Synthetic antiferromagnets (S-AFMs) composed of strongly correlated oxides have recently been demonstrated to show potential applications in spintronic devices. However, the tunability for the interlayer exchange coupling (IEC) in these all-oxide S-AFMs remains unclear. Here, we report that the IEC in La0.67Ca0.33MnO3/CaRu1-xTixO3 [LCMO/CRTO(x), (0 ≤ x ≤ 0.5)] superlattices (SLs) grown on NdGaO3 (NGO) substrates can be tuned via altering the composition of the spacer layer as well as the growth orientation. The IEC changes from ferromagnetic to antiferromagnetic (AF) type upon doping the spacer CRO with Ti. As the Ti doping level (x) increases, the AF-IEC field (Hex) peaks at x = 0.2, while the Curie temperature (TC) and coercivity (HC) decrease monotonously. Also, we find that the SLs grown on NGO(110) substrates possess larger Hex and smaller HC compared with those grown on NGO(001). Based on these observations, we further fabricate a “hybrid” heterostructure in the form of CRO/LCMO/CRTO(x = 0.5)/LCMO/CRO. Thanks to the collective roles of CRO and CRTO layers, the AF-IEC is maintained and meanwhile the TC is greatly enhanced. The observed high tunability of AF-IEC in LCMO-based S-AFM can primarily be ascribed to the highly tunable properties of the oxide constituents in the AFMs, which are sensitive to both the chemical composition and the growth orientation. Our work paves a way to control the AF-IEC behavior in all-perovskite-oxide S-AFMs, and the results may be instructive to the design of oxide spintronic devices.

Spintronics, termed the spin-controlled electronics, is expected as a next-generation technology after the silicon-based ones. Generally, spintronic devices rely primarily on ferromagnets, and the antiferromagnets (AFMs) are employed as passive parts, such as pinning layers to harden the operating ferromagnetic layers in spin valves and magnetic tunnel junctions.1 With the advancement of nanotechnology, AFMs are now revisited as potential candidates for the active constituents in spintronic devices, such as bistable AFM memory resistors,2 multiple-stable AFM-based memory devices,3 and AFM-based tunnel junctions.4 In fact, the inert reaction to a disturbing magnetic field and the lack of stray fields make the AFMs more promising for high-density storage compared with the ferromagnets.

A drawback of intrinsic AFMs is that they require enormously large external field (H) to induce spin-flop transitions, limiting their broad applications. In this regard, synthetic antiferromagnets (S-AFMs) can be an alternative option which show moderate switching fields. S-AFMs are constructed with ferromagnetic layers periodically interleaved with nonmagnetic spacers, where the adjacent ferromagnetic layers exhibit alternate magnetization due to the antiferromagnetic interlayer exchange coupling (AF-IEC).5–9 The magnetic properties of the S-AFMs, including Fe/Cr, Fe/Au, and CoFeB/Ru multilayers with metallic spacers,5–7 as well as Fe/Si and Fe/MgO multilayers with insulating spacers,8,9 can be manipulated with a moderate magnetic field around a few hundred or thousand gauss. Their AF-IEC behaviors are closely related to the nature of the spacers. For S-AFMs with metallic non-magnetic spacers, the IEC strength oscillates with the spacer thickness due to the Ruderman-Kittel-Kasuya-Yosida (RKKY)-type interaction.6 The periods of the oscillatory coupling are mainly determined by the spanning vectors of the Fermi surface of the spacer layer.10 For those S-AFMs with insulating spacers, the IEC strength decays exponentially as the thickness of the spacer increases, resulting from the spin-polarized quantum tunneling of electrons between the ferromagnetic layers.9 In addition, the AF-IEC behaviors of the S-AFMs can be modified via altering the composition of the non-magnetic spacers.11–13 

The aforementioned S-AFMs are mainly constituted with metals or alloys. They are difficult to be integrated with correlated oxides, which show a diverse array of functional and emergent properties due to the strong correlations between charge, spin, lattice, and orbital degrees of freedom.14 Recently, we have realized all-oxide S-AFMs in the form of all-perovskites La0.67Ca0.33MnO3 (LCMO)/CaRu0.5Ti0.5O3 or La0.7Sr0.3MnO3/CaRu0.5Ti0.5O3 superlattices (SLs),15,16 which display a robust AF-IEC behavior with layer-resolved magnetic switching. In addition to the epitaxial growth with atomic-layer control, the spacer, which helps the ultrathin ferromagnetic layers to retain ferromagnetism at nanoscale thickness and show robust in-plane uniaxial magnetic anisotropy, is of key importance to realize the AF-IEC.15–17 In general, for correlated oxide spacers, their physical properties should sensitively rely on the chemical compositions.17–20 Therefore, it is intriguing to investigate the doping-effect of the spacer layer on the AF-IEC behavior in these manganite-based S-AFMs.

In this paper, we fabricate the S-AFM by using La0.67Ca0.33MnO3 (LCMO) films as the ferromagnetic layer and CaRu1−xTixO3 [CRTO(x), 0 ≤ x ≤ 0.5] as the spacer. We grow the LCMO and CRTO(x) layers alternatively on the orthorhombic NdGaO3 (NGO) substrates to form SLs. The interfacial charge transfer from CRTO(x) to LCMO can effectively maintain the ferromagnetism of the ultrathin manganites.21–23 Upon changing the Ti doping level x, the evolution of the AF-IEC behavior in LCMO-based S-AFMs is systematically investigated. Except for the x = 0 case, all the LCMO/CRTO SLs exhibit clear layer-resolved AF-IEC behavior. However, the dominated mechanism for AF-IEC behavior seems to alter as x increases. Meanwhile, their TC, the IEC field (Hex), and the coercivity (HC) can be significantly tuned by Ti doping. Moreover, we also varied the substrate orientations from NGO(001) to NGO(110) and found that the growth orientation can lead to sizable changes in Hex and HC but negligible changes in TC. Furthermore, we fabricated a “hybrid” CRO/LCMO/CRTO(x = 0.5)/LCMO/CRO heterostructure. The collective effect of CRO and CRTO layers can effectively raise the TC and meanwhile maintain the layer-resolved AF-IEC behavior.

