Manipulating magnetoelectric properties by interfacial coupling in La0.3Sr0.7MnO3/Ba0.7Sr0.3TiO3 superlattices

Artificial superlattices constructed with ferromagnetic La0.7Sr0.3MnO3 layer and ferroelectric Ba0.7Sr0.3TiO3 layer were designed and fabricated on SrTiO3 substrates. An epitaxial growth with sharp interfaces between La0.7Sr0.3MnO3 and Ba0.7Sr0.3TiO3 layers was confirmed by scanning transmission electron microscopy and x-ray diffraction. An unambiguous charge transfer involving an electron transferring from the La0.7Sr0.3MnO3 layers to Ba0.7Sr0.3TiO3 layers (Mn3+→Mn4+; Ti4+→Ti3+) across the interface were resolved by electron energy loss spectra analysis. These observations are attributed to the possible modification in the stereochemistry of the Ti and Mn ions in the interfacial region. The out-of-plane lattice parameter, Curie temperature, and magnetoresistance are strongly affected by the thicknesses of the La0.7Sr0.3MnO3 and Ba0.7Sr0.3TiO3 layers. Huge magnetoresistance subsisting to low temperature was also observed in the La0.7Sr0.3MnO3/Ba0.7Sr0.3TiO3 superlattices. All spectral changes identified at a nanometer scale and their potential effect on the degradation of magnetic and transport properties at a macroscopic level. These findings highlight the importance of dependence on sublayer thickness, illustrating the high degree of tenability in these artificially low-dimensional oxide materials.

Controlled interfaces between complex oxide systems with the perovskite oxides, lead to a variety of novel and unexpected electronic and magnetic phenomena and functionalities at the interface that are absent in the respective bulk materials 1 . Representative aspects of recent finding include a two-dimensional electron gas with magnetic and superconducting ground states at strong electric insulating interfaces between SrTiO 3 and LaAlO 3 2-5 , emergent ferromagnetism at antiferromagnetic interfaces between LaMnO 3 /SrMnO 3 6, 7 , interfacial orbital and spin polarization at the interface of a superconductor YBa 2 CuO 7 and a ferromagnet La 0.7 Ca 0.3 MnO 3 8, 9 , orbital reconstruction at the interfaces of LaNiO 3 and LaAlO 3 10,11 , and enhanced polarization in BaTiO 3 /SrTiO 3 /CaTiO 3 superlattice 12, 13 , etc. These results also generated an intense debased on the origin of the unexpected properties at the interfaces. However, the competition between multiple order constants induces a great challenge on our ability to predict the resulting novel electronic phases and functionalities and thus to create artificial materials with meet specific application requirements, and these artificial materials in the form of either composites, superlattices, or multilayers. Recently, the superlattices consisting of alternating ferroelectric and ferromagnetic layers yielded unusual electrical and magnetic transport properties that cannot be obtained in either of their constituents, such as Pr 0.85 Ca 0. 15 18,19 , etc. However, most of the research was devoted to evaluate the relationship between the whole magnetic and transport properties and individual superlattice layers quality, and it was proposed that these unusual electrical and magnetic transport properties are attributed to the ferroelectric spacer layer and the associated magnetoelectric coupling with the ferromagnetic layer [14][15][16][17][18][19] . The microscopic structure of the layers and interfaces (e.g. residual strain, strain gradient, atoms arrangement etc.) and electric structures (valence of the ions and charge transfer at the interfaces) should play crucial roles. One the other hand, oxygen vacancies seem to play a key role in these structures, but the puzzle is far from being understood 20 .
