Gelation: Grow with the flow.

381 news & views two ferromagnetic layers. In the present case, this is the ferroelectric BaTiO 3 sandwiched between magnetic Fe and La 0.67 Sr 0.33 MnO 3 layers. Indeed, theoretical calculations 6 have predicted the possibility of significant changes in the interfacial magnetization and spin polarization in a ferromagnet in response to the ferroelectric polarization state across the interface. Thus, the interplay between the non-volatile polarization states of the ferroelectric and its influence on the spin polarization can lead to unusual tunnelling characteristics. This is the focus of the paper by Garcia and co-workers. Using a combination of epitaxial synthesis, scanned probe imaging and high-resolution electron microscopy, they have demonstrated a large tunnelling magnetoresistance that is sensitive to the polarization state of the ferroelectric. The fundamental origin of this large effect arises from the asymmetric screening of the polarization at the two interfaces, albeit at a low temperature of 4.2 K. So, what are the implications of this approach? It is now becoming increasingly clear that ferromagnetism, and thus spin polarization, can be controlled and manipulated through a coupling of a magnetic layer to a ferroelectric or a multiferroic. Thus, more than one possible pathway to control magnetism with an electric field exists. Clearly, room-temperature operation is a critical step, although it is probable that that issue will be solved in the near future. However, an aspect that clearly emerges from this work, as well as from previous studies involving the manganites, is the observation that they are not the ideal candidates for room-temperature devices as the transfer of spin polarization at the interface invariably is corrupted. Thus, the discovery of alternative ferromagnets would be a fruitful future research direction to pursue. If using epitaxial oxide heterostructures is a possible pathway, the possibility of using the orbital degree of freedom in oxides as an active component of their interface interactions can also be a fertile area of research. Some initial results in this direction are already available. The robustness of these interactions, particularly with respect to electric-field cycling, long-term stability and speed of operation are all issues that need to be addressed for real-world devices. Finally, by far the most important issue will probably be this: what are the device structures and applications that will be enabled by these new observations? In summary, the paper by Garcia and co-workers has demonstrated, in a convincing way, the great potential of …