Fundamental spatial limits of all-optical magnetization switching

Modern magnetic hard drives can store more than one terabit of data per square inch, which means that the smallest unit of information can be encoded on an area smaller than 25 nanometers by 25 nanometers. Therefore, to realize the full potential of laser-based all-optical switching (AOS), particularly in terms of faster write/erase cycles and improved power efficiency, we need to understand whether a nanoscale magnetic bit can still be all-optically reversed.

For AOS to occur, the magnetic material has to be heated up to high temperatures in order to reduce its magnetization close to zero. Only then can its magnetization reverse. The twist in AOS is that it is sufficient to heat the electrons of the material while leaving the lattice of atoms cold. This is exactly what an optical laser pulse does: it primarily interacts with the electrons, allowing much higher electron temperatures to be reached with very low power levels. However, since hot electrons cool down very rapidly by scattering with the cold atoms, the magnetization must be reduced within the characteristic time scale of such a cooling process, i.e., AOS relies on a careful balance between the evolution of the (electron) temperature and the loss of magnetization. It is easy to see that this balance is altered when the excitation is confined to the nanoscale: electrons cannot only lose energy by interacting with cold atoms but also by diffusing out of the nanometer-small hot regions. At the nanoscale, all these processes occur on comparable and ultrafast time scales: for instance, the electrons may cool down too quickly, the magnetization is not sufficiently decreased, and AOS breaks down.

Figure 1: Dots are the measured diffracted intensity off the transient magnetic grating with a period of 17 nanometers. The diffracted intensity decays as the grating washes out because of lateral ultrafast electron diffusion. The model (dashed line) describes our data very well. Please not the broken linear/logarithmic horizontal scale.

An international team of researchers from the Max Born Institute in Berlin, Germany, the Instituto de Ciencia de Materiales in Madrid, Spain, and the free-electron laser facility FERMI in Trieste, Italy, has successfully addressed for the first time the question: how small can AOS work? The answer was provided by combining soft X-ray transient grating spectroscopy with atomistic spin dynamics calculations. Using the unique capabilities of the EIS-TIMER instrument of FERMI, they produced an extremely short-lived pattern of dark and bright stripes of laser light by interference at the sample surface of the prototypical magnetic material GdFe. The novelty of the experiment lies in the use of laser light in the soft X-ray spectral range, allowing the distance between dark and bright areas to be reduced to just 8.7 nanometers. This leads a lateral modulation the (electron) temperature and a corresponding localized loss of magnetization. Subsequently, the ultrafast dynamics of the magnetization grating was probed by diffracting a third soft X-ray pulse with a wavelength of 8.3 nm; the experimental signal is displayed in Figure 1. At this particular wavelength, an electronic resonance at the gadolinium atoms allows the soft X-ray pulse to “feel” the presence of magnetization, enabling the detection of the change in the magnetization with femtosecond temporal and nanometer spatial resolution. Combining the experimental results with state-of-the-art calculations, the researchers could determine the nanoscale and ultrafast energy transport process, and proposing a “phase diagram” for AOS, as a function of the excitation energy density and of the spatial periodicity of the magnetization pattern (see Figure 2).

Figure 2: The phase diagram for AOS as a function of absorbed energy density and the period of the excited transient magnetic grating, suggesting that the minimum size for AOS in GdFe alloys, induced by a nanoscale periodic excitation, is around 25 nm.

It turns out that the minimum size for AOS in GdFe alloys, induced by a nanoscale periodic excitation, is around 25 nm. This limit is due to ultrafast lateral electron diffusion, which rapidly cools the illuminated regions on these tiny length scales, ultimately preventing AOS. The faster cooling due to electron diffusion can be compensated to some extent by increasing the excitation power, but this approach is inherently limited by the structural damage caused by the intense laser beam. The researchers expect that the 25 nm boundary is rather universal for all metallic magnetic materials.

Felix Steinbach¹, Unai Atxitia², Kelvin Yao¹, Martin Borchert¹, Dieter Engel¹, Filippo Bencivenga³, Laura Foglia3, Riccardo Mincigrucci³, Emanuele Pedersoli³, Dario De Angelis³, Matteo Pancaldi³, Danny Fainozzi³, Jacopo Stefano Pelli Cresi³, Ettore Paltanin³, Flavio Capotondi³, Claudio Masciovecchio³, Stefan Eisebitt^1,4, and Clemens von Korff Schmising¹

¹ Max-Born-Institute for Nonlinear Optics and Short Pulse Spectroscopy, Berlin, Germany.
² Instituto de Ciencia de Materiales de Madrid, CSIC, Madrid, Spain.
³ Elettra - Sincrotrone Trieste S.C.p.A., Trieste, Italy.
⁴ Institut für Optik und Atomare Physik, Technische Universität Berlin, Berlin, Germany.

Contact persons email:
Local contact person emails: