Xray Imaging of Magnetic Nanostructures and their Dynamics

X-ray Imaging of Magnetic Nanostructures and their Dynamics Joachim Stöhr Stanford Synchrotron Radiation Laboratory X-Rays have come a long way…… 1 m 1895 1993 2003

Andreas Bauer 1, 2 Jan Lüning 2 Andreas Scholl 1 Howard A. Padmore 1 3 Yves Acremann 2 Aaron Lindenberg Andrew Doran 1 Sug-Bong Choe 1 Hendrik Ohldag 2 Squaw Valley, April 2003 Advanced Light Source 2 Stanford Synchrotron Radiation Laboratory 3 UC Berkeley 1

Fe metal – L edge Kortright and Kim, Phys. Rev. B 62, 12216 (2000) Soft X-Rays are best for magnetism!

Imaging by Coherent X-Ray Scattering Phase problem can be solved by “oversampling” speckle image Transmission X-ray Microscope Reconstruction from Speckle Intensities 5 m (different areas) S. Eisebitt, M. Lörgen, J. Lüning, J. Stöhr, W. Eberhardt, E. Fullerton (unpublished)

Magnetic Spectroscopy and Microscopy

Spectromicroscopy of Ferromagnets and Antiferromagnets AFM domain structure at surface of Ni. O substrate s [010] s 2 m Ni. O XMLD Co XMCD H. Ohldag, A. Scholl et al. , Phys. Rev. Lett. 86(13), 2878 (2001). FM domain structure in thin Co film on Ni. O substrate

Magnetic characterization of interfacial spins Co/Ni. O Co/Ir. Mn Co Ni. O Publications: Stöhr et al. , Phys. Rev. Lett. 83, 1862 (1999) Thomas et al. , Phys. Rev. Lett. 84, 3462 (2000) Scholl et al. , Science 287, 1014 (2000) Nolting et al. , Nature 405, 707 (2000) Regan et al. Phys. Rev. B 64, 214422 (2001) Ohldag et al. , Phys. Rev. Lett. 86 2878 (2001) Ohldag et al. , Phys. Rev. Lett. 87, 247201 (2001) Ohldag et al. , Phys. Rev. Lett. 91, 017203 (2003) loop of interfacial spins only 4% are pinned

Exchange Bias Model from X-Rays ideal AFM poly AFM

Present limitations of magnetic recording • Present method of magnetic switching is unfavorable: – present recording time ~1 ns – unfavorable torque and dependent on thermal activation

Fast Magnetization Dynamics is governed by Landau-Lifschitz-Gilbert equation: Angular momentum change Precession torque 1 Tesla field: 90 o rotation in 10 ps Gilbert damping torque Typically a << 1, 100 ps We want to understand a on atomic level a controls switching time, a~1 optimal

Time Resolved X-Ray Microscopy Laser pump – x-ray probe synchronization excitation laser pulse t observation x-ray pulse 328 ns < 1 ps < 100 ps

Production of Magnetic Field Pulses Photoconductive switch 100 m 2 m H ~ 200 Oe Conducting wire 50 => I = 200 m. A, 10 V bias Magnetic Cells Current 10 m

Sample and Magnetic Field Pulse 20 nm Co 90 Fe 10 films with in-plane anisotropy (1 m) x (1 -3 m) rectangles Current magnetic field M Magnetic Field Pulse ~ 150 Oe at Maximum < 50 ps rising time > 300 ps decaying time with some reflection

Observation of Vortex Motion H 1 m x 1 m 2 m x 1 m 1. 5 m x 1 m Vortices rotate oppositely - vortex cores point in opposite directions Vortex speed ~ 100 m/s

Conclusions The challenge of the future is to control the magnetization on the nanometer length scale and picosecond/femtosecond time scale Our current capabilities are: • image the magnetization with 50 nm spatial resolution, • image the response of the magnetization with 100 ps time- and 100 nm spatial resolution Outlook into the future: • 5 nm spatial resolution – PEEM 3, under construction • 100 fs time resolution: pump-probe excitations single snapshots of equilibrium dynamics Modern x-ray sources offer unique opportunities for studies of the ultrafast magnetic nanoworld

The End

Vortex Structure And Vortex Elevation view Motion torque H Plane view Landau-Lifshitz equation: (neglect damping) Motion antiparallel to field! The field acts like a screw driver. Depending on the orientation of the thread pitch, the screw (vortex) will move either forward or backward

Vortex Precession M Under a field pulse, the vortex moves from the center. Happlied After the field pulse, the vortex continues to move radially due to the magnetostatic energy. Induced magnetostatic field is always perpendicular to the vortex motion. Hmagnetostatic x dx H d. H Magnetostatic field is always perpendicular to the vortex deviation Vortex will precess forever if there is no damping.

Incoherent vs. Coherent X-Ray Scattering Small Angle Scattering Coherence length larger than domains, but smaller than illuminated area scattering vector q ( m-1) 40 information about domain statistics 20 0 -20 -40 -20 0 20 40 scattering vector q ( m-1) Speckle 40 scattering vector q ( m-1) Coherence length larger than illuminated area true information about domain structure 20 0 -20 -40 -20 0 scattering vector q 20 ( m-1) 40

Pulse Structure Possible solutions: - gated detector, pulse picker - pump at 500 MHz