Magnetization and spin density in [FeCp* 2] [rad]+: A look-back at the magnetic interactions in the chains [FeCp* 2] [rad]+[TCNE] [rad

Abstract

The discovery of ferromagnetism in the charge transfer salt [FeCp* 2] +[TCNE] − has raised a controversy concerning the nature of the magnetic coupling. The group of J.S. Miller invoked a McConnell II mechanism, with configuration interactions. O. Kahn and his coworkers proposed a McConnell I mechanism, implying a positive exchange due to an overlap between spin densities of opposed signs. In the first interpretation there would be a positive spin density located on the carbon of the Cp* rings, but in the second interpretation, this spin density should be negative. To clarify this mechanism, it was decided to independently determine, by polarized neutron diffraction, the spin density of [TCNE] −, associated with a nonmagnetic donor and the magnetization density of [FeCp* 2] +, associated with a nonmagnetic acceptor. For [TCNE] −, the measurement was straightforward, but for [FeCp* 2] +, we had several failures due to a loss of symmetry on cooling the samples. Finally, we could carry on this measurement on a triclinic crystal where [FeCp* 2] + was associated with a polyoxotungstate anion. On [TCNE] −, most of the spin density is located on the central carbons, with a noticeable amount on the terminal nitrogens. On [FeCp* 2] +, the Fe atoms carry a moment of 2.0 μ B, a combination of both spin and orbital moments, the carbon of the rings carry −0.005±0.001 μ B and, surprisingly, the methyl carbons carry +0.008±0.001 μ B. The signs of the spin populations support the McConnell I, but their magnitude accounts for a part of the experimental intrachain exchange interaction only. Ab initio DFT calculations predict the very small densities on the carbon atoms of [FeCp* 2] +, with the correct signs, but with magnitudes which are three times larger than the experimental ones.