Page 1 of 2

Semiconductor Could Create SuperComputers by stopping light

Posted: Mon Apr 24, 2006 9:40 am
by XMEN Iceman
http://www.physorg.com/news64851319.html
'The speed of light' is a byword for the extremes of rapidity: nothing travels faster than light. But Chris Phillips of Imperial College in London and co-workers have found a new way to apply the brakes to light. As Phillips explains on Friday 21 April at the Institute of Physics Condensed Matter and Materials Physics conference, at the University of Exeter, he and his colleagues have shown that light passing through a sandwich of wafer-thin films of semiconductors can be slowed to less than 1/40th of its speed in empty space. And the researchers think that ultimately their semiconductor sandwiches could bring light to a complete standstill.

This kind of manipulation of the speed of light could be useful in schemes for processing information in the form of light pulses, rather than the electrical currents of conventional silicon-chip electronics. So-called optical information technology is already the standard means of transmitting information over large distances, by sending light down optical fibres; but if signals encoded in light particles (photons) can also be shunted around 'photonic circuits', equivalent to today's microelectronic circuits, that might make information technologies faster and more powerful.

Controlling the speed of light in such circuits could provide a way of synchronizing the signals, and even of storing information in 'frozen photons'. Devices that manipulate photons could also be used to create super-powerful quantum computers, which use the laws of quantum physics to perform calculations much more efficiently than today's supercomputers.

Making 'slow light' has been done before. It was first achieved in exotic materials: ultracold gases of metal atoms, cooled to within a whisper of absolute zero. Subsequently, researchers figured out how to use lasers to 'tune' the light-transmitting properties of solid, crystalline materials such as ruby so that a light beam passing through gets slowed down by its interactions with the atoms of the crystal.

Now Phillips and colleagues have found how to make 'designer' slow-light materials from the kind of semiconductors routinely used by microelectronic engineers. Light generally propagates through a material by bouncing between its atoms: each photon, a packet of oscillating electrical and magnetic fields, interacts with the electrons in the atoms in a way that is described by quantum theory. Exotic 'quantum optical' effects such as slow light can be created by using laser beams to alter the atoms' electronic states and thus to tamper with their interactions with photons, in effect making the photons dally longer with each interaction.

In a slab of semiconducting material such as silicon, the electronic states are too smeared out to permit this kind of fine-tuning. But in very narrow films of semiconductor, just a few nanometres thick, the electronic states are more sharply defined, and they can be adjusted by altering the film's thickness. Such thin slices are called quantum wells, and in effect they act like 'artificial atoms'.

Phillips and colleagues have shown that stacks of quantum wells made from the semiconductors indium gallium arsenide and aluminium indium arsenide have electronic states that can be tailored to manifest quantum-optical phenomena such as slow light. The layered films also display an unusual effect called 'gain without inversion', which enables a light signal to be amplified - the basic requirement for generating laser light - without first having to create the preponderance of high-energy electronic states required in conventional lasers.

Source: Institute of Physics
Discuss....

Posted: Mon Apr 24, 2006 12:28 pm
by TimberWolf
I see a technology revolution coming our way. Also it will require education to become better to help keep up the pace.

Posted: Mon Apr 24, 2006 5:54 pm
by XMEN Gambit
Yup. Dis be good stuff.

Posted: Tue Apr 25, 2006 11:18 am
by Spinning Hat
This is all well and good, but until it either hits the mainstream market from Dell, HP, etc, I doubt we'll see a shift to them, just because they'd be so incredibly expensive. Also, how do you think it would affect the enthusiast market, when there are so may people out there who build and tweak their own machines? This technology seems like a way to remove the user from actually building their own PC in a conventional way, like we do now... I'm excited about, I belive that we should try and find better, faster, cheaper ways of increasing computing power, and reduce heat generation and power coinsumption. It's a great step forward, and I hope they're successful.

Posted: Tue Apr 25, 2006 12:22 pm
by XMEN Iceman
Currently...no pun intended...there are heat and speed limitations based on electricity. Aluminum has been the conductor of choice for years in motherboards, memory, and processors, but it has a limited capability of heat resistance. Then we developed copper in silicon, albeit shielded from the silicon for contamination. But there is a point of where you can only push so much signal over the copper conductor before it heats and melts. Optical is the future of all circuits. Light travels faster than an electron flowing over wire. There is no heat generated. Light can travel farther without requiring some type of signal boost. Using current technology with LED diode lasers, optical switches, etc. we will soon be dealing with fully digital optical systems. I can't wait. Talk about Front Side Bus' running in the gigahertz transfer rate range or even higher. The bottle neck will still be storage transfer rate...this new technology has the answer for it.

Two thumbs up!

Computing at the speed of light.

Posted: Tue Apr 25, 2006 1:41 pm
by DarkKnife
Sooo.... This is Laymens terms as I see it. In order to do computing at the speed of light or with optical drives.... Youve got to drop the speed of light to near zero? That'll be a hell of an advertisement. "We do buisness at the speed of light. Which is now zero."

Posted: Tue Apr 25, 2006 5:43 pm
by XMEN Gambit
Heh. Interesting way to look at it, DK. But try this way:

The basic thing you need for modern computer design is a transistor. Transistors serve a number of purposes, including amplifying and gating signals. By gating I mean basically playing traffic cop - this signal can go here, this one has to wait, etc. They do this with (typically) three inputs - power in, power out, and a "signal" line. You build simple logic gates (AND/OR/NOT/XOR) with a handful of transistors, and modern integrated circuit tech lets us put millions of transistors in a square inch of silicon. Because of the material limitations Iceman mentioned, we run into heat and power issues when we do that. If we didn't, we could build faster and smaller chips. (Let's leave out the lithography issue for the moment...)

