THOMAS YOUNG CENTRE
THE LONDON CENTRE FOR THE THEORY AND SIMULATION OF MATERIALS
Advances in metal/oxide interface science – the power of a few in-diffused metal atoms
Professor Scott Chambers, Physical Sciences Division, Pacific Northwest National Laboratory, Richland, WA, USA
Most metals form Schottky barriers when placed in contact with most semiconductors. This unfortunate fact of life makes it difficult to form low-resistance Ohmic contacts, which are essential for device fabrication. To get around this, a mix of metals is typically co-evaporated and then annealed to the eutectic point of the mix, resulting in a rather messy interface. This situation exists for compound III-V and II-VI, as well as for oxide semiconductors. Against this backdrop, we ask whether an epitaxial metal might yield better results. We consider contacting Nb-doped SrTiO3(001) (n-STO). Au, which should be epitaxial on STO by virtue of its symmetry and lattice parameters, is not epitaxial  and generates a classic Schottky barrier . However, Cr, which has nearly the same lattice mismatch to STO as Au, does grow epitaxially and forms a nearly ideal Ohmic contact, at least when deposited in ultrahigh vacuum at a substrate temperature of ~550oC . The contact resistance at room temperature, as measured by the van der Pauw method, is ~65 mW/square, the lowest every reported to the best of our knowledge. What is the difference between these two heterojunctions? Previous experimental studies have concluded that both interfaces are atomically abrupt [1,4]. Yet, their electronic properties are completely different. We have carried out a detailed investigation of Cr/ n-STO(001) using scanning transmission electron microscopy/electron energy loss spectroscopy, together with x-ray and ultraviolet photoelectron spectroscopy (including band mapping), and first-principles theory (in collaboration with Prof. Peter Sushko) to gain understanding . ~1-2 monolayers of Cr are found to diffuse into the STO and occupy interstitial sites within the first three atomic planes (two TiO2 and one SrO). Density functional theory calculations reveal that these Cr interstitial species stabilize the entire film by acting as an “anchor”. Both theory and experiment show that Cr epitaxial film growth with Cr in-diffusion results in charge transfer from interstitial Cr to structural Ti(IV) near the interface, effectively metalizing the STO. This in turn prevents Schottky barrier formation, and generates an ultra-low resistance Ohmic contact. In contrast, Au interacts very weakly with STO and does not diffuse into the STO, precluding both strong adhesion (a requirement of epitaxy) and near-interface metallization in the STO.
 F. Silly and M. R. Castell, Phys. Rev. Lett. 96, 086104 (2006).
 Y. Hikita, M. Kawamura, C. Bell, H. Y. Hwang, Appl. Phys. Lett. 98, 192103 (2011).
 C. Capan, G. Y. Sun, M. E. Bowden, S. A. Chambers, Appl. Phys. Lett. 100, 052106 (2012).
 Q. A. Fu and T. Wagner, Surf. Sci. 601, 1339 (2007).
 S. A. Chambers, M. Gu, P. V. Sushko, H. Yang, C. Wang, N. D. Browning, submitted (2013).