Follow Slashdot stories on Twitter

 


Forgot your password?
Close
typodupeerror
×
Technology

Molecule Sized Transistors 132

Posted by CmdrTaco from the little-is-better dept.
IceFoot writes "Bell Labs announced it has created organic transistors with a single-molecule channel length, more than a factor of ten smaller than anything that has been demonstrated even with the most advanced lithography techniques. The really cool part is the transistors assemble themselves: the molecules do the work of finding the electrodes and attaching themselves. Webcast on Wednesday, October 17, 2001 at 3:00 p.m. Eastern time"
This discussion has been archived. No new comments can be posted.

Molecule Sized Transistors More

Molecule Sized Transistors

Comments Filter:

  • Old news (Score:2, Funny)

    by Anonymous Coward
    God created these when he invented viruses.

    Slash Doter
    (Back from hiatus!)
    • True. God did invent this - and it appears that IBM's closing the gap. A giant tower of technology reaching for the heavens with IBM at the peak. It's only a matter of time before programmers around the world are struck down with a curse of multiple languages and different protocols so that they will no longer be able to communi....aw crap.
  • Imagine a Beowulf of these! Mod me down - I deserve it :P

  • Step one accomplished... (Score:4, Insightful)

    by InfinityWpi ( 175421 ) on Wednesday October 17, 2001 @04:25PM (#2443075)
    Now for step 2: Can we do this and make it cost-efficient?

    And Step 3: Can we make new things with this that we couldn't before, or will this just help us shrink down current things?

    And Step 4: How can this make us more money (only the salespeople worry about this one)?

    And don't forget Step 5: How can this get us laid (only the engineers typically worry about that one)?
    • Can we make new things with this that we couldn't before, or will this just help us shrink down current things?

      Assuming these get cheap enough for research purposes in the near future, imagine a few pounds worth of sub-cellular sized nodes arranged in a neural net processing pattern, maybe attached to some video and audio imput.

      I see into minds, you see, and you have no idea how complicated they are. -- Isaac Asimov
    • Step 3: Well, yes and no. Yes we will do more calculations in a smaller space than ever before. We've been doing that for a while now. Wether or not we will do anything truly revolutionary depends on visionaries in the software biz. The SETI@home project is a good example of revolutionary application of computer technology. Some imaging stuff was unthought-of at the time. Other than that, honestly no. Most likely this new technology will be used to kill people when placed in smarter guns. Sad, but true.
    • The only substantial difference between this and the currently used transistors from a materials point of view is the organic component, and that is dirt cheap and easy to make in a lab. If there isn't significant deviations from current designs, ie. "Oh yeah, this requires 95% rhodium backings" then the costs should be comparable.
  • If Moore was a pessimist?
  • I should never have let the PR people make me put this silly molecule model together with superglue ....

  • Mr. Moore (Score:3, Insightful)

    by Mr. Sketch ( 111112 ) <<moc.liamg> <ta> <hcteks.retsim>> on Wednesday October 17, 2001 @04:28PM (#2443096)
    So how do we get smaller than this? Smaller molecules? Atoms? Sub-atomic particles? Photons?

    It just seems like there is a finite limit to how small we can make these things, and it looks like we're approaching that limit, so where to next? Get rid of transistors all together and use a different method? Optical matricies that perform switching functions? Who knows, but I know I'm excited to find out what happens next.
    • Which store multiple states in very small things.

      I'm sure there will always be room for more.

      • Re:Quantum Thingies (Score:2, Interesting)

        by babymac ( 312364 )
        You're not too far off base! Some research has been done that suggests that an infinite amount of data can be stored within the wave of a SINGLE electron! Check it out:

        http://www.eet.com/story/OEG20000831S0019 [eet.com]

      • OK. Warning: Unicode follows:

        The main problem with quantum computers is contianed in the follong equasion:
        xpÖh

        The real problem is why? Why is Planck's constant roughly equal to the product of the imprecisions of the position and velocity. Is this due to quantum "noise" or some other form.

