Current knowledge of the intrinsic defect properties of ${\text{Cu}}_2{\text{ZnSnS}}_4$ (CZTS) is limited, which is hindering further improvement of the performance of CZTS-based solar cells. Here, we have performed first-principles calculations for a series of intrinsic defects and defect complexes in CZTS, from which we have the following observations. (i) It is important to control the elemental chemical potentials during crystal growth to avoid the formation of secondary phases such as ZnS, CuS, and ${\text{Cu}}_2{\text{SnS}}_3$. (ii) The intrinsic $p$-type conductivity is attributed to the ${\text{Cu}}_{\text{Zn}}$ antisite which has a lower formation energy and relatively deeper acceptor level compared to the Cu vacancy. (iii) The low formation energy of many of the acceptor defects will lead to the intrinsic $p$-type character, i.e., $n$-type doping is very difficult in this system. (iv) The role of electrically neutral defect complexes is predicted to be important, because they have remarkably low formation energies and electronically passivate deep levels in the band gap. For example, $\left[{\text{Cu}}_{\text{Zn}}^{-}+{\text{Zn}}_{\text{Cu}}^{+}\right]$, $\left[V_{\text{Cu}}^{-}+{\text{Zn}}_{\text{Cu}}^{+}\right]$, and $\left[{\text{Zn}}_{\text{Sn}}^{2-}+2{\text{Zn}}_{\text{Cu}}^{+}\right]$ may form easily in nonstoichiometric samples. The band alignment between ${\text{Cu}}_2{\text{ZnSnS}}_4$, ${\text{CuInSe}}_2$ and the solar-cell window layer CdS has also been calculated, revealing that a type-II band alignment exists for the $\text{CdS}/{\text{Cu}}_2{\text{ZnSnS}}_4$ heterojunction. The fundamental differences between CZTS and ${\text{CuInSe}}_2$ for use in thin-film photovoltaics are discussed. The results are expected to be relevant to other ${\text{I}}_2{\text{-II-IV-VI}}_4$ semiconductors.

• Received 8 April 2010

DOI:https://doi.org/10.1103/PhysRevB.81.245204