Detailed Mapping of the Local ${\mathrm{Ir}}^{4+}$ Dimers through the Metal-Insulator Transitions of ${\mathrm{CuIr}}_2{\mathrm{S}}_4$ Thiospinel by X-Ray Atomic Pair Distribution Function Measurements

The evolution of the short-range structural signature of the ${\mathrm{Ir}}^{4+}$ dimer state in ${\mathrm{CuIr}}_2{\mathrm{S}}_4$ thiospinel has been studied across the metal-insulator phase transitions as the metallic state is induced by temperature, Cr doping, and x-ray fluence. An atomic pair distribution function (PDF) approach reveals that there are no local dimers that survive into the metallic phase when this is invoked by temperature and doping. The PDF shows ${\mathrm{Ir}}^{4+}$ dimers when they exist, regardless of whether or not they are long-range ordered. At 100 K, exposure to a 98 keV x-ray beam melts the long-range dimer order within a few seconds, though the local dimers remain intact. This shows that the metallic state accessed on warming and doping is qualitatively different from the state obtained under x-ray irradiation.

Figure 1

Changes in PDF across the MI transition in ${\mathrm{CuIr}}_2{\mathrm{S}}_4$ when induced by temperature and Cr doping: experimental PDFs of (a) metallic 230 K (red) and insulating 220 K (blue), (b) metallic 230 K (red) and metallic 240 K (blue), (c) insulating undoped (blue) and metallic 5% Cr doped (red) at 200 K temperature. Difference curves are shown offset below (green). The corresponding side panels emphasize the short-range scale: the appearance of excess PDF intensity corresponding to the short ${\mathrm{Ir}}^{4+}\mathrm{\text{−}}{\mathrm{Ir}}^{4+}$ dimer distance around 3.0 Å in the insulating phase is indicated by arrows. The existence of this peak indicates the presence of local dimers. For reference, a sketch of Ir-chains in M and I phases is given on top. See text for details.

Figure 2

Evolution of the short-range dimer structure across the MI transition in ${\mathrm{CuIr}}_2{\mathrm{S}}_4$ induced on cooling (blue) and warming (red) cycles. The insets show the evolution of the 3.0 Å dimer PDF peak on cooling (bottom left) and warming (up right). The main panel shows the excess height of this peak with respect to the 230 K data (see text for details).

Figure 3

Raw 2D diffraction data at 100 K of ${\mathrm{CuIr}}_2{\mathrm{S}}_4$ for successive (a) 0.5, (b) 1.0, (c) 30.0, and (d) 0.5 s exposures to x rays, with 2 min separations between the exposures. Notable is an extra diffraction ring indicated by the arrow in panel (b) present for all exposure times shown, except for 30 s. This corresponds to an unresolved multiplet of SLPs, including ($3/2$, $3/2$, $3/2$), coming from long-range charge and dimer order. Panels (e) and (f) show the 1D powder diffraction pattern obtained by integrating around the rings from panels (a),(d) and (b),(c), respectively. The inset in panel (f) emphasizes the disappearance of the SLP multiplet. (g) Dependence of the normalized SLP intensity vs x-ray exposure time.

Figure 4

Experimental PDFs of ${\mathrm{CuIr}}_2{\mathrm{S}}_4$ at 100 K corresponding to short (250 msec, blue) and long (30 sec, red) exposure times. The difference curve is offset below for clarity. The side panel emphasizes the short-range structure, with the arrow denoting the position of the 3.0 Å PDF peak corresponding to the local dimers. The differences are much smaller than seen in Fig. 1.