Low-temperature exchange coupling between Fe2 O3 and FeTi O3

Insight into the mechanism of giant exchange bias in a natural nanoscale intergrowth

Richard J. Harrison*, Suzanne A. McEnroe, Peter Robinson, Brian Carter-Stiglitz, Erika J. Palin, Takeshi Kasama

*Corresponding author for this work

Research output: Contribution to journalArticle

29 Citations (Scopus)


Exchange bias (>1 T at 10 K) has been observed in natural sample of Fe2 O3 containing abundant nanoscale exsolution lamellae of FeTi O3. Exchange bias is first observed below the Néel temperature of FeTi O3 (55 K). Possible interface magnetic structures are explored within the framework of a classical Heisenberg model using Monte Carlo simulations. The simulations predict a threshold value of the Fe2 O3 anisotropy constant, below which Fe3+ spins become tilted out of the basal plane in the vicinity of the interfaces. This tilting creates a c -axis component of magnetization in the Fe2 O3 host that couples to the c -axis magnetization of the FeTi O3 lamellae. Exchange interactions across the interfaces are frustrated when the FeTi O3 lamellae contain an even number of Fe2+ layers, resulting in zero net exchange bias. Lamellae containing an odd number of Fe2+ layers, however, are negatively exchange coupled to the Fe2 O3 host across both (001) bounding surfaces, and are the dominant source of exchange bias. Exchange bias is observed whenever there is a significant c -axis component to both the Fe2 O3 magnetization and the applied field. An exchange bias of 0.9 T was obtained with an anisotropy constant of 0.1 K.

Original languageEnglish
Article number174436
Pages (from-to)1-10
Number of pages10
JournalPhysical Review B: Condensed Matter and Materials Physics
Issue number17
Publication statusPublished - 28 Nov 2007
Externally publishedYes

Fingerprint Dive into the research topics of 'Low-temperature exchange coupling between Fe2 O3 and FeTi O3: Insight into the mechanism of giant exchange bias in a natural nanoscale intergrowth'. Together they form a unique fingerprint.

Cite this