Polycrystalline LCMO and CRTO(0 ≤ x ≤ 0.5) targets were prepared by the conventional solid state reaction method, of which the procedural details can be found in the supplementary material. All the SLs and multilayers were deposited on NGO substrates by pulsed laser deposition using a 248 nm KrF excimer laser with an energy density of 2 J/cm2 and deposition frequency of 3 Hz. During deposition, the substrate temperature (T) and oxygen pressure were maintained at 700 °C and 30 Pa, respectively, yielding a deposition rate of 2 nm/min. After deposition, these heterostructures were in situ annealed for 15 min before cooling down in the same deposition atmosphere. The structures of the films were characterized by x-ray diffraction (XRD) 2θ-ω linear scans at room temperature performed on a high-resolution x-ray diffractometer using Cu Kα1 radiation (Panalytical X’pert). All the magnetization measurements were conducted on a Quantum Design vibrating sample magnetometer (Quantum Design, VSM). For the magnetic field- and temperature-dependent magnetization (M-H and M-T) measurements, H was applied along the in-plane easy-axis [1-10] and [010] directions for the (110)- and (001)-oriented samples, respectively.

The samples are labeled as [m/n]N, where m and n denote the thicknesses of LCMO and CRTO(x) in nanometers, respectively, and N denotes the period, i.e., the repeat number of LCMO layers. Both the bottom and topmost layers are set to be CRTO(x) to ensure the symmetric interfacial environment for each LCMO layer, as schematically shown in Fig. 1(a). First, we study the IEC as a function of the Ti doping, x. The LCMO layer is set at 3.2 nm, and the CRTO(x) layer thickness is optimized to be either 1.6 or 1.2 nm. Figure 1(b) shows the typical XRD 2θ-ω scans around (004) reflections measured from the SLs grown on NGO(001) substrates. The clear Laue fringes and two satellites (indexed as −1 and +1) with high intensity attest the high epitaxial quality with uniform thicknesses and sharp interfaces.

FIG. 1.

Doping-dependent AF-IEC behavior of LCMO/CRTO AFMs. (a) Schematic stacking sequence of the LCMO/CRTO(x) superlattices (SLs), [m/n]10, where m and n are the LCMO and CRTO(x) thicknesses, respectively. Both the bottom and topmost layers are set to be CRTO(x) to ensure a symmetric interfacial environment for each LCMO layer. The Ti (x) doping contents are kept at 0, 0.2, 0.375, and 0.5, with m = 3.2 nm and n ranging from 1.6 nm to 1.2 nm (as denoted). (b) High-resolution x-ray diffraction (XRD) linear scans from the LCMO/CRTO(x) SLs. The (004) main reflections and satellite peaks are indexed as “0” and “±1,” respectively. For clarity, the curves are shifted vertically. (c) Temperature-dependent magnetization (M-T) measured from the SLs. The paramagnetic background from the NGO(001) substrate is also included for comparison. During these measurements, a cooling magnetic field (H) of 200 Oe is applied along the in-plane easy-axis [010]. (d) The corresponding magnetic hysteresis (M-H) loops measured at 100 K from the SLs with the paramagnetic background from the NGO substrate subtracted. Noting that the interlayer exchange coupling fields (Hex1 and Hex2) and coercivities (HC1 and HC2) of the SLs are marked.

FIG. 1.

Doping-dependent AF-IEC behavior of LCMO/CRTO AFMs. (a) Schematic stacking sequence of the LCMO/CRTO(x) superlattices (SLs), [m/n]10, where m and n are the LCMO and CRTO(x) thicknesses, respectively. Both the bottom and topmost layers are set to be CRTO(x) to ensure a symmetric interfacial environment for each LCMO layer. The Ti (x) doping contents are kept at 0, 0.2, 0.375, and 0.5, with m = 3.2 nm and n ranging from 1.6 nm to 1.2 nm (as denoted). (b) High-resolution x-ray diffraction (XRD) linear scans from the LCMO/CRTO(x) SLs. The (004) main reflections and satellite peaks are indexed as “0” and “±1,” respectively. For clarity, the curves are shifted vertically. (c) Temperature-dependent magnetization (M-T) measured from the SLs. The paramagnetic background from the NGO(001) substrate is also included for comparison. During these measurements, a cooling magnetic field (H) of 200 Oe is applied along the in-plane easy-axis [010]. (d) The corresponding magnetic hysteresis (M-H) loops measured at 100 K from the SLs with the paramagnetic background from the NGO substrate subtracted. Noting that the interlayer exchange coupling fields (Hex1 and Hex2) and coercivities (HC1 and HC2) of the SLs are marked.

Close modal

Figure 1(c) shows the M-T curves of the corresponding samples, and the data from a bare NGO(001) substrate are also inserted for comparison. For the x = 0 sample, the Curie temperature (TC) is stabilized at 263 K, which is close to the bulk value. Such a high TC has been ascribed to the interfacial charge transfer from the CRO [CRTO(x = 0)] to the LCMO layer.21–23 As x increases, TC decreases gradually from 240 K (x = 0.2) to 197 K (x = 0.5). This behavior could be attributed to the suppression of the interfacial charge transfer due to the depletion of the Ru 4d itinerant band. For the SLs with x = 0.2, 0.375, and 0.5, the M-T curves exhibit a sudden drop below TC, signifying the existence of AF-IEC in these samples. By contrast, the x = 0 sample does not show any signal related to the AF-IEC in the M-T curve. These observations are also confirmed by the magnetic hysteresis loops, as shown in Fig. 1(d). For the x = 0 sample, the magnetizations of LCMO layers in LCMO/CRO SL behave as ferromagnetic coupling with a standard square loop (Fig. S4 of the supplementary material). For all the other samples with x > 0, the M-H curves show a negligible remanent magnetization (MR). And the loops present multiple discrete steps with two magnetization plateaus at ∼±1/5 of the saturation magnetization (MS). The multiple discrete magnetization plateaus have been attributed to the layer-resolved magnetization reversal as previously reported for the x = 0.5 spacer (Fig. S2 of the supplementary material).15,24 It should be noted that for different Ti contents, the samples show a similar magnetic switching process [Fig. 1(d) and Fig. S3 of the supplementary material].

