Interface characteristics for graphene contact to n-type ...

J Mater Sci: Mater Electron DOI 10.1007/s10854-015-2796-7

Interface characteristics for graphene contact to n-type and p-type GaN observed by X-ray photoelectron spectroscopy Chia-Lung Tsai • Yow-Jon Lin • Jian-Huang Lin

Received: 19 November 2014 / Accepted: 31 January 2015 Ó Springer Science+Business Media New York 2015

Abstract The interface characteristics of graphene/GaN samples using X-ray photoelectron spectroscopy (XPS) measurements are investigated. XPS makes it possible to extract a reliable barrier-height value. For graphene/n-type GaN (graphene/p-type GaN) samples, the Schottky barrier height is 0.85 (2.50) eV. To determine the Fermi-level pinning/unpinning at the graphene/GaN interfaces, an analysis is conducted according to the Schottky–Mott limit. It is shown that the Fermi energy level is unpinned and the barrier-height value is dependent on the work function of graphene. Investigation of graphene/GaN interfaces is important, and providing the other technique for surface potential control is possible.

1 Introduction Graphene is a zero-gap semiconductor and has a very large intrinsic carrier mobility, which makes it a very promising material for incorporation into devices ranging from diodes to transistors [1–5]. Many efforts have been made to investigate the physical properties of Schottky contacts and their interfacial properties of graphene on semiconductors, such as GaN, Si, and MoS2 [6–18]. Due to the technological importance of Schottky diodes which are among the most

C.-L. Tsai Electronics and Optoelectronics Research Laboratories, Industrial Technology Research Institute, Hsinchu 31040, Taiwan Y.-J. Lin (&) J.-H. Lin Institute of Photonics, National Changhua University of Education, Changhua 500, Taiwan e-mail: [email protected]

simple of the graphene-semiconductor contact devices, a full understanding of the nature of their electrical characteristics is of great interest. In this paper, we report transformation of epitaxial graphene layers onto the n-type GaN (n-GaN) and p-type GaN (p-GaN) surfaces. Investigation of graphene/GaN interfaces is important, and providing the other technique for surface potential control is possible. The formation of Schottky and ohmic contacts between graphene and GaN is crucial to their wide application in the area of electronic and optoelectronic devices [12–16]. However, the behavior of Schottky barriers at the graphene/ GaN interfaces is complicated and not well understood. In general, the Schottky barrier height is extracted through traditional current–voltage (I–V) and capacitance–voltage (C–V) measurement methods. However, it is difficult to obtain reliable C–V measurements because of high series resistance [16] and high junction conductance [19]. In addition, the derived ideality-factor (g) values from I–V measurements are much larger than 2 [13], suggesting the existence of the large tunneling factor (Eoo) and making the thermionic emission theory inapplicable in the accurate barrier-height extraction [20–22]. The interface characteristics of graphene/n-GaN and graphene/p-GaN samples using X-ray photoelectron spectroscopy (XPS) measurements are investigated in this study. XPS makes it possible to extract a reliable barrier-height value. We found that the Fermi energy level (EF) is unpinned and the barrier-height value is dependent on the work function of graphene.

2 Experimental procedures The n- and p-GaN epitaxial layers used in this study were grown on sapphire substrates using an atmospheric pressure metalorganic chemical vapor deposition system

