The experimental properties are critiqued that relate the midgap flaw concentration in semi-insulating GaAs, and the resulting near-infrared transmittance of a polished wafer. Since quantitative information of such flaw concentrations is desirable even for an optical thickness αt≪1, a highly stable and accurate experimental arrangement is described, which permits a meaningful evaluation even when αt≂0.01. (The transmittance is then almost Tmax, as set by the substantial reflectance losses.) This system permits mapping over a wafer’s area, by translation of the wafer with respect to the optical path. Calibration of absorption into flaw concentration is discussed for the midgap EL2 donor defect, and (in an appendix) for chromium-doped GaAs. Representative wafer maps for EL2 are used as illustrations, some as mosaic grey-scale matrix plots, and others as pseudo-three-dimensional contour plots.
Despite several detailed theoretical analyses of the stress distribution expected for Czochralski grown GaAs crystals, experimental verification of these calculations has hitherto relied on dislocation density measurements. The present work shows that weak photoelastic patterns are resolvable in the near-infrared transmittance (typically near 1.4 μm) of semi-insulating GaAs wafers. Mapping of these patterns reveals the contours of constant shear stress, with results generally supporting the calculated models for the stress distribution.
Distributions of stress, dislocations, and the EL2 midgap defect have been optically mapped in semi-insulating GaAs wafers, from [100]-grown crystals created by the liquid-encapsulated Czochralski method. The evolution of EL2 along the growth axis indicates that assessment of this property through the majority of the crystal volume is often poorly represented by wafers from near the two end regions. A comparison of maps for stress, dislocation and EL2 patterns as all measured with a given wafer does not support hypotheses that EL2 is a direct consequence either of stress or of dislocations. Other mechanisms, such as segregation and melt dynamics, thus appear more likely to control the formation and distribution of EL2.
A near-infrared transmittance mapping system, previously described for quantitative macroscale (∼1 mm) mapping of the midgap EL2 defect in thin semi-insulating GaAs wafers, has been modified to provide such quantitative data on a finer scale (∼50 μm). This allows EL2 mapping on both macro- and microscales to be compared, quantitatively; and these also to be compared with dislocation micromaps and microphotographs. The well-known cellular structures for both EL2 and dislocations are resolved, and it is possible to measure the optically detectable neutral EL2 concentration inside cells (where dislocations are fewest) and in the cell walls (where they are most numerous). Comparisons of micromaps for various small parts of a wafer, and with a macromap showing the broad low-resolution variations, indicates that the local EL2 values inside cells and at cell walls vary across a wafer with the same spatial trend as given by a low-resolution macromap. This is consistent with the hypothesis that EL2 simply getters around any dislocation as a Cottrell atmosphere, as contrasted with expectations from an alternative model which postulates that EL2 (or the point defects from which EL2 is able to construct itself) will be produced by dislocation climb.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.