De Beers's Nightmare

[Iconoclast]

What's the big bruhaha here? There have been cheap, synthetic, industrial diamonds for decades now. Where do you think all that diamond jewelry from Service Merchandise came from?

The big "bruhaha" is that up until a few years ago we weren't able to grow large single crystal diamonds.

 

[Ziggurat]

I'm curious how noisy a fan will have to be to cool a diamond based processor? And what kind of fire insurance will I need?
Nitrogen filled case?

I don't see diamond chips happening any time soon. And a big reason is that silicon manufacturing technology is very mature, and still rapidly advancing. There's a lot of money being spent optimizing silicon technology, and there's simply no economic way to get to even a fraction of that funding with a new, unproven technology. And any time you start out with something totally different like this, the number of problems you need to solve is enormous, problems that silicon technology has already spent decades overcoming. No one is going to risk spending the money, and time, necessary to develop an uncertain technology to compete against a moving target like silicon. So nothing is going to replace silicon until the exponential increase in silicon technology starts to level off. The only hope for diamond chips in the near future is the possibility of niche applications, where it isn't performance or cost, but something completely different (like, say, heat tolerance) that lets it be used for applications that silicon can't do at all. But even then, don't count on it, because the money available for such a development isn't going to be very large.

 

[Yaotl]

I don't see diamond chips happening any time soon. And a big reason is that silicon manufacturing technology is very mature, and still rapidly advancing. There's a lot of money being spent optimizing silicon technology, and there's simply no economic way to get to even a fraction of that funding with a new, unproven technology. And any time you start out with something totally different like this, the number of problems you need to solve is enormous, problems that silicon technology has already spent decades overcoming. No one is going to risk spending the money, and time, necessary to develop an uncertain technology to compete against a moving target like silicon. So nothing is going to replace silicon until the exponential increase in silicon technology starts to level off. The only hope for diamond chips in the near future is the possibility of niche applications, where it isn't performance or cost, but something completely different (like, say, heat tolerance) that lets it be used for applications that silicon can't do at all. But even then, don't count on it, because the money available for such a development isn't going to be very large.

But wouldn't it be a trade off between the cost of coming up with new tech to create <90nm processors and the like and beginning anew with a material that doesn't need to be that small to be usable?

 

[phildonnia]

"It is not a symbol of eternal love if it is something that was created last week." So goes the De Beers-backed line.

I guess they decided against:

"It's not a symbol of eternal love unless it helped finance a bloody civil war in West Africa"


For those of us who get the dead-tree version of Wired, there was a cover with this naked babe covered with diamonds. And the funny thing was, she appeared to be one of those computer-generated images; creating a more thorough "synthetic" motif.

 

[TillEulenspiegel]

A few posters have noted the application in chips by this created diamond as an active element. That isn't really the case per se.

Modern Mil-Spec LSI chips use a Sapphire substrate as a foundation to deposit layers of conductors to formulate chips used in warheads, satellites and other high stress environments. I'm not quite sure how the constraints work , but in consumer electronics the intrinsic silicone substrate is used to create a Quantum "potential wells"which is utilized as an active part of the chip.

Sapphire is a bit more dense but diamond is stronger I'm not sure of the considerations of a trade off between the two since the info is not readily available, interesting tho. I'll have to track down some hard factual data to understand the case fully.

 

[Yaotl]

A few posters have noted the application in chips by this created diamond as an active element. That isn't really the case per se.

Modern Mil-Spec LSI chips use a Sapphire substrate as a foundation to deposit layers of conductors to formulate chips used in warheads, satellites and other high stress environments. I'm not quite sure how the constraints work , but in consumer electronics the intrinsic silicone substrate is used to create a Quantum "potential wells"which is utilized as an active part of the chip.

Sapphire is a bit more dense but diamond is stronger I'm not sure of the considerations of a trade off between the two since the info is not readily available, interesting tho. I'll have to track down some hard factual data to understand the case fully.

It would replace silicon, wouldn't it? And there are already a few places that are making the attempt. Newer article that gives a name or two.

 

[TillEulenspiegel]

I'm not sure Yaotl. As I stated most of the design characteristics are not published, but as Ziggurat mentioned there is an economy of scale. Who could be a more perfect consumer and driver of research then the DOD ?

The mention of Intel gives the study credence. There are questions I have which are not answerable by news releases.
The main one being the process of "doping" of silicon. The whole enterprise of semi-conductors is based on a potential difference between two similar materials. It's long and pedantic, bottom line being , I'm not sure how you could accomplish this in our diamond scenario. The questions are for material scientists.

I will poke around tho.

 

[Yaotl]

I'm not sure Yaotl. As I stated most of the design characteristics are not published, but as Ziggurat mentioned there is an economy of scale. Who could be a more perfect consumer and driver of research then the DOD ?

The mention of Intel gives the study credence. There are questions I have which are not answerable by news releases.
The main one being the process of "doping" of silicon. The whole enterprise of semi-conductors is based on a potential difference between two similar materials. It's long and pedantic, bottom line being , I'm not sure how you could accomplish this in our diamond scenario. The questions are for material scientists.

I will poke around tho.

The Wired article has something about the potential difference I think. Something about boron or other. It's beyond my ken.

 

[Beanbag]

The usual dopants for silicon are boron for a positive charge, and phosphorus for negative. I spent a few years as a semiconductor diffusion process tech.

Regards;
Beanbag