First-Order Metal-Insulator Transitions in Manganites: Are They Universal?

Conductivity data for ${\mathrm{La}}_{2-2x}{\mathrm{Sr}}_{1+2x}{\mathrm{Mn}}_2{\mathrm{O}}_7$ ($x=0.6$) show a first-order transition from an orbital- or charge-ordered insulator to a metal as the temperature falls below $\sim 160\mathrm{K}$. The change in conductivity is 100 times larger than that seen previously in any single-phase manganite in zero field. The metallic low-temperature state is similar to $x=0.58$, but $x=0.58$ shows no evidence of orbital or charge order. This result supports a conclusion that strongly coupled magnetic-conductive transitions are first order.

Figure 1

(c) Temperature dependence of ${\sigma }_{ab}$ and ${\sigma }_c$ for $x=0.6$ showing transitions at $\sim 160\mathrm{K}$ from a low-temperature $ab$-plane metal into a charge or orbital ordered insulator (with hysteresis of $\sim 5\mathrm{K}$) and at $\sim 250\mathrm{K}$ into a high-temperature paramagnetic insulator. As explained in the text, ${\sigma }_c$ comes from method B (b) while ${\sigma }_{ab}$ also requires the top voltage of method A (a), and these results have been averaged over current configurations. For comparison, data for $x=0.36$ show only one transition at $\sim 125\mathrm{K}$ from a low-temperature 3D metal into a high-temperature paramagnetic insulator. Inset: magnetization at a field of 3 T for the same $x=0.6$ crystal (A) used for conductivity and x-ray scattering and a second crystal (B) with the same nominal composition.

Figure 2

Conductivity results are shown for two principal current configurations for methods A and B. The peaks and dips near the transition at $\sim 160\mathrm{K}$ (most readily seen by method B data) are most likely due to slightly different $T_C$ values along the length and through the thickness of the crystal. The integrated intensity of BIS superlattice reflection for the same crystal as used for conductivity is shown as large solid circles. Inset: x-ray CCD images near the $(2,0,0)$ peak showing the presence of the $(2.21,0.21,0)$ superlattice peak at 185 K due to orbital ordering and its absence at 145 and 300 K.

Figure 3

Temperature dependence of ${\sigma }_{ab}$ and ${\sigma }_c$ for $x=0.6$ showing an approximately linear dependence of ${\sigma }_c$ that is consistent with spin-wave fluctuations spoiling the perfect AFM alignment between neighboring planes of the bilayer. For $x=0.36$ and 0.6, ${\sigma }_{ab}$ extrapolates to a finite value at $T=0$, but the data of Ref. 14 for nominal composition $x=0.5$ do not extrapolate to a constant and are orders of magnitude smaller.