Atomistic Structure of Cobalt-Phosphate Nanoparticles for Catalytic Water Oxidation

Xiao Hu, Simone Piccinin, Alessandro Laio, Stefano FabrisACS 2012 · DOI

TL;DR

This paper tackles the challenge of understanding the atomic structure of cobalt-phosphate (Co-Pi) nanoparticles, crucial for efficient solar-driven water splitting. By combining advanced computational simulations, they reveal that these amorphous nanoparticles actually contain stable, ordered crystallites composed of cobalt-oxo cubane units and layered structures, which are key to their catalytic activity. This breakthrough provides a realistic structural model for the catalyst's active sites, paving the way for deeper mechanistic studies and improved catalyst design.

What Problem Does This Paper Address?

Solar fuels, produced through processes like water splitting, hold immense promise for sustainable energy. A major bottleneck in this process is the efficient oxidation of water to produce oxygen, protons, and electrons. Cobalt-phosphate (Co-Pi) catalysts, made from abundant elements, have emerged as highly effective for this reaction, operating under mild conditions. However, their amorphous nature and complex composition have made it incredibly difficult to determine the precise atomic arrangement of the active catalytic sites. Previous studies suggested the presence of ordered cobalt-oxo units, like cubanes or layered structures, but a definitive, statistically meaningful model was lacking. This paper aims to bridge this knowledge gap by building and validating a realistic atomistic model of Co-Pi nanoparticles.

Methods

The researchers employed a multi-pronged computational approach:

  • Shell Model Empirical Potential: To handle the large system sizes and complex dynamics, they first developed and fitted an empirical potential based on the shell model. This potential was trained using ab initio DFT-PBE calculations on known cobalt oxides and potassium phosphate to accurately reproduce their energies and structures.
  • Metadynamics Simulations: This enhanced sampling technique was crucial for exploring the complex energy landscape of the Co-Pi nanoparticles. They used collective variables (CVs) to guide the simulations, focusing on:
  • CV1 & CV2: To explore the local atomic environment around cobalt ions, specifically their coordination numbers with oxygen and the formation of Co-O-Co bridges. This helped identify stable amorphous grain structures.
  • CV3: A novel CV designed to specifically promote the formation of cubane motifs, allowing them to generate nanoparticles with a higher fraction of these ordered structures.
  • Ab Initio DFT-PBE Calculations: Structures identified as promising from the metadynamics simulations were further optimized using spin-polarized Density Functional Theory (DFT-PBE) to obtain highly accurate atomic positions and energies.
  • EXAFS Simulation: To validate their structural models against experimental data, they simulated the Extended X-ray Absorption Fine Structure (EXAFS) spectra of their computational nanoparticles using the FEFF code.

Datasets

Not discussed in the paper. The researchers generated their own computational models and compared them to experimental EXAFS data from existing literature.

Key Insights and Ideas

  • Amorphous Exterior, Crystalline Interior: The study reveals that seemingly amorphous Co-Pi nanoparticles are not entirely disordered. Instead, they contain stable, molecular-sized crystallites within the amorphous matrix.
  • Emergence of Ordered Structures: These crystallites are not just random arrangements but are composed of specific, ordered motifs: bis-oxo-bridged cobalt centers forming layered structures (edge-sharing CoO₆ octahedra) and corner- and face-sharing cubane units (C₄O₄).
  • Cubane-Layered Coexistence: Importantly, these layered and cubane motifs are not mutually exclusive; they coexist within the same crystallites, suggesting a complex but stable structural organization.
  • Phosphate Decoration: Disordered phosphate groups are consistently found at the edges of these crystallites, interacting with the cobalt centers. This highlights the crucial role of phosphates in stabilizing the structure and potentially in the catalytic mechanism.
  • Structural Stability: The identified crystallites are remarkably stable, maintaining their structure even at elevated temperatures (up to 900 K), which is essential for a functional catalyst.
  • EXAFS Validation: The simulated EXAFS spectra of nanoparticles containing a significant number of cubane motifs showed excellent agreement with experimental data, providing strong validation for the proposed structural models.

Key Results

  • The computational models successfully predicted the formation of stable crystallites within Co-Pi nanoparticles.
  • These crystallites consist of a combination of layered cobaltite structures and C₄O₄ cubane units.
  • The simulated EXAFS spectra of models with cubane motifs closely matched experimental spectra, confirming the presence of these ordered structures in real Co-Pi catalysts.
  • The study provides atomistic models of the catalyst surface, which are proposed to be reliable representations of the active sites.

Limitations and Future Work

  • Nanoparticle Size: The computational nanoparticles were limited in size (~1 nm). The authors acknowledge that this size limitation might lead to an overestimation of the amorphous region compared to the actual material. Larger crystallites were not stabilized, possibly due to size effects, open boundary conditions, local minima trapping, or inaccuracies in the force field.
  • Future Work: The authors suggest that their validated structural models will be invaluable for future mechanistic studies of water oxidation, allowing for a more quantitative approach to understanding the reaction mechanism and the role of phosphate groups. They also propose that these models can guide the design of improved Co-Pi catalysts.

Summary updated

Abstract

Solar-driven water splitting is a key photochemical reaction that underpins the feasible and sustainable production of solar fuels. An amorphous cobalt-phosphate catalyst (Co-Pi) based on earth-abundant elements has been recently reported to efficiently promote water oxidation to protons and dioxygen, a main bottleneck for the overall process. The structure of this material remains largely unknown. We here exploit ab initio and classical atomistic simulations combined with metadynamics to build a realistic and statistically meaningful model of Co-Pi nanoparticles. We demonstrate the emergence and stability of molecular-size ordered crystallites in nanoparticles initially formed by a disordered Co-O network and phosphate groups. The stable crystallites consist of bis-oxo-bridged Co centers that assemble into layered structures (edge-sharing CoO(6) octahedra) as well as in corner- and face-sharing cubane units. These layered and cubane motifs coexist in the crystallites, which always incorporate disordered phosphate groups at the edges. Our computational nanoparticles, although limited in size to ~1 nm, can contain more than one crystallite and incorporate up to 18 Co centers in the cubane/layered structures. The crystallites are structurally stable up to high temperatures. We simulate the extended X-ray absorption fine structure (EXAFS) of our nanoparticles. Those containing several complete and incomplete cubane motifs-which are believed to be essential for the catalytic activity-display a very good agreement with the experimental EXAFS spectra of Co-Pi grains. We propose that the crystallites in our nanoparticles are reliable structural models of the Co-Pi catalyst surface. They will be useful to reveal the origin of the catalytic efficiency of these novel water-oxidation catalysts.
BibTeX
@inproceedings{hu2012atomistic,
  author = {Xiao Hu and Simone Piccinin and Alessandro Laio and Stefano Fabris},
  title = {{Atomistic Structure of Cobalt-Phosphate Nanoparticles for Catalytic Water Oxidation}},
  booktitle = {ACS},
  year = {2012},
  doi = {10.1021/nn3044325},
}
Figure