Band dispersion in the deep 1s core level of graphene

Abstract

It is generally assumed that electrons in deep atomic core states are highly localized and do not participate in the bonding of molecules and solids. This implies well-defined core-level binding energies and the absence of any splitting and band-like dispersion, a fact that is exploited in several powerful experimental techniques, such as X-ray photoemission spectroscopy. Violations of this assumption have been found for only a few small molecules in the gas phase such as C 2 H 2 or N 2 with much stronger bonding and shorter bonding distances than present in solids 1,2 . Here we report the observation of a sizeable band-like dispersion of the C 1s core level in graphene, a single-layer honeycomb net of carbon atoms that is attracting considerable attention at present in the scientific community 3 . The dispersion is observed as an emission-angle-dependent binding-energy modulation and it is shown that under appropriate conditions only the bonding or antibonding states can be observed. A very similar dispersion is also found by ab initio calculations.The binding-energy modulations of the C 1s core state are illustrated in Fig. 1b-f. Each panel shows a group of spectra taken at a fixed polar emission angle θ as a function of azimuthal emission angle φ. Clear shifts of the peak position are observed. Figure 1g shows a comparison of the spectrum taken at normal emission and one taken at θ = 25 • together with the result of a peak fit to these two spectra.The binding-energy variation obtained from the peak fitting is given in Fig. 1h with the markers corresponding to the spectra in Fig. 1b-f. Strong changes are evident with the largest difference of binding energies spanning ≈60 meV. The variation is consistent with the point symmetry of the graphene lattice. Figure 1i shows the intensity variations of the peaks, shown as the modulation function. Strong intensity modulations are observed, caused by photoelectron diffraction in the final state. The variations also follow the point symmetry of the graphene lattice, but they do not seem to be correlated with the binding energy in Fig. 1h, with the main structures being in phase for some polar emission angles and out of phase for others.Although the data of Fig. 1 serve to illustrate the nature of the effect and the fitting procedure, they are insufficient to pin down the physical origin of the modulation. Figure 2 therefore shows a much more extensive data set measured over many polar and azimuthal angles and at different photon energies. Again, a good fit to all of the spectra in the data set was obtained for a single C 1s component using always the same line shape. Note that this excludes the existence of unresolved components, because their intensities would modulate differently, changing the shape of the peak. The left panel of the figure shows the resulting intensity modulation function and the right panel gives the binding-energy modulation. The intensity modulation function is compared to the simulation for a flat, free-standing layer of graphene. The agreement b...

Influence of elastic scattering on the measurement of core-level binding energy dispersion in X-ray photoemission spectroscopy

“…This has been demonstrated for gaseous molecules like C 2 H 2 1 and N 2 2 and in recent measurements on graphene by Lizzit et al 3 who were able to determine the bandwidth of the C 1s core level to be 60 meV. The corresponding theoretical prediction from ab initio calculations support the claim, that these orbitals are not completely degenerated.…”

Wet-transfer of CVD-grown graphene onto sulfur-protected W(110)

“…In this case, a shift of the Fermi level relative to the Dirac point in graphene will cause an equivalent shift of the C-1 s binding energy [17]. Note that C-1 s measurements were taken at normal emission and thus the reported dispersion of the C-1 s core level as a function of emission angle of less than 0.05 eV [34] does not contribute to the uncertainty of the measurement. This is schematically shown in Fig.…”

Calculation of the graphene C1score level binding energy

“…For graphene, values found in the literature range from 283.97 eV for graphene on Pt(111) 44,45 , 284.15 on Ir(111) 6,45 , 284.2 eV on Auintercalated Ni(111) 46,47 , 284.47 eV for suspended few-layer graphene 48 , 284.6 eV on hydrogen-intercalated SiC 49 , 284.7 eV on Ni(111) 50 , to 284.8 eV on SiC 5,51 . While it is thus clear that charge transfer from and screening by the substrate affect the measurements significantly, the exact value for freestand- ing single-layer graphene has not been fully established.…”