ELECTRONIC-PROPERTIES OF MXC60 AS SEEN BY NMR AND EPR AND COMPARED TO GRAPHITE-INTERCALATION COMPOUNDS

The knowledge of the wave function near the Fermi level is very important to understand conductivity and superconductivity mechanisms. NMR and ESR are able to describe how the wave function near E(F), shares among the various atomic orbitals. The Knight shift measures the projection of the density of states at E(F) on the orbitals of the observed nucleus. So we prepared under vacuum the samples M(x) C60 with K, Rb(x = 3) and Cs (x = 1,3) which were first characterized by X-rays, and Na9.8C60 with the help of high pressure. In this paper we present the various NMR shifts and lineshapes measured in our lab and concerning C-13 and alkali metals in M(x)C60 and RT ESR. We compare with similar measurements in GICs yet to be published. In M(x)C60 alkali compounds, the density of states at E(F) passes through a maximum at x = 3 which corresponds to the composition with the highest superconductive T(c) in agreement with an e-phonon mechanism: The C-13 NMR spectra reveal essentially a C(2p) state at E(F) with the same range of values for the density of states as for GICs, but with an additional shift resulting from a small C(2s)-hybridization. Variable temperature Cs-133 spectra, and RT ESR of the series K, Rb and Cs, with x = 3 are also compared to GIC ones and show the same residual contribution of alkali s-orbital to the free electron wavefunction near E(F). An axial symmetry is seen by Cs-133 in CsC60 with an important thermal shift. In low temperature prepared Cs3C60, two sites at least are occupied by Cs, with a bad symmetry. We evidenced at 25 K, 8.5 T, the superconductive transition in Cs3C60 prepared at low temperature, and deduce some conclusions relative to the origin of differences in T(c) between GICs and M(x)C60. We conclude that the superconductive mechanism is a classical one, with a good e-coupling to the hard ''in-plane'' longitudinal phonons thanks to the C(2s)-hybridization near E(F). Annealing Cs3C60 at 380-degrees-C destroys the superconductive phase and the stability domain of superconductive Cs3C60 phase needs to be more precisely evaluated.