Follow Slashdot blog updates by subscribing to our blog RSS feed

 


Forgot your password?
Close
typodupeerror
×
Science Technology

Optical Waveguides in Photonic Crystals 55

Posted by michael from the wave-of-the-future dept.
KeelSpawn sent in a short article talking about creating the equivalent of etched silicon for light, using a method intended to be cheap enough for commercial applications.
This discussion has been archived. No new comments can be posted.

Optical Waveguides in Photonic Crystals More

Optical Waveguides in Photonic Crystals

Comments Filter:

  • costs $$$ (Score:1, Troll)

    by ceejayoz ( 567949 )
    "It's a wonderful technique--if you don't care what it costs," Braun says.
    *sigh*

  • Other Technology (Score:2, Informative)

    by Anonymous Coward
    How 'about this [opticalcrosslinks.com] instead!

    • Re:Other Technology (Score:2, Informative)

      by Pooua ( 265915 )
      How 'about this [opticalcrosslinks.com] instead!

      The site to which you linked refers to a completely different subject, that of interconnecting between discrete devices. The subject of the original article is interconnection within a single, monolithic chip. One of the points of the original article is that the new technology can also make tighter turns than regular optical fibers.

  • Entertainment applications? (Score:3, Interesting)

    by ipmcc ( 466386 ) on Tuesday April 30, 2002 @07:00AM (#3434782) Homepage Journal
    Although this is obviously aimed at more "business-productive" markets, I'd be interested to see how (or if) this technology affects the entertainment industry. Just yesterday, on the star wars topic, we saw lots of good banter about DLP, which is made up of millions of minature mirrors. I wonder if technology like this could take the mechanics out of something like DLP. Or perhaps, on a further refinement, this technology could supercede the entire concept of things like galvanometers in things like laser shows.

    The major obstacle here seems to be cost, but what if making the waveguides so small wasnt the challenge?
    • Sad I missed that on the Star Wars thread, as I have a DLP projector for my home theater system and therefore some hands-on experience from them. Anyway, what can be said about them is;

      - less color saturation than LCD projection (colors are not as vivid)
      - no burn-in (as opposed to LCD)
      - better longevity of colors (no fade over time)
      - MUCH better brightness, in fact, black becomes dark-grayish (this is a problem)
      - bulbs cost you an arm and a leg
      - less need for cooling => less noise
      - crispness is so good you have to deliberately DE-focus to get a good movie viewing experience

      Everything of course from my own personal experience with them. I could recommend a DLP projector to anyone who wants to set up a home system.
    • Although this is obviously aimed at more "business-productive" markets

      I think that the primary long-term use for these techniques will be scientific applications in quantum computing. Being able to manufacture small and cheap optical computing components will make the miniaturization of quantum computers practical for smaller scale research institutions.
    • I wonder if technology like this could take the mechanics out of something like DLP. Or perhaps, on a further refinement, this technology could supercede the entire concept of things like galvanometers in things like laser shows.

      DLP is a completely different application, on a vastly different physical scale of application. Whereas, in DLP, the object is to project a beam of light onto a screen several meters distant for viewing by people, the integrated light guides are guiding light through paths that are less than a millimeter long, on static routes. There really isn't any interchangeability between these applications.

  • Wave guides that small are useless, you need big wave guides to improve the experiance of surfing the net.

  • Trek (Score:2, Insightful)

    by lute3 ( 72400 )
    Isolinear chips are on the way, baby!

    I thought they had something of a chance about ten to fifteen years ago when I saw an episode of Beyond 2000.. They discussed a method of changing the color of a transparent polymer at the molecular level via laser--in three dimensions. This soooo would be an improvement over CDs. And what's with the spinning!? [slashdot.org] Solid state needs to be the ultimate goal.

    Looks like it's finally three years down the road--according to the article, anyway.. I saw it will be at least five years. Manufacturers are just now getting to the point of making DVD-writing standardized and afforadable.. Sony asks, "Why on Earth would we want something new?"

