Spin doctors

A major European initiative is underway to develop a new breed of faster, low-power computing devices based on the physical phenomenon of spintronics.

Computing experts believe that integrated circuit technology will soon run into constraints preventing further miniaturisation. Various candidate technologies are being considered to take over the role of complementary metal oxide semiconductors (CMOS) in integrated circuits. Spintronics is one of them.

Spintronics exploits the spin quantum property of electrons, detectable as a weak magnetic force, a property already used on some hard disks. It also features in emergent magnetic random access memory (MRAM), non-volatile electronic memory that makes use of spin-valves and related devices. Future spintronic devices should be much quicker than conventional electronics and require much less power.

Nanospin is a European Commission project bringing together eight academic and industrial collaborators to develop new types of spintronic nanoscale devices using ferromagnetic semiconductors. The University of Würzburg will co-ordinate the project.

'In spintronics we try to combine the properties of electronic devices with those of magnetic ones,' said Laurens Molenkamp, professor of physics at Würzburg. 'This means you can have memory systems which you can control electronically, but which have a permanent memory because they are intrinsically magnetic.'

According to Dr Charles Gould, a post-doctorate researcher on the project: 'Programmable logic applications could enable you to have your computer applications instantly available. You wouldn't have to wait for it to boot up, you just store where you were in memory at night, and the next morning you start immediately where you left off.'

The project will use gallium manganese arsenide, a ferromagnetic semiconductor that is well understood but only operates at extremely low temperatures, to prove the technology. The team hopes that the resulting technology will in the longer term work with room-temperature semiconductors.

Another collaborator in the project is Nottingham University, whose role is to supply and investigate suitable materials. Bryan Gallagher, professor of physics and a consultant for project industrial partner Hitachi Cambridge Laboratory, said: 'We grow perfect crystals monolayer-by-monolayer using molecular beam epitaxy. We deposit a set number of monolayers of one semiconductor, then more monolayers of another on top to make high-quality materials.'

As well as growing materials to make the nanoscale devices, Nottingham will also carry out electron beam lithography to fabricate the devices being studied.

'We're taking a dual-track approach,' said Gallagher. 'We're trying to develop the new materials which will be commercially useful, and are pressing ahead with what we've already got to achieve new types of devices and functions. So the devices we'll be working will not be of immediate commercial interest, but we'll be proving the technology.'

The researchers believe the technology could take 10 years to come to market if a suitable room temperature semiconductor can be found, but they emphasise that spintronics is not the only technique being investigated.

'Other people are taking different approaches,' said Gallagher. 'Beyond the 10-year horizon, most believe there will have to be a real shift in the basis of the technology and spintronics is just one possibility. There are many contending future technologies, including carbon nanotubes and organic electronics. We're just accepting that we have to work on a range of technologies to get faster processors and bigger memory storage.'

'What we're doing at the moment is at the forefront of solid-state physics,' said Molenkamp, 'and it's encouraging to see that there is so much interest from industry. It realises that the current technology is not going to live forever, so they are looking at new concepts. It gives us a lot of feedback on the type of things we should incorporate in devices, and the kind of direction we should be looking at.'