AVS 59th Annual International Symposium and Exhibition
    Spectroscopic Ellipsometry Focus Topic Monday Sessions
       Session EL+TF+BI+AS+EM+SS-MoA

Paper EL+TF+BI+AS+EM+SS-MoA9
Preparation of Abrupt LaAlO3 Surfaces Monitored by Spectroscopic Ellipsometry

Monday, October 29, 2012, 4:40 pm, Room 19

Session: Spectroscopic Ellipsometry: From Organic and Biological Systems to Inorganic Thin Films
Presenter: S. Zollner, New Mexico State University
Authors: C.M. Nelson, New Mexico State University
M. Spies, New Mexico State University
L.S. Abdallah, New Mexico State University
S. Zollner, New Mexico State University
Y. Xu, New Mexico State University
H. Luo, New Mexico State University
Correspondent: Click to Email
LaAlO3 is a polar perovskite oxide, used as a single-crystal substrate in oxide epitaxy. It has created much interest for novel electronic applications, because a two-dimensional electron gas is formed at LaAlO3/SrTiO3 heterostructures. The purpose of our work is twofold: First, we are interested in an accurate determination of the complex refractive index of LaAlO3 at room temperature. Second, we studied the impact of various cleaning methods on the abruptness of the LaAlO3surface.

We obtained a commercial single-side polished LaAlO3substrate with 2-inch diameter and a (100) pseudo-cubic surface orientation.The surface was polished with an rms roughness below 0.8 nm. We determined the ellipsometric angles ψ and Δ for LaAlO3 at 300 K from 0.7 to 6.5 eV. For a bulk insulator with a clean smooth surface, the phase change Δ should be zero or π below the band gap. In practice, this never happens, because surfaces are covered with overlayers (adsorbed organic or water vapors). Surface roughness has a similar effect on the ellipsometric spectra as a surface overlayer. Even for an abrupt bulk/air interface, there is a thin (~0.5nm) transition region where the electron wave functions leak from the crystal into the ambient. For the as-received sample, the data were described with a Tauc-Lorentz model for LaAlO3, plus 2.1 nm of surface layer thickness (described as an effective medium with 50% density of the bulk). After ultrasonic cleaning in acetone, the overlayer thickness decreased to 1.8nm. Next, we mounted the wafer in a UHV cryostat, pumped down to below 10-8Torr, and acquired an ellipsometric spectrum at 70°. The surface layer thickness was reduced to 1.2 nm, presumably because a part of the adsorbed surface layer (especially water) desorbed under vacuum.

So far, everything worked as expected, but here it gets interesting: We heated the sample to 700 K for about an hour to desorb the remaining surface overlayer. After cooling down to 300 K, we measured the ellipsometric angles again at 70° angle of incidence from 0.7 to 6.5 eV. The ellipsometric angle Δ at 2 eV was reduced to below 0.2°, consistent with a surface layer thickness of less than 1 Å, much less than the surface roughness specified by the supplier (8 Å).

In conclusion, a macroscopically smooth and clean LaAlO3surface was prepared by ultrasonic cleaning of the wafer in acetone, followed by heating in UHV to 700 K. The resulting surface layer thickness was below 1 Å, as measured by spectroscopic ellipsometry. We will report Tauc Lorentz parameters. We will also describe the temperature dependence of the LaAlO3dielectric function from 77 to 700 K. This work was supported by NSF (DMR-11104934).