Page 1 of 1
Posted: Wed Feb 06, 2002 7:25 pm
by IntelMole
If any of you don't know about resistivity, I'm currently doing it for an A-level.
Resistivity is the amount of resistance per metre of a wire, obviously, the lower the reading, the lower the resistance for any given length...
So, I'm looking through a few reference tables in my Physics notes and I come across this...
Resisitivity (Ohm metres) of semiconductors:
Silicon - 2300
Germanium - 0.6 !!!!!
Okay, so if this is the case, and I'm reading the stuff properly, and I'm not drugged or anything...
Germanium has a far far lower resitivity than silicon, and hence a lower resistance...
So why don't we use it? I'm thinking that because the resistance is lower, chips can be run at higher speeds etc. with lower temps and less cooling...
Two ideas come to mind on this one:
Cost - I don't know the cost of germanium but it might well be more expensive than silicon...
Future idea: This could be something that will be used when the limits of silicon start to be approached.
Ideas people?,
IntelMole
Edit: Spelling mistake of Ohm meters
<font size=-1>[ This Message was edited by: IntelMole on 2002-02-06 18:26 ]</font>
Posted: Wed Feb 06, 2002 8:55 pm
by lenzenm
Germanium...ugh. I may be a bit off here, but from what I recall, germanium is:
1) Fairly rare.
2) Difficult to refine/make pure.
3) Toxic (adds to the problems in #2).
In spite of all this, IBM & Motorola (that I know of, there may be others) are playing with germanium and are very excited by the possibiliies it offers, now they're working on making it economical.
Posted: Wed Feb 06, 2002 9:45 pm
by mesh
Unless I have suddenly been overcome by memory loss, it seems to me that you have got it rather backwards.
Silicon has, as you have noted, a very high resistivity. That is why it is used as the *insulator* that the wires are etched onto. The wires themselves are usually aluminum, although copper interconnects are just now starting to get used.
I seem to remember that gallium-arsenide has been used/suggested as a even better replacement for the aluminum interconnects, so we are getting there.
Posted: Thu Feb 07, 2002 4:23 am
by Bruce
Intelmole, see if you can find resistivity figures for doped silicon - I suspect they'll be quite different.
Bruce
Posted: Thu Feb 07, 2002 7:17 am
by newbie_of_jan0502
Intrinsic (pure) silcone and germanium are poor conductors. This is due partially to the number of valence electrons, the covalent bonding, and the relatively large energy gap. Because of their poor conductivity, intrinsic silicon and germanium are of little use.
Doping is the process of adding impurity atoms to intrisic silicon or germanium to improve the conductivity of the semiconductor. The term impurity is used to describe the doping elements because the silicon or germanium is no longer pure once the doping has occurred. Since a doped simiconductor is no longer pure, it is called an extrinsic semiconductor.
Got this from a text book that I've kept since college. I always new it would come in handy some day.
Posted: Thu Feb 07, 2002 11:12 am
by dolemitecomputers
[Intrinsic (pure) silcone and germanium are poor conductors. This is due partially to the number of valence electrons, the covalent bonding, and the relatively large energy gap. Because of their poor conductivity, intrinsic silicon and germanium are of little use. ]
That would make sense why it is not used in thermal paste like Arctic Silver.
Posted: Thu Feb 07, 2002 11:20 am
by Bruce
I think you're confusing electrical conduction with thermal conduction - the above discussion is about electrical conduction.
Posted: Thu Feb 07, 2002 5:29 pm
by IntelMole
Okay, so I admit, I got it wrong, they use the silicon to separate all the pathways right?
Ahhhhhhhh.... now it makes sense, I thought they used them for a part of the actual transistor or something, which would be incredibly stupid
Oh, and the fact that germanium is rare would make it expensive, which means I was partially right in one of my ideas
Definitly on the back foot here
,
IntelMole
Posted: Thu Feb 07, 2002 5:54 pm
by dolemitecomputers
Ok then why is silicon not used in thermal paste or not good to use in that?
Posted: Thu Feb 07, 2002 6:00 pm
by IntelMole
Hey I just had an idea for a thermostat, which is probably majorly remote and already done
Don't hate me cos I'm slow...
Silicon and germanium are semiconductors, so as the heat being emitted by them increases, so does their resistors, and the next bit, the reasoning, I CAN GET RIGHT, cos I'm gonna copy it straight from the notes...
Nah, dat'd b easy... The resistance decreases because more charge carriers become avaialable, whatever that means
So why not make a thermostat out of it, as the resistance decreases, the current increases, so they could use this in a thermally controlled fan. Or do they do that already?
Posted: Thu Feb 07, 2002 7:25 pm
by newbie_of_jan0502
Transistors (such as a typical pnp) if you tie the base and collector together as temperature rises so does the voltage drop across the transistor. Transistors are used extensively in analytical applications which require temperature compensation.
The thermal properties of silicon are decent kinda between gold and aluminium.
Silicon Thermal Properties
Thermal Conductivity (solid) 1.412 W/cm-K
Thermal Conductivity (liquid) 4.3 W/cm-K
Specific Heat 0.70 J/g-K
Thermal Diffusivity .9 cm**2/s
Melting Point 1683 K
Boiling Point 2628 K
Critical Temperature 5159 K
Density (solid) 2.33 g/cm**3
Density (liquid) 2.53 g/cm**3
Vapor pressure
at 1050C 1e-7 Torr
at 1250C 1e-5 Torr
Molar heat capacity 20.00 J/mol-K
On 2002-02-07 17:00, IntelMole wrote:
Hey I just had an idea for a thermostat, which is probably majorly remote and already done
Don't hate me cos I'm slow...
Silicon and germanium are semiconductors, so as the heat being emitted by them increases, so does their resistors, and the next bit, the reasoning, I CAN GET RIGHT, cos I'm gonna copy it straight from the notes...
Nah, dat'd b easy... The resistance decreases because more charge carriers become avaialable, whatever that means
So why not make a thermostat out of it, as the resistance decreases, the current increases, so they could use this in a thermally controlled fan. Or do they do that already?
Posted: Thu Feb 07, 2002 7:38 pm
by Veritas
I thought they used them for a part of the actual transistor or something, which would be incredibly stupid
Actually silicon is used as part of the transistor. When fabricating a transistor you start with a silicon substrate. From there you apply lithographic masks and begin etching and doping things. I don;t have time to explain it all. But here is a power point presentation about it for you:
http://www.ee.fit.edu/courses/ece4312/f ... n_Flow.ppt<font size=-1>[ This Message was edited by: Veritas on 2002-02-07 18:39 ]</font>
Posted: Fri Feb 08, 2002 1:14 am
by Bruce
Dolemite, there are simply betters materials to use in thermal paste - metal oxides, for example.
Posted: Sun Mar 10, 2002 2:39 am
by HowardDrake
And if I recall, germanium isn't used because its doped with arsenic, major toxic stuff.
Posted: Sun Mar 10, 2002 3:51 am
by Speed
I actually remember a time when germanium semiconductors were more common than silicon. That was back in the discrete transistor days.