Nanocrystalline Silicon as a Thermoelectric Material -Bring the nanotechnological advantage into bulk-

The worldwide issue of the growing energy demand, along with the rapid decrease of the conventional fossil sources and the global warming, is one of the key challenges of this century. Among the solutions studied and developed to face these problems, reduction of the energy wastes and a strong energetic efficiency improvement appear to be necessary. Since it is well known that nearly the 60% of the energy generated around the world is rejected as heat, the possibility to recover even just a small percentage of this huge amount of energy could lead to a significant reduction of the consumption of fossil fuels and CO2 emissions. Thermoelectric devices are one of the viable options, especially because they can generate electrical power even with small temperature gradients and without moving parts. Energy conversion ratio for a thermoelectric material is related to its figure of merit: ZT=(α^2 σ)/κ, where the numerator, often call power factor (PF), is the product between the square of the Seebeck coefficient (α) and the electrical conductivity (σ), and the denominator is the thermal conductivity. Although it have been shown that the thermoelectric efficiencies, achievable in case of bulk materials, are small compared to that of conventional power sources, nanotechnology has opened in the last decade new ways to increase thermoelectric performances. In this PhD work we have studied a non-toxic abundant nanostructured material for thermoelectric applications, based on nanocrystalline silicon thin films highly doped with boron. In the first part we demonstrated the possibility to induce a great enhancement of the film power factors, due to a simultaneous increase of the Seebeck coefficient and of the electrical conductivity, by annealing between 500 and 1000 °C. These enhancements has been shown to be caused by the dopant segregation (induced by thermal annealing) and the creation of a second phase, which lead to an energy filtering effect enhancing the Seebeck coefficient of the films. These evidences along with other characterizations led to a model which was confirmed within the picture of a computational work. In the second part we studied the possibility to modulate the film thermal conductivity by generating a dispersion of nanovoids (NVs) using helium implantation and subsequent thermal treatments. The applicability of this technique, known only for the single crystal case, was tested on the silicon thin films. Concerning void morphology, we observed that while the helium dose rules the film final porosity (in our case ≈0.5 %), the characteristics of the annealing processes can be used to tune the NV morphology. In particular, since the NV creation is a coalescence process we observed an increase of the NV diameter, ranging between 1 and 20 nm, w the annealing temperature. Furthermore we characterized the compatibility of the energy filtering effect with the NV generation process, noting a reduction of PF enhancement by a factor two due to the slower dopant segregation. Finally we measured the thermal conductivity of a batch of samples treated at different temperatures (hence with different NVs morphologies) but with the same porosity, observing a proportionality between κ and the distance between NVs. This dependency was demonstrated to break down when the NV diameter is small because of the frequency dependence of the phonon−NV scattering mechanism. Unfortunately, although we showed that helium implantation is a useful technique to modulate the κ of nanocrystalline silicon thin films and that it is compatible with the mechanism of PF enhancement, minimal κ and maximum PF occur at different annealing temperatures. This mismatch, although seemingly surmountable, limited ZT values to ≈0.12. Overall, it may be concluded that heavily boron-doped nanocrystalline silicon may qualify as an interesting system to develop thermoelectric devices meeting acceptable performances with remarkably low material costs.