A radical new kind of computer memory will be a million times faster than
existing hard-drives, a leading expert in the field of nanotechnology announced
today in Sydney.
It will use nanotechnology to manipulate data like cars on tiny
Many IT researchers have predicted the end of Moore's Law - which essentially
says that computers will double in speed every two years. They've told us we'll
need light or quantum computers.
But Dr Stuart Parkin, an experimental physicist at IBM in San Jose, California, is
performing miracles with more conventional electronics. He told the ICONN
conference that the "racetrack memory" chips he and his team are developing will
be dramatically faster, more powerful and more reliable than today's hard
"We want to replace the entire disk drive with a chip that is solid state,"
Dr Parkin says. "Basically it's a disk drive on a chip. It would be entirely
reliable, a million times faster and use a lot less energy."
To make the new racetrack memory, Dr Parkin's team uses nanotechnology to
build a forest of tiny metal wires that stand up from a silicon wafer.
"You store the data in the magnetic nanowires," he says, "and you bring the data
up and down the tracks like race-cars."
The data itself is encoded using a new form of technology called
"spintronics", which uses one of the fundamental properties of electrons, known
Dr Parkin's team has already transformed computing once before with a
combination of spintronics and nanotechnology. About a decade ago they developed
a new kind of hard disk reader called a "spin-valve" or magnetic tunnelling
These readers, made up of metallic sandwiches built from layers of
single atoms, increased the storage capacity of hard drives 1000-fold.
Most digital data today, such as the information that makes up the internet,
is stored in these magnetic hard disk drives. But their rotating disks and
moving read/write heads make these drives unreliable and slow. Crashes happen
relatively often, sometimes resulting in the catastrophe of lost data.
It can also take these drives up to 10 milliseconds to read the first bit of
requested data. "In computers, 10 milliseconds is an eon," Dr Parkin says. "A
modern processor can perform 20 million operations in that time."
That's why computers also use a second type of storage, solid-state memory,
for actually doing their computational operations. Solid-state memories read and
write data with great speed, but they have their own problems, losing data when
the computer powers down or crashes.
A third kind of memory can retain data when the power is off. This is used in
smart phones and other handheld devices, but there is a trade-off between cost
and performance. The cheapest of this kind of memory is a kind called flash
memory, which is the basis of flash drives. But there are problems with this
kind of memory, too, as it is slow and unreliable in comparison with other
memory chips, and becomes unusable relatively quickly.
Racetrack memory could overcome all these problems and, in doing so,
transform the computing world, Dr Parkin says. "It will put a greater richness
of information at your fingertips." It could also make computers themselves
cheaper and more robust, he says.
Over the past three or four years, Dr
Parkin's group have shown in principle that their nanotech racetrack chips work.
He estimates that it could take another five to eight years before a product
will be ready for manufacturing.
Looking even further ahead, Dr Parkin will tell the conference about a more
futuristic idea he has for using spintronics to build what he calls a "brain in
a box" that uses spintronics to mimic the way human brain cells are
"It's possible that we could build computers that might think like the
brain," he said. "But that's a very long way off."
"Stuart Parkin and his
team's remarkable work is a great demonstration of nanotechnology in action,"
says Prof. Andrew Dzurak, ICONN co-chair and director of the Semiconductor
Nanofabrication Facility at UNSW.
For more information contact Niall Byrne
for Science in Public: 0417 131 977, email@example.com
Or Michelle Kovacevic for Science in Public: 0433 496 728, firstname.lastname@example.org