Waste-heat harvesting in nano-structures

Researchers demonstrate how voltage fluctuations in a micro-fabricated structure can be converted to electricity. This work paves the way for electronic circuits capable of converting excess heat into usable energy. 


Anyone working in an office has witnessed how the heat generated from computers, printers and photocopiers can noticeably increase the ambient temperature. Data storage servers and computer clusters, in particular, generate an enormous amount of heat that ought to be removed by costly air conditioning means. Heat-recycling circuits integrated into everyday electronics could lower our air-conditioning needs and put a fraction of the wasted energy back into use. In this issue of Physical Review Letters, Hartmann and colleagues report a proof-of-principle experiment that brings us a step closer to such a technology [1].


Conversion of heat to electrical power can be achieved through a phenomenon known as the Seebeck effect [2]. The ideal convertor must be a poor heat conductor but have excellent electrical conductivity such that the temperature gradient is maintained despite the current flow. As heat and charge transport are closely linked, no naturally occurring material meets this criterion.  A series of pioneering experiments in the 1990s demonstrated the Seebeck effect in nano-structures [3-4], and researchers began to wonder whether some of the limitations of conventional materials could be overcome through utilization of quantum mechanics at the nano-scale.  In 2011, Sánchez and Büttiker put forward a new design that ingeniously circumvented the conductivity dilemma by decoupling the direction of the heat and the charge currents.

The proposal is a three terminal device consisting of two quantum dots (QDs) and three thermal reservoirs (Fig. 1). The upper QD, known as the conductor, is in contact with hot and cold reservoirs. The lower dot is in contact with a gate reservoir that provides the necessary thermal fluctuations. When the conductor dot is empty, electrons flow from the hot reservoir into the conductor.
The energetic barrier between the conductor and the cold reservoir prevents the tunneling of the electron out of the conductor.
The electron will occupy the conductor until fluctuations in the gate reservoir excite an electron into the lower dot. The resulting Coulomb repulsion forces the electron out of the conductor. The temperature difference between the two reservoirs ensures that on average the flow is from the hot to the cold reservoir.

Schematic diagram of the experiment. The upper QD is known as the conductor dot and is in contact with hot and cold reservoirs. The lower QD is in contact with a fluctuating reservoir, known as the gate. The temperature fluctuations are simulated by applying a voltage with a white noise spectrum.  I and J represent the direction of the current and the heat flow respectively. 

To realize this scheme, Hartmann and colleagues utilized micro-fabricated GaAs/AlGaAs heterostructure as quantum dots.  A high mobility two-dimensional electron gas (2DEG) located about 80 nm below the surface took the role of the electron reservoir. The components were connected via quantum point contacts to allow electron tunneling in either direction.  As achieving thermal fluctuations is difficult experimentally, the fluctuating thermal reservoir was simulated through a gate voltage with a white noise spectrum. Similarly, thermal reservoirs were simulated by applying voltage biases to the 2DEG. The researchers observed output powers (currents) in the pW (nA) region.

For practical applications, to make the device economically viable, the power required to maintain the thermal reservoirs must be lower than the output power of the device.  This demands channeling the heat generated in the electronic circuitry into a small volume, and maintaining a source of cold electrons that are isolated from the ambient temperature. Further experiments, ideally with thermal electron sources, are thus required to assess the efficiency of such a converter.

 This technology may find applications in nano-scale temperature stabilization where a bias is generated in response to temperature fluctuations thus counter-acting undesirable thermal changes. From a theoretical perspective, a working prototype of a quantum thermoelectric device can provide the ideal test-bed of quantum thermodynamic concepts. It would be interesting, for instance, to investigate under what circumstances quantum effects can enhance the classical efficiencies, and whether the efficiencies can be improved via quantum reservoir engineering.

1. F. Hartmann, P. Pfeffer, S. Höfling, M. Kamp, and L. Worschech, Phys. Rev. Lett. 114, 146805 (2015).

2. L. E. Bell, Science 321, (2008).

3. S. F. Godijn, S. Moller, H. Buhmann, L. W. Molenkamp, and S. A. van Langen, Phys. Rev. Lett. 82, 2927 (1999).

4. L. W. Molenkamp, Th. Gravier, H. van Houten, O. J. A. Buijk, M. A. A. Mabesoone, and C. T. Foxon. Phys. Rev. Lett. 68, 3765 (1992).

5. R. Sánchez and M. Büttiker, Phys. Rev. B 83, 085428 (2014).