[PPT] - Spin tunnel and Spin Polarisation Laurent Ranno Laboratoire Louis PowerPoint Presentation

SLIDE 1

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Ecole Franco-Roumaine Magnétisme des systèmes nanoscopiques et structures hybrides Brasov, septembre 2003

Spin tunnel and Spin Polarisation

Laurent Ranno Laboratoire Louis Néel, Grenoble

SLIDE 2

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Summary

I- Introduction to Tunnel Effect II-Magnetic Tunnel Effect III-Bias Voltage and Temp Dependence IV-Spin Polarisation V- Half Metals

SLIDE 3

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

I- Introduction to Tunnel Effect

SLIDE 4

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

∆E w E E0

ψ ψ ψ E x V dx d m = + − ) ( 2

2 2 2

Tunnel Effect has a Quantum Mechanics Origin

A classical electron with energy E<E0 cannot enter the barrier zone However a quantum electron obeys the Schrödinger equation ! (1D model)

SLIDE 5

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

2

E

m q ∆ ± =

t i qr b

e

ω

ψ

−

=

Off the barrier In the barrier

) ( t kr i

e

ω

ψ

−

=

ψ ψ E dx d m = −

2 2 2

2

ψ

ψ ) ( 2

2 2

E E dx d m − = −

and

2

mE

k ± =

0 <

− E E

and and Plane waves Evanescent waves

SLIDE 6

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

nm q electron free m eV E 2 . 1 1 =

=

= ∆

2

E

m q ∆ ± =

Tunnel barriers must be very thin insulating layers Width = w < 10 nm

SLIDE 7

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

ψ ψ ψ E x V dx d m = + − ) ( 2

2 2 2

∆E

w x V(x)

=

±

x

du u k

e x

) (

) ( ψ

W. K. B. Approximation

ψ ψ ψ ) ( ) ) ( ( 2

2 2 2 2

x k E x V m dx d = − =

For a more general barrier shape E

SLIDE 8

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

What is neglected ?

=

±

x

du u k

e x

) (

) ( ψ ) ( ) ( ) ( x x k dx x d ψ ψ ± = ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) (

2 2 2

x x k x dx x dk dx x d x k x dx x dk dx x d ψ ψ ψ ψ ψ + ± = ± ± =

) ) ( ( 2 ) (

2

E x V m x k − =

The barrier potential should vary smoothly

SLIDE 9

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

We are dealing with transport. What about the current ? To pass a current, we must apply a bias voltage across the barrier. the barrier has a transmission coefficient T eV

SLIDE 10

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Simmons The probability to find the electron (energy E) on the other side of the barrier is :

=

= =

− −

w

du E u V m

e E P

2

) ) ( ( 2 2 2

) (

ψ

ψ ψ

dt dV E P V n V dN

x x x x

∞

= ) ( ) (

Electrons coming from the left to the right

∞

=

// 3 2

) ( ) ( 4 dE E f dE E P h m dt dN

E x x

π dt Vx

SLIDE 11

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

∆E w E E0 No bias voltage Current 1 2 = Current 2 1 J(V=0)=0 1 2

SLIDE 12

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

w E Bias voltage eV<<E0 eV

∞

+ − = − =

// 3 2 2 1

)] ( ) ( [ ) ( 4 ) ( dE eV E f E f dE E P h m e dt dN dt dN e J

E x x

π

E0

02 . 1 10

10 16 . 3

E S

e S V E J

−

=

Linear J(V) at low bias

Simmons (1963) has calculated approximate recipes

SLIDE 13

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

w E Bias voltage 0<eV<E0 eV E0 Simmons ’ parabolic fit

) (

3

V V J β α + = ) 3 1 (

2

V dV dJ β α + =

β contains the barrier height and the barrier width

SLIDE 14

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

dV dJ

Why ? From the experimental point of view :

) cos( ) ( t v V t V ω + = ) ( ) cos( ) ( )) cos( ( ) ( V dV dJ t v V J t v V J V J ω ω + = + =

Measuring the ω component of the signal with a lock-in amplifier gives directly the differential conductance Filter all the constant voltage and non-ω noise Using a voltage source :

V v <<

SLIDE 15

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Reduced effective w E Bias voltage eV>E0 eV E0 J increases rapidly but in fact such a bias voltage corresponds to the electrical breakdown regime

SLIDE 16

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

V V J dJ/dV

SLIDE 17

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

0.24
0.16
0.08

0.08 0.16

0.24
0.16
0.08

0.08 0.16 0.24 0.32

courbe I(V) à 2 K et 52 K

V (2 K) V (52 K)