Mixed-valence manganite La 0.7 Sr 0.3 MnO 3 exhibits fantastic rich variety of attractive properties, such as colossal magnetoresistance (MR), a Curie temperature above room temperature (T C ~ 360 K), half-metallicity, high conductivity, and low coercive field [21][22][23] . While Ba 0.7 Sr 0.3 TiO 3 as good a candidate for application in high-density dynamic random access memories and microwave devices is due to their high relative dielectric constant (ε r ) values and tunable ε r by applying an electric field 24 . In this work, we have grown the artificial superlattices consisting with a ferromagnetic La 0.7 Sr 0.3 MnO 3 (LSMO) layer and ferroelectric Ba 0.7 Sr 0.3 TiO 3 (BSTO) layer on (100) SrTiO 3 (STO) substrates, which are of special relevance in practical dielectric and spintronics devices. The aim in this work is to investigate the crystalline microstructure quality and the electrical structures of the individual layers and interfaces of the LSMO/BSTO superlattices (SLs), and to evaluate their effect on the magnetic and electrical properties and the associated magnetoelectric coupling of the LSMO/BSTO superlattices.

Results
Epitaxial nature and crystallinity revealed by SXRD. The LSMO/BSTO SLs composed of individual LSMO layers with 7 and 14 unit cells (u.c.) and BSTO layer thickness of 6, 12, and 24 u.c. with a total periodicity of 15 were designed and fabricated. The LSMO layers with 7 and 14 u.c. were chosen because the dead layer of LSMO is just below 7 u.c. [25][26][27] . Figure 1S shows the high-resolution synchrotron x-ray diffractometry (SXRD) results of the various samples, including the LSMO/BSTO SLs, the LSMO and BSTO films, and the LSMO/BSTO/ STO bilayers. It can be seen from Fig. 1S that all samples show only the (00 l) Bragg reflections, and no diffraction peaks from secondary phase or randomly oriented grains are observed, indicating that all the films, bilayers, and superlattice were epitaxially grown along the c-axis orientation with a good single phase. The SXRD pattern of the SL [LSMO 7 /BSTO 6 ] 15 (L7B6) around (001) and (002) reflections was exhibited in Fig. 1(a). The presence of higher-order satellite peaks (1, 2, 3…) in the LSMO/BSTO SL, which arising from the periodic chemical modulation of multilayer structures, clearly provides evidence for the structural coherency and the good crystalline quality of the LSMO/BSTO SLs. The high quality of the LSMO/BSTO SL can be appreciated from oscillations of integrated x-ray intensity, known as Pendellösung fringes, as shown in the inset of Fig. 1(a). X-ray φ scan was carried out to confirm the epitaxial growth of SLs. The (202) φ scan result of L14B12 is plotted in Fig. 2S, and four equally spaced peaks separated by 90° were observed, indicating that L14B12 was epitaxially grown on the ST0 substrate. Moreover, the SL peaks are the same with respect to the substrate peaks, suggesting the epitaxial relationships: From the position of the fundamental diffraction peak of multilayers, the out-of-plane lattice constant c of these SLs, together with the bilayers and films, were calculated and shown in Fig. 1(b). The pseudo-cubic lattice constants of the bulk LSMO and BSTO are 3.876 28 and 3.972 Å 29 , respectively, and that of the STO substrate is 3.905 Å. The lattice mismatch is −0.74% between LSMO and STO, 1.72% between BSTO and STO, and 2.48% between BSTO and LSMO, respectively. Therefore, the LSMO layers and film experience compressive stress along out-of-plane direction, and the BSTO layers and film experience the tensile stress along out-of-plane direction, indicating by red arrows in Fig. 2(a). Figure 3S exhibits an x-ray reciprocal space map of L14B12 around the STO (013) Bragg peak. Except the STO (013) peak (red arrow), the superlattice satellite peaks were readily apparent (white arrows). All the superlattice peaks are on the same H line with the substrate peak, indicating that the superlattices are under the fully coherent strain.
