Plasmonics could boost computer capacity by a million

Plasmonics could boost computer capacity by a million

Summary: A research team at Penn State has developed a plasmonic switch that may help pave the way for the next generation of super-fast computers."If plasmonics are realized, the future will have circuits as small as the current electronic ones with a capacity a million times better," said Tony Jun Huang, James Henderson assistant professor of Engineering Science and Mechanics.

SHARE:
TOPICS: Hardware, CXO
5

A research team at Penn State has developed a plasmonic switch that may help pave the way for the next generation of super-fast computers.

"If plasmonics are realized, the future will have circuits as small as the current electronic ones with a capacity a million times better," said Tony Jun Huang, James Henderson assistant professor of Engineering Science and Mechanics. "Plasmonics combines the speed and capacity of photonic -- light based -- circuits with the small size of electronic circuits." – Science Daily

The researches created the switch from switchable bistable rotaxanes. Rotaxanes are mechanically-interlocked molecules that consist of a dumbbell shape with a ring or rings encircling the shaft and are sometimes called molecular machines. The movement of the rings along the shaft is the basis of the plasmonic switch.

“Computers, in their simplest form, are machines that can say yes or no multiple times to transfer information. The motion of a molecule can serve the same purpose as the on off switch on a light.”

An article from Penn State explains how the research team developed the plasmonic switch:

The researchers attached their molecular machines to gold-coated nanodiscs fabricated on glass. The machines were attached with disulfide functional groups. The dumbbell shaped molecules have two areas of the shaft primed with two different chemicals. The ring is initially drawn to circle at one primed area. When the chemical there is oxidized, the ring is repelled and moves to the other primed area, flipping the switch. The process is reversible, so the ring returns to its original state to switch on again later. When the molecule moves, it changes the surface plasmon resonance in that tiny area of the metal where it is attached. This change in resonance is what would send the signal on the circuit. The plasmonic switch that Huang and his team developed is not yet part of a circuit.

Plasmonics, sometimes referred to as “light on a wire”, is an emerging area that uses metal nanostructures to transmit optical signals through minuscule nanoscale structures. It can be used to bridge the worlds of electronic and optical circuits to transmit electrons and light at the same time using the surface of a circuit. This overcomes the capacity and speed limitations of pure electronic circuits as well as the the large size restriction of pure optical circuits, which send information at the speed of light with bulky reflective interiors.

All-plasmonic chips have yet to be developed, but the switch is a step in the right direction. The technology has other applications such as biosensors, and in 2007, researchers have merged plasmonics with another emerging field called spintronics to give birth to “spinplasmonics”, an approach that can lead to ‘computers with extraordinary capacities.’

Topics: Hardware, CXO

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.

Talkback

5 comments
Log in or register to join the discussion
  • Dear Chris Jablonski

    You are a telented writer and intelligent human.

    Please keep up the good work!!!!!

    Your work gives the next generation insight into current understanding.

    Don't underestimate the importance of your work!
    javajunkie@...
    • next generation insight

      I'll do my best, and thanks for the feedback javajunkie.

      -Chris
      chris jablonski
  • Sounds like a scam.

    This sounds very fishy. "Fishy" meaning "invented, it really doesn't exist" fishy.

    -Unfamiliar words, even to the tech industry. Invented words are a big red flag.

    -Little mention of how exactly the technology works. No mention of being able to perform the basic operations needed by basic computers.

    -Big claims. Even for the fast-moving tech industry this has a "too good to be true" ring to it.

    -Even if it did exist, it sounds too complex to be feasible. Frankly, my money would be on carbon nanotubes before I'd put money into this.
    CobraA1
  • RE: Still doesn't solve the multi-core dilemma.

    If you replace electons with photons but use the same basic chip design you still have the problem of diminishing returns past 8 cores.

    If you are going to build on the nanoscale level, what better to do it with than organic material. Nerve cells.


    Bozzer
    • Diminishing returns?

      "you still have the problem of diminishing returns past 8 cores."

      I'd think that has a bit more with how the algorithms are written and how often you lock things.

      Frankly, I think most multithreaded development is done the wrong way - I see quite a few developers trying things to fine grained and lock things too many times.

      Hello? The whole idea of multiple cores is to be able to do things in parallel! If you're using locks and mutexes all the time, of course the performance is going to suffer, because you're forcing a parallel task to become serial. That's just plain stupid, IHMO.

      The best way to handle parallel tasks is to prevent locking - threads should communicate with each other as little as possible, and whatever communication they do have should be non-locking. I use non-blocking queues and they work great.

      A couple more things that could look like "diminishing returns":

      -The fact that we don't have very many truly parallel applications yet. Yeah, if all you have are a bunch of serial apps, then the returns will be diminishing. It's the applications that can actually use the parallel processing that will shine when you begin to crank the number of cores up.

      -The i7s aren't really 8 core processors. They're 4 core with hyperthreading. Hyperthreading is not going to give you as much of a boost as you'd get with a true 8 core processor. That could easily give the impression of diminishing returns for people who aren't aware of how hyperthreading works.

      "If you are going to build on the nanoscale level, what better to do it with than organic material. Nerve cells."

      Unfortunately, we know almost nothing about nerve cells, and even less about how they form intelligence. Not to mention neural networks can't be programmed in conventional ways and have a tendency to inherit all of the drawbacks humans have that we're trying to avoid with computers.
      CobraA1