As illustrated in Fig. 1(d), the IEC fields Hex1 and Hex2 are defined as the offset of the separated loops from zero field, where Hex1 (Hex2) corresponds to the outer (interior) LCMO layers (Fig. S3 of the supplementary material). The interior LCMO layers can receive interactions from the two adjacent LCMO layers, while the outer LCMO layers only receive interaction from one side. To overcome the IEC interaction, the interior layers require twice as much energy as for the outer LCMO layers, yielding Hex2 ≈ 2Hex1. Based on Hex1, the IEC strength can be calculated as J = mMSHex1.25,26 Apparently, when m is fixed, J is in proportion to Hex1. At x = 0, the IEC between the LCMO layers is ferromagnetic with Hex1 = 0. With increasing x, Hex1 (J) changes from 416 Oe (0.072 erg/cm2) at x = 0.2 to 382 Oe (0.066 erg/cm2) at x = 0.375 and then to 143 Oe (0.025 erg/cm2) at x = 0.5. This variation should be determined by the physical nature of the CRTO(x) layers induced by the chemical doping. Slonczewski has proposed that the magnitude of the IEC is closely correlated with the barrier height of the insulating spacer layer,27 which has been corroborated experimentally in the Fe/Fe1−xSix/Fe system.11,13 And the localized impurity or defect states in the insulating barrier layer could also play a crucial role through the resonant tunneling mechanism.28 For example, it has been revealed that the IEC in S-AFMs with insulating MgO spacers could be driven to be AF type due to the oxygen vacancies located in the middle of the MgO layer29 or due to the interfacial oxygen.30 Therefore, we deem that the varied Hex in LCMO/CRTO(x) S-AFMs might be closely linked to the varied barrier height and the impurity located states within the bandgap of the CRTO(x) spacers induced by the Ti doping. Unlike Hex, the HC (including HC1 and HC2) of the SLs decrease monotonously with increasing x. The prominently enhanced HC was also reported in the Ru doped La0.6Sr0.4MnO3 due to the charge transfer and exchange interactions between Mn and Ru sites.31 For the present LCMO/CRTO(x) multilayers, the charge transfer from Ru to Mn across the heterointerfaces enables the stabilization of high-TC ferromagnetic state.21–23 Meanwhile, the exchange coupling between the interfacial Mn and Ru could give rise to the enhancement of HC due to the strong single-ion magnetic anisotropy of Ru4+.32 When the Ru concentration decreases, both the TC and HC decrease remarkably.

It should be noticed that for the samples with different x, the critical spacer-layer thickness for the presence of AF-IEC behavior is different. We have systematically explored the influences of both magnetic and spacer-layer thicknesses (m and n) on the AF-IEC of the SLs. It is found that for all S-AFMs of various Ti doping, the dependence of AF-IEC behavior on the LCMO thickness is similar (Fig. S6 of the supplementary material). Here, we focus on the impact of the CRTO(x) spacer-layer thicknesses, as shown in Fig. 2. We set N = 2 for all the SLs, which are the smallest structural units that display the AF-IEC behavior, as illustrated in the inset of Fig. 2(d). For x = 0.375, the AF-IEC occurs at n = 1.2 and disappears at 0.8 and 1.6 [Fig. 2(a)]. This is similar to the case with x = 0.5, where the AF-IEC is maintained at n = 1.2 and ferromagnetic coupling appears as n varies.15 For x = 0.2, the AF-IEC occurs only at n = 1.6. When n changes, ferromagnetic IEC with a square loop appears [Fig. 2(b)]. The altered spacer thickness, where the AF-IEC occurs, might be ascribed to the varying barrier heights of the CRTO(x) spacers.11 Intriguingly, the case for x = 0.125 seems remarkably different. The LCMO/CRTO(x = 0.125) SLs, [2.8/n]2, with n varying from 0.8 to 4.0 nm show an oscillation-like AF-IEC behavior, resembling the results obtained from the S-AFMs with metallic spacers.6 These LCMO/CRTO(x = 0.125) SLs maintain AF-IEC for 1.6 ≤ n ≤ 3.2 and ferromagnetic IEC for n < 1.6 and n > 3.2 [Fig. 2(c) and Fig. S5 of the supplementary material]. When the CRTO(x = 0.125) layer thickness is 2.4 nm, the largest Hex is observed, as summarized in Fig. 2(d).

FIG. 2.

Doping-dependent and spacer-layer-thickness effects in LCMO/CRTO AFMs. [(a)-(c)] M-H curves measured from the LCMO/CRTO(x) SLs, [2.8/n]2, with x fixed at 0.375, 0.2, and 0.125. (d) The corresponding Hex1 as a function of n. The inset shows the schematic of the SLs with N = 2.

FIG. 2.

Doping-dependent and spacer-layer-thickness effects in LCMO/CRTO AFMs. [(a)-(c)] M-H curves measured from the LCMO/CRTO(x) SLs, [2.8/n]2, with x fixed at 0.375, 0.2, and 0.125. (d) The corresponding Hex1 as a function of n. The inset shows the schematic of the SLs with N = 2.

