Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 


Forgot your password?
Close
typodupeerror
×
Science Technology

Silicon LED 70

Posted by michael from the mommy,-is-my-RAM-supposed-to-glow-like-that? dept.
Ian writes "Scientists at the University of Surrey have developed an LED made entirely from silicon. This is a different approach to optoelectonics which had previously concentrated on nanocrystals. Full report from Nature, also coverage from the BBC, stand back and watch the patents fly (although in this case they are much more deserved)."
This discussion has been archived. No new comments can be posted.

Silicon LED More

Silicon LED

Comments Filter:
  • NAPIER UNI [napierengineering.org.uk] has some info on a pretty funky technology I first came across at an Investor Conference in Edinburgh last Autumn.

    Essentially, this stuff gathers all the available light and throws it back out within a narrower range of wavelengths - making it apparently brighter. Especially when you consider it takes UV and (_I think_)some IR and throws it back in the visible. You can make signs out of it, so goodby neon, and the brighter the sun gets the brighter the sign gets. You can make ANYTHING out of it, and it will glow in daylight, artificial light, moonlight... v.cool.

    Photons - officially the new SEX
  • The Nature 'article' linked appears to just be a summary of the actual paper.

    Either get it in paper form, or wait for NewScientist to cover the story.
  • Ah yes. I remember going to the National Cryptological Museum a few years ago. It's right next to the NSA, of course - they run it, and probably swipe your DNA while you're there...

    Anyway, in their hall of obselete computers was a CM5. Impressive, yes. A little humbling though. I can't imagine what they've got in there, these days.
  • So they can generate the light signal on chip now, which is a big step. But for light to replace electrical current on/around a chip, how that signal is transmitted? As far as I know, there is not any way to create a fiber optic 'wire' on chip.
  • And what if I like my oversized, very heavy, power sucking 19" relic? I really have no desire for the flat pannels I've seen so far, I like my oversized box that sits on my desk.

  • wow, I am suprozed how many here have no clue on LED's or their history. Check out

    http://www.marktechopto.com/HistoryofLEDTech.htm

    for some good history.

  • Hear, Hear!

    I like the fact that I can work on my suntan in a windowless office. If I need some exercise, there's heavy lifting right in front of me.

    Just think if flat panel displays become the norm, there are gonna be a lot of disappointed cats (except for maybe the really skinny ones).

  • You wouldn't believe it, but there is life outside the US [surrey.ac.uk], too! ;-)

  • The 56K modems push up against the limits of the telephone voice network which has a maximum bandwidth designed to carry voice transmissions. There is no way to push more information through the voice telephone system than this bandwidth limitation of the system itself. (ok, there are ways of using multiple lines/circuits, but that is cheating...)

    Well, that's not really true. Todays 56K modems push up against the limits of the old telephone network and even those are theoretical limits. Basically, the phone company could push more speed through the phone line by increasing the maximum allowed voltage. But they don't; because they don't have to.

    Later...
  • Have you heard the theory of rapitivity?

    E = MC Hammer.

    Sirch
  • Basically, the phone company could push more speed through the phone line by increasing the maximum allowed voltage.
    Increasing the allowed voltage would allow some modem designs to get closer to the maximum set by the bandwidth set up for voice transmission - it would only be a small gain. A voice channel in the voice network is less than 60K of information and without changing the design of that voice network, you can't get past that. You cannot use a voice channel at anything greater than the voice channel's capacity - the pipe is only so big.

    This limitation is not a theoretical one, it is a designed one. As I understand things, the system was designed to take the voice noises, encoded them digitally at less than 60K, transmit them through those fancy new digital thingies that the phone companies installed in the 1970's (or earlier? or later? I don't know) and route them to the other phone location. Transfer the data back into analogue and pump it out the recipient's phone.

    Thus, since the sound is encoded at about 60K, there ain't no way possible to take your data and encoded it into a sound to send through your phone such that you can push more than 60K through the equipment the phone company has. Unless, as is done with the DSL stuff, you do not use that equipment with those limitations.

    At least that's the model working in my mind - I could be totally wrong.

  • I guess what QwkHyenA had in mind, but i guess we could detect light 'leaking' from the circuitry, thus running the board in the dark we'd be able to see where the light is leaking...
  • Why does this article and a lot of the others on the front page today have a huge block of italic text?

