Mechanism of ferromagnetic coupling in copper(II)-gadolinium(III) complexes

J Am Chem Soc. 2004 Mar 17;126(10):3321-31. doi: 10.1021/ja030628k.

Abstract

This paper offers the first series of state-of-the-art quantum chemical calculations (CASSCF, CASPT2, MS-CASPT2) and analytical models for the well-known problem of quasi-general ferromagnetic coupling in copper-gadolinium complexes. A system chosen from the chemical report of Costes et al. was taken as prototype. At the CASSCF level, calculated results for the experimental structure reproduced the magnetic coupling constant well (J(calcd)( )()= +7.67 cm(-)(1) vs J(exp)( )()= +7.0 cm(-)(1)). For more insight, the study molecule was further idealized by geometry optimization to C(2)(v)() symmetry. Systematic ab initio computation experiments were designed and performed. Owing to specific problems related to the non-aufbau ground configuration of the [CuL-Gd] complexes, the calculations were conducted in a nonstandard manner. We found that the qualitative mechanism of Kahn, assigned to the electron jump from 3d of Cu(II) to 5d shell of Gd(III), can be presented effectively as the cause of the phenomenon, if CASPT2 MOs are taken as magnetic orbitals. We showed that the ferromagnetic coupling is also matched and magnified by spin polarization effects over the ligand, in line with the early assumption of Gatteschi. To be distinguished from the initial hypothesis of Gatteschi, which assumed the role of 6s AO of Gd(III), we found that one 5d-type AO is actually involved in the polarization scheme. In fact, the Gatteschi and Kahn mechanisms are not mutually contradictory, but are even interconvertible with appropriate changes of the magnetic orbitals. Within C(2)(v)() symmetry of complexes, the ferromagnetic coupling can be qualitatively regarded as the preponderant influence of interaction channels exhibiting orbital orthogonality (four 3d-4f contacts) over the nonorthogonal ones (two 3d-4f contacts). The effective preponderance from ferromagnetic pathways is supported by CASPT2 results. One may explain the generality of Cu(II)-Gd(III) ferromagnetic coupling as being correlated with the large occurrence of approximate pseudo-C(2)(v)() geometry of complexes. The observed orbital regularity is lost in lower symmetries. Thus, the antiferromagnetic exceptions occur when the molecular asymmetry is advanced (e.g., owing to strong chemical nonequivalence of the donor atoms).