No abstract

No abstract

Four interface models for crystalline oxynitride on (001)Si substrates are proposed and investigated. All four models are proposed to model thin oxynitride films on Si substrates, according to experimental findings and theoretical studies on amorphous oxynitride films and nitrogenated SiO 2 . State-free insulating interfaces were obtained by expanding the bulk oxynitride cell by approximately 12% and 1% along the [100] and [010] axes, respectively, and interfacing it with (001)Si. Results demonstrate state-free insulating interfaces for all models with, however, valence-band offsets slightly above or below 1 eV. The significant decrease in the valence-band offsets is mainly attributed to the significant expansion of the oxynitrideÕs lattice constant to lattice-matched (001)Si, as well as to the high concentration of nitrogen atoms in the oxynitride bulk.

This paper analyses the sampling uncertainty in regenerative comparators due to substrate noise coupling and provides a model for the resulting sampling distortion power. The analysis identifies two contributors of the total sampling uncertainty: the input signal-dependent one and the substrate noise-related one. The two disturbances of the ideal operation of the sampling transistors cause a non-uniform sampling operation, whose properties depend on the frequencies of the disturbing signals. The non-uniform sampling introduces distortion components mainly at the harmonics of each interference signal, and other components located at frequencies related to the spectral content of the interference signals and the sampling frequency. The experimental results indicate that the developed model manages to capture accurately all the aforementioned distortion components in the presence of any input and any substrate noise signal, and therefore to predict the overall sampling distortion power.

This paper presents a model to evaluate the impact of substrate noise on a CMOS regenerative comparator and on a flash A/D converter. The proposed approach initially relates substrate noise with the induced timing uncertainty of the comparator. Subsequently, the obtained expression for the timing uncertainty is used to derive a generalized expression for the SNR reduction in flash A/D converters, relating the resulting SNR directly to substrate noise. The developed approach is validated on a 10-bit flash A/D and is utilized to investigate performance degradation of practical converters.