from the now-waiting-for-microchips-to-grow-on-sharks dept.
Phys.org is reporting that researchers at the University of California, Santa Barbara (UCSB) have successfully integrated billions of light-emitting dots on silicon to create an efficient silicon-based laser. Previously-existing methods involved separately fabricating lasers on expensive wafers, which then had to be bonded onto silicon chips. This introduced bandwidth and latency limitations as compared to microsystems integrated onto a single chip.
In a paper published by Applied Physics Letters, engineers at UCSB showed it was possible to "grow" or deposit successive layers of indium arsenide material directly on silicon wafers to form billions of light-emitting dots — known as "quantum dots." This method of integrating electronic and photonic circuits on a common silicon substrate promises to eliminate wafer bonding, and has application in numerous military and civilian electronics where size,weight, power and packaging/assembly costs are critical.
This breakthrough is a result of work on the DARPA E-PHI (Electronic-Photonic Heterogeneous Integration) program to develop chip-scale electronic-photonic/mixed-signal integrated circuits on a common silicon substrate which began in November, 2011.
The E-PHI program includes efforts led by the following organizations:
- Aurrion, Inc. (Goleta, CA)
- Massachusetts Institute of Technology (Cambridge, MA)
- University of California, Berkeley (Berkeley, CA)
- University of California, San Diego (La Jolla, CA)
"It is anticipated that these E-PHI demonstrator microsystems will provide considerable performance improvement and size reduction versus state-of-the-art technologies," said Josh Conway, DARPA program manager for
E-PHI. "Not only can lasers be easily integrated onto silicon, but other components can as well, paving the way for advanced photonic integrated circuits with far more functionality than can be achieved today."
(Score: 0) by Anonymous Coward on Friday September 12 2014, @05:43AM
bag of sharky chips with laser dip
(Score: 1) by q.kontinuum on Friday September 12 2014, @06:39AM
There is still the bonding layer between the chip and the sharks head. Why wouldn't the project directly aim for growing the lasers on the sharks head?
Registered IRC nick on chat.soylentnews.org: qkontinuum
(Score: 1) by c0lo on Friday September 12 2014, @06:54AM
They started with a chip and grew a laser on the chip. The next logical step is to grow a shark (head) on the laser (instead of a laser on a shark head - please pay more attention next time).
https://www.youtube.com/watch?v=aoFiw2jMy-0
(Score: 2) by q.kontinuum on Friday September 12 2014, @07:27AM
You got it all wrong. Previously they constructed the laser, then mounted it to the chip, then mounted the construct to the sharks head. Now they are moving the laser down the stack, constructing the laser already one layer closer to the sharks skin.
If we follow your logic, the chip is the center, and they would start growing a shark head on the other side of the chip. Hmm. Intriguing.
Registered IRC nick on chat.soylentnews.org: qkontinuum
(Score: 2) by GreatAuntAnesthesia on Friday September 12 2014, @10:57AM
So, I understand that in principle you could create logic machines that use photons rather than electrons, and that this development is a step in that direction.
However, what's the big advantage of photons over electrons? Is it just a case of moar processing speed? If so, I'm not particularly excited - I already have more processing power than I know what to do with. Or are there other benefits that I'm missing? Increased bandwidth? Reduced power consumption? Built-on Quantum Crypto? Shiny?
Also, can these lights-on-a-chip be used for anything other than photonic circuitry? Might they eventually replace LEDs?
(Score: 2) by Taibhsear on Friday September 12 2014, @12:59PM
Think less for laptops, more for supercomputers.
(Score: 3, Informative) by Taibhsear on Friday September 12 2014, @01:06PM
More information from wikipedia [wikipedia.org] for anyone curious:
"Photonic integrated circuits can allow optical systems to be made more compact and higher performance than with discrete optical components. They also offer the possibility of integration with electronic circuits to provide increased functionality.[5]
One challenge to achieving this level of integration is the size discrepancy between electronic and photonic components.[6] The emerging field of nanoplasmonics is focused on creating ultracompact components for realizing truly nanoscale photonic devices to match their electronic counterparts.
...
Photonic integrated circuits should also be immune to the hazards of functionality losses associated with electromagnetic pulse (EMP), though may not be immune to high neutron flux."
(Score: 3, Interesting) by q.kontinuum on Friday September 12 2014, @01:24PM
If you want to develop SW in a bigger team in agile mode and still want to maintain a stable mainline, you want to evaluate each commit as fast as you can, as in compiling the whole product, run all unit tests and as many functional tests as possible. If you really manage to do that within a reasonable time, you can think about a) redefine "reasonable time" (developer will be happy if his commit is accepted/rejected within 5 seconds rather then 10 minutes) or b) throwing more tests at the software (like integration tests etc.).
Registered IRC nick on chat.soylentnews.org: qkontinuum
(Score: 2) by hoochiecoochieman on Friday September 12 2014, @02:24PM
You, my friend, are a very, very rare phenomenon.
(Score: 2) by GreatAuntAnesthesia on Friday September 12 2014, @04:13PM
Thanks. I'm going to have that tune in my head all weekend now...
http://www.youtube.com/watch?v=g4l1k4XLvo0&t=0m58s [youtube.com]
(Score: 2) by HiThere on Friday September 12 2014, @06:23PM
He's not that rare, but perhaps he's rare among programmers. I could certainly use more speed, not at the moment, but as soon as I get to the next stage of my project.
Javascript is what you use to allow unknown third parties to run software you have no idea about on your computer.