Close modal

We believe that the different CRTO(x)-thickness-dependent AF-IEC could be ascribed to the altered conductivity of the spacer layers at various Ti doping (Fig. S7 of the supplementary material). For the SLs with high Ti doping levels, the spacers are insulating, and the IEC is mainly controlled by the spin-polarized tunneling as suggested for the Fe/MgO/Fe system.9,28 The tunneling probability of the electrons is exponentially decreased with increasing spacer thickness. As a result, a non-oscillatory dependence with exponentially decreasing IEC is observed for insulating CRTO(x) spacers. However, as the Ti dopant is reduced, the spacers become more conductive. The contribution from electron tunneling becomes negligible. Instead, the magnetization alignment in the adjacent ferromagnetic layers is primarily mediated by the mobile carriers in the spacer layer.33 Based on the RKKY theory, all the electrons scatter from the interface, and the interference between the incoming and scattered waves generates oscillatory probability densities for each electron. The states at the Fermi energy facilitate an oscillatory spin density, while the oscillations of the states below the Fermi energy cancel each other out. The second interface couples to the spin density set up by the first. Since the spin density oscillates with the spacer layer thickness, the IEC strength oscillates as well.34,35 Along these lines, for the LCMO/CRTO(x) system, by utilizing a wide-range chemical doping in the spacer, one may successfully alter the type of the AF-IEC via changing the spacer from the metallic to the insulating state.

We further compare the AF-IEC behavior of LCMO/CRTO(x) SLs grown on orthorhombic NGO(001) and NGO(110) substrates. Since the magnetization reversal in LCMO/CRTO(x) SLs highly relies on the in-plane uniaxial magnetic anisotropy, the IEC may be controlled by different strain states of the SLs. Figure 3(a) illustrates the M-T curves of LCMO/CRTO(x = 0.375) SLs, [2.8/1.2]8, grown on the two different oriented NGO substrates. The identical TC (∼215 K) suggests that the growth orientation has a negligible influence on the ferromagnetism. However, the magnetization dropping temperature of the SL grown on the NGO(110) substrate is higher than the (001)-oriented SL. This means that the AF-IEC behavior can also be affected by the substrate orientation. It is verified by the magnetic hysteresis loops [Fig. 3(b)]. The (110)-oriented SL exhibits smaller HC (HC1 and HC2) but larger Hex (Hex1 and Hex2) compared with the (001)-oriented SL. For the S-AFMs constructed with metals, the AF-IEC strength J is intimately associated with the band structure of the spacer material and the interfacial intermixing.36,37 Given the high-quality epitaxial structure of our all-perovskite SLs grown on both the two oriented substrates, the extrinsic interfacial intermixing could be excluded for the altered Hex. However, the electronic structure of correlated oxides is easily tailored via the epitaxial strain and the interfacial octahedral coupling.38–40 The modulated Hex of the two oriented SLs could be ascribed to the varied band structure of the CRTO(x) layers induced by the different strain states. On the other hand, the epitaxial strains could efficiently modify the magnetic anisotropy of the LCMO layers, hence the HC.41,42 The above observations are consistent with all the LCMO/CRTO(x) SLs with various doping levels, as shown in Fig. 3(c).

FIG. 3.

Dependence of the IEC on the growth orientations. (a) M-T curves measured from LCMO/CRTO(x = 0.375) SLs, [2.8/1.2]8, grown on (110)- and (001)-oriented NGO substrates. During the measurements, H is applied along the in-plane easy-axis [1-10] for (110)-oriented SL and [010] for (001)-oriented SL. (b) The corresponding M-H curves measured at 100 K from the SLs. For each M-H hysteresis loop, the paramagnetic background from the NGO substrate is subtracted. (c) M-H curves measured from the LCMO/CRTO(x) SLs, [2.8/n]2, grown on NGO(110) and NGO(001) substrates and at various doping concentrations x.

FIG. 3.

Dependence of the IEC on the growth orientations. (a) M-T curves measured from LCMO/CRTO(x = 0.375) SLs, [2.8/1.2]8, grown on (110)- and (001)-oriented NGO substrates. During the measurements, H is applied along the in-plane easy-axis [1-10] for (110)-oriented SL and [010] for (001)-oriented SL. (b) The corresponding M-H curves measured at 100 K from the SLs. For each M-H hysteresis loop, the paramagnetic background from the NGO substrate is subtracted. (c) M-H curves measured from the LCMO/CRTO(x) SLs, [2.8/n]2, grown on NGO(110) and NGO(001) substrates and at various doping concentrations x.

Close modal

We now move to map the temperature-dependent magnetic switching and AF-IEC in the LCMO/CRTO(x) system with various x. Figure 4(a) shows the M-H curves of the LCMO/CRTO(x = 0.375) SL, [3.2/1.2]10, grown on the NGO(001) substrate measured at various temperatures. At first glance, both the Hex and HC decline monotonously with increasing temperature, which is consistent with the previous reports.43,44 Based on the magnetic hysteresis loops [Fig. 4(a) and Fig. S8 of the supplementary material], the temperature-dependent Hex1 and HC1 of the LCMO/CRTO(x) SLs with various x and growth orientations are plotted in Fig. 4(b), and the magnetic configurations at various temperatures and magnetic fields are mapped in Figs. 4(c)–4(f) (see the procedural details in Fig. S9 of the supplementary material). For all the samples, the Hex1 and HC1 increase monotonously with decreasing temperature due to the increased AF-IEC strength and magnetic anisotropy energy, which are the two main parameters controlling the magnetic switching for an in-plane magnetized film.45 However, HC1 increases more rapidly in an exponential form as compared with the Hex1. As the Ti dopant increases, both the Hex1 and HC1 decrease over the whole temperature range. For (001)-oriented SLs, the higher the Ti doping level, the thinner the region of the AF state [Figs. 4(c)–4(e)]. When x is increased to 0.5, the different dependence of Hex1 and HC1 on temperature results in a crossover from the AF state to the intermediate state (IS) at ∼50 K and 0 Oe, as marked by the arrows in Figs. 4(b) and 4(e). Although at various x the Hex1 and HC1 have a similar dependence on the temperature, for x = 0.375 and 0.2, the large difference between Hex1 and HC1 makes them difficult to intersect even at low temperatures, and no crossover from the AF state to the IS is observed. This means that for the low-doping SLs, the AF-IEC strength is always higher than the magnetic anisotropy energy even at low temperatures. For the LCMO/CRTO(x = 0.2) SL grown on the NGO(110) substrate, however, under a varied epitaxial strain state, the larger AF-IEC strength leads to a large Hex1 and a relatively large AF state region [Figs. 4(b) and 4(f)]. In sharp contrast to the LCMO/CRTO(0 < x ≤ 0.5) SLs with abundant magnetic configurations, the LCMO/CRO(x = 0) SLs have only the FM state in the magnetic phase diagram [Fig. S10 of the supplementary material].