The out-of-plane lattice constant of the BSTO film is as large as 4.169 Å, a little larger than that (4.145 Å) of BSTO in LSMO/BSTO/STO bilayer. Both the out-of-plane lattice constants of the BSTO film and the BSTO layer in the LSMO/BSTO/STO bilayer are larger than that of the bulk BSTO due to the tensile stress along the out-of-plane direction. In contrast, the out-of-plane lattice constants of the LSMO film and the LSMO layer in LSMO/BSTO/STO bilayer are both smaller than that of the bulk LSMO due to the compressive stress in along the out-of-plane direction. The out-of-plane lattice constants of the LSMO/BSTO superlattice are determined by the out-of-plane cumulative stress (i.e. the substrate induced stress and interfacial stress). The out-of-plane lattice constants of the LSMO/BSTO superlattices are among 3.917 and 3.986 Å, tuned by both the thicknesses of the LSMO (t LSMO ) and BSTO (t BSTO ) layers, i.e., the out-of-plane lattice constants of the LSMO/BSTO superlattices slightly decrease with increasing the thickness ratio of t LSMO /t BSTO , or to say that an enhanced out-of-plane lattice constants for a superlattice are obtained with thicker BSTO layers or thinner LSMO layers.

STEM and EELS.
Atomic resolution characterization is crucial to investigate the interfacial structure and to demonstrate interfacial electronic states. This can be acquired by the state-of-the-art sophisticated aberration-corrected scanning transmission electron microscopy (STEM) using a high-angle annular dark-field (HAADF) detector in combination with electron energy loss spectroscopy (EELS). Figure 2(a) exhibits the schematic stacking sequence of LSMO/BSTO SLs grown on (001) STO substrates, with a total periodicity of 15. The stacking sequence of the LSMO/BSTO SL is shown in Fig. 2(b), showing a single supercell that is repeated the indicated number of times. The contrast of HAADF image exhibits ~Z 1.7 dependence, where Z is the atomic number 30 , therefore, the brighter bands corresponding to the nominal 7-u.c.-thick LSMO layers and the darker ones to the 6-u.c.-thick BSTO layers. A low-magnification cross-sectional HAADF-STEM image of L7B6 are shown in Fig. 2(c), clearly exhibiting well-defined superlattice structure consisting of 15 LSMO (bright) /BSTO (dark) bilayers, which is consistent with our design (as shown in Fig. 2(a)). The layers and interfaces are flat over long lateral distances. Higher magnification images allow the study of the interfacial structure in detail. High-resolution HAADF-STEM image is shown in Fig. 2(d). In Fig. 2(d), atomic columns of La(Sr), Mn, Ba(Sr), and Ti are resolved while O columns are not observed due to weak scattering of light atomic columns. The STEM image exhibits atomically sharp interfaces and coherent, which indicates the epitaxial growth of the LSMO/BSTO SLs. STEM imaging of SL L7B6 was also conducted under annular-bright-field (ABF) mode, as shown in Fig. 3S. The ABF-STEM images also exhibit atomically sharp interfaces and lateral coherency, indicating excellent epitaxial growth and high-quality of the LSMO/BSTO superlattices. Moreover, the epitaxial growth was also confirmed by the results of corresponding Fast Fourier transform (FFT) images (as shown in Fig. 5S).
The identification of the oxidation states of cations at atomic resolution combined with STEM has become a useful tool to identify the charge states in recent years [31][32][33][34] . EELS under STEM mode with high spatial resolution was used to examine the each element of individual atomic columns. EEL spectrum imaging was used to ascertain the interface structure and to probe any chemical disorder (e.g., interdiffusion). The Ti-L 2,3 , Mn-L 2,3 , Ba-M 4,5 , and La-M 4,5 edges can be simultaneously recorded in a line scan across the interface of L7B6. Figure 3(a) exhibits the HAADF-STEM image marked with the position of the EELS line scan and EELS spectral maps for Ti-L 2,3 , Mn-L 2,3 , Ba-M 4,5 , and La-M 4,5 edges across the STO substrate, three layers of 6-u.c.-thick BSTO, and two layers of 7-u.c.-thick LSMO. From the EELS spectral maps it can be concluded that the interfaces are very sharp and little chemical interdiffusion is present. EELS line scans were used to characterize not only the cations content profiles but also the local Mn and Ti valence along the thickness of LSMO/BSTO superlattices. Figure 3(b) shows the Mn L 2,3 edges from the layer of 7-u.c.-thick LSMO. The spectra of the Mn L edges do not exhibit strong outstanding features as one might expect for these spectroscopic features dominated by multiplet effects 31 . The line shape of the spectrum depends strongly on the multiplet structure given by the Mn 3d-3d and 2p-3d Coulomb and exchange interactions, as well as by the local crystal fields and the hybridization with the O 2p ligands. It can be seen from Fig. 3(b) that no resolvable shifts on the Mn-L-edge fine structures among the Mn-O planes of the LSMO layer can be resolved, which is generally expected when the local Mn valence changes 9,35 . No obvious chemical shift of the Mn L 3 peak was also observed in previous work on LaMnO 3 -SrMnO 3 superlattice samples 36,37 . It was proposed if the Mn concentration was constant in their LMO-SMO superlattice, the lack of a chemical shift has been suggested to result from either oxygen vacancies in the LSMO film or the covalent Mn-O bond rather than a purely ionic bond 37,38 . An additional peak about 1.5 eV below the main L 3 peak can be seen, which is as a signature of Mn 4+ ion 39 . It can be noted that the additional peaks (indicated by the black arrow) of the end interfacial MnO 2 planes show a slight enhancement, which can be ascribed to a slightly increased concentration of Mn 4+ ion in the two end interfacial MnO 2 planes.