V (Volt)

I-V non linear curves (ferromagnetic/insulator/ferromagnetic)

Thèse E. Favre-Nicolin (Grenoble 2003)

No temperature dependence of tunnel effect (1st order)

SLIDE 18

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

II-Magnetic Tunnel Effect

SLIDE 19

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel) In 1972 Gittleman et al. measured the resistance and MR of Ni grains in a SiO2 matrix

Gittleman et al. Phys. Rev. 5 (1972) 3609

50% Ni Longitudinal MR is <0 contrary to the sign of the bulk Ni AMR ρ// > ρ perp

R > R M J M J

SLIDE 20

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

2 2 1 1

) ( ) ( ) , ( m m m T T H T

⋅

+ = σ σ σ

Gittleman et al. 1972 :

SLIDE 21

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Development of film deposition techniques Trilayer : better characterisation of electrodes and control of magnetisation R changes by 14% at low temperature depending on the magnetic configuration

Fe Co Ge (10-15nm) +dry oxygen

SLIDE 22

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

11.8% at 300 K 24% at 24 K φ=1.9 eV and t=1.6 nm V50%=200mV 200µm x 300 µm

And also Miyazaki, Tezuka JMMM 139 (1995) L231

1995

SLIDE 23

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

IBM 1997

Ferro (free) Ferro (pinned) Insulator Magnetic Tunnel Junction

Same technical solutions as GMR structures to get 2 different coercive fields i.e. well defined parallel and antiparallel states. Hard - Soft materials (Co - NiFe) Different shape anisotropies for both electrodes Pinning to AF layer (MnFe) or Artificial AF layer (Co/Ru/Co)

SLIDE 24

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Jullière ’s model (1975) Not magnetisation BUT Polarisation of electrodes is the parameter

) ( ) ( ) ( ) (

F F F F

E N E N E N E N P

↓ ↑ ↓ ↑

+ − =

) ( ) ( ) ( ) (

2 1 2 1 F F F F

E N E N G E N E N G G

↓ ↓ ↑ ↑ ↑↑

+ = ) ( ) ( ) ( ) (

2 1 2 1 F F F F

E N E N G E N E N G G

↑ ↓ ↓ ↑ ↑↓

+ =

2 ) 1 (

i i i

P N N + =

↑

2 ) 1 (

i i i

P N N − =

↓

Assume : No spin-flip transition across the barrier at low voltage 2 parallel channels (spin up and spin down) Conductance is the sum of spin up and down conductances Conductance is proportional to the density of state (d.o.s.) 1 and d.o.s. 2

) ( ) (

2 electrode i 1 electrode i i F spin F spin spin

E N E N G G =

SLIDE 25

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Jullière ’s model

(M. Jullière, Phys. Lett. 54 A, 225 (1975))

2 1

N N G G G = +

↑↑ ↑↓ 2 2 1 1 2 2 1 2 2 1

P N P N G P N N G P N N G G G = − = −

↓ ↑ ↑↓ ↑↑ 2 1P

P G G G G = + −

↑↓ ↑↑ ↑↓ ↑↑ 2 1 2 1

1 2 P P P P G G G + = −

↑↑ ↑↓ ↑↑

r

Does depend on Pi Does not depend on the barrier (height, width) because of assumption about G0 i.e. no spin dependence of transmission

2 1 2 1

1 2 P P P P R R R − − = −

↑↑ ↑↓ ↑↑

TMR ratio : (pick your definition )

SLIDE 26

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

69.1% at 4.2 K From Jullière ’s formula PCoFe=50.7% similar to expected CoFe TMR junction (Tohoku 2000) Fe/a-Ge/Co (Jullière) Exp : TMR=14% Theor : PCo34% +PFe44% TMR 26%

SLIDE 27

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Slonczewski ’s model (1989) Barrier in the model m

↑ F

k ↓ ↑

↓ F

k

exch

E m k E ± = 2

2 2

Ferromagnetic electrodes

↓ ↑ ↓ ↑ ↓ ↑ ↓ ↑

+ − + − =

F F F F F F F F

k k k k k k q k k q P

2 2

Solve Schrödinger for both channels, calculate conductances

SLIDE 28

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

↓ ↑ ↓ ↑ ↓ ↑ ↓ ↑

+ − + − =

F F F F F F F F

k k k k k k q k k q P

2 2

) ( ) ( ) ( ) (

F F F F

E N E N E N E N P

↓ ↑ ↓ ↑

+ − =

2

E

m q ∆ ± =

High barrier

↓ ↑ ↓ ↑ ↓ ↑ ↓ ↑

+ − = + − ⋅ =

F F F F F F F F

k k k k k k k k q q P

2 2

Free electrons k mk mE m E DOS ∝ = =

2 2 2 3

2 ) ( π π

Back to Jullière ’s formula

SLIDE 29

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Improved models : Bratkovsky : Correction to Slonczewski’s model (different effective mass in the barrier) mb/m=0.4 for Fe/Al2O3 Ab initio band structure calculations : to get the band structure close to the interface

