Irreversible phase reconfiguration and thermal-memory effects in a highly-correlated manganite

TL;DR

The study addresses how competing electronic and structural orders in a phase-separated manganite couple to produce metastable states and memory effects. Using temperature-cycling Raman spectroscopy on LPCMO, it identifies a regime of structural irreversibility arising from the competition between FMM and AFMI-COO phases across the DPS, delineated by $T_C$ and $T_{COO}$, which stores a thermal memory of the perturbation history. Complementary magnetization and resistivity data confirm strong lattice–electron coupling and reveal a nonequilibrium phase dynamics regime in mixed-valence oxides. The findings demonstrate non-volatile resistive switching tied to cumulative thermal cycling within $T_C < T_T < T_{MI}$, with laser-driven localized heating reconfiguring the phase balance and enabling laser-assisted phase engineering with potential neuromorphic applications, thereby linking fundamental phase-transition physics to functional oxide devices.

Abstract

Phase separated manganites provide a unique platform to study the dynamics of competing electronic and structural orders in correlated systems. In $La_{0.275}Pr_{0.35}Ca_{0.375}MnO_{3}$ (LPCMO), we use temperature cycling Raman spectroscopy to uncover a previously unidentified regime of structural irreversibility, emerging from the interplay between lattice distortions and phase competition across the phase separation and charge and orbital ordering temperatures. This irreversible behavior encodes a thermal memory effect reflecting the system's history dependent energy landscape. Correlated magnetic and transport responses confirm the coupling between lattice and electronic degrees of freedom, revealing a nem form of nonequilibrium phase dynamics in mixed valence oxides. These results advance the understanding of metastability and memory phenomena in strongly correlated materials, opening pathways toward adaptive and neuromorphic functionalities in quantum materials.

Figures (13)

• Figure 1: Raman spectroscopy as a function of cumulative thermal cycles. Figure a shows the optical setup, and its main components, used for the Raman spectroscopy measurements. Figure b shows the adopted experimental procedure, consisting of cyclic variations of temperature between a base, T$_B$, and target, T$_T$, values. Figure c shows the LPCMO "as-cooled" Raman spectra together with illustrations of the AS-JT and S-JT modes, identified by blue dashed vertical lines. Figures d - f reproduces the "as-cooled" spectrum together with spectra collected at T$_B$ after cycling to T$_T$ = 102(4), 124(5), 145(5), 167(5) and 188(5) K, respectively.
• Figure 2: Standard thermal evolution of the AS- and S-JT modes. Figure a show the LPCMO Raman spectra for 38, 121, 221 and 321 K, vertically translated. The position of the main rotational ($\nu_{rot}$), bending ($\nu_{bend}$), AS-JT ($\nu_{AS-JT}$) and S-JT ($\nu_{S-JT})$ modes are indicated by black dashed vertical lines. In figure b, all spectra are interpolated to produce the displayed heat map. The horizontal white lines indicates the sample's transitions temperatures. Figures c and d shows the FWHM and the normalized intensities of the AS- and S-JT modes. The intensities of the modes at each temperature were normalized by the $\sim$ 251 cm$^{-1}$ rotational mode intensity. In these figures, the black vertical dashed lines indicates the transition temperatures of the sample. All measurements were taken with fixed laser power ($P=10$ mW).
• Figure 3: Evolution of the AS- and S-JT modes at T$_B$ as a function of thermal cycling. Figure a shows a heat map produced by the interpolation of the Raman spectra at T$_B$ as a function of T$_T$, with the white horizontal dashed lines indicating the sample's transitions temperatures. Figures b and c show the FWHM and the normalized intensities of the AS- and S-JT modes as a function of T$_T$. The intensities of the modes at T$_B$ after each cycle were normalized by the corresponding $\sim$ 250 cm$^{-1}$ rotational mode intensity. Figures d and e display the normalized difference in intensity, $(I_F - I)/ I$, in both modes. The difference between each individual cycle, with $I=I_i$ the initial intensity before each cycle is shown in d. While e shows the normalized difference between the final intensities values after each cycle and the initial, as-cooled, intensity, labeled $I =I_0$. In figures b - e the black vertical dashed lines correspond to the sample's transitions temperatures.
• Figure 4: Magnetic and electrical response of LPCMO as a function of thermal cycling. Panel a shows magnetization curves as a function of temperature, in units of $\mu_B$/Mn, in a cooling and heating regimes between 300 and 10 K and between T$_B$ = 30 K and T$_T$ = 70, 90, 100, 150, and 210 K, with no applied DC-field. Panel b, similarly, shows resistivity curves as a function of temperature, in units of $\Omega$ mm, in a cooling and heating regime between 300 and 10 K and between T$_B$ = 30 K and T$_T$ = 70, 90, 100, 150, and 210 K. Panel c shows the magnetization and resistivity values at T$_B$ before (M$_i$ and $\rho_i$) and after (M$_f$ and $\rho_f$) each thermal cycle. Panel d shows the normalized difference, $(F_F-F)/F$, between final ($F_F$) and initial ($F$) magnetization (M) and resistivity ($\rho$) values for each individual cycle (left figure, $F=F_i$) and in respect to the initial values in T$_B$ = 30 K before thermal cycling commencing (right figure, $F=F_0$). The vertical dashed lines correspond to the respective magnetic and electrical transitions of the compound and the dotted line indicate T$_B$.
• Figure 5: Schematization of the evolution of FMM percolation paths in the AFMI-COO matrix as a function of T$_T$. A schematic and qualitative representation of the changes in the FMM regions in the AFMI-COO matrix at T$_B$ after temperature cycling to T$_T$ is shown, illustrating the observed phenomena.