Two 'failed' materials can perform much better when united. Such a combination exhibits magnetization and electric polarization up to room temperature, providing a basis for new magnetoelectric devices. See Letter p.523
Materials in which the electric dipoles or magnetic moments associated with atoms, ions or molecules are ordered are of immense technological value. Multiferroic materials unite these two types of order in a single material and are therefore highly desirable. However, because simultaneous electric and magnetic order is difficult to achieve, multiferroics — especially those that function at or approaching room temperature — are extremely rare1. On page 523, Mundy et al.2 detail an effort to build such a material. Remarkably, they achieved this by combining two 'failed' multiferroics, forming a new compound with excellent multiferroic properties.
Electrons are subatomic particles that have a magnetic moment and an electric charge. In a crystal, if the electron clouds that surround atomic nuclei orient themselves in the same way, the crystal can develop macroscopic magnetization (ferromagnetism) or electric polarization (ferroelectricity). Ferromagnetism is essential for technologies such as power generators, sensors and computer hard disks. And ferroelectrics are often used in devices that generate or detect a small mechanical deformation, such as loudspeakers and pressure sensors3.
The idea of creating multiferroic materials that simultaneously exhibit magnetic and electric order was first suggested in the late 1950s, and they began to be constructed a few years later4. In these materials, the magnetic and ferroelectric properties are manifested together and, more importantly, the two types of order are often coupled to each other.
Why would such a coupling be useful? As an example, many computer memory elements are operated by electric currents. These currents limit the computer's processing speed, and produce so much heat that energy consumption and the need for cooling become serious problems. However, in a multiferroic memory element, a voltage pulse could be used to control the ferroelectric state and, through an internal magnetoelectric coupling, activate the ferromagnetic state that represents the memory bit. Voltage pulses can be transmitted more quickly than electric currents and consume less power5.
Unfortunately, building a multiferroic material is difficult because the conditions that favour magnetic and electric order tend to be mutually exclusive1. As a result, the use of known multiferroics is technologically infeasible for various reasons — for example, the magnetization and electric polarization are too small to be used in devices, occur far below room temperature or are too weakly coupled. Even the most promising room-temperature multiferroic, bismuth ferrite (BiFeO3), is intrinsically not ferromagnetic6.
Mundy and collaborators' work introduces a way to engineer materials that have coupled magnetization and electric polarization. Early attempts to produce such multiferroics focused on bulk synthesis — magnetic and electric moments were brought together in the same unit cell (the smallest periodically repeating structure in a crystal), usually resulting in weak manifestations of multiferroic order4. However, spiral-like arrangements of magnetic moments across many unit cells provided a new and general source of such order7. A later development was to apply small changes to multiferroics at inter-atomic distances (for example, using external pressure8), which could modify the materials' magnetization, polarization or ordering temperature. Finally, moving from bulk materials to thin films of multiferroics had advantageous effects resulting from the limited thickness of the material and the presence of surfaces or interfaces4.
Building on this earlier work, Mundy and colleagues create a multiferroic that has large magnetization and electric polarization at room temperature, and a strong coupling between the two up to at least 200 kelvin. They achieve this by combining two 'failed' multiferroics — LuFeO3 and LuFe2O4 — unit cell by unit cell, such that the deficiency of each material is compensated for by the desirable property of the other. LuFeO3 is multiferroic but lacks pronounced magnetization9, whereas LuFe2O4 has magnetization but no ferroelectric order10.
Using a technique called molecular-beam epitaxy, the authors build a film of repeating units that consist of a single layer of LuFe2O4 and nine layers of LuFeO3 (Fig. 1). The LuFeO3 has a corrugated structure, which acts as a template for the atomic arrangement of LuFe2O4, allowing the latter material to become ferroelectric. In turn, the multiferroic order of the entire structure is reinforced. The authors show that when an electric field is used to reverse the direction of the polarization, the magnetization direction is also reversed, which suggests that the multiferroic has a strong magnetoelectric coupling.
Such a coupling at the atomic level is reflected in the macroscopic properties of the authors' material. A sharp tip, to which a positive or negative voltage is applied, can be used to draw an electric polarization pattern in the material. The authors show that this pattern is complemented by an identical magnetization pattern, even though a magnetic field was not applied. The magnetization is determined entirely by the sign of the electric voltage, which is exactly the functionality that is required for magnetoelectric devices.
It remains to be seen whether Mundy and colleagues' atomic-template approach can be used to create multiferroics in general. However, their work seems to show that the box of tricks for improving such materials is not yet empty. Multiferroics are now migrating to a wide variety of research disciplines, such as electronics, photonics and even high-energy physics, in which they are studied for properties that are, at best, indirectly related to their multiferroic order4. Hence, although they originated as a specialist's topic, multiferroics are now a substantial part of materials research.