Abstract
Silicon has become the material of choice for fabrication of high circuit density, low defect density and high speed integration devices. CMOS technology has been favoured as an attractive candidate to take advantage of the performance enhancements available through miniturisation. However, hot carrier effects in general, and hot electron currents in particular, are posing as the main obstacle to a new era of sub-micron architecture in semiconductor device technology. Electron transport in modern sub-micron device is often governed by mechanisms that were not relevant to long-channel devices. Many of the classical device models are based upon such convenient assumptions as "thermal equilibrium" and "uniform local electric field". With the downscaling of devices, hot electron currents are becoming increasingly inherent. These currents arise from the fact that electrical fields in small geometry devices can reach very high values and can vary rapidly in space. The large electric field can Impart significant kinetic energies to the carriers. In thermal equilibrium, all elementary excitations in a semiconductor (eg. Electrons, holes, phonons) can be characterised by a temperature that is the same as the lattice temperature. Under the influence of large electric fields, however, the distribution function of these elementally excitations deviate from those in thermal equilibrium. The term "Hot Carriers" is often used to describe these non-equilibrium situations. In this thesis hot electron currents, in particular their physical origins and dependence upon various operational and geometrical parameters, have been discussed and then quantified in a number of models based on the "Lucky Drift" theory of transport. Temperature is then used as a tool to differentiate between the underlying physical processes, and to determine if reliability problems related to hot electron effects would improve under cryogenic operation. It has been the prime objective of this work from the outset to concentrate on the study of N-channel devices. This is primarily due to the fact that N-channel MOSFET's are more prone to hot electron effects, and therefore, studies in the nature of this enhanced susceptibility could prove to be more fruitful.