FIG. 4.

T-dependent IEC of the LCMO/CRTO(x) SLs. (a) M-H curves measured at various T from LCMO/CRTO(x = 0.375) SL, [3.2/1.2]10, grown on NGO(001). Note that the paramagnetic signal from the NGO substrate is subtracted. (b) T-dependent Hex1 and HC1 of the LCMO/CRTO(x) SLs, [3.2/n]10, grown on NGO(110) and NGO(001) substrates with different x. [(c)–(f)] Magnetic configurations mapped at various T and H for LCMO/CRTO(x = 0.2), LCMO/CRTO(x = 0.375), and LCMO/CRTO(x = 0.5) SLs grown on NGO(001) substrates and for LCMO/CRTO(x = 0.2) SL grown on the NGO(110) substrate. Note that “FM” denotes the parallel magnetic alignments of all LCMO layer. “IS” is an intermediate state with antiparallel alignments of all the interior LCMO layers but parallel alignments of the two outer LCMO layers (top and bottom), and “AF” means antiparallel alignments of all adjacent LCMO layers (Fig. S9 of the supplementary material).

FIG. 4.

T-dependent IEC of the LCMO/CRTO(x) SLs. (a) M-H curves measured at various T from LCMO/CRTO(x = 0.375) SL, [3.2/1.2]10, grown on NGO(001). Note that the paramagnetic signal from the NGO substrate is subtracted. (b) T-dependent Hex1 and HC1 of the LCMO/CRTO(x) SLs, [3.2/n]10, grown on NGO(110) and NGO(001) substrates with different x. [(c)–(f)] Magnetic configurations mapped at various T and H for LCMO/CRTO(x = 0.2), LCMO/CRTO(x = 0.375), and LCMO/CRTO(x = 0.5) SLs grown on NGO(001) substrates and for LCMO/CRTO(x = 0.2) SL grown on the NGO(110) substrate. Note that “FM” denotes the parallel magnetic alignments of all LCMO layer. “IS” is an intermediate state with antiparallel alignments of all the interior LCMO layers but parallel alignments of the two outer LCMO layers (top and bottom), and “AF” means antiparallel alignments of all adjacent LCMO layers (Fig. S9 of the supplementary material).

Close modal

So far, we have demonstrated that a highly tunable AF-IEC with layer-resolved magnetic switching can be achieved for LCMO/CRTO(x) SLs via controlling the Ti-doping level of the CRTO spacer and the growth orientation. Now, we propose another approach to tune the AF-IEC and extend the applications of this system. Based on the result that the LCMO/CRO SLs possess a much higher TC than the LCMO/CRTO(x = 0.5) SLs, a “hybrid” structure composed of a LCMO/CRTO(x = 0.5)/LCMO sandwich in the middle of two CRO epilayers is fabricated, which could maintain both high TC and layer-resolved magnetization reversal. A comparative study on the magnetic properties of “hybrid” CRO/LCMO/CRTO(x = 0.5)/LCMO/CRO and “simple” CRTO(x = 0.5)/LCMO/CRTO(x = 0.5)/LCMO/CRTO(x = 0.5) multilayers grown on NGO(110) substrates is conducted. The structures are schematically shown in Fig. 5(a). For the “hybrid” multilayer (sample 1), [1.6/2.8/1.2/2.8/1.6], a high TC of ∼234 K is observed, which is remarkably improved in contrast to that (188 K) of the “simple” multilayer (Sample 2) [Fig. 5(b)]. Indeed, sample 1 shows similar AF-IEC behavior with layer-resolved magnetization reversal as sample 2 [Fig. 5(c)].

FIG. 5.

AF-IEC of the “hybrid” S-AFMs. (a) Schematic of the “hybrid” CRO/LCMO/CRTO(x = 0.5)/LCMO/CRO structure (sample 1), as compared with the CRTO(x = 0.5)/LCMO/CRTO(x = 0.5)/LCMO/CRTO(x = 0.5) epitaxial multilayer (sample 2). (b) and (c) show M-T and M-H curves measured from the two multilayers grown on NGO(110) substrates. (d) M-T curves measured from the CRO/LCMO/CRTO(x = 0.5)/LCMO/CRO multilayers grown on NGO(110) substrates with various LCMO layer thicknesses (m). The inset shows the TC variation with m. The corresponding M-H curves measured at 100 K are shown in (e). (f) M-T curves measured from the CRO/LCMO/CRTO(x = 0.5)/LCMO/CRO multilayers grown on NGO(110) and NGO(001) substrates. The inset shows the corresponding M-H curves measured at 140 K.

FIG. 5.

