Control of magnetic anisotropy by orbital hybridization in (La0.67Sr0.33MnO3)n/(SrTiO3)n superlattice

The asymmetry of chemical nature at the hetero-structural interface offers an unique opportunity to design desirable electronic structure by controlling charge transfer and orbital hybridization across the interface. However, the control of hetero-interface remains a daunting task. Here, we report the modulation of interfacial coupling of (La0.67Sr0.33MnO3)n/(SrTiO3)n superlattices by manipulating the periodic thickness with n unit cells of SrTiO3 and n unit cells La0.67Sr0.33MnO3. The easy axis of magnetic anisotropy rotates from in-plane (n = 10) to out-of-plane (n = 2) orientation at 150 K. Transmission electron microscopy reveals enlarged tetragonal ratio>1 with breaking of volume conservation around the (La0.67Sr0.33MnO3)n/(SrTiO3)n interface, and electronic charge transfer from Mn to Ti 3d orbitals across the interface. Orbital hybridization accompanying the charge transfer results in preferred occupancy of 3d3z2-r2 orbital at the interface, which induces a stronger electronic hopping integral along the out-of-plane direction and corresponding out-of-plane magnetic easy axis for n = 2. We demonstrate that interfacial orbital hybridization in superlattices of strongly correlated oxides may be a promising approach to tailor electronic and magnetic properties in device applications.


Abstract:
The asymmetry of chemical nature at the hetero-structural interface offers an unique opportunity to design desirable electronic structure by controlling charge transfer and orbital hybridization across the interface. However, the control of hetero-interface remains a daunting task. Here, we report the modulation of interfacial coupling of (La0.67Sr0.33MnO3)n/(SrTiO3)n superlattices by manipulating the periodic thickness with n unit cells of SrTiO3 and n unit cells La0.67Sr0.33MnO3. The easy axis of magnetic anisotropy rotates from in-plane (n = 10) to out-ofplane (n = 2) orientation at 150 K. Transmission electron microscopy reveals enlarged tetragonal ratio > 1 with breaking of volume conservation around the (La0.67Sr0.33MnO3)n/(SrTiO3)n interface, and electronic charge transfer from Mn to Ti 3d orbitals across the interface. Orbital hybridization accompanying the charge transfer results in preferred occupancy of 3d3z 2 -r 2 orbital at the interface, which induces a stronger electronic hopping integral along the out-of-plane direction and corresponding out-of-plane magnetic easy axis for n = 2. We demonstrate that interfacial orbital hybridization in superlattices of strongly correlated oxides may be a promising approach to tailor electronic and magnetic properties in device applications.
The asymmetry at hetero-structural interface of 3d transitional metal ABO3 oxides, including the mismatch of lattice constant, oxygen octahedral rotation and distortion, and chemical environment, has profound influences on spin and orbital coupling, yielding emerging phenomena, such as enhanced ordering temperature, induced interfacial magnetism at superconductor/manganite interface and orbital reconstruction. [1][2][3][4][5] Coupling between crystal structures [6][7][8][9][10][11][12] across the heterostructure interface leads to new properties, for example, filmthickness dependent interfacial ferromagnetic-polaronic insulator in Pr0.67Sr0.33O3/SrTiO3, 13 spontaneous deep polarization 9 in SrTiO3 near the La2/3Sr1/3MnO3/SrTiO3 interface, rotation of magnetic easy axis of La2/3Sr1/3MnO3 14 and SrRuO3 15 thin film. In strongly correlated 3d oxides, the properties of transition metal oxides are sensitive to electronic structure, especially the occupation of 3d orbitals, 4 which may be manipulated by doping with different covalent ions, 16 hydraulic pressure, 17 and external electric bias. [18][19] Control of interfacial electronic structure may also be exploited using chemical asymmetry, such as polar discontinuity, to induce charge transfer and orbital hybridization across the interface. [20][21][22] La0.67Sr0.33MnO3 (LSMO) has a pseudo-cubic lattice of 3.880 Å at room temperature, high Curie temperature and high spin polarization. 22 SrTiO3 (STO) is cubic 23 at room temperature with lattice constant ~ 3.905 Å close to that of LSMO. For (00l) orientated heterostructure, the TiO2 and SrO atomic plane of STO are charge-neutral (no planar charge), while MnO2 and La/SrO atomic plane of LSMO have negative and positive planar charge, respectively, as illustrated as blue bar at right bottom of Fig. 1A. The polar discontinuity exists around the LSMO/STO heterostructure, and the charge transfer across the interface occurs in order to avoid the polar catastrophe. [24][25] The final distribution of planar charge (charge on each atomic layer) is illustrated as pink bar at right bottom of Fig. 1A, and the short-range charge transfer is constrained to several unit cells (UC) within the proximity of the interface, [25][26][27] as illustrated as the probability of charge transfer at left bottom of Fig. 1A. Artificial superlattices of ABO3 with multiple interfaces provide a viable way to manipulate the strength of interfacial coupling and corresponding properties. Although properties of LSMO/STO system, such as insulator-metal transition temperature, 28 magnetoelastic effect, 29 and octahedral rotation 30 have been investigated, to date, the electronic charge transfer and related properties at LSMO/STO interfaces remain largely unexplored. Increased number of LSMO/STO interfaces in superlattices may enhance the probability of short-range charge transfer, offering an unique approach to investigate the effects of charge transfer.
