Follow Slashdot stories on Twitter

 



Forgot your password?
typodupeerror
×
Data Storage IBM Science Technology

IBM Researchers Prove It Is Possible To Store Data In a Single Atom (techcrunch.com) 84

In an experiment published today in Nature, IBM researchers have managed to read and write data to a single atom. A previous atomic storage technique, as mentioned by TechCrunch, doesn't actually store data in the atom, but moves them around to form readable patterns. "This means that imbuing individual atoms with a 0 or 1 is the next major step forward and the next major barrier in storing data digitally, both increasing capacity by orders of magnitude and presenting a brand new challenge to engineers and physicists," reports TechCrunch. From the report: It works like this: A single Holmium atom (a large one with many unpaired electrons) is set on a bed of magnesium oxide. In this configuration, the atom has what's called magnetic bistability: It has two stable magnetic states with different spins (just go with it). The researchers use a scanning tunneling microscope (also invented at IBM, in the 1980s) to apply about 150 millivolts at 10 microamps to the atom -- it doesn't sound like a lot, but at that scale, it's like a lightning strike. This huge influx of electrons causes the Holmium atom to switch its magnetic spin state. Because the two states have different conductivity profiles, the STM tip can detect which state the atom is in by applying a lower voltage (about 75 millivolts) and sensing its resistance. In order to be absolutely sure the atom was changing its magnetic state and this wasn't just some interference or effect from the STM's electric storm, the researchers set an iron atom down nearby. This atom is affected by its magnetic neighborhood, and acted differently when probed while the Holmium atom was in its different states. This proves that the experiment truly creates a lasting, stored magnetic state in a single atom that can be detected indirectly. And there you have it: a single atom used to store what amounts to a 0 or a 1. The experimenters made two of them and zapped them independently to form the four binary combinations (00,01,10,11) that two such nodes can form.
This discussion has been archived. No new comments can be posted.

IBM Researchers Prove It Is Possible To Store Data In a Single Atom

Comments Filter:
  • "in" an atom or on it or to it. atoms want to know.
  • Oh no, does this mean we will soon have a shortage of atoms?
  • by gringer ( 252588 ) on Wednesday March 08, 2017 @08:57PM (#54003921)

    Richard Feynman's talk discussed manipulation at the atomic level as a target to strive for, demonstrating how much room there is for miniaturisation.

    Now it seems that we're going to need to drop to the sub-atomic level for further manipulation.

    • by Anonymous Coward

      the actual Feynman's transcript of the lecture is here, it's not a hard read:

      https://www.pa.msu.edu/~yang/RFeynman_plentySpace.pdf

    • For now atomic manipulation at large quantities is very far off (they only used 2 atoms in that experiment).

      Our problem currently with storage is that we are kind of bound to 2D, only in the last couple of years we started actually build storage that has bits layered with 3D NAND, but the Z axis is still much, much smaller than the others. In my opinion, the future of large storage should more a cube than an atom.

  • Very interesting. (Score:4, Insightful)

    by Nutria ( 679911 ) on Wednesday March 08, 2017 @09:11PM (#54003977)

    Where does this research fall on the Munroe Scale?

    https://xkcd.com/678/ [xkcd.com]

  • It sounds like we may have finally hit the limit for density of data storage. Kinda hard to get below the atomic level.

    Think of the amount of data you could store in a single copper BB if the atoms could be used as memory. Holy fuck.

    Ten million Libraries of Congress? 100 million? A billion?

    • That's why I advocate research into creating pocket universes, especially ones with a flow of time faster than our own. Then the amount of data and processing can become practically infinite at the point where the pocket universe intersects our own.
      • The problem with pocket universes is always laundry. You forget to take them out and then when you open the drier you've got universe smeared all over your clothes.

    • Re:Liit hit! (Score:4, Insightful)

      by kenai_alpenglow ( 2709587 ) on Wednesday March 08, 2017 @10:26PM (#54004205)
      One word: Quarks.
    • by msauve ( 701917 )
      "Kinda hard to get below the atomic level."

      That's so old school (really, really, old!). Since then, we've discovered electrons, protons, neutrons, and even more turtles holding them up.

      Heck, on a day-to-day basis we transfer info using sub-atomic photons (your TV remote!). The real limit, as far as we know it, would be something at Planck scale.
    • Think of the amount of data you could store in a single copper BB if the atoms could be used as memory.

      Yes, but how much shielding against stray radiation will be required to preserve the integrity of all that data? Not to mention protection against magnetic fields - I imagine it would be pretty easy to induce eddy currents that would provide 150mV worth of potential at the requisite 10 uA to flip a few bits. I'm sure they'll solve such problems, but I think we're a long way from seeing a practical implementation.

