The world's smallest magnetic data storage unit

Jan 12, 2012
The world's smallest magnetic data storage unit
welve iron atoms make up the world's smallest magnetic data storage unit to date. The antiferromagnetic order in the iron atom array is revealed by spin-polarized imaging with a scanning tunneling microscope. Credit: Sebastian Loth/CFEL

Scientists from IBM and the German Center for Free-Electron Laser Science (CFEL) have built the world's smallest magnetic data storage unit. It uses just twelve atoms per bit, the basic unit of information, and squeezes a whole byte (8 bit) into as few as 96 atoms. A modern hard drive, for comparison, still needs more than half a billion atoms per byte. The team present their work in the weekly journal Science this Friday (13 January 2012). CFEL is a joint venture of the research centre Deutsches Elektronen-Synchrotron DESY in Hamburg, the Max-Planck-Society (MPG) and the University of Hamburg. "With CFEL the partners have established an innovative institution on the DESY campus, delivering top-level research across a broad spectrum of disciplines," says DESY research director Edgar Weckert.

The data storage unit was built atom by atom with the help of a (STM) at IBM's Almaden Research Center in San Jose, California. The researchers constructed regular patterns of , aligning them in rows of six each. Two rows are sufficient to store one bit. A byte correspondingly consists of eight pairs of atom rows. It uses only an area of 4 by 16 nanometres (a nanometre being a millionth of a millimetre). "This corresponds to a that is a hundred times higher compared to a modern hard drive," explains Sebastian Loth of CFEL, lead author of the Science paper.

Data are written into and read out from the nano storage unit with the help of an STM. The pairs of atom rows have two possible , representing the two values '0' and '1' of a classical bit. An from the STM tip flips the magnetic configuration from one to the other. A weaker pulse allows to read out the configuration, although the nano magnets are currently only stable at a frosty temperature of minus 268 degrees Centigrade (5 Kelvin). "Our work goes far beyond current ," says Loth. The researchers expect arrays of some 200 atoms to be stable at room temperature. Still it will take some time before atomic magnets can be used in data storage.

The world's smallest magnetic data storage unit
This graphic shows the atomically precise assembly of an atomic-scale antiferromagnet with the tip of a scanning tunneling microscope. Iron atoms are placed onto a copper nitride surface and bound by two nitrogen atoms (blue rods) into a regular array separated by one copper atom (yellow). Credit: Sebastian Loth/CFEL

For the first time, the researchers have managed to employ a special form of magnetism for data storage purposes, called antiferromagnetism. Different from ferromagnetism, which is used in conventional hard drives, the spins of neighbouring atoms within antiferromagnetic material are oppositely aligned, rendering the material magnetically neutral on a bulk level. This means that antiferromagnetic atom rows can be spaced much more closely without magnetically interfering with each other. Thus, the scientist managed to pack bits only one nanometre apart.

"Looking at the shrinking of electronics components we wanted to know if this can be driven into the realm of single atoms," explains Loth. But instead of shrinking existing components the team chose the opposite approach: "Starting with the smallest thing - single atoms - we built data storage devices one atom at a time," says IBM research staff member Andreas Heinrich. The required precision is only mastered by few research groups worldwide.

"We tested how large we have to build our unit to reach the realm of classical physics," explains Loth, who moved from IBM to CFEL four months ago. Twelve atoms emerged as the minimum with the elements used. "Beneath this threshold quantum effects blur the stored information." If these quantum effects can somehow be employed for an even denser data storage is currently a topic of intense research.

With their experiments the team have not only built the smallest unit ever, but have also created an ideal testbed for the transition from classical to quantum physics. "We have learned to control quantum effects through form and size of the iron atom rows," explains Loth, leader of the Max Planck research group 'dynamics of nanoelectric systems' at CFEL in Hamburg and the Max-Planck-Institute for Solid State Research at Stuttgart, Germany. "We can now use this ability to investigate how quantum mechanics kicks in. What seperates quantum magnets from classical magnets? How does a magnet behave at the frontier between both worlds? These are exciting questions that soon could be answered."