For investigating the variation of the valence state of Mn, except for the change of the intensity of the shoulder at higher energy corresponding to Mn 4+ , the ratio of the integrated intensity of the Mn L 3 and L 2 lines (L 23 ratio) has been used to quantify the Mn valence state 39 . The Mn EELS data close to the LSMO/BSTO interface is shown in Fig. 6S(a). After scaling and subtraction of a Shirley function (black dotted line), the remaining signals under the L 3 and L 2 lines are integrated, and their ratio (L 23 ratio) is calculated. The calculated L 23 ratio is about 2.64. The nominal Mn oxidation state for our La 0.7 Sr 0.3 MnO 3 is +3.3. According the Varela's results, the L 23 ratio of the nominal Mn oxidation state for +3.3 is about 2.5, and the larger the Mn L 23 ratio, the smaller the Mn oxidation state is. The smaller Mn oxidation state should be attributed to the oxygen vacancies present in the LSMO layers. Figure 6S(b) shows the dependence of the calculated L 23 ratio with the distance of LSMO layer from the interface. It can be seen from Fig. 5S(b) that the L 23 ratio close to the interfaces are a little smaller than that inside MnO 2 layers, indicating the Mn valence state close to the interfaces is a little higher than that inside MnO 2 layers.
Individual EELS spectra of Ti-L 2,3 edge obtained from line scan cross the 6-u.c.-thick BSTO layer of are shown in Fig. 3(c). Unlike Mn, there is an empty d orbital for Ti 4+ while there is one d electron in d orbital for Ti 3+ , and the EELS spectra are quite different for Ti 4+ and Ti 3+ with large shift and different line shapes, The reference EELS spectra for Ti 4+ and Ti 3+ are taken from the SrTiO 3 and LaTiO 3 . It can be seen from Fig. 3(c) that every Ti-L 2,3 EELS spectroscopic features are the mixture of those from Ti 4+ (black dashed lines) and from Ti 3+ (black arrows), indicating that the Ti sites exhibit mixed valence between 3+ and 4+. The spectral shape of the Ti EELS changes very tiny across the entire layers, unlike the case of the Mn EELS. This difference maybe come from the different mechanism of the electrons transferring between the t 2g 0 /t 2g 1 orbitals of the Ti 4+ /Ti 3+ ions and the t 2g 3 e g 0 /t 2g 3 e g 1 orbitals of the Mn 4+ /Mn 3+ ions, and the t 2g 0 /t 2g 1 orbital configuration is more efficient electron transfer. The presence of Ti 3+ features at interface should be caused by an electron transfer to Ti 4+ ion across the interfaces, since the valence of Ti in the stoichiometric BSTO is +4. Such charge-transfer mechanisms across the oxide interface were also occurred at the other oxide interfaces of the heterostructures and superlattices evidenced by the STEM-EELS spectra, and the typically range is within ~1 nm of the interface 2, 33, 34, 36 . Combination of the EELS analysis of Mn and Ti, an unambiguous charge transfer involving an electron transferring from the LSMO layers to BSTO layers (Mn 3+ →Mn 4+ ; Ti 4+ →Ti 3+ ) across the interface are confirmed. Finally, both the oxidization states of the Mn and Ti ions should be changed due to the existence of the oxygen vacancies. The existence of the oxygen vacancies in the LSMO layers were confirmed by the analysis of the EELS data, as shown in

Magnetic Properties. Charge distribution at the interfaces of the LSMO/BSTO SLs could be expected to