    It's not needed, ugly, and hard to read.
  • The article states that integrated circuits are all made from silicon. This is not true. A majority of integrated circuits are silicon, but a lot of high speed CPUs are gallium-arsenide.

    And I believe a lot of LEDs are made from gallium arsenide. Maybe a mix of germanium and gallium too.
  • You might find that this link [wordsmyth.net] offers further insight into his post.

  • The parent is the obligatory "This is not news" post to any story about technology.

  • Fraid not d0000d; only the semi-conductor transistors have very minute amounts of germanium and what not. For the most part they are silicon.

    I imagine there will be some capacitance problems if they make say an entire chip out of silicon

  • stand back and watch the patents fly (although in this case they are much more deserved).

    Why are they "much more deserved"? Is it because those researchers work in a field that you don't understand? Why does working on computers mean that our experts don't deserve patents? It seems that computer geeks have a fundamental misunderstanding of their place in the world of science...

  • I thought the main reason for using other materials in optoelectronics was that Si have an indirect bandgap. From what I could se in the article, this isn't meantioned at all.
  • If someone gets going a transparent cube with lots of transparent chips blinking very fast in different colors...

    ...It's going to be Apple's intellectual property.

    iCube, anybody?
  • Hey! Come on! They say this kind of stuff every time! Ok, they have to come up with something new to acchive that "next, impossible" result. But they always do!

    I remember when 56K modems came out and we were told they would virtually rip a hole in the space-time continuum. Suddenly we've got 2mb ASDL sockets in the wall, that can do simultaneous voice & data. What da f????

    In summary, I agree entirely with you. It only looks impossible now. Tonight some bright spark will have a dream about how it can be bettered.

  • It is mentioned in the actual research paper here [nature.com]. They explain the band structure can be modified by introducing an array of dislocations, which create a special kind of strain field in the crystal structure.

    --
  • My impression was that this invention would enable us to integrate fiber-optic connections into the silicon circuits, thus removing some bottlenecks.

    --
  • "Chips are getting smaller and more powerful every year, but ultimately, and probably in the next few years, there will come a point where the complexity of the contacts and the wire doesn't scale, but stays the same"

    Hey! Come on! They say this kind of stuff every time! Ok, they have to come up with something new to acchive that "next, impossible" result. But they always do! It's getting boring, like in all those TV-Show, where you know that the hero is going to survive

    On the other hand I'm happy that they allways find a way to work it out ;-)

  • Read the article carefully...

    Britain's University of Surrey...


    aztek: the ultimate man

  • Most probably you are referring to blue SiC LEDs. These are however not related to the article.
  • Maybe just cut a groove, and fill it in with something transparent?
    ___
  • Ehm, well I must be part of the 1% you mention then. The day I got my first laptop with TFT (must be about 8 years ago, not sure) I always have sworn by TFT. Now that I became more "sedentary", I bought a nice 15" flatscreen for my desktop. Agreed, it costed more than the box itself but I do care about my eyes.

    At work I have one of those nice 19" aquariums to develop on... If I look around me: heck, I see 7 developers here that have such a screen. I saw dozens of flatscreens here in the company, on places where no customer should ever come. Guess they care about our health around here. ;-)

  • Also, think of the implications in debugging which board/circuit went bad?? If they can do this cheap enough, we could have our electronics tell us which section went bad based on what color it's emitting!

  • Now we can all finally build those kewl androids like 'Data' that have those kewl flashy head circuitry implants!

    Just think, you'll be able to tell how your bot is doing based on the different colors emitting from its cranium!

    Light emitting diodes are yummy!

  • But the light-emitting devices needed to do this can't be built into silicon circuits.

    This is talking about older style LEDs. Consider the next paragraph:

    The ideal solution would be to make light emitters from silicon itself, but silicon does not glow efficiently. Various tricks have been tried to squeeze light from silicon.

    And what is the article about?

    A light-emitting device based on silicon promises to end what has become an uncomfortable marriage.

    (perhaps silicon can help your uncomfortable marriage? ;)

    Nothing that I could see in the article about how they overcame "silicon does not glow efficiently," or even if they did. Perhaps the BBC one was different, but the Nature article is a headline with some background information. Nothing really about the headline.

  • There's no such place as Surrey in America.

