Nonpercolative nature of the metal-insulator transition and persistence of local Jahn-Teller distortions in the rhombohedral regime of ${\mathrm{La}}_{1-x}{\mathrm{Ca}}_x{\mathrm{MnO}}_3$

Evolution of the average and local crystal structure of Ca-doped ${\mathrm{LaMnO}}_3$ has been studied across the metal to insulator (MI) and the orthorhombic to rhombohedral (OR) structural phase transitions over a broad temperature range for two Ca concentrations $\left(x=0.18,0.22\right)$. Combined Rietveld and high real space resolution atomic pair distribution function (PDF) analysis of neutron total scattering data was carried out with aims of exploring the possibility of nanoscale phase separation (PS) in relation to MI transition, and charting the evolution of local Jahn-Teller (JT) distortion of ${\mathrm{MnO}}_6$ octahedra across the OR transition at $T_S\sim 720$ K. The study utilized explicit two-phase PDF structural modeling, revealing that away from $T_{\text{MI}}$ there is no evidence for nanoscale phase coexistence. The local JT distortions disappear abruptly upon crossing into the metallic regime both with doping and temperature, with only a small temperature-independent signature of quenched disorder being observable at low temperature as compared to ${\mathrm{CaMnO}}_3$. The results hence do not support the percolative scenario for the MI transition in ${\mathrm{La}}_{1-x}{\mathrm{Ca}}_x{\mathrm{MnO}}_3$ based on PS, and question its ubiquity in the manganites. In contrast to ${\mathrm{LaMnO}}_3$ that exhibits long-range orbital correlations and sizable octahedral distortions at low temperature, the doped samples with compositions straddling the MI boundary exhibit correlations (in the insulating regime) limited to only $\sim 1$ nm with observably smaller distortions. In the $x=0.22$ sample local JT distortions are found to persist across the OR transition and deep into the $R$ phase (up to $\sim 1050$ K), where they are crystallographically prohibited. Their magnitude and subnanometer spatial extent remain unchanged.

Figure 1

${\mathrm{MnO}}_6$ octahedra—basic building block of the ${\mathrm{La}}_{1-x}{\mathrm{Ca}}_x{\mathrm{MnO}}_3$ structure: (a) JT distorted, with pairs of long (green), intermediate (gray), and short (red) Mn-O distances; (b) undistorted, with all Mn-O distances equal (gray).

Figure 2

Average structure behavior (Rietveld). Top row: Temperature evolution of the average Mn-O distances of ${\mathrm{MnO}}_6$ octahedron for (a) $x=0.18$ (blue symbols) and (c) $x=0.22$ (red symbols). Horizontal dashed lines represent these distances in ${\mathrm{LaMnO}}_3$ endmember with vertical gray double arrows indicating the distortion size. Bottom row: Temperature evolution of uncorrelated isotropic atomic displacement parameter (ADP) of $8d$ oxygen ($Pnma$ model) for (b) $x=0.18$ (blue symbols) and (d) $x=0.22$ (red symbols). Sloping dashed gray lines denote uncorrelated isotropic ADPs of the same oxygen site in ${\mathrm{CaMnO}}_3$ endmember as a reference in the absence of disorder [36]. Sloping dashed red line in (d) represents the same, but with added constant offset to match the high temperature end of the ADP data. Double arrows indicate excess disorder in doped samples as compared to ${\mathrm{CaMnO}}_3$. The vertical dashed red line marks $T_{\text{MI}}$.

Figure 3

Comparison of experimental PDFs over $r$ range describing ${\mathrm{MnO}}_6$ octahedron, for $x=0.18$ (solid blue line) and $x=0.22$ (solid red line) at various comparable temperatures: $T<T_{\text{MI}}$ (left column), and $T>T_{\text{MI}}$ (right column). (a) 10 K, (b) $\sim 50$ K, (c) $\sim 150$ K, (d) $\sim 250$ K (e) 300 K, and (f) 550 K. Notable difference can be observed at low temperature where the $x=0.18$ sample is insulating, whereas the $x=0.22$ sample is metallic. At $T>T_{\text{MI}}$ visible difference vanishes.

Figure 4

Local structure behavior (PDF). Top row: Temperature evolution of the local Mn-O distances of ${\mathrm{MnO}}_6$ octahedron for (a) $x=0.18$ (blue symbols) and (c) $x=0.22$ (red symbols). Horizontal dashed lines represent these distances in ${\mathrm{LaMnO}}_3$ endmember. Bottom row: Temperature evolution of correlated isotropic ADPs of $8d$ oxygen ($Pnma$ model) for (b) $x=0.18$ (blue symbols) and (d) $x=0.22$ (red symbols). Dashed lines sketch the uncorrelated ADPs shown in Figs. 2 and 2 for reference. The vertical dashed red line marks $T_{\text{MI}}$.