have a large impact on the transport and magnetic properties of the LSMO/BSTO SLs. In order to understand these, transport and magnetic measurements were performed using PPMS. Figure 4 exhibits the magnetization of the LSMO/BSTO SLs along with comparison with the LSMO film as a function of temperature and magnetic field (M(T)). The magnetization is normalized by the volume of the LSMO layer. Magnetization was measured at field-cooled mode in a magnetic field of 0.01 T. It can be seen from Fig. 4(a) that the magnetizations of all LSMO/ BSTO SLs exhibit paramagnetic-to-ferromagnetic transition with decreasing temperature, in accordance with the behavior of the temperature-dependent magnetization of the single-layer LSMO film. The ferromagnetic (FM) properties of all LSMO/BSTO SLs at low temperature were confirmed by exhibiting ferromagnetic hysteresis loops at low temperature, as shown in Fig. 4(b). Table 1 shows the dependences of paramagnetic-to-ferromagnetic transition temperature (T C ), the saturated magnetization moment (M S ) under 1 T at 10 K, and corrective field Resistivity and MR. The resistivity, ρ, of each superlattice is calculated from its resistance using the simplified assumption that the current flows only in the LSMO layers since the resistivity of BSTO is orders of magnitude higher than that of LSMO and the entire stack rather than only the LSMO layers contribute to the conduction. The temperature dependence of the resistivity of the superlattices and the LSMO film under the magnetic fields of 0, 0.5, 1, 3, 5 T are presented in Fig. 5(a-c). The LSMO film exhibits insulating characteristics  with a large negative MR which monotonically increases in magnitude with decreasing temperature, and the resistivity at lower temperature is above the instrumental limit, as shown in Fig. 5(a). The similar transport and MR behaviors have been observed in ultrathin LSMO films 40 as well as LSMO films with oxygen vacancies grown under low oxygen pressures 41 . Since the thickness of the LSMO film is 130 nm, the oxygen vacancies should play the key role in the strong depression of the double-exchange magnetotransport properties in the LSMO film. The SLs of L14B24, L7B12, and L7B6 exhibit similar insulating behaviors as the LSMO film across the entire measurable temperature range, as shown in Fig. 5(a). The resistivity ρ of the SLs increases with increasing the thickness of the BSTO layers, and larger than that of the LSMO film. This indicates that the thickness of the BSTO layer plays a determining role in the transport properties of the system, since the BSTO layers act as barrier layers to prevent electron hopping through BSTO layers. Quite different from other LSMO/BSTO SLs, SL L14B6 displays a transition from semiconducting to metallic behavior at 160 K, followed by an increase for temperatures below 40 K, as shown in Fig. 5(b). This qualitative behavior is also observed in low Sr-doped La 1−x Sr x MnO 3 which is attributed to the weak localization of carriers due to disorder 42 , and in ultrathin LSMO films which is attributed to the presence of quantum interference effects due to weak localization 43 . SL L14B12 also displays similar features, as shown in Fig. 5(c), although with much larger resistivity and lower transition temperature.
