I saw some very good talks this week. First up was a physics colloquium by Stuart Parkin from IBM Almaden. In some very real sense, you're reading this because of Parkin - he and his team were the people who first took giant magnetoresistance (GMR) and developed it into a useful technology in the read heads of hard disk drives. The remarkable explosion in data storage capacity over the last decade and a half is largely due to this advance, possibly the best example of true nanotechnology (the film thicknesses involved in spin valves are a few nm) making it out of the lab and into manufacturing and consumer products. Their later work on tunneling magnetoresistance has also now been transferred into hard drive read heads. In fact, TMR heads with MgO tunnel barriers between ferromagnetic layers can have room temperature resistance changes of several hundred percent in the presence of few-Oersted fields like those from drive media. After reviewing all of this at just the right level, Parkin went on to talk a bit about his latest ideas and work on high performance "racetrack" memory. In this idea, a single transistor cell can be responsible for reading and writing tens of bits of memory (as opposed to one in current RAM designs). The bits are stored as domain walls in a ferromagnetic nanowire. The walls can be detected through their local change in the magnetization, and they can be moved by pulsing spin-polarized currents through the ferromagnetic wires. All in all, a great colloquium - one of my colleagues wished that we'd taped it so that we could show it to job candidates as an example of a real general audience colloquium.
There was also a workshop on campus this week about probabilistic and nanoscale computing that featured some nice talks. One of the best was by Tom Theis, head of physical sciences research at IBM, who reviewed their latest developments and the future of the field-effect transistor from his perspective. Anyone who has alternative ideas in mind about computing technologies really needs to do their homework by listening to someone like Theis, who has perspective about the science as well as the economic and manufacturing issues.