Imaging the dynamics of 3D magnetic systems

Researchers at the University of Cambridge and Glasgow in the UK, and the ETH Zurich and the Paul Scherrer Instiute in Switzerland have developed a new technique to image magnetisation dynamics in 3D. Here we tell the story of the role of collaboration in the work.

Like Comment
Read the paper

In magnetism, there is growing interest in three-dimensional magnetic systems. An example are bulk micromagnets with technological applications (such as in hard drive read heads or sensors). In addition, patterned 3D systems are expected to support chiral effects, which can be exploited to create functional devices and applications. In both cases, the dynamic response of the magnetisation to both currents, and external magnetic fields, are key to understanding, and exploiting, these new functionalities. In our recently published paper “Time-resolved imaging of three-dimensional nanoscale magnetization dynamics” [1], we present magnetic laminography, a new method that provides the possibility to visualise magnetic configurations – and their dynamics – in three dimensions. 

This work combines multiple areas of expertise at the forefront of synchrotron techniques, including three-dimensional magnetic imaging and dynamical measurements, as well as the fabrication of 3D magnetic systems. As a result, this work is a perfect example of what can be achieved with strong collaborations and joint efforts, that form perhaps the most crucial, yet uncelebrated, part of the scientific method of today. Here, we take the opportunity to highlight the role of collaborative work in the story behind the paper.  

The story begins with two of the authors, Claire Donnelly and Simone Finizio, meeting on the way to an international conference: as is often the case, although both were based at the same institute (the Paul Scherrer Institute in Switzerland), this was the first time that we properly met one another, and took the time and space to discuss, and dream, about possible projects. Specifically, Simone was to present recent work on the X-ray imaging of magnetisation dynamics in strained magnetic films [2], whilst Claire would present her work on the X-ray imaging of three dimensional magnetic systems [3]. Over the course of the week's conference, we realised that our interests - that is, dynamic measurements, and 3D magnetic imaging, could be combined to describe one of biggest challenges facing the experimental investigation of three dimensional magnetism: being able to image the dynamics of a magnetic system in three dimensions. But, how?

After discussing the idea with our supervisors, Jörg Raabe and Laura Heyderman, the next few months were spent exploring possibilities for time-resolved magnetic tomography. Progress was initially slow: we weren’t awarded synchrotron beamtime, and our first attempts at combining the sample fabrication for dynamic measurements with the constraints for magnetic tomography – i.e. e.g. building circuits that traverse micrometre-wide pyramids on a Si chip, to be able to measure the sample from different perspectives – and factoring the manual reorientation of the sample part way through the experiment that is needed to probe all components of the magnetisation, were not promising. It seemed as if our ideas were to remain just that. 
Before all hope was lost, however, a breakthrough occurred: while performing some exploratory simulations, we discovered that an alternative 3D imaging geometry, pioneered by Mirko Holler, Manuel Guizar-Sicarios, and Michal Odstrčil at the cSAXS beamline at the Swiss Light Source [4], was ideally suited for magnetic imaging. Laminography, where the rotation axis of the sample is no longer at 90° to the X-rays, is compatible with flat samples, and provides access to all three components of the magnetisation (which is a vector field, requiring the measurement of three spatial components) with only one axis of rotation, therefore removing the need to reorient the sample (and interrupt the experiment) half-way through. As if by magic, this new geometry appeared to address all feasibility concerns simultaneously. 

Following a first confirmatory experiment of static magnetic laminography, plans were made to extend this to time-resolved measurements to investigate magnetization dynamics. The sample involved a 1.2 μm thick ferrimagnet, sputtered on a membrane, with a varying anisotropy introduced through the thickness during the deposition by controlling the rotation of the sample. Depositing such a thick film required not only dedicated optimisation, but the final deposition took over 8 hours, a challenge in itself, which luckily Ales Hrabec was willing to take on. On the other side of the membrane, a copper stripline was patterned lithographically by Sina Mayr, and the magnetic film was subsequently patterned in into a disc that was aligned with the stripline in order to be able to excite the magnetisation in the disc with the magnetic field produced by the neighbouring stripline.

