Nano-Science Center > Seminarer og events > First Principles Model...

First Principles Modelling of the Electrical Properties of Nanoscale Devices

Inaugural lecture, Honorary Professor to the NBI Kurt Stokbro,  Department of Computer Science, DIKU, University of Copenhagen.

Kurt Stokbro The minimum feature size of electronic devices is approaching the atomic scale and at this scale the quantum nature of electrons start to alter the device behaviour in new ways. The new effects requires modifications of the device geometries and introduction of new device materials and I will in my introduction give a brief overview of some of the most recent developments in the semiconductor industry and what can be anticipated in the future.

For successful introduction of the new technologies it is important with reliable modelling tools that allow for testing and optimizing the new materials and device geometries. However, many of the atomic-scale effects are not included in the current modelling tools used in the semi-conductor industry and as member of the International Technology Roadmap for the Semiconductors (ITRS - http://www.itrs.net/). I have been involved in putting forward a list of modelling challenges that must be addressed in order to model semiconductor devices below the 32 nm node.

In this talk I will discuss my research in the field of quantum transport based on Density Functional Theory within the Non Equilibrium Greens Function formalism, NEGF-DFT and how this methodology can be used to address the challenges in the ITRS roadmap. The methodology has so far been applied to new emerging device geometries based on carbon nanotubes, 1-D graphene sheets, molecules, atomic wires, or magnetic materials and I will present results from these studies.

These emerging devices are still rather exotic and are expected to enter the market more that 10 years out in the future. In order to make the tools able to model more traditional semiconductor device geometries and thereby address more near time challenges in the semiconductor industry, it is necessary to develop new more efficient and highly parallel algorithms that can treat systems with many thousand atoms. In the Nanopar (www.nanopar.dk) project at Copenhagen University we are looking into new algorithms that give O(N) scaling of computational time with system size and which speed-up state of the art algorithms by orders of magnitude. I will present the most recent results in this project and discuss our future path towards enabling large scale device simulations comprising millions of atoms.