AF-IEC of the “hybrid” S-AFMs. (a) Schematic of the “hybrid” CRO/LCMO/CRTO(x = 0.5)/LCMO/CRO structure (sample 1), as compared with the CRTO(x = 0.5)/LCMO/CRTO(x = 0.5)/LCMO/CRTO(x = 0.5) epitaxial multilayer (sample 2). (b) and (c) show M-T and M-H curves measured from the two multilayers grown on NGO(110) substrates. (d) M-T curves measured from the CRO/LCMO/CRTO(x = 0.5)/LCMO/CRO multilayers grown on NGO(110) substrates with various LCMO layer thicknesses (m). The inset shows the TC variation with m. The corresponding M-H curves measured at 100 K are shown in (e). (f) M-T curves measured from the CRO/LCMO/CRTO(x = 0.5)/LCMO/CRO multilayers grown on NGO(110) and NGO(001) substrates. The inset shows the corresponding M-H curves measured at 140 K.

Close modal

We further explore the AF-IEC behavior as a function of the LCMO layer thickness in the “hybrid” multilayers. M-T curves of these multilayers, [1.6/m/1.2/m/1.6], with m varying from 2.4 to 3.6 nm are shown in Fig. 5(d). As the thickness of LCMO layer m is reduced from 3.6 to 2.4 nm, TC of the samples decrease gradually from 236 to 217 K. Indeed, such a finite-size effect is commonly observed in ferromagnetic oxide thin films, although the CRO layers have a stronger effect to suppress the “dead layer” effect than the CRTO(x = 0.5) layers.21–23 All the M-T curves show a sudden drop below TC, signifying the existence of AF-IEC in these “hybrid” multilayers. Figure 5(e) shows the corresponding hysteresis loops. Regardless of the thickness of LCMO layers, an AF state with magnetization step at MR ∼ 0 is observed, which is in good agreement with the drop of magnetization below the TC. As m increases, Hex decreases monotonously, which coincides with all-CRTO(x = 0.5)-based S-AFMs (Fig. S4 of the supplementary material).

Finally, the AF-IEC behavior in two “hybrid” multilayers grown on (110)- and (001)-oriented NGO substrates is scrutinized. Albeit with an identical TC of ∼234 K, the epitaxial strain shows a strong impact on the magnetization dropping temperature. The magnetization of the (110)-oriented sample drops sharply at ∼215 K, while the (001)-oriented one drops at a much lower temperature (133 K) [Fig. 5(f)]. It means that the AF-IEC in the “hybrid” S-AFMs is also manipulated by the epitaxial strain. To validate this statement, M-H measurements of these two multilayers were performed, and the measuring temperature is marked in the M-T curves. As shown in the inset of Fig. 5(f), when the magnetic field is ramped down from the positive saturation field, the magnetization of the (110)-oriented sample drops abruptly to zero at a positive magnetic field, implying that the magnetizations of the two LCMO layers are antiparallel around zero field. However, for the (001)-oriented sample, the magnetization drops at a negative field. In other words, magnetizations cannot be totally compensated without the help of a field applied in the reverse direction. Clearly, the (110)-oriented sample maintains the AF-IEC over a much wider temperature range compared to the (001)-oriented one. This distinction is provoked by the fact that the (001)-oriented sample has larger HC and smaller Hex compared with the (110)-oriented one, resulting from the competition between the AF-IEC strength and magnetic anisotropy energy as explained above for the “simple” S-AFMs.

Therefore, for the “hybrid” structure, although the CRO epilayers enhance the HC to some extent, the nature of the AF-IEC is essentially the same, such as the layer-resolved magnetization reversal, the orientation-dependent AF-IEC, and the layer-thickness dependence of the AF-IEC. Meantime, the TC of this structure is effectively raised. This result would shed some light on designing new type functional oxide devices. The properties of the S-AFMs can be improved purposefully on the basis of the LCMO/CRTO(x)/LCMO AFM unit. For example, one can construct a “hybrid” structure by heterostructuring the LCMO/CRTO(x)/LCMO sandwich with superconductive or ferroelectric oxide epilayers and select the appropriate Ti dopant and substrate to achieve new functionalities.

In summary, we have studied the IEC in LCMO/CRTO(0 ≤ x ≤ 0.5) SLs. Robust AF-IEC with layer-resolved magnetization reversal is observed for the whole spacer composition range except for x = 0. We demonstrate that a broad tunable range of the AF-IEC behaviors including the IEC type, Hex, HC, and TC is effectively modulated by the Ti doping and growth orientation (Fig. 6). By means of increasing the Ti dopant, the AF-IEC shows a sign of crossover from an oscillation-type to an exponential-decay-type as a function of the CRTO(x) layer thickness. This transition could be accompanied with the transition from metallic to insulating state of the spacer. Owing to the controllable properties of the CRTO(x) spacer (depending on the barrier height and impurity located states, which need further investigations), the magnetic anisotropy of the LCMO layer, the interfacial charge transfer between LCMO and CRTO(x) on the chemical composition of the CRTO(x) spacer, Hex, HC, and TC are highly tunable by the Ti doping. The growth orientation also has a significant impact on the band structure of CRTO(x) and magnetic anisotropy of LCMO, thus the Hex and HC. Moreover, the “hybrid” AFM structures provide us an alternative path to modify the AF-IEC, thus engineering the devices.

FIG. 6.

Modulations of the AF-IEC behavior in LCMO/CRTO S-AFMs. (a) TC as a function of x for LCMO/CRTO(x) SLs, [2.8/n]2. In (b) and (c), HC1 and Hex1 (J) of the LCMO/CRTO(x) SLs grown on NGO(110) and NGO(001) substrates are plotted against x. The IEC type of the LCMO/CRTO(x) SLs with various Ti doping is roughly indicated by the arrows on the top of the figure. Note that all the SLs presented here possess the optimal AF-IEC behavior. The thicknesses of CRTO(x) layers are optimized to be 2.4 nm (x = 0 and 0.125), 1.6 nm (x = 0.2), and 1.2 nm (x = 0.375 and 0.5), respectively.

FIG. 6.