        Of course, the whole universe could be one gigantic quantum computer and we could just be programs.
        • Why? Because the position and momentum operators don't commute, so an eigenstate of one is not an eigenstate of the other. Thus a particle cannot have a well-defined position and momentum at the same time.
    • Mind you, I'm no mathemitician, but the limiting factor that has created the science of binary logic (which shouldn't be confused with pure Boolean logic) is the fact that transistors operate in two states when used as logic gates. On or Off. 1 or 0. Powers of 2 are so important to computer people because that's the way our machines think.

      It's going to be much easier to build a decimal or hexidecial adder/math device with optical components since light can be controlled more precisely than electricity. This is, assuming of course, that the process can be carried out on a truly optical chip. Currently the only benefit to optical technology when used inside a logic system is speed of transmission.
      • Electricity can always be split up into individual electrons, but light sometimes acts as a wave, and is thus harder to manipulate in small increments.

        And kudos on realizing the whole 'powers of 2' thing. That had us all stumped for a while.
        • Electricity can always be split up into individual electrons, but light sometimes acts as a wave, and is thus harder to manipulate in small increments.

          Uh, not quite. Single electrons have half-integer spin and thus obey the pauli exclusion principle [ucl.ac.uk], while photons (spin zero) obey bose-einstein statistics [ucl.ac.uk]. This makes electrons "avoid" each other while photons "congrigate" (condense). But both may be viewed as either waves or as particles, and the WP duality per se has no effect on their behaviour.

          -- MarkusQ

        • Even a single electron will act like a wave in the dual-slit experiment. Individual or not has nothing to do with it.

          For a good explanation of this phenomena, I sugest feynman's QED.

          Tim
    • One-molecule logic gates. Then, one-molecule FPGA's. Then one-molecule processors. Then one-molecule computers. Then one-molecule beowulf clusters. Then we will be like unto gods.
      • One-molecule logic gates. Then, one-molecule FPGA's. Then one-molecule processors. Then one-molecule computers. Then one-molecule beowulf clusters. Then we will be like unto gods.

        To paraphrase Harlan Ellison, I have no mod points, and I must laugh.

        -- MarkusQ

    • I'd guess that we would store the bulk of the processor hardware in a pocket dimension (a la a tesseract [google.com], or a superstring dimension [att.net]). Essentially the only thing you would need in the Real world would be the interface, and whatever hardware is needed to provide the dimensional interface.

  • Liked the quote (Score:2, Insightful)

    by IdIoTt ( 130358 )
    where it said that this discovery would be a historical "bookend" in the transistor world.
    I wonder if this will prove to be true, or if, like so many feats of miniaturization, be proven incorrect further down the line.

    And God said, "Let there be Windows"
    And behold, the earth was without form and void,
    save a giant blue glow upon the heavens.
    • Actually, you could probably make a transistor out of a single molecule but it's not clear how you would connect it to anything. The nice thing about this technique is not so much that single molecules are involved, but rather that this is a cheap way to get a very uniform, very thin layer of insulator.

      Smaller, thinner layers of insulator are possible maybe, but 1/1000000000 meter is about as small as layers made out of atoms and molecules are going to get. In other words, one molecule thick is as good as chemistry can do even in principle. Hence the "bookend" comment, probably.
    • You might be able to make sub-atomic switching devices, but they wouldn't properly be called "transistors." Then again, these might be stretching the term a bit. At least they work on electrons, though.

  • And what are the specs? (Score:5, Informative)

    by wowbagger ( 69688 ) on Wednesday October 17, 2001 @04:28PM (#2443099) Homepage Journal
    What is the transconductance? The maximum switching speed? The gain/bandwidth product? In short, where are all the specs on this transistor that a real engineer would need to evaluate it?

    I don't care if you can make a transistor with a gate length of .1 Planck length, if the thing only has gain below 1 Hz it won't be very useful.

    Until Bell releases some more data on how this device can perform, don't get too excited....
    • Well actually this tech is years away from being usable for a "real engineer" it was only just demonstrated to work. Before everyone go screaming for "engineering data" please, remember, people still have to finish thier research on this stuff before there is any "practical" applications. Sheesh, as if having to fend off all the posts about "could BEC be used as a switching device" Just leave it to /.ies to rate any accomplishment on it's ability to be put into thier pet project...

