a starting material. The wafer was cut into small pieces and some of them were loaded into an ALD chamber after cleaning process. Then, the temperature was ramped up to 250 °C for 5  min in an N2 ambient. After H2O prepulse treatment for 3  min, about 2.5  nm thick Al2O3 layer was deposited at 250 °C. For Schottky contacts, 50 nm thick Au metal films were deposited by using electron beam evaporation through a shadow mask. For ohmic contacts, In metal was rubbed over the entire back surface Fig. 1 Temperature dependent current–voltage (I–V) characteristics for a Au/ZnO and b Au/ Al2O3/ZnO junctions

13

of the samples. Current–voltage (I–V) measurements were performed with a Keithley 238 current source. I–V measurements under various temperatures were performed using a hot chuck connected with a temperature controller.

3 Results and discussion Figure 1 shows the semi-logarithmic I–V curves measured at different temperatures for both samples. Compared to Au/ZnO junction, Au/Al2O3/ZnO junction showed much lower current values. In addition, the variation of current values with varying temperatures was larger for Au/Al2O3/ ZnO junction. Regardless of the temperature, the rectifying ratio at ±2 V was about ~5 for Au/ZnO junction. The current transport is ohmic-like, which can be associated with the electron accumulation layer on the ZnO surface [13]. Coppa et al. have investigated Au/n-ZnO Schottky contacts, after exposure to O2/He plasma. Significant improvement in the I–V characteristics of plasma treated sample was found compared to the as-received samples which showed the microampere level leakage currents and ideality factors of n > 2 due to the OH− conducting layer [14]. Such conducting layer was found to affect the electrical properties of ZnO Schottky diodes significantly according to the measurement ambient condition [15, 16]. Because I–V measurements were performed in air in this work, acceptor-like adsorbates such as O2 and H2O could (if possible, mainly occurred in the periphery of the contacts) passivate the surface conduction layer. Monotonic increase of the current values for Au/ZnO junction also implies that such passivation behavior became weaker with increasing

Author's personal copy J Mater Sci: Mater Electron

the temperature. For Au/Al2O3/ZnO junction, the rectifying ratio ±2 V increased to ~60 at 300 K. The conductive accumulation layer might be compensated to some extent by Al2O3 interlayer. However, the rectifying ratio became about ~3 above 338 K, indicating the activation of conduction layer. From the forward bias I–V curves, the electrical parameters were determined based on the thermionic emission (TE) model [17]

I = I0 [exp(q(V − IRS )∕nkT) − 1]

(1)

(2) where I0 is the reverse bias saturation current, A is the diode area, A** is the effective Richardson constant (32 Acm−2K−2 for n-ZnO), ϕB is the effective barrier height, n is the ideality factor, and RS is the series resistance. Note that before performing current–voltage–temperature (I–V–T) measurements, we measured 5–6 diodes for each sample at room temperature and the barrier heights were obtained as 0.57 (±0.04) eV and 0.69 (±0.02) eV, respectively, for Au/ZnO and Au/Al2O3/ZnO junctions. From the I–V data in Fig. 1, the temperature dependences of the ideality factor and the barrier height are shown in Fig.  2. Compared to Au/ZnO junction, Au/Al2O3/ZnO junction showed the increase in both the barrier height and ideality factor. At 300 K, the barrier heights (ideality factors) were 0.52  eV (3.28) and 0.69  eV (3.45) for Au/ZnO and Au/ Al2O3/ZnO junctions, respectively. The variation of barrier height according to temperature is relatively small for Au/ Al2O3/ZnO junction. Figure  2 also shows that the ideality

I0 = AA∗∗ T 2 exp(−q𝜙B ∕kT)

factor decreased and the barrier height increased with the temperature, generally explained by lateral barrier inhomogeneity [18]. By assuming that the inhomogeneous barriers are distributed with a Gaussian function over the Schottky contact area, the relation between the effective barrier height, φB and the temperature can be given as (3) From Eqs.  (2) and (3), the modified Richardson plot can be derived as follows