      • Yes, but how much shielding against stray radiation will be required to preserve the integrity of all that data?

        You'd just use error-correcting code. Store 3 copies (or 5 or 10 or whatever) and do comparisons...it would be easy to spot the copy that's changed from the ones that hadn't. The old "tell me three times" routine:

        1) My dog likes to play in the yard.
        2) My dog likes to play in the yard.
        3) My dog likes to play in the yard.
        4) My dog likes to play iK the yard.
        5) My dog likes to play in the yard.
        6) My dog likes to play in the yard.
        7) My dog likes to play in the yard.

        Which one got corrupted? The one that doesn't m

        • You'd just use error-correcting code. Store 3 copies (or 5 or 10 or whatever) and do comparisons...

          Yeah, you're right - I guess with that high a data density you can afford to give some up for the sake of redundancy. Slows down the throughput a bit though.

          • Yeah, you're right - I guess with that high a data density you can afford to give some up for the sake of redundancy. Slows down the throughput a bit though.

            Yeah, there's always a trade-off, but in most cases accuracy takes precedence over speed. Of course, by the time this is a reality we'll probably have banks of quantum computers built in to the BB, each doing a kabillion operations per nanosecond. :)

          • by Imrik ( 148191 )

            To up the throughput (for read options) you can use just a few bits to detect if there is an error and then go do a full cross check if one is detected.

    • by Imrik ( 148191 )

      Ignoring that it is probably possible to get below the atomic level, it's also possible that we might be able to induce more than two states in an atom. This could happen by either having an attribute with more than two states that could be changed between, or by having multiple attributes on the same atom.

    • by krray ( 605395 )

      > Think of the amount of data you could store in a single copper BB if the atoms could be used as memory. Holy fuck.

      > Ten million Libraries of Congress? 100 million? A billion?

      Well -- copper BB's are usually copper coated. Let's assume it is SOLID copper... BB's weigh 5.28 grains at ~6mm; but that is copper coated. I don't remember my density formula at the moment. Let's call it 5 grains.

      Copper has a molar mass of ~63.5 g/mol. One mole of an element is defined as 6.022 x 10^23 -- so there's that many

      • That's 40 zettabytes or 40 sextillion bytes.

        Roughly. :)

        So by the time this becomes commercially available, that would be roughly 2 cat pictures and a Word document.

        • So by the time this becomes commercially available, that would be roughly 2 cat pictures and a Word document.

          Probably, but instead of two cat pictures it will be two interactive cat holograms and the "Word document" will be a quasi-intelligent construct capable of tailoring itself to suit readers across a broad spectrum of languages, cultures, and prior training.

          It's true that new media tends to take up more space, but we're not just doing the same thing less efficiently. New media takes up more space because it contains more information. Is this extra information capability sometimes wasted? Sure. But it's not li

          • Probably, but instead of two cat pictures it will be two interactive cat holograms

            Cat holograms! Be still mah heart!

            Will the cats have flying cars? Because that would be worth waiting for.

    • Think of the amount of data you could store in a single copper BB

      Think of the amount of data you could store in the number of atoms equivalent to a single copper BB... arranged in such a way your "read/write device" can access those atoms.

      How small can you make THAT contraption?
      • Zero because they've done it with holmium atoms on a Magnesium Oxide bed. I don't claim to know how even platter based storage works to know how to answer that question. But the amount of atoms in a BB is theoretically 3 Zetabytes (3 billion terabytes) worth of information could be stored. Pie in the sky, but it's a beautiful sky....
  • by dwywit ( 1109409 )

    This is why I don't normally object to the volume of patents filed by and granted to IBM. It's R&D that actually leads to useful things.

    • Re:R&D (Score:4, Interesting)

      by Goldsmith ( 561202 ) on Thursday March 09, 2017 @01:47AM (#54004759)

      IBM certainly has a well-earned reputation of being the premier industrial research lab in nanotechnology, but they also have a well-earned reputation for keeping the technology at the grant and publication stage much longer than necessary.

      IBM invented the STM, but it was about 15 years before someone else brought one to market. IBM invented carbon nanotube transistors, ran the premier group in CNT research for over 20 years, and then shut it down without attempting to develop a product.

      This would be ok, but they've also sucked up a tremendous amount of grant money and investment targeting nanotech commercialization over the last 30 years without actually commercializing any of the technologies they've worked on.

      I am a nanotechnology researcher. I know and greatly respect many researchers at IBM. It's disappointing that the company has decided not to participate in developing products using their technologies.

