Metastable ordered states induced by low-temperature annealing of $\delta -{\mathrm{Ag}}_{2/3}{\mathrm{V}}_2{\mathrm{O}}_5$

In $\delta \text{−}{\mathrm{Ag}}_{2/3}{\mathrm{V}}_2{\mathrm{O}}_5$ with charge degrees of freedom in V, it is known that the charge ordering state and physical properties of V that appear at low temperatures depend strongly on the ordering state of Ag. In this study, we focus on the Ag ions in the interlayer and study the structure using synchrotron radiation powder diffraction depending on temperature. We find that when the sample is slowly cooled from room temperature and ordering occurs at the Ag sites, ${\mathrm{V}}^{4+}/{\mathrm{V}}^{5+}$ charge ordering of V and subsequent ${\mathrm{V}}^{4+}\text{−}{\mathrm{V}}^{4+}$ structural dimers are produced. Although quenching the sample from room temperature suppresses the ordering of Ag, annealing at around 160 K promotes partial ordering of Ag and allows a metastable phase to be realized. This metastable phase is maintained even when the temperature is lowered again, producing a remarkable change in low-temperature properties. These results indicate that the ordered state of Ag, which is the key to control the charge-ordered state and physical properties, can be controlled by low-temperature annealing. The results of this study may provide a methodology for the realization of metastable states in a wide range of material groups of vanadium compounds, in which competition among various charge-ordered states underlies the physical properties.

Figure 1

(a) Results of Rietveld analysis of synchrotron x-ray diffraction data in the high-temperature phase at 300 K. Data were collected at the BL02B2 beamline at SPring-8, and the x-ray energy was 23 keV. (b) Crystal structure of the high-temperature phase. (c) Temperature dependence of the anisotropic thermal parameters $U_{11}, U_{22}$, and $U_{33}$. The inset shows the thermally oscillating ellipsoid of an Ag atom. (d) Coordination relationship between Ag and oxygen obtained with Rietveld structural analysis of the Ag two-site model. (e) Temperature dependence of the occupancy at the two Ag sites determined in (d), six-coordinated Ag-I and five-coordinated Ag-II.

Figure 2

(a) Top: Color map showing the change in diffraction patterns during the slow-cooling process from room temperature. Bottom: Results of Rietveld analysis of synchrotron x-ray diffraction data in the intermediate-temperature phase of 200 K. Data were collected at the BL5S2 beamline at Aichi SR, and the x-ray energy was 20 keV. (b) Crystal structure of the intermediate- and low-temperature phases and occupancy of three Ag sites. (c) Temperature dependences of the isotropic displacement parameters of the Ag1 and Ag2 sites. The isotropic displacement parameter of the Ag3 site is not shown because the occupancy of the Ag3 site is almost zero.

Figure 3

(a) Temperature variation of synchrotron x-ray diffraction patterns from the intermediate- to low-temperature phases. Data were collected at the BL02B2 beamline at SPring-8, and the x-ray energy was 23 keV. (b) Temperature dependence of six different V-V distances normalized at 220 K in the intermediate-temperature phase. Blue, yellow, and red correspond to the colors of V-V bonds shown in Fig. 2. (c) Results of neutron diffraction experiments performed just below the transition between the low- and intermediate-temperature phases.

Figure 4

(a) Top: Color map showing the change in diffraction patterns during the slow-heating process of the quenched sample. Data were collected at the BL5S2 beamline at Aichi SR, and the x-ray energy was 20 keV. Bottom: Results of Rietveld analysis of synchrotron x-ray diffraction data in the quenched sample at 110 K. (b) Enlarged image of the 010 superstructure peak. (c) Temperature dependences of 010 peaks. (d) Temperature dependence of the x-ray diffraction pattern using high-energy x rays at 67.42 keV. The development of the 011 and $01\overline{1}$ peak intensity is consistent with the 010 peak obtained from measurements at 20 keV, confirming that it is independent of the x-ray energy used. (e) Temperature dependences of 011 and $01\overline{1}$ peaks.

Figure 5

(a)–(d) Time variation of the 010 superlattice peak due to long time annealing. Data were collected at the BL5S2 beamline at Aichi SR, and the x-ray energy was 20 keV. (e) Diffraction patterns at 110 K for slow-cooled samples, annealed samples at various temperatures, and quenched samples. Annealed samples were held at a specific temperature for 1 h. (f) Temperature dependence of magnetic susceptibility of slow-cooled, annealed, and quenched samples during the temperature increase from the lowest temperature to 300 K. Annealed samples were obtained by holding them at a specific temperature for more than 2 h. The measurements were performed under a magnetic field of 1 T during the temperature increase process from 5 to 300 K. The inset shows an example of the evolution of magnetic susceptibility with temperature and time.

Figure 6

(a) Schematic of the relationship between Ag ion sites with atomic displacements in the high-temperature phase. (b) Relationship between the energies of the five- and six-coordination sites, where the six-coordination site has a shorter distance from the neighboring Ag site and is expected to have strong Coulomb repulsion. (c) Pattern of four Ag ion configurations expected to occur in two adjacent Ag site pairs. (d) Schematic of the relationship between Ag ion sites with lost atomic displacement in the intermediate-temperature phase. (e) Patterns of three Ag ion configurations occurring in two adjacent Ag site pairs. The realization of these three patterns is confirmed by Rietveld analysis, and the reason why both six- and five-coordinated sites appear can be understood based on the Coulomb repulsion argument. (f) The temperature dependence of the integral of the background data during the slow-cooling process. (g) The temperature dependence of the integral of the background data during the slow-heating process of the quenched sample.