𝜙B = 𝜙̄ B − q𝜎0 2 ∕2kT

ln(I0 ∕T 2 ) − q2 𝜎0 2 ∕2k2 T 2 = ln(AA∗∗ ) − q𝜙̄ B ∕kT (4) where 𝜙̄ B is the zero-bias mean barrier height and σ0 is the standard deviation with the negligible temperature dependence [18]. As shown in Fig. 3a, the linear fits to the plots of ϕB versus 1/2kT based on Eq. (3) yielded the σ0 values of 0.182 and 0.097  V for Au/ZnO and Au/Al2O3/ZnO junctions, respectively. The smaller σ0 value for Au/Al2O3/ZnO junction indicates that the higher degree of uniformity in the local barrier heights was obtained with an Al2O3 interlayer. Figure 3b shows the ln(I0 ∕T 2 ) − q2 𝜎0 2 ∕2k2 T 2 versus 1/kT plots. The intercepts at the ordinate from the linear fits produced the Richardson constant of A** as 37.6 and 30.0 Acm−2K−2 for Au/ZnO and Au/Al2O3/ZnO junctions, respectively. The obtained Richardson constants are close to the theoretical value of 32 Acm−2K−2 for n-ZnO, implying that the barrier inhomogeneity can explain the current transport properties.

Fig. 2 Temperature dependence of a barrier height and b ideality factor for both samples

13

Author's personal copy J Mater Sci: Mater Electron Fig. 3 a Plots of barrier height versus 1/2kT, b modified Richardson plot of ln(I0/T2) − q2σ2/2k2T2 versus 1/kT

When the thickness of interfacial layer is considered, the forward bias current density–voltage (J–V) characteristics for the Schottky diodes is given as [17] √ J ≈ A∗∗ T 2 exp(− 𝜁𝛿) exp[−q∕kT(𝜙B − V∕n)] (5) where ζ (in eV) and δ (in Å) are the effective barrier and effective thickness of the interfacial layer, respectively. As shown in Fig.  4a for Au/Al2O3/ZnO junction, plots of ln(J/T2) versus 1/kT (Richardson plot) at different forward biases produced a set of different Ea values (Ea = 𝜙B − V∕n ) from the slopes of the Richardson plots. Then, the barrier height was determined to be 0.63  eV from the linear Fig. 4 a Richardson plot of Au/Al2O3/ZnO junction under different forward biases and b plot of the activation energy (Ea) versus the corresponding forward biases

13

fitting to the Ea versus VF plot shown in Fig. 4b. This value is still larger than the barrier height of 0.52 eV at 300 K for Au/ZnO junction. Except the effect of interfacial layer, the modulation of interface properties occurred at the Al2O3/ ZnO interface, enhancing the barrier height. The detailed interface characteristics were investigated using the cross-sectional scanning transmission electron microscope (STEM). Figure  5a shows the STEM image around the Al2O3 interlayer region and the distribution of each component across the Au/Al2O3/ZnO interface was conducted by using energy-dispersive X-ray spectroscopy (EDS) for the elements of Au, Al, O and Zn as shown in

Author's personal copy J Mater Sci: Mater Electron

Fig. 6 Photoluminescence intensity measured at room temperature and the atomic force microscope (AFM) image in the inset for bare m-plane ZnO

Fig. 5 a Cross-sectional scanning transmission electron microscope (STEM) image across the Al2O3 layer, and b energy-dispersive X-ray spectroscopy (EDS) depth profiles for each element

Fig.  5b. The beam size of STEM mode is 1  nm and the step size for EDS scan is about 0.38  nm. The EDS depth profiles revealed that the diffusion of Au atoms even into the ZnO subsurface region (denoted as no. 1 in Fig. 5a) is significant. Furthermore, the diffusion of Al atoms into the ZnO subsurface is not negligible. Hence, region 1 can be regarded as Au and Al doped ZnO layer with the thickness of ~5  nm. Meanwhile, the thickness of region 2 is measured to be about 7 nm, thicker than the originally designed 2.5  nm thick Al2O3 layer. Figure. 6 shows the photoluminescence (PL) intensity for bare m-plane ZnO measured at room temperature. PL spectrum consists of a UV near band edge (NBE) emission and a broad deep band emission around 2.1–2.5  eV (500–590 nm: green luminescence) related to point defects. Recently it was suggested that, green luminescence in ZnO has multiple origins and consists of a band at 2.3 eV related zinc vacancy acceptors and a band at 2.47  eV related to oxygen vacancies [19]. Although the nature of these defects is still controversial,

