Transcript Spin Hall Effect

Observation of the spin-Hall Effect
2005. 4. 18
M1 Colloquium in the Frontier Materials Division
Soichiro Sasaki
Suzuki-Kusakabe Lab.
Graduate School of Engineering Science
Osaka University
Abstract
• To realize spin-electronics, techniques to accumulate electron
spin in semiconductor devices are required. In the usual spinelectronics devices, spin injection is usually realized using
ferromagnets.
• The spin-Hall effect is another possibility and we can substitute
the effect for ferromagnets. Spin current flows in the direction
perpendicular to the direction of charge current when the voltage
is applied to an electron system in a semiconductor device. Spin
accumulation would occur at two edges as a result. This is called
the spin-Hall effect.
• The experiments [1,2] I'll introduce succeeded in observation of
the spin-Hall effect that was predicted theoretically. Spin
accumulation is observed with a Kerr rotation microscope in the
first one. Spin-dependence of polarization in a light-emitting
diode structure is an evidence in the second one. There are two
origins of spin-Hall effect, extrinsic spin-Hall effect and intrinsic
spin-Hall effect. The former experiment shows domination in the
extrinsic spin-Hall effect. The latter claims observation of
intrinsic
one.
References
[1] Y.K. Kato, et al., Science 306 (2004) 1910.
[2] J.Wunderlich,Phy. Rev. Lett. Vol.94 (2005) 047204
Contents
1. Introduction
2. Observation of the spin-Hall Effect
2.1 Kato’s spin-Hall device
2.2 Wunderlich’s spin-Hall device
3. Summary
1-1 Spin-electronics devices
Spin-FET
GMR device
Gate bias (Off)
Gate electrode
Source (FM)
Drain (FM)
Spin-FET by
Datta & Das
Current flows
TMR device
Channel
Gate bias (On)
Electrode
(Ferromagnel)
Insulator
(Oxides)
Electrode
(Ferromagnet)
No current flows
These devices are realized by using ferromagnets.
Problems of these devices
• We use ferromagnets to fabricate spin-electronic
devices. However, it is troublesome to fabricate
ferromagnets in semiconductor devices.
• We must use magnetic field to change the
direction of a magnetic moment. We can’t make
a device structure, which is controllable by
applied magnetic field, smaller than a predicted
limit.
Solution
Spin-Hall effect
1-2 semiconductor hetero junction
(a)
(b)
Ordinal heterojunction to realize
the two-dimensional electron gas.
(The 2DEG)
2DEG
Heterojunction may used as a n-type device.
The electronic system at the interface in AlGaAs/
GaAs junction produces the two-dimensoinal hole
gas (The 2DHG), when the bias is introduced as
shown by the black lines.
2DHG
1-3 The Hall effect
(Carriers are electrons)
Hall voltage ;
Lorentz force ;
Applied current ;
Density of carriers ;
Velocity of a carrier ;
n

v
Induced voltage ;
(Hall voltage)
I x Bz
Vy  
end
 evx Bz  eFy

 
(ev  B  eF )
I x  enhdvx
Vy  hFy
The elementary electric charge ;
e
1-4 Spin Hall Effect
In a nonmagnetic metal, the spin-orbit interaction causes
a pure spin current perpendicular to the applied electric field.
the vector of spincurrent whose spin is
polarized along z axis.
Spin accumulation
Motion of up-spins
Motion of down-spins
Extrinsic Spin-Hall effect
• This effect is caused by spin-orbit interaction of
impurities.
A phenomenological continuous equation of spin density and spin-flux density.
S 
t

q
x
   S  
S
s
0
Spin relaxation
q : Spin flux density
S : The spin density
Magnetic field effect
=mBgH/h
H: Magnetic field
Intrinsic Spin-Hall effect
• The electrons at the valence band
top, i.e. the holes follow an
equation of motion in the k-space,
in which a band-touching point the
acts as a Dirac magnetic monopole
in the momentum space.
• Valence band top of GaAs is a
typical example. Thus p-type
doped semiconductor is required to
have intrinsic SHE.
1-5 Hall effect v.s. Spin-Hall effect
Spin-Hall effect
Applied field
Accumulat
ed
quantity
Hall effect
Magnetic field
& Electric
field
Elctric
carge
Spin-Hall
effect
Spin-orbit
interaction
&
Electic field
Electron
spin
1-6 Kerr rotational microscopy
Kerr effect
Magnetic substance
• When incident linearly
polarized light (i) interacts
with a magnetic system the
reflected light (r) turns out
to be elliptically polarized
(the orientation of the
magnetization M is
perpendicular to the
surface).
Kerr rotational microscope
Kerr rotaional microscopy
(in the Santa Barbara)
Schematic of the experimental geometry
Phys. Today Vol. 58 Nov.2 17
(Feb. 2005)
Observations of spin-Hall effect
1. Y. K. Kato, R. C. Myers, A. C. Gossard, D.
D. Aschalom, Science Vol.306,1910
(2004);published online 11 November
2004 (10.1126/science.1105514).
2. J. Wunderlich, B. Kaestner, J. Sinova,
T.Jungwirth, Phys. Rev. Lett. Vol.
94,047204 (2005).
2.1 Y. K. Kato’ s experiment
Method of detection ;
Kerr rotational microscopy
Strong point ;
It can observe spatial map of spin.
Weak point ;
The limit of detected spot
(more than 1.1μm)
Long spin life time is needed to observe
the spin-Hall effect.
Kerr rotation in an unstrained sample
Schematics of the unstrained GaAs sample
Kerr rotation, peak KR,