Insight on Hole-Hole Interaction and Magnetic Order from Dichroic Auger-Photoelectron Coincidence Spectra

The absence of sharp structures in the Auger line shapes of partially filled bands has severely limited the use of electron spectroscopy in magnetic crystals and other correlated materials. By a novel interplay of experimental and theoretical techniques we achieve a combined understanding of the photoelectron, Auger, and Auger-photoelectron coincidence spectra (APECS) of the antiferromagnetic CoO. A recently discovered dichroic effect in angle resolved (DEAR) APECS reveals a complex pattern in the Auger line shape, which is here explained in detail, labeling the final states by their total spin. Since the dichroic effect exists in the antiferromagnetic state but vanishes at the Néel temperature, the DEAR-APECS technique detects the phase transition from its local effects, thus providing a unique tool to observe and understand magnetic correlations where the usual methods are not applicable.

Figure 1

Comparison of experimental and calculated valence-band XPS profiles. Left panel: the spectrum excited by 1253.6 eV x rays from Ref. 24 (dashed line) and our calculated Co-removal spectrum (solid line). Right panel: the spectrum excited by 250 eV x rays from Ref. 25 (dashed line) and our calculated ($0.8\mathrm{Co}+0.2\mathrm{O}$)-removal spectrum (solid line). In the calculated spectra a 1 eV Lorentzian broadening was used.

Figure 2

Top panel: Experimental DEAR-APECS spectra of CoO vs Auger electron kinetic energy for the AN and NN geometry, and their difference (dashed). The continuous AN and NN lines are obtained by best fit procedures based on a representation of the experimental line shapes by six Voigt profiles (see main text). The difference (or asymmetry curve) displays structures labeled by $A\mathrm{\text{−}}E$. Bottom panel: Calculated total Auger spectrum (solid line) and the different contributions corresponding to the spin $S=5/2$ (dot-dashed), $S=3/2$ (dotted), $S=1/2$ (dashed) of the final state. The structures $A\mathrm{\text{−}}E$ are reproduced by the theory. In addition, $C$ has a shoulder at $\sim 43\mathrm{eV}$ which is not evident in the asymmetry curve because it appears in both the $S=5/2$ and $S=3/2$ spectra, but is clearly seen in both the $AN$ and $NN$ data.