an overview of spintronics in 2d...

© 2014 ICQM

An Overview of Spintronics in 2D

Materials

Wei Han (韩伟)

1

2

Outline

I. Introduction to spintronics (Lecture I)

II. Spin injection and detection in 2D (Lecture I)

III. Putting magnetic moment in 2D (Lecture II)

IV. Spin Hall effect and spin orbit torque in 2D (Lecture II)

V. Acknowledgement

3

Summary of Lecture I

Electron

Spin

Magnetic structure

1988 年GMRHistory

～1997 年Hard drive

1991-2004TMR2001- present

MRAMLow R

High R

4

Summary of Lecture I

Graphene is a very good candidate material for spin channels

Large spin signal (with tunnel barrier)

Long spin lifetime (6.2 ns in BLG)

Long spin diffusion length (> 10 micro meters at RT)

Easy to manipulate (Gate)

Electrical detection of spin --momentum locking in TI

Graphene “wins” the match for tunnel barrier

5

Outline

I. Introduction to spintronics (Lecture I)

II. Spin injection and detection in 2D (Lecture I)

III. Putting magnetic moment in 2D (Lecture II)

IV. Spin Hall effect and spin orbit torque in 2D (Lecture II)

V. Acknowledgement

6

Outline

I. Introduction to spintronics (Lecture I)

II. Spin injection and detection in 2D (Lecture I)

III. Putting magnetic moment in 2D (Lecture II)

Graphene

Topological insulator

7

Why make graphene magnetic

8

How to make graphene magnetic

Ferromagnetic oxides/grapheneAdatom and moleculedoping of graphene

Doping effect

Examples:Mn doped GaAsZnO

Proximity effect

Examples:Co-Pt (Magnetism in Pt)

9

Making graphene magnetic

Yazyev and Helm, PRB (2007)

H vacancy

10

Making graphene magnetic

Ferromagnetic ??

Cervenka, et al, Nature Physics (2009)

11

Making graphene magnetic

Paramagnetic ??

F doped

Nair, et al, Nature Physics (2012)

Defects

12

Making graphene magnetic

Question

Ferromagnetic ??

Paramagnetic ??

vacancy

13

Nonlocal spin transport geometry

IINJ VNL

spininjector

spindetector

chargecurrent

pure spincurrent

• Localized measurement

• Direct coupling of spin to magnetic moment

With magneticmoment Magnetic moment could

scatter pure spin currentthrough exchange interaction:

pure spincurrent

Hex = Aex Se SM

graphene

Making graphene magnetic

14

Making graphene magnetic

Si

SiO2

(backgate)

Graphene spinvalve device

I V+ -

Atomic hydrogen source

Hydrogen

• All measurements done in ultrahigh vacuum (UHV)

• Compare immediately before and after hydrogen doping

15

Aha! A dip in the nonlocal spin signal.

This may be the magnetic moment!

Making graphene magnetic

Orange arrows = amount of spinat the spin detector

pure spincurrent

At zero field At high field

Due to exchange coupling, pure spin current is scattered by magnetic moment

Hex = Aex Se SM

Fewer spins at detector

Scattering by exchange coupling is suppressed

Se and SM

More spins at detector

decouple!

16

Making graphene magnetic

17

Making graphene magnetic

H vacancy

Paramagnetic !!

McCreary, et. al, PRL (2012)

Doping effect

18

Making graphene magnetic

Proximity effect with EuO

T. Santos, et. al., PRL 101, 147201 (2008)

TC = 69 K

0.1 nm rms 0.2 nm rms 0.5 nm rms

19

Making graphene magnetic

A. Swartz, et. al, ACS Nano (2012)

Proximity effect with EuO

Graphene on SiO2 5 nm EuO on Graphene on SiO2

• No success of observing magnetic graphene/EuO in our group

20

Making graphene magnetic

Wang, et al, PRL (2015)

The first observation of Proximity effect in graphene with YIG

21

Making graphene magnetic

What if we try harder on EuO/Graphene?

