Silicon is known to be a poor thermoelectric material. Despite its high power factor, its large thermal conductivity makes its thermoelectric figure of merit zT as low as 0.01 around room temperature.
Over the last decade an effort has developed to increase the thermoelectric figure of merit of nanocrystalline silicon by increasing its power factor. Energy-selective charge carrier scattering (also referred to as energy filtering) was demonstrated long ago to provide a route to enhance power factor in nanocomposites. In the presence of potential barriers at phase boundaries, scattering of charge carriers occurs with an efficiency dependent upon carrier kinetic energies. Low-energy (‘cold’) carriers are scattered more efficiently than high-energy (‘hot’) ones, causing the average carrier mobility to increase while mobile (non-localized) charge carrier density decreases. This makes the Seebeck coefficient increase while keeping the electrical conductivity about steady. Such an effect was never reported in silicon till 2010 by this laboratory. Very heavily boron-doped nanocrystalline silicon films display a remarkable increase of their power factor when extensively annealed at temperatures above 800 °C. Transmission electron microscopy revealed that annealing promotes the precipitation of silicon boride around grain boundaries. Computational and theoretical analyses showed that the potential barriers generated at the interphases filter out cold holes, semi-quantitatively explaining why power factor increases from a few mW/mK2 in as-implanted films to > 20 mW/mK2 in fully annealed samples. However, boron precipitation occurs only when hydrogen (embedded in thin films upon CVD deposition) is let diffuse out of the sample, since hydrogen forms stable complexes with boron, preventing its precipitation.
The availability of high-zT nanocrystalline silicon thin films paves the way to remarkably relevant novel applications of thermoelectrics, sensibly impacting microharvesting for a wide range of miniaturized devices.
Due to its high thermal conductivity, single-crystalline silicon is not suitable as such for thermoelectric application. However, nanostructuration offers efficient and innovative ways to lower its thermal conductivity, opening relevant opportunities to its usage as thermoelectric material. Among the several possible routes, Silver-assisted Chemical Etching (SaCE) discloses a simple way to obtain Si nanowires (SiNWs) with a solution-based one-step procedure. SaCE is a metal-catalyzed reaction occurring at the outer substrate surface leading to the spontaneous formation of Si nanowires. We could obtain SiNW with lengths exceeding 10 micrometers. Quite astonishingly, even fully etched Si wafers could be attained, with NWs being self-sustained through tip bundling. This, along with the possibility of making electrical contacts localized at Si tips, discloses the opportunity of making 'bulk' SiNW pads fully made of high-efficiency SiNWs - to be embedded into thermoelectric devices, either for cooling or for heat harvesting.
Despite an outstanding increase of conversion efficiency, photovoltaic (PV) generators are still substantially ineffective at converting solar energy into electricity, as more than half of the solar power that could be converted is degraded to heat. Over the last years, many ways of pairing photovoltaic and thermoelectric stages have been considered, modeled and, to a lesser extent, experimentally validated. A detailed analysis of the energy balance in both photovoltaic-thermoelectric solar generators shows that solar harvesting might be improved through an accurate coupling of PV cells to thermoelectric generators, especially when a moderate solar concentration is considered. Further to energy profitability, hybrid harvesters also meet criteria of economic convenience, which is required for pairing to be practically successful.
This lab actively participates in the effort to devise and prototype hybrid solar converters, focusing on the direct pairing of specially designed Bi2Te3 thermoelectric generators with perovskite-based photovoltaic cells. Currently, research is focused on the development of suitable heat mirrors minimizing radiative heat dissipation at the cell front.
Microgeneration is a specific instance of thermoelectric generation that may complement or compete with conventional batteries as power sources in distributed low-power sensing. Our energetic and economic analyses carried suggest that microgeneration may be competitive as an alternate powering strategy in the enormously growing fields of the Internet of Thing and of Industry 4.0.
An endeavor to validate these conclusions through suitable proofs of concept is under way.
Thermoelectric low efficiencies are basically due to the difficulty of decoupling thermal and electrical conductivity in real systems. Nanotechnology has shown innovative ways to reach such a target – but it is possibly unrealistic to expect further major improvements in this direction. On the other side, power output depends on heat input power, so that only 3D systems may provide useful outcomes.
A possibility to decouple thermal and electrical conductivities in bulk thermoelectrics (and possibly also to enhance the Seebeck coefficient) might arise from operating thermoelectric generators in non-steady state mode. Large differences between the relaxation times of phonons and electrons might enable heat to be converted into electric energy with larger efficiencies when heat is supplied with periods intermediate between phonon and electron relaxation times, therefore silencing phonon modes while not perturbing charge carrier transport. At such frequencies, however, classic irreversible thermodynamics is no longer valid. We are exploring the use of the Extended Irreversible Thermodynamics to analyze thermoelectric devices operating in non-steady state conditions in view of the possible enhanced conversion efficiencies achievable under such operating conditions.