diamond semiconductor processors

[TanisHalfElven]

hi guys. i justy read in popular science that the next semiconductor material would be diamonds and since they can work at higher tempratures than silicon thry can reach speeds of upto 89 GHz while silicon is limited to 10 GHz. is this true ?

and how much will these things cost. what advantages will they offer. will they realy run at 1300 F?

 

[potato28]

Links?

 

[TanisHalfElven]

oops sorry. i sort of assumed everyone would know about it except me.

well searching google.
http://www.geek.com/news/geeknews/2003Aug/gee20030827021485.htm
http://www.diamondsemiconductor.com/
http://www.aist.go.jp/aist_e/latest_research/2005/20050615/20050615.html


basically this info was hidden in a article about synthetic diamond production by a process called vapour deposition or something. i personally found it fascinating.

edit.
more links
http://net127.com/2003/08/26/ntt-verifi...ond-semiconductor-operation-at-81-ghz/
http://www.ferret.com.au/articles/e9/0c019ee9.asp

 

[Maximilian]

How? Diamonds already cost too much.

 

[Duvie]

Originally posted by: Soviet
How? Diamonds already cost too much.

The diamonds would be on the size of being smaller then dust...Ever heard of diamond coated saw blades??? They dont cost thousands of dollars...

I also dont think the entire thing would be made of diamonds...

Also I read awhile back we were moving towards a more biological type of design...



http://ieeexplore.ieee.org/xpls/absprintf.jsp?arnumber=1512497
http://www.cscamm.umd.edu/programs/smc03/
http://www.drproctor.com/os/amorphous.htm

this last one is really interesting and not as far over my head...

http://web.engr.oregonstate.edu/~rorrergl/bio_synth.html

 

[BrownTown]

IBM just made a 500GHz transistor, so stuff like this is nothing new. Problem is that they don't work in real life. Also, In a CPU you have many transistors in series per pipeline stage. So if each one works at 90Ghz, and you have 15 per stage then you can only run at 6G, then you have wire delay and such, maybe get you down to 4G.

 

[Griswold]

Originally posted by: Soviet
How? Diamonds already cost too much.

Industrial made diamonds cost next to nothing (compared to natural diamonds) but have the same physical and technical qualities as "real" diamonds.

 

[RichUK]

Originally posted by: Griswold

Originally posted by: Soviet
How? Diamonds already cost too much.

Industrial made diamonds cost next to nothing (compared to natural diamonds) but have the same physical and technical qualities as "real" diamonds.

Thats correct, they take granite, which is the next densest material to diamond, compress it at a certain temperature which converts it to a diamond (physically). The only thing is although it has the same physical properties to a diamond it is no way near as pure as a real diamond, hence why they are worthless, and can?t on a monetary value be compared. However due to its physical make up, is perfect for industrial purposes.

 

[Fenixgoon]

Originally posted by: BrownTown
IBM just made a 500GHz transistor, so stuff like this is nothing new. Problem is that they don't work in real life. Also, In a CPU you have many transistors in series per pipeline stage. So if each one works at 90Ghz, and you have 15 per stage then you can only run at 6G, then you have wire delay and such, maybe get you down to 4G.

that transistor operates at near absolute 0 temps, probably superconducts

 

[Trippytiger]

Originally posted by: RichUK

Thats correct, they take granite, which is the next densest material to diamond, compress it at a certain temperature which converts it to a diamond (physically). The only thing is although it has the same physical properties to a diamond it is no way near as pure as a real diamond, hence why they are worthless, and can?t on a monetary value be compared. However due to its physical make up, is perfect for industrial purposes.

[/quote]

I believe you mean graphite, which is used because, like a diamond, it consists entirely of carbon atoms. The intense heat and pressure of that particular process for creating synthetic diamonds (there is another, called chemical vapour deposition) causes the molecular structure of the graphite to change into that of a diamond. Granite would be absolutely unsuitable for this because it generally contains, as far as I know, no carbon whatsoever.

 

[BrownTown]

Originally posted by: Fenixgoon

Originally posted by: BrownTown
IBM just made a 500GHz transistor, so stuff like this is nothing new. Problem is that they don't work in real life. Also, In a CPU you have many transistors in series per pipeline stage. So if each one works at 90Ghz, and you have 15 per stage then you can only run at 6G, then you have wire delay and such, maybe get you down to 4G.

that transistor operates at near absolute 0 temps, probably superconducts

at room temperature it was at 300Ghz, semiconductors don't run faster at colder temperatures, only the metal interconnects.

 

[dmens]

since they can work at higher tempratures than silicon thry can reach speeds of upto 89 GHz while silicon is limited to 10 GHz.

The fact that CPU/VLSI designers look for different electrical characterisitcs compared to the RF world is horribly mangling the tech media coverage of device "speed records". From a digital circuits point of view, silicon is in no way restricted, digital designers couldn't give a hoot about frequency response and all that jazz. The 89ghz/500ghz number has absolutely no meaning to the digital VLSI world.

For reference, single digit picosecond delay FO1 inverters already exist on productized silicon today, so that's like what, 100+ ghz, using the convoluted metric of switching delay inversed? LOL.

 

[RichUK]

Originally posted by: Trippytiger

Originally posted by: RichUK

Thats correct, they take granite, which is the next densest material to diamond, compress it at a certain temperature which converts it to a diamond (physically). The only thing is although it has the same physical properties to a diamond it is no way near as pure as a real diamond, hence why they are worthless, and can?t on a monetary value be compared. However due to its physical make up, is perfect for industrial purposes.

I believe you mean graphite, which is used because, like a diamond, it consists entirely of carbon atoms. The intense heat and pressure of that particular process for creating synthetic diamonds (there is another, called chemical vapour deposition) causes the molecular structure of the graphite to change into that of a diamond. Granite would be absolutely unsuitable for this because it generally contains, as far as I know, no carbon whatsoever.

[/quote]

Yep graphite was what i meant, thanks for the correction.

 

[coldpower27]

Here's a topic I read several years ago on synthetic diamond machines, something about producing diamonds which are even more perfect then natural ones.

"Unbelievable," he says slowly as he peers through the lens. "May I study it?" I agree to let him keep the gems overnight. When we meet the next morning in the lobby of the High Council, Van Royen looks tired. He admits to staying up almost all night scrutinizing the stones. "I think I can identify it," he says hopefully. "It's too perfect to be natural. Things in nature, they have flaws. The growth structure of this diamond is flawless."

http://www.wired.com/wired/archive/11.09/diamond.html?pg=4&topic=&topic_set=

 

[coldpower27]

Double Post.

 

[imported_inspire]

Originally posted by: Fenixgoon

Originally posted by: BrownTown
IBM just made a 500GHz transistor, so stuff like this is nothing new. Problem is that they don't work in real life. Also, In a CPU you have many transistors in series per pipeline stage. So if each one works at 90Ghz, and you have 15 per stage then you can only run at 6G, then you have wire delay and such, maybe get you down to 4G.

that transistor operates at near absolute 0 temps, probably superconducts

I read that article to say that the transistor would hit 300GHz at room temperature.