      • they've also sucked up a tremendous amount of grant money and investment targeting nanotech commercialization over the last 30 years without actually commercializing any of the technologies they've worked on.

        Interesting. Why do you think this is? Are the technologies simply not commercially viable? Is it government interference?

  • That's either going to bring the price of STMs way down.

    Or the price of memory is going to go through the roof.

  • So Atom Ant is real!

  • This and other similar forms of radical memory storage are frightening to me. Not because of anything they do or are, but because they carry a huge implication.

    If we humans, certainly not all that advanced, have come up with ways to store data in single atoms like this, or in forms like grains of sand, then what does this suggest about how more advanced species might store and distribute their data? And would we even know it was there?

    Suppose right now, the beaches of earth were covered in grains of sand

    • by Nutria ( 679911 )

      I bet you're also one of the 3% that expresses fear at least once a week at losing your job to a robot.

      • Nope. Never once. But I will happily build a robot to take your job. Just business.

        • But I will happily build a robot to take your job. Just business

          There's nothing wrong with automation and robotics, of course. However, when someone says "just business", they're trying to justify something they find ethically dubious by claiming it's OK because it's for money.

          It reminds me of the line from Grosse Point Blank:

          "No, no, no, a psychopath kills for no reason. I kill for money. It's a job. That didn't come out right."

        • by Wulf2k ( 4703573 )

          I would love having a robot do my job.

          When can you get started?

    • We could be surrounded by alien data sand waiting to be discovered, except somebody vacuumed it up from the floor mats in their car and threw it out.

      This is just retarded. Anyone storing data on something so incredibly volatile deserves to have their data destroyed.

      • We could be surrounded by alien data sand waiting to be discovered, except somebody vacuumed it up from the floor mats in their car and threw it out.

        This is just retarded. Anyone storing data on something so incredibly volatile deserves to have their data destroyed.

        Retarded? You seem to be. Storing or distributing data like that would be a great test. If a feeble species like ours never finds it, oh well, we were just too retarded. But if we did find it, hey, maybe we are not all morons like you after all. Eh. Well, we probably are.

        • by Anonymous Coward

          You should definitely get a microscope and go start looking at all the sand! You don't want to be morons like us!

    • by Anonymous Coward

      I believe the phrase "crying over spilt milk" applies well here. If this is causing you nightmares, I suggest visiting a psychiatrist.
      So what if we are accidentily destorying alien knowledge? If we cannot tell difference between the encoded forms of matter and unencoded forms, and are unable to read or use it, what use is it?
      It is unlikely that the sand and dirt in your floor mats and car is special alien data compared to sand and dirt in some remote area untouched by civilisation since the aliens visited,

  • by Anonymous Coward

    If one atom holds one bit, then one gram of holmium holds about 456 exabytes of data.

    For comparison, DNA has sucessfully held data at 215 petabytes/gram.

  • by Anonymous Coward

    It's hardly just a single atom involved if they required a magnesium oxide bed for it to rest on.

    • This is a poorly worded (and deceptive) headline. Each Holmium atom is only encoded with a 1 or a 0 depending on the spin or magnetic orientation of the atom (up or down). Therefore only one bit encoded on each atom, hence datum. The headline makes it seem like they are encoding more than that on each atom, which if you learn a bit about electron orbitals is unlikely (at least as a long-term storage) especially at room temperature.
  • It's amazing to think that we've gone from the discovery of the electron in (? 1897) to this...

    However, if we extrapolate forwards from this discovery, then is it theoretically possible to construct atomic-scale logic gates? Could we conceivably construct a lattice or matrix of atoms - perhaps held on some form of crystalline substrate, in which we could "inject" a signal in the form of a single free electron, only to have that propagate through the structure in a similar manner to the way that logic flo
  • You can only read the data if you don't know the location of the atom.and the data becomes unreadable if you know the location of the atom
  • Is this the return of the Millipede project?
    Let's hope they can get it out the door this time.

  • Every few months for the past 20 years there is some news article that IBM has found a way to store data in some strange new way. I remember decades ago something about storing data in crystals with lasers or some shit. Yet what technology are we using today? The same as 20 years ago, semiconductors and disk drives. Can IBM please STFU until they have a shippable product?

  • Same story on Engadget: "IBM built an atomic hard drive! It's 100,000 times more efficient than the state-of-the-art."
  • The good news: new harddrives will have the size of a pin. The bad news: it'll require the scanning tunneling microscope attached to work: http://www-03.ibm.com/ibm/hist... [ibm.com]

    So, now, it'll require more 10 years to work to reduce the scanning tunneling microscope to the size of current harddrives.

Some people manage by the book, even though they don't know who wrote the book or even what book.

Working...