williams@kirk.dec.com (John Williams DTN 223-2163) (09/12/86)
GaAs is 10-100 times faster than Silicon. Part of the problem with manufacturing it is that the covalent bonds between atoms is much weaker.( part of the reason it's so much faster ) As you put it through implantation, the crystal structure is broken. When you go to anneal it, there is a greater amount of diffusion. John.
hmtriese@water.UUCP (Marc Riese) (09/13/86)
>> From: jeff@gatech.CSNET (Jeff Lee) >> Subject: Re: VERY LARGE main memories >> Message-ID: <4496@gatech.CSNET> >> >> >Then there's GaAs... So fast you can spend a lot of time converting >> >to a different logic family. I like GaAs. Expensive, though. >> >> I know absolutely nothing about GaAs except that Seymour is planning >> to do his cray-3 in it. What are the speeds and costs of some "typical" >> GaAs chips? What sort of power do they dissipate? What is the >> difficulty in processing GaAs as opposed to silicon? Also, is anybody >> doing anything with InP (Indium Phosphide) yet? > > From: williams@kirk.dec.com (John Williams DTN 223-2163) > Subject: GaAs > Message-ID: <5300@decwrl.DEC.COM> > > GaAs is 10-100 times faster than Silicon. Part of the problem with > manufacturing it is that the covalent bonds between atoms is much > weaker.( part of the reason it's so much faster ) As you put it through > implantation, the crystal structure is broken. When you go to anneal it, > there is a greater amount of diffusion. There was an article about GaAs in IEEE Spectrum. I believe it was the December, '84 issue, or some issue close to then. -- USENET: {ihnp4,allegra,decvax,utzoo,utcsri}!watmath!water!hmtriese CSNET : hmtriese%water@waterloo.csnet ARPA : hmtriese%water%waterloo.csnet@csnet-relay.arpa
mangler@cit-vax.Caltech.Edu (System Mangler) (09/22/86)
In article <5300@decwrl.DEC.COM>, williams@kirk.dec.com (John Williams) writes: > GaAs is 10-100 times faster than Silicon. Most GaAs and silicon devices work by field effect (MESFET, JFET, MOSFET) wherein a charge on the gate induces an equal & opposite charge in the channel, and this induced charge is what carries the output current. Switching speed is determined by how fast you can move this induced charge, i.e. how much output current you can get for a given charge applied on the input. To move the charge faster, you can use a higher-mobility material, such as GaAs, whose (electron) mobility is 5 to 6 times that of silicon. You can also apply a stronger electric field (usually the same voltage over a shorter distance). But the charges will only move just so fast (velocity saturation). Most modern devices run at that limit. This limiting velocity is approximately 50 km/s for Ge, 100 for Si, 220 for GaAs, and 250 for InP. Mobility is a measure of how far the electrons can fly between collisions. For very short gates in GaAs or InP, some electrons will make it all the way across without a collision ("ballistic transport") which allows much higher velocities. This is still a research topic. The way GaAs currently gets most of its speed is the short gate length. But silicon is catching up in that respect. Silicon is so successful primarily because it is so easy to make superb silicon dioxide gate insulation. (Any freshman can do it). GaAs has no such natural insulator, so the gate is isolated with a Schottky diode. The diode only insulates in one direction; the inputs have to stay below +0.7 volts or the diode will conduct. This means the inputs either can't swing very far, or they swing negative. Swinging negative requires level-shifting, which eats chip area and power; small swings require tolerances tighter than can be readily manufactured. GaAs is also brittle and poisonous. Reference: "Physics of Semiconductor Devices" (2nd ed), S. M. Sze, Wiley & Sons 1981. Don Speck speck@vlsi.caltech.edu seismo!cit-vax!speck
henry@utzoo.UUCP (Henry Spencer) (09/26/86)
> > GaAs is 10-100 times faster than Silicon. > > [...description of some of GaAs's details and problems...] Perhaps a year ago there was an excellent survey article in Science -- if enough people are interested, I'll dig out a precise reference -- on the fundamental device technologies underlying computers. It discussed the basic requirements for a good computing technology, and then spent some time explaining why silicon is still king and many supposedly superior technologies have fallen by the wayside. It is pessimistic about large-scale replacement of silicon by any of the major competitors visible now, including GaAs. -- Henry Spencer @ U of Toronto Zoology {allegra,ihnp4,decvax,pyramid}!utzoo!henry