News & Analysis
Comment
wilber_xbox
@pixies, 'never' is a strong word and i have learned during my studies to not ...
iniewski
I belive that people typically underesitmate efforts required in materials ...
Spintronics aims for atomic memories
R Colin Johnson
12/16/2010 2:01 PM EST
PORTLAND, Ore.—Spintronics memories should store quantum information on the individual atoms, rather than electrons, according to University of Utah researchers. Led by professor Christoph Boehme, a team of researchers demonstrated how it's done by reading and writing spin onto phosphorus atoms in a silicon substrate, achieving a refresh time of 112 seconds—thousands of times longer than memories storing information on electron spin.
"Nuclear spins where we store information have extremely long storage times before the information decays," said Boehme. "Its spin isn't messed with by what's going on in the clouds of electrons around the nucleus."
To prove that atomic nuclei store spins more reliably than electrons, the Boehme group's experimental demonstration used a phosphorus-doped silicon chip measuring just 1 millimeter square. After supercooling the material to 3.2 degrees Kelvin, an intense magnetic field of nearly 8.6 Tesla aligned the spins of the phosphorus atom's electrons, which was then transferred to the nuclei by FM-band radio waves.
Up to 112 seconds later, the group was able to demonstrate that near-terahertz wavelengths could then be used to transfer the nuclei spins back onto the atom's electrons, allowing its value to be electrically read-out as a current. The researchers showed that the mechanism was reliable by reading and writing the phosphorus nuclei 2000 times. Next the researchers are going to try the same demonstration with a single phosphorus atom.
"Nuclear spins where we store information have extremely long storage times before the information decays," said Boehme. "Its spin isn't messed with by what's going on in the clouds of electrons around the nucleus."
To prove that atomic nuclei store spins more reliably than electrons, the Boehme group's experimental demonstration used a phosphorus-doped silicon chip measuring just 1 millimeter square. After supercooling the material to 3.2 degrees Kelvin, an intense magnetic field of nearly 8.6 Tesla aligned the spins of the phosphorus atom's electrons, which was then transferred to the nuclei by FM-band radio waves.
Up to 112 seconds later, the group was able to demonstrate that near-terahertz wavelengths could then be used to transfer the nuclei spins back onto the atom's electrons, allowing its value to be electrically read-out as a current. The researchers showed that the mechanism was reliable by reading and writing the phosphorus nuclei 2000 times. Next the researchers are going to try the same demonstration with a single phosphorus atom.
A phosphorus-doped one millimeter square silicon chip demonstrate how data can be stored in magnetic "spins" on the nuclei of phosphorus atoms.
|
|
|
|
|
Navigate to related information
R_Colin_Johnson
12/16/2010 4:55 PM EST
NIST has been demonstrating how to store quantum information on atoms and last week Duke University and the University of Wisconsin showed how lasers could be focused with MEMS micro-mirrors to read and write them in their gaseous state: http://bit.ly/NextGenLog-h8ti
However, this University of Utah researcher takes the next step--holding out the promise of being able to store quantum information on solid-state atoms that can be read and written electrically, opening the door to quantum computers fabricated on chips just like conventional computers.
Sign in to Reply
wilber_xbox
12/16/2010 6:52 PM EST
quantum computers are still decades away, in my opinion, as simultaneous manipulation of group of atoms/electrons/photons is required. this article has good news about the long decay time of the information, which is very crucial for any kind of practical use.
Sign in to Reply
tedt
12/17/2010 11:08 AM EST
After sucessfully developing a practical quantum entanglement technology, I wonder what will happen when like material is placed near the device. Will alpha/beta/gamma radiation affect atomic spin? Will radiation affect the ability to refresh data over time? Plastic or epoxy packaging can have a small radioactive component. These and other questions must be answered to achieve a working, reliable product.
Sign in to Reply
yalanand
12/18/2010 7:09 AM EST
I have the same doubt. Are there any factors which could affect the spin of the nuclei and in-turn loose your data ?
Sign in to Reply
pixies
12/17/2010 4:59 PM EST
It seems that this device can never operate at room temperature. I do not see an application in the foreseeable future. But it is really cool physics.
Sign in to Reply
wilber_xbox
12/19/2010 6:46 AM EST
@pixies, 'never' is a strong word and i have learned during my studies to not say never as in Science impossible is just a breakthrough result away!
Sign in to Reply
prabhakar_deosthali
12/18/2010 5:17 AM EST
What I see is the demise of the current semiconductor memory technologies by 2020!
Sign in to Reply
iniewski
12/18/2010 12:20 PM EST
I belive that people typically underesitmate efforts required in materials science, this is not software or another chip design in well established CMOS process. It might easily take 20 years to make high volume, manufacturable spin based device. Severeal examples of new material developments that took 30-40 years do exist...Kris
Sign in to Reply