Time Resolve Studies of Spin Transfer Interactions

1. Time Resolve Studies of Spin Transfer Interactions Huanlong Liu Daniel Bedau Advisor: Andrew Kent Department of Physics New York University

2. Outline

3. Motivation How will the spinof itinerant electrons interactwith the background magnetization? B

4. Magnetization Spin influence • Electrons scattering with lattice not only exchange their energy, but also their angular momentum. • Background magnetization influences the spin of electrons. Spin diffusion length Spin polarized Motivation

5. GMR effects • We can measure the influence of the background magnetization to the spins of electrons by adding another ferromagnetic layer. Giant Magneto-Resistance (GMR) Discovered in 1988 Nobel Prize in 2007 Motivation

6. Spin Torque effects • Spins of the electrons can also influence the background magnetization. Spin torque effect 0 Spin filtering Conservation of angular momentum Spin torque effect Motivation

7. Introduction How to describe the dynamics of magnetization ? Ferromagnet Temperature Spin current ? Magnetic field

8. Magnetic Energy • Exchange energy • Dipole energy 1 2 Coulomb interaction + Pauli exclusion principle for ferromagnetic materials 2 1 Introduction

9. Magnetic Energy • Zeeman energy • Uniaxial anisotropy energy • Total energy for one spin is the sum of all the four contributions: Spin – orbit coupling Introduction

10. Domains Exchange energy dominates • For a ferromagnetic system with many spins: • Exchange energy uniformly distributed spins • Dipole energy multi-domain structure Simplify Introduction

11. Single Domain – Macro Spin Model • is a constant everywhere. • Sum of dipole energy demagnetization energy • Zeeman energy and uniaxial anisotropy energy take a similar form after summation. is the demagnetization factor depending on the shape of the FM material Introduction

12. Spin Torque • Spin torque – the amount of transverse angular momentum transferred in unit time. e mp electron m magnetization Introduction

13. LLG Equation Time scale nanosecond • Landau-Lifshitz-Gilbert (LLG) equation + Spin Torque Normalize with the magnitude of the magnetization Introduction

14. Thermal Effects • Thermal effects Langevin random field • The LLG equation will be: Introduction

15. Experimental Techniques

16. Sample Structure We need at least two FM layers to detect magnetization reversal by GMR effect Now add another polarizing layer to maximize spin torque effect Pt [Co/Ni]x4 Cu [Co/Ni]x2 [Co/Pt]x4 Pt Experimental Techniques

17. Probe Station I V Arbitrary Waveform Generator AWG7120 Digital Phosphor Oscilloscope DPO72004 Signal Generator MG3692B Projected field electromagnet GMW5201 S N Bipolar Operational Power Supply BOP20_20 Bias Tee Bias Tee Lock-in Amplifier SR830 Source Meter Keithley2400 Experimental Techniques

18. Experiment Types • DC measurements: • Hysteresis measurements • Current sweep measurements • High speed measurements: • Pulse measurements • Frequency sweep measurements Experimental Techniques

19. DC Measurements Initial Results and Analysis

20. Pulse Measurements Flow Pulse (1 mA, 1 ns) 1 2 3 ? B = 0.2 T B = 0 T Apply measurement field and current Saturate Check if switched Apply pulse If switched go here If NOT switched go here 4 Pulse 100 times and we have :switching probability = # of switched / 100 for AP to P at 0 T, 1 mA, 1 ns

21. Short Time Pulse Measurements Initial Results and Analysis

22. Switching Boundary – Short Time LLG equation + macro spin model: A is the dynamic parameter Initial Results and Analysis

23. Long Time Pulse Measurements The finite life time of being inside the potential well obeys a Boltzmann distribution. Switching boundary: Initial Results and Analysis

24. Long Pulse Scans Initial Results and Analysis

25. Results and Fitting

26. Determine Different Regimes Dynamic Regime: I ~ 1 / duration Thermal Regime: I ~ log(duration) Theoretical values start to deviate fromexperimental data Dynamic Intermediate Thermal Initial Results and Analysis

27. Summary • What do we know: • There are three distinguishable regimes of the switching boundary due to whether the thermal effects influence the switching process or not. • In the short time (dynamic) regime, the switching boundary goes as , which satisfies the LLG equation without thermal effects. • In the long time (thermal) regime, the switching current goes as , which indicates that the switching process is just like a particle escaping a potential barrier from thermal fluctuations.

28. Summary • What do we want to know: • How do the dynamic parameter A and the critical current Ic0depend on the material parameters? • How can we obtain the correct energy barrier from the material parameters? • A better model to describe the switching dynamics since the macro spin model gives unphysical values.

29. Further Plans • Analyze data • Find the probability distribution for short time switching • Energy barrier for long time – is there any domain wall motion? • Do the same measurements for different sample sizes • how do the dynamics change with the volume of the junction? • Low Temperature measurements • how much does temperature influence the short time switching process • New Structure with non-collinear magnetization configurations. • deterministic switching, shorter switching time and lower switching current.