INTEGRATION OF FUNCTIONAL OXIDES WITH SILICON

Wang, Zhe

INTEGRATION OF FUNCTIONAL OXIDES WITH SILICON

Abstract

Perovskite oxides exhibit the full spectrum of physical properties including ferroelectricity, ferromagnetism, multiferroicity, and superconductivity with unparalleled performance in many cases. In recent decades, there has been tremendous interest both in the physical origin of these properties and fabricating devices to make use of them. High structural quality usually correlates with maximizing these properties and thus growing thin films of perovskite oxides with high structural perfection is relevant to both fundamental physics research and device applications. To date, most films of oxide perovskites with high structural quality have been grown on single-crystal oxide substrates that typically have excellent crystalline perfection, but suffer from being costly and available in only small sizes (typically no larger than 10x10 mm^2). To exploit the stellar functional properties of perovskite oxides in semiconductor devices—properties that conventional semiconductors do not possess—and the mature fabrication processes developed by the semiconductor industry, it is desirable to grow these functional oxides directly on silicon, the workhorse of the semiconductor industry. The growth of perovskite oxides or for that matter any epitaxial oxides directly on silicon is, however, difficult due to the reactive nature of the silicon surface to oxygen. Formation of an amorphous SiO2 layer on top of silicon can interfere with the epitaxial growth. To date only a few oxides have been epitaxially grown directly on silicon; these include BeO, MgO, SrO, BaO, .... Among these SrTiO3 stands out as its use has yielded overlying perovskite oxide layers with the best functional properties on silicon substrates. This is due to the relatively high crystalline perfection of (001) SrTiO3 on (001) Si and its chemical and structural compatibility with overlying functional perovskite oxides. Beginning in the 1980’s significant effort has been dedicated to the growth of SrTiO3 on silicon. Despite this prior work, there is plenty of room for improvement as the structural perfection of SrTiO3 films on silicon pales in comparison to that of SrTiO3 single crystals. In the first part of this thesis, I present my work on improving the crystalline perfection of epitaxial (001) SrTiO3 on (001) Si. The films are grown by molecular-beam epitaxy (MBE). Using a carefully controlled growth recipe consisting of multiple steps, a rocking curve full width at half maximum (FWHM) below 0.03° for the SrTiO3 002 peak has been achieved for SrTiO3 films ranging from 2 to 300 nm thick. These narrow rocking curves are by far the narrowest ever achieved for SrTiO3/Si and are comparable to SrTiO3 single crystals. Although this might at first sound like I solved the SrTiO3=Si growth problem and have made SrTiO3 films with comparable structural perfection to SrTiO3 single crystals, measurements by scanning transmission electron microscopy reveal that my films contain huge densities of threading dislocations (> 10^11 cm^-2). I show that the narrow rocking curve of the 002 SrTiO3 peak in my films is because of the insensitivity of this particular peak to threading dislocations with pure edge character. Because of the highly anisotropic strain introduced by the defects in my SrTiO3/Si films, characterizing them in a single direction (as is typically done) is insufficient. I conclude that the out-of-plane orientation variation of the subgrains (mosaic spread) of my SrTiO3 films on silicon has been reduced, but the in-plane mosaic spread remains the biggest challenge to overcome. To improve the in-plane order, I developed a growth method for SrTiO3 on silicon that not only works at record low growth temperatures (below 450 °C), but also provides the highest in-plane and out-of-plane structural perfection (lowest mosaic spreads) ever achieved. Specifically, the in-plane FWHM of the 101 SrTiO3 peak in phi is under 0.1°, which is far narrower than the previous record of 0.5°. At the same time that these films have narrow in-plane FWHM, their out-of-plane FWHM (a rocking curve in omega) is under 0.01°. This means that these SrTiO3/Si films grown with my new low-temperature method possess the highest structural perfection of all SrTiO3/Si films reported to date. In the second part of this thesis, three examples of applications of SrTiO3-buffered silicon are shown. Epitaxial, twin-free LaAlO3 with record low out-of-plane mosaic spread (FWHM of 002 peak is 0.02°) on silicon was grown on a 5 unit-cell-thick SrTiO3 buffer layer. Epitaxial 14 nm thick SrRuO3 with record residual resistivity ratio on silicon of 11 was grown on a 14 nm thick SrTiO3 buffer layer. This value is 3 times higher than the best reported SrRuO3 grown on silicon and is comparable to the very best SrRuO3 films grown on oxide single crystals by all thin film growth methods other than MBE. Lastly, a high mobility epitaxial La-doped BaSnO3 film was grown on silicon using an intervening 18 nm thick SrTiO3 buffer layer. At a carrier concentration of 1.4e10^20 cm^-3, the room temperature mobility of the La-doped BaSnO3 film is 128 cm^2 V^-1 s^-1, and the resistivity is 3.6e10^-4 cm. Theses values are again comparable to those of the best reported La-doped BaSnO3 films grown by all methods other than MBE on single-crystal oxide substrates. I conclude this thesis by discussing possible device applications that build upon the materials that I have epitaxially integrated with silicon.

Date Issued

2018-12-30

Keywords

silicon; Applied physics; SrTiO3; oxides; Materials Science; molecular-beam epitaxy

Committee Chair

Schlom, Darrell

Committee Member

Kourkoutis, Lena Fitting
Fennie, Craig James

Degree Discipline

Applied Physics

Degree Name

Ph. D., Applied Physics

Degree Level

Doctor of Philosophy

Rights

Attribution-NonCommercial-NoDerivatives 4.0 International

Rights URI

https://creativecommons.org/licenses/by-nc-nd/4.0/

Types

dissertation or thesis