Figure 5

Mn-O bond lengths from the refined structure model to PDF data as a function of upper refinement limit $r_{\text{max}}$ for (a) $x=0.18$ and (b) $x=0.22$ compositions. Assessment for $x=0.18$ was done at 10 K (solid symbols) and 300 K (open symbols), whereas for $x=0.22$ at 300 K for PDF data obtained at $Q_{\text{max}}$ values of $25{\AA }^{-1}$ (open symbols) and $35{\AA }^{-1}$ (solid symbols). Dashed lines indicate Mn-O bond lengths in ${\mathrm{LaMnO}}_3$ at 300 K, where octahedral distortions are long-range ordered. See text for details.

Figure 6

(a) $T$ evolution of undistorted fraction from two-phase short-range PDF modeling for $x=0.18$ (solid blue symbols, no MIT), and for $x=0.22$ (solid red symbols, MIT at $\sim 180$ K indicated by a vertical dashed red line). Horizontal solid black lines are guides to the eye. Typical short-range refinements of the two-phase model are shown in (b) for $x=0.18$ and in (c) for $x=0.22$ data at 10 K. In both panels open blue symbols represent the data, solid red line is the model, and solid green line is the difference (offset for clarity).

Figure 7

(a) Comparison of $T$ evolution of uncorrelated ADPs (Rietveld) for $x=0.18$ (open blue symbols), $x=0.22$ (open red symbols), and $x=1.0$ (dashed black line, reference) compositions. Correlated ADPs obtained from two-phase refinements (PDF) are shown by solid symbols for $x=0.18$ (blue) and $x=0.22$ (red). (b) Evolution of long Mn-O bond with temperature from short-range PDF modeling. Solid blue symbols represent data for $x=0.18$ (no MIT), solid red symbols represent data for $x=0.22$ (MIT at $\sim 180$ K, indicated by vertical dashed red line). Horizontal dashed black line marks long Mn-O bond for $x=0$ (reference) composition. Solid blue line through $x=0.18$ points is guide to the eye.

Figure 8

Rietveld refinements of the average structure models to ${\mathrm{La}}_{1-x}{\mathrm{Ca}}_x{\mathrm{MnO}}_3 x=0.22$ data at (a) 550 K $\left(Pnma\right)$ and (b) 1050 K $\left(R\overline{3}c\right)$. Blue symbols represent the data, solid red lines are the model, and solid green lines are the differences (offset for clarity). Vertical ticks mark reflections. (c) Experimental diffraction patterns over a narrow range of $d$ spacing sensitive to O-R average structural transition for temperature as indicated. Profiles have been offset for clarity. The bottom and top profiles correspond to the same 550 K temperature collected on warming and cooling, respectively. Oxygen content refined for these two data sets does not change within the experimental uncertainty, verifying that the sample integrity has been preserved through the heating cycle.

Figure 9

(a) Temperature evolution of the average Mn-O distances of ${\mathrm{MnO}}_6$ octahedron for the $x=0.22$ sample, as obtained from Rietveld refinements across the entire $T$-range studied. Vertical dashed red lines at $\sim 180$ and $\sim 720$ K mark M-I and O-R transitions, respectively. Vertical dashed lines mark 300 K values for $x=0$ reference. (b) $T$ evolution of $8d$ oxygen ($Pnma$ model) ADP as obtained from Rietveld refinements. Inset to (a) sketches reflections in the diffraction data characteristic for O and R phases, as shown in Fig. 8.

Figure 10

Temperature evolution of the local Mn-O distances of ${\mathrm{MnO}}_6$ octahedron for the $x=0.22$ sample, as obtained from short-range PDF refinements. Local distortions that are onset upon entering the insulating state persist also in the rhombohedral phase, and are seen up to 1050 K—the highest temperature studied. Vertical dashed red lines mark $T_{\text{MI}}$ and $T_S$.

Figure 11

(a) Comparison of experimental PDF data for $x=0.22$ at 650 K ($Pnma$ phase, open blue symbols) and 750 K ($R\overline{3}c$ phase, open red symbols). (b) and (c) Comparison of the data (open gray symbols) and the $Pnma$ model based on intermediate range structure parameters (solid red line) for 650 and 850 K, respectively. In all panels the difference curve (solid green line) is offset for clarity. Light-red solid lines above the differences represent a 4 Å running average of the absolute value of the difference curve, multiplied by 10 and offset for clarity. Shaded areas emphasize the regions of disagreement of the compared PDFs.