Temperature-independent ytterbium valence in YbGaGe

We have determined the ytterbium valence as a function of temperature in the reported near-zero thermal expansion material YbGaGe using x-ray photoemission at various incident photon energies. The Yb $3d$, $4d$, and $4f$ levels, which directly yield the Yb valence, have been measured. Careful analysis enabled the clear separation of surface and bulk contributions. Resonant photoemission at the $4d\text{−}4f$ absorption edge was used to enhance the low contribution of the ${\mathrm{Yb}}^{3+}$ component. Contrary to the initially proposed Yb valence transition, we find no change in the valence from room temperature down to $115\mathrm{K}$.

Figure 1

The valence-band and shallow-core levels of YbGaGe measured at room temperature with a photon energy of $72\mathrm{eV}$.

Figure 2

The valence-band and shallow-core levels of YbGaGe as a function of temperature, measured on-resonance (see text), at $182\mathrm{eV}$. All spectra were taken on the same cleaved surface.

Figure 3

The Yb valence as a function of temperature, as determined by the two methods in the text, using a photon energy of $182\mathrm{eV}$. The dashed lines are zero-slope fits to the data.

Figure 4

Determination of the surface and bulk components of the $4f^{13}$ state for clean and contaminated surfaces. The hatched doublet is the bulk component. The surface component is clearly observed to be substantially broader.

Figure 5

Fit of the $4f^{12}$ component on an oxidized YbGaGe surface. The hatched peak is that which does not grow upon contamination.

Figure 6

A comparison of the valence band of clean YbGaGe with that of a slightly oxidized sample, taken with a photon energy of $182\mathrm{eV}$.

Figure 7

The valence-band and shallow-core levels of YbGaGe as a function of temperature, measured at a photon energy of $1421\mathrm{eV}$.

Figure 8

The two different background subtractions used to determine the ${\mathrm{Yb}}^{2+}$ spectral weight in the valence band when using a photon energy of $1421\mathrm{eV}$. The sample temperature was $158\mathrm{K}$.

Figure 9

The intensities, as a function of temperature, for the $4f^{13}$ and $3d_{5∕2}$ $(2+)$ parts of the spectrum. The error bars for the $4f^{13}$ intensities have been calculated from the variations in the parameters of the background fits and those for the $3d_{5∕2}$ from the variations in the fits of the core level itself. The dashed lines are zero-slope fits to the data.

Figure 10

The Yb $3d_{5∕2}$ core level as a function of temperature, taken with a photon energy of $1650\mathrm{eV}$, shown together with fits to the experimental data.

Figure 11

The Yb $4d$ core level taken at a temperature of $300\mathrm{K}$ and a photon energy of $1423\mathrm{eV}$, shown together with spectra of a divalent and a trivalent compound (see text).