Growth of Perovskites with Crystalline Interfaces on Si(100)
- PDF / 2,555,856 Bytes
- 8 Pages / 612 x 792 pts (letter) Page_size
- 12 Downloads / 183 Views
E7.3.1
Growth of Perovskites with Crystalline Interfaces on Si(100) G. J. Norga1, ‡, A. Guiller1, C. Marchiori1,2, J. P. Locquet1, H. Siegwart1, D. Halley1, C. Rossel1, D. Caimi1, J. W. Seo3, and J. Fompeyrine1 1
IBM Research, Zurich Research Laboratory, 8803 Rueschlikon, Switzerland Laboratorio MDM - INFM, Via C. Olivetti 2, 20041 Agrate Brianza (MI), Italy 3 IPMC, Ecole Polytechnique Fédérale de Lausanne, 1015 Ecublens, Switzerland 2
ABSTRACT The main challenges involved in the growth of an epitaxial oxide film with a crystalline interface to silicon are reviewed: (1) structural matching of the oxide and semiconductor lattices; (2) thermodynamic energy stabilization at the semiconductor–oxide interface, and (3) kinetic control over oxygen motion throughout the deposition process. We report on how this approach can be used to grow epitaxial perovskites of high structural quality from the (Ba, Sr)(Zr,Ti)O3 family with crystalline interfaces on Si (100) by molecular-beam epitaxy.
INTRODUCTION ABO3 perovskites such as SrTiO3, BaZrO3 and SrZrO3 grown hetero-epitaxially on Si(100) by molecular-beam epitaxy (MBE) have received significant interest in the high-k community because their large dielectric constants open the perspective of gate dielectrics with very low equivalent oxide thickness (EOT < 5 Ǻ) [1-6]. Moreover, leakage can be kept low, mainly because a substantially greater physical thickness can be afforded than in “conventional” amorphous high-k materials. A crucial requirement for obtaining such a low EOT is that the formation of amorphous SiO2-containing interfacial phases be kept under tight control. The materials reported to date can be subdivided in two broad classes: (1) crystalline oxides grown epitaxially on silicon but with a thin amorphous SiO2 layer at the interface [2-6], and (2) crystalline oxides with a crystalline interface on Si [1, 4]. This second category represents a novel type of insulator–semiconductor heterointerfaces. Only a very limited set of experimental data is currently available to suggest how bonding and structural imperfections affect the metaloxide-semiconductor (MOS) properties of the epitaxial oxide/semiconductor interface (such as the interface trapped charge, Dit and band offsets) [7]. Different classes of oxides exist that can be structurally matched with Si. For NaCl- and fluorite-type oxides, the lattices can be made to match cube-on-cube with Si. For perovskitestructure oxides, such cube-on-cube matching is not possible. However, by a 45° rotation in the plane of the film, the Si(100) and perovskite planes can be made to match approximately (orientational relationships: (001)perovskite || (001)silicon and [110]perovskite || [100]silicon). Examples of systems that can be matched to Si(100) are (1) (Ba, Sr)O (NaCl structure); (2) perovskites of the (Ba,Sr)(Zr,Ti)O3 family, and (3) fluorite-structure oxides such as (La, Y)2O3 and La2Zr2O7 (pyrochlore). Because of the minimum surface energy of the (111) planes in the fluorite structure, fluorite-structure oxides are espe
Data Loading...
