First-principles Study Of Boron-oxygen And Their Related Complexes In Silicon

Solar energy is an important renewable energy. Photovoltaic technology is regarded as one of the most effective ways to utilize the solar energy. At present, more than80%of solar cells are based on crystalline silicon. However, the solar cells made from boron-doped Czochralski silicon (CZ-Si) inevitably suffer the light-induced degradation (LID), leading to a loss of3-5%in solar cell efficiency. This largely limits the performance of solar cells. Initial researches indicated that this phenomenon is caused by boron-oxygen complex in silicon. However, there still exists some controversy about the nature of LID. Meanwhile, even though it has been reported that the LID can be suppressed by doping isovalent atoms into silicon, understanding the mechanism still needs more investigation.In this thesis, we have investigated the components and configurations of boron (B) and oxygen (O) complexes in crystalline slicon. Using first-principles calculation, the relation between boron, oxygen dimer and isovalent elements have been clarified. Below is the innovative results achieved in this thesis:(l)The stable configuration and electronic properties of B;O2i have been investigated. We found that the stable configuration of B;O2i keeps constant in various charge states, thus there is no reconfiguration among different charge states. The electronic level induced by B;O2i is calculated to be (Ec-0.4eV)-(Ec-0.5eV). All these results are against the newly proposed latent BjO2i defect model. Therefore, we can draw the conclusion that the latent BiO2i defect cannot be responsible for LID.(2) The interaction between substitutional carbon/germanium/tin (C/Ge/Sn) and substitutional boron has been studied. It is found that the C and B tend to repel each other. It is energetically favorable for them to locate far away from each other in silicon lattice. Ge/Sn and boron can form stable complexes in silicon since they can compensate the stress in silicon lattice caused by each other. Using action law, the equilibrium concentrations of Ge-B (Sn-B) related complexes are considerably low, compared to the doping concentrations of Ge (Sn) and B. This lead us to believed that most of Ge (Sn) atoms should not form complexes with B.(3) The interaction between substitutional C/Ge/Sn and oxygen has been investigated. It was found that C can easily capture interstitial oxygen (Oi) and oxygen dimer (On) to form stable COj and CO2i complex, respectively. However, it is difficult to form GeOi (SnOi) and GeO2i (SnO2i) complexe since both Ge (Sn) and O; or O2i cause the compressive stress in the lattice. Meanwhile, Ge (Sn) can also suppress the formation of O3i. All these three isovalent elements can improve the migration energy (Em) of both Oi and O2i.(4) The mechanism for the suppression of LID in crystalline silicon with isovalent element doping has been proposed. In C-doped CZ-Si (CCZ), the diffusion of Oi become more difficult, and meanwhile, the C atoms can capture O2i and form stable CO2i complex. This will result in the decrease of the O2i concentration ([O2i]), and therefore reduce the concentration of B-O defects (Nt). In Ge-doped CZ-Si (GCZ) or Sn-doped CZ-Si (SCZ), since both Ge and Sn atoms enhance the Em of O; and suppresses the formation of O2i, the [O2i] is much lower than that in the conventional CZ-Si. The larger Em of O2i in the crystalline silicon with isovalent doping necessarily cause the higher generation activation energy of B-O complexes.