A new CFEL laboratory offering ideal conditions for this research will enable Loth to follow up these questions. "With Sebastian Loth, one of the world's leading scientists in the field of time-resolved scanning tunneling microscopy has joined CFEL," stresses CFEL research coordinator Ralf Köhn. "This perfectly complements our existing expertise for the investigation of the dynamics in atomic and molecular systems."

Explore further: The unifying framework of symmetry reveals properties of a broad range of physical systems

More information: "Bistability in atomic-scale antiferromagnets," Science ,Bd. 335, S.196, DOI: 10.1126/science.1214131

Related Stories

Single atom stores quantum information

May 02, 2011

(PhysOrg.com) -- A data memory can hardly be any smaller: researchers working with Gerhard Rempe at the Max Planck Institute of Quantum Optics in Garching have stored quantum information in a single atom. ...

Physicists capture first images of atomic spin

Apr 26, 2010

(PhysOrg.com) -- Though scientists argue that the emerging technology of spintronics may trump conventional electronics for building the next generation of faster, smaller, more efficient computers and high-tech ...

Nanoscale Dominoes: Magnetic Moments Topple Over in Rows

Sep 24, 2008

Physicists at the Institut für Festkörperforschung in Germany have discovered a type of domino effect in rows of individual manganese atoms on a nickel surface. They determined that the magnetic arrangement of these nanowires ...

Researchers discover 'superatoms' with magnetic shells

Jun 08, 2011

(PhysOrg.com) -- A team of Virginia Commonwealth University scientists has discovered a new class of 'superatoms' – a stable cluster of atoms that can mimic different elements of the periodic table – ...

Recommended for you

Single laser stops molecular tumbling motion instantly

3 hours ago

In the quantum world, making the simple atom behave is one thing, but making the more complex molecule behave is another story. Now Northwestern University scientists have figured out an elegant way to stop a molecule from ...

What time is it in the universe?

Aug 29, 2014

Flavor Flav knows what time it is. At least he does for Flavor Flav. Even with all his moving and accelerating, with the planet, the solar system, getting on planes, taking elevators, and perhaps even some ...

Watching the structure of glass under pressure

Aug 28, 2014

Glass has many applications that call for different properties, such as resistance to thermal shock or to chemically harsh environments. Glassmakers commonly use additives such as boron oxide to tweak these ...

User comments : 24

Adjust slider to filter visible comments by rank

Display comments: newest first

hopefulbl
4.3 / 5 (4) Jan 12, 2012
really cool
Vendicar_Decarian
2.7 / 5 (7) Jan 12, 2012
really unreliable
Callippo
2 / 5 (4) Jan 12, 2012
Why not - at the zero Kelvin temperature? To maintain the reproducibility at the room temperature would require many thousands of atoms - and this is where the size limit of contemporary storage technologies begins. You cannot beat the uncertainty principle so easily under situation, when the contemporary technology is already limited with it. It's one of fundamental physical principles.
antialias_physorg
5 / 5 (10) Jan 12, 2012
To maintain the reproducibility at the room temperature would require many thousands of atoms

I think you missed this part of the article:
The researchers expect arrays of some 200 atoms to be stable at room temperature.

I'd hazard that they are more hip to physics than you are.
Callippo
1 / 5 (7) Jan 12, 2012
I'd hazard that they are more hip to physics than you are
He laughs best, who laughs last. http://www.realcl...170.html
Sonhouse
5 / 5 (5) Jan 12, 2012
I'd hazard that they are more hip to physics than you are
He laughs best, who laughs last. http://www.realcl...170.html


What has that got to do with the amount of atoms it would take to make the smallest bits on a hard drive?

Since it now takes about a billion atoms per bit on present day hard drives, you are just quibbling if you say thousands and he says hundreds, in either case you are talking about EXAbyte hard drives at that rate.
Norezar
1 / 5 (1) Jan 12, 2012
I'd hazard that they are more hip to physics than you are
He laughs best, who laughs last. http://www.realcl...170.html


What has that got to do with the amount of atoms it would take to make the smallest bits on a hard drive?