Magnetoresistance, MR = [ρ(H)−ρ(0)]/ρ(H) × 100, as a function of temperature for the LSMO/BSTO SLs and the LSMO film are shown in Fig. 6. MR of SL L7B6 shows similar features as that of the LSMO film, exhibiting negative MR at the measureable temperature range. However, SL L14B6 shows a negative peak in the magnetoresistance at the temperature close to a transition from semiconducting to metallic behavior, and large MR extends over low temperatures. Especially in SL L14B12, MR increases with temperature and huge MR subsist to low temperature, as −1000% under 5 T at 10 K. These results are obviously different from those of LSMO bulk, which exhibit negligibly small MR at low temperature far below T C , because the local moments within each magnetic domain orientate rather unanimously and the suppression of the local moment fluctuations by a magnetic field can be neglected [21][22][23]42 . In additional, an anomalous positive MR is observed both in SLs L14B6 and L14B12 at low temperature and under low magnetic fields, seen in the inset of Fig. 6(d).

Discussion
The lattice parameters of the LSMO/BSTO SLs are determined by the interfacial strain, tuned by both the thicknesses of the LSMO (t LSMO ) and BSTO (t BSTO ) layers, i.e., the out-of-plane lattice constants of the LSMO/BSTO superlattices slightly decrease with increasing the thickness ratio of t LSMO /t BSTO , or to say that an enhanced out-of-plane lattice constants for a superlattice are obtained with thicker BSTO layers or thinner LSMO layers. On the other hand, an unambiguous charge transfer involving an electron transferring from the LSMO layers to BSTO layers (Mn 3+ →Mn 4+ ; Ti 4+ →Ti 3+ ) across the interface are confirmed by EELS. All above observed Moreover, the interfacial effects also affect the transport and magnetic properties of the LSMO/BSTO SLs. The interfacial effects cause suppression of the FM state of the LSMO/BSTO SLs. The results of the magnetic property measurements indicate that the overall magnetization is dominated by the ferromagnetic LSMO layers. It is also worth noting that the magnetization of the LSMO/BSTO bilayer is larger than that of LSMO/BSTO SLs with the same thickness of the LSMO monolayer. Since the magnetizations of LSMO/BSTO SLs are normalized by the volume of the LSMO layer, the normalized magnetizations of LSMO/BSTO SLs should be same if there are no any interfacial effects. The magnetic profiles would be modulated by the interfaces, i.e., the presence of spin frustration at the interfaces or phase separation 44,45 . Finite-size effects in the superlattices will decrease the long-range interactions, and cause the rapid decrease of T C and magnetization with decreasing the LSMO layer thickness. Another reason may come from the interfacial orbital-lattice coupling induced by the biaxial lattice strain, which tends to increase the in-plane Mn-O bond length, reducing the effective electron hopping and reducing the effect of the double exchange mechanism 46,47 . It should be noted that a significant reduction can be seen both in the magnitude of magnetization and T C of all of the LSMO/BSTO SLs compared with the single-layer LSMO film, indicating that all of the LSMO/BSTO SLs present signatures of the important interfacial effects.
The oxygen vacancies always play a key role in the strong depression of the double-exchange magnetotransport properties in the manganite oxides 40,41 , therefore the oxygen vacancies should also affect the magnetotransport properties of the LSMO/BSTO SLs. The effects of the oxygen vacancy on the magnetotransport properties are as following. Firstly, the local Jahn-Teller effect is enhanced by oxygen vacancies, and the strong interaction between the electron and the Jahn-Teller lattice distortion is generally regarded as one main reason for the insulating state of LSMO film. Secondly, the oxygen vacancies break the octahedron symmetry, therefore the energy of the e g orbital pointing at the oxygen vacancy is reduced, which introduces a strong electron-lattice interaction and tends to localize the electrons so that the superexchange tends to dominate the magnetic interaction. Thirdly, the oxygen vacancies lead to the local distortion and rotation of the oxygen octahedrons, which leading to decrease the Mn-O-Mn bond angle and consequently the probability of hopping between Mn ions. Finally, the oxygen vacancies also introduce disorder. Except for the oxygen vacancies, the interface effects are believed to be response to the strong depression of the magnetotransport properties of these SLs. These interface effects include charge redistribution at the interface triggered by a polarity discontinuity 1,4 , and the STEM-EELS results confirmed an unambiguous charge transfer involving an electron transferring from the LSMO layers to BSTO layers (Mn 3+ →Mn 4+ ; Ti 4+ →Ti 3+ ) across the interface. Another is a preferential out-of-plane 3d e g (3z 2 −r 2 ) Mn orbital occupation, attributed to symmetry breaking at the interfaces 48 .
