By the addition of selected impurities to silicon it is possible to affect its physical, optical and electrical properties to a remarkable extent. Of great technological importance are the elements of groups III and V which introduce shallow acceptor and donor levels and control the equilibrium charge density. These levels have been studied extensively and are well understood. However, other impurities and structural defects are known to introduce deep levels. These levels may act as acceptors or donors in the traditional sense but they can also control the recombination of non-equilibrium charge carriers. The properties of these recombination levels reflect the processes of energy exchange and provide information on the defect structure and energy spectrum and are thus of physical interest. Also since they control the excess carrier lifetime which is a critical parameter in many semiconductor devices they are of interest to the technologist. The major difficulty in analyzing these levels arises from the catalytic nature of the recombination centers which allows an extremely low density of centers to significantly affect the recombination lifetime. Most bulk single crystal silicon techniques applied to date use activation processes with high resistivity, high recombination center density samples. This severely limits the range and sensitivity of analysis. In this dissertation is detailed a new approach to more accurately determine lifetime parameters in conjunction with a phenomenological model which describes the recombination using a set of characterizing parameters. The technique is used to characterize the levels responsible for recombination in single crystal silicon after controlled heat treatments and, also, in gamma irradiated and impurity containing silicon. / Ph. D.

Bidirectional reflectance is a fundamental radiative property of rough surfaces. Knowledge of the bidirectional reflectance is crucial to the emissivity modeling and heat transfer analysis. This thesis concentrates on the modeling and measurements of the bidirectional reflectance for microrough silicon surfaces and on the validity of a hybrid method in the modeling of the bidirectional reflectance for thin-film coated rough surfaces. The surface topography and the bidirectional reflectance distribution function (BRDF) of the rough side of several silicon wafers have been extensively characterized using an atomic force microscope and a laser scatterometer, respectively. The slope distribution calculated from the surface topographic data deviates from the Gaussian distribution. Both nearly isotropic and strongly anisotropic features are observed in the two-dimensional (2-D) slope distributions and in the measured BRDF for more than one sample. The 2-D slope distribution is used in a geometric-optics based model to predict the BRDF, which agrees reasonably well with the measured values. The side peaks in the slope distribution and the subsidiary peaks in the BRDF for two anisotropic samples are attributed to the formation of {311} planes during chemical etching. The correlation between the 2-D slope distribution and the BRDF has been developed. A boundary integral method is applied to simulate the bidirectional reflectance of thin-film coatings on rough substrates. The roughness of the substrate is one dimensional for simplification. The result is compared to that from a hybrid method which uses the geometric optics approximation to model the roughness effect and the thin-film optics to consider the interference due to the coating. The effects of the film thickness and the substrate roughness on the validity of the hybrid method have been investigated. The validity regime of the hybrid method is established for silicon dioxide films on silicon substrates in the visible wavelength range. The proposed method to characterize the microfacet orientation and to predict the BRDF may be applied to other anisotropic or non-Gaussian rough surfaces. The measured BRDF may be used to model the apparent emissivity of silicon wafers to improve the temperature measurement accuracy in semiconductor manufacturing processes. The developed validity regime for the hybrid method can be beneficial to future research related to the modeling for thin-film coated rough surfaces.