Y. Acremann, Sara Gamble, Mark Burkhardt ( SLAC/Stanford ) Exploring Ultrafast Excitations in Solids with Pulsed e-Beams Joachim Stöhr and Hans Siegmann Stanford Synchrotron Radiation Laboratory Collaborators: A. Vaterlaus (ETH Zürich) A. Kashuba (Landau Inst. Moscow) ; A. Dobin (Seagate) D. Weller, G. Ju, B.Lu (Seagate Technologies) G. Woltersdorf, B. Heinrich (S.F.U. Vancouver) samples theory magnetic imaging

The ultrafast technology gap want to reliably switch small “bits” The Technology Problem: Smaller and Faster

Faster than 100 ps…. Mechanisms of ultrafast transfer of energy and angular momentum Electric fields Optical pulse Electron pulse Magnetic fields Shockwave IR or THz pulse ElectronsPhonons Spin  ~ 1 ps  = ?  ~ 100 ps

The simplest case: precessional magnetic switching M Conventional (Oersted) switching

Creation of large, ultrafast magnetic fields Conventional method - too slow Ultrafast pulse – use electron accelerator C. H. Back et al., Science 285, 864 (1999)

Torques on in-plane magnetization by beam field Max. torque Min. torque Initial magnetization of sample Fast switching occurs when H M ┴

Precessional or ballistic switching: 1999 Patent issued December 21, 2000: R. Allenspach, Ch. Back and H. C. Siegmann

In-Plane Magnetization: Pattern development Magnetic field intensity is large Precisely known field size 540 o 360 o 180 o 720 o Rotation angles:

Origin of observed switching pattern In macrospin approximation, line positions depend on: angle  = B  in-plane anisotropy K u out-of-plane anisotropy K ┴ LLG damping parameter   from FMR data H increases 15 layers of Fe/GaAs(110) Beam center

Breakdown of the Macrospin Approximation H increases With increasing field, deposited energy far exceeds macrospin approximation this energy is due to increased dissipation or spin wave excitation

Magnetization fracture under ultrafast field pulse excitation Macro-spin approximation uniform precession Magnetization fracture moment de-phasing Breakdown of the macro-spin approximation Tudosa et al., Nature 428, 831 (2004)

Results with ultrafast magnetic field pulses Compare 5ps pulse with 160fs bunch – same charge 1.6x10 10 e - Tudosa et al. Nature 428, 831 (2004) 160 fs – 5 ps Magnetic pattern with 5 ps electron bunch ~ 20  m

Magnetic pattern of 10 nm Fe film with 160 fs bunch length Pattern size 470 µm by 980 µm New results with a 10 times shorter ( 167 fs ) pulse Pattern is severely asymmetric – should follow circles No switching for certain magnetic field orientations ! Violates conventional laws of angle-dependence of magnetic torque Puzzle is unresolved at present…..

5.5 ps, 1.6x10 10 e fs, 1.6x10 10 e - ablation Magnetic flash images with different bunch lengths characteristic demagnetization domains slow pulsefast pulse magnetic topographical

Surprising result: No beam damage at ultrafast time scales Tpographical image of beam impact area reveals no damage ! Magnetic pattern indicates no heating ! Maximum power density is 4 x W / cm 2 Sub-picosecond energy dissipation must exist (photons, electrons) Dissipation faster than electron-phonon relaxation time (ps) Results of importance for LCLS, as well.

Example: PbTiO 3 From Ferro-magnetic to Ferro-electric Materials FM materials respond to magnetic field (axial vector) - broken time reversal symmetry of electronic system FE materials respond to electric field (polar vector) - broken inversion symmetry of lattice Materials important as storage media – but how fast do they switch?

Ferro-electric domains can be observed with x-rays E E PEEM Reverses with change in polarization direction. Dichroism

Use E-field of SLAC beam to switch In lab (Auston switch): ~70 ps with the E-field amplitude of 25 MV/m SLAC Linac: femtosecond pulses up to several GV/m

Pump-Probe Dynamics with SLAC-pulses

Summary The breakdown of the macrospin approximation for fast field pulses limits the reliability of magnetic switching At ultrafast speeds (< 1 ps) new ill-understood phenomena exist one approaches timescales of fundamental interactions between electrons, lattice and spin Future experiments to explore the details using both H and E fields In the future, e-beam “pump”/ laser “probe” experiments are of interest, as well For more, see: and J. Stöhr and H. C. Siegmann Magnetism: From Fundamentals to Nanoscale Dynamics 800+ page textbook ( Springer, 2006 )