Maxwell wrote in Theory of Heat that the Second Law of Thermodynamics requires the Zeroth Law of Thermodynamics. Carnot pointed to the observed atmospheric temperature lapse rate as an example of what we now call adiabatic expansion, thus rejecting the Zeroth Law. We find similar Zeroth Law violations in LEDs, which says that solid state is not bound by the Second Law. This is the essence of our technology.
Our technology converts environmental heat directly to electricity without a cold side. It makes use of quantum-mechanical behavior of electrons in semiconductors between materials of differing composition. Stacks of thin layers of selected semiconductor materials generate persistent voltages across themselves, driven by differences in band gap (energy needed to move electrons between valence and conduction bands in the materials).
Device operation springs from two well-known features of semiconductors:
- Differences in carrier densities-of-state between differing semiconductor compositions, by as much as two orders of magnitude.
- Diffusion of free electrons from high to low densities-of-state materials, causing a standing voltage.
The figures below portray a heterojunction of two compositions of Hg1-xCdxTe and the consequent net carrier flows Jn and Jp as driven by diffusion, per Yang. By default there is a continuous flow of carriers because the materials are inherently out of equilibrium. This dynamic equilibrium of charge results from:
- Electrons tunneling from the charge
- Holes diffusing toward the charge
The flows result in a churn of electrons across the heterojunctions and a standing negative charge on the low density-of-state side.
We exploit those tendencies to flow by directing them with alternately-doped layers between the high and low density-of-state layers, blocking electrons from going right and (less significantly) holes from going left. Now we have both a net electrical charge and a current.
Electrons and holes flow in opposite directions in a series of high and low density-of-state with p- and n-doped layers.
We assume energy balance within and around our device, but can’t yet confirm it experimentally.
This table develops the carrier flows and current across heterojunctions of AlGaAs and HgCdTe using Yang and accepted parameters of these semiconductors. Yang confirmed AlGaAs flows experimentally.
The flows and corresponding band gaps for HgCdTe are:
The predicted net flow of electrons is in the order of 6×1023, or 105 amps per cm2.
The critical issues for our device are:
- Whether electrons flow continuously across the heterojunction
Yang validated his model with AlGaAs measurements.
ThermaWatts showed LEDs self-establish two temperatures, a Zeroth Law violation.
ThermaWatts shows that zero-biased LEDs emit more light than reverse-biased LEDs.
ThermaWatts showed light emitted near and below zero bias can be well above the lattice temperature, a Zeroth Law violation.
- Whether doping works:
Prototypes exhibited their signature voltages – stable, repeatable. Serial connection of prototypes #7 and #9 added their voltages (to 13 microvolts).
This is a truly innovative technology that, if brought to fruition, will have a huge societal impact. Prototyping is costly, and the work remains speculative. NSF’s mission aligns with such speculative and potentially world-changing research.