Fast Growth of Strain-Free AlN on Graphene-Buffered Sapphire

Transcription

1 Fast Growth of Strain-Free AlN on Graphene-Buffered Sapphire Yue Qi, 1,2,3 Yunyu Wang, 4,5 Zhenqian Pang, 6,7 Zhipeng Dou, 1,8,9 Tongbo Wei, *,4,5 Peng Gao, *1,8,9,10 Shishu Zhang, 1,2 Xiaozhi Xu, 1,3,11 Zhenghua Chang, 6,7 Bing Deng, 1,2 Shulin Chen, 1,8,9 Zhaolong Chen, 1,2 Haina Ci, 1,2,3 Ruoyu Wang, 1,2 Fuzhen Zhao, 12 Jianchang Yan, 4,5 Xiaoyan Yi, 4,5 Kaihui Liu, 1,10,11 Hailin Peng, 1,2,13 Zhiqiang Liu, 4,5 Lianming Tong, 1,2 Jin Zhang, 1,2,13 Yujie Wei, *, 6,7 Jinmin Li, *, 4,5 and Zhongfan *,1,2, 3,13 Liu 1. Center for Nanochemistry (CNC), College of Chemistry and Molecular Engineering, Peking University, Beijing , China 2. Beijing National Laboratory for Molecular Sciences, Peking University, Beijing , China 3. Academy for Advanced Interdisciplinary Studies, Peking University, Beijing , China 4. Research and Development Center for Solid State Lighting, Institute of Semiconductors, Chinese Academy of Sciences, Beijing , China 5. University of Chinese Academy of Sciences, Beijing , China 6. LNM, Institute of Mechanics, Chinese Academy of Sciences, Beijing, , China 7. School of Engineering Sciences, University of Chinese Academy of Sciences, Beijing , China 8. Electron Microscopy Laboratory, School of Physics, Peking University, Beijing , China 9. International Center for Quantum Materials, Peking University, Beijing , China 10. Collaborative Innovation Centre of Quantum Matter, Beijing , China. 11. State Key Laboratory for Mesoscopic Physics, School of Physics, Peking University, Beijing , China 12. College of Science, China University of Petroleum, Qingdao, , China 13. Beijing Graphene Institute (BGI), Beijing , China Corresponding authors: Zhongfan Liu Tongbo Wei Peng Gao Jinmin Li Yujie Wei 1. Transferred graphene for AlN growth. S1

2 Figure S1: AlN growth on transferred graphene on sapphire substrate. (a) SEM image for grphene grown on Cu(111) foils. (b, c) Large-scale SEM images of the transferred graphene from copper foils onto sapphire. (d) Magnified SEM image of the transferred monolayer graphene onto sapphire. (e) Raman spectrum of graphene transferred onto sapphire before AlN growth. (f) SEM image for AlN grown on graphene/sapphire. 2. AlN nucleation on graphene/sapphire and bare sapphire with the growth duration of ~3 min. Figure S2: (a) AlN nucleation on graphene/sapphire and bare sapphire with the nucleation time of ~3 min. (b, c) Magnified SEM images from the regions in the yellow and red rectangles in (a), showing the amounts of AlN nuclei of 76 and 47, corresponding to the nuclei density of 11/µm 2 and 7/µm 2, respectively. S2

3 3. SAEDs for different AlN islands on graphene/sapphire. Figure S3: SAED for different AlN islands on graphene/sapphire. (a) SEM image of AlN islands on graphene/sapphire. (b, c, d) SAED from different AlN islands show the same pattern, indicating the orientations of different AlN islands are well aligned. 4. Vertical one-dimensional (1D) Raman line mapping in the coalesced and non-coalesced AlN regions. When the growth of AlN films was finished, graphene became defective with intense D peak and peak broadening, which are mainly attributed to three reasons: (I) Etching effect introduced by NH3 (500 sccm) and H2 (12 SLM), which are used as the Al, N precursors and carrier gas, respectively, with the growth temperature of 1200 oc. Most of the defects are possibly terminated with N atoms, considering the N-rich atmosphere in the reaction system, and then bond with Al atoms to form the initial nuclei of AlN. (II) Serious strains introduced into graphene due to the mismatch in lattice and thermal expansion coefficient along with the strong G-AlN, G-Al2O3 binding (Figure 4g). Figure S12 is the DFT calculated result for the system of AlN/graphene/sapphire, which shows significant structural S3

