Active materials of Nb-doped rutile TiO2 powders were synthesized a typical sol–gel method. By using a gas-deposition method, we prepared Nb-doped TiO2 thick-film electrodes which were consisted only of the active materials and titanium current collectors to clarify an original anode performance of its active materials. We assembled 2032-type coin cells consisted of the thick-film electrodes as working electrode, Na foil as counter electrode, electrolyte, and propylene-based separator. The electrolytes used in this study were 1 M NaClO4-dissolved in propylene carbonate. Galvanostatic charge–discharge tests were carried out using an electrochemical measurement system at 303 K.
As the results of charge–discharge tests, all the electrodes showed potential shoulders in the charge (sodiation) reaction at 0.1–0.7 V vs. Na/Na+ and the discharge (desodiation) reaction at 0.5–1.0 V. However, a Na-insertion/extraction mechanism has never been explored to the best of the authors’ knowledge. To elucidate the mechanism, we performed ex-situ XRD measurements for the electrodes during the charge–discharge reactions. During the charge process, the diffraction peaks of the rutile TiO2 shifted toward lower angles. Although the diffractions significantly weakened at 0.005 V, the (101) and (211) reflections still remained. During the discharge process, the peak positions shifted toward higher angles associated with the increasing diffraction intensities. The change in the patterns revealed that the crystal phase of rutile was basically maintained during the charge–discharge process, and that there are reversible reactions of Na-insertion and Na-extraction into/from crystal lattice of the rutile TiO2. Reversible lattice expansion and contraction were clearly observed. In the tetragonal rutile structure, the lattice parameter c was not changed during the Na-insertion/extraction. In contrast, the lattice parameter a showed about a 1.1% increasing by the Na-insertion. From this result, we firstly revealed the reversible Na-insertion/extraction of rutile TiO2(Fig.1).
Figure 2 shows the cycling performances of the Ti1−xNbxO2 electrodes with various Nb amounts between x = 0 and x = 0.18. All the electrodes showed a gradual increase in the discharge capacities until the 5th–15th cycles. It is suggested that the side reactions were gradually suppressed by the 5th–15th cycles. The discharge capacity increased with x. At x= 0.11 and 0.18, the electrode temporarily showed higher capacities in the initial several cycles and a greater capacity decay after the cycles compared to other electrodes. The capacity decay is presumably caused by the reduced crystalline sizes and the formation of an impurity crystal phase, as already mentioned. Among all the Ti1−xNbxO2s, the electrode with x = 0.06 exhibited the best performance with the reversible capacity of 160 mA h g–1 at the 50th cycle. It is a noteworthy result that our rutile Ti1−xNbxO2 electrode delivered a comparable performance even though its active material layer is a thick film of approximately 14 µm and does not contain any conductive material and binder. These improved performances are mainly attributed to the three orders of magnitude higher electronic conductivity of Ti0.94Nb0.06O2 compared to TiO2. These results suggest that Nb-doped TiO2 is an attractive candidate as a conductive-additive-free active material for the Na-ion battery anode as well as for the Li-ion battery anode.