Skip to main content

Chris Lutz discusses where moving atoms could take us


“A Boy and His Atom,” the animated movie that IBM researchers made by moving individual atoms, is certainly popular–it's generated more than 4.5 million hits on YouTube so far. But it's more of a coming attraction than a feature film.

“The movie was a demonstration of capabilities; it wasn't in itself a breakthrough,” says Chris Lutz, an IBM staff scientist who worked on the project. “It was a way to reach a wider audience and to show our ability to manipulate atoms and build with them.”

Taking on Moore's Law

Why should we care about moving atoms around? Lutz cites Moore's Law, the observation that computational power and storage double roughly every two years. “For years now, we've been putting more power and more storage on the same size device. If we continue this trend, we'll ultimately come to the level of atoms to accomplish it.” He concedes that at this rate atomic computing may be 20 to 40 years away, but “we're jumping ahead. We're exploring what we will want to build then and what properties we will want.”

In fact, IBM researchers, including Lutz, have already demonstrated the potential of atomic memory. Early in 2012 they stored the information needed to spell out “THINK” one letter at a time on eight bits of 12 atoms each. This demonstration of atomic-scale magnetic memory showed just how compact such an approach could be. Storing a bit–the basic unit of information in computers and digital communications–currently requires 1 million atoms. If commercialized, this atomic memory could one day store all of the movies ever made in a device the size of a dime.

“For years now, we've been putting more power and more storage on the same size device. If we continue this trend, we'll ultimately come to the level of atoms to accomplish it.”  -Chris Lutz
Decrease in silicon transistor size over the years. Current Silicon transistor storage technology size has continued to shrink over the years and will approach the atomic level.

Magnifying 100 million times

Both the tiny movie and the 12-atom magnetic memory approach are possible thanks to a device called a scanning tunneling microscope. Invented in the early 1980s by two IBM scientists (who won the 1986 Nobel Prize in physics for their work), one such microscope sits in a special room at IBM's Almaden Research Center. The instrument looks more like a prop from the “Alien” films than a microscope. “It weighs two tons, operates at a temperature of negative 268 degrees Celsius {-450 degrees Fahrenheit} and magnifies the atomic surface 100 million times,” says Lutz. “To give you an idea of that, if an atom were magnified to be the size of an orange, then an orange would be the size of Earth.”

The microscope has a needle-like electrode that attracts an atom when brought close enough to it. That it gives scientists the means not only to see atoms and their arrangement, but to move them around. That's no easy task when you consider how much effect temperature, vibrations and air pressure can have at such a small scale.

“The ability to reduce the temperature, pressure and vibrations to miniscule levels makes our lab one of the few places in the world where atoms can be moved with such precision,” Lutz notes.

IBM´s approach. IBM researchers start at the atomic level to see how many atoms are needed to store 1-bit of data. What is a bit? A bit is the most basic piece of information that a computer understands. A bit has one or two values, on or off, like a light switch: '1' or '0'.

12 atoms needed to store one bit of data. IBM researchers have succesfully used 12 atoms to store one bit of data by aligning their magnetic properties so that the group of atoms would not interfere with their neighboring group of atoms.

Advancing a step at a time

Magnetic memory also holds promise for smaller power consumption. “The magnets hold information without the use of power, but you'd need power to read and write the information,” says Lutz.

And don't look for the world's film library on a smart phone any time soon. “You have to do things a step at a time,” he says. “Sure, we can store information in 12 magnetic atoms, but before it can be functional we need a way to read and write the information. And we need a way to manufacture them cheaply and reliably. At this point, we don't yet know what sort of device we'd like to manufacture.

“No memory today or in the near future works at the atomic level. The bits in existing disks and tapes, and in MRAM (magnetic random access memory) and racetrack memory under development, all use little magnets to store information. All benefit by shrinking their size to fit more bits in the same space. But they still all take about a million atoms to hold one bit.”

Forgetful magnets

What's more, all of these technologies face what's called the paramagnetic limit. If the bits are too small, the magnets start to flip their orientation spontaneously. As a result, “they forget the bit they are supposed to remember,” says Lutz. “We're working on how small one could make a magnetic bit that holds its orientation stably, at the extreme limit where we place every atom exactly where we want it, and every atom counts. Conceivably, what we learn could be applied to any of these memory and storage technologies and to technologies we haven't even thought of that would use the magnetization direction to store information.

“We are taking some of the first steps into the land of the atomic-scale magnets, studying how to make a stable magnet in the first place. We have gone on to demonstrate that the state of these bits can be written and read back using a small electric current from a nearby electrode, the microscope's tip, so electrical reading and writing might be used in future atomic-scale memories.”

100x Atomic-scale magnetic memory is potentially 100x denser than today´s hard disk drive technoloy.

Learn more

Read more about atom-scale magnetic memory at this web page and in this press release.


Experts on this topic

100x Atomic-scale magnetic memory is potentially 100x denser than today´s hard disk drive technoloy.

Chris Lutz
IBM staff scientist