Key Parameters for Single Crystalline ZnO Film Growth by Magnetron Sputtering via Inverted Stranski-Krastanov Mode

ZnO has attracted much attention as a material for optoelectronic devices because of the wide and direct band gap of 3.37 eV and the large exciton binding energy of 60 meV[1]. For these applications, ZnO films are often grown on sapphire substrate by radio frequency (RF) magnetron sputtering, but the large lattice mismatch of 18% between ZnO and sapphire brings about Stranski-Krastanov (SK) mode growth, two-dimensional (2D) growth of a few monolayers followed by 3D island formation, resulting in polycrystalline ZnO films [3]. Recently, single crystalline ZnO films have been successfully fabricated on sapphire substrates via 3D to 2D growth mode transition: inverted SK mode[4,5], which has been realized controlling the surface energy through the manipulation of adsorption-desorption behavior of nitrogen atoms on the surface[5]. However, the key parameters for the growth of high-quality ZnO films grown via the inverted SK mode are unclear. Here, we investigate effects of 3D island layers on the quality of the subsequently grown 2D layers to figure out the desirable properties of the buffer layers. We define a figure of merit (FOM) of ZnO films as a reciprocal of a product of full width at half maximum (FWHM) of (002) plane x-ray rocking curves (XRC) and root-mean-square (RMS) roughness and discuss the correlations between the FOM and the properties of the buffer layers.<br/>All films were fabricated by RF magnetron sputtering using two ZnO ceramic targets (2 inch, &gt; 99.99 purity), which were sputter-cleaned for 15 min prior to deposition. First, 10-nm thick buffer layers packed with 3D islands were grown on sapphire substrates (10×10 mm²). The substrate temperatures (<i>T</i>_buffer) were 760–830°C. The RF power supplied to each cathode was 100 W. Ar and N₂ gasses were used, and the flow rates were 24.0, and 1.0 sccm, respectively. The total gas pressure was 0.35 Pa. Next, 1000-nm-thick ZnO films were deposited on the buffer layers at 800°C. Ar and O₂ gasses were used, and the gas flow rates were 45.0, and 5.0 sccm, respectively. The total gas pressure was 0.70 Pa.<br/>Atomic force microscopy (AFM) revealed that single-crystalline 1000-nm-thick ZnO films grow in inverted SK mode (3D to 2D transition) for <i>T</i>_buffer = 760– 800°C. A positive correlation with the coefficient of 0.63 is observed between the FOMs of the 1000-nm-thick films and 1/(FWHM_XRC(101)×<i>R</i>_q) of the buffer layers. Here FWHM_XRC(101) and <i>R</i>_q are the (101) plane XRC FWHM and the surface RMS roughness, respectively. On the other hand, no clear correlation is observed between the FOM and 1/(FWHM_XRC(002)×<i>R</i>_q), where FWHM_XRC(002) is the (002) plane XRC FWHM. These results indicate that the in-plane orientation of the buffer layers is a key for the quality of ZnO films. The correlation coefficient further increases to 0.80 when the FOMs are plotted against <i>ξ</i>/(FWHM_XRC(101)×<i>R</i>_q), where <i>ξ </i>is the lateral correlation length of the surface of the buffer layers. We determined <i>ξ</i> from AFM images deriving the height–height correlation functions. Since <i>ξ</i> provides a measure of the surface diffusion length of adatoms, we consider that the buffer layers with large <i>ξ</i> allow adatoms to reach the lowest thermodynamically favored lattice positions. As a result, high-quality single crystalline ZnO films, where the (002) plane XRC-FWHM and the RMS roughness are 0.05° and 1.5 nm, respectively, have grown on the buffer layers with <i>ξ</i> of 18.7 nm. Our findings provides a new measure of the buffer layers for the growth of high-quality single-crystalline films in the inverted SK mode.<br/>This work was supported by JSPS KAKENHI Grant Numbers JP21H01372, JP21K18731, NTT collaborative research, Toyota Riken Scholar, The Murata Science Foundation.<br/><b>Reference</b><br/>[1] A. Moezzi <i>et al</i>., Chem. Eng. J. <b>185</b>, 1 (2012).<br/>[2] Ü. Özgür <i>et al.</i>, J. Appl. Phys. <b>98</b>, 041301 (2005).<br/>[3] R. D. Vispute <i>et al.</i>, Appl. Phys. Lett. <b>70</b>, 2735 (1997).<br/>[4] T. Tinjod <i>et al.</i>, J. Alloys Compd. <b>371</b>, 63 (2004).<br/>[5] N. Itagaki <i>et al.</i>, Sci. Rep. <b>10</b>, 4669 (2020).