Wspólnotowy Serwis Informacyjny Badan i Rozwoju - CORDIS

Final Activity Report Summary - COMSON (Coupled multiscale simulation and optimisation in nanoelectronics)

Performing the step from microelectronics to nanoelectronics, the semiconductor industry was confronted with very high levels of integration, introducing coupling effects that were not observed before. Currently, the overall complexity of this problem is beyond the capabilities of any industrial software and design environment. Furthermore, in the near future, researchers must understand all aspects of the problems faced by the industry.

To meet these new scientific and training challenges, the COMSON research training network (RTN) on 'Coupled multiscale simulation and optimisation in nanoelectronics' merged the know-how of three major European semiconductor companies with the combined expertise of university groups specialised in developing adequate mathematical models, numerical schemes and e-learning facilities, covering all relevant fields of interest. In COMSON, academia and industry joined their efforts to realise a common demonstrator platform. On the one hand, they aimed to test mathematical methods and approaches, so as to assess whether they were capable of addressing the industry's problems. On the other hand, they aimed to adequately educate young researchers by hands-on experience on state of the art problems and beyond.

During the four and a half years duration of the network, research was carried out along the lines of five work packages, linking the research topics 'Modelling, discretization and dynamic coupling for partial differential algebraic (PDAE) systems' (WP1), 'Model order reduction' (MOR) in WP3 and 'Optimisation' (WP4) to the methodologies 'Demonstrator Platform' (WP2) and 'E-Learning' (WP5). Experienced researchers associated to the different work packages jointly worked on developing the demonstrator platform and the link to e-learning. They were actively involved in supervising the early stage researchers in the three research topics and mapping the outcome of this research to the demonstrator platform.

In WP1, PDAE models for describing the coupling of circuits, devices, electromagnetic (EM) and heat effects were developed, analysed in terms of well-posedness and index concepts, and numerically simulated by dynamic iteration schemes. In addition, more sophisticated semiconductor models were developed to cope with heating effects of the crystal, for example. Being only interested in an adequate input-output behaviour, reduced models developed in WP3 were an efficient alternative to the PDAE approach of WP1. In this case COMSON made valuable contributions to four areas, namely parameterised MOR, nonlinear MOR for DAEs, sensitivity analysis exploiting MOR and structure preserving MOR. In WP4, the problem of optimal semiconductor doping profiling was tackled into the framework of optimal control by using the space mapping and the manifold mapping optimisation method. Optimisation based on a genetics algorithm framework fully integrated within the design flow was developed to analyse the hot spot occurrence in silicon power metal oxide semiconductor field effect transistors (MOSFETs) operated at high switching frequency.

The network fellows were heavily involved in the network research, especially in contributing to the final outcome of the project, the COMSON handbook. Thus, along with the research carried out in the network, a new generation of European researchers was trained. These researchers were specialized in different aspects of coupled multiscale simulation and optimisation in nanoelectronics. At academic and industry nodes, during internships, summer and autumn schools and various specialized transfer of knowledge (ToK) events, they were given a broad training in this area and so that they were ready to play an important role in modelling, numerical simulation and optimisation of coupled problems in nanoelectronics at both academia and industry by the time of the project completion.

Reported by

Śledź nas na: RSS Facebook Twitter YouTube Zarządzany przez Urząd Publikacji UE W górę