A step toward correct interpretation of XPS results in metal oxides: A case study aided by first-principles method in ZnO

Metal oxide semiconductors constitute a vast group of materials whose physical properties are greatly affected by native defects. For decades, x-ray photoelectron spectroscopy (XPS) has been widely used in defect analysis. However, correct interpretation of XPS results remains a difficult task. In this work, we present a detailed first-principles study on the core-level shift of the most stable and commonly cited crystal imperfections in ZnO, including O and -OH species at the surface with different coverages and bulk defects, including O interstitial (O-i), O vacancy in the +2 charge state (Vo(2+)), and the neutral vacancy (Vo(0)). The O1s core level spectrum is simulated and compared with experiments to understand the correlation between local atomic structures and features in the O1s spectrum. In particular, our results indicate that the widely adopted assignment in the defect analysis of ZnO, which links the defect peak in XPS to Vo, the most stable defect, is very likely a misinterpretation. Theoretical analysis indicates that there are no distinguishable XPS features arising from the Vo defect. Furthermore, we show that the commonly observed defect-related peak instead arises due to O-i or specific surface configurations. Given the importance of native defects in materials performance, misinterpretation of XPS results may lead to erroneous conclusions regarding materials properties. This work provides a first-principles basis for the analysis of oxide defects through XPS.

Automated summary

Metal oxide semiconductors are greatly influenced by native defects, making the interpretation of X-ray photoelectron spectroscopy (XPS) results challenging. In this study, a first-principles analysis was conducted on the core-level shift of various crystal imperfections in ZnO. The results suggest that the commonly adopted assignment of XPS defect peaks to the most stable defect (Vo) in ZnO is likely a misinterpretation, and instead, the defect-related peak arises from O-i or specific surface configurations. The misinterpretation of XPS results may lead to erroneous conclusions about material properties.