      No .sig found. Terminating user
    • buy a subscription to nature. I think they only cost a shitload.

      shitload n. more of a thing than i have at any one time

    • But if they are easy to make then imagine highly parallel procesor that still fit your palm-size computer.

      Or big memories (static, no capacitors). Wasn't there article about solid state memories as replacement for harddrives on /. recently?

    • Also, am i missing something, or did they not say how they could get these things to form into the structure necessary for processing? They did say that they could attach themselves to electrodes, but that won't help quake run at 10,000 fps.

      Although, the fact that they are organic is very interesting. Imagine, organic molecules with the processor's structure coded into their DNA, all you do is go down to the local computer parts store and pick up an eyedropper. then you put it's contents into a little dish lined with electrodes on your mother board. Next throw on some Miracle Grow, and wait overnight.
  • I mean, look, what credit did Zhenan Bao get besides holding that damn model?

  • hot digiity (Score:3, Funny)

    by mjackso1 ( 14092 ) <mjackson2317@@@yahoo...com> on Wednesday October 17, 2001 @04:29PM (#2443106)
    Finally, a silver lining behind the huge cloud that is the ownership of Lucent stock...

  • I saw the webcast (Score:5, Interesting)

    by CmdrTroll ( 412504 ) on Wednesday October 17, 2001 @04:29PM (#2443108) Homepage
    My employer [national.com] has a substantial stake in the chip business, and most of us from the Custom ASIC department huddled around a large monitor over lunch today to watch this. Their techniques are fascinating and may someday revolutionize the way we build circuits.

    However, the thing that they do not mention in the announcement is that Bell Labs continues to have problems with the chemical bonds between molecules decaying quickly on these transistors. It is similar to the problems that plague engineers of DNA processors, another cutting-edge-but-hopelessly-broken technology. In fact, despite all of these new achievements that promise to revolutionize the industry, silicon is still king and will be king for many years to come.

    -CT

  • Where are have the good electrodes gone? (Score:4, Funny)

    by 12agnarok ( 226284 ) on Wednesday October 17, 2001 @04:30PM (#2443112)

    The really cool part is the transistors assemble themselves: the molecules do the work of finding the electrodes and attaching themselves

    Great, now even my transistor will find a significant other before I do.

  • control any Assitant Directors of the FBI, Skinner style?
    Yes, this was somewhat offtopic, but it does have to do with nano-tech... so sue me. hehe
  • Just pour some yoghurt over a thin plate of glass and let your processor grow!
  • Webcast on Wednesday, October 17, 2001 at 3:00 p.m. Eastern time

    [flame]It only got posted on the front page at 4:20 Eastern. Awesome timing as usual.[/flame]

  • Reliability (Score:4, Insightful)

    by BlackGriffen ( 521856 ) on Wednesday October 17, 2001 @04:35PM (#2443158)
    With single molecule transistors, wouldn't there be reliability concerns? After all, the uncertainty principle could wreak havoc on a circuit that is too small. Both information processing and durability would be hampered.

    BlackGriffen

    • uncertainty principle (Score:2, Informative)

      by daghlian ( 113201 )
      Do not worry about Heisenberg's uncertainty principle.

      Self-assembling layers that are one-molecule thick are really common and very stable. Examples include cell walls (I know where several trillion are), soap bubbles, etc.

      Thermal fluctuations have many orders of magnitude more energy (1/40 kT, where k is Boltzmann's constant) than the quantum mechanical fluctuations associated with the uncertainty principle. Since room temperature doesn't make these things fall apart we can immediately stop worrying about quantum fluctuations. Also, 1 nm is big enough that quantum tunneling of electrons isn't a problem, especially at 5 V (or whatever they use in chips). The scanning tunneling microscope uses gaps that are ten times smaller and voltage differences that are thousands of times larger.

      • I doubt they'll use 5 V. 5 mV is more likely.

