Please activate JavaScript!
Please install Adobe Flash Player, click here for download
ePaper created 2013-02-01, 09:35:07 | version 1.24.0
Research in Jülich 3|201220 Focus on Watson’s Successor The European Research Council has faith in Prof. Paul Kögerler. The Jülich chemist has been awarded € 1.5 million – Europe’s most generous grant for early-career scientists. His research aims to pave the way for the next generation of computers. W atson is the name of the IBM supercomputer that outwitted two winners on the US quiz show Jeopardy! in 2011. It proved that ar- tificial intelligence is capable of beating human intelligence, even when it comes to complex linguistic and knowl- edge-based tasks. However, what the public did not see were the 80 quintillion processes per second that Watson need- ed to do so and its excessive power con- sumption. This is where Prof. Paul Kögerler from Jülich’s Peter Grünberg Institute – Elec- tronic Properties (PGI-6) comes into play. “Microelectronics must break away from conventional computer designs and let itself be inspired by the non-binary workings of the brain. Only then will it become dramatically more energy-effi- cient and be in a position to process even more complex tasks than Watson’s artificial intelligence,” says Kögerler de- scribing the envisaged revolution. While the Watson experiment proves that artificial intelligence can logically combine facts from a huge database, the human brain is much more economical: it consumes only about as much energy as a 50-watt incandescent bulb. When it comes to associative processes such as understanding language, our brain is still head and shoulders above any comput- er. Kögerler’s research therefore aims at more powerful and energy-efficient tran- sistor technology. He believes magnetic molecules that can be activated with minimal voltages will be the basis of this technology. The European Research Council (ERC) has now confirmed that his research approach is extremely promising by awarding him an ERC Start- ing Grant. With the funds from the European Union, the race for transistors of the next generation is ready to go into the next round. Throughout the world, scientists are tinkering with the basic idea that Kö- gerler and his team are also working on, namely replacing the silicon in transis- tors with magnetic molecules. “What is interesting about these molecules are their nonlinear effects,” says Kögerler. “Nonlinear means that these molecules exhibit sharp, or clearly measurable, changes in their behaviour even if the voltage applied only changes a tiny bit.” As a consequence, a transistor based on a magnetic molecule could be con- trolled precisely with a minimal voltage – and thus extremely low energy require- ments. CONTACTING AS THE CRUX The hitch: nobody has yet managed to reliably use the switching properties that result from the magnetic states of individual molecules and can be very complex. This is because any external influence – including contact with an electrode – changes both the magnetic In his laboratory, Kögerler studies magnetic molecules that can be activated precisely with very small voltages. This will be the basis for novel transistors for energy-saving computers.