Electronic structure of titania-based nanotubes investigated by EELS spectroscopy

We present an electron-energy-loss spectroscopy study of the O-K and $\text{Ti}{L}_{2,3}$ edges for anatase-, rutile-, and titania-based nanotubes. The titania-based tubes are composed of tetravalent titanium ions in an octahedral symmetry with the oxygen ligands, however the electronic structure does not correspond to that of either of the titania precursors. Crystal-field splitting is comparable with anatase but the $3d$ occupation number is closer to that of rutile. In addition, $\text{Ti}{L}_{2,3}$ spectroscopic signatures indicate that the octahedron connectivity of the tubes is different to that of the reference titania. We also report the first measurement at the scale of an individual nanometrical object of the charge-transfer satellites for the transition metal $2p$ edges. This may open new opportunities for the estimation of $d$-orbital occupation numbers for transition-metal or rare-earth elements occurring in local grains, phase-segregated regions, nanomaterials, or in the vicinity of interfaces.

Figure 1

Left, HREM micrograph of the titanate nanotube. The square indicates the spatial resolution of the EELS spectroscopy. Right, EELS O-K edges for anatase-, rutile-, and the titania-based nanotube. Energy range has been set at 530 eV on the onset of the rutile edge. The dotted lines indicate a 2.5 eV separation.

Figure 2

Experimental EELS $\text{Ti}{L}_{2,3}$ main edges for anatase-, rutile-, and the titania-based nanotube. Energy range has been set at 456 eV at the sharper $L_3$ line for rutile.

Figure 3

Simulated EELS $\text{Ti}2p$ spectra for a ${\text{Ti}}^{3+}$ ground state and a ${\text{Ti}}^{4+}$ ground state in tetrahedral and octahedral symmetry. In this example, the crystal-field splitting has been set to $\left|10\text{Dq}\right|=E\left(e_{\text{g}}\right)-E\left(t_{2\text{g}}\right)=1.8\text{eV}$, typical value for transition-metal oxide. An ${\text{Ti}}^{4+}$ ground state in lower symmetry showing “fortunate” agreement with the rutile experimental data is also shown. In the latter case, a crystal-field subsplitting of $\text{Ds}=-90\text{meV}$ and $\text{Dt}=-140\text{meV}$ have been set. Spectra have been aligned in accordance with the experimental energy.

Figure 4

DFT-LDA electronic structure for the anatase and rutile. Only the $\text{Ti}3d$ projected density of state are shown.

Figure 5

Experimental EELS $\text{Ti}{L}_{2,3}$ satellite edges for anatase-, rutile-, and the titania-based nanotube. CTM calculations for two hybridization strengths $V\left(e_{\text{g}}\right)$ are also reproduced. The CTM edges have been rigidly 2.5 eV shifted to the higher energy.