5 nm EuO on Graphene

However, there is NO if.

??

Note: Sometimes, you are so close to the peak. Just try harder and move one more step!!

22

Making graphene magnetic

Note: Sometimes, you are so close to the peak. Just try harder and move one more step!

23

Making graphene magnetic

Moving forward

Wei Yuan, Yuelei Zhao, Chi Tang, Tang Su, Qi Song, Jing Shi*, and Wei Han*, Appl. Phys. Lett. 107, 022404 (2015).

24

Why make TI magnetic

Quantum Anomalous Hall effect

25

How to make TI magnetic

Doping effect by Cr/V

Chang, et al, Science (2013)

More information, please see the results of Q. Xue group (Tsinghua University), K. Wang group (UCLA), J. Moodera (MIT), etc

Observation of the QAH

26

How to make TI magnetic

Proximity effect

Wei, et al, PRL (2013)

27

Outline

I. Introduction to spintronics (Lecture I)

II. Spin injection and detection in 2D (Lecture I)

III. Putting magnetic moment in 2D (Lecture II)

IV. Spin Hall effect and spin orbit torque in 2D (Lecture II)

V. Acknowledgement

28

Introduction to spin-orbit torque

Spin transfer torque

Brataas, et al. Nature Mater. 11, 372-381 (2012)

tST =ℏ

2�� × (�� × ��)

29

Experimental observation of the spin transfer torque

Ralph, D. C. & Stiles, JMMM. 320, 1190-1216 (2008).

Charge Current + Spin Current

Introduction to spin-orbit torque

30

Ralph, D. C. & Stiles, JMMM. 320, 1190-1216 (2008).

Charge Current

Introduction to spin-orbit torque

Disadvantage

Heat!!!

31

D'yakonov, M. I. & Perel', J. Exp. Theor. Phys. Lett. 13, 467-469, (1971).Hirsch, J. E. Phys. Rev. Lett. 83, 1834-1837, (1999).Zhang, S. Phys. Rev. Lett. 85, 393-396, (2000).

How about pure spin current?Spin orbit coupling (spin Hall effect)

Introduction to spin-orbit torque

32Kato et al., Science 306 (5703) 2004

The first observation of spin Hall effect

Introduction to spin-orbit torque

33

Materials with large spin orbit couplings – high Z materials

Low

Large

Introduction to spin-orbit torque

34

Materials with large spin orbit couplings – high Z materials

Spin Hall current torque to FM

tSTM

JC

tST =ℏ

2�� × (�� × ��)

NM

FM

Introduction to spin-orbit torque

35Liu et al., PRL 109, 096602 (2012)

Spin Hall current torque to FM

Introduction to spin-orbit torque

36

The main Challenge

How to get larger spin orbit torque?

FM

PtJC

Increase the Efficiencyof the SOT

Search for larger SOT in quantum materials

37

FM

PtJC

Question: 100%?

Interface transparency of the spin current

Introduction to spin-orbit torque

38

Efficient spin orbit torque in Pt/Co and Pt/Py

Py

PtJC

Co

PtJC

eg: Different spin density states

Book: Magnetic Multilayers and Giant Magnetoresistance(editor: Uwe Hartmann & R. Coehoorn )

EF (eV)

Our approach

39

x

yz

V

HexttH

tSTM

JC

Au

Pt/FM

9GHz 10GHz 11GHz 12GHz 13GHz

10000 30002000Hext (Oe)

-10

0

10

-20

20

30

Vm

ix(µ

V)

Liu et al., PRL 106, 036601 (2011)

Photolithography Ion Beam etching/deposition

ST-FMR

Spin Torque FerroMagnetic Resonance

400 1000 2000 3000

-20

-10

0

10

20

30

Hext

(Oe)

Data fit S A

Vm

ix (V

)

tSTtH

���

��= −��� × ����

+��� ���

��+�

ℏ

������ × ��,��(�� × �� × ��)

