Role of Cd interstitial in group-V doped CdTe
Role of Cd interstitials in group-V doped CdTe
Engineering topics
Publications and source records attributed to Chatratin, Intuon.
Role of Cd interstitials in group-V doped CdTe
Cadmium Telluride is at the core of an important thin-film technology for photovoltaics that is already commercially available, yet the CdTe-based solar cell efficiency remains limited at 22%, well below the theoretical limit of ~30%. Increasing the hole concentration is crucial for higher efficiency, and group-V elements such as As, P, and Sb are potential dopants as they are shallow acceptors. Nevertheless, group-V doped p-type CdTe often exhibits low doping activation, and the compensation source remains debated. Here, we performed hybrid density functional calculations with spin-orbit coupling to investigate possible sources of hole compensation in group-V doped CdTe. First, regarding possible self-compensation of the group-V dopants, we find that the formation of AX centers is unlikely since they are found to be unstable relative to the shallow acceptor forms. However, if the group-V dopants come in during growth (such as dimer molecules As2, P2, and Sb2), we find that the impurity atoms, which would occupy nearest neighbor sites, maintain the V-V bonds, limiting the hole density. For the native defects, our study reveals that Cd interstitial is the lowest energy donor defect in p-type CdTe. Still, it has a small migration barrier of 0.5 eV, making it highly mobile and unstable at room temperature. The Te vacancy is the next lowest formation energy donor. The migration barrier of 1.4 eV indicates that the Te vacancy is stable at room temperature. The antisite CdTe is also a donor, with low formation energy and stable at room temperature, potentially limiting the hole concentration. Our results, therefore, shed light on possible compensation centers and some guidance on how to avoid them.
Theory of formability and diffusivity of point defects and impurites in Cd(Se,Te)
Compensation centers in group-V doped CdTe
CdTe is a leading thin-film technology for solar-cell applications. Basic features such as the optimum band gap of 1.5 eV and simple manufacturing process have been driving the research on improving the current record efficiency of 22% for CdTe solar cells toward the theoretical limit of 32%. Low p-type doping and short carrier lifetimes in the CdTe absorber have been suggested as the main limiting factors and attributed to defect compensation and carrier recombination. Understanding the role of point defects and impurities in CdTe is crucial to solving these problems. Despite many years of research, the sources of compensation and the microscopic recombination mechanisms are still unclear. By using hybrid density functional calculations with spin-orbit coupling and large supercells, we investigate the electronic properties, the formation energies, and the diffusivities of the native defects in CdTe. As possible sources of compensation for p-type conductivity, we discuss the structural and electronic properties of Cd interstitial, Te vacancies, and TeCd antisites. The stability of each defect is discussed in terms of both the defect formation energy and the migration barrier. We also address possible mechanisms that lead to non-radiative recombination.
Polycrystalline cadmium telluride (CdTe) is a leading material in photovoltaic technology due to its high absorption coefficient and near-optimum bandgap of 1.44 eV. It is known that CdTe film processing can promote surface oxidation depending on the growth environment and upon exposure to different oxidation conditions. For example, CdTeO3 forms when the CdTe film is treated in heated dry air, while in humid air, CdTe2O5 is detected. These oxides feature tellurium in the oxidation state +4 compared to the +2 state in CdTe. Other possible relevant oxides are CdO, TeO2, and TeO3. Using hybrid density functional calculations, we studied the electronic structure of these oxide materials and their band alignment to CdTe, which are essential parameters in the characterization of the interfaces at grain boundaries. The goal is to understand their stability and possible effects on passivating grain boundaries. The results are compared to the available experimental data.
Predicting accurate shallow donor and acceptor levels in semiconductors has been quite challenging using periodic boundary conditions as implemented in current density functional theory codes. The reason is that the wave functions associated with the shallow centers are quite extended and are not fully contained in the supercell (typically with a few hundred atoms) so impurities in the periodically repeated image cells interact with each other. Errors of ~0.1 eV are expected, and these are of the same order of magnitude as the ionization energies themselves. In the case of acceptors in CdTe, this problem is exacerbated by the strong spin-orbit coupling that split the Te-related states at the top of the valence band, making the calculations at least 8 times more expensive. CdTe is an important solar-cell material with record high efficiency of 22%. One of the main limiting factors to increasing the efficiency towards the theoretical limit of ~30% is the often reported very low hole concentration. Pushing to the limit of computational capability by using very large supercells with spin-orbit coupling we report the results of hybrid functional calculations of group-V acceptors in CdTe. We show that extrapolation to the dilute limit leads to an interpretation of the experimental data that is qualitatively different from previous DFT and hybrid functional calculation reports. We find that the group-V impurities indeed behave as shallow acceptors and that the corresponding compensating AX-centers are unstable and do not limit p-type doping. We address the differences between our results and previous theoretical predictions and show that our calculated ionization energies predict hole concentrations that are in excellent agreement with recent temperature-dependent Hall measurements on high-quality single-crystal samples.
Low p-type doping is a limiting factor to increase CdTe thin-film solar-cell efficiency toward the theoretical Shockley-Queisser limit of 33%. Previous calculations predict relatively high ionization energies for group-V acceptors (P, As, and Sb) and they are plagued by self-compensation, forming AX-centers, severely limiting hole concentration. However, recent experiments on CdTe single crystals indicate a much more favorable scenario, where P, As, and Sb behave as shallow acceptors. Using hybrid functional calculations, we solve this puzzle by showing that the ionization energies significantly decrease with the supercell size. When including the effects of spin-orbit coupling and extrapolating the results to the dilute limit, we find these impurities behave as hydrogenic-like shallow acceptors, and AX-centers are unstable and do not limit ptype doping. We address the differences between our results and previous theoretical predictions and show that our ionization energies predict hole concentrations that agree with recent temperature-dependent Hall measurements.
CdTe-based solar cells are leading thin-film photovoltaic technology, with efficiencies over 22%, but still much lower than the theoretical maximum of 29%. Further improvements will rely on increasing the open-circuit voltage Voc, which, in turn, depends on carrier density and lifetime. Using hybrid density functional calculations, we investigate Sb doping of CdTe, focusing on its limitation as shallow acceptor and the formation of compensating AX center. Paying special attention to supercell size and effects of spin-orbit coupling, we predict an ionization energy of 116 meV in the dilute limit, much closer to recent experimental value of 103 meV from temperature-dependent Hall measurements in bulk single crystals, and in contrast to much larger values from previous calculations. We also find that the Sb-related AX centers are not major compensation centers in Sb-doped CdTe.
Group-V acceptor ionization energies and compensation centers in CdTe revisited: Spin-orbit coupling is crucial to describe the band structure of CdTe and the properties of group-V acceptors in CdTe; explains the difference between present work and previous calc. P, As, and Sb are shallow acceptors in CdTe, with the ionization energies ~100 meV, in agreement with recent Hall measurements in bulk crystals AX center is not the dominant compensation center in p-type CdTe; unstable in the case of As and P, barely stable in the case of Sb doping Intrinsic defects, such as VTe, are potentially important compensation centers Cdi is unstable with low migration barrier