Magnetism to its lowest terms
An international research team reports the first observation of ferromagnetism in one-dimensional monatomic chains of metal atoms.
Progress in atomic engineering makes it now possible to produce one-dimensional (1D) nanostructures tailor-made on the ultimate atomic scale and to probe their physical properties. While an ideal infinite chain of atoms cannot sustain ferromagnetic order at nonzero temperature, an international team of physicists led by Prof. Klaus Kern, Director at the Max-Planck-Institute for Solid State Research in Stuttgart and Professor at the Swiss Federal Institute of Technology in Lausanne, including Dr. Pietro Gambardella at the Swiss Federal Institute of Technology in Lausanne, Prof. W. Eberhardt at the Forschungszentrum Jülich, and Dr. Carlo Carbone at the National Research Council of Italy shows that the finite length of the chains and magnetic anisotropy barriers stabilize ferromagnetism at finite temperatures in monatomic cobalt chains. The persistence of ferromagnetism in monatomic chains has important consequences for the design and properties of magnetic nanostructures, a very active field of physical research and development today. Their results are published this week in Nature (March 21, 2002).
Who hasn't been fascinated as a kid by the spell of the invisible and yet powerful attraction of two magnets brought close to each other? Men have wondered about magnetism for more than 3000 years now, while making the best of it in their daily life. We see manifestations of magnetism in many aspects of modern life; its effects are used to drive electric motors, transformers, telephones, and to store digital information on credit cards and computers. Yet its understanding, even in the simplest materials, is still incomplete.
We know that the magnetic properties of a given material can be attributed to its electrons. A single atom is magnetic when the sum of the spin and orbital magnetic moments of its electrons is nonzero. For a solid, however, two additional conditions are required: first, its constituent atoms must not loose all their magnetic moment as their outer electrons mix together; second, every single atomic moment must align parallel to its neighbors in order to produce a net macroscopic magnetization in the material. The first condition is rarely found and explains why only a few chemical elements and compounds exhibit ferromagnetic properties: iron, cobalt, nickel, some rare earth elements, and some of their alloys. The second condition, on the other hand, is made possible in the aforesaid metals by the exchange interaction, which is a purely quantum mechanical effect that couples neighboring spins favoring their parallel alignment. The effectiveness of the exchange interaction depends on the number of neighbor atoms and on the temperature, since thermal agitation tends to destroy the spin alignment. All ferromagnets loose their magnetic properties when heated to a temperature known as the Curie point, which varies from substance to substance. Magnetism in 2D systems such as thin films is much more sensitive to temperature effects compared to bulk systems due to the reduced number of neighbors contributing to the exchange interaction. Ultimately, the famous Ising model predicts that an infinite 1D linear chain of atoms cannot be magnetic since thermal fluctuations would be strong enough to destroy the spin alignment at any temperature.
Physicists have always sought to make things more comprehensible by looking at simple cases. Models describing the persistence of magnetism can often be solved in one dimension but not two, while others can be solved in two but not three. Progress with atomically-engineered materials allows today to test the predictions made for ideal low-dimensional systems in a real laboratory. Much insight can be gained by doing so, not only in retrospective, but also for the conception of new nanoscale magnetic structures which have useful properties distinct from those of bulk materials. 2D systems such as ultrathin epitaxial films and superlattices have already found widespread application in data storage and magnetic sensor devices. New frontiers are now open to investigation. In the race to miniaturization, atomic-scale growth and characterization methods are eventually leading us to dimensions comparable to the exchange interaction length scale. What happens then to the magnetization? Can we control the interplay between atomic structure and exchange interaction to obtain useful magnetic properties?
Fig.: Scanning tunnelling microscope image of monatomic cobalt chains decorating the steps of a regularly stepped Platinum surface. The average distance between neighboring cobalt chains is 20 Å.
Photo: Max Planck Institute for Solid State Research
The results are not only important for a fundamental understanding of low-dimensional magnetism but bear also important implications for magnetic data storage technology. Currently more than 105 atoms (spins) are needed for the construction of a stable magnetic bit in the hard disc of a personal computer. If the number of spins needed for establishing a bit can be scaled down, the storage density automatically goes up. How far can this scaling go? In the experiment it is shown that by decreasing the coordination of the magnetic atoms values of the magnetic anisotropy energy are obtained that are two orders of magnitude larger than those so far encountered in transition metal systems. The measured anisotropy energies for the monatomic cobalt chains indicate that not more than a few hundred cobalt atoms might be needed in tailor-made structures for constructing a stable magnetic bit at room temperature. A brilliant perspective for the race to terabit memories!
Max Planck Society