Magnetic thin films: Physics from basic research to spintronics Christian Binek 201H 11/18/2005 Why thin films Size matters Length (and time) scales determine the physics of a system Quantum mechanics tells us: Confinement of electrons by lowering dimensions affects the electronic states Electronic states 3D bulk 2D film 1D wire 0D quantum dots artificial atoms all macroscopic properties Physics 201H 11/18/2005 When can films considered to be thin or thin with respect to what d

dcharacteristic length Thin in comparison with the characteristic length scale Examples: -Superconducting thin film thickness correlation length -optical thin film like dielectric mirrors Length scale /4 500nm/4 Physics -Magnetic thin films approach the ultimate extreme 201H 11/18/2005 thickness quantum mechanical exchange interaction length a few atomic layers Exchange J(d) ferromagnet spacer d nonmagnetic ferromagnet Spacer thickness d in # of atomic layers

J(d=8)>0 d=8 monolayer Ferromagnetic coupling J(d=10)<0 d=10 monolayer Antiferromagnetic coupling How to grow magnetic heterostructures > 250 000 ? Molecular Beam Epitaxy Thin film growth @ low deposition rate Ultra high vacuum condition 10 10 mbar ( 10 8 Pa ) Important growth modes in heteroepitaxy specific free energy

B A Layer-by layer (Frank van der Merwe) substrate Monolayer followed by 3D islands (Stranski Krastanov) deposited material interface B A 3D islands (Volmer weber) Reflection High-Energy Electron Diffraction RHEED 3o Electron gun up to 50 keV sample RHEED Eye screen camera What are the magnetic heterolayers good for ? Basic components of modern spintronic devices Conventional electronics has ignored the spin of the electron Advantages using spin degree of freedom: magnetic field sensors

M-RAM Spin-transistor semiconductor Quantuminformation Impact of GMR based field sensors on magnetic data storage Areal density [Mb/in2] Evolution of magnetic data storage on hard disc drives 10 5 10 4 10 3 10 2 10 1

10 0 10 -1 10 -2 10 -3 Superparamagnetic effect GMR Magnetoresistive heads inductive read head 1960 1970 1980 Year 1990

2000 2010 rotating sensor layer FM1 fixed layer FM2 How to pin FM2 while the sensor layer FM1 rotates? Exchange Bias! Pinning of the ferromagnet by an antiferromagnet from T>TN to T

JS AFSFM 0MFM t FM 5 H -20 20 AF 40 0H [mT] -5 -10 Hfc TN T Meiklejohn Bean: uniform magnetization reversal of a pinned FM FM interface magnetization: SFM MFM coupling constant: J AF interface magnetization: SAF

tFM KFM, H MFM :saturation magnetization of FM layer Exchange bias field: 0 H E J S AF S FM M FM t FM 2 2 F t cos K t s i n F --(0 HM H M t J S S

) co s K t sin -J S S cos FM FM FM FM 0 FM FM AF FM FA MF FM FM JS AF S FM H 0 M FM t FM M FM t FM

AF/FM-interface coupling Stoner-Wohlfarth Electric control of the Exchange Bias Investigated multilayer system: Cr2O3(0001)/Pt0.67nm/(Co0.35nm/Pt1.2nm)3/Pt3.1nm tPt=1.20nm Co Co U Co Pt tCo=0.35nm Pt M Pt Cr2O3: Magnetoeletric Magnetoeletric effect AF, TNof =308K Cr2O3 * SQUID-magnetometry @ T=290K

M IIE electric field E=U/d Cr2O3 (0001) Magnetization M=m/V U Idea: E M contributes to SAF 0 H E FM thin film with perpendicular magnetic anisotropy JS AFS FM M FM t FM Cr2O3 (0001) * Change of the exchange bias field as a function of the electric field at T = 150K (0HE) [mT] 0.04

0.00 U=Ed Co Pt Cr2O3 (0001) -0.04 -300 0 300 E [kV/m] 2 Magnetoelectric Switching of Exchange Bias*: Control via field-cooling * P. Borisov, A. Hochstrat, Xi Chen, W. Kleemann and Ch. Binek, PRL 94 117203 (2005) Magneto-optical Kerr measurements @ T = 298 K after cooling from T>TN in 0Hfr = 0.6 T Magnetic Field Cooling (MFC) 1.0 (+,-) M / MS 0.5

