Coupling of single molecule magnets to antiferromagnetic substrates

TbPc2 single molecule magnets coupled to oxide (left) and metal (right) antiferromagnetic films. The non-hysteretic magnetization loops of Tb in (a) and (b) for different field-cooling conditions indicate that there is no exchange bias in TbPc2/CoO. The vertically shifted Mn loop in (c) and hysteretic Tb magnetization in (d), on the other hand, show that the molecules are exchanged bias by an AFM Mn film. All data measured by XMCD at 8 K.

The possibility of inducing exchange bias between a molecule and an antiferromagnet is also very interesting in this context. Antiferromagnets have zero total magnetization because the magnetic moments of neighbor atoms point in opposite directions, that is, they compensate each other. However, because this compensation is not perfect at interfaces, ferromagnetic layers can lock their magnetization to the spins in the antiferromagnets, which gives rise to the phenomenon of exchange bias. Exchange bias typically occurs in either ferromagnetic/antiferromagnetic (FM/AFM) bilayers or FM/AFM core−shell nanoparticles cooled in a magnetic field from below the Curie temperature of the FM through the Néel temperature (TN) of the AFM. Unidirectional exchange anisotropy sets in below TN, as the interfacial spins of the AFM align with the magnetization of the FM and henceforth remain pinned in the direction of the cooling field. The tell-tale signature of exchange bias is a shift of the hysteresis loop of the FM along the field axis by an amount HE, termed the exchange field, often accompanied by an enhancement of the coercivity HC. These effects offset the response of a FM to applied magnetic fields, currents, and temperature, leading to prominent applications of exchange bias in, for example, spin valve and magnetic tunnel junction devices.

Does exchange bias with antiferromagnets also work for molecules? There are rather reasons to doubt it. First, exchange bias is triggered locally by the presence of pinned uncompensated spins in the AFM. As the molecules constitute discrete magnetic elements, there is no mechanism guaranteeing that the sparse pinning centers of an AFM may bias a single molecule, unless this adsorbs on or creates a pinning site. Second, biasing is unlikely to extend from individual sites to a continuous molecular layer, since the magnetic moments of molecules adsorbed next to each other are usually uncoupled. Third, thermal fluctuations tend to randomize the orientation of the molecular magnetic moment down to temperatures T ≪ TN, which hinders the alignment of the pinned spins in the AFM during the field cooling process. Despite such unfavorable premises, we found that exchange bias can be induced in TbPc2 adsorbed on antiferromagnetic Mn films, resulting in enhanced coercivity and a shifted hysteresis loop of the molecular magnetization [2,3]. Only a fraction of the TbPc2 molecules is biased, however, which calls for further investigations of the microscopic mechanism leading to the pinning of molecular spins to antiferromagnets. Further, we found no evidence of exchange bias when TbPc2 is deposited on insulating (oxide) antiferromagnets rather than metallic ones, probably due to the weaker electronic hybridization between molecules and substrate in the former compared to the latter [2,3]. These results will help to address several outstanding problems in the field of molecular spintronics related to the incorporation of magnetic molecules in practical devices. The pinning of SMM, and paramagnetic molecules in general, to AFM represents a convenient way to stabilize and control their magnetic properties using substrates with no net magnetization.