Mechanism Of Grain Boundaries On CuInSe₂ Thin Film Solar Cells

CuInSe2 and Cu(In,Ga)Se2 have a direct band gap and a high optical absorption coefficient, and are the suitable absorbing layer materials for thin film solar cells. In recent years, they have been investigated extensively. However, the materials were found to behavior different from traditional semiconductors:the devices made from the traditional polycrystalline semiconductors usually have better performance than that from their single-crystal counterparts, while the CuInSe2 and Cu(In,Ga)Se2 polycrystalline thin films solar cells have higher conversion efficiencies than single crystal ones. This phenomenon is very interesting and has attracted a lot of attentions.Generally, the grain boundary in thin films is considered as the key factor to give the polycrystalline thin films solar cells a better performance, but the mechanism of how the grain boundary influences the solar cells is still unclear. This thesis will focus on this problem with the aim to clarify the role of the grain boundary.The thesis uses the first principle theory to investigate the grain boundary. The research includes:1. The electronic structure of CuInSe2 crystal was studied. The ligand field theory and hybrid orbital theory are used to explain the bonding in CuInSe2. The calculation indicates that the Cu-Se bond is covalent, and is a hybrid between Cu-d and Se-p orbitals. The In-Se bond is more like an ionic bond, is formed through sp3 hybridization. The top of the valence band of the CuInSe2 is determined by Cu-d and Se-p. Different from the traditional points, the calculation shows that the bottom of the conduction band is determined by In-s, Cu-s and Se-p.2. The atomic structure of (112) surface in CuInSe2 is investigated. After the surface relaxation of the clean (112) surface (no defects), the structural change is similar to that of Si (111) surface, which is in accordance with the sp3 hybridization degeneration. On the basis of the relaxed (112) surface structure, the introduction of a copper vacancy results in a lower surface energy. This indicates that the copper-lacking (112) surface is a stable structure, and there should be many Cu vacancies at the grain boundaries.3. By comparing the formation energy of copper vacancy Vcu at different positions in CuInSe2, it is found that Vcu tends to be distributed in the grain interiors instead of the grain boundaries. After the introduction of copper vacancies, the reduction in the p-d repulsion and the weakening of the chemical bond strength lead to the downward of the valence band maximum and the upward of the conduction band minimum. So the hole barrier and the electron barrier are generated simultaneously. The same method has also proved that the defects pair 2V-Cu+InCu2+ (two Cu vacancies and In replacing Cu) also tend to be distributed in the grain interiors, and result in the downward of the valence band maximum and the conduction band minimum. Both VCu and 2V-Cu+InCu2+ can reduce the recombination of carriers.4. By comparing the formation energy of Na related defects at different positions in polycrystalline CuInSe2, the doping and distribution of Na in CuInSe2 are studied. The results show that Nacu (Na replacing Cu) has a low formation energy, and is more likely to be distributed in the grain boundaries instead of the grain interiors. The Na atom and the Se atom form an ionic bond. Due to reduced p-d repulsion, the valence band maximum moves down and results in a hole barrier. So the grain boundaries with Nacu prevent only holes to pass through, and reduce the recombination of carriers. This provides an explanation why the grain boundary in CuInSe2 can improve the performance of the solar cells.5. The influence of Cd atoms doped in CuInSe2 is investigated. The formation energies of defects Cdcu (Cd replacing Cu), CdIn (Cd replacing In) and Cdi (Cd between atoms) in the bulk CuInSe2 are calculated. The results show that the Cd atoms like to replace Cu and In atoms due to lower formation energies. On the (112) surface, the Cd atom tends to be combined with the Vcu to form Cd+Cu+V-Cu. The defect pair makes the valence band maximum downward and result in a hole barrier. Cd atoms doped on the CuInSe2 (112) surface can reduce the carrier recombination in the CuInSe2/CdS heterojunction and improve the output current of the solar cells.