back to article Much more Moore's Law: Wonder-stuff graphene transistor trickery

Scientists have developed a way to “chemically grow” transistors that are only a few atoms thick in a bid to give poor old battered Moore's Law another reprieve, according to new research published in Nature Nanotechnology. Packing more transistors on a microchip becomes increasingly difficult as shortening the distance …

  1. NoneSuch

    Well, yeah. But after that they are out of tricks. :)

    1. BebopWeBop Silver badge

      Well I am not a physicist or materials scientist, but it is remarkable how often the demise/brick wall for all sorts of things - oil, computers, disks, communications over copper ect ect ect have been predicted, only to have science/technology ride to the rescue.

      1. Pascal Monett Silver badge
        Thumb Up

        That's because we are continually discovering new things about how Nature works.

        Science is a wonderful thing.

      2. Charles 9 Silver badge


        Except they haven't tried to do this in volume. How are you going to precisely cut the channels in the graphene? Make sure the nucleation doesn't leave holes or impurities, and so on?

        This is nothing more than yet another "work in progress" article that whets our appetites but in the end leaves us hungry. When's one of these things actually going to hit mass production?

        1. Mage Silver badge
          1. asdf Silver badge

            Re: Except...

            Also the funny thing about mass production is when a market is fairly mature (sadly low single digit growth is the new norm in ICs) and the current products meet almost all needs (except of shareholders) and the new technology is orders of magnitude more expensive to develop or produce it tends to not come to market at light speed. See 450 mm wafer production.

        2. Anonymous Coward
          Anonymous Coward

          Re: Except...

          When's one of these things actually going to hit mass production?

          All the time, just look at the generational changes in (for example) CPUs or GPUs. Or HDD technology, or NAND. But you're choosing not to notice?

          I can remember the days of the 8086, the earliest consumer PCs with (gasp!) 40 MB hard disks....

          1. asdf Silver badge

            Re: Except...

            The step up in capabilities between what you could do with a 8086 vs 286 vs 386 vs Pentium Pro is a heckuva lot bigger than between generations of microprocessors today.

          2. Malcolm Weir Silver badge

            Re: Except...

            So... you're a recent arrival, then: the earliest PCs had 8088's and 10MB HDDs!

            1. Deltics

              Re: Except...

              Nope. The earliest PC's didn't have HDD's (though one could be added with later expansion boxes and/or as long as you also upgraded the powersupply).

              It was 2 years before HDD's appeared as standard equipment (in the XT)

              1. Jeffrey Nonken Silver badge

                Re: Except...

                Son, move over. It's time for your nap.

                The earliest commercial Personal Computers ran on 8-bit micros (typically the Z-80 or 6502) and had either 8" or 5 1/4" floppies as a high-priced option; otherwise, your mass storage was audio tape. The 8" floppies had more storage until they invented High Density encoding.

                Of course, there were Personal Computers before that, but those were pretty much just for dedicated hobbyists. I myself owned a KIM-1 and still own an IMSAI 8080.

                1. mosw

                  Re: Except...

                  And I remember when you had to start your PC with a hand crank and spin the hard disk by hand. You tell kids these days and they don't believe you!

          3. Charles 9 Silver badge

            Re: Except...

            "All the time, just look at the generational changes in (for example) CPUs or GPUs. Or HDD technology, or NAND. But you're choosing not to notice?"

            Yes, until I can actually SEE the technology face to face. GPUs and the like I can at least actually BUY. Why get all worked up on something that may not show up for years...if at all.. STREET release; THAT'S when I'll pay attention.

          4. John Brown (no body) Silver badge

            Re: Except...

            I can remember the days of the 8086, the earliest consumer PCs with (gasp!) 40 MB hard disks...."

            Sheesh! Kids today! <shakes head>

          5. Mage Silver badge

            Re: Except... 40 MB hard disks?

            First desktop 1980s had 5M Byte full height drives. They didn't have intel CPUs very often (8085)., 6502, 6809, Z80 etc.

            I have a 10M MFM drive on an 8088 in the attic.

            By 1991 the 20M was common and you had to use evil disk compression to get the 40M you needed for Windows.

            The Actual IBM PC had one or two 180K SS or 360K SD SD 5.25" Floppies. Though Apple II had only 100K floppies. Some non-IBM computers had much higher capacity floppies (8" or 5.25") and hard drives. The XT was first IBM PC with an official hard drive (and was 10Mbyte).

  2. Mage Silver badge

    Very thin?

    However the issue isn't reducing the thickness of devices, but the area that is required on the chip.

    I maybe scanned the article too quickly, but what area is the entire device compared to current ones?

    1. Anonymous Coward
      Anonymous Coward

      Re: Very thin?

      The thick they talk about is the thickness of the conducting paths being left in the chip. The narrower these paths, the less heat you generate from the conductor's resistance...but also the greater the chance electrons, so to speak, derail.

      1. Mage Silver badge

        Re: Very thin?

        Thinner conventional connections means higher resistance and more losses turned into heat, or indeed too slow or too much voltage drop (noise margin).

        1. Anonymous Coward
          Anonymous Coward

          Re: Very thin?

          Not if the current drops with it. Consider, how are larger and larger CPUs and GPUs still able to control their thermal profile so they don't melt down?

  3. SkippyBing Silver badge


    No attempt at a 'Give give give me more Moore's law - Wonderstuff graphene trickery' headline...


  4. Anonymous Coward
    Anonymous Coward


    Still pretty impressive, this is one step closer to super-dense 3D neural nets on a single chip.

    I have my suspicions that at some point a trivial graphene dopant modification will get it to superconduct at useful temperatures.

  5. Disk0
    Thumb Up


    isn't it always the answer?

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

Biting the hand that feeds IT © 1998–2019