Transition metal oxides: multiferroics go high-TC.

The discovery by Tsuyoshi Kimura and colleagues of a multiferroic coupling in cupric oxide up to 230 K, as reported on page 291 of this issue1, is all the more remarkable as physicists had been wondering for a long time if it is possible at all to align electric dipoles by magnetic fields or change the direction of magnetization by applying voltage. To some, such a magnetoelectric coupling may seem to be an oddity, as static magnetic and electric dipoles do not interact. Others may find it to be the most natural thing in the world, because electrons responsible for physical properties of solids carry both charge and spin. But the fact is that until recently the number of known materials showing pronounced magnetoelectric effects was very small. However, the situation changed when in 2003, Kimura et al. found that magnetic ordering in TbMnO3 induces electric polarization2, which soon led to the discovery of a dozen similar materials — called multiferroics, as they are both magnetic and ferroelectric3. The direction of the electric polarization vector and dielectric susceptibility of these materials can be very efficiently controlled by applied magnetic fields. The discovery of this magnetically induced ferroelectricity helped to resolve an old problem of chemical incompatibility in the attempts to simultaneously realize ferroelectric and magnetic orders, which for many years hampered the progress in the search for multiferroics. There are still a few hurdles left on the way to practical applications of these materials: small values of induced electric polarization and low transition temperatures, which, with one exception4, are limited to 40 K. The discovery that cupric oxide (CuO, also known as tenorite) is a multiferroic material with a high antiferromagnetic transition temperature TN of 230 K therefore represents a significant advance. However, this finding doesn’t seem accidental in light of the fact that CuO is also the starting material for the synthesis of copper-oxidebased superconductors, which twenty years ago broke ‘the theoretical upper bound’ on superconducting transition temperatures. At temperatures below 230 K, CuO shows antiferromagnetic order, and, as for most of the recently discovered multiferroics, spins in the ferroelectric state of CuO form a periodic spiral superstructure between 213 and 230 K (Fig. 1). Here, electric polarization is induced in the plane of the spin spiral, but perpendicular to the spiral’s wave vector. If the temperature is dropped below 213 K, the spiral ordering is replaced by a collinear spin structure, and the electric polarization disappears. The origin of spiral ordering was first discussed fifty years ago and two alternative explanations were proposed. One, put forward by Villain5, Kaplan6 and Yoshimori7, is based on magnetic frustration, where geometric constraints and competing interactions make it impossible to select a single lowest-energy configuration of collinear spins. In this situation, spiral ordering emerges as a compromise. Another idea was proposed by Dzyaloshinskii, who noticed that the free energy of magnets without inversion symmetry can have a phenomenological term that is the source of inhomogeneous magnetic order8. Its microscopic origin was elucidated by Moriya, who showed that relativistic corrections led to a new interaction that favoured non-collinear spins9. In the course of time both these explanations were proved to be right: there are many frustrated magnets showing spiral spin ordering and there are materials, such as ferroelectric BiFeO3, where the breaking of inversion symmetry transforms a uniform antiferromagnetic state into a low-pitch spin spiral. In spiral multiferroic materials these two mechanisms work together. Frustration induces spiral ordering, which breaks inversion symmetry. This activates the Dzyaloshinskii–Moriya interaction, which forces positive and negative ions to shift away from each other and makes the magnetic spiral state ferroelectric. The magnetically induced ferroelectricity is one of many beautiful phenomena found in frustrated magnetic systems. But as ordering temperatures in frustrated systems are inherently low due to the competition between different spin states, finding a room-temperature multiferroic among these materials is not trivial. Kimura and colleagues have now shown that this is possible1, provided that spins in frustrated magnets interact strongly, as they do in copper oxides. High transition temperature is not the only condition necessary for multiferroic materials to be used in devices. The hexaferrite Ba0.5Sr1.5Zn2Fe12O22 orders magnetically well above room temperature4; in a magnetic field it becomes ferroelectric, but its electric polarization can only be reliably measured up to 130 K. The problem is that insulators such as hexaferrite and even CuO become conducting at high temperatures, making it difficult to align electric dipoles by an applied electric field. Nevertheless, the discovery of hightemperature multiferroic coupling in CuO shows that the potential of geometrically frustrated magnets to generate magnetoelectric phenomena is nowhere near exhausted, and that novel lattice geometries, magnetic The discovery of magnetically induced electric polarization in cupric oxide at 230 K has uncovered a new class of multiferroics with significantly higher ordering temperatures. TransiTion MeTal oxides