    Actually, there is... At least three of them, actually. Surrey, ND; Surrey, IN; and Surrey, IL.

    MapQuest [mapquest.com]

    Though, I doubt any of them have Universities.
  • Actually, I think it has more to do with creating a pure optical based laser so fibre doesn't have to be terminated back into electronics at routers (this slows everything down precipitously).

    This is very much "pure research" though, so it could be applied or adapted to anything, including CPU's.
  • This is an interesting breakthrough. Display technology just keeps advancing at such a pace it reminds me of moore's law in the microprocessor world. First it was LEPs and now this.

    It cannot be long before these things will be as throwaway as the daily newspaper.

  • There's no such place as America. All these 4-letter abbreviations (DMCA, MPAA, RIAA etc) are obviously the ravings of a deranged mind.
    --

  • Glass optical fibres can carry much more information than copper wires. So most long-and medium-range data transmission is now conducted in pulses of laser light rather than pulses of electricity.
    It is also rather expensive to attempt to replace copper with fibre cabling

    When the charge carriers that ferry an electrical current around are confined within such narrow channels or boxes, the processes that prevent them from producing photons of light are suppressed. So 'nanostructured' silicon glows.
    Not really an effective way to bundle strands of fibre together in order to make this work. What can end up happening is, once strands start dying out, the entire emission of the light becomes lower and less useable.

    Pardon me for being the troll here, but why would their be patent suits or other on this?
  • now all we have to do is find intelligent Silicon-based light forms.
  • Laptops... Now they'll be even smaller! Great, I can't stand typing on one as it is..
  • The BBC article says that it works with current fabrication plants.
  • They have forced silicon to emit light, a discovery that may make it possible to use light rather than electricity to carry signals around a chip.

    This sounds very similar to fibre optic technology.

    Does anyone know if this new technology would have a reduced power consumption ? It all sounds really promising. Well done University of Surrey (15 miles from where I live).

    Claric
    --

  • Of course, the fact that the Nobel prize is Swedish to begin with must have escaped both of you...
  • It might be interesting to put indicators on your chip based on these devices to show state during chip test. You could then view that with a camera to determine the state of lots of nodes at the same time. Of course I doubt it would be cheaper than boundary scan, but I suspect that there would be some useful applications. As it is, I doubt it would be useful for much more than fibercomm chips.

  • True, but from my understanding of the optical/electronic interface (which is far from complete) generating laser pulses (which is required for data transmission (LED's simply don't cut it) isn't the problem. It's going from photons to electrons that's the slow process.

    If they can do this (p2e) with this technology then they are will on the way to using photons in the data bus, either between chips or on the chip. Unfortunately, from the paper, it appears that this can only be used to generate light, so they have only one half of the equation.

    Also from the picture they published, the doped tracks glow, so it's not even a point source. There seems to be a lot more promise vertical-cavity serface emitting lasers [sciam.com].

  • > They're making pixels out of *transistors* now?

    See TFT: Thin Film Transistor.
  • Comment removed based on user account deletion
  • It's not new that silicon can be made into LED's. At Rochester Institute of Technology (for one), they've been doing it with porous silicon. While it's an expensive and difficult way to get a blue LED, it does work.

    -h-

  • I think Britain has actually won more noble prizes in total than the US
    Trinity College, Cambridge has won more than France. Err... that wasn't really relevant, was it?
  • There's no such place as Surrey in America.

  • I remember when 56K modems came out and we were told they would virtually rip a hole in the space-time continuum. Suddenly we've got 2mb ASDL sockets in the wall, that can do simultaneous voice & data. What da f????

    The 56K modems push up against the limits of the telephone voice network which has a maximum bandwidth designed to carry voice transmissions. There is no way to push more information through the voice telephone system than this bandwidth limitation of the system itself. (ok, there are ways of using multiple lines/circuits, but that is cheating...)

    The various "DL" technologies do not use the telephone voice system in this way, and thus are not limited by this bottleneck. In fact, most of the "DL" systems have enough room to provide the bandwidth necessary for voice in addition to the larger amount provided for data - which is why you can talk on the phone and use your internet connection at the same time with only one line.

