Due to the aggressive miniaturization of memory and logic devices, the current technologies based on silicon have nearly reached their ultimate size limit. One method to maintain the trend in device scaling observed by Moore's law is to create a heterostructure from existing materials and utilize the underlying electronic and optical properties. Another radical approach is the conceptualization of a new device design paradigm. The central objective of this thesis is to use both of these approaches to address issues associated with the aggressive scaling of memory and logic devices such as leakage current, leakage power, and minimizing gate oxide thickness and threshold voltage. In the first part of the dissertation, an atomistic, empirical tight binding method was used to perform a systematic investigation of the effect of physical (shape and size), and material dependent (heterogenity and strain) properties on the device related electronic and optical properties of the Germanium (Ge)/Silicon (Si) nanocrystal (NC) or quantum dot (QD). The device parameters pertaining to Ge-core/Si-shell NC-based floating gate memory and optical devices such as confinement energy, retention lifetimes and optical intensities are captured and analyzed. For both the memory and optical device applications, regardless of the shape and size, the Ge-core is found to play an important role in modifying the confinement energy and carrier dynamics. However, the variation in the thickness of outer Si-shell layer had no or minimal effect on the overall device parameters.

In the second part of the dissertation, we present a systematic study of the effect of atomistic heterogeneity on the vibrational properties of quasi-2D systems and recently discovered 2D materials such as graphene, while investigating their applicabilities in future devices applications. At first, we investigate the vibrational properties of an experimentally observed misoriented bilayer graphene (MBG) system, a heterostructure where two graphene layers are rotated by a relative angle, using molecular dynamic (MD) method. The MD method includes temperature dependent phonon anharmonicity which correctly predicts misorientation angle (θ) dependent low-energy breathing modes, and establishes a correlation between the experimentally observed low frequency Raman modes. Using a similar method, we have also conceptualized a phononic circuit using quasi-2D materials constructed from group IV elements of the periodic table, mainly carbon (C), Germenium (Ge) and Silicon (Si) by modifying the phononic bandgap (PBG). We successfully demonstrated the realization of various phononic interconnects such as nano-scaled phononic resonators, waveguides and switches by simultaneously introducing defects of different types at various locations on the 3C-SiC and 3C-GeSi surfaces.

Finally, we have conceptualized a novel low power device called TMDC Excitonic Field Effect Transistor (TExFET), using other 2D materials namely, hexagonal boron nitride (h-BN) and Transition Metal Dichalcogenides (TMDC) by creating a TMDC/h-BN/TMDC heterstructure system. The characteristics of the TExFET device is explored with a combination of the variational principle and the mean field approximation. Our variational principle based calculation of the unscreened interlayer Coulombic forces in the TMDC/h-BN/TMDC system gives us an upper bound exciton gap in the order of 100 meV, mainly due to the isotropic electron and hole effective masses of the TMDC layers. Due to an effective exciton radius in the range of 2 nm, the TExFET could also be a device of choice for maintaining the device scaling trend. Further, when the effect of static screening between the layers is considered during self-consistent calculations, the interaction strength is reduced by ~ 40% to 60 meV, producing an excitonic gap suitable for low temperature, low power device applications.