− ��� × ���

spin torque tST

torque from RF field tH

damping

external effective field

Symmetric Component: Spin Hall torque

Antisymmetric Component: Torque from RF field

Landau-Lifshitz-Gilbert equation

Spin torque ferromagnetic resonance

41

0 1000 2000 3000

-20

-10

0

10

20

30

Vm

ix (V

)

Hext

(Oe)

tSTM

JC

Spin torque

��� =�

�

�������

ℏ[1 + (4�����/����)]

�/�

0 1000 2000 3000

-20

-10

0

10

20

30

Hext

(Oe)

Vm

ix (V

)

tH

M

JC

Symmetric Component Antisymmetric Component

Field torque

Efficient spin orbit torque in Pt/Co and Pt/Py

42

109 1211

Freq (GHz)

0.0013

0.04

0.08

0.12

109 1211

Freq (GHz)

0.0013

0.04

0.08

0.12

ƟS

H

ƟS

H

Py

PtJC

Co

PtJC d

t

d

t

6 nm Pt/6 nm Py

6 nm Pt/6 nm Co

Efficient spin orbit torque in Pt/Co and Pt/Py

43

Interface transparency

��� � = −

�

2����� − ���,����

FM

PtJC

In Pt layer:

�� � = �⃗��� �⁄ +��� �⁄

��� 0 = ��

� �� = ��(�� × �� × �� ) ∗ e h⁄

In FM layer:

d

t

z = 0x y

z

44

Interface transparency

The continuity of the spin current density at the interface(z = 0, and z = -d)

�� � = −��(2� � �⁄ )���,�� tanh��2�

sinh2� + ���

2�

sinh���2�

− ��� 0

cosh� + ���2�

sinh���2�

��� = ���,��(�� × �� × �� )�

��

�� coth����

+��ℎ2��

FM

PtJC d

t

z = 0x y

z

45

Interface transparency model

�����

���

Interface transparency of the spin current

� =�����

���=

�↑↓tanh(���2�)

�↑↓coth����

+��ℎ2��

�↑↓:intrinsic spin mixing conductanceλ: spin diffusion length of Pt σ: conductivity of Pt

FM

PtJC d

t

46

Interface transparency

���(�)

��� ∞= 1 − sech

�

�

Data

=1.2 nm =1.4 nm =1.6 nm

400 10060

0.00

0.02

0.04

0.06

d (Å)

0.08

20 80

� = 1.4 nm

d Pt/6 nm Py

Spin diffusion length

47

Interface transparency

Damping parameters for FM and Pt/FM obtained from conventional FMR

Pt/Py Pt/Co

Geff (1019m-2) 1.52±0.34 3.96±0.39

6 Pt/t Py, Co

Intrinsic mixing conductance

48

Interface transparency

Pt/Py Pt/Co

Geff (1019m-2) 1.52 ± 0.34 3.96 ± 0.39

σPt (µΩ*cm) 15 ± 1 15 ± 1

� (nm) 1.4 ± 0.2 1.4 ± 0.2

�↑↓ (1019m-2) 2.4 ± 0.4 11.1 ± 3.1

T 0.25 ± 0.05 0.65 ± 0.06

1

����=

1

�↑↓+

1

����

ℎ2��

SHA of Pt ~ 0.19

� =�����

���=

�↑↓tanh(���2�

)

�↑↓coth����

+��ℎ2��

Intrinsic mixing conductance

49

Interface transparency

0.0 0.2 0.4 0.6 0.8 1.00.00

0.05

0.10

0.15

x

higher G larger T higher SHA

� =�����

���=

�↑↓tanh(���2�

)

�↑↓coth����

+��ℎ2��

Co Ni

SHA of Pt in 6 nm Pt- 6 nm Co1-xNix

50

Interface transparency plays an important role for spin orbit torque

Py

PtJC

Co

PtJC

W. Zhang*, Wei Han*, Xin Jiang, See-Hun Yang, Stuart Parkin, Nat Phys. (2015)

Interface transparency

51

The main Challenge

How to get larger spin orbit torque?