123

J Mater Sci: Mater Electron

(Taiyo Nippon Sanso SR-4000). A 20-nm-thick low-temperature GaN nucleation layer (510 °C) was first grown on the c-plane sapphire substrate, and a 3,000-nm-thick undoped GaN layer (1130 °C) was then grown. Next, the 1,000-nm-thick Si-doped n-GaN (1,130 °C) and 300-nmthick Mg-doped p-GaN layers (1,025 °C) were grown on the undoped GaN layer, respectively. The grown p-GaN samples were annealed for the purpose of generating holes, at 650 °C for 20 min in ambient N2. Using Hall measurement at room temperature, for n-GaN samples, the electron concentration and mobility were determined to be 5.73 9 1017 cm-3 and 560 cm2 V-1 s-1, respectively. For p-GaN samples, the hole concentration and mobility were determined to be 2.43 9 1017 cm-3 and 4.7 cm2 V-1 s-1, respectively. After epitaxial growth, the samples were first cleaned with chemical solutions of trichloroethylene, acetone, and methanol, then rinsed with deionized water, and immediately blow dried with N2. Next, the surfaces of nand p-GaN samples were treated with aqua regia for 10 min. Then, the n- and p-GaN samples were dipped into a 60 °C-(NH4)2Sx solution (with 6 % S, Nippon Shiyaku Co., Ltd.) for 30 min [20, 23]. The graphene sheet was grown by chemical vapor deposition (CVD). Before the CVD growth of graphene, the copper foil (90 lm thick) was preannealed at 1,000 °C for 30 min under a flow of H2 = 9 SCCM (SCCM denoted standard cubic centimeter per minute) in order to prepare a high-density terrace structure on Cu. A gas mixture of CH4 (120 SCCM) and H2 (40 SCCM) was used for the growth of graphene at 66.7 Pa. After 40 min of growth, the system was cooled to room temperature under H2. To transfer the as-grown graphene sheets, a polymethylmethacrylate (PMMA) layer was spin coated on the graphene/Cu sample. The PMMA/graphene/Cu sample was then baked at 100 °C for 1 min. The procedures from coating to drying were repeated five times. Next, the sample was immersed in FeCl3 solution (0.1 g/cm3) for 6 h overnight to remove the Cu substrates. The PMMA/graphene layers were respectively transferred to n- and p-GaN substrates. The PMMA/ graphene/n-GaN and PMMA/graphene/p-GaN samples were then dried at 50 °C for 30 min on a hotplate. Next, the PMMA layer was dissolved by acetone. The graphene/nGaN and graphene/p-GaN samples were then inserted into a furnace and annealed in pure nitrogen ambient at 200 °C for 20 min. The graphene area is 0.25 cm2. The structural property of graphene was examined using Raman spectroscopy (Ramboss 500i, DongWoo Optron). A 532-nm laser was used for excitation. The graphene work function was examined with the scanning Kelvin probe (KP Technology). Systems offer very a high work-function resolution of 1–3 meV (2 mm tip). The XPS study is used to determine the band-structure lineup of the graphene/GaN heterojunction structure. XPS measurements (ULVAC-

123

PHI, PHI 5000) were performed using a monochromatic Al Ka X-ray source. The photon energy was calibrated with the Au 4f core-level line.

3 Results and discussion Figure 1 shows the Raman spectra of graphene films. Graphene displays a band at *1,345 cm-1, a band at *1,580 cm-1 and a band at *2,700 cm-1 corresponding to the well-documented D, G and 2D bands [4, 5]. The G band is assigned to the E2g mode of the relative motion of sp2 carbon atoms. The intensity of the 2D band is related to the layer numbers of graphene. The ratio of the 2D to G peak intensities was calculated to be close to 1, suggesting that four layers of graphene formed [24, 25]. In addition, the full width of half-maxima (FWHM) of the 2D peak is 75.1 cm-1, which is close to the reported value of 75.0 cm-1 for few-layer graphene on SiO2 [26]. Our graphene film shows FWHM of the 2D peak as 75.1 cm-1 which is broader than monolayer graphene on top of the Si surface (28.3 cm-1 [27]). This enhancement shows the formation of multilayer graphene. Thus, it is reasonable to conclude that four layers of graphene formed in this study. The D band is disorder-induced and caused by phonon scattering at defect sites and impurities. The ratio of the D to G peak intensities (ID/IG) is usually used to evaluate the disorder of graphene. The observed ID/IG value (0.42) in this study is larger than the reported value by Luo et al. [28]. The Raman spectrum is sensitive to growth conditions. The different growth conditions may lead to the change in the ID/IG ratio. In this study, the Schottky barrier height (q/n) at the graphene/n-GaN interface was determined from the XPS data using the relation shown in Eq. (1) [21, 29, 30].