Modulations of the AF-IEC behavior in LCMO/CRTO S-AFMs. (a) TC as a function of x for LCMO/CRTO(x) SLs, [2.8/n]2. In (b) and (c), HC1 and Hex1 (J) of the LCMO/CRTO(x) SLs grown on NGO(110) and NGO(001) substrates are plotted against x. The IEC type of the LCMO/CRTO(x) SLs with various Ti doping is roughly indicated by the arrows on the top of the figure. Note that all the SLs presented here possess the optimal AF-IEC behavior. The thicknesses of CRTO(x) layers are optimized to be 2.4 nm (x = 0 and 0.125), 1.6 nm (x = 0.2), and 1.2 nm (x = 0.375 and 0.5), respectively.

Close modal

See the supplementary material for characterizations of targets; the magnetization switching behavior of the S-AFMs; the doping-dependent magnetization switching process in LCMO/CRTO(x) S-AFMs; the CRTO(x)-thickness-dependent AF-IEC behavior of LCMO/CRO and LCMO/CRTO(x = 0.125) SLs; the LCMO-thickness-dependent AF-IEC behavior of LCMO/CRTO(x) SLs at various doping contents; resistivity measurements of the CRTO(x) single-layer films; and the mapping of magnetic configurations at various temperatures and fields.

This work was supported by the National Basic Research Program of China (Grant Nos. 2016YFA0401003 and 2017YFA0403502), the NSF of China (Grant Nos. 11474263, 11574324, U1432251, 11804342, and 51872278), Hefei Science Center CAS, and Postdoctoral Science Foundation (No. 2018M632557).

