Are you pursuing any patents?

Our U.S. Patent number is 8624100, issued 1/7/2014.  
International patents are pending.

Do you have a prototype in the works?

We’ve built three prototypes out of AlGaAs.  
We selected AlGaAs for ease of manufacture rather than performance.

Do you have any quantitative data from theory or experiments?

Yes.  http://www.thermawatts.com/atec-test-results.html
Theoretical results can be found in Section 3.  
Experimental results can be found in Section 4.

If I have a waste heat stream, how much of the energy can be recovered from the stream using ATEC?

We can recover a majority of the heat from a heat exhaust stream.  The amount recovered depends on the nature of the exhaust stream and how the ATEC is packaged to absorb heat.  If it’s a combustion product rich in carbon dioxide, you probably want to leave it above ambient so it gets carried away by convection.

How does this scheme compare with old style Seebeck junctions?

Seebeck effect results from a combination of differences in carrier density based on temperature, and carrier selection based on differences in carrier mobility because of the band structures of the materials. We use differences in carrier density based on the band structures of the materials, and carrier selection based on doping.

Does this technique produce greater electron-hole pairs per material volume?

Electron-hole pairs are primarily a function of bandgap.  For ATEC, the more carriers there are, the higher the performance.  This is really only constrained by the doping limits of the material and Auger recombination in the ‘low recombination material’.  Auger recombination in the ‘high recombination material’ is preferred.

How many sq cm of junction are needed to produce 1000 watts?

Our typical design for modeling uses many repetitions of the cycle, each of which is about 0.5 microns thick.  Each cycle generates only a small voltage, limited to about 50 millivolts (2kBT).  The highest performance models suggest we can produce 8 amps per sq. cm.  If you want 1000 watts from a unit, you will need something the size and thickness of a credit card.

What temperature is the sweet spot?

It varies with material set.  Maximum performance is where the intrinsic carrier density is close to the doping level.  HgCdTe should work pretty well across the -20 to +30 degrees C range.

Does the output increase with temperature (or does the carrier lifetime drop off reduce the output)?

Output increases exponentially with temperature until the carrier density approaches the doping levels. The carrier lifetime drop off also improves performance.

Since silicon is limited in temperature range, does one have to use higher temp semiconductors like GaN?

The ATEC structure requires either two different materials, or some way of manufacturing the material with different properties in one of the layers.  Silicon by itself doesn’t exhibit the effect.  Also, Silicon has too wide a band-gap to be useful as one of the materials.  Silicon starts to anneal before it gets hot enough to have a useful carrier density.

What is the temperature difference required to generate useful power?

The ‘band structure’ of the materials acts like temperature difference, so no external difference is required.

What about the Second Law of Thermodynamics?

The Second Law of Thermodynamics was conceived half a century before semiconductors were discovered.  We (engineers and scientists) work the the Second Law, and it works very well.  It has not complete.

What we have built incorporates elements of both Maxwell’s Demon and a Brownian Ratchet in a way that fixes their well-known technical problems.