Spin order and lattice frustration in optimally doped manganites: A high-temperature NMR study

Understanding the complex glassy phenomena, which accompany polaron formation in optimally doped manganites (ODMs) is a cumbersome issue with many unexplained perspectives. Here, on the basis of ${}^{139}\text{L}\text{a}$ and ${}^{55}\text{M}\text{n}$ nuclear magnetic resonance (NMR) measurements, performed in the temperature range 80–900 K we show that glass freezing, observed in the paramagnetic (PM) phase of ODM ${\text{La}}_{0.67}{\text{Ca}}_{0.33}{\text{MnO}}_3$, is not a random uncorrelated process but the signature of the formation of a genuine spin-glass state, which for $T<T_c$ consolidates with the ferromagnetic (FM) state into a single thermodynamic phase. Comparison with NMR measurements performed on ${\text{La}}_{1-x}{\text{Ca}}_x{\text{MnO}}_3$ systems for $0.0\le x\le 0.41$ and ODM ${\text{La}}_{0.70}{\text{Sr}}_{0.30}{\text{MnO}}_3$, demonstrates the key role played by the local lattice distortions, which control (i) the stability of the spin-glass phase component and (ii) the kind (first or second order) of the PM-FM phase transition. The experimental results are in agreement with the predictions of the compressible random bond-random field Ising model, where consideration of a strain field induced by lattice distortions is shown to invoke at $T_c$ a discontinuous first-orderlike change in both the FM and the “glassy” Edwards-Anderson order parameters.

Figure 1

(Color online) The phase diagram of ${\text{La}}_{1-x}{\text{Ca}}_x{\text{MnO}}_3$ for $0.0\le x\le 0.5$. The polaron glass and dynamic polaron regimes are defined according to Ref. 23. The green squares (line) define the PM to FM transition line in magnetic field 9.4 T. The inset shows the corresponding phase diagram for ${\text{La}}_{1-x}{\text{Sr}}_x{\text{MnO}}_3$ from Ref. 31. The blue squares define the transition line between the $R\overline{3}c$ and $Pnma$ crystal structures.

Figure 2

$H/M$ vs $M^2$ plots of isotherms in the vicinity of the Curie temperature $T_c$ for LCMO(0.23,0.33,0.41) and LSMO(0.30) systems.

Figure 3

${}^{139}\text{L}\text{a}$ NMR spectra for LCMO $x=0.0$ at 293 K and 0.11 at 250 and 500 K, in 9.4 T external magnetic field. Spectra are presented relatively to the central line frequency ${\nu }_0$. Open circles are experimental data and solid lines theoretical simulations.

Figure 4

(Color online) ${}^{139}\text{L}\text{a}$ NMR spectra for LCMO $x=0.11$, 0.17,0.23, 0.33, and 0.41 at 500 K, in 9.4 T external magnetic field. Spectra are presented relatively to the central line frequency ${\nu }_0$. Open circles are experimental data and solid lines theoretical simulations. The inset indicates the electric quadrupolar coupling ${\nu }_Q=\frac{e^{2}qQ}{h}$ as a function of doping.

Figure 5

${}^{139}\text{L}\text{a}$ NMR spectra for optimally doped LCMO(0.33), in 9.4 T external magnetic field, as a function of temperature. Spectra are presented relatively to the central line frequency ${\nu }_0$. The dotted line shows schematically the border line between the $R\overline{3}c$ and $Pnma$ phase regimes. The left inset depicts the calculated powder pattern spectrum in comparison to the experimental spectrum at $T=730\text{K}$. The right inset shows a comparison of spectra at 80 K (open circles) and 400 K (solid line).

Figure 6

${}^{139}\text{L}\text{a}$ NMR spectra for LCMO, $x=0.23$ (left), 0.41 (middle), and LSMO $x=0.3$ (right) at various temperatures. Spectra are presented relatively to the central line frequency ${\nu }_0$.

Figure 7

(Color online) The ${}^{139}\text{L}\text{a}$ NMR frequency as a function of temperature for LCMO(0.23) (◆), LCMO(0.33) (●), LCMO(0.41) (○), and LSMO(0.30) (▲). In the inset the width of the NMR frequency distribution is shown for (i) LCMO(0.23) central line (◆) and satellite powder pattern (◇), (ii) LCMO(0.33) central line (●) and satellite powder pattern (○), and (iii) for LSMO(0.30) central line (▲) and satellite powder pattern (△).

Figure 8

${}^{55}\text{M}\text{n}$ NMR spectra in zero external magnetic field (open circles) of LCMO(0.33) at 80 K and 200 K and LSMO(0.30) (left figures) at 80 K and 300 K, respectively. The spin-spin $T_2$ relaxation times (crosses) as a function of frequency are also shown in the same figures. By increasing temperature spectra shift to lower frequencies while for LCMO(0.33) they become asymmetric broader on approaching $T_c$ from below. The $T_2$ minimum at the peak of the spectra at 80 K is due to the Shul-Nakamura interactions.

Figure 9

(Color online) The hyperfine field $B_{hf}$ as a function of temperature, obtained from 55 Mn NMR measurements, for LCMO(0.33) (blue,○) and LSMO(0.30) (black, ○). The corresponding (blue and black, ●) are the $B_{hf}$ values as obtained by Mössbauer spectroscopy on lightly Sn-doped samples from Ref. 22. The solid lines are theoretical fits by applying the self consistent Eqs. (5, 6). Green filled squares (◼) present the EA order parameter, ${\overline{q}}_{\text{EA}}=q_{\text{EA}}-M^2$, as obtained by Eqs. (5, 6). Green open squares (◻) correspond to the intensity of the quasielastic neutron-scattering component from Ref. 49. In conventional spin glasses, the quasielastic neutron-scattering component is considered to provide a measure of the EA order parameter. The inset shows the corresponding spectra for LCMO(0.33) (blue lines) and LSMO(0.30) (black lines).

Figure 10

(Color online) Experimental ${}^{55}\text{M}\text{n}$ NMR line shapes (left) and theoretical simulations (right) for LCMO $x=0.33$ at various temperatures, as obtained by formula (7).