↓ ↑ ↓ ↑ ↓ ↑ ↓ ↑

+ − ⋅ + ⋅ − =

F F F F F F b F F b

k k k k k k m q k k m q P

2 2 2 2

2

E

m q

b∆

± =

SLIDE 30

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

III-Bias and Temp Dependence

SLIDE 31

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

TMR temperature dependence Resistance Temperature dependence

Miyazaki group (Tohoku) APL 2000

SLIDE 32

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Temperature dependence of resistance

SLIDE 33

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

2

E

m q ∆ =

qw

e T

n

Transmissi

2 −

∝ =

w ∆E Temperature dependence of the barrier transmission Going from 0 Kelvin to 300 Kelvin Wavevector in the barrier (evanescent wave)

n

Transmissi e Conductanc ∝

SLIDE 34

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

2

E

m q ∆ =

qw

e T

n

Transmissi

2 −

∝ = E E d wq dq w qw d T dT ∆ ∆ − = ⋅ − = − = 2 ) 2 (

meV kT E eV E nm w q 25 d , 2 , 1 , 1

1

= = ∆ = ∆ = Α =

− °

% 5 . 12 = ∆ = ∆ = ∆ R R G G T T

w ∆E

SLIDE 35

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Temperature dependence of TMR

SLIDE 36

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Polarisation is related to magnetisation

) 1 )( ( ) (

2 / 3

T T M T M

s

α − = ) 1 )( ( ) (

2 / 3

T T P T P

s

α − =

Inelastic processes can appear

magnon e e +

↓

↑ ↓ ↑

+

e magnon e

Opens a spin-flip conductance channel conductance increases TMR decreases Surface magnetisation is less robust to thermal fluctuations No general results, depends on the studied system (Tc, surface state …) Different contributions may rule this behaviour :

SLIDE 37

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Voltage dependence of TMR

SLIDE 38

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Jullière, Phys. Lett. 1975

50% MR at 3mV TMR50%=3mV

SLIDE 39

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

TMR - bias voltage dependence 50% decrease TMR for 400mV R - bias voltage dependence

Miyazaki group (Tohoku) APL 2000

SLIDE 40

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Inelastic processes can be activated

magnon e e +

↓

↑ ↓ ↑

+

e magnon e

Opens a spin-flip conductance channel TMR decreases with V The voltage decrease depends on experimental systems and years At large bias voltages, hot electrons are introduced in the second electrode : 0.1 V = 1200 Kelvin 1975 : TMR50%=2 mV 1995 : TMR50%=200 mV 2000 : TMR50%=450 mV 2003 : TMR50%>1000 mV May be due to non perfect samples, which improve with time

SLIDE 41

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Epitaxial NiFe electrode : 50 % decrease of TMR at 750 mV

Yu et al. APL 2003

SLIDE 42

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Ding et al. (M.P.I. Halle) PRL 2003

Cobalt (0001) Magnetic amorphous tip Tunnel junction with « perfect barrier » : STM with ferromagnetic electrodes in vacuum Voltage dependence of TMR not related to magnon excitations or surface magnetisation but more likely to defects in the barrier to be confirmed ...

SLIDE 43

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

A few words about The insulating barrier

SLIDE 44

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Making the barrier Aluminium Film (0.7-2 nm) + thermal oxidation in oxygen atmosphere or air Aluminium Film + oxygen plasma Deposition of Alumina (Al2O3) produces a less dense barrier with poor electrical properties Good recipe #1 Good recipe #2 Bad recipe #1

SLIDE 45

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

S R 1 ∝

Constant thickness barrier

Gallagher et al. JAP 1997

5 orders of magnitude 100 nm 100 µm

SLIDE 46

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Electrical Breakdown 1 Volt across a 1 nm barrier is E = 1 GigaV/m This is the order of magnitude of the electrical field necessary to ionise an atom The barrier should be uniform

flat
0 roughness
fully oxidised, homogeneously
no oxidation of the ferromagnetic metals

A 0.1 nm fluctuation means 100 mV decrease of the breakdown voltage !