    • Re:Trek (Score:4, Insightful)

      by Psion ( 2244 ) on Tuesday April 30, 2002 @07:37AM (#3434866)
      HA! I remember an article in Byte magazine from ten or so years ago detailing a holographic terabyte storage medium on a piece of glass the size and shape of a microscope slide -- hows that for your isolinear chip? The article said such devices could be available for commercial use within one to three years.

      One must approach these kinds of announcements with a degree of skepticism. Sometimes they are little more than fishing expeditions intended to drum up a little shareholder interest. Sometimes, they are completely legitimate, but other market pressures prevent the technologies from coming out in anything close to the stated time frame.

      Not that I disagree in any way with your solid state goal! I'm with you 99.9997% on that one!

    • Re:Trek (Score:2, Informative)

      by tsa ( 15680 )
      For a good photonic crystal, with or without incorporated waveguide, you need three things:
      1) Two substances with an as high as possible difference in refractive index. Normally air and a semiconductor are used.
      2) The substances that are used must not absorb light, so semiconductors with a high bandgap are necessary (depending on wavelength of courase)
      3) A VERY regular crystal structure. This is very hard to achieve. Most research groups in the world use the trick with the spheres as shown in the article. The problem is that these spheres form fcc lattices and what you really want is a diamond lattice, which can not be obtained as easily.
      So the five years you mention may be a little optimistic!

  • Substitution? (Score:3, Funny)

    by grondak ( 80002 ) on Tuesday April 30, 2002 @07:12AM (#3434806) Homepage
    I just hope the ad guys don't make a mistake and try to substitute photonic crystals for my coffee!

  • The equivalent, or the same as (Score:3, Insightful)

    by gewalker ( 57809 ) <Gary DOT Walker AT AstraDigital DOT com> on Tuesday April 30, 2002 @07:15AM (#3434813)
    The article does say that the current process is based on a laser etching in a polymer, but Paul Braun also suggests that the ultimate goal of usefulness will would probably be made of a material "such as silicon" that transmit light more reliably.

    I fail to see a huge advantage in a photonics circuit based on this technology. Braun has perhaps developed a new method that could replace the complex multistep photochemical etching process of todays microprocessors. But it would appear to be harder to scale for production if the laser has to draw the circuit (or the inverse of the circuit) on the chip. Its like the difference between stamping a CD & burning a CDR. Stamping scales for production, and burning one at a time does not. Could be a real innovation for small-run custom circuits, but that does not seem to be where the money is.
    • Could be a real innovation for small-run custom circuits..

      Does that mean that this could be a step toward Desktop Fab ? [slashdot.org] The part of the article about growing the crystals in a particular fashion sounds hard to automate cheaply.

      • Actually not.

        As stated in the article, the silica-spheres are self organizing. You only need to give them time (and a CLEAN! environment) and then they will produce perfect crystals (although a small degree of disorder would always be pressent due to thermodynamical reasons).


        Yours Yazeran


        Plan: To go to Mars one day with a hammer.

    • No, in this case the laser process is cheaper and more reliable. One reason for this, as stated in the article, is that there is only one pass of the laser to make the pathways. The difference is that the current "complex multistep photochemical etching process" is just that complex and multistep.

  • The article really didn't mention any specific applications of the technology. I assume this would be aimed toward optical processors, but does anyone know any links to more information on that kind of application?
    • http://www.ece.ucsb.edu/MOST/research.html#optical

      looks like a pretty likely source of further info but, as it's entirely comprised of large graphic images of text (rather than just the text like any real webmaster would have known to use) I reckon it must all be too secret for us.

      Perhaps Bill Gates hasn't approved the release yet.
    • Probably the biggest application for this technology will be optoelectronic integrated circuits. One of the things that makes optical communications so expense and limits there speed is the need to convert from a light signal to an electronic signal and vice versa. Not only are electronic cictuits slower than there optical counter parts, all the separate modules are expensive. This technology would reduce the need for conversion, at least part of the time, by allowing one to make optical circuits instead. Once that can be done, integrating the wave guides, pump lasers, and amplifiers would be not too far down the road. This could make fiber to the home a reality.

