Spintronics exploits the spin of an electron as well as its charge. The spin of an electron can point in an "up" or "down" direction, and this property could be used to store and process information in spintronics circuits. Such circuits would be smaller and more efficient than conventional electronic circuits – which rely on switching charge alone – because switching spins from up to down can be done using very little energy.

Graphene is a sheet of carbon just one atom thick and, in principle, should be a very good conductor of spin and so ideal for making spintronics devices. This is because, in theory, electron spins in the material should maintain their direction for a long time thanks to the fact that carbon has a low atomic number (and therefore low spin-orbit coupling) and the main isotope (carbon-12) lacks nuclear spin. But the reality is quite different: graphene has revealed itself to be a rather bad conductor of spin.

Indeed, researchers were surprised to find that electron spins in graphene only last around 100 picoseconds (rather than the micro- or even milliseconds predicted by theory). These values are similar to those expected in conventional semiconductors and metals, and are nothing to write home about. “Fast spin relaxation, or spin decoherence, makes spintronics very difficult,” explained team leader Joshua Folk. “The orders-of-magnitude discrepancy between spin relaxation measurements and those predicted by theory is one of the most important puzzles in graphene research today.”

To better understand this disagreement, the researchers performed quantum interference measurements on graphene. Quantum interference is a good way to study how conduction-electron charges and spins interact in materials.

More magnetic defects than expected

“Spin interactions affect quantum interference of charge carriers (electrons and holes) and we have measured ultralow temperature transport in varying magnetic fields to distinguish between magnetic and non-magnetic decoherence effects in graphene,” said Folk. “Our technique allowed us to discover that magnetic defects are the main cause of spin decoherence in the carbon material.”

But that is not all. The researchers also found that there appears to be more magnetic defects than expected in a sheet of unadulterated graphene and they believe that this is one of the main reasons why theory and experiments do not match. Finally, spin-orbit interactions may be unexpectedly strong at the magnetic defects too, something that will lead to faster spin coherence as well.

“These results are important because they put us on the road to improving the usefulness of graphene as a spintronics material,” Folk told nanotechweb.org.

So, could the findings be used to improve the spin transport of graphene and make the material a better conductor of spin? “My guess – and this is just a guess – is that chemical passivation of the magnetic defects might prolong spin lifetimes,” he added.

The team says that it is now busy looking at how defects and spin-orbit coupling vary in samples of graphene prepared by different methods – such as graphene made by chemical vapour deposition or grown on silicon carbide wafers (in contrast to exfoliation – the famous “sticky tape” method by which graphene was first isolated).

The current work is detailed in Physical Review Letters.