Lanthanoid ions have been a recent target for single-molecule magnets (SMMs) with several examples found for all the lanthanoid Kramers’ ions except for Yb(III). This ion has only produced SMMs that function while under an applied magnetic field, a significant deficiency compared to the other Kramers’ ions.1 The cause of this discrepancy is absent in the literature, despite its theoretical potential for high performing SMMs (large crystal field splitting and easily stabilized mJ = ±7/2 ground state). We have, to date, identified a systematic underperformance of Yb(III) compared to Er(III) in terms of SMM properties. Basis set variations were performed which produced minimal changes in computational output, falling well short of closing the gap between Yb(III) and Er(III).2 We shift focus from model to real world computations and tackle the inaccuracy of Complete Active Space Self-Consistent Field (CASSCF) computations of the ground J-multiplet. While CASSCF produces excellent predictions for other lanthanoid ions,3 it underpredicts crystal field splitting by approximately 50% for Yb(III) and is consistent across a range of Yb(III) complexes. It was found that incorporating dynamic electron correlations using Second-Order Complete Active Space Perturbation Theory (CASPT2) produces almost exact agreement with crystal field splitting energies obtained from luminescence measurements (Figure 1). This was compared with inelastic neutron scattering (INS) and electron paramagnetic resonance spectroscopic data, which validated both the higher energy splitting and accurate prediction of ground state anisotropy. We outline the necessity to employ CASPT2 calculations for all future computational studies of Yb(III) coordination complexes regardless of the field of research.
(1) M. J. Giansiracusa, G.K. Gransbury, N. F. Chilton, D. P. Mills, Single-Molecule Magnets. In: Encyclopedia of Inorganic and Bioinorganic Chemistry, 2021, 1–21.
(2) V. K. Fagundes, E. Dias Viegas, M. J. Giansiracusa, submitted
(3) L. Ungur, L. F. Chibotaru, Chem. Eur. J. 2017, 23, 3708–3718.