C 1s photoemission and shake-up features of (C59N)2
Schulte, Karina1; Wang, Li2; Moriarty , Philip1; Prassides, Kosmas1; Tagmatarchis, Nikos3
1United Kingdom;
2Singapore;
3Greece

The archetypal fullerene C60 can be placed within the class of correlated electron systems on account of its relatively large ratio of U, the Coulomb interaction between electrons in the valence band, and the bandwidth W of U/W ~ 3-4, comparable to the values for high Tc cuprates. Even upon electron doping the picture continues to hold, as has been studied for three phases of KxC60, where U approximately has a value of 1.5 eV for all three compounds, but activation of intermolecular charge-transfer screening reduces the measured energy value in K3C60 to 0.6 eV. One might then ask how these fundamental processes and parameters are influenced by different forms of doping, such as in endohedral fullerenes or substitutionally doped C59N.
With this in mind we have conducted a spectroscopic study of C59N, or azafullerene. In this molecule one carbon atom is replaced by a nitrogen atom producing, effectively, a form of n-type doping. This substitution affects the bonding of the carbon cage in such a way that the preferred structure at room-temperature is a (C59N)2 dimer, where an intermolecular sp3-like bond between two carbons neighbouring the nitrogen atoms is formed.
Combining valence band photoemission and Auger spectra to obtain U, we report values ranging from U=1.10 ± 0.10 eV for the highest occupied molecular orbital (HOMO) to 1.35 ± 0.10 eV for the deeper lying orbitals, comparable to values found in C60. The on-site Coulomb interaction between a carbon core hole and valence electrons, Uc, is, however, substantially lower than in C60 at 1.35 ± 0.07 eV. Resonant photoemission (RESPES) results show a weakened participator decay channel, especially around the N 1s threshold, where resonance of the highest occupied molecular orbital shoulder is entirely absent. Near-edge X-ray absorption fine structure (NEXAFS) and constant initial state (CIS) measurements, taken in parallel with the RESPES data, indicate, however, that matrix element effects cannot be ruled out.
back