Quantum and atomistic effects in nanoelectronic transport devices

As devices scale towards atomistic sizes, researches in silicon electronic device technology are investigating alternative structures and materials. As predicted by the International Roadmap for Semiconductors, (ITRS), structures will evolve from planar devices into devices that include 3D features, strong channel confinement, strain engineering, and gate all around placement for better electrostatic control on the channel. Alternative channel materials such as carbon nanotubes (CNT), nanowires (NW) and III-V based channel materials are considered to be possible candidates for future device technology nodes because of their potentially superior to silicon transport properties. For nanoscale dimensions, and under the operating conditions mentioned above, both atomistic and quantum effects become important in determining the electronic structure and transport properties of the devices. Detailed modeling and simulation that capture these new physics will be essential in providing understanding and guidance to the device operation and optimization. We have used the non-equilibrium Green's function (NEGF) formalism for quantum transport simulations and real space atomistic tight-binding techniques (p z, sp3d5s*-SO) to investigate transport properties in CNT, NW and III-V HEMT field-effect transistors. Specifically, we have investigated the effect of atomistic defects such as atomic vacancies, and charged impurities in 1D CNT, and dangling bonds in NW channels. It was found that the presence of single defects, severely degrades the transport performance of 1D channels. We have further investigated the effect of physical quantization on the electronic structure of NW field-effect transistors and identified the main electronic structure factors that influence their performance. It was found that structural and quantization below 10nm can severely affect the electronic properties of NW channels by changing the effective masses and altering degeneracies through valley splitting. Different wire orientations have different transport properties. The [110] and secondly [100] oriented nanowires are found to perform better than the [111] wires in terms of ON-current capabilities for n-type wires, whereas the [111] and [110] significantly outperform the [100] wires in the case of p-type nanowires. Explanations for this behavior can be extracted from the non-parabolicity and anisotropy of the Si 3D bulk E(k). Finally, we present an analysis of recent experimental data for III-V HEMT devices using the NEGF formalism and address several issues related to the operation of HEMT devices. Interestingly, a 60nm HEMT device can be though to first order as a ballistic channel connected to two series resistances.