VOLUME 61, NUMBER 2 PHYSICAL REVIEW LETTERS 11 JULY 1988 Defects and Impurities at the SilSi(100) Interface Studied with Monoenergetic Positrons Peter J. Schultz and E. Tandberg Department of Physics, The University of Western Ontario, London, Ontario, Canada N6A 3K7 K. G. Lynn and Bent Nielsen Brookhaven National Laboratory, Upton, Ne~ York 11973 and T. E. Jackman, M. W. Denhoff, and G. C. Aers Microstructural Sciences Laboratory, National Research Council of Canada, Otta~a, Ontario, Canada K1AOR6 (Received 21 March 1988) Positrons implanted with varying energies (0-20 keV) have been used to study silicon epilayers grown by molecular-beam epitaxy on Si(100) substrates. Defects at the initial growth interface and throughout the overlayer have been observed and depth profiled. In addition, field-driven positron drift observed in some of the epilayers is shown to be consistent with estimated concentrations of (active) in- terfacial impurities. The study demonstrates that positrons can be used nondestructively to profile structural defects and electric fileds in thin films and at interfaces. PACS numbers: 68.35.Dv, 68.55.8d, 78.70.8j Molecular-beam epitaxy (MBE) is now extensively used for creating group-IV multilayer structures. The composition of these superlattices can be tailored to pro- duce structures with specific physical or electronic prop- erties. As in any growth process, impurities, disloca- tions, or other types of structural defects formed in the overlayers can have a detrimental effect on the electronic properties of the material. The origin and consequence of damage in artificially grown semiconductors is an area of intense study, even though there are relatively few ways to profile buried defects in dilute quantities. 4 In the last few years, several studies have demonstrated that positrons implanted with varying energies can be used to study defects nondestructively in the near-surface region of solids. ' In the present Letter, we report on the first results of a study of the properties of MBE-grown Si/Si(100) epilayers (single layer). The data show that positrons are sensitive to some types of structural defects concentrated at the initial growth interface and distri- buted throughout the overlayer of some Si/Si(100) sam- ples. The data also show evidence of electric-field- induced positron mobility, which can be directly attribut- ed to active impurities known to be at the interface. The epilayers were grown at =700'C without inten- tional doping to a thickness of = 3000 A at a rate of = 3 A/s. Substrates were prepared by growth of an ex situ sacrificial oxide in an ultraviolet/ozone reactor which was removed by heating to =900'C for 5 min, followed by 10 min at the same temperature with a low ( &0.1 A/s) Si fiux. " Two positron beams were used to study the epilayers, which provide beams of monoenergetic positrons (=10 to 10 e+/s) that are tunable from near 0 to several tens of kiloelectronvolts. These facilities are described elsewhere, 'z" and studies of solids with vari- able-energy positrons (including details of the analysis summarized below) are reviewed by Schultz and Lynn. ' The depth distribution or profile of positrons implant- ed into a monatomic material such as Si can be approxi- mated by the derivative of a Gaussian, 7' with a mean depth (in microns) of z =(0. 04/p)E" where p gm/cm3 is the mass density of the target and E keV is the incident positron energy. Once thermalized, the positrons diffuse through the solid, annihilating from this freely diffusing state, or after being trapped in bulk-defect or surface localized states. When a positron annihilates with an electron, the 511-keV y rays are Doppler broadened because of the finite momentum of the annihilation pair. This distribution is measured with two different high-resolution intrinsic Ge detectors (reso- lution & 1. 5 keV at 511 keV) and is characterized by the "5 parameter, "' defined as the integral of a central fixed portion of the annihilation line normalized by the total intensity. The probability of positron annihilation from various states can be predicted by the steady-state diffusion- annihilation equation, ' which is given (in one dimen- sion) by D+ t)'n(z)/8z' — [). tt+ vC(z)1n(z) — Bvqn(z )/8z+ P(z ) =0, (2) where n(z) is the positron density as a function of depth, 1988 The American Physical Society 187