strong deep PL emission indicates that lots of point defects are present in m-plane ZnO. Atomic force microscopy (AFM) image in the inset of Fig. 6 shows that m-plane ZnO has a corrugated morphology. Such corrugated surface morphology might increase the thickness of surface oxide layer. Furthermore, we performed H2O prepulse treatment at 250 °C for 3  min before Al2O3 deposition. During this treatment, oxygen species were adsorbed on ZnO surface. These oxygen species as well as the native oxide (namely, oxygen-contained layer) might affect the subsequent Al2O3 deposition (for example, deposition rate and film quality). It is known that Au and Al act as acceptor and donor, respectively, in ZnO [20]. Significant diffusion of Au atoms into ZnO indicates the formation of p-type doped ZnO layer. Probably this layer also affected the barrier height. The Al2O3 interlayer did not block the diffusion of Au atoms into ZnO effectively, which is quite different from the result for InP [21]. Resultantly, oxygen-contained layer on ZnO mainly degraded the film quality of overgrown Al2O3 layer. Auret et al. showed that the higher forward and reverse currents were formed for electron beam deposited Au contacts on n-type Ge as compared to that for resistively deposited Au contacts, which was associated with the process-induced defects acting as generation centers [22]. The electron beam induced defects are caused by ionized particles that are accelerated from the region near the filament [23] and impinge on the semiconductor surface. The induced defects by electron beam exposure in Au/n-GaAs [24] and Ni/Au/n-GaN [25] contacts were electrically characterized. These results suggest the possibility that Au atoms may also diffuse into ZnO layer and induce defects during electron beam evaporation process. Meanwhile, it was shown in Au/ZnO contacts that thermally driven defect formation in high defect ZnO is more significant than in

13

Author's personal copy J Mater Sci: Mater Electron

low defect ZnO [26]. In addition, lower barrier height and higher reverse leakage current in O-polar ZnO as compared to m-plane ZnO was associated with the different surface morphology and the surface defects [27]. Therefore, it is possible that ionized Au particles impinge on m-plane ZnO surface and will diffuse into the subsurface more easily. In order to check the repeatability of the electrical data, we deposited about 2.5 nm thick Al2O3 layer on ZnO wafer again and then fabricated Au/Al2O3/ZnO junction. Based on the I–V data measured at room temperature (typical I–V curve is shown in Fig. 7), we obtained the average barrier height of 0.66 (±0.04) eV. This is very similar to the value obtained from first measurements. Hence, our electrical data from Au/Al2O3/ZnO junction is repeatable. Based on the results, the possible mechanism can be suggested as follows: (i) Before Al2O3 deposition, oxygen-contained layer (native oxide) was present on the ZnO surface. (ii) During Al2O3 deposition, some amount of Al atoms diffused into the ZnO subsurface region, degraded the film quality of Al2O3 layer (decrease of [Al]/[O] atomic percent ratio) and produced the thicker Al2O3 than expected. (iii) During Au deposition, Au atoms diffused into the ZnO subsurface region. Therefore, the preparation of fairly smooth surface for m-plane ZnO before ALD process is essential to improve the device performance. Finally, in order to rule out any possibility that the 2.5  nm thick Al2O3 layer would not provide continuous coverage on the m-plane ZnO due to the corrugated morphology, we investigated the Au/ZnO junction with a 10  nm thick Al2O3 layer. Figure  7 shows the typical I–V data measured at room temperature. From the I–V data, the barrier height was found to be 0.67 (±0.05) eV for the sample with a 10 nm thick Al2O3 layer, which is similar to the sample with a 2.5  nm thick Al2O3 layer. This assures

Fig. 7 Typical current–voltage (I–V) characteristics for the samples with different Al2O3 thicknesses

13

again that the electrical properties in Au/ZnO junction was enhanced with an Al2O3 layer.

4 Conclusion We carried out temperature dependent current–voltage (I–V) measurements to characterize the electrical properties of Au/nonpolar m-plane ZnO Schottky diodes with an Al2O3 interlayer prepared by atomic layer deposition (ALD). Compared to Au/ZnO junction, Au/Al2O3/ZnO junction was found to have higher barrier heights and higher rectifying ratio. Modified Richardson plots produced an effective Richardson constant of 30.0 Acm−2K−2 for Au/Al2O3/ZnO junction, similar to the theoretical value of 32.0 Acm−2K−2 for n-ZnO. STEM results indicate that the oxygen-contained layer on ZnO surface degraded the film quality of Al2O3 layer. Inter-diffusion of Au and Al atoms into the ZnO subsurface region also modulated the electrical properties of Au/ZnO contacts. Acknowledgements This study was supported by the Research Program funded by the Seoul National University of Science and Technology.