4 deformation in graphene. (III) The other reason to cause the low-quality Raman spectra of graphene are the measurement technique. During the characterizations, the laser has to penetrate thick AlN films (~1.5 µm), the not ideal focusing and attenuation of the laser may also affect the quality of Raman spectra. Figure S4: (a) Vertical 1D Raman line mapping in the coalesced AlN regions. (b) Magnified image for the regions in the red box in (a). S4

5 Figure S5: (a) Vertical 1D Raman line mapping in the non-coalesced AlN regions. (b) Magnified image for the regions in the red box in (a). 5. CL spectra of AlN from AlN/sapphire and AlN/graphene/sapphire. Figure S6: (a) SEM image for non-coalesced and coalesced AlN domains on sapphire and S5

6 graphene/sapphire, respectively. (b) CL spectra of AlN from AlN/sapphire (red) and AlN/graphene/sapphire (black). The CL spectrum from AlN/graphene/sapphire shows a narrower FWHM of ~8.2 nm and stronger intensity than that from AlN/sapphire with the FWHM of ~10.4 nm (peaks at ~210.6 nm). 6. XRD and LEED characterizations of AlN/graphene/sapphire. (1) To characterize the large-scale epitaxial growth of AlN on graphene-covered sapphire, the XRD phi scan was preformed (inset in Figure 2a). The AlN epilayer was measured across the {10-11} Bragg reflections in skew geometry. Six peaks were observed and each one is 60 degrees apart, which reflects the 6-fold symmetry of the hexagonal AlN. For sapphire substrate, the measurement was conducted across {01-12} Bragg reflections. Three peaks were observed and each one is 120 degrees apart, indicating the 6-fold symmetry. It is clearly that the AlN peak is 30 degrees away from the neighbor sapphire peak, which means there is a rotation of 30 degrees between the AlN epilayer and sapphire substrate. The AlN epilayer was thus in registry with the sapphire substrate in such a way that [1-100]AlN [11-20]sapphire, which is the same as traditional AlN epitaxy on sapphire substrate. The XRD phi scan of AlN/graphene/sapphire presents the good epitaxial relationship between AlN and sapphire in a large scale. Figure S7a is the XRD characterization of AlN/graphene/sapphire, which confirms the epitaxial relationship between AlN and sapphire and indicates that growth direction (out-of-plane orientation) is (0002). In addition, LEED characterization was also shown in Figure S7b, which presented the same patterns of AlN at different locations on the sample, indicating the large-scale epitaxial growth of AlN on graphene-covered sapphire. S6

7 Figure S7: (a) XRD characterization of AlN/graphene/sapphire. (b) LEED characterizations of AlN/graphene/sapphire at different locations on the sample, showing the same patterns of AlN, indicating the large-scale epitaxial growth of AlN on graphene-covered sapphire. (2) Analysis for the lattice parameters change based on XRD data: Analysis for a-axis strain: The lattice parameters of AlN can be calculated through the equations below according to the XRD data. S7

8 Based on the equations above, θ (002) was achieved from (002) ω-scan, then c parameter is acquired. From (102) ω-scan, θ (102) was achieved, and then a parameter is determined based on c. The diffraction peaks and calculated lattice parameters are listed in Table 1. Table 1: Diffraction peaks and the calculated lattice parameters in the systems of AlN/graphene/sapphire, AlN/graphene and bulk AlN. From Table 1, we can conclude that the AlN films on sapphire is largely compressed in-plane. In contrast, the compressive strain is significantly relaxed by introducing graphene between AlN and sapphire. Analysis for c-axis strain: The out-plane c-axis strain in AlN is the tensile strain in AlN/sapphire, which is also well released after the introduction of graphene between AlN and sapphire. 7. Element mapping of AlN/sapphire. S8

9 Figure S8: Element mapping of AlN/sapphire. (a) TEM image for AlN/sapphire. (b, c, d) Al, N, O mapping of AlN/sapphire, respectively. 8. SAED patterns of single AlN and graphene/sapphire Figure S9: Cross sectional SAED patterns of single AlN (a) and graphene/sapphire (b). 9. Raman spectra of bare sapphire, AlN/sapphire and AlN/graphene/sapphire. For bare sapphire, three are three typical peaks located at ~419.2 cm -1, ~580.1 cm -1 and ~751.2 cm -1 (Figure S7a). The peaks around cm -1 (E 2 (low)), cm -1 (E 2 (high)) and cm -1 (A 1 S9

