back to article Compound that 'remembers' phase transitions could have uses in computer memory

Researchers in Switzerland have discovered a compound that can "remember" its previous phase transition states, offering potential applications in computer memory and information processing. Vanadium dioxide's (VO2) crystalline structure has long intrigued scientists. It changes structure (phase transition) at around 67°C (152 …

  1. Paul Herber Silver badge

    up to three hours

    'up to three hours'

    I can forsee a time when a 3 hour memory will, er, something, and who are you?

    1. IgorS

      Re: up to three hours

      3 hours is plenty for active working memory.

      At worst, you just refresh it once an hour, or so (like we do (much more frequently) with DRAM).

  2. Oglethorpe


    I wonder how accurate the timing has to be to distinguish between states compared to the jitter of a standard system clock.

  3. John Brown (no body) Silver badge

    This is the kind of science I love!

    I've often wondered about some of the SciFi speculations on weird compounds doing stuff we otherwise can't imagine and what we might be missing "because that's impossible" based on accepted science.

    As Isaac Asimov is often credited with, it's not "Eureka!" that precedes the best scientific discoveries. It's more likely "That's odd" or "That's funny".

    Maybe if we irradiate just the right type of Cheddar Cheese in just the right way with just the right amount and type of radiation, we too can make Cheddite and install an interstellar drive on a Jumbo Jet, or discover that there really is a liquid allotrope of iron to power spaceships :-)

    (probably showing my age here, also the age of those who get the references)

    1. Anonymous Coward
      Anonymous Coward

      Re: This is the kind of science I love!

      OK - I admit that I’m getting old.

      Ageing is compulsory, but growing up is optional,

    2. Twanky Silver badge

      Re: This is the kind of science I love!

      I have an unfortunate* habit of keeping nearly every book I buy. I'll have to dig those out of their packing cases from the last time we moved house about 10 years ago.

      *She who mustshould be obeyed uses a stronger word than 'unfortunate'.

    3. NightFox

      Re: This is the kind of science I love!

      Just as long as it's not that bloody awful ubiquitous Catherdral City stuff.

      Cheese already exhibits unique characteristics. For example, the induced force (expressed in Cheddars [Ch])* it exerts away from the block on, and perpendicular to the motion of, a knife when attempting to cut a straight slice for cheese on toast,

      *1 Ch being a derived SI vector quantity equal to 1 NSm^(-1), with S being the standardised Cheese Strength Index (1-5) as shown on the packaging and m being the thickness of the slice.

      1. low_resolution_foxxes Silver badge

        Re: This is the kind of science I love!

        Ah Cathedral City, blessed are thee thy cheese which is manufactured by a $14bn Canadian stock market company.

        And thy sales tactics, pretending to be a 'premium cheddar', overcharging a ridiculous price for cheese, with the apparent sole intent to sell at a 50% discount on a near-permanent basis.

        Truly, a company that could only be championed by Tesco.

        1. Anonymous Coward
          Anonymous Coward

          Re: This is the kind of science I love!

          Canadian Cheddar, British Sherry and French Letters - three things that you should never find in your mouth.

          I don't know if there are North American "microdairies" that do make quality cheese, but whatever makes it over this side of the pond is either industrial plastic optimised for melting onto burgers, or the same contaminated with jalapeno peppers, presumably intended for purposes of torture.

          1. Mike 137 Silver badge

            Re: This is the kind of science I love!

            "three things that you should never find in your mouth"

            There's another one worth noting - any of the supposedly foreign varietal beers that are now all brewed under license in one or other of the two giant dog piss factories (Hein.... or Carls....). Not surprisingly, they all taste the same.

            1. Muscleguy Silver badge

              Re: This is the kind of science I love!

              The solution is to brew your own. I have in my glass a porter which tastes distinctly of coffee but which does not contain coffee. I have in my beer fridge a barrel of live Pilsner which has never been warm since it was made. Dropped to 4C for serving I had it earlier. Extremely crisp, clean but tasty Pilsener.

              Next weekend I will brew my oatmeal stout which will taste of marmalade on toast. The result of taking a stout recipe adding oats and trying out an interesting hop.

              Just remember you MUST boil all the water you will use open for at least 15min the night before you brew. This removes chlorine which reacts with hops to give a TCP/medicinal taste which is redolent of much home brew. Not mine because I never deviate from this.

              I have reminders set on my phone so I do not fail to boil the brewing water.

              Ultra tasty live Real Ale at a fraction of the cost of buying it.

    4. Ian Johnston Silver badge

      Re: This is the kind of science I love!

      As Isaac Asimov is often credited with, it's not "Eureka!" that precedes the best scientific discoveries. It's more likely "That's odd" or "That's funny".

      One of the original purposes of the Journal of Irreproducible Results (which later became the Annals of Improbable Research) was to give researchers a place to publish strange observations which might be worth someone investigating, or which might fit in with somebody else's oddities to start a new train of thought.