FM

PtJC

Increase the Efficiency of the SOT

Search for larger SOT in 2D Quantum materials

52

1) Large SHE in 2D Ir-Mn

Motivated by a recent theoretical work of AHE in IrMn3

Large spin orbit coupling of Ir transfer to Mn.

Non-collinear antiferromagnetism

53

PML10

AFM

PM

Ir1-xMnx

c/a=1

L12

AFM

c/a>1

Ir47Mn53Ir

Mn

Ir25Mn75

Ir14Mn86

1) Large SHE in 2D Ir-Mn

54

Ir1-xMnx Grown on SiO2/Si substrates

Related work on Ir20Mn80 and IrMn have also been seen by other groups.IrMn: Zhang, W. et al. Phys. Rev. Lett. 113, 196602, (2014).Ir20Mn80: Mendes, J. B. S. et al. Phys. Rev. B 89, 140406, (2014)

1) Large SHE in 2D Ir-Mn

55

Ir1-xMnx Grown on SiO2/Si substrates

Related work on Ir20Mn80 and IrMn have also been seen by other groups.IrMn: Zhang, W. et al. Phys. Rev. Lett. 113, 196602, (2014).Ir20Mn80: Mendes, J. B. S. et al. Phys. Rev. B 89, 140406, (2014)

1) Large SHE in 2D Ir-Mn

56

Facet dependent SHE

1) Large SHE in 2D Ir-Mn

57

Facet dependent SHE

20 40 60 80 100 1200.00

0.05

0.10

0.15

0.20

(100) IrMn3

(111) IrMn3

p-IrMn3

S

HA

d (A)

1) Large SHE in 2D Ir-Mn

W. Zhang*, Wei Han*, Xin Jiang, See-Hun Yang, Stuart Parkin, under review

58

2) Spin orbit Torque in Topological insulators

59

2) Spin orbit Torque in Topological insulators

Mellink, et al. Nature (2014)

Spin Hall angle: 2.0-3.5

Bi2Se3

60

Spin orbit Torque in Topological insulators

Fan, et al, Nature Mater. (2014)

T=1.9 KSHA > 100

(Bi0.5Sb0.5)2Te3

61

Current Status of Spin orbit Torque

Materials Effective SHA Research group

Semiconductor GaAs 0.0005-0.005 UCSB (Awschalom)

Metal Pt 0.19 PKU (Han) & IBM (Parkin)

Cornell (Ralph & Burhman)

β-Ta 0.15 Cornell (Ralph & Burhman)

β-W 0.3 Cornell (Ralph & Burhman)

Bi doped Cu 0.24 Japan (Otani) & France (Fert)

Quantum

Materials

TI 2.0-3.5 (~500?)

>100 (1.9 K)

Cornell (Ralph & Burhman)

UCLA (Wang)

IrMn3 ~0.3 PKU (Han) & IBM (Parkin)

62

Summary of Lecture II

Doped graphene, Paramagnetic moment

H vacancy

FM in graphene by proximity effect

63

Summary of Lecture II

Doped TI, QAH

FM in TI by proximity effect

64

Interface Transparency for SOT

Facet dependent SOT in IrMn3

Py

PtJC

Co

PtJC

20 40 60 80 100 1200.00

0.05

0.10

0.15

0.20

(100) IrMn3

(111) IrMn3

p-IrMn3

S

HA

d (A)

Summary of Lecture II

TI has very large SOT

65

Summary of Lecture II

Questions still to be answered:

FM in graphene by doping

QAH in Graphene heterostructures

Robust QAH at higher temperature

More effective SOT in quantum materials

66

Acknowledgement

To the mentors for my research life

67

Roland K. Kawakami

Stuart. S. P. Parkin

Acknowledgement

To the mentors for my research life

68

Acknowledgement

The good team@PKU

69

Summary

70

Summary

Thank you for your attention!

Email: weihan@pku.edu.cn

71