Non-Invasive Optical Characterization of Defects in Gallium Arsenide.
Abstract
This work is concerned with the development of a non-invasive comprehensive defect analysis system based on computer-assisted near infrared (NIR) microscopy. Focus was placed on the development of software for quantitative image analysis, contrast enhancement, automated defects density counting, and two-dimensional defect density mapping. Bright field, dark field, phase contrast, and polarized light imaging modes were explored for the analysis of striations, precipitates, decorated and undecorated dislocations, surface and subsurface damage, and local residual strain in GaAs wafers. The origin of the contrast associated with defect image formation in NIR microscopy was analyzed. The local change in the index of refraction about a defect was modelled as a mini-lens. This model can explain reversal of image contrast for dislocations in heavily doped n-type GaAs during defocusing. Defect structures in GaAs crystals grown by the conventional liquid encapsulated Czochralski (LEC) method are found to differ significantly from those grown by the horizontal Bridgman (HB) or vertical gradient freeze (VGF) method. Dislocation densities in HB and VGF GaAs are one to two orders of magnitude lower compared to those in conventional LEC GaAs. The dislocations in HB and VGF GaAs remain predominantly on the {111}/<1 |10> primary slip system and tend to form small-angle subboundaries. Much more complicated dislocation structures are found in conventional LEC GaAs. Dislocation loops, dipoles, and helices were observed, indicating strong interaction between dislocations and point defects in these materials. Precipitates were observed in bulk GaAs grown by the LEC, HB, and VGF methods. Precipitation was found to occur predominantly along dislocation lines, however, discrete particles were also observed in dislocation-free regions of the GaAs matrix. The size of discrete precipitates is much smaller than that of the precipitates along dislocations. Quenching after high temperature annealing at 1150^ circC was found effective in dissolving the precipitates but glide dislocations are generated during the quenching process. STEM/EDX analysis showed that the precipitates are essentially pure arsenic in both undoped and doped GaAs. NIR phase contrast transmission microscopy was found to be very sensitive in detecting surface and subsurface damage on commercial GaAs wafers. Wafers from a number of GaAs manufacturers were examined. It was shown that some GaAs wafers exhibit perfect surface quality, but in many instances they exhibit, to various extents, subsurface damage. Computer-assisted NIR transmission microscopy in a variety of modes is found to be a rapid and non-invasive technique suitable for wafer characterization in a fabline environment. (Copies available exclusively from MIT Libraries, Rm. 14-0551, Cambridge, MA 02139-4307. Ph. 617-253-5668; Fax 617-253-1690.) (Abstract shortened by UMI.).
- Publication:
-
Ph.D. Thesis
- Pub Date:
- 1993
- Bibcode:
- 1993PhDT.......281C
- Keywords:
-
- Engineering: Materials Science; Physics: Condensed Matter; Engineering: Electronics and Electrical