References 1. S. Pearton, D. Norton, K. Ip, Y. Heo, T. Steiner, Prog. Mater. Sci. 50, 293 (2005) 2. A. Chakraborty, H. Xing, M. Craven, S. Keller, T. Mates, J. Speck, S. DenBaars, U. Mishra, J. Appl. Phys. 96, 4494 (2004) 3. X. Wu, G. Siu, C. Fu, H. Ong, Appl. Phys. Lett. 78, 2285 (2001) 4. C. Gorla, N. Emanetoglu, S. Liang, W. Mayo, Y. Lu, M. Wraback, H. Shen, J. Appl. Phys 85, 2595 (1999) 5. S. Han, S. Hong, J. Lee, J. Song, Y. Nam, S. Chang, T. Minegishi, T. Yao, J. Cryst. Growth 309, 121 (2007) 6. E. Cagin, J. Yang, W. Wang, J. Phillips, S. Hong, J. Lee, J. Lee, Appl. Phys. Lett. 92, 233505 (2008) 7. B. Tsui, M. Kao, Appl. Phys. Lett. 103, 032104 (2013) 8. L. Chauhan, S. Gupta, P. Jaiswal, N. Bhat, S. Shivashankar, G. Hughes, Thin Solid Films 589, 264 (2015) 9. S. Zheng, Q. Sun, W. Yang, P. Zhou, H. Lu, D. Zhang, Nanoscale Res. Lett. 8, 116 (2013) 10. Y. Zhou, M. Ogawa, X. Han, K. Wang, Appl. Phys. Lett. 93, 202105 (2008) 11. T. Krajewski et al., Appl. Phys. Lett. 98, 263502 (2011) 12. G. Wilk, R. Wallace, J. Anthony, J. Appl. Phys. 89, 5243 (2001) 13. M. Allen, C. Swartz, T. Myers, T. Veal, C. McConville, S. Durbin, Phys. Rev. B 81, 075211 (2010) 14. B. Coppa, R. Davis, R. Nemanich, Appl. Phys. Lett. 82, 400 (2003) 15. H. Kim, H. Kim, J. D. Kim, Korean Phys. Soc. 61, 1314 (2012) 16. H. Von Wenckstern, S. Muller, G. Giehne, H. Hochmuth, M. Lorenz, M. Grundmann, J. Electron. Mater. 39, 559 (2010). 17. S. Sze, Physics of Semiconductor Devices. (Wiley, New York, 1981). 18. R. Tung, Mater. Sci. Eng. R 35, 1 (2001) 19. J. Cızek, J. Valenta, P. Hruska, O. Melikhova, I. Prochazka, M. Novotny, J. Bulır, Appl. Phys. Lett. 106, 251902 (2015)

Author's personal copy J Mater Sci: Mater Electron 20. A. Janotti, C. Van de Walle, Rep. Prog. Phys. 72, 126501 (2009) 21. H. Kim, M. Kim, S. Ryu, B. Choi, Semicond. Sci. Technol. 32, 025011 (2017) 22. F. Auret, S. Coelho, P. Rensburg, C. Nyamhere, W. Meyer, Mater. Sci. Semicond. Process. 11, 348 (2008). 23. C. Christensen, J. Petersen, A. Larsen, Appl. Phys. Lett. 61, 1426 (1992) 24. S. Tunhuma, F. Auret, J. Nel, E. Omotoso, H. Danga, E. Igumbor, M. Diale, Nucl. Instrum. Methods Phys. Res. B (2017). 10.1016/j.nimb.2017.05.041

25. P. Ngoepe, W. Meyer, F. Auret, E. Omotoso, M. Diale, Mater. Sci. Semicond. Process. 64, 29 (2017). 26. H. Mosbacker, C. Zgrabik, M. Hetzer, A. Swain, D. Look, G. Cantwell, J. Zhang, J. Song, L. Brillson, Appl. Phys. Lett. 91, 072102 (2007) 27. H. Kim, Trans. Electr. Electron. Mater. 16, 285 (2015)

13