Critical focused issues
Role of the antiferromagnetic bulk spins in exchange bias

https://doi.org/10.1016/j.jmmm.2016.04.065Get rights and content

Highlights

  • We address the role of bulk antiferromagnetic spins in the exchange bias phenomenon.

  • Significant experiments on how bulk AFM spins determine exchange bias are highlighted.

  • We explain the model that accounts for experimental results.

Abstract

This “Critical Focused Issue” presents a brief review of experiments and models which describe the origin of exchange bias in epitaxial or textured ferromagnetic/antiferromagnetic bilayers. Evidence is presented which clearly indicates that inner, uncompensated, pinned moments in the bulk of the antiferromagnet (AFM) play a very important role in setting the magnitude of the exchange bias. A critical evaluation of the extensive literature in the field indicates that it is useful to think of this bulk, pinned uncompensated moments as a new type of a ferromagnet which has a low total moment, an ordering temperature given by the AFM Néel temperature, with parallel aligned moments randomly distributed on the regular AFM lattice.

Introduction

Exchange bias (EB) is characterized by the shift of the magnetic hysteresis loop along the field axis, generally observed in Antiferromagnetic (AFM)/Ferromagnetic (FM) bilayered hybrids [1]. This interesting, basic research effect is also the basis for many applications in the spintronics area such as magnetic data storage and sensor devices. The essential characteristics which determine the properties of an exchange biased system are: the magnitude of the shift, its sign, asymmetry of the hysteresis loop, blocking temperature (above which the EB disappears), training effect and time dependence. Although much work has been dedicated to understand the phenomenology of EB [2], each one of these important characteristics presents interesting puzzles, which give complementary clues regarding the essential physics of the effect. EB is generally considered to be a consequence of the interfacial interaction between the FM and AFM constituents [3]. This is attributed to the pinned, uncompensated magnetic moments [1], [4], [5], [6] at the interface originating from the AFM.

Originally it was postulated that only the AFM interface controls the EB, i.e. EB is a purely interfacial phenomenon in which the role of the AFM bulk is restricted to pinning the interfacial magnetic moments. However, the interface is always coupled to the AFM bulk. Therefore the AFM bulk may affect the precise magnetic state of the interface with the consequent effect on the exchange bias. There is by now much compelling evidence that the bulk magnetic state of the AFM may affect the exchange bias, which implies that EB is not a purely interfacial phenomenon. Although its ultimate origin is the exchange interaction at the AFM/FM interface, the pinned, uncompensated spin distribution at the interface might be determined by the AFM bulk. In this “Critical Focused Issue” we highlight the role and microscopic origin of the pinned, uncompensated moments (PUM) present in the bulk of the AFM. More specifically, we emphasize the important experiments, which provide clues regarding the microscopic mechanism that governs exchange bias. We conclude by describing potential new directions in which this field can move and connected open questions.

EB is initiated by cooling the FM/AFM bilayer in an externally applied magnetic field below the AFM Néel temperature. The exchange coupling between the FM and the AFM, which shifts the hysteresis loop along the field axis, is determined by an effective “exchange field” or by a “unidirectional” anisotropy energy. The AFM crystallinity, its morphology (e.g., grains) and intrinsic anisotropy are crucial parameters which determine the magnitude of the EB. In general, two types of exchange-biased systems, which show distinctly different behavior, can be distinguished. Type 1 are highly textured or epitaxial systems such as FeF2 or CoO. On the other hand, type 2 are usually small-grained, polycrystalline systems, such as the classic archetypes IrMn or FeMn. It should be noted that the anisotropy energy, the central quantity determining the EB magnitude, depends on both the effective anisotropy constant as well as the crystal volume. Consequently, some polycrystalline AFMs may behave as either type 1 or type 2 depending on the crystallite size, the inter-crystallite magnetic coupling, which may lead to a larger effective particle volume, and the microstructure (e.g., growth mode), which may yield an effective increase of the anisotropy constant [7], [8]. It is also important that the properties of the exchange bias bilayers are not only determined by the AFM's physical structure, but also its magnetic structure. Even if the crystallographic orientation of the AFM/FM interface is well defined, the spin orientation may become very complicated since in some cases equivalent crystallographic directions may not be magnetically equivalent. For instance, NiO is a classic example in which the (111) crystallographic plane has 4 structurally equivalent, but magnetically inequivalent directions.

Type 1 systems have in general a large exchange bias, the blocking temperature coincides with the AFM Néel temperature, and training and time dependences are practically absent. In type 2 systems the blocking temperature can be considerably reduced compared to the Néel temperature, and they may exhibit large training as well as time dependent effects. The change from negative (NEB) to positive (PEB) exchange bias shift can be present in both types of systems if the exchange coupling at the interface is antiferromagnetic and the surface layer of the AFM couples to the increasing external cooling field [9], [10]. Like with many other situations in physics, there is no clear demarcation between type 1 and type 2 systems; these are just two extreme cases. For instance, there may be situations in which the blocking temperature coincides with the AFM Néel temperature but the systems exhibit large training effects [11], [12], [13]. This may occur even within the same combination of materials, since sometimes there are large structural differences within the same system. A classic example is Co/CoO where the CoO may be polycrystalline, textured, or epitaxial depending on the specific preparation method.

