ABSTRACT
We show that in spatially resolved reflectance anisotropy (RA) spectrometers, off-axis optical rays introduce a spurious signal component that cannot be addressed by optical alignment. Such a component is associated with the difference between the reflectivities s and p of the sample and depends, in a complex manner, on the incidence position of the incident light on the surface of the sample. We report a data-reduction procedure to easily identify and remove spurious RA signals associated with the off-axis optical rays, based on the singular value decomposition analysis of spatially resolved RA spectra. We validated this approach by developing a spatially resolved RA spectrometer based on an 8 × 8 multi-anode photomultiplier (PMT). The PMT allowed the use of phase-sensitive detection techniques to enhance the signal-to-noise ratio, which is essential for the evaluation of the proposed data reduction procedure.
ABSTRACT
Reflectance anisotropy spectroscopy (RAS) is a highly sensitive optical probe for the real-time study of the epitaxial growth of zincblende semiconductors. Here we report on (1) non-equilibrium RAS spectra acquired in real time during the homoepitaxial growth of GaAs, and (2) RAS spectra for GaAs surfaces under equilibrium with several arsenic overpressures. We show that in both cases RAS spectra can be decomposed into two basic components, each with a characteristic line shape. We further show that both dynamic and equilibrium RAS spectra are described by the same pair of basic components. We conclude that the time evolution of non-equilibrium RAS spectra acquired during the epitaxial growth can be described in terms of RAS spectra for equilibrium surfaces. The results reported here should be useful for the interpretation of the physics underlying the rapid time evolution of dynamic RAS spectra during the first monolayer growth. Thus, we show that RAS constitutes a valuable tool for the study of epitaxial growth mechanisms.
ABSTRACT
We report on a rapid, 32-channel reflectance-difference (RD) spectrometer with sub-second spectra acquisition times and ΔR/R sensitivity in the upper 10(-4) range. The spectrometer is based on a 50 kHz photo-elastic modulator for light polarization modulation and on a lock-in amplifier for signal harmonic analysis. Multichannel operation is allowed by multiplexing the 32 outputs of the spectrometer into the input of the lock-in amplifier. The spectrometer spans a wavelength range of 230 nm that can be tuned to cover E(1) and E(1) + Δ(1) transitions for a number of III-V semiconductors at epitaxial growth temperatures, including GaAs, InAs, AlAs, and their alloys. We present two examples of real-time measurements to demonstrate the performance of the RD spectrometer, namely, the evolution of the RD spectrum of GaAs (001) annealed at 500 °C and the time-dependent RD spectrum during the first stages of the epitaxial growth of In(0.3)Ga(0.7)As on GaAs (001) substrates.
ABSTRACT
We describe a microreflectance difference (microRD) spectrometer based on a charge coupled device (CCD), in contrast to most common RD spectrometers that are based on a photomultiplier or a photodiode as a light detector. The advantage of our instrument over others is the possibility to isolate the RD spectrum of specific areas of the sample; thus topographic maps of the surface can be obtained. In our setup we have a maximum spatial resolution of approximately 2.50 microm x 2.50 microm and a spectral range from 1.2 to 5.5 eV. To illustrate the performance of the spectrometer, we have measured strains in mechanically polished GaAs (001) single crystals.
ABSTRACT
We report photoreflectance-difference and reflectance-difference measurements on reconstructed GaAs (001) surfaces. From these data the linear and quadratic electro-optic coefficient spectra are determined in the important 2.8-3.4 eV spectral region. The surface strain and fields induced by the surface reconstruction are also determined. We show experimentally that between c(4x4) and (2x4) surfaces, there is an inversion of the surface electric field which we attribute to a direct piezo-electric effect related to the surface strain induced by reconstruction.