      1. Michael H.F. Wilkinson

        Re: This is the kind of science I love!

        I love Annals of Improbable Research. I have three papers published there: "A Quantum Mechanical Interpretation of Homeopathy" (homeopathy can only work if you don't look), "Zero-Tolerance Math: a Defense of No-Math" (only the number zero should ever be used at school) and "Cogno-Intellectualism, Retorical Logic, and the Craske-Trump Theorem" (introducing the logical equivalent of Skinner's constant).

        It is great fun writing these things, and even greater fun when someone misses the joke and cites your work as if it was serious.

    5. Michael H.F. Wilkinson

      Re: This is the kind of science I love!

      The real danger of faffing around with cheese in the laboratory is that one might end up with something like Horace, the Lancre Blue, which you wouldn't like when it is angry

      Doffs hat to the late, great Terry Pratchett

    6. John Smith 19 Gold badge

      As Isaac Asimov is often credited with

      And in fact he has a short story for this.

      IIRC it's something called "Thiotonite" and dissolves in water.

      Except it will dissolve in water without water being present, if it has dissolved before.

      A sort of "Temporal memory"

      And IRL the classic physical example is NiTiNOL with it's "shape memory" (actually a range of Nicklel Titanium mixes).

      I will raise a glass to this curious phenomena and wish them good luck explaining it (and exploiting it maybe).

      1. Mike 137 Silver badge

        Re: As Isaac Asimov is often credited with


        Thiotimoline. A molecule "so distorted that one bond is forced into extension through the temporal dimension into the past; and another into the future 1"

        It actually races through time seeking water to dissolve in, provided you're convinced that you are going to provide it. (and incidentally it can carry anything to which it's attached physically with it through time).


        "The endochronic properties of resublimated thiotimoline" (Astounding, March 1948)

        "The micropsychiatric applications of thiotimoline" (Astounding, December 1953)

        "Thiotimoline and the Space Age" (Analog, September 1960)

        "Thiotimoline to the stars" (??, 1972) (also appeared in "Buy Jupiter", Granada Publishing 1976)

        1Thiotimoline to the stars

  4. Roger Greenwood


    But from Nature Electronics:- "Your institute does not have access to this article"


    (on the same day).

    1. navidier

      Re: Interesting

      > But from Nature Electronics:- "Your institute does not have access to this article"

      You too, huh?


      Obvious riposte: EPFL is not a US institution, but a CH one (as the F implies). However, there is a definite trend to Open Access. CERN publishes everything in OA, but then it's international, not CH. Research UK is moving strongly towards OA as well.

  5. Kevin McMurtrie Silver badge


    So it takes ns to write a '1' but writing a zero means waiting many minutes for memory to fade away? That's crap for traditional storage but maybe good for analog AI computations where it's advantageous to have unused paths fade away.

  6. StargateSg7

    You actually COULD use Nitinol (aka Nicket Titanium or ANY OTHER shape memory metal or even a ceramic) by using tiny nano-sliver (say 120 nanometres long) trapped in a well surround by ceramic or glass walls and floors and a clear EUV transparant ceiling with a nano-lens attached and then use pulses of Extreme-UV laser (aka E-UV laser) to twist or bend the trapped nano-sliver into a specific torsional shape (i.e. degree of twist), 3D-XYZ orientation/rotation or even bend (aka tensile deflection) to create a multi-bit memory cell. The difference in reflection between the shape from a baseline would indicate a specific bit value.

    The DEGREE of twist, rotation, orientation or bend could be a specific bit value at say 1 to 8 bits per memory cell depending on accuracy of the laser pulse and the continuous shape memory of the nano-sliver alloy or ceramic. At nano-scale sizes the SPEED AND STAYING POWER of any shape change can get into sub-ten nanosecond ranges for the change and the STAYING POWER can range into the many decades which is more than fast enough for use as a long term or short term memory cell and more than long enough staying power for archival uses.

    And at 120 nm for the nano-sliver of shape memory alloy or ceramic, means that data densities would be into the TENS of BILLIONS per square cm if you do stacked layering too! It would also be MUCH MORE RUGGED than ANY SSD drive. Size-wise, I would expect a single 3.5 inch SSD storage card using Nitinol-like Shape memory nano-slivers to be on the order of ONE PETABYTE PER DRIVE!

    For REALLY LARGE DATA DENSITY, use carbon nanotube rods trapped within a nanowell as the bit field value where the orientation or deflection from a plane IS the bit value and you would get TRILLIONS of bytes per square cm! Even though the wavelength of the Laser is much larger than the individual carbon nano-sliver/nano-rod size, you could use interferometry techniques that sample the SHAPE of the reflected interference patterns to represent the bit value and even write to it and NOT disturb nearby nanoslivers/nanorods using interferometry techniques with EUV lasers.

    The above idea and implementation is now FULLY FREE AND OPEN SOURCE UNDER GPL-3 Licence Terms!


POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like