Since it now takes about a billion atoms per bit on present day hard drives, you are just quibbling if you say thousands and he says hundreds, in either case you are talking about EXAbyte hard drives at that rate.


I shudder at the thought of fragmentation...
PhotonX
2.5 / 5 (2) Jan 13, 2012
Sets the upper boundary to Moore's Law, I guess, at least for disc surface storage. Ah, how I long for the heady and hearty days of the 5 1/2' disk and 8-bit data flags!

I can't wait to see how they manage to miniaturize scanning tunneling microscopes into the read/write heads in those drives. Except...I probably won't life that long, and we'll likely have a whole different way of reading bits by then. Nano-machines to the rescue.

Since it now takes about a billion atoms per bit on present day hard drives, you are just quibbling if you say thousands and he says hundreds, in either case you are talking about Exabyte hard drives at that rate.


I think the article says half a billion per byte, but even so I think we get an exabyte drive if the limit can be maxed out, and if we consider the current limit at 1 TB per platter. (I'm using all my fingers and toes here, so someone correct me if I'm wrong.) Awesome in either case.
finitesolutions
4 / 5 (5) Jan 13, 2012
Mechanical spinning hard drives are NOT the future.
Having a spinning disk is slow and erroneous.
SSD is the way forward. SSD densities only get better and their performance already surpasses the best plate disks.
Norezar
3.7 / 5 (3) Jan 13, 2012
Mechanical spinning hard drives are NOT the future.
Having a spinning disk is slow and erroneous.
SSD is the way forward. SSD densities only get better and their performance already surpasses the best plate disks.


Agreed.

It's a waste of effort fooling with this.
kaasinees
2.3 / 5 (6) Jan 13, 2012
Wrong, it will be the future for making backups. Tapes are still used as well.
antialias_physorg
3 / 5 (2) Jan 13, 2012
I can't wait to see how they manage to miniaturize scanning tunneling microscopes into the read/write heads in those drives. Except...I probably won't life that long,

Probably much like the millipede drive which already exists (though it has not been turned into a commercial product because it is too expensive compared to other forms of memory).
http://en.wikiped...e_memory
kaasinees
1.8 / 5 (5) Jan 13, 2012
for teh bloody idiot down rating me:
http://www.tested...ls/1926/

There you go, you cant make backups on SSD's. Idiots be idiots.
So this research is still very usefull. Yes the speeds may be slow be thats not the point of hard drives.
Deathclock
1 / 5 (1) Jan 13, 2012
I shudder at the thought of fragmentation...


With such high density and therefore capacity one could use very large cluster sizes to combat fragmentation.
antialias_physorg
not rated yet Jan 13, 2012
I shudder at the thought of fragmentation...

Why? SSDs do well without defragmentation (actually you shouldn't defrag an SSD at all). Nothing in the article indicates that this would be used for rotary disc storage. There are other geometries available.
kaasinees
1 / 5 (2) Jan 13, 2012
I shudder at the thought of fragmentation...

Why? SSDs do well without defragmentation (actually you shouldn't defrag an SSD at all). Nothing in the article indicates that this would be used for rotary disc storage. There are other geometries available.

There are special file systems for SSD's that deal better/faster with these devices. Also if you do not want fragmentation use a different file system. Fragmantation is mainly a windows issue.
Deathclock
3 / 5 (2) Jan 13, 2012
I shudder at the thought of fragmentation...

Why? SSDs do well without defragmentation (actually you shouldn't defrag an SSD at all). Nothing in the article indicates that this would be used for rotary disc storage. There are other geometries available.

There are special file systems for SSD's that deal better/faster with these devices. Also if you do not want fragmentation use a different file system. Fragmantation is mainly a windows issue.


Uhh... what?