  • So what? So there's the podunk place called "eng-bland" or some such -- all the electronics work worth doing is done right here in America: all microchip research is done by Intel, all operating systems are made by Microsoft, and all boxes are designed by Dell.
  • Please tell me you're joking. There are many universities outside of America doing very good research. In fact I think Britain has actually won more noble prizes in total than the US, although someone might like to correct me on this. Anyway, I'm going off-topic.
  • Display technology just keeps advancing at such a pace it reminds me of moore's law in the microprocessor world. First it was LEPs and now this.

    No, if display technology kept up with Moore's law, we would all have sub £100 29" flat-panel monitors on our desks!

    An amazing percentage of computer systems still use extremely old-tech monitors (maybe 99%?) I don't actually know anywhere (outside of dealing rooms or comms rooms) where people order modern-tech monitors rather than some so-big-it's-got-gravity cream-coloured monitor!

    I also apologise for over use of the "-" char today ... I don't know what's come over me.

  • Well ... an how about the rest of the world? Sometimes I really hate all those america-centric ppl.

    http://www.surrey.ac.uk/ [surrey.ac.uk] (notice the .uk?)


  • but it's a well known fact that the best engineers are imported from england.

    For sure!

    Why is it that the British like warm beer?

    Because Lucas makes refrigerators.


    Hmmm...

    British Reliant [reliants.co.uk]. American Reliant [stationwagon.com]. (Yeah, okay, it's actually a Canadian Reliant now living in Scotland.)

    At least North American Reliants have 4 wheels. So there.

  • Do you have a very dumb bimbo girlfriend ?

    Do you find yourself hiding her in the closet when your family comes around ?

    Do you wish she could run Linux ?

    Well now she can thanks to Fnarg

    Introducing the MPU : Mammary Processing Unit. Thanks to silicone-based transistors, your blonde bimbo can now harness the power of Linux using the processing power found inside her breast implants. The benefits are limitless: human-powered text-to-speech engine, self-powered design, intuitive touch-sensitive GUI, and don't forget our revolutionary S.L.U.T. module (Super Lightweight Undressing Terminal). Never again will you want to call your girlfriend a dumb whore once she gets a dual P3-Xeon installed on her ribs. And best of all, the MPU is available in a wide variety of shapes and sizes to fit all your business needs.

    Call 1-888-GOAT-SEX now for more information

    Fnarg is an equal opportunity satire.

    Yes I know Silicon and Silicone aren't the same.. but hey, someone's gotta laugh!
  • Unless they've been giving Noble Prizes in Funkology to Prince or Sir Mix-a-Lot... :)

    They have.

    I myself was recently awarded Harvard's prestigous "Noble Prize in the Musical Arts" for my hit song "Baby Got Back." In my acceptance speech, I spoke on how honored I was to contribute to the degredation of America's culture.

  • Britain has everything America has plus we have and electricity aswell to run our new silicon LEDs with... :->

  • IANAEE, but it seems to me that this won't really help. Reading the Nature article reveals that creating these SLEDs (new acronym? Silicon LED?) requires doping the silicon with boron, then heating it to 1000 deg C.

    The stated purpose of this invention is to ease the integration of optics and silicon-based electronics - ostensibly to allow chip designers to fabricate an LED directly on the same chip, without having to "scab" on a separate LED to talk to the optics.

    Is this boron doping and superheating process really going to be compatible with general chip fabrication procedures? Maybe a real EE can answer that.
    * ~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-~-
    * Split Infinity Music [simusic.com]

  • After reading both the Nature and BBC articles, I'm not sure that they are really talking about opto-electonics, although it might be used for such.

    Instead I think they may be using this to emit the energy loss of silicon circuits in the visible spectrum, rather than IR.

    As we all probably know, today's CPU's run extremely hot, just look at all the comments about the new Apple titanium notebook. Getting rid of this excess as light instead would be an immense benefit. And would mean small faster CPU's without the need for a cyrogenic cooling system.

  • Great news (Score:3)

    by ralmeida ( 106461 ) on Thursday March 08, 2001 @04:29AM (#376895) Homepage
    Perhaps this will enlighten Silicon Valley a little bit?

    --

  • From the BBC radio interview... (Score:3)

    by Andy_R ( 114137 ) on Friday March 09, 2001 @01:23AM (#376896) Homepage Journal
    ...which went into more depth than the article:

    The real benefit of this is that these silicon light emitters can fairly easily be fabbed in existing chip plants wthout requiring the 'start from scratch' of other optical computing tecnologies.

