Duke of Edinburgh gives spintronics researchers a pile of cash

The Duke of Edinburgh is funding research into spintronics that has demonstrated the reading and writing of data from the spin of the nuclei of phosphorous atoms.

The grand old D of E, through the Royal Commission for the Exhibition of 1851 which he heads, is but one of many funders of the research which has been carried out by boffins at the University of Utah, University of Sydney, Florida State University and the London Centre for Nanotchnology. The findings were published in the 17 December issue of Science (subscription access only).

Previously, information has been stored momentarily in the spin of electrons. Professor Christoph Boehme ran a 2006 project at the University of Utah which demonstrated the reading and writing of binary data from around 10,000 electrons in phosphorous atoms which had been inserted into a silicon substrate. The initial spin of the electrons had been lined up by a high-strength magnetic field.

The latest study took the same basic approach but looked at the atomic nuclei spin, using electron spins to do so. A 1mm square chip containing phosphorous atoms in a silicon substrate was used, with the material cooled to a few degrees above absolute zero, (3.2 degrees Kelvin, 454 degrees less than zero Fahrenheit), and surrounded by an 8.6 Tesla magnetic field, some 200,000 times stronger than that of planet Earth. This lined up the atoms' electron spins, which could then be changed by writing data to them. FM radio waves were then used to transfer the electron spins to the atomic nuclei.

Radio wavelengths in the high hundreds of gigahertz area were then used to transfer the up or down nuclei spins, representing binary ones or zeros, back to the electrons and the electrons' spin value read out as variations in an electrical current, thus integrating spintronics and electronics. The kickers here were that this could be done up to 112 seconds after the nuclei spin had been set, and the nuclei spins could be read and re-read 2,000 times, meaning that the technique is reliable and durable enough, the researchers say, for use as a computer memory technology.

Electron spins are unstable because of interference from surrounding electrons, whereas an atomic nucleus is a pretty isolated entity. A 112 second refresh time is more than enough for computer main memory, where millisecond refresh rates are the rule.

The researchers dangle the carrot of having both computer memory and processing engine in the same silicon chip. There are the somewhat tricky problems of getting this to work without an enormous magnetic field and near-absolute-zero-capable refrigeration unit, but researchers hunting funding dollars allow for no such obstacles.

Indeed, they say the technique could be used both for binary computers and for the glamorous prospective quantum computer with qubits being one and zero simultaneously. However, this would require the ability to read and write the spin of a single atomic nucleus. That requires more research. ®