Fig. 1 Raman spectra of graphene films


G q/n ¼ Eg EiV þ Eicore EG core ¼ Eg Ecore EVC

ð1Þ

where Eg (Eg = 3.4 eV) is the energy band gap of GaN, EG core is the binding energy of the n-GaN core-level peak following transfer of epitaxial graphene layers onto the n-GaN surfaces, Eicore is the initial binding energy of the core-level peak, EiV is the initial binding energy of the valence band maximum (EV) of GaN, and EVC is equal to (Eicore EiV ). All binding energies are measured relative to EF. For the calculation of the barrier height (q/p) at the graphene/p-GaN interface, the equation is expressed as q/p ¼ EG core EVC

ð2Þ

where EG core is the binding energy of the p-GaN core-level peak following transfer of epitaxial graphene layers onto the p-GaN surfaces. Figure 2a shows an example of the Ga 3d core level and the valence-band spectrum collected on a n-GaN sample without a graphene overlayer. EVC is calculated to be about 17.80 eV. This derived value is in good agreement with the reported value of 17.80 eV [21, 31, 32]. Figure 2b shows the Ga 3d core level at the graphene/n-GaN interface. The spectra determine the Ga 3d binding energy EG core relative to G EF. We can see that Ecore is equal to 20.35 eV. Therefore, q/n was calculated to be 0.85 eV, according to Eq. (1). Figure 3a shows an example of the Ga 3d core level and the valence-band spectrum collected on a p-GaN sample without a graphene overlayer. EVC is calculated to be about 17.80 eV. Figure 3b shows the Ga 3d core level at the

Fig. 3 The surface and interface characteristics are investigated using XPS measurements. a The left-hand spectrum shows the Ga 3d core-level peak on p-GaN without a graphene overlayer. The righthand figure presents the spectrum of the valence-band region. A linear fit is used to determine the energy of the valence-band edge. b Ga 3d core level at the graphene/p-GaN interface. The binding energy is referenced to the Fermi energy level

graphene/p-GaN interface. The spectra determine the Ga G 3d binding energy EG core relative to EF. We can see that Ecore is equal to 20.30 eV. Therefore, q/p was calculated to be 2.50 eV, according to Eq. (2). Schottky-barrier values are well described using either the Bardeen or Schottky limits [16]. In the Bardeen limit, the interface physics is mostly governed by interface states which, by accumulating free charge, change the charge distribution at the interface and cause EF of the semiconductor to be fixed (Fermi-level pinning). To determine the Fermi-level pinning/unpinning at the graphene/n-GaN and graphene/p-GaN interfaces, an analysis is conducted according to the Schottky–Mott limit. For graphene/n-GaN samples, the graphene work function (WG) is given by WG ¼ q/n þ v

ð3Þ

where v is the electron affinity of GaN (v = 4.10 eV). For graphene/p-GaN samples, WG is given by WG ¼ v þ Eg q/p ð4Þ

Fig. 2 The surface and interface characteristics are investigated using XPS measurements. a The left-hand spectrum shows the Ga 3d core-level peak on n-GaN without a graphene overlayer. The righthand figure presents the spectrum of the valence-band region. A linear fit is used to determine the energy of the valence-band edge. b Ga 3d core level at the graphene/n-GaN interface. The binding energy is referenced to the Fermi energy level

It is shown that the derived WG value (4.95 eV) from Eq. (3) is close to that the derived WG value (5.00 eV) from Eq. (4). To confirm the Fermi-level pinning/unpinning at the graphene/n-GaN and graphene/p-GaN interfaces, the actual work function of graphene was examined with the SKP5050 Scanning Kelvin probe. Figure 4a shows the images of the work-function difference (DW) between the probe and graphene (Au). The Au and graphene samples were placed together. The graphene sample is on the right-