We currently have the technology to build an optical computer, but it'd be something like ENIAC and its brethren - those building-size machines that a hand calculator today can outperform. They built those with vacuum tubes. When the transistor was invented it paved the way for modern semiconductor-based integrated circuits.

So we need an optical transistor. The tech in this article, or something like it, allows that "traffic cop" behavior I mentioned above to happen with photons instead of electrons. Won't happen tomorrow, and Dell won't be selling all-optical home PCs for $500 within 10 years, but it's a step in the right direction.

And Hat, there's nothing intrinsic about optics that would put a hobbyist out of business. They'd just be different parts. It really depends on how the technology is developed and released. An odd combination of events allowed the x86 PC to be the interchangable, upgradable behemoth it is today. It could have happened with Apples, the PCjr, or any of the other computers developed in the 80's, but the publicity and use of standards and the wresting of the proprietary BIOS away from... ah, whoever it was, really let the PC take off.

Posted: Tue Apr 25, 2006 11:56 pm
by XMEN Ashaman DTM
You'll see these things first in military, nuclear, and astrophysics uses. There are some required shifts in thinking when doing quantum math, and some shifts in thinking when it comes to assembling the logic gates that will make up an optical computer. Mostly the thinking shifts involve realizing that quantum calculations happen ALL at once, EVERYWHERE in your computing domain. This behaviour solves the single hardest problem in normal computational math: solving the n-body problem quickly. Think of a matrix of hundreds of thousands of rows and columns. In todays fastest methods for solving the equations that such a matrix describes, you are at m*n at the minimum (where m and n represent the number of rows and columns in the matrix). The m*n solution speed is also for extremely trivial problems. Typically to solve such a system of equations, you add a row to another row, swap rows, and so on until you get a solution. It can take a VERY long time.

Quantum math accounts for a huge system all at once by including all possible "states" of the system; which are actually the solutions to the system. When you solve the system of equations in quantum math, you are taking all of the states and doing the operations necessary to solve them all at once.

It's like looking at how to read data from a memory chip: you need the row and column of the location of the data, then you need to get the data. So, you scan down the rows of the array of the memory chip, then scan across to the appropriate column, then you read the data (copy it from there to another location). It takes a certain amount of time to find the column and the row, and then to read the data. Now, with a quantum mathematical process, you'd instantly go straight to the row+column that you need. You would not scan down each row and across each column.

The solution is there, in the system of equations, you just pick the right one by applying the correct operator. That kind of math makes today's hardest problems trivial.


Though you would first see processors based on current layouts of logic gates; an optical version of an electronic circuit, if you will. Then, as people begin to investigate quantum math logic, you will see truly quantum math processors. It's fascinating, and we'll likely see such things in our lifetimes. Though if that happens it makes some of the work that I have been doing as part of a personal research project fairly pointless. ;)

Posted: Wed Apr 26, 2006 8:03 pm
by BlackRider
Wait... if nothing can travel faster than the speed of light.... and we stop light..... do we die? :P

Posted: Thu Apr 27, 2006 6:10 am
by TimberWolf
How do you slow down a photon that has no mass and thus no direct means to manipulate? How does a photon carry its energy? I know it is an Electro-Magnetic wave, but how does that exist without some kind of outside force? How long is the wavelength of a car?

Posted: Thu Apr 27, 2006 6:33 pm
by XMEN Gambit
IIRC, it has been shown that photons do have mass; they are affected by gravity after all. It's just a very very very small number. :)

Posted: Thu Apr 27, 2006 7:07 pm
by TimberWolf
That isn't what they are teaching me in my Physics class. I even asked about light being effected by gravity and they told me that it wasn't. Proof to the contrary is a black hole.

Posted: Thu Apr 27, 2006 9:41 pm
by XMEN Gambit
Ask about "gravitational lensing". That's when light from a distant galaxy passes by a nearer one and we end up seeing multiple distorted images.

http://antwrp.gsfc.nasa.gov/apod/ap050327.html
http://antwrp.gsfc.nasa.gov/apod/ap040807.html

Of course, the other explanation is that gravity doesn't actually attract mass, but it warps space itself and the light just follows the curve. In practical terms, that's much the same thing. :)

Posted: Fri Apr 28, 2006 1:43 am
by XMEN Ashaman DTM
Photons have zero REST MASS.

When they move they have a mass that is associated with their kinetic energy. But you have to remember that because mass and energy are equivalent, the photon can have a rest mass of zero (like many particles) and still be affected by gravity.

Posted: Fri Apr 28, 2006 1:50 am
by XMEN Ashaman DTM
And technically, when a photon is at rest, you aren't slowing down the information that said photon encodes. You are just reducing the wave speed to less than normal. Wave speed can be anything from zero to superluminal. You see it all the time in the ionosphere when electromagnetic waves are propagated through it.


All photons have zero rest mass. It's likely that gravitons have zero rest mass (but a spin of 2, if I recall correctly), and that neutrinos have a very slight rest mass. The hard part with measuring the rest mass of neutrinos is that they interact very weakly with ordinary matter. The hard part with measuring the rest mass of a graviton is that no one has identified what a graviton is from particle accelerator data. Some experiments in the next decade may prove that gravitons definitely exist, and we can measure them. And those same experiments may allow us to probe higher dimensions, if string theory is correct.