Diffusion mechanisms of Mo contamination in Si

Molybdenum contamination of silicon can have serious detrimental consequences for the efficiency of solar cells, raising durability concerns for novel solar cell designs that utilize $\mathrm{Mo}{\mathrm{O}}_{3}$ in contact with Si. Density functional theory simulations of Mo defects in Si revealed that Mo is preferentially accommodated in tetrahedrally coordinated interstitial sites and that the contamination may reach a sufficiently high concentration to cause a 20% relative solar cell efficiency degradation if processing steps are performed between 950 and 1300 K. The formation energy of the most energetically favored Mo defect in Si has a minimum value of 1.58 eV at the valence band maximum and a maximum of 2.10 eV at higher Fermi levels, indicating that higher Mo defect concentrations may occur in $p$-type Si than intrinsic or $n$-type Si. The diffusion processes for Mo in Si were investigated, and it was identified that interstitial diffusion dominates over a vacancy-mediated mechanism under all equilibrium conditions. Migration barriers were calculated to be 2.29 eV for charge neutral and 2.03 eV for charge +1 defects, occurring under $n$-type and $p$-type doping, respectively, indicating that Mo diffusion is faster in $p$-type Si, and hence potentially more effectively gettered than it would be in $n$-type Si.