10 (LO)) are from AlN. Figure S7d and e shows the E 2 (high) peaks and A 1 (LO) peaks of AlN from AlN/sapphire (red lines) and AlN/graphene/sapphire (black lines), respectively. Figure S10: (a-c) Full spectra from the samples of bare sapphire (a), AlN/sapphire (b) and AlN/graphene/sapphire (c), respectively. (d, e) Raman spectra of E 2 (high) (d) and A 1 (LO) (e) of AlN, respectively. 10. Raman peak positions of AlN, sapphire and graphene from the pristine samples, the samples of AlN/sapphire and AlN/graphene/sapphire. Figure S11: Raman peak positions of AlN, sapphire and graphene from the pristine samples, the samples S10

11 of AlN/sapphire and AlN/graphene/sapphire. 11. DFT calculations about the AlN-Al 2 O 3 binding energy (BE) and the internal stress in AlN in the systems of AlN/sapphire and AlN/graphene/sapphire. First-principles density functional theory (DFT) calculations for the interaction between AlN and sapphire were performed using the Vienna Ab initio Simulation Package (VASP) code. 1, 2 The projector augmented wave (PAW) pseudopotentials 3, 4 and the generalized gradient approximation (GGA) of the Perdew-Burke-Ernzerhof (PBE) functional 5, 6 are used. A plane-wave basis set with a kinetic-energy cut-off of 520eV and a Monkhorst-Pack 7 k-point mesh of are used for all the interaction calculations. Periodic boundary conditions are applied in the two in-plane directions for all the calculations conducted here. All structures are relaxed using a conjugate gradient (CG) algorithm until the atomic forces are converged to 0.01eV/Å. To eliminate the interactions between periodic images of system, a vacuum space of larger than 10Å was used. The lattice constants of AlN and sapphire obtained by DFT calculations are 3.13 Å and 4.80 Å, separately. Based on the previous experimental 8 and theoretical 9, 10 epitaxial relationships, we use 3 3 cell of AlN matching a 1 1 cell of sapphire. In this case, the lattice constant of AlN is restrained as the constant of sapphire. So it is equal to apply about 11% strain to the AlN and it have no impact to the following research of the interaction between these two materials. 11 Similar to the AlN/sapphire systems, the systems with the introduction of a 2 2 cell of graphene are adopted the same calculation conditions. Considering the experiment results, we calculated the interface interaction using the end faces of N for AlN and Al for Al 2 O 3. To eliminate the influence of interaction calculation from thickness, the thickness of AlN and Al 2 O 3 are Å and Å, respectively, which is large enough to calculate the interfacial effect. Three layers of bottom atoms in sapphire are fixed in all the simulations. There exists charge exchange between N from AlN and Al from Al 2 O 3, as illustrated in Figure 4b. This binding effect leads to the epitaxial growth of AlN. When we introduced the graphene between AlN and sapphire, the charge exchange happened among all the interfaces (Figure 4e). The corresponding strong interaction leads to the distortion of graphene layer (Figure S12). S11

12 Figure S12: Distorted graphene located between AlN and sapphire obtained after DFT calculations. References (1) Kresse, G.; Furthmüller, J. Comp. Mater. Sci , (2) Kresse, G.; Furthmüller, J. Phys. Rev. B , (3) Blöchl, P. E. Phys. Rev. B 1994, 50, (4) Kresse, G.; Joubert, D. Phys. Rev. B 1999, 59, (5) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996, 77, (6) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1997, 78, (7) Monkhorst, H. J.; Pack, J. D. Phys. Rev. B 1976, 13, (8) Ponce, F. A.; Major, J. S.; Plano, W. E.; Welch, D. F. Appl. Phys. Lett. 1994, 65, (9) Efimov, A. N.; Lebedev, A. O. Thin Solid Films 1995, 260, (10) Dovidenko, K.; Oktyabrsky, S.; Narayan, J.; Razeghi, M. J. Appl. Phys. 1996, 79, (11) Felice, R. D.; Northrup, J. E. Appl. Phys. Lett. 1998, 73, S12