Professor S Hall, Dr I Mitrovic and Professor S Taylor
The transistors in successive generations of integrated circuits have increased in speed, use less power and cost less because of diminishing transistor sizes. Unfortunately, it is not possible to reduce the operating voltage with the same scaling factors as the physical size so the electric fields in the transistor have increased dramatically. The silicon dioxide which is used as the gate of these transistors is now so thin that a significant current flows through this insulating layer at normal operating voltages because of tunnelling. This increases power consumption and reduces reliability. Future transistor generations will need a gate dielectric of higher permittivity (high-k) that will allow a larger physical thickness to be used without increasing the equivalent electrical thickness. This is the most important single issue facing the development of integrated circuits at the moment. The high-k materials proposed by the industry (materials based on hafnium oxide are favoured at the moment) are physically and chemically very different to silicon dioxide. Considering the importance of this technological leap, surprisingly little is known about their interface characteristics with silcon or about the trapping sites in the oxide. Work now encompasses high-k oxides on Ge for application at the 8nm ITRS technology node.
The research concerns screening of novel dielectrics prepared by ALD at Liverpool, with other samples provided by our collaborators. We investigate the samples with a wide range of physical and electrical characterisation techniques; namely, XTEM, XRD, MEIS, Raman, XPS, Spectro-ellipsometry (in collaboration with Dr Vin Dhanak, Dept. Of Physics), C-V and I-V(T). We also address reliability.
We are also engaged in research looking at 'end of roadmap' solutions in collaboration with the high-k gang (http://www.high-k-gang.eu) comprising Chalmers (Sweden), AMO/Aachen University (Germany) and Tyndall Institute (Ireland) together with IMEC. The work has been funded by EPSRC, and EC through projects NANOSIL/SINANO and PULLNANO.