EfrHfr<0 cooling from M E F C (+,-) T>TN a l i o in 0Hfr = +0.6 T gn ec el ol e t d i and n t r g i o Efr=-500 kV/m (+,+) EfrHfr>0 0.0 c -0.5 -1.0 -0.2 0.0 [T] 0H [mT]

0.2 cooling from M E F C (+,+) T>TN a l i o in 0Hfr =+0.6 T g ec e ol n l t e d i and n r t g Efr=+500 kV/m o i The sign of the Exchange bias follows the sign of EfrHfr c Spintronic applications* * Ch. Binek and B. Doudin, J. Phys.: Condens. Matter 17 (2005) L39L44 V V FM 2

FM 2 ME ME FM 1 FM 1 R H U U V V FM2 FM2 NM NM FM1 FM1

ME ME R -He-Hi He-Hi H Exclusive Or x|y | 0|0| 0|1| 1|0| 1|1| xORy 0 1 1 0 Example: 0 +V -H X:= Voltage

Input Y:= magn. field +V 0 -V 1 +H 1 -H 0 R high 0 R low 1 Output R 0

H Basic research with magnetic heterostructures generalized Meiklejohn Bean approach J finite anisotropy KAF0 3 S3AF 3 SFM J S AF SFM J 0 He MFM tFM 8 K 2AF MFM tFM t 2AF :coupling constant SAF/FM :AF/FM interface magnetization tAF/FM :AF/FM layer thickness MFM :saturation magnetization of FM layer Experimental check of advanced models understanding the basic microscopic mechanism of exchange bias Exchange bias is a non-equilibrium phenomenon new approach to relaxation phenomena in non-equilibrium thermodynamics

The training effect: a novel approach to study relaxation physics Training effect: reduction of the EB shift upon subsequent magnetization reversal of the FM layer - origin of training effect - simple expression for 0HEB vs. n Relaxation towards equilibrium F S S Landau-Khalatnikov :phenomenological damping constant Training not continuous process in time, but triggered by FM loop discretization of the LK- equation Discretization:

S AF S AF (n 1) S AF (n) LK- differential equation difference equation Comparison with experimental results on NiO-Fe 1st& 9th hysteresis of NiO(001)/Fe (001) compensated 12nm Fe NiO experimental data recursive sequence 2 e EB f HEB (n) 0 HEB (n) H n

f f , 0 0 e (0HEB ) 3 HEB (n 1) 2 min. 0.015 (mT) -2 and 0HeEB 3.66 mT Magnetic Nanoparticles Collaborations self-assembled Co clusters

I thermally decompose You want to know metal whatcarbonyls I am doing? in the presence of appropriate surfactants 25nm Transmission electron microscopic image ~5nm Fundamental questions Which magnetic interactions dominate the system What kind of magnetic order can we observe For large particle distances the dipolar interaction will dominate Here is a real fundamental question: Do dipolar systems still obey extensive thermodynamics What does this mean: Magnetic moment ,T,H

= 2 Magnetic moment Simulations suggest: Yes: for a 2 dimensional array of dipolar interacting particles but No: for a 3 dimensional array of dipolar interacting particles Modifications of conventional thermodynamics required ,T,H Summary MBE is a technology at the forefront of modern material science magnetic heterolayers are basic ingredients for spintronic applications magnetism of thin films and nanoparticles provides experimental access to fundamental questions in statistical physics 25nm Mechanical analogy V(X) Fx dV dx equilibrium F() dV

0 dx xeq d x -eq dV d 1 Dx 2 dx dx 2 Damped harmonic oscillator: mx equilibrium dF m x Dx 0 eq

0 2 D 1 2 Solution for: : 0 m 2 x(t) e t 2 A e 2 1 2 2 0 t

x0 1 A x0 x0 2 2 2 1 2 0 2 x0 1 B x0 2 2 1 2

2 0 2 2 B e 1 2 2 0 t x(0) 0 x(0) x 0 with

2 1 2 0 2 A 0 B x0 x(t) x 0 e t 2 x0 e 2 2 e 1 2 2 0 t 02 t 1

1 2 2 ... 0 0 2 2 x(t) x 0e dV x dx 2 0 t also derived from integration of: 0 m x x

m / dV Dx dx where dV dx Temporal evolution of X with increasing damping: x t dx D dt x 0 x0 1,0 0,8 X(t) 0,6 x(t) x 0 e

0,4 0,2 0,0 0 5 10 t 15 2 0 t P 10 11 mbar (1nPa ) 384,400 km Near earth outer space: P 10 6 mbar (100 Pa )