Circumstantial evidence for hydrogen-induced surface magnetism on Pd(110)

The surface of clean and hydrogen-covered Pd(110) was investigated by angle-resolved photoemission spectroscopy. At a coverage of 1.5 ML, where the surface adapts a pairing-row reconstruction, the spectra show a peculiar peak splitting of the $\overline{S}$-surface resonance. The band dispersion of this satellite is parallel to the dispersion of the main resonance but shifted by 400 meV at $T=100$ K. An interpretation in terms of magnetic exchange splitting is supported by the appearance of a kink structure of the photoemission data around 350 meV binding energy. This unexpected behavior for H adsorption is attributed to the reduced coordination of the Pd atoms at the surface.

Figure 1

Structural investigation of clean Pd(110) [(a), (b)] and the 1.5 ML hydrogen covered system (1$\times$2)-H/Pd(110)-pr (c). The STM image is recorded at liquid N${}_2$ temperature.

Figure 2

(a) Fermi-surface mapping of clean Pd(110); the surface Brillouin zone (SBZ) is indicated by the black rectangle. (b) ARUPS intensity map taken along the horizontal red line in (a). (c) Same cut on the (1$\times$2)-H/Pd(110)-pr system. (d) Cut through $\overline{S}$ in direction $\overline{X}\overline{S}$, perpendicular to the cut in (c). To guide the eye, the dispersion of the main resonance and the satellite is indicated in (c) and (d) by the dashed line (black). Red lines and circles in (b), (c), and (d) correspond to the surface-state bands found in the LDA calculation. Color scale: Blue corresponds to low, white [red in (a)] to high photoemission intensity.

Figure 3

ARUPS dispersion plots taken at 100 K around $\overline{S}$ ($k_x=1.142$ Å${}^{-1}$). The direction of scanning is toward $\overline{Y}$. The dispersion is fitted and compared to the clean and the (1$\times$2)-H/Pd(110)-pr surface [black squares in (a) and (b), respectively]. For the hydrogen-covered system a double-peak structure can be identified (c), which is indicative of electron-magnon coupling. This is not the case for clean Pd. White corresponds to high, blue to low photoemission intensity.

Figure 4

Temperature-dependent evolution of spectra taken at $\overline{S}$. In (a) the raw data of the ARUPS measurements are shown. (b) and (c) show the temperature dependence of the peak position and distance $\Delta E$ between the peaks, as determined by fitting the data of (a). The splitting in (c) shows a critical behavior indicating a phase transition (see text).