The long-range exchange bias in ultrathin paramagnetic LaNiO3-based heterostructure

Abstract

With the rapid development of synthesis techniques, it has become possible to control transition metal oxide heterostructures by manipulating atomic flatness. 1 In the nickelate family, LaNiO 3 (LNO), is an exception since it exhibits metallic paramagnetic behavior over all temperatures. 2 Recently, the antiferromagnetic (AFM) state in ultrathin LNO layers has been theoretically predicted. 3 However, as far as we know, this has not yet been experimentally realized. The long-range exchange bias effect 4, where a non-magnetic material is inserted into ferromagnetic/AFM heterostructure, is a useful way to exclude the contribution of exchange bias by the interfacial coupling between Ni and Mn ions. 5, 6 In this work, high quality LNO/STO $(n)$/LSMO superlattices are grown using a pulse laser deposition system. Both the exchange bias and coercive fields decrease monotonically with the spacer thickness. The role of antiferromagnetic behavior in the ultrathin LNO layer is essential for understanding the behavior of this exchange bias system. Hopefully, this work reveals new perspectives for simplifying the behavior of confusing systems. Fig. 1a shows the configuration of the investigated LNO/STO $( n)$/ LSMO superlattices, which consists of a non-magnetic STO layer. The STO layer changed from 1 u.c. to 7 u.c. and inserted in every interface between the LNO and LSMO layers. Typical x-ray diffraction patterns measured with Cu-K α radiation around the (002) STO Bragg peak of the investigated specimens are shown in Fig. 1b. Satellite peaks SL+1 and SL+2 are observed, indicating smooth interfaces in the LNO/STO/LSMO superlattices.25 To support of the claims of coherent and epitaxial heterostructure growth and the absence of secondary phases, an atomically resolved high-angle annular dark field scanning transmission electron microscopy image of the LNO(2)/STO(2)/LSMO(5) superlattice is shown in Figure 1c. The good layering with atomically flat interfaces is readily identified from this image. Typical in-plane magnetic loops at 5 K in the LNO(2)/LSMO(5) superlattice after positive field cooling (FC) and negative field cooling processes are shown in Figure 2a. In this field-cooling magnetic loop, the exchange field and coercivity are 603 Oe and 1999 Oe, respectively.8 In Figure 2b, the $H_{\mathrm{EB}}$ and $H_{\mathrm{C}}$ decrease to 230 Oe and 1181 Oe for the LNO(2)/STO(2)/LSMO(5), respectively. In Figure 2c, when the thickness of the inserted STO spacer layer is 6 u.c., the negative coercive field and positive coercive field values after the ±5 kOe FC procedures are consistent with each other. This indicates that the exchange bias effect is absent. The dependence of the exchange bias field and coercivity on the LNO/STO $( n)$/LSMO system STO spacer layer thickness is shown in Figure 2d. As the spacer layer thickness increases, the exchange bias field decreases monotonically and smoothly. At a maximum thickness of 6 u.c., the exchange coupling vanishes. According to these magnetic measurements, long-range exchange coupling can be confirmed in the LNO/STO $( n)$/LSMO system. The work was supported by National Key R&D Program of China (No. 2017YFB0405703), NSFC (Nos. 61434002, 51571136, 11274214), and the Special Funds of Sanjin Scholars Program.