Magnets could make chips faster, smaller

Magnets made from "loadstone" are being used to change the way electric currents flow across tiny circuits one million times …

Magnets made from "loadstone" are being used to change the way electric currents flow across tiny circuits one million times smaller than a pinhead. These nano-scale devices can be made to store and recall digital data up to 10 times faster than an average PC and point the way towards faster, smaller storage devices.

Prof Michael Coey and PhD student Janko Versluijs of Trinity College's Physics Department are pioneers of the new science of magnetic spin electronics that lies behind these miniature devices.

At present, components are still relatively crude, Prof Coey said. "In the commercial development of spin electronics we've only got as far as the equivalent of a resistor, the simplest electronic device," he explained.

"In these, a magnetic field is used to change their electrical resistance. This is called magneto-resistance." Even so, this achievement has already led to the "spin valve", a vital component in computer hard-disc drives that is manufactured and sold in huge quantities.

READ MORE

Conventional electronic circuits work by controlling the flow of electrons using their negative charge, but electrons have another important characteristic. As well as charge, electrons have "spin". They can rotate either clockwise or anti-clockwise. A normal electric circuit doesn't make any use of this rotating behaviour, but if instead a circuit is made from magnetic materials then spin becomes important.

The spin-doctors at TCD have found a way of boosting the effects caused by rotating electrons to give them a "giant magneto-resistive effect", Prof Coey explained. "Spin valves give a resistance change about 10 per cent. Our experiments give a change of 500 per cent for the same level of magnetic field."

The secret to the team's success is to use microscopic slivers of tapered magnetic material which barely touch. The contact point between them is only about one nanometre (a thousandth of a millionth of a metre) across and makes a tight bottleneck for electrons. Each sliver is a magnet in its own right, with the magnetic poles of one pointing in the opposite direction to the other.

As a pulse of current forces electrons across this nano-scale bottleneck, the electrons sense the change in magnetic direction. This forces them to flip their spin. They are reluctant to do this and so the result is electrical resistance. The tighter the bottleneck, the higher the resistance.

When three pieces of magnetic material are brought together in a line, giving two contact bottlenecks, one bigger than the other, then a basic digital computer memory element is possible. Prof Coey refers to this circuit as a "peanut".

As a current is pulsed through the peanut, the region of spin-flip can be made to jump from the smaller contact zone to the larger zone. This switches the resistance of the device between the two conditions of high and low resistance - the essence of a digital device.

"The trick is to get the switch to happen between the two states as smoothly and as coherently as possible," Prof Coey said. "This will happen at least 10 times faster than other devices."

The memories need only low powers to work because the magneto-resistance effect is so large in these experimental devices. This will make any future devices highly efficient.

"Computer memories made this way would be non-volatile, so there would be no need to re-boot when you switch on," said Prof Coey. This is because the memories work using magnetisation: they keep their stored data even when the power is removed.

Prof Coey and his team are experimenting with trial circuits made of thin layers of magnetic materials on a flat wafer. Although commercial memories made using this method are some way off, their efforts will be boosted by a £5 million research grant funded by Science Foundation Ireland.

Dr Peter Foote is a research scientist with BAE Systems in Bristol. He is participating in the Media Fellow programme of the British Association for the Advancement of Science