The magnetoelectric coupling associated with the multilayers of magnetostrictive and piezoelectric perovskite oxides could attribute to these effects [14][15][16][17][18][19]49 . The electric field can be generated by ferroelectric BSTO layers in the superlattice structure by producing the charges at the interface, which in turn changes the resistance of LSMO layer. Usually negative colossal MR is observed in manganite bulks and thin films arising from the magnetic field suppressing the scattering of carriers by localized magnetic moments [21][22][23]42 . Similar phenomenon was also been observed in La 0.7 Pb 0.3 MnO 3 bulk 48   be attributed to quantum interference effects due to the Coulomb interactions between carriers enhanced by disorder and the interface effect. The charge and spin states are reconstructed at the interfaces in our system and hence affect the electronic and magnetic properties of the entire system, which should be affected the observed positive MR.
In summary, an epitaxial growth with sharp interfaces between LSMO and BSTO layers was confirmed by STEM and SXRD. An unambiguous charge transfer involving an electron transferring from the LSMO layers to BSTO layers (Mn 3+ →Mn 4+ ; Ti 4+ →Ti 3+ ) across the interface were resolved by EELS. These observations are attributed to the possible modification in the stereochemistry of the Ti and Mn ions in the interfacial region. The out-of-plane lattice parameter, Curie temperature, and magnetoresistance are strongly affected by sublayer thickness. Huge magnetoresistance subsisting to low temperature is also observed in the LSMO/BSTO superlattices. All spectral changes identified at a nanometer scale and their potential effect on the degradation of magnetic and transport properties at a macroscopic level. These findings highlight the importance of dependence on sublayer thickness and temperature, illustrating the high degree of tenability in these artificially layered materials.

Methods
Fabrication. LSMO/BSTO superlattices were grown on STO single crystals by a laser molecular-beam epitaxy system (Laser-MBE) combined with a reflected high energy electron diffraction (RHEED). The LSMO/BSTO superlattices were deposited at 590 °C using a XeCl excimer laser (308 nm, 1.5 J/cm 2 , 2 Hz) at oxygen partial pressure of 5 Pa, which represented the best compromise for both obtaining the smooth interfaces and surfaces and high-quality of the LSMO and BSTO layers. The LSMO/BSTO superlattices composed of individual LSMO layers with 7 and 14 u.c. and BSTO layer thickness of 6, 12, and 24 u.c. with a total periodicity of 15 were realized. For comparison, the LSMO and BSTO thin films and LSMO/BSTO bilayer were also grown at the same conditions. RHEED oscillations were observed and were used to calibrate the number of layers grown. After growth, all samples were in-situ annealed for 30 minutes.
SXRD and RSM. The epitaxial nature and structural characterization of the LSMO/BSTO superlattices were identified by high-resolution SXRD and x-ray reciprocal space map (RSM) at the BL14B1 beam line of Shanghai Synchrotron Radiation Facility (SSRF), using a 1.24 Å X-rays with a Huber 5021 six-axis diffractometry.

STEM and EELS.
To characterize the atomic resolution characterization of the LSMO/BSTO superlattices, state-of-the-art sophisticated aberration-corrected STEM using a HAADF detector were utilized. The HAADF imaging was executed using an ARM-200F (JEOL, Tokyo, Japan) STEM operated at 200 kV with CEOS Cs corrector (CEOS GmbH, Heidelberg, Germany) to cope with the spherical aberration of the probe-forming condenser lens. More detailed chemical composition and bonding was probed on an atomic scale using spatially resolved EELS.
Magnetization and resistivity measurements. Temperature dependences of magnetization, resistivity, and MR of the films and LSMO/BSTO superlattice were performed using a Physical Properties Measurement System (PPMS, Quantum Design Inc.). The standard four-point probe method was used to measure the film resistivity versus temperature (ρ(T)). The magnetotransport properties as a function of temperature and magnetic field were measured by a standard four-point probe method using PPMS.