        I expect these things will eventually be useful as implants. Think of it as a VERY high bandwidth link to a computer :-)

        I was reading the other day about some other things that "aren't quite ready for prime time" but should fit right in here. Neurons cultured on layers of silicon. Engines that can turn blood sugar into electricity. Specialized robots that can slide into a cell and self-assemble. (No real time control proposed here. Just do what you're designed for where you are targeted. This one uses chemical targeting, either custom designed [look for a molecule like this: ..], or specialized [i.e., other special methods I don't remember]. Uses include drug dispensing and apoptosis (sp.), i.e., killing off the targeted cells [think cancer].)

        But if you merge those ideas with a computer built from molecular circuits... Live Action 3D with SmelloVision isn't even in the ballpark.
        • 5 mV? you have to be out of your mind. That implies a 5meV gap for the electrons. That's the thermal energy of particles around 60 degrees kelvin (cooler than liquid nitrogen). Getting an electronic state that small is near impossible. Getting a vibrational state that small is near impossible. No, gaps of closer to 1-5 V are much, much, more likely.

    • Re:Reliability (Score:2, Informative)

      by Fastball ( 91927 )
      Defect tolerance is central to the science of nanotechnology and addresses the very concern you raised. Essentially, when you're working with stuff this small, you assume that some things are going to be defective. The real juicy idea is writing software for these kinds of systems that assumes a certain amount of defects and works around them. Try writing a garbage collector for that!
  • It biodegrades within 6 months forcing an upgrade :D

  • Organic transistors (Score:3, Funny)

    by sulli ( 195030 ) on Wednesday October 17, 2001 @04:36PM (#2443169) Journal
    So now there will be an "Organic" aisle at Fry's, with pretty green labels and much higher prices?
  • Ok, I can definately see how this is cool, both in terms of miniaturization and the number of transistors that can fit into large spaces (Such as big-assed Pentium IV CPUs.).

    What I want to know is; how does power consumption scale in regards to transistor miniaturization? If I can make a chip require ten times less space, how much less power will it need? If size and power scale ~equally, how does this mean that Dragonball CPUs will soon be crammed into watches?
    • does this mean that Dragonball CPUs will soon be crammed into watches?

      Nahhh ... gotta get the color camera out of the way first. :)

      On a serious note, would static be even more dangerous of shorting out the circuits?

    • What I want to know is; how does power consumption scale in regards to transistor miniaturization?

      it depends, if they're running at higher frequencies the power usage will increase. Plus with smaller transistors they usually pack more on a circuit, so more power is required.

      However, the power required to switch a transistor is proportional to the *square* of the voltage, so if you can lower the voltage required you've got a big win. And with smaller feature size (miniaturization) they might lower the operating voltage a bit. So the answer is....it depends.

  • Molecular Transistors have already been done (Score:5, Informative)

    by caesar-auf-nihil ( 513828 ) on Wednesday October 17, 2001 @04:39PM (#2443188)
    While the work done at Bell labs does indeed look unique, this experiment and breakthrough has technically already been done by Prof. James Tour (at Rice University) and Prof. Mark Reid of Yale who, in a very high-tech experiment, showed that a single molecule can conduct. It was similar to the structure shown in the Bell labs work, except it was one benzene rather than two. Tour and Reid also used self-assembly to get the molecules to line up to check conductance. The work was published in Science in late 1999.
    Further, Tour and his group have synthesized molecular transistors (he calls them "Moleisters") about a year and a half ago. Unfortunately, I can't bring up his web pages to find the reference to the papers.

  • Shape of turn-on curve... (Score:3, Interesting)

    by dpilot ( 134227 ) on Wednesday October 17, 2001 @04:41PM (#2443207) Homepage Journal
    I wonder what the shape of the turn-on curve is like?

    Modern short-channel MOSFETs are an ungainly compromise between being on and off. We can manipulate the threshold voltage, but so far we can't change the shape of the turn-on (essentially the gain) curve very well. The practical upshot is that modern transistors are perched somewhere between leaking too much DC current and not being strong enough to drive signals with the strength we need.

