A new theoretical framework was developed to investigate the electronic and defect properties of ceria (CeO₂) by integrating shell-model interatomic potentials, hybrid quantum mechanics/molecular mechanics (QM/MM) embedded-cluster approach, density functional theory (DFT) calculations, and Monte Carlo simulations. This integrated approach allowed clear differentiation between bulk, surface, and environmental contributions to the absolute band edge positions, thus resolving persistent experimental ambiguities for CeO₂ and other metal oxides. Comprehensive agreement among classical mechanical models, quantum mechanical calculations, and experimental measurements establishes a robust foundation for future research and design of advanced materials in catalysis, optoelectronics, and energy conversion applications.

Figure 1. Combined theoretical approaches for studying defect and electronic properties of cerium oxides.

CeO₂ is a technologically important metal oxide widely used in catalysis and ionic conduction. The complex defect and surface chemistry in ceria result in fascinating properties such as non-stoichiometry, polaron formation, and variable band edge positions, which are strongly related to its technological applications. On ARCHER2 and Young, we combined several state-of-the-art computational techniques, including density functional theory (DFT) calculation, shell-model interatomic potential, the hybrid quantum mechanics/molecular mechanics (QM/MM) embedded-cluster approach, and large-scale Monte Carlo simulations, specifically targeting understudied aspects of ceria with substantial relevance to its technological applications.

Atomistic simulations using IAPs have long been playing an active role in disentangling the mysteries in materials science, especially for understanding complex materials and interfaces. In the first part of our study, we developed a robust set of shell-model interatomic potentials capable of accurately reproducing the structural, elastic, dielectric, phonon, defect, thermodynamic, and surface properties of CeO₂, as validated against experiment and high-level DFT results. This set of potentials enables accurate modelling of intrinsic charged point defects in CeO­₂ using the MM Mott-Littleton approach, generating results in good agreement with QM/MM calculations with hybrid DFT functionals. It can also be extended to model complex systems, including reduced and doped ceria, nanoparticles, and reconstructed surfaces.

A critical issue in ceria research involves significant discrepancies among experimental measurements of fundamental electronic properties such as ionisation potential (IP), electron affinity (EA), work function (Φ), and band alignment with other materials. Experimental measurements of these quantities show substantial discrepancies, while theoretical studies further raise ambiguities regarding the mixing concepts of bulk and surface IPs. We provided a theoretical framework to understand the origins of uncertainty in experimental measurements by separating the pure bulk and surface contributions to the IPs of metal oxides (Figure 2a). By isolating bulk and surface contributions using shell-model-based electrostatic analyses and QM models, we determined the intrinsic bulk IP of stoichiometric CeO₂ to be 5.38 eV, whereas surface-dependent IPs obtained via periodic slab calculations ranged extensively between 4.2 and 8.2 eV. These variations arise from orientation-dependent stacking sequences, polarisation effects, and atomic relaxation, applicable not only to CeO₂ but also to other metal oxides, particularly those with high dielectric constants such as TiO₂, ZrO₂, and HfO₂.

Furthermore, we integrated theoretical predictions with experimental observations to elucidate how environmental conditions affect the IP and Φ of ceria (Figure 2b). Despite significant structural and stoichiometric changes during reduction from CeO₂ to Ce₂O₃, the intrinsic bulk IP exhibited only modest variations from 5.38 to 5.0 eV. Instead, experimentally observed variability predominantly originated from surface chemistry-induced shifts in electrostatic potentials. Combined theoretical calculations with experimental XPS/UPS measurements on ceria thin films revealed that oxygen-deficient conditions lower the IP and Φ due to the formation of surface oxygen vacancies and Ce³⁺ polarons, while oxygen-rich environments elevate these parameters primarily via surface peroxide formation. The near-surface defect distribution profoundly affects electronic properties, resulting in IP and Φ fluctuations of approximately 1 eV even under consistent defect concentrations. Additional variability arises from surface adsorbates and impurities, dependent on their coverage, electronegativity, and molecular orientation. Our work demonstrates that the IP and Φ of metal oxides are surface-related parameters, highly sensitive to material morphology, processing history, and operating conditions.

Figure 2. Understanding bulk, surface, and environmental contributions of band edge positions in metal oxides.

Outcomes:

• New knowledge: Separation of bulk, surface, and environmental contributions to ionisation potentials of metal oxides and their correlation with in-lattice electrostatics.
• New technology: Development of a robust set of shell-model interatomic potentials and hybrid QM/MM models for studying CeO₂ towards a consistent prediction of its defect chemistry.

Papers:

• Zhang X. et al., Chem. Mater., 2022, 35(1): 207-227.
• Zhang X. et al., Angew. Chem. Int. Ed., 2023, 135(40): e202308411.
• Zhang X. et al., J. Am. Chem. Soc, 2024, 146, 24, 16814–16829.

Collaborators:

Xingfan Zhang,* Lei Zhu, Taifeng Liu, Jingcheng Guan, You Lu, Thomas W. Keal, John Buckeridge, Qing Hou, Xu Zhang, Shijia Sun, Christopher Blackman, Robert G. Palgrave, Sobia Ashraf, Avishek Dey, Matthew O. Blunt, C. Richard A. Catlow,* and Alexey A. Sokol*

Funding:

• U.K.’s HEC Materials Chemistry Consortium funded by EPSRC (Grant Nos. EP/P020194 (£4,000,000), EP/T022213 (£5,710,207), and EP/R029431 (£489,315)).
• EPSRC (Grant Nos. EP/W014580 (834,868), EP/W014378 (£682,674), EP/R001847 (£1,022,120), EP/K038419 (£474,548), and EP/I030662 (£474,548))

Contacts:

Prof. Richard Catlow: c.r.a.catlow@ucl.ac.uk

Dr. Alexey A. Sokol: a.sokol@ucl.ac.uk

Dr. Xingfan Zhang: xingfan.zhang.20@ucl.ac.uk