123


Fig. 4 a Images of the work-function difference between graphene (Au) and probe. b C 1 s core-level spectra of graphene films

hand side and Au is on the left-hand side. It is found that the work function of graphene is lower than that of the probe (DW = 320 ± 10 meV) and the work function of Au is lower than that of the probe (DW = 130 ± 10 meV). If the work function of Au is 5.10 eV [33], the work function of graphene is calculated to be 4.91 ± 0.02 eV. This value is close to the derived values from Eqs. (3) and (4), indicating the Fermi-level unpinning at the graphene/ n-GaN and graphene/p-GaN interfaces. In addition, the determined work-function value (4.91 eV) of graphene from Kelvin probe measurements is larger than the reported value (4.6 eV [16] ) of undoped graphene. The deviation from this ideal graphene work function can be attributed to the lowering of the Fermi energy level due to hole doping of the graphene during the fabrication process. We deduce the induced structural imperfection (Fig. 1) by incorporating oxygen [28] that serves to increase the work function of graphene. Figure 4(b) shows the C 1 s corelevel spectra of graphene films. The signal of the C1 s was divided into two major peaks near 284.5 and 286.0 eV, corresponding to the sp2 hybrid (C–C) and the sp3 hybrid (C–O) bonds, respectively [34–36]. The existence of C–O bonds indicated the incorporation of oxygen into graphene. We suggest that the incorporation of oxygen into graphene may lead to p-type doping, thus increasing the work function of graphene. For graphene/n-GaN and graphene/pGaN heterojunctions, the schematic diagrams of the band alignment are shown in Fig. 5.

4 Conclusion In summary, the interface characteristics of graphene/n-GaN and graphene/p-GaN samples using XPS measurements are

123

Fig. 5 Energy band diagrams of a graphene/n-GaN and b graphene/ p-GaN heterojunctions (EC: the conduction band minimum)

investigated in this study. Quantitative analysis of heterojunction conditions has become important to develop highperformance electronic and optoelectronic devices. XPS makes it possible to extract a reliable q/n (q/p) value. For graphene/n-GaN (graphene/p-GaN) samples, the Schottky barrier height is 0.85 (2.50) eV. To confirm the Fermi-level pinning/unpinning at the graphene/GaN interfaces, an analysis is conducted according to the Schottky–Mott model. It is found that the Schottky barriers are sensitive to the work function of graphene and the Fermi energy level is not pinned at the graphene/GaN interface.

J Mater Sci: Mater Electron Acknowledgments The authors acknowledge the support of the Ministry of Science and Technology, Taiwan (Contract No. 103-2112-M-018-003-MY3) in the form of grants.

References 1. K.N. Parrish, D. Akinwande, Appl. Phys. Lett. 98, 183505 (2011) 2. D. Dragoman, M. Dragoman, R. Plana, J. Appl. Phys. 108, 084316 (2010) 3. C. Yim, N. McEvoy, G.S. Duesberg, Appl. Phys. Lett. 103, 193106 (2013) 4. Y.J. Lin, J.J. Zeng, Appl. Phys. Lett. 102, 183120 (2013) 5. J.J. Zeng, Y.J. Lin, Appl. Phys. Lett. 104, 233103 (2014) 6. C.C. Chen, M. Aykol, C.C. Chang, A.F.J. Levi, S.B. Cronin, Nano Lett. 11, 1863 (2011) 7. M. Mohammed, Z. Li, J. Cui, T. Chen, Nanoscale Res. Lett. 7, 302 (2012) 8. J.H. Lin, J.J. Zeng, Y.J. Lin, Thin Solid Films 550, 582 (2014) 9. J.J. Zeng, Y.J. Lin, Appl. Phys. Lett. 104, 133506 (2014) 10. C.C. Chen, C.C. Chang, Z. Li, A.F.J. Levi, S.B. Cronin, Appl. Phys. Lett. 101, 223113 (2012) 11. Y.F. Lin, W. Li, S.L. Li, Y. Xu, A. Aparecido-Ferreira, K. Komatsu, H. Sun, S. Nakaharaia, K. Tsukagoshi, Nanoscale 6, 795 (2014) 12. H. Zhong, Z. Liu, L. Shi, G. Xu, Y. Fan, Z. Huang, J. Wang, G. Ren, K. Xu, Appl. Phys. Lett. 104, 212101 (2014) 13. S. Tongay, M. Lemaitre, T. Schumann, K. Berke, B.R. Appleton, B. Gila, A.F. Hebard, Appl. Phys. Lett. 99, 102102 (2011) 14. S. Kim, J.M. Lee, D.H. Lee, W.I. Park, Thin Solid Films 546, 246 (2013) 15. H.W. Shiu, L.Y. Chang, K.H. Lee, H.Y. Chen, S. Gwo, C.H. Chen, Appl. Phys. Lett. 103, 081604 (2013)