    At the moment, there is no such thing as low-power, high-performance deep submicron logic. It's the nature of the transistor, not the clock speed.
  • soon... molecule sized phones.

  • We've got good news and bad news... (Score:4, Funny)

    by hoggoth ( 414195 ) on Wednesday October 17, 2001 @05:11PM (#2443352) Journal
    The good news:
    Bell Labs scientists Zhenan Bao and Hendrik Schon have fabricated molecular-scale organic transistors.

    The bad news:
    As you can see in the picture [bell-labs.com], they are REALLY BIG molecules!

  • OK, this sounds good but I think that there is one problem: with my luck my chip would be defective. And then I wouldn't be able to return it because it had a "genetic defect" and I'm not allowed to "discriminate" on that basis. Of course, with organic chips the term "virus" will take on a whole new meaning.
  • To a surgical center near you.

    Molecule sized cell phones implanted while you wait!
    GPS tracking implanted while you wait!
    PS9 seems to be closer than you think!

    But really. This is a great advance. It could lead to pacemakers or other medical equipment less likely to be rejected by the body and also last longer.Very good news indeed.

    Goran
  • I'm in love! [bell-labs.com] .. at least her invention will help me run my 'how the hell do lush/pothead programmer/musicians meet hot molecular physicists' algorithm. (Yeah, it sucks, UD turned down my application to turn the problem into a UD Job ... )

  • Vague and Over-Hyped (Score:3, Insightful)

    by DumbSwede ( 521261 ) <slashdotbin@hotmail.com> on Wednesday October 17, 2001 @06:41PM (#2443900) Homepage Journal
    Hopefully one day, one of these things will have the promised impact on computing, but like many promised breakthroughs, this report leaves many details vague.

    molecular-scale transistors that rival conventional silicon transistors in performance
    And rival here means what? Slower, but will become faster; Slower, but less power? Aren't we aiming for superior in all facets, not just size?

    factor of ten smaller...
    A Million Times Smaller Than a Grain of Sand
    In what dimensions? Width, Area, Volume? What are the actual dimensions, and not just the molecule switch, but the whole gate arrangement?

    A good switch perhaps, but it looks as though it is still hooked up to conventional circuitry, unless I'm reading there diagram incorrectly. This isn't so much nano-technology, as a nano-coating on existing technology. Self assembly doesn't impress me in a coating.

    Self-assemble the paths in and out of the circuit and then we'll talk.

  • Organic transistors? This could suggest that Bell has been working with an outside company. Can anyone remember "Ch-ch-ch-chia"
  • Hendrik is a cool guy. I remember meeting him a couple of years ago because I'm acquainted with his girlfriend. When I asked him about his surely interesting job at Bell Labs, he told me that it really wasn't anything special:

    "I basically just go through lots of samples of different materials, connect each one to electrodes, and see if it conducts. Nothing too exciting." I guess now I know what understatement means.
  • Really, when are these things ever going to be put to use. IMHO, there are two major issues that will prevent these being used in any kind of circuitboard. First, the issue of crosstalk. At the molecular level, especially with the "self-assemblage," there's no way that major crosstalk could be avoided. Secondly, what happens when one of these things breaks down? Obviously it's not going to be easy to fix, and the whole circuit board would need replacing. Don't get me wrong, this is quite cool, but it's always good to keep one foot on the ground when you're reaching for the sky.
  • Microsoft Labs never makes discoveries like this.
    Where's the "innovation"?

  • Re: (Score:2, Funny)

    by account_deleted ( 4530225 )
    Comment removed based on user account deletion
  • This is truly Geek cool!

  • Could this mean the creation of Neural Network pathways are closer than we think?

  • There's got to be something that comes before this (they always say the PDAs are first): "While the molecular transistor technology is still in its infancy, the scientists expect that the first applications of organic semiconductors will be in flexible flat screens-PDAs, for instance, that could be folded up and put in a pants pocket. Ultrafast, ultracheap, plastic microprocessors would likely follow many years later." See Small Times [smalltimes.com]

Slashdot Top Deals

Say "twenty-three-skiddoo" to logout.

Close