Insulator-metal transition in TiGePt: A combined photoelectron spectroscopy, x-ray absorption spectroscopy, and band structure study

We present a combined experimental and theoretical study of the electronic structure of the intermetallic compound TiGePt by means of photoelectron spectroscopy, x-ray absorption spectroscopy, and full potential band structure calculations. It was recently shown [S.-V. Ackerbauer et al., Chemistry - a European Journal 18, 6272 (2012)] that TiGePt undergoes a structural phase transition by heating which is accompanied by a large volume contraction and a drastic change of physical properties, in particular a large decrease of the electrical resistivity. The present study revealed substantial differences in the electronic structure for the two TiGePt modifications, although they have the same nominal composition and show similar electron counts for particular valence band states. Our photoemission experiments and band structure calculations establish that an insulator-to-metal transition occurs with an appreciable band broadening and closing of the band gap.

Figure 1

Crystal structures of the LT phase (cubic MgAgAs-type, left) and the HT phase (orthorhombic TiNiSi-type, right) of TiGePt.[1] Ti, Ge, and Pt atoms are shown as white, blue, and gray spheres, respectively.

Figure 2

Total and partial density of states (DOS) from fully relativistic electronic structure calculations of TiGePt: the results for the low-temperature (LT) phase are shown in the bottom panel and the high-temperature (HT) phase in the top panel. The common vertical dashed line indicates the position of the Fermi level.

Figure 3

Valence band spectra of the low-temperature (LT) phase (bottom panel) and the high-temperature (HT) phase of TiGePt (top panel) in comparison with the broadened and photoionization-cross-section weighted partial DOS. The spectra were taken with 700 eV photon energy.

Figure 4

Pt 4$f$ core level photoemission spectra of TiGePt in the low-temperature (LT, top, blue solid line) and high-temperature (HT, top, red dashed line) phase, and of elemental Pt metal (bottom, black solid line). The spectra were taken with 700 eV photon energy. Solid vertical lines represent the peak positions of the $4f_{7/2}$ levels; dashed vertical lines represent the center of gravity positions (see text).

Figure 5

X-ray absorption spectra of the Ti $L_{2,3}$-edge of TiGePt in the low-temperature (LT) and high-temperature (HT) phase in comparison with Ti${}_2$O${}_3$ and SrTi${}_2$O${}_3$ as references for Ti $d^1$ and Ti $d^0$, respectively.

Figure 6

X-ray absorption (XAS) spectra at the Ge $K$ edge for TiGePt in the low-temperature (LT) and high-temperature (HT) phase (black solid lines with experimental points), together with the data for the reference system Ge (blue dashed lines with experimental points) and with the calculated XAS spectra (red solid lined). The experimental spectra were normalized using a standard method as implemented in the Athena program (Ref. 30). The position of the absorption edges determined by taking the maximum in the first derivative of the normalized spectra is indicated by vertical dashed lines. The theoretical curves were scaled to match the maximum in the near edge XAS region of the experimental spectra.

Figure 7

Density of states and photoemission spectra near the Fermi level of the low-temperature (LT, left panel) and high-temperature (HT, right panel) phase of TiGePt. Top curves: density of states; middle curves: density of states multiplied by the Fermi distribution function at 300 K; bottom curves: photoemission spectra taken using 190 eV photons (dots) and density of states multiplied by the Fermi function and broadened by the experimental resolution.