There are a number of additional extrinsic experimental complications, which may cause confusion. Sometimes the exchange bias is much smaller than the coercivity, the hysteresis loops are sheared and/or there are large contributions from other (presumably irrelevant) parts of the sample such as substrates. In either case, small shifts along the field axis may be caused by artifacts such as vertical loop shifts and may complicate the identification of the EB. Other important issues which are not discussed here include intrinsic and extrinsic effects such as interfacial roughness, interdiffusion, variation in thickness and reduced magnetization and/or formation of interfacial compounds at interfaces and surfaces. Of course, in order to avoid erroneous conclusions the physical and chemical properties of these systems must be thoroughly characterized quantitatively using a comprehensive battery of tests.

Section snippets

Issues

In spite of all the above-mentioned complications, it seems that a single physical mechanism determines the exchange bias. There is overwhelming evidence that the origin of EB resides in the pinned, uncompensated moments (PUM) present in the AFM. The only possible exceptions are interfaces with sizeable Dzyaloshinskii–Moriya interaction [14], which breaks mirror symmetry and may lead to EB at perfectly compensated interfaces. Moreover, it is crucial that in addition to the PUM there is evidence

Experimental evidence

In this section, we will summarize the different classes of experiments which imply that bulk PUM play a major role in exchange bias.

Models

Several models have been developed in the past to account for effects related to the EB phenomenon. Many of these models make different assumptions regarding the nature of the pinned, uncompensated moments (PUM) that give rise to EB and some of them have never been tested by numerical simulations. To achieve EB there must be some interface magnetization in the AFM which couples to the FM. A purely interfacial moment maybe expected from atomically flat interfaces with geometrically (perhaps

Conclusions

Exchange bias systems were divided into two classes: type 1, highly ordered epitaxial or textured systems whose behavior is largely governed by spin configurations produced by the structure, and type 2, polycrystalline systems for which thermal fluctuations play a major role. We have presented experimental and theoretical evidence that exchange bias in type 1 ferromagnetic/antiferromagnetic heterostructures is in grand part affected by the behavior of pinned uncompensated moments in the bulk of

Acknowledgments

This is a highly collaborative manuscript. The outline was conceived jointly, the data and models were extensively debated and the paper was written by multiple iterations between all the coauthors. The research at UCSD was supported by the Office of Basic Energy Science, U.S. Department of Energy, BES-DMS funded by the Department of Energy's Office of Basic Energy Science, DMR under grant DE FG02 87ER-45332. Support from the Spanish MINECO (Grant nos. MAT2015-68772-P, FIS2013-45469 and

References (52)

  • J. Sort et al.

    Exchange bias in antiferromagnetic–ferromagnetic–antiferromagnetic structures with out-of-plane magnetization

    Phys. Rev. B

    (2005)
  • J. Nogues et al.

    Positive exchange bias in FeF2-Fe bilayers

    Phys. Rev. Lett.

    (1996)
  • T. Gredig et al.

    Unidirectional coercivity enhancement in exchange-biased Co/CoO

    Appl. Phys. Lett.

    (2002)
  • A. Hoffmann

    Symmetry driven irreversibilities at ferromagnetic–antiferromagnetic interfaces

    Phys. Rev. Lett.

    (2004)
  • S. Brems et al.

    Reversing the training effect in exchange biased CoO/Co bilayers

    Phys. Rev. Lett.

    (2005)
  • C. Binek

    Training of the exchange-bias effect: a simple analytic approach

    Phys. Rev. B

    (2004)
  • R. Yanes et al.

    Exchange bias driven by Dzyaloshinskii–Moriya interactions

    Phys. Rev. Lett.

    (2013)
  • S. Roy et al.

    Depth profile of uncompensated spins in an exchange bias system

    Phys. Rev. Lett.

    (2005)
  • Y. Imry et al.

    Random-field instability of the ordered state of continuous symmetry

    Phys. Rev. Lett.

    (1975)
  • M.R. Fitzsimmons et al.

    Influence of interfacial disorder and temperature on magnetization reversal in exchange-coupled bilayers

    Phys. Rev. B

    (2001)
  • H. Ohldag et al.

    Spectroscopic identification and direct imaging of interfacial magnetic spins

    Phys. Rev. Lett.

    (2001)
  • C. Chien et al.

    Hybrid domain walls and antiferromagnetic domains in exchange-coupled ferromagnet/antiferromagnet bilayers

    Phys. Rev. B

    (2003)
  • Y. Fan et al.

    Exchange bias of the interface spin system at the Fe/MgO interface

    Nat. Nanotechnol.

    (2013)
  • M. Ali et al.

    Exchange bias using a spin glass

    Nat. Mater.

    (2007)
  • P. Miltényi et al.

    Diluted antiferromagnets in exchange bias: proof of the domain state model

    Phys. Rev. Lett.

    (2000)
  • H. Shi et al.

    Exchange bias in FexZn1−xF2/Co bilayers

    J. Appl. Phys.

    (2002)
  • Cited by (54)

    • Exchange bias and interface-related effects in two-dimensional van der Waals magnetic heterostructures: Open questions and perspectives

      2023, Journal of Alloys and Compounds
      Citation Excerpt :

      In FM/AFM systems, the EB effect was initially ascribed to a unidirectional pinning of the FM layer by the adjacent AFM layer causing a shift in the switching field of the FM layer away from the H = 0 axis by an amount HEB (the EB field) and is accompanied by an increase in the coercive field (HC) of the FM layer [9]. However, recent research has shown that the pinned uncompensated moments in FM/AFM systems generated by defects at the FM/AFM interface play a dominant role in determining the EB magnitude [28,29]. These disordered interfacial spins behave like “spin clusters” (SCs), which are analogous to spin glasses (SGs) at the FM/AFM interface.

    View all citing articles on Scopus
    View full text