Computers powered by frickin’ laser beams just came a step closer. Light-based, or photonic, computers would theoretically be much faster and smaller than the electronic computers we use today, but researchers have had a hard time putting theory into action. Now, a new study has shown that two laser beams can be harnassed to turn a single molecule into a transistor. However, the specialized conditions necessary for the trick to work mean that computer stores won’t have photonic sections anytime soon.
Conventional computers are based on transistors, which allow one electrode to control the current moving through the device and are combined to form logic gates and processors. The new component achieves the same thing, but for laser beams, not electric currents. A green laser beam is used to control the power of an orange laser beam passing through the device [New Scientist]. In the study, published in Nature, the green beam could make the orange beam either weak or strong, which is analagous to an electronic transistor turning a current on or off.
In the experiment, the researchers suspended tetradecane, a hydrocarbon dye, in an organic liquid. They then froze the suspension to -272 °C using liquid helium – creating a crystalline matrix in which individual dye molecules could be targeted with lasers. When a finely tuned orange laser beam is trained on a dye molecule, it efficiently soaks up most of it up – leaving a much weaker “output” beam to continue beyond the dye [New Scientist]. But when a green laser beam is also focused on the dye molecule, the molecule absorbs the green light instead and releases energy in the form of orange light, which effectively boosts the orange output beam.
Such a system could theoretically be used in an all-optical circuit, researchers say, although the cost of keeping the materials so cold would be quite high. But the system still holds plenty of interest for researchers, who say that the experiments could lead to high-density packaging of nanometre-sized all-optical transistors. Theoretical calculations by the team also show that it should be possible to build complex circuits where many single emitters are coupled to tiny waveguides that carry optical signals on a chip [Physics World].
Related Content:
80beats: This May Sound Strange: Sonic Lasers and Sonic Black Holes
80beats: Can Laser Treatment Rejuvenate the Incandescent Bulb?
80beats: Biocomputer Made of RNA Understands Boolean Logic
80beats: IBM to Build “Thinking” Computers Modeled on the Brain
80beats: To Cool Computer Chips, Tiny Water Pipes
Image 1: Martin Pototschnig, Image 2: Robert Lettow
July 2nd, 2009 at 10:40 pm
Very complicated.
July 3rd, 2009 at 2:41 am
Where are all the trolls to decry your use of the word “frickin’” is a scientific milieu?
The most fricken awesome thing about light is the ability to be able to encode information in different frequencies/wavelengths (think of being able to listen to all the different instruments in an orchestra at the same time – it all comes through the same carrier, air molecules, but carries many different signals). So you might be able to multiplex data through this system at the buses using multi-wavelength laser encoding. So not only would the processing be fast, the connections would be smaller and faster! Fiber optic technology already has this ability (or so I’ve heard in the past), but it may not work at such small scales, or some weird quantum effects may be exhibited. Regardless, I look forward to the future of computing. Especially if it’s at -272 degrees, because my current computer is 100+ and where I live that’s also the daily outside temperature in the summer. Not fun!
July 3rd, 2009 at 10:10 am
It would seem that photo devices would be more susceptible to the Heisenberg uncertainty principle since the momentum is so much smaller. So it would seem anything using a laser can’t be much smaller than 4000 Angstroms; whereas, a transistor could be only several atoms across.
July 3rd, 2009 at 1:55 pm
If the computer is subject to the heisenburg uncertainty principle, then sci-fi logic would dictate that the computer would develop sentience. real logic would dictate it wouldn’t work very well.
July 3rd, 2009 at 5:33 pm
Hasn’t the HUP been disproved?
Interesting thought about combining light frequencies and amplitudes to add computational dimension to the process…
Depending upon the degree of differentiation, a single transistor could have nearly limitless simultaneous “switches” thrown.
Cool!!
July 5th, 2009 at 8:22 am
If cooling is to expensive cant they (when they’ve made one) put one of these photonic supercomputers in space our free freezer in the sky?
July 5th, 2009 at 10:11 am
@#2: blimey, 100°C or more outside where you live? Now that is frickin’ boiling…
March 8th, 2010 at 4:36 pm
@#7: I think he was referring to Fahrenheit, unless he lives somewhere near the lower crust. It’s hot, here, in Texas, but it isn’t quite boiling.