In this letter, we report the manipulation of interfacial electronic properties and corresponding magnetic anisotropy in [(La0.67Sr0.33MnO3)n/(SrTiO3)n]m superlattices (SL) with different periodic thickness (n UC LSMO and n UC STO in one period, referred to SLn hereafter). The total thickness was kept at 2n×m ~ 120 UC. The periodic thickness of SLn was systematically changed to obtain different numbers of LSMO/STO interfaces, as illustrated in Fig. 1A. The experimental details are available in supplementary information. The magnetic anisotropy was measured by a superconducting quantum interference device (SQUID) at 150 K and 10 K. At 150K, the magnetic easy axis of SL10 (in-plane) is clearly different from SL2 (outof-plane), as shown in Fig. 1B-1C. The element-specific magnetic information was collected using X-ray magnetic circular dichroism (XMCD). The difference in magnetic easy axis of different samples is further confirmed by angular-dependent Mn L3,2 edge XMCD in Fig. S1.
The contribution from Ti sites to the total magnetic moment is insignificant as calculated by the sum rules 31 from Ti L3, 2 XMCD (Fig. S2). Hence the discussion of magnetic properties in this work will mainly focus on the contribution from Mn sites hereafter. The effect of interfacial octahedral rotation on magnetic anisotropy has been reported, [14][15]33 and current work has measured volume-averaged x-ray half-integer diffraction [34][35][36] to retrieve the information of MnO6 rotation (Fig. S5). There is a significant difference in Mn-O bond angle and bond length in SL10 for the in-plane and out-of-plane directions. In SL2, the bond lengths of both directions are similar, yet the bond angles are different. The larger in-plane bond angle for SL2 (Fig. S5B) suggests a higher in-plane hopping integral that should favor in-plane magnetic easy axis instead of the out-of-plane direction as observed in current work. Therefore it is necessary to consider other factors that are responsible for the observed magnetic anisotropy.
Then high resolution scanning transmission electron microscopy (STEM) were used to probe the local structural information at the LSMO/STO interfaces for n = 2, 6, 10, as shown in Fig. 2. The STO substrate was treated with a TiO2 termination layer before deposition, and an atomically-flat interface between LSMO/STO existed in these samples. is used to separate the contribution of diffraction peak between the film and the substrate in Fig.   S6. It is obvious that the 002 diffraction peak of SL2 film (blue line) shifts to the left with respect to that of the substrate (red line), confirming that the cint of SL2 film is slightly larger than that of STO substrate. The SAED reveals that a = 3.905 Å, and cint = 3.911 Å for SL2, which is consistent with the XRD results. The shoulder at the right side of 002 X-ray diffraction reflection, corresponding to the film, is quite clear for n  6, but disappears for n < 4 ( Fig. S7), revealing that the out-of-plane lattice constant increases with the decrease of n for SLn. Similar enlarged tetragonal ratio around heterostructural interface has been reported in other systems, 39-40 that may affect the electronic structure. SL2 shows the preferred occupancy of 3d3z 2 -r 2 orbitals. [44][45] The resultant depletion of the 3dx 2 -y 2 orbitals weakens the ferromagnetic double-exchange interaction 45 and lowers the paramagneticferromagnetic phase transition temperature in SL2, as shown in Fig. S4.
The coupling of electronic structure across the interface has an obvious effect on the angular-dependent magnetoresistance (AMR) at 10 K. In the AMR measurement, the applied magnetic field was rotated from the in-plane to out-of-plane directions. The in-plane current is always perpendicular to the applied magnetic field as shown in the inset of Fig. 3C. At 10 K, all films show in-plane magnetic anisotropy (Fig. S3). Normally the minimum resistance should occur when the applied magnetic field is in the film plane ( = 90). However, with increasing magnetic field from 1 kOe to 80 kOe, the minimum resistance point of SL3 changes from the inplane to the out-of-plane ( = 0) directions (Fig. 3C). This phenomenon is related to orbital reconstruction 46 with electronic charge transfer across the LSMO/STO interface, especially the compared to that with magnetic field along the out-of-plane direction due to the decreasing electronic velocity. Fig. 3D shows that the AMR effect is largest for SL3 and it decreases with increasing n. Although the resistance of SL2 exceeds the low temperature measurement capability as shown in Fig. S4, the trend of AMR indicates that the effect of orbital reconstruction at LSMO/STO interface increases with decreasing n (enhanced interfacial coupling).
