Posted on 04/15/2005 2:06:56 PM PDT by LaserLock
A pair of physicists in the US has built the fastest ever transistor: one that can operate at a frequency of over 600 gigahertz. Developed by Walid Hafez and Milton Feng at the University of Illinois at Urbana-Champaign, the device is made from the semiconductors indium phosphide and indium gallium arsenide (Appl. Phys. Lett. 86 152101). The work demonstrates the feasibility of making transistors that can operate at frequencies of several terahertz, which could be used in ultrafast communications, high-speed computing, medical imaging and sensors.
The new device is a so-called bipolar transistor, which is very different from the more well-known field-effect transistor. In it, electrons are injected from the "emitter" terminal, travel towards the "base" and are then received by the "collector", an arrangement that allows the device to work faster than a field-effect transistor.
Hafez and Feng have previously built a high-frequency bipolar transistor, but this earlier work focused on reducing the time it takes electrons to pass through the device by minimizing the device's vertical thickness. Their new research further increases electron speeds through the device by slightly varying, or "grading", the composition of the semiconductor layers. This, say the researchers, lowers the band gap in selected areas of the transistor and makes it easier for electrons to travel across the device.
The two physicists have shown their transistor can operate at a frequency of 604 gigahertz, a new record. However, according to Hafez, what is more important is that they have developed a technology that could be used to build transistors operating in the terahertz range. "Projections from our earlier high-frequency devices indicated that in order to create a transistor with a cutoff frequency of 1 terahertz, the devices would have to operate above 10,000 degrees C," he says. "By introducing the grading into the layer structure of the device, we have been able to lower the potential operating temperature for a terahertz transistor to within an acceptable range."
Devices operating at terahertz frequencies (the far infrared) could be used in communications applications or as sensors to detect toxic gases. They could also be used for medical imaging, since the radiation is long enough to penetrate skin and image what lies underneath.
The researchers' next step is to show that their devices can be assembled into circuits.
IIRC, RTL was in production in '65, and the development team fot TTL shifted to production in early '66. In '65, I was already working on dilectric-isolated ECL for speed.
AFAIK, I tested the very first ECL gate to break the nanosecond propagation delay "barrier".
For other trivia, at that time, wafers were 3/4" in diameter, and the ECL chips had a single ECL gate -- with a few ancillary conductors -- used to hook up 3X3 or 4X4 arrays using multilevel metalization. And we used "KPR" photoresist as the interlevel dielectric!
I haven't worked much with ECL ... I designed a MECL to PECL translator not to long ago but not much more. Now it's mostly CMOS.
Yer making me feel old here!
2N2222's
I've built those too. STOP IT!!!
Yes.
I am now hoping the IBM powerPC architecture eclipses them both.
Wasn't DTL out there during that time too .. BTW ... I started engineering in earnest in 78
Odd. No Exar...
Ahhhh! KPR. Kodak 747. Shipley 1350J...
IBM works with Mot Somerset on the PowerPC ... It's all LSSD based ... I worked on the project for a short while in Austin .. I have my little Mighty Mouse logo ... thats the Power PC logo ... on my laptop case. I also have a pair of boxer shorts Mot made with little Mighty Mouse logo's all over them
I have my old Exar book at home ...
Cool.
I build the viewer that IBM uses to look at their GL1 data. My buddy who co-founded K2 with me does the software to make their reticle sets. Quite fancy things, those reticles. Our software is used for all their production.
Moto uses our tools too, I used to build the Caeco layout editor Moto used to do all their 68000 series chips through the 68040. Worked with Dennis Armstrong at Moto who passed away a few years back.
We sold our company, K2, to Cadence two years ago. Ver lives on, TI's old verification program, with a development team that has been together for 25 years. We acquired them and the latest and greatest Cadence has, is using it.
Amazing industry, how it recycles ideas.
Yeah. The good ones. And the bad ones...
I guess Verilog is a spin-off of Ver. I used to date a lady in the HR department at Cadence
And that wonderfully noxious, phenol-laden "JP-100" resist stripper (used hot)!
No matter how good the scrubbers were, you could always tell when you were downwind of a Wafer Fab (AKA "Front End" at TI)...
Even many miles -- and 40 years -- away ...I can still smell the stuff!
Hi Nully,
You ever notice how you have to have been around for a turn or two to figure out which ones are the bad ones?
I don't know how to explain to some of the juniors what exactly constitutes a bad idea.
Solomon -- There is nothing new under the sun.
It's amazing how many of us ol codgers of the electronics industry are out here. Remember when IBM was at the end of Monterey Road just past Frontier Village in San Jose ??? (Ford Road)
Hey ... was that the old stinky brown syrup stuff ...
The 747 was amber-to-brown and smelled like xylene.
The 1350J was dark red and smelled a bit like slightly sweet acetone.
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.