No abstract

A new paradigm in multiferroics is observed in BaCoX2O7 (X = As, P) compounds. They consist of one dimensional (1D) antiferromagnetic chains undulated by incommensurate structural modulations with unusually large atomic displacive waves, giving a mixed 1D/2D “real” magnetic topology. The magnetic ground state is antiferromagnetic (AFM) with k = [½ 0 0], leading to a nonmodulated collinear spin lattice despite the aperiodic atomic framework, and allows developing spin‐induced multiferroicity below TN. Severe arguments against the identified mechanisms for type‐II multiferroics, i.e., by inverse Dzyaloshinskii–Moriya, exchange striction and spin‐dependent p–d hybridizations, suggest an original scenario in which the atomic waves, the collinear magnetic structure, and magnetic dipole–dipole interactions may interact as crucial ingredients of the spin‐induced ferroelectric phase. Here, the specific role of the Co2+ spin–orbit coupling in the magnetoelectric (ME) phase diagram is demonstrated by comparison with the novel Heisenberg BaFeP2O7 isomorph, similarly structurally modulated. This compound shows a noncollinear modulated AFM ordering, while no ME coupling is detected in its case. Accordingly, both BaCoX2O7 and BaFeP2O7 also undergo metamagnetic transitions above 5–6 T promoted by the modulated distribution of spin exchanges, but the spin‐flop progressive alignment of the spins in the noncollinear spin structure (Fe2+ case) turns into an abrupt flip‐like transition in the uniaxial spin structure (Co2+ case).

Keywords: exchange bias, cluster spin glass, magnetic memory effect, density functional theory Exchange bias (EB) as large as ~5.5 kOe is observed in SrLaCo0.5Mn0.5O4 which is the highest ever found in any layered transition metal oxides including Ruddlesden-Popper series. Neutron diffraction measurement rules out long-range magnetic ordering and together with dc magnetic measurements suggest formation of short-range magnetic domains. AC magnetic susceptibility, magnetic memory effect and magnetic training effect confirm the system to be a cluster spin glass. By carrying out density functional calculations on several model configurations, we propose that EB is originated at the boundary between Mn-rich antiferromagnetic and Co-rich ferromagnetic domains at the sub-nanoscale. Reversal of magnetization axis on the Co-side alters the magnetic coupling between the interfacial Mn and Co spins which leads to EB. Our analysis infers that the presence of competing magnetic interactions is sufficient to induce exchange bias and thereby a wide range of materials exhibiting giant EB can be engineered for designing novel magnetic memory devices.

No abstract