No, not at all... fragmentation is a fundamental problem with storing information as a linear array of binary information. All file systems deal with fragmentation differently, but it is an issue for all of them to overcome because the problem is not introduced with the design of the file system but, as I said, is a fundamental problem when you combine variable length files and the ability to create and delete files at will.
finitesolutions
4.5 / 5 (2) Jan 13, 2012
As far as backup concerns I would use a medium on which some permanent irreversible damage can be done. Backup is written only once for long term storage ( centuries ). ROM mediums qualify better than magnetic mediums.
Deathclock
1 / 5 (1) Jan 13, 2012
When a linear array of storage is empty you can easily start one file right after the last file and the fill the array with zero fragmentation. However, as soon as you start deleting files at will variable sized "holes" are left in the array. This leads to conditions where a new file needs to be saved, and there are enough total free slots in the array, but no single contiguous section large enough to store the file. There are a few things you can do to solve this:

1 - Split the file up into smaller chunks that will fit in the various size "holes" in the array (this is fragmentation)
2 - Compact the free space into a single contiguous array by physically moving data around (this is compaction/defragmentation)
3 - Prevent the user from saving the file (typically considered unacceptable)

If you can think of another option that works better than the two already in use you will be a millionaire.
Deathclock
1 / 5 (1) Jan 13, 2012
I shudder at the thought of fragmentation...

Why? SSDs do well without defragmentation (actually you shouldn't defrag an SSD at all). Nothing in the article indicates that this would be used for rotary disc storage. There are other geometries available.


That does not mean that fragmentation does not occur. Fragmentation on an SSD is less of a concern because there is virtually no seek time so having one file split up in multiple places on the array has a less significant impact on performance.
antialias_physorg
5 / 5 (1) Jan 13, 2012
That does not mean that fragmentation does not occur.

Fragemantation does occur on SSDs. That's the point. But with the limited amount of rewrite cycles you really don't want to defragment them (defragmentation usually puts all the data at the 'beginning' of the medium...this makes for uneven read/write use over the lifetime of the medium.)

SSDs use lookup tables to see where the fragged pieces are - which is fine. (Also one reason why your OS will report that your SSD has filled up, when it actually hasn't)

The type of memory described in the article could well be manufactured in array form instead of disc form and use similar methods like the millipede link I posted for reading/writing. In this case defragmentation would not be an issue because it would not be needed.
Deathclock
3 / 5 (2) Jan 13, 2012
It's all array form... memory, regardless of the physical medium providing it, is always considered a linear array of bits. Like I said the reason SSD's don't have to worry about fragmentation is due to their practically non-existent seek time. (they still require a finite time to address the memory, but it is orders of magnitude less than to physically move the R/W heads of a traditional HDD). I deal with fragmentation in the flash memory of the products I design, but I do it intelligently in real time, moving one file to a more efficient location to combine two blocks of free space into one as needed. It's completely transparent to the user. This is not implemented for performance, more due to my laziness in developing a mechanism to keep track of the location of file fragments.
Eikka
not rated yet Jan 16, 2012
Like I said the reason SSD's don't have to worry about fragmentation is due to their practically non-existent seek time.


It does come at a cost though. Once you have a file in 10,000-100,000 little pieces, it takes considerable time from the operating system to combine the data from the file table, compute where the file apparently is, and send the read requests one by one.

Waiting for small bits of data is inefficient because of the latencies of the path between the disk, CPU, and the system RAM. You're basically hauling gravel with a wheelbarrow a single rock at a time. It's not just the seek latency of the drive, although it comes at a greater cost with ordinary spinning hard drives.

The usual suspect in a PC is the virus protection suite, which gets regular updates and appends them to a database file. Sooner or later it's going to be in half a million little pieces and the whole computer bogs down trying to access it all the time.
MarkyMark
not rated yet Jan 19, 2012
I'd hazard that they are more hip to physics than you are
He laughs best, who laughs last. http://www.realcl...170.html

I love it when Callipo pretnds to know stuff. Hows things in Aether land ( or whatever it is called) these days?