  • Vaporware (Score:2, Interesting)

    by Iberian ( 533067 )
    Unless an invention comes from IBM I just assume it is vaporware or that it won't be invented until some company with a lot of capital comes in and buys it. Now if MS buys the technology to make 10+ ghz chips and then bundles them together with their software then we are in trouble. With a proprietary CPU 5 times faster even XP will run faster then the leanest self compiled Linux kernel.
    • Now if MS buys the technology to make 10+ ghz chips and then bundles them together with their software then we are in trouble.

      But we all know that Microsoft is the DARK SIDE. They can't use anything connected with the LIGHT.
      :)

  • Interestung Note (Score:1, Interesting)

    by Anonymous Coward
    It should be noted that photonic crystals are not actually as strong as what one might think. By keeping the material in a frozen state, the item becomes increasingly more fragile and diminnishes it's ability to reflect sound waves - HOWEVER, Just the oppostite will occur when heating up to over 50'C - The item will then start to amplify that original sound. Quite amazing isn't it?!

    jESUS the Monkey

  • quiet machines (Score:2, Interesting)

    by blind_abraxas ( 446151 )
    If you can replace the cpu with an optical chip clocking at the same speed or faster, you would avoid issues like heat, because the photons don't encounter things like electrical resistance. So you don't need a fan now. And you probably could seriously boost the clock speed. The better this manufacturing process gets, the closer you get to a machine you don't hear. No fans and flash memory sounds like a wearable computer as strong as a pc. And the price is dropping.
    • photons don't encounter things like electrical resistance

      One may not have to worry about electrical resistance in a photonic circuit, but one must still be concerned with optical absorption. Some of the photons that are absorbed in the chip end up heating the chip.

      (I vaguely recall that some of the other absorbed photons wouldn't necessarily heat the chip, but would, instead, give rise to other products, like phonons, which themselves may or may not heat the chip.)

    • One thing I hope to see also is asyncronous design, where there is no longer a reliance on a clock, each element just tells the ones next to it when it is done, thus everything runs at optimal speed, research is being done in this area, which looks quite promising.
      • every few years comes some researcher and suggests asynchronous design and every time the industry ignores him.

        the reason is simple : verification.

        it is tough enough as it is to verify a synchronous design (exponentially hard, in fact.)
        it takes about least 1/2 of development time. (actually more like 2/3 most of the time.)

        adding to this the complexity of many time scales and asynchronous tasks and you very quickly reach something which cannot be verified.

        moreover, the actual value of asynchrounicity in terms of performance cannot be predicted in advance , so what you're asking for is changing a complete industry to a much riskier method, with no quantitatively defined profit.

        cool idea, impractical.
  • This article is the standard babble one sees from Professors trying to drum up research grants for research that will eventually lead to a very nice way to manufacture something... but ends up never being used because the technology is superceded by something created in the industry at lower cost and better performance. Scanning a laser across a 12 inch wafer will never be cheaper than doing it by lithography.
    • Scanning a laser across a 12 inch wafer will never be cheaper than doing it by lithography.

      The laser doesn't have to scan across the wafer; one may use the same masking process used in lithography. The difference between conventional lithography and the new technique is that the new technique can complete the entire waveguide in one step, simply by exposing the polymer to laser light. The old lithographic technique required a step to build the waveguide and several more steps to build the reflectors, besides any other components (such as lenses) that might be in the waveguide.

  • When I was in high school, my 1984 Science Fair project was on optical computers. My presentation was based on a rectangular cell in which tiny bubbles reflected light along the parameter of the 2-D rectangle. Changes in the dimensions of the cell result in changes in the interference of the light, which represented data.

    Eventually, I figured out that the bubbles don't have to be a discrete substance. It should be possible to use changes in density in the substrate as if they were bubbles. The changes in density would be produced by interference waves, sort of a dynamic hologram. Indeed, my optical computer would ideally have been an analog computer based on dynamic holograms.

  • Significant! (Score:1, Insightful)

    by Anonymous Coward
    Doesn't anyone realize how important this will be for the future of the internet?! (or whatever augments/replaces it)
    Petabit workgroup switches for your LAN...
    Exabit (or beyond)aggregate switches/routers for backbone...

