Doping and temperature dependence of the Kohn anomaly in the phonon dispersion of graphite

Correlation effects and the electron-phonon coupling are at the heart of modern condensed matter physics and the key for understanding the optical and transport properties which also depend strongly on doping.Carbon materials provide an ideal system to probe the doping dependence of the physical properties in the regime of both, electron and hole doping. As a consequence, doping has been extensively applied to graphite, fullerenes, carbon nanotubes and recently also to graphene. Especially the field of doping graphite [i.e. graphite intercalation compounds (GICs)] has been intensively studied for over three decades[1]. The recent revival of graphite [2] and GICs is also strongly motivated by the fact that stage-I GICshave an electronic band structure most similar to a doped monolayer of graphene which was proven by angle-resolved photoemission spectroscopy (ARPES) for a potassium doped GIC [3]. The electron-phonon interactions in these doped graphene layers are also evident from the ARPES spectra as kinks in the quasiparticle dispersion relations [3]. Much less is known on the corresponding phonon branches and their intercalant and staging dependence. A Kohn anomaly in the TO phonon branch has been found in recent inelastic x-ray scattering (IXS) experiments on pristine graphite [4] which agrees with the observed kink position measured in ARPES of stage-I GICs. Furthermore, IXS experiments on alkali(ne) metal GICs are currently being prepared by several groups. These experiments will shine further light on the questions related to the superconducting pairing mechanism in alkali(ne) metal doped GICs. However, a comprehensive understanding of how the intercalant atom type and the staging order affects the interacting electron-phonon system is still missing. Theory suggests that the Kohn anomaly becomes weaker upon doping due to better screening but there are no experiments to directly verify this claim. This holds in
particular for hole doped systems. Due to fewer intercalants and experimental difficulties, there are much less experimental data than for the electron doped system even though the involved physics is extremely challenging.