Finally, the allocated week for the beamtime arrived. A diamond phase plate used to produce circular polarised X-rays was integrated into the setup by Valerio Scagnoli, allowing us to probe the magnetisation of the sample. The time-resolved setup, in which the microwave excitation signal and the X-ray pulses of the synchrotron SLS have to be synchronized, was integrated with the laminographic setup by Jörg Raabe and Simone Finizio, and the measurements could begin. Perhaps anticlimactically, thanks to months of preparation and a couple of intense days working to get everything up and running, the laminography measurements themselves went without a hitch. The full set of laminographic images used for reconstructing the dynamics consisted of more than 1000 2D magnetic projections, took more than four days of continuous measurements. This is in fact the main constraint of the technique, at the moment. With the added coherent flux provided by the next generation of synchrotrons, the measurement time is expected to decrease by orders of magnitude!

Following the beamtime, efforts turned to the reconstruction, and interpretation, of the seven dimensional data, three dimensions for the position, three for the direction and one for the time.  We first considered a single time step (see figure 1), where we could see a strong influence of the variable anisotropy on the magnetic state: in the lower half of the sample, we found a double vortex state, whilst in the upper half, the strong magnetic anisotropy led to the vortices being expelled out the sides of the disc, and a single domain state forming. To understand how the observed magnetic configuration could have formed, co-author Sebastian Gliga performed micromagnetic simulations. 

When it came to dynamics of the 3D magnetic configuration, things were even more interesting. While the simulations were not included in the paper, they provided insight into the dynamics and their representation: by performing Fourier analysis of the evolution of the magnetization, we could plot a map of the dynamics throughout the structure, which revealed the presence of so-called edge modes that occur along a spiral around the edge of the sample, as well as in the vicinity of the vortex walls – which we found to gyrate from side to side at speeds of over 100 m/s. It took a great deal of discussions between the co-authors to unravel exactly what exactly was going on – but it soon became clear that the dataset provided an ideal demonstration of our technique, not only allowing us to observe, but also gain insight into complex 3D magnetisation dynamics.

By bringing together scientists from different areas of expertise – synchrotron X-ray science, 3D nanomagnetism, magnetisation dynamics – and at different stages in their career – PhD, postdoc, and permanent staff – we developed a method that provides new capabilities that are vital for the young and growing field of 3D magnetic systems. If you’re interested in making use of the technique for your own samples, or finding out more, please get in touch!

[1] Claire Donnelly, Simone Finizi, Sebastian Gliga, Mirko Holler, Aleš Hrabec, Michal Odstrčil, Sina Mayr, Valerio Scagnoli, Laura J. Heyderman, Manuel Guizar-Sicairos and Jörg Raabe, Nature Nanotechnology (2020) 
[2] Simone Finizio, Sebastian Wintz, Eugenie Kirk, Anna K. Suszka, Sebastian Gliga, Phillip Wohlhüter, Katharina Zeissler, and Jörg Raabe, Phys. Rev. B 96, 054438 (2017)
[3] Claire Donnelly, Manuel Guizar-Sicairos, Valerio Scagnoli, Sebastian Gliga, Mirko Holler, Jörg Raabe and Laura J. Heyderman, Nature 547, 328 (2017)
[4] Mirko Holler, Michal Odstrcil, Manuel Guizar-Sicairos, Maxime Lebugle, Elisabeth Müller, Simone Finizio, Gemma Tinti, Christian David, Joshua Zusman, Walter Unglaub, Oliver Bunk, Jörg Raabe, Albrecht FJ Levi and Gabriel Aeppli, Nature Electronics 2, 464 (2019)

This Behind the Paper article tells the story behind our recent publication: Time-resolved imaging of three-dimensional nanoscale magnetization dynamics, Nature Nanotechnology (2020)

by Claire Donnelly (University of Cambridge, UK) and Simone Finizio (Paul Scherrer Institute, Switzerland)

Claire Donnelly

Early Career Researcher, University of Cambridge