Ceria (CeO₂), as a typical rare earth metal oxide, has been widely applied in the CO oxidation and other catalytic reactions due to its excellent oxygen storage/release capacity, reversible Ce⁴⁺/Ce³⁺ redox properties and abundant surface defect structures^[1−4]. Studies have indicated that the activity of CeO₂-based catalysts in CO oxidation can be significantly enhanced by adjusting the exposed crystal facets^[5−6], controlling the surface oxygen vacancy concentration^[7] and finely designing the local coordination environment of active sites^[8−9]. In recent years, transition metals in single-atom doped CeO₂ catalysts have exhibited excellent catalytic performance^[10−12]. The incorporation of single atoms into the CeO₂ lattice, where individual metal atoms substitute the surface Ce sites, is expected to retain the inherent redox properties of CeO₂ while tuning the electronic structure of active sites through the electronic and geometric effects of the dopant atoms, providing new solutions for achieving efficient low-temperature CO oxidation^[13−16]. For example, theoretical calculations have shown that the Cu center in the Cu₁/CeO₂(111) single-atom catalyst (SAC) reversibly switches between Cu(I) and Cu(II) states; such a change is not random but precisely matches the reaction steps, thereby simultaneously optimizing reactant activation and product desorption, breaking the traditional “scaling relation” limitation in catalysis^[17]. However, systematic computational studies of single-atom doped CeO₂ catalysts (M-CeO₂) in CO oxidation remain relatively scattered, lacking unified data standards and high-throughput screening frameworks. Therefore, constructing a computational data system for the CeO₂-based SACs to systematically evaluate the stable configurations, electronic structures and CO oxidation reaction activities of different single metal atoms on the CeO₂ surface will not only help fill the gap in systematic theoretical data but also provide a reliable computational basis for revealing the catalytic mechanisms of single-atom doped CeO₂.

In this study, 20 transition metals elements spanning the 3d to 5d series are selected, including Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Y, Zr, Mo, Ru, Rh, Pd, Ag, Ta, W, Re, Pt and Au. The dopant atoms directly substitute surface Ce atoms of CeO₂(111), forming M-CeO₂ models with the M−O−Ce local coordination structures. This broad elemental selection aims to systematically investigate the regulatory effects of dopant atoms from different periods and with different electronic configurations on the catalytic performance of M-CeO₂(111) surface in the CO oxidation, providing a theoretical basis to reveal the intrinsic activity of dopant elements and rationally design efficient CeO₂-based catalysts for CO oxidation.

Hence, a specialized database was established for the single-atom doped CeO₂ catalysts, which will provide a solid theoretical foundation for the screening and preparation of efficient catalysts in the CO oxidation. In addition, this database can also serve as an example of doped material databases, which may significantly promote the research progress in the ceria-based catalysts and related fields.

This database covers calculation parameters, structural information, software environment details and electronic properties for the single-atom doped CeO₂(111) catalysts in the CO oxidation. All calculation files are stored by catalyst type and organized in a unified, standardized directory structure. The Ti-CeO₂ system is used as an example to illustrate the file storage structure (Figures 2−6). All data files follow a consistent two-level folder naming convention: Level-1 folders are named by configuration type (Type 1, Type 2-1, Type 2-2, Type 2-3, Type 2-4), whereas Level-2 folders by single-atom metal type (e.g., CeO₂, Ti-CeO₂, Y-CeO₂, Zr-CeO₂ and Mo-CeO₂). The data contents are specified as follows:

(1) VASP calculation data (Figure 2). The folder “DFT calculation data of M-CeO₂ by VASP” contains calculation folders for 21 catalysts, covering the catalyst model geometry optimization and energy data (slab), reactant/intermediate/product adsorption structures and energy data, and TS search calculation results.

(2) Frequency analysis data (Figure 3). The folder “Frequency analysis data of M-CeO₂ by VASP” contains frequency analysis results for the elementary reactions on 21 stable catalyst structures, covering the IS, TS and FS structures for each reaction.

(3) Bader charge data (Figure 4). The folder “Bader charge data of M-CeO₂ by VASP” contains the Bader charge calculation and data processing results for 21 stable catalyst structures.

(4) COHP data (Figure 5). The folder “COHP data of M-CeO₂ by VASP” contains the COHP bond strength calculation and data processing results for 21 stable catalyst structures.

(5) DOS data (Figure 6). The folder “DOS data of M-CeO₂ by VASP” contains the DOS calculation and data processing results for 21 stable catalyst structures.

To ensure the stability and reliability of the calculated data, in the process of constructing the data set, this research has established strict convergence tests for key calculation parameters and results, ensuring that they all meet the strict convergence standards. All files have passed the convergence criteria specified in Section 3.1.

This dataset systematically collects the complete input and output files for the DFT structural optimization, transition state searches and electronic property calculations (including DOS, COHP and Bader charge) of pristine CeO₂(111) and 20 types of metal-doped M-CeO₂(111) catalysts for the CO oxidation reaction. The initial computational configurations and output files provided in this dataset hold significant value for in-depth structural analysis and subsequent research. Users can directly access the raw data for result validation or data mining, enabling the reproduction of key reaction steps or the extraction of specific thermodynamic and kinetic parameters. Moreover, the comprehensive calculation files can serve as a foundation for extended investigations, facilitating advanced analyses such as barrier corrections and density of states interpretation.

For instance, researchers can directly invoke the frequency analysis data from the “Frequency analysis data of M-CeO₂ by VASP.zip” folder, and combine it with tools such as VASPKIT to perform free energy corrections under specific reaction temperatures, thereby accurately assessing the thermodynamic feasibility and kinetic barriers under reaction conditions. By retrieving the activation barriers and reaction energies of elementary reactions from the “DFT calculation data of M-CeO₂ by VASP.zip” folder, promising M-CeO₂ catalysts with high activity and superior catalytic performance potential for low-temperature CO oxidation can be rapidly screened, providing theoretical guidance for the experimental synthesis and effectively circumventing the limitations of the traditional trial-and-error approach for its long cycle times and high costs. Furthermore, by extracting the geometric structural features of stable catalyst configurations, the Bader charge transfer distribution in the active region, and electronic structure parameters such as bond lengths from the COHP database, researchers can systematically mine descriptors capable of accurately predicting the intrinsic activity of catalysts, thereby enabling efficient screening and rational design of catalysts.

The files in this database are available for direct download and are compatible with computational formats used by Materials Studio and VASP software. Due to copyright restrictions, pseudopotential files are not included in this database.

All first-principles calculation data (with no external experimental data were used for comparison) presented in this paper—including structure optimization, frequency analyses, Bader charges analyses, COHP and DOS calculations for pure CeO₂(111) and 20 M-CeO₂(111)—are stored in the Science Data Bank (SDB) operated by the Computer Network Information Center, Chinese Academy of Sciences. The dataset is titled “Database of CO oxidation over single-atom doped CeO₂ catalysts” and is assigned with the following identifiers: Database DOI: https://doi.org/10.57760/sciencedb.35455; Database CSTR: https://cstr.cn/31253.11.sciencedb.35455. The dataset is accessible for download upon request via the platform and may be used for subsequent research.

Example 1 (Data retrieval). Researchers can retrieve the target data by following the directory structure illustrated in Figures 2−6. For example, to locate the structural and energetic data for the stable pristine surface configuration of the Pt-CeO₂(111) system: First, select the folder “DFT calculation data of M-CeO₂ by VASP” within the compressed package; Second, following the directory hierarchy shown in Figure 2, navigate successively into the Pt-CeO₂ subfolder corresponding to the metal-doped system; Third, further locate the Pt-CeO₂ folder within the “slab” directory; Finally, invoke the “CONTCAR” file from this directory to obtain the optimized atomic coordinates and cell parameters and extract the total energy from the “OUTCAR” file.

Example 2 (Machine learning data support). Researchers can systematically invoke the geometric structural descriptors, energy data, Bader charge transfer data, and relevant feature values from the COHP and DOS folders for the 20 types of M-CeO₂ catalysts in the database, serving as the input feature set for machine learning models. Based on these multi-dimensional data, one can not only analyze the microscopic origins of coordination environment differences among various M-CeO₂ catalysts, revealing the intrinsic correlations among metal dopant type, local coordination structure, and electronic properties, but also screen and construct a descriptor system capable of accurately predicting the catalytic activity for low-temperature CO oxidation, such as intrinsic parameters including the Bader charge of the metal center, the d-band center, or specific bond lengths. Furthermore, by leveraging the trained machine learning model, rapid prediction and preliminary screening of the catalytic performance of other single-atom-modified CeO₂ catalysts can be achieved, providing a data-driven theoretical support for the rational design of novel catalysts of high-efficiency and effectively shortening the experimental exploration cycle.

To address the problems that systematic computation studies of M-CeO₂ in the low-temperature CO oxidation are relatively scattered and lack unified data standards, this study employs first-principles calculations to construct a database for the M-CeO₂ single-atom catalysts, systematically revealing the differences in catalytic performance of single-atom types and doping structures in various elementary reaction steps. This database can not only be used to evaluate the catalytic activity in the CO oxidation by referencing reaction energies, but also to establish structure-performance relationships among “dopant metal—electronic properties—catalytic activity” by correlating reaction energies with electronic properties. With the aid of COHP, DOS and Bader charge analyses, the mechanisms underlying performance differences are explained at the electronic level, providing theoretical basis for the rational design of efficient in the catalysts CO oxidation. This work demonstrates the feasibility of generating well-structured, information-complete and mechanism-explicit domain-specific datasets through large-scale calculations. Furthermore, this scientific database will be continuously enriched and optimized through experimental comparison and validation, which helps to promote the research of low-temperature catalytic CO oxidation toward a new data-driven paradigm.