Data-Driven Discovery of Structure–Reactivity Relationships in Oxidative Addition at Zerovalent Nickel Centers

Summary

Combines MACE machine-learning-potential thermochemistry of bipyridine–Ni–alkyne complexes and nickelacycle products, IR and DFT validation, and equivariant graph neural networks with GNNExplainer to identify the structure–reactivity relationships that govern CO₂ insertion and regioselectivity in Ni-catalyzed reductive carboxylation.

Keywords

nickel catalysis CO2 insertion reductive carboxylation machine-learning interatomic potentials MACE graph neural networks interpretable machine learning high-throughput thermochemistry

Cite

@article{2026_elliott_nickel_oxidative_addition_ml,
  title = {Data-Driven Discovery of Structure–Reactivity Relationships in Oxidative Addition at Zerovalent Nickel Centers},
  author = {Sarah N. Elliott and Rosmi Reji and Qingzhi Chen and Xin Zheng and Simah Riaz and Xiaoli Yan and Thang Pham and Eliu Huerta and Ksenija D. Glusac and Murat Keçeli},
  journal = {ChemRxiv preprint (2026)},
  year = {2026},
  doi = {10.26434/chemrxiv.15004105/v2}
}