Since graphene (a sheet of carbon just one atom thick) was discovered in 2004, a whole new family of 2D materials has appeared. Such crystals have a broad range of electronic properties that rarely exist in the bulk material from which they are obtained. Recent additions include semiconductors with spin-valley coupling, Ising superconductors that can be tuned into a quantum metal and topological semimetals. However, until now, no intrinsically magnetic 2D crystal had been seen.

Ferromagnetic materials are those in which magnetic moments (spins) are aligned, something that produces uniform, permanent magnetization. A team led by Xiadong Xu of the University of Washington in Seattle and Pablo Jarillo-Herrero from the Massachusetts Institute of Technology, has now used magneto-optical Kerr effect (MOKE) measurements to determine the extent of ferromagnetic order in chromium triiodide (CrI3). This technique uses the rotation of linearly polarized light to produce a spatial map of the direction of spins in a material. By making measurements on the sample as a function of applied magnetic field perpendicular to the sample plane, the researchers obtained a hysteresis curve that shows all the hallmarks of ferromagnetic ordering.

Layer-dependent ferromagnetic ordering

The material displays ferromagnetic order below 45 kelvin (known as the Curie temperature), a value that is only slightly lower than that of the bulk crystal (61 kelvin). “The magnetization forms spontaneously without applying an external magnetic field,” explains Xu. “This indicates that the magnetism in monolayer CrI3 is very robust. Spins lie perpendicular to the material’s crystal plane with large magneto-anisotropy. It can thus be described by the 2D Ising model of magnetism.”

But things become even more surprising, explain the researchers, since the nature of the ferromagnetic ordering depends very much on the number of layers in the system. In a bilayer, for example, the remnant magnetization in a single layer is suppressed (which implies that the two layers have oppositely oriented – antiferromagnetic – spins). In contrast, in a trilayer this property is lost and the net magnetization comes back.

“These layer-dependent magnetic phases are remarkable and remain to be understood,” says Xu. “Recent theoretical calculations have predicted that the interlayer coupling in CrI3 should be ferromagnetic. However, our results show that the coupling in a CrI3 bilayer is antiferromagnetic. What is even more puzzling (and striking) is that the interlayer coupling becomes ferromagnetic again in trilayer and in bulk crystals. Lots of theorists are now working on understanding this, and the possible consequences for novel devices.”

Good for exploring the fundamental physics of magnetism

“The 2D crystal offers us several truly unique opportunities for exploring the fundamental physics of magnetism,” he tells “For one, its atomically thin nature, together with the fact that CrI3 is an insulator/semiconductor means that we may be able to control its magnetic properties and phases by applying electric field and strain. Such magneto-electrical coupling effects, in particular, are very difficult to obtain in bulk crystals.

“Second, the van der Waals nature of the material (that is, the weak interlayer coupling and the thickness-dependent optical, electronic and mechanical properties) means that we may be able to design heterojunction/heterostructures between these 2D magnets and other van der Waals materials (such as a different 2D magnet, a 2D superconductor or a strong spin-orbit coupled 2D material) without worrying about lattice mismatch.”

The researchers, reporting their work in Nature doi:10.1038/nature22391, say that they will now be looking at how layer stacking order in CrI3 affects the magnetic phases in the material. They will also be looking at controlling its magnetic properties in situ using electrical, optical and strain effects. “We will also see if magnetism and superconductivity are possible in CrI3 at the same time (normally, they don’t like each other)”, says Herrero. “In the long-term, we would like to look at Curie-temperature engineering, the development of topological magnets and superconductors, and the novel interfacial physics between this material and other 2D van der Waals crystals.”