Pairing a Global\nOptimization Algorithm with EXAFS\nto Characterize Lanthanide Structure in Solution

<i>Ensemble</i>-average sampling of structures\nfrom <i>ab\ninitio</i> molecular dynamics (AIMD) simulations can be used\nto predict theoretical extended X-ray absorption fine structure (EXAFS)\nsignals that closely match experimental spectra. However, AIMD simulations\nare time-consuming and resource-intensive, particularly for solvated\nlanthanide ions, which often form multiple nonrigid geometries with\nhigh coordination numbers. To accelerate the characterization of lanthanide\nstructures in solution, we employed the Northwest Potential Energy\nSurface Search Engine (NWPEsSe), an adaptive-learning global optimization\nalgorithm, to efficiently screen first-shell structures. As case studies,\nwe examine two systems: Eu(NO<sub>3</sub>)<sub>3</sub> dissolved in\nacetonitrile with a terpyridine ligand (terpyNO<sub>2</sub>), and\nNd(NO<sub>3</sub>)<sub>3</sub> dissolved in acetonitrile. The theoretical\nspectra for structures identified by NWPEsSe were compared to both\nexperimental and AIMD-derived EXAFS spectra. The NWPEsSe algorithm\nsuccessfully identified the proper solvation structure for both Eu(NO<sub>3</sub>)<sub>3</sub>(terpyNO<sub>2</sub>) and Nd(NO<sub>3</sub>)(acetonitrile)<sub>3</sub>, with the calculated EXAFS signals closely matching the experimental\nspectra for the Eu-ligand complex and showing good similarity for\nthe Nd salt; the better agreement with the ligand-containing structure\nis attributed to a less dynamic coordination environment due to the\nrigid ligand. The key advantage of the global optimization algorithm\nlies in its ability to sample the coordination environment across\nthe potential energy surface and reduce the time required to identify\nstructures from generally a month to within a week. Additionally,\nthis approach is versatile and can be adapted to characterize main-group\nmetal complexes.