16. S. Tongay, M. Lemaiyre, X. Miao, B. Gila, B.R. Appleton, A.F. Hebard, Phys. Rev. X 2, 011002 (2012) 17. J.H. Lin, Y.J. Lin, H.C. Chang, Appl. Phys. A 118, 361 (2015) 18. Y.J. Lin, J.J. Zeng, H.C. Chang, Appl. Phys. A 118, 353 (2015) 19. J.J. Zeng, Y.J. Lin, Mater. Chem. Phys. 145, 250 (2014) 20. Y.J. Lin, C.F. You, C.S. Lee, J. Appl. Phys. 99, 053706 (2006) 21. Y.J. Lin, Appl. Phys. Lett. 86, 122109 (2005) 22. Y.J. Lin, W.X. Lin, C.T. Lee, H.C. Chang, Jpn. J. Appl. Phys. 45, 2505 (2006) 23. Y.J. Lin, C.T. Lee, Appl. Phys. Lett. 77, 3986 (2000) 24. Q. Yu, J. Lian, S. Siriponglert, H. Li, Y.P. Chen, S.S. Pei, Appl. Phys. Lett. 93, 113103 (2008) 25. S.J. Chae, F. Gu¨nes¸ , K.K. Kim, E.S. Kim, G.H. Han, S.M. Kim, H.J. Shin, S.M. Yoon, J.Y. Choi, M.H. Park, C.W. Yang, D. Pribat, Y.H. Lee, Adv. Mater. 21, 2328 (2009) 26. R. Zhang, Z.S. Wang, Z.D. Zhang, Z.G. Dai, L.L. Wang, H. Li, L. Zhou, Y.X. Shang, J. He, D.J. Fu, J.R. Liu, Appl. Phys. Lett. 102, 193102 (2013) 27. Y. Wang, Z. Ni, T. Yu, Z.X. Shen, H. Wang, Y. Wu, W. Chen, A.T.S. Wee, J. Phys. Chem. C 112, 10637 (2008) 28. Z. Luo, N.J. Pinto, Y. Davila, A.T. Charlie, Johnson. Appl. Phys. Lett. 100, 253108 (2012) 29. J.R. Waldrop, R.W. Grant, Appl. Phys. Lett. 52, 1794 (1988) 30. K.M. Tracy, P.J. Hartlieb, S. Einfeldt, F. Davis, E.H. Hurt, R.J. Nemanich, J. Appl. Phys. 94, 3939 (2003) 31. V.M. Bermudez, J. Appl. Phys. 80, 1190 (1996) 32. J.R. Waldrop, R.W. Grant, Appl. Phys. Lett. 68, 2879 (1996) 33. H.B. Michaelson, J. Appl. Phys. 48, 4729 (1977) 34. Y.J. Lin, J.H. Lin, Appl. Surf. Sci. 311, 224 (2014) 35. Y.C. Lin, C.Y. Lin, P.W. Chiu, Appl. Phys. Lett. 96, 133110 (2010) 36. C. Petridis, Y.H. Lin, K. Savva, G. Eda, E. Kymakis, T.D. Anthopoulos, E. Stratakis, Appl. Phys. Lett. 102, 093115 (2013)

123