Mega-breakthrough? Super-efficient thermo-electrics?

Brian H · Post by **Brian H** » Sun Sep 26, 2010 6:05 pm

This is beyond huge: http://www.physorg.com/news204552797.html .
A "ZT" rating of 50 or so, maybe?

Micron-thick ultra-efficient super-cheap heat->power conversion. Everything from engines to refrigerators to computer chips to solar panels to reactors to coal-fired power plants to clothing to ....

chrismb · Post by **chrismb** » Sun Sep 26, 2010 8:26 pm

"With a very efficient thermoelectric device based on our design, you could power about 200 100-Watt light bulbs using the waste heat of an automobile," he said.

20kW!!!!!....

arff, arff, arffffff

Yeah, sure. Are people so daft that they actually believe this stuff that comes out of computers is actually the real deal?

So let's just say it's 100% efficient (!

!), there's gonna be this nano-'stuff' that can take on 20kW of power and it is a thousandth of a micron in thickness!?!? Anyone care to tell me something about thermodynamics here! How hot is this stuff gonna get?

JohnFul · Post by **JohnFul** » Mon Sep 27, 2010 12:44 am

Hmm.

I work for the company that has produced the most energy efficient data centre on earth http://www.marketwatch.com/story/netapp ... _news_stmp http://www.netapp.com/us/communities/te ... -0509.html . If even 25% of the waste heat could be converted to electricity ...

I run 96 cores and 100TB of SAN/NAS at home in my personal compute environment, The heat generated is amazing, The combined direct electric consumption/cooling load is something to be considered. I think it would be a much more lucrative market for a startup than automotive.

Interesting technology to watch.

J

Damon Hill · Post by **Damon Hill** » Mon Sep 27, 2010 3:12 am

Never trust extrapolations based on theoretical models, when working devices have yet to be built.

Betruger · Post by **Betruger** » Mon Sep 27, 2010 4:23 am

chrismb wrote:"With a very efficient thermoelectric device based on our design, you could power about 200 100-Watt light bulbs using the waste heat of an automobile," he said.

20kW!!!!!....

arff, arff, arffffff

Yeah, sure. Are people so daft that they actually believe this stuff that comes out of computers is actually the real deal?

So let's just say it's 100% efficient (!!), there's gonna be this nano-'stuff' that can take on 20kW of power and it is a thousandth of a micron in thickness!?!? Anyone care to tell me something about thermodynamics here! How hot is this stuff gonna get?

Was going to say, maybe they'd given the nanolayer example for lower energy drains, and extrapolated 20kW for some more substantial heat sink. But that really is what the article says. Weird.

Haven't read the actual paper yet, but Rcain mentions a much smaller ZT over in Giorgio's thread (incl link to paper).
viewtopic.php?p=49137

Brian H · Post by **Brian H** » Mon Sep 27, 2010 4:35 am

chrismb wrote: ...

So let's just say it's 100% efficient (!!), there's gonna be this nano-'stuff' that can take on 20kW of power and it is a thousandth of a micron in thickness!?!? Anyone care to tell me something about thermodynamics here! How hot is this stuff gonna get?

Numbers matter. They say one-millionth of an inch, which is about 1/40 of a micron, not a thousandth. The structure is so simple (metal/polymer/metal sandwich) I'd expect, like the experts they consulted, that it can be made very easily and cheaply. Probably it's being worked on as we speak, given the lag between completion of a report and its publication date.

chrismb · Post by **chrismb** » Mon Sep 27, 2010 8:08 am

Brian H wrote:Numbers matter. They say one-millionth of an inch, which is about 1/40 of a micron, not a thousandth.

Silly me! That'll be OK then...

Brian H wrote:The structure is so simple (metal/polymer/metal sandwich) I'd expect, like the experts they consulted

Brian H wrote: that it can be made very easily and cheaply. Probably it's being worked on as we speak, given the lag between completion of a report and its publication date.

In my humble experience, computers are great for taking a conceivable construction and refining it to the peak efficiency possible. I have never, ever yet seen a computer used to design something that has simply never been built before. If this principle was viable, then such a blanket could have been built by now, even if it was only 1% efficient and needed improvement by computer design.

If ever there was a case for building a working prototype BEFORE working on the computer simulation, this is it!!!!

Betruger · Post by **Betruger** » Mon Sep 27, 2010 8:27 am

Sounds like chicken and egg - computer modeling is what prompted this thing. Hasn't software modeling successfully been that kind of influence for superconductor research? In this article IIRC some researchers say the production folks told them they were ready to build the software's suggested formula.

chrismb · Post by **chrismb** » Mon Sep 27, 2010 9:05 am

Betruger wrote:Hasn't software modeling successfully been that kind of influence for superconductor research?

No. Super conductor research was prompted by physical experiments done in 1911. No computers then.

Can anyone name a device or technology for which computer simulations were done prior to a technological demonstration [that worked!].

Betruger · Post by **Betruger** » Mon Sep 27, 2010 10:01 am

Obviously.. I do mean recent superconductor research. That this kind of software-directed research development hasn't happened yet is no guarantee it won't ever. Even if only minor milestones.

GIThruster · Post by **GIThruster** » Mon Sep 27, 2010 3:37 pm

For software to help in superconductor research we'd need to know how they work, and we don't.

Betruger · Post by **Betruger** » Mon Sep 27, 2010 4:08 pm

I thought I'd seen reports of software modeling successfully predicting where to look for new/better SCs. Either way I don't see why software won't ever be up to the task of accurately suggesting research leads, or outright novel solutions. It's just a matter of time.

MSimon · Post by **MSimon** » Tue Sep 28, 2010 5:44 am

Brian H wrote:
chrismb wrote: ...

So let's just say it's 100% efficient (!:wink:!), there's gonna be this nano-'stuff' that can take on 20kW of power and it is a thousandth of a micron in thickness!?!? Anyone care to tell me something about thermodynamics here! How hot is this stuff gonna get?
Numbers matter. They say one-millionth of an inch, which is about 1/40 of a micron, not a thousandth. The structure is so simple (metal/polymer/metal sandwich) I'd expect, like the experts they consulted, that it can be made very easily and cheaply. Probably it's being worked on as we speak, given the lag between completion of a report and its publication date.

Care to explain how much delta T you can maintain across such a thin layer?

MSimon · Post by **MSimon** » Tue Sep 28, 2010 5:47 am

There used to be a lot more rational scepticism (an engineering mind set) around here.

Hope springs eternal. But Murphy is even older.

Brian H · Post by **Brian H** » Tue Sep 28, 2010 8:13 am

MSimon wrote:
Brian H wrote:
chrismb wrote: ...

So let's just say it's 100% efficient (!!), there's gonna be this nano-'stuff' that can take on 20kW of power and it is a thousandth of a micron in thickness!?!? Anyone care to tell me something about thermodynamics here! How hot is this stuff gonna get?
Numbers matter. They say one-millionth of an inch, which is about 1/40 of a micron, not a thousandth. The structure is so simple (metal/polymer/metal sandwich) I'd expect, like the experts they consulted, that it can be made very easily and cheaply. Probably it's being worked on as we speak, given the lag between completion of a report and its publication date.
Care to explain how much delta T you can maintain across such a thin layer?

Yes, thermal insulation is an issue. Be interesting to see the entropy calcs on that.