    The implications for computing are even better~
    All optical CPUs with optical interconnects means no more heat/size trade offs for portable devices and no more "jet thrust" air conditioning for the server farms.
    This means you can increase the server density (how many processor blades you can stuff into a 7 foot cabinet)
    (oh, and cheaper Co-Location spaces for broadband startups (like Covad..whoever) because they will only need a fraction of the floor space for their gear!)

    All in all, this IS effieciency!

    Low power consumption (and very little waste heat), nearly unlimited performance....
    Ok, i need to go change my pants now...
    ta ta~

  • This is super (Score:2, Funny)

    by adamy ( 78406 )
    so when does Kal-el file for patent infringement.

    When this is done, will I be able to take a green glowing crystal, throw it at an ice field, and have a huge building grow?

  • Please, I just had an exam on this stuff and it raped my brain! Please have pity of me as I have nightmares of forward and reverse propagating solutions capturing and grilling me on the dielectrig grating of a DFB.

    I need a break,
    Edo

  • a brief intro to photonic crystals (Score:3, Informative)

    by stevenj ( 9583 ) <stevenj@al[ ]mit.edu ['um.' in gap]> on Tuesday April 30, 2002 @01:23PM (#3437219) Homepage

    Since I actually do research [mit.edu] in this area and there is some confusion here, let me give a very brief introduction [mit.edu] to photonic crystals [pbglink.com] (which can be studied using free software [mit.edu]).

    Photonic crystals are periodically-structured optical media that, with the right structure, completely forbid the propagation of light in a certain range of wavelengths (analogous to electronic band gaps). They form a sort of "optical insulator" that you can use to trap, guide, and control light. The work at essentially any wavelength (in contrast to metallic waveguides) provided that you can fabricate a periodic structure with periodicity on the order of half a wavelength, and have a number of potential applications, including:

    • Integrated optics: optimally miniaturized networks of optical devices to offload some analog signal-processing or telecommunications tasks, circumventing e.g. bandwidth limitations of electronic circuits. (Few these days are predicting all-optical computers.)
    • Optical fibers (from 2d [bath.ac.uk] or 1d [omni-guide.com] patterns) that circumvent fundamental loss/nonlinearity/PMD limits of silica fiber, and for other novel applications (e.g. high-power or highly-nonlinear fiber devices).
    • More-efficient LEDs and lasers, both by enhancing the optical density of states and by making the light go where you want it to go.
    • Slow-light devices for time-delays, nonlinear interactions...
    • Super-lenses (that can focus beyond the diffraction limit via negative effective indices of refraction).
    • Super-prisms (very wavelength-sensitive refraction, e.g. for wavelength demultiplexing).
    • ...

    1d photonic crystals (multilayer films) have been known since Rayleigh in 1887 (although there are new twists [mit.edu]) but 2d and 3d crystals weren't conceived until 1987, via a marriage of solid-state physics and electromagnetism.

    The paper Slashdot linked to is considering photonic-crystals made by self-assembly of microspheres into close-packed lattices. A perfect crystal has limited use; you need to make defects to carve devices out of it, and that is what they are doing here. (There are many problems of precision, etcetera, that still need to be overcome for practical integrated devices, I think.)

    Note that one can also make photonic crystals with traditional lithography, but that poses its own set of challenges (especially for full 3d-periodic crystals).

  • I think the real advancement here is the possibility for quantum computing as soon as we learn to harness interference paterns. With this technology, creating pathways for coherent light to be channeled, combined, interfere, give and end result, and then be re-channeled recombined and so on ad infinitum... untill an end result is met. This allows for a computing cycle that is as short as t=C*L where L is the entire path length. The cycle would by necesity be slightly longer than that to allow the result to be recorded, but it need not be much longer. The only thing that is stoping us now is our lack of knowledge of photon interactions, and that is growing quickly. When this is finaly harnessed it will blow everything we could have dreamed of out of the water, then again it might take 100 years to work out the science and build the damn thing... but at the rate of advancement I doubt it.
  • Last Post You WHORES!!!!

    I claim this post for Kathleen Fent.

Slashdot Top Deals

There is no opinion so absurd that some philosopher will not express it. -- Marcus Tullius Cicero, "Ad familiares"

Close