Silicon surface passivation by hydrogen termination: A comparative study of preparation methods

Abstract

Clean silicon surfaces having low carbon and oxygen contamination are necessary for good epitaxial overgrowth. Methods for low-temperature preparation of clean surfaces are needed to fully enable low-temperature fabrication processes on silicon. In this paper silicon surfaces are compared for residue and chemical passivation after (i) hydrogen termination by various low-temperature, wet-chemical techniques, (ii) cleaving in ultrahigh vacuum (UHV), and (iii) ion sputtering. Surface residue was characterized with x-ray photoelectron spectroscopy (XPS) and small-spot Auger electron spectroscopy (AES). Low-energy electron diffraction (LEED) was also used. Evaluations of passivation were done by exposing the cleaned surfaces to various environments, e.g., UHV, N2 gas, and room air. We have obtained especially promising results with a technique whereby the wafer surface oxide is etched using an HF-alcohol reagent in a flowing nitrogen atmosphere at room temperature while the wafer is spinning, i.e., a spin etch. We have evaluated the tolerances for this process and have found quite practical requirements, which we discuss in detail. Typically, after spin etching the XPS of Si(100) surfaces indicated 0.03 ML (monolayer) of total carbon residue (from C 1s core-level emission) and about 0.005 ML each of oxygen and fluorine, while all such residue was below our AES detection limits. The Si 2p core level at shallow emission angles was free of any chemically shifted components within the noise and linewidth resolution limits. Brief exposure of these passivated surfaces to room air increased the total C and O residue slightly, while LEED patterns remained unreconstructed. By comparison, samples etched by dipping in HF or sputtered by Ar+ ions showed tenfold more surface residue, while cleaved sample surfaces were vastly more reactive.