The origin of the preferred occupancy of 3d3z 2 -r 2 at the heterostructure interface had been attributed to arguments such as the interfacial symmetry breaking and new crystal structure phases. 37,44 Based on the existence of the electronic charge transfer and AMR at low temperature, a mechanism due to the orbital hybridization across the interface 47-48 is proposed here (Fig. 4A).
From the perspective of molecular orbitals, the interfacial bonding lowers the energy of 3d orbitals, favoring the occupancy of the 3d3z 2 -r 2 orbital due to a large spatial overlap. When this interfacial orbital forms, the electronegativity of interfacial bonds may facilitate the charge transfer between Mn and Ti 3d orbitals. Due to the correlation between the crystal structure and the electronic structure in 3d oxides, the preferred 3d3z 2 -r 2 orbital occupancy may enlarge the outof-plane cint in order to lower the total energy, as supported by the enlarged cint/a > 1 around the left LSMO/STO interface of each LSMO layer (Fig. 2). The oscillating pattern of cint for SL10 and SL6 ( Fig. 2A-2B) indicates a strong coupling across the interface, which may be understood     The element-specific magnetic information was measured using X-ray magnetic circular dichroism (XMCD). The XMCD at the L3,2 edges of Mn were measured with magnetic field of 5 kOe, which were conducted at beamline 4-ID-C at the APS in ANL using total electron mode for data collection at 150 K.
The XANES spectra and XMCD difference spectra were simulated using the FDMNES code. 51 The final states of electron in the cluster with radius of 6 angstrom were calculated using the multiple-scattering approach with spin-orbit coupling included. To account for the core-hole The SQUID was used to provide the overall signal from the whole sample, and the X-ray magnetic circular dichroism (XMCD) yielded element-specific information. Mn L3,2 edge XMCD was measured to investigate the magnetic anisotropy for n =2, 4, 6, 10, as shown in Fig. S1A-S1D. The measurement setup is illustrated in Fig. S1E. The wave-vector (k) of circular Xray is always parallel to the magnetic field of 5 kOe. The incidence angle θ = 0 corresponds to the magnetic field normal to the film plane and 90 corresponds to magnetic field in the film plane. The magnetic easy axis is identified by maximum XMCD intensity. Fig. S1F shows the summary of XMCD intensity at L3 edge, ~ 639 eV measured at 150 K. When n decreases from 10 to 2, the maximum XMCD intensity changes from θ = 60 for SL10 to θ = 15 for SL2. The magnetic anisotropy at 10 K was measured by SQUID as shown in Fig. S3. The effective anisotropy is separated into constant bulk term Hb and demagnetization term Hd favoring in-plane direction, and interfacial contribution Hi favoring out-of-plane direction. With decreasing n, the interfacial contribution Hi to out-of-plane anisotropy increases, and it shows the same trend with varying n at 150 K. Although the easy axis of all SLn still lies in the film plane at 10 K due to the strong demagnetization effect, these results suggest that the effect of interfacial term Hi on magnetic anisotropy with varying periodic thickness exists at a low temperature of 10 K. Figure S4: The temperature dependence of (A) magnetic and (B) transport properties for SLn. With decreasing n, the Curie temperature decreases and the resistivity increases. See text for more discussion. Half-integer diffraction was used to measure the octahedral rotation pattern. The information of bond angle and bond length was obtained by fitting the intensity of half-integer diffraction peaks. Taking SL4 as example, the rotation pattern of MnO6 is aacaccording to Glaze's note. [34][35][36] The sub-peaks, in Fig. S5A, around L ~ 1.362, 1.638 are the satellite peaks due to the superlattice configuration. The appearance of these satellite peaks shows the good quality of the sample. The octahedral rotations between neighboring manganite layers are correlated, leading to coherent interference of octahedral rotation. The bond angle and bond length were retrieved by fitting peak intensities using a maple code, and the results are shown in Fig. S5B. There is no direct connection between the octahedral rotation and the magnetic anisotropy. Hence, the coupling of crystal structure across the LSMO/STO interface is unlikely the origin of interfacial term Hi. See text for more details.

S5. Half-integer diffraction at room temperature
In addition, according to the fitting results, the MnO6 rotations along the two in-plane axes do not change significantly with periodic thickness (4.3 for SL2 and 4.5 for S10), however, the indicate that as the Mn valence increases from 3.3 to 3.8, the higher energy shoulder at both the L3 and L2 edges, indicated by the green arrow, increases and dominates for valences greater than 3.6. Compared to the measured XAS (Fig. 3A), it suggests that the Mn chemical valence is higher than 3.3. Although the shape of simulated XAS for Mn 3.8+ is closer to the measured XAS compared to other cases, the simulation may overestimate the interfacial coupling and hence the exact chemical valence in SL is not necessarily 3.8. The simulation on Mn L edge XMCD in Fig.   S8B supports the same argument. Figure S9: The comparison of XAS for three samples in perpendicular configuration, which relates to the in-plane electronic hopping integral tin.