Elevating Curie Temperature through Orbital Ordering in La0.67Sr0.33MnO3/BaTiO3 Superlattices

Background

Perovskite manganite films routinely suffer a pronounced decline in T_C as the film thickness is reduced, a phenomenon often termed the “dead layer” [1–5]. This thickness‑dependent loss of ferromagnetism limits the performance of spintronic devices such as field‑effect transistors, magnetic tunnel junctions, spin valves, and nonvolatile magnetic memories [6–8]. Possible origins include electronic and/or chemical phase separation [9,10], growth‑induced microstructure changes [11,12], or reconstruction of the Mn e_g orbitals [13,14]. Recent efforts have focused on superlattice engineering and strain tuning to restore or even surpass the bulk T_C in ultrathin films [15–18]. Among the perovskite manganites, La_0.67Sr_0.33MnO₃ (LSMO) is of particular interest due to its colossal magnetoresistance, high T_C, and half‑metallicity [19–23]. Heterostructures incorporating LSMO have revealed rich interfacial physics, including atomic intermixing and novel couplings [24–28]. Notably, M. Ziese et al. reported ferromagnetic order persisting down to two unit cells in LSMO/SrRuO₃ superlattices, with T_C exceeding room temperature [29]. First‑principles calculations predict that intercalating suitable layers—such as BaTiO₃ (BTO)—into LSMO can raise T_C dramatically by controlling orbital ordering, specifically favoring the in‑plane e_g(x²–y²) state that promotes strong double‑exchange interactions [30]. This theoretical insight has now been confirmed experimentally in the present work.

In this paper, we synthesize LSMO/BTO superlattices by pulsed laser deposition (PLD) and employ XLD to correlate the enhanced T_C with Mn e_g orbital occupancy. We find that regular BTO interlayers not only preserve high‑quality epitaxy but also impose a tensile strain that stabilizes the in‑plane e_g(x²–y²) orbital across both central and interfacial LSMO layers, leading to an elevated T_C that can exceed 310 K. These results provide a concrete strategy for tuning magnetism in artificial oxide structures, with implications for high‑temperature spin‑valve devices and nonvolatile magnetic memories.

Methods

We fabricated (001)-oriented [(LSMO)₃/(BTO)₃]_n superlattices (SL‑n, n = 3, 4, 10) on (001) SrTiO₃ (STO) substrates using PLD. A stoichiometric polycrystalline target was ablated in 100 mTorr O₂ at 725 °C for LSMO and 780 °C for BTO. A KrF excimer laser (λ = 248 nm) at 2 Hz, with 350 mJ for LSMO and 300 mJ for BTO, produced layer‑by‑layer growth. Post‑growth, the samples were annealed in 300 Torr O₂ for 1 h to reduce oxygen vacancies before cooling to room temperature. For comparison, we also prepared single‑layer LSMO films of 3 u.c. and 40 u.c. (LSMO(3) and LSMO(40)). To achieve atomically flat, single‑terminated STO surfaces, we etched in NH₄F‑buffered HF (BHF) and annealed at 960 °C in O₂. Atomic force microscopy (AFM) confirmed step‑terrace morphology (Fig. 1d). In‑situ reflection high‑energy electron diffraction (RHEED) monitored growth dynamics and ensured unit‑cell precision (Fig. 1a). Crystal structure and morphology were characterized by X‑ray diffraction (XRD) and transmission electron microscopy (TEM). Raman spectroscopy (514.5 nm Ar⁺ laser) verified strain states (Fig. 1c). Magnetic properties were measured with a SQUID magnetometer under a 3000 Oe in‑plane field, after subtracting the STO diamagnetic background. Transport was assessed via the Van der Pauw four‑point probe in a PPMS from 20 K to 365 K. XAS and XLD were performed in total electron yield mode at the Shanghai Synchrotron Radiation Facility (BL08U1A) and the National Synchrotron Radiation Laboratory (U19) at room temperature.

Results and Discussion

RHEED oscillations during SL‑3 growth (Fig. 1a) confirm layer‑by‑layer deposition and precise control of thickness. XRD patterns (Fig. 1b) show sharp (001) reflections for all SL‑n samples; the LSMO peaks shift to higher 2θ while BTO peaks shift lower, indicating a tensile strain in LSMO and compressive strain in BTO that is preserved across the multilayers. Raman spectra (Fig. 1c) reveal a slight low‑frequency shift (~252 cm⁻¹) in SL‑10 relative to LSMO(40), consistent with tensile strain imposed by BTO [31–33]. High‑resolution TEM (Fig. 2a–b) confirms atomically sharp, fully strained interfaces without interdiffusion, corroborating the XRD findings.

Temperature‑dependent magnetization (Fig. 3a) shows a dramatic increase in T_C for the SL‑n series compared with ultrathin LSMO(3) (T_C ≈ 45 K). SL‑10 reaches T_C ≈ 310 K, well above room temperature. Hysteresis loops at 5 K (Fig. 3b) exhibit clear ferromagnetism with a saturation magnetization M_s ≈ 1.5 μ_B/Mn (excluding LSMO(3)). The easy axis lies in‑plane for all SL‑n, consistent with the observed orbital occupancy.

XLD spectra (Fig. 4a–b) for SL‑3 and SL‑10 reveal negative ΔXLD at the Mn L₂ edge, indicating preferential occupancy of the e_g(x²–y²) orbital. The magnitude of ΔXLD is larger in SL‑10, matching the higher T_C. The tensile strain from the 4 % lattice mismatch (BTO a = 0.397–0.403 nm vs. LSMO a = 0.387 nm) forces the c < a condition, stabilizing the in‑plane orbital and enhancing the double‑exchange pathway within the (001) plane. Consequently, the high T_C originates from strengthened in‑plane ferromagnetic coupling rather than interfacial layers alone, contrasting with earlier theoretical predictions that emphasized only the central LSMO layers [30].

Resistivity measurements (Fig. 4c) show metal–insulator transitions at T_MI ≈ 178 K (SL‑3) and 310 K (SL‑10), coincident with their respective T_C values, reinforcing the link between ferromagnetism and itinerant electron transport. The enhanced double‑exchange in the (x²–y²) channel explains the observed high‑temperature metallicity.

Conclusions

We have demonstrated that inserting BaTiO₃ layers into La_0.67Sr_0.33MnO₃ superlattices systematically elevates the Curie temperature above 310 K. The key mechanism is strain‑induced preferential occupancy of the Mn e_g(x²–y²) orbital, which amplifies in‑plane double‑exchange ferromagnetism. This approach provides a practical route to engineer high‑temperature ferromagnetism in ultrathin manganite films, paving the way for robust magnetoelectronic devices that function far above room temperature.

Abbreviations

AFM:

Atomic force microscopy

BHF:

NH₄F-buffered HF solution

FFT:

Fast Fourier transform

M_s:

Saturation magnetization

PLD:

Pulsed laser deposition

PPMS:

Physical properties measurement system

RHEED:

Real‑time reflection high‑energy electron diffraction

SQUID:

Superconducting quantum interference device

T_C:

Curie temperature

TEM:

Transmission electron microscopy

TEY:

Total electron yield

T_MI:

Metal‑to‑insulator transition temperature

XAS:

X‑ray absorption spectroscopy

XLD:

X‑ray linear dichroism

XRD:

X‑ray diffraction