Posted on 05/30/2005 7:08:57 PM PDT by eagle11
by Bill Steele Ithaca NY (SPX) May 20, 2005
Much of our electronics could soon be replaced by photonics, in which beams of light flitting through microscopic channels on a silicon chip replace electrons in wires.
Photonic chips would carry more data, use less power and work smoothly with fiber-optic communications systems. The trick is to get electronics and photonics to talk to each other.
Now Cornell University researchers have taken a major step forward in bridging this communication gap by developing a silicon device that allows an electrical signal to modulate a beam of light on a micrometer scale.
Other electro-optical modulators have been built on silicon, but their size is on the order of millimeters, too large for practical use in integrated circuit chips. (a micrometer, or micron, is one millionth of a meter, or one thousandth of a millimeter.)
Smaller modulators have been made using compound semiconductors such as gallium arsenide, but silicon is preferable for its ability to be integrated with current microelectronics.
The work is described in a paper published in the May 19, 2005, issue of "Nature" by Michal Lipson, Cornell assistant professor of electrical and computer engineering, and her research group.
Their modulator uses a ring resonator - a circular waveguide coupled to a straight waveguide carrying the beam of light to be modulated. Light traveling along the straight waveguide loops many times around the circle before proceeding.
Schematic layout of the ring resonator based modulator. The inset shows a cross-section of the ring. See larger image. Credits: Cornell Nanophotonics Group. The diameter of the circle, an exact multiple of a particular wavelength, determines the wavelength of light permitted to pass.
For the experiments reported in Nature, the ring used was 12 microns in diameter to resonate with laser light at a wavelength of 1,576 nanometers, in the near infrared.
The ring is surrounded by an outer ring of negatively doped silicon, and the region inside the ring is positively doped, making the waveguide itself the intrinsic region of a positive-intrinsic-negative (PIN) diode.
When a voltage is applied across the junction, electrons and holes are injected into the waveguide, changing its refractive index and its resonant frequency so that it no longer passes light at the same wavelength. As a result, turning the voltage on switches the light beam off.
The PIN structure has been used previously to modulate light in silicon using straight waveguides. But because the change in refractive index that can be caused in silicon is quite small, a very long straight waveguide is needed. Since light travels many times around the ring resonator, the small change has a large effect, making it possible to build a very small device.
In tests, the researchers found that the device could completely interrupt the propagation of light with an applied voltage of less than 0.3 volts.
The researchers note in their paper that devices using a PIN configuration have been relatively slow in switching but that the ring resonator configuration also eliminates this problem.
Tests using a pulse-modulated electrical signal produced an output with a very similar waveform to the input at up to 1.5 gigabits per second.
The Nature paper is titled "Micrometer-scale Silicon Electro-Optic Modulator." Co-authors are Cornell graduate students Qianfan Xu and Bradley Schmidt and postdoctoral researcher Sameer Pradhan, now at Intel Corp.
Why cling to any electronics? Why not build photonic systems from scratch, with little or no electronic support?
exactly
They have been working on a photonic computer for at least 20 years. It looks like they are making progress but still a ways away.
Sigh...I guess this means tube type radios really are obsolete.
How large is a photon compared to an electron? How small can the channels practically get?
It's an intermediate step that would allow the new photonic stuff to quickly interface with current (no pun) electronics rather than scrapping everything.
Also, most of the things you want to eventually run are electrical, such as motors, lights, speakers, etc. and they can't (today) be run by the low intensity light that is in the photonic devices.
Large motors and speakers powered by light are quite a stretch for today's technology.
I'd like to place an order for photon torpedos with nanite guidance, please.
Get in line. I've got 25,000 terrorists who want a dozen each.
One, it's a pretty neat idea.
Two, they have been claiming photonics is going to take over from electronics for decades, now. Your PC CPU is going to be electronics for many more decades.
One, it's a pretty neat idea.
Two, they have been claiming photonics is going to take over from electronics for decades, now. Your PC CPU is going to be electronics for many more decades.
Neither one occupies any space. They are "infinitely small" unless space itself is quantized. They are "particles" in the sense that they both have a mathematical definition.
I see. Electrics provide power, photonics produce logic. Got it.
OR:
"Electrons for work, photons for play." -copyright, me
Integrated electronic circuit elements are already significantly smaller than the wavelength of visible light.
The wavelength of an electron depends on its energy, but for energies of the order of atomic binding, 1-10 electron volts, the wavelength is on the order of an atomic size, less than a nanometer, compared to 500 nanometers for visible light.
I'm not sure if this is air-tight reasoning, but to me it implies a big advantage for electronics over photonics.
That's kinda what I was thinking. What is to be gained, except less heat buildup? How fast do electrons travel in Silicon currently?
Storage.
Nobody knows how to store photons.
IIRC, electrons physically travel in semiconductors on the order of a kilometer per second, in electric fields created by typical working voltages. Holes, which are the positively charged quantum counterparts to electrons, propagate somewhat slower.
However, electrons and holes need only travel a few nanometers--partway across a junction or channel--to do their job. The associated electric fields and currents, somewhat paradoxically, travel at an appreciable fraction of the speed of light. The exact fraction depends on the resisitivity and dimensions of the conductor, and the characteristics of the dielectric (insulating) substrate.
(A few years ago, IC makers made the metallurgically difficult transition from aluminum to copper for the interconnect layers, because copper, having higher conductivity, results in faster signal propagation, all other things being equal.)
Interesting. Though I'm confused what you mean by the terms "the" and "voltages".
Well, the vacuum tube is immune to the effects of EMP but transistors and integrated circuits are not. If they get the photonic computer working, then the tube WILL be obsolete.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.