1.
S. A.
Wolf
,
D. D.
Awschalom
,
R. A.
Buhman
,
J. M.
Daughton
,
S. v.
Molnar
,
M. L.
Roukes
,
A. Y.
Chtchelkanova
, and
D. M.
Treger
,
Science
294
,
1488
(
2001
).
2.
X.
Marti
,
I.
Fina
,
C.
Frontera
,
J.
Liu
,
P.
Wadley
,
Q.
He
,
R. J.
Paull
,
J. D.
Clarkson
,
J.
Kudrnovsky
,
I.
Turek
,
J.
Kunes
,
D.
Yi
,
J. H.
Chu
,
C. T.
Nelson
,
L.
You
,
E.
Arenholz
,
S.
Salahuddin
,
J.
Fontcuberta
,
T.
Jungwirth
, and
R.
Ramesh
,
Nat. Mater.
13
,
367
(
2014
).
3.
D.
Kriegner
,
K.
Vyborny
,
K.
Olejnik
,
H.
Reichlova
,
V.
Novak
,
X.
Marti
,
J.
Gazquez
,
V.
Saidl
,
P.
Nemec
,
V. V.
Volobuev
,
G.
Springholz
,
V.
Holy
, and
T.
Jungwirth
,
Nat. Commun.
7
,
11623
(
2016
).
4.
B. G.
Park
,
J.
Wunderlich
,
X.
Marti
,
V.
Holy
,
Y.
Kurosaki
,
M.
Yamada
,
H.
Yamamoto
,
A.
Nishide
,
J.
Hayakawa
,
H.
Takahashi
,
A. B.
Shick
, and
T.
Jungwirth
,
Nat. Mater.
10
,
347
(
2011
).
5.
S. S.
Parkin
,
N.
More
, and
K. P.
Roche
,
Phys. Rev. Lett.
64
,
2304
(
1990
).
6.
J.
Unguris
,
R. J.
Celotta
, and
D. T.
Pierce
,
J. Appl. Phys.
75
,
6437
(
1994
).
7.
N.
Wiese
,
T.
Dimopoulos
,
M.
Rührig
,
J.
Wecker
, and
G.
Reiss
,
J. Appl. Phys.
98
,
103904
(
2005
).
8.
G. J.
Strijkers
,
J. T.
Kohlhepp
,
H. J. M.
Swagten
, and
W. J. M. d.
Jonge
,
Phys. Rev. Lett.
84
,
1812
(
2000
).
9.
J.
Faure-Vincent
,
C. T.
Bellouard
,
E.
Popova
,
M.
Hehn
,
F.
Montaigne
, and
A.
Schuhl
,
Phys. Rev. Lett.
89
,
107206
(
2002
).
10.
M. D.
Stiles
,
Phys. Rev. B
48
,
7238
(
1993
).
11.
Y.
Endo
,
O.
Kitakami
, and
Y.
Shimada
,
J. Appl. Phys.
87
,
6836
(
2000
).
12.
C.
Quirós
,
J. I.
Martín
,
L.
Zárate
,
M.
Vélez
, and
J. M.
Alameda
,
Phys. Rev. B
71
,
024423
(
2005
).
13.
Y.
Endo
,
O.
Kitakami
, and
Y.
Shimada
,
Phys. Rev. B
59
,
4279
(
1999
).
14.
H. Y.
Hwang
,
Y.
Iwasa
,
M.
Kawasaki
,
B.
Keimer
,
N.
Nagaosa
, and
Y.
Tokura
,
Nat. Mater.
11
,
103
(
2012
).
15.
B.
Chen
,
H.
Xu
,
C.
Ma
,
S.
Mattauch
,
D.
Lan
,
F.
Jin
,
Z.
Guo
,
S.
Wan
,
P.
Chen
,
G.
Gao
,
F.
Chen
,
Y.
Su
, and
W.
Wu
,
Science
357
,
191
(
2017
).
16.
H.
Xu
,
F.
Chen
,
B.
Chen
,
F.
Jin
,
C.
Ma
,
L.
Xu
,
Z.
Guo
,
L.
Qu
,
D.
Lan
, and
W.
Wu
,
Phys. Rev. Appl.
10
,
024035
(
2018
).
17.
H.
Xu
,
S.
Wan
,
B.
Chen
,
C.
Ma
,
F.
Jin
,
Z.
Guo
,
D.
Lan
,
F.
Chen
, and
W.
Wu
,
Appl. Phys. Lett.
110
,
082402
(
2017
).
18.
V.
Hardy
,
B.
Raveau
,
R.
Retoux
,
N.
Barrier
, and
A.
Maignan
,
Phys. Rev. B
73
,
094418
(
2006
).
19.
B.
Chen
,
P.
Chen
,
H.
Xu
,
F.
Jin
,
Z.
Guo
,
D.
Lan
,
S.
Wan
,
G.
Gao
,
F.
Chen
, and
W.
Wu
,
ACS Appl. Mater. Interfaces
8
,
34924
(
2016
).
20.
G.-L.
Liu
,
J.-S.
Zhou
, and
J.
Goodenough
,
Phys. Rev. B
70
,
224421
(
2004
).
21.
M.
Ziese
,
F.
Bern
,
E.
Pippel
,
D.
Hesse
, and
I.
Vrejoiu
,
Nano Lett.
12
,
4276
(
2012
).
22.
P. F.
Chen
,
B. B.
Chen
,
X. L.
Tan
,
H. R.
Xu
,
X. F.
Xuan
,
Z.
Guo
,
F.
Jin
, and
W. B.
Wu
,
Appl. Phys. Lett.
103
,
262402
(
2013
).
23.
B. B.
Chen
,
P. F.
Chen
,
H. R.
Xu
,
X. L.
Tan
,
F.
Jin
,
Z.
Guo
,
B. W.
Zhi
, and
W. B.
Wu
,
Appl. Phys. Lett.
104
,
242416
(
2014
).
24.
M.
Charilaou
,
C.
Bordel
, and
F.
Hellman
,
Appl. Phys. Lett.
104
,
212405
(
2014
).
25.
S. S.
Parkin
,
Phys. Rev. Lett.
67
,
3598
(
1991
).
26.
L.
Li
,
F.
Zhang
,
N.
Wang
,
Y.
Lv
,
X.
Han
, and
J.
Zhang
,
J. Appl. Phys.
108
,
073908
(
2010
).
27.
J. C.
Slonczewski
,
Phys. Rev. B
39
,
6995
(
1989
).
28.
M. Y.
Zhuravlev
,
E. Y.
Tsymbal
, and
A.
Vedyayev
,
Phys. Rev. Lett.
94
,
026806
(
2005
).
29.
T.
Katayama
,
S.
Yuasa
,
J.
Velev
,
M. Y.
Zhuravlev
,
S. S.
Jaswal
, and
E. Y.
Tsymbal
,
Appl. Phys. Lett.
89
,
112503
(
2006
).
30.
H.-C.
Wu
,
S.
Arora
,
O.
Mryasov
, and
I.
Shvets
,
Appl. Phys. Lett.
92
,
182502
(
2008
).
31.
H.
Yamada
,
M.
Kawasaki
, and
Y.
Tokura
,
Appl. Phys. Lett.
86
,
192505
(
2005
).
32.
T.
Harano
,
G.
Shibata
,
K.
Ishigami
,
Y.
Takashashi
,
V. K.
Verma
,
V. R.
Singh
,
T.
Kadono
,
A.
Fujimori
,
Y.
Takeda
,
T.
Okane
,
Y.
Saitoh
,
H.
Yamagami
,
T.
Koide
,
H.
Yamada
,
A.
Sawa
,
M.
Kawasaki
,
Y.
Tokura
, and
A.
Tanaka
,
Appl. Phys. Lett.
102
,
222404
(
2013
).
33.
M. D.
Stiles
,
J. Magn. Magn. Mater.
200
,
322
(
1999
).
34.
P.
Bruno
,
J. Appl. Phys.
76
,
6972
(
1994
).
35.
K.
Yoshimura
,
T.
Imai
,
T.
Kiyama
,
K.
Thurber
,
A.
Hunt
, and
K.
Kosuge
,
Phys. Rev. Lett.
83
,
4397
(
1999
).
36.
M.
van Schilfgaarde
,
F.
Herman
,
S. S.
Parkin
, and
J.
Kudrnovsky
,
Phys. Rev. Lett.
74
,
4063
(
1995
).
37.
J.
Blinowski
and
P.
Kacman
,
Phys. Rev. B
64
,
045302
(
2001
).
38.
A. T.
Zayak
,
X.
Huang
,
J. B.
Neaton
, and
K. M.
Rabe
,
Phys. Rev. B
74
,
094104
(
2006
).
39.
J.
Nichols
,
J.
Terzic
,
E. G.
Bittle
,
O. B.
Korneta
,
L. E.
De Long
,
J. W.
Brill
,
G.
Cao
, and
S. S. A.
Seo
,
Appl. Phys. Lett.
102
,
141908
(
2013
).
40.
J. M.
Rondinelli
and
N. A.
Spaldin
,
Adv. Mater.
23
,
3363
(
2011
).
41.
L. M.
Berndt
,
V.
Balbarin
, and
Y.
Suzuki
,
Appl. Phys. Lett.
77
,
2903
(
2000
).
42.
H.
Boschker
,
M.
Mathews
,
E. P.
Houwman
,
H.
Nishikawa
,
A.
Vailionis
,
G.
Koster
,
G.
Rijnders
, and
D. H. A.
Blank
,
Phys. Rev. B
79
,
214425
(
2009
).
43.
N.
Moutis
,
C.
Christides
,
I.
Panagiotopoulos
, and
D.
Niarchos
,
Phys. Rev. B
64
,
094429
(
2001
).
44.
Z.
Celinski
,
B.
Heinrich
,
J.
Cochran
,
W.
Muir
,
A.
Arrott
, and
J.
Kirschner
,
Phys. Rev. Lett.
65
,
1156
(
1990
).
45.
P.
Poulopoulos
,
U.
Bovensiepen
,
M.
Farle
, and
K.
Baberschke
,
Phys. Rev. B
57
,
R14036
(
1998
).

Supplementary Material