    This technique should (according to the invetor) provide a way of building hybrid practical electronic/optical chips very soon. He particularly mentioned the clock distribution problem that /. had a big discussion on few days back as being one of the first applications he expects to see for this technology.

  • If someone gets going a transparent cube with lots of transparent chips blinking very fast in different colors... well, all I can say is they will get a lot of free coverage in the next major films.

    It's already been done in Blakes 7 [cultv.co.uk] (spectacular BBC TV sci-fi series from the 80s).

    Computer was called Orac [btinternet.com], and had a spectacularly cheesy pseudo-computer-generated voice!

    To quote:
    Orac was described by its creator, Ensor, as being beyond a simple computer but rather being a brain, a genius.

    Sounds just like b1ll Gates describing Win2K ;-)

  • Optical Computing (Score:3)

    by Alien54 ( 180860 ) on Thursday March 08, 2001 @04:38AM (#376898) Journal
    Plainly a step in the direction of optical computing.

    Although details like the reaction time of the optical effect are missing. In a world where many things are measured in nanoseconds, if these things react in millionths, for example, then this will limit the applications.

    Interesting all the same.

  • More info (Score:3)

    by stubob ( 204064 ) on Thursday March 08, 2001 @08:27AM (#376899) Homepage
    I submitted this story too. More information from Yahoo(via Reuters) [yahoo.com]. They mention how it works, something with dislocations, loop-flaws in the silicon. The press release from U-Surrey is here [surrey.ac.uk]. Google also claims to have indexed their paper here [surrey.ac.uk] but accesss is forbidden.

  • Re:this is clearly a hoax... (Score:3)

    by DetritusX ( 319569 ) on Thursday March 08, 2001 @05:04AM (#376900)
    "...I think Britain has actually won more noble prizes in total than the US..."

    I would imagine they would, seeing as the US doesn't have a recognized nobility. Unless they've been giving Noble Prizes in Funkology to Prince or Sir Mix-a-Lot... :)

  • I mean, they really miss the old times when they could represent computers with lots of blinking red lights!

    If someone gets going a transparent cube with lots of transparent chips blinking very fast in different colors... well, all I can say is they will get a lot of free coverage in the next major films.

  • Silicone LEDs? (Score:4)

    by Anonymous Coward on Thursday March 08, 2001 @07:51AM (#376902)
    "Flashing your tits" gets a whole new meaning :)

  • They exist! (Score:4)

    by HarryTuttle ( 69566 ) on Thursday March 08, 2001 @04:33AM (#376903)
    It's long been a rumour, but finally...

    Silicon-based light forms!

  • Re:Let there be light (Score:4)

    by trentfoley ( 226635 ) on Thursday March 08, 2001 @05:45AM (#376904) Homepage Journal
    Perhaps I'm missing something here, but I don't think the point of this research directly relates to transmission of data over fibre. I see it more as an improved interface between the optical and electrical world that has potential to be much more efficient. My reason for thinking this way has to do with the paragraph in the Nature article:

    But the light-emitting devices needed to do this can't be built into silicon circuits. They rely on semiconducting materials that don't sit comfortably on silicon chips, and so they have to be mounted separately in the awkward hybrid technology 'optoelectronics'. The physical separation of light emitters and electronic circuitry is a bottleneck to further miniaturization, and so to greater processing power and speed.
    Nor do I see this applying to flat-panel display technology, or electronic paper for that matter.
    Then again, having been married for over 10 years, I know that I'm almost never right about anything.

  • Re:Indirect bandgap. (Score:5)

    by swm ( 171547 ) <swmcd@world.std.com> on Thursday March 08, 2001 @05:09AM (#376905) Homepage
    That's the whole point of this invention.

    The number of momentum states is (essentially) equal to the number of Si atoms in the crystal. So if you make a crystal with only a few atoms, you only get a few momentum states.

    That may push the bands around so that you get a direct band gap,

    OR

    That may make it possible to get a significant carrier population in the zero-momentum state, even though that isn't the lowest energy state.

Slashdot Top Deals

Top Ten Things Overheard At The ANSI C Draft Committee Meetings: (5) All right, who's the wiseguy who stuck this trigraph stuff in here?

Close