SLIDE 47

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel) Ta2O5 -1023 HfO2

1144

Nb2O5

949

ZrO2

1100

CeO2

1088

Y2O3

952

TiO2

944

Gd2O3

909

SiO2

910

Nd2O3

903

NbO2

796

La2O3

896

CrO2

598

Al2O3

837

MnO2

520

V2O3

609

Cr2O3

570

MgO

601

Ga2O3

544

NbO

405

Mn2O3

479

MnO

385

Fe2O3

412

ZnO

350

FeO

272

Mn3O4

462

NiO

238

Fe3O4

372

CoO

237

Co3O4

297

CuO

157

SiO

99

Ga2O

178

Cu2O

84

Formation enthalpy ∆H (298K) kJ/mol metallic atom Al-O bond is stronger than 3d metal -O bonds Better oxides exist : HfO2, , Rare earth-O But kinetics : (Al passivated « automatically » 1 nm thick), local state : (amorphous, nanocrystallised, crystallised) when RT deposition

SLIDE 48

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Crystallised barrier Amorphous barrier

Smith et al. JAP,83 ,5154 (1998)

SLIDE 49

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Role of annealing a tunnel junction

Sousa et al. (Lisbon)

Barrier (Simmons’ fit) after anneal : thickness :0.87 to 0.77 nm wide barrier height :1.8 eV to 2.5 eV

SLIDE 50

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

RBS measurement of Al and O distributions

Sousa et al. APL 98

O from the CoFe electrodes goes back to Al2O3

SLIDE 51

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel) Miyazaki group (Tohoku) APL 2000

Ta (5 nm)/ Ni79Fe21 (3 nm)/ Cu (20 nm)/ Ni79Fe21(3 nm)/ Ir22Mn78(10nm)/ Co75Fe25(4 nm)/ Al (0.8 nm)-oxide/ Co75Fe25(4 nm)/ Ni79Fe21(20 nm)/ Ta (5nm) Free layer Pinned layer Best present TMR junctions : 50% Room temperature (non exotic materials)

SLIDE 52

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

IV-Spin polarisation

SLIDE 53

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

We have been using values for P. Where do they come from ? How to measure them ?

) ( ) ( ) ( ) (

F F F F

E N E N E N E N P

↓ ↑ ↓ ↑

+ − =

SLIDE 54

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Density of States Band structure calculations (spin resolved)

Moroni et al. PRB56 (1997)15629

[100] [000] [000] [000] [000] [110] [111] [100] [111] [210] [110]

SLIDE 55

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Nickel integrated density of states Polarisation at Fermi level should be NEGATIVE

SLIDE 56

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Be careful (M. B. Stearns JMMM,5 ,167 (1977))

2 2 2 *

k E m ∂ ∂ =

Bands do not have the same effective mass at EF Electrons at Fermi level have different mobilities d-like electrons are more localised (narrow bands) s-like electrons are less localised (wide bands) and more mobile

SLIDE 57

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

[100] [000] [000] [000] [000] [110] [111] [100] [111] [210] [110]

s-like d-like

SLIDE 58

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Be careful (M. B. Stearns JMMM,5 ,167 (1977))

2 2 2 *

k E m ∂ ∂ =

Bands do not have the same effective mass at EF Electrons at Fermi level have different mobilities d-like electrons are more localised (narrow bands) s-like electrons are less localised (wide bands) and more mobile Ferromagnetism comes from the 3d bands but transport comes from s-electrons. s electrons are not supposed to be polarised ! Difficult to predict the polarisation of conduction electrons

SLIDE 59

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Ferromagnetic / insulating / superconducting junctions

M. I. T. speciality (Tedrow/Meservey)
∞

+ − ∝

2 1 2

)] ( ) ( )[ ( ) ( dE eV E f E f E N E N M J

qw

e M

2 2 −

∝

N(E) non magnetic metal

constant ) ( ∝ E N

When it is difficult to predict, let us measure !

SLIDE 60

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

∞

+ − ∝

2 1

)] ( ) ( )[ ( ) ( dE eV E f E f E N E N J ∆ < = E if ) (E N ∆ > ∆ − = E if E ) (

2 2

E E N

N(E) in a superconductor ∆ −∆ E N(E) f(E) f(E+eV) E E EF f(E)-f(E+eV) E

SLIDE 61

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

∞

+ − ∝

super metal

)] ( ) ( )[ ( ) ( dE eV E f E f E N E N J

∆/e V J ∆/e V

dV dJ

SLIDE 62

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

V

dV dJ

If a magnetic field is applied B B E B E

B

µ ± = = ) ( ) ( If 50% electrons spin up, 50% electrons spin down V

dV dJ

V

dV dJ

µB/e If not 50%-50% ….

SLIDE 63

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

M.I.T group JAP 1979

SLIDE 64

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel) Meservey et al; J.A.P. 50 (1979)

Fe Co Ni Gd +44% +34% +11% +4%

SLIDE 65

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Andreev reflexion : metal / superconductor clean interface

Soulen et al. J.A.P. 85 (1999)

SLIDE 66

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Nb gap I(V) dI/dV Normal metal : 1 electron arrives, 2 electrons enter

SLIDE 67

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

SLIDE 68

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Nb gap I(V) dI/dV No conductance at small bias

SLIDE 69

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

10 20 30 40 50 60 70 80 90 100 N i F e C

N

i F e N i M n S b L a S r M n O C r O 2 Spin Polarisation (%) Tunneling Andreev

Soulen et al. 1999

But it only works at 1 Kelvin ! Andreev Reflexion does not give the sign of P

SLIDE 70

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Material Surface (UHV conditions) Photon hν Photoemitted electron Energy+spin analysis Photon penetration length = 0.5 nm Very clean surface, UHV prepared surface UV To analyse the conduction electrons : spin resolved photoemission

SLIDE 71

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Jonker et al. PRL57 (1986)

Example : Epitaxial Fe on Ag Spin-resolved photoemission

SLIDE 72

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Park et al. PRL 81 (1998)

Photoemission signal La0.7Sr0.3MnO3 film Photon hν = 40eV 40K

SLIDE 73

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Polarisation at 0K for Ni, Co and Fe is known Its temperature dependence is still unclear Surface polarisation still a mystery (enough to keep you busy after Brasov school) What about the other magnetic metals ?

SLIDE 74

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

V-Half Metallic Ferromagnets

SLIDE 75

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

D(E) E

↓ ↑ > s s

λ λ

d d s s Ferromagnetic band structure

SLIDE 76

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Half metallic ferromagnets

D(E) E d d s s

Fully spin polarised conduction band

No s-band at Fermi level Large band splitting

Gap = not usual metals = less metallic than metals

SLIDE 77

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Predicted HMF 1984 : ab initio calculcations (De Groot et al.) NiMnSb (Tc=720 K), PtMnSb (Tc=575 K) CrO2 La0.7Sr0.3MnO3 Fe3O4 FexCo1–xS2 …………. magnetic semiconductors ?

SLIDE 78

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Compound Curie temperature Magnetisation µoMs Crystallographic structure La0.7Sr0.3MnO3 350 K 0.74 T 3.7 µB/Mn Perovskite NiMnSb 728 K 0.89 T 4 µB/Mn Semi Heusler CrO2 396 K 0.81T 2 µB/Cr Rutile Fe3O4 860 K 0.63 T 4 µB/F.U. Inverse Spinel PtMnSb 572 K 0.9 T 4 µB/Mn Semi Heusler Sr2FeMoO6 415 K 0.73 T 4 µB/F.U. Double perovskite

SLIDE 79

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Spin-polarized band structure of La0:7Sr0:3MnO3. The majority bands are shown as solid lines, and the minority as dashed lines Livesay et al. J. Phys.: Condens. Matter 11 (1999) L279–L285

Predicted Half metallic character of La0.7Sr0.3MnO3 Zero K picture

SLIDE 80

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

EF ∆ δ E D (E) D (E) E(q) q E=Dq2 ∆ δ Stoner continuum Spin waves

Electronic excitations in a HMF Still exist at T>0

SLIDE 81

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

HMF are a very interesting class of materials Useful to study HMF state Are they really 100% polarised ? Up to Tc ? Is the surface polarised ? Useful to study other materials Source of d-like electrons Highly spin polarised source

SLIDE 82

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Inverse TMR Cobalt : positive P (previous studies) LaSrMnO : positive P (no down bands) and negative TMR !

DeTeresa et al. PRL 82(1999)

d electrons tunnel

SLIDE 83

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Conclusions : Tunnel is old but TMR is quite a recent field Spin-dependent transport properties are not fully understood HMF is still a mystery A lot of material science is necessary to control the structures Only a few materials and structures have been controled Bringing together magnetism and electronics is a fruitful playground ...

SLIDE 84

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

SLIDE 85

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Improved structures : resonant junctions

SLIDE 86

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Ivo Sturm Manouk Rijpstra

SLIDE 87

Spin Tunnel Course Brasov sept. 2003 Laurent Ranno (Lab. Louis Néel)

Bowen et al. APL 2001