Berkeley (CA) - A team of scientists at UC-Berkeley have successfully stored and retrieved information using the nucleus of an atom. A paper outlining the achievement entitled "Solid-state quantum memory using the 31P nuclear spin" appears in yesterday's Nature journal. This is the first ever demonstrable proof that the nucleus of an atom can be used to store digital data reliably, bringing the dream of quantum computers one step closer to reality.





Phosphorous



Phosphorus-31 was chosen as a doping agent to an exceptionally pure and isotopically controlled silicon-28 crystal. Normal silicon is comprised of several isotopes. For these tests, however, some of the most pure silicon ever created was required. This was the first big hurdle the team encountered. The second was specifically and deliberately placing phosphorus atoms at known locations on the surface. One researcher compared this to inserting a single person into the world's population, and doing so at exactly the correct address - i.e., very hard to do.

The quantum information was first encoded into the outer electron shell. Using microwave and radio transmissions, it was then transferred from the electrons to the nucleus, resulting in nuclear spin. After that, it was once again transferred back out of the nucleus into the outer shell where it could be read as data.



The electrons themselves never entered the nucleus, but served basically as a type of read-write head as the digital data stored in the electron spin was injected into the nuclear spin. Oxford University's John Morton, lead researcher on the project, said, "The electron acts as a middle-man between the nucleus and the outside world. It gives us a way to have our cake and eat it - fast processing speeds from the electron, and long memory times from the nucleus."

The team used the stable isotope phosphorus-31 in conjunction with silicon-28 because it made an almost ideal configuration to allow this type of data transfer. However, it requires almost completely pure silicon-28, and that is a product which is extremely difficult to produce in any large quantity.





Eyeing the nucleus



One of the major obstacles this team overcame was the fragile state of previous attempts to use electrons for quantum memory. Electrons have proven unreliable due to frantic interaction with other electrons. The team chose the nucleus to store data, therefore bypassing that risk. The nucleus, while still very active, is much more stable than individual electrons.



Electrons are still used in this kind of quantum computer, but for data processing. The team refers to electrons as "processing qubits" and nucleus as a "memory qubit," due to their much greater storage rigidity.



In these early tests, the electron spin was successfully transfered to the phosphorus nucleus, and maintained in its transferred state for up to two seconds. This is thousands of times longer than any previous effort has ever achieved. In addition, the nuclear spin was transferred back to the electrons with a 90% accuracy over repeated tests.



Note:  Compare the two seconds of nuclear spin storage to the capacitor-based DRAM computer memory inside our PCs today which have to be refreshed many hundreds of thousands of times per second, or else they lose their data (charge). In early PCs, with 60- to 80-ns DRAM chips, this "refresh cycle" could be decreased in frequency slightly to gain a few percentage points in computer performance.





Future research



Now that it's been demonstrated atomic nuclei can be used to store data, the team's future efforts are focused on improving the purity of silicon to maintain better atomic and electron spin control, further developing readout mechanisms and increasing the atomic storage lifespan from just two seconds to far longer.

The researchers believe if they can get isotopically pure enough silicon crystals, with appropriately placed phosphorus atoms, then data could be stored for literally years without losing its state. This would make active computer memory devices also be solid state devices, meaning one form of memory for both processing and permanent storage - the instant on computer.





Uber-small, yet uber-powerful



There's a line in the first feature length Star Trek movie when Bones examines the artificial Lt. Ilia created by V-ger. He reports to Kirk that she has (among other things), "Molecule-sized multi-processor chips."

At the time of that movie (1980), our best integrated circuits were using a 10,000 nm manufacturing process. Compare that to today's 45nm and you'll see how far we've come in about 30 years. But still, how far are we away from molecule-sized multi-processors? Well, that's a question that can only be answered by work like this in the domain of quantum.


The team

Berkeley Lab research was funded in part by the U.S. Department of Energy’s Office of Science, through the Materials Sciences and Engineering Division of its Basic Energy Sciences programs, and also in part by the National Security Agency.

John Morton of Oxford University was the paper's lead author. Additional authors from Berkeley's Lab were Thomas Schenkel, Eugene Haller and Joel Ager. Additional authors included Richard Brown, Brendon Lovett and Arzhang Ardavan of Oxford University, and Alexei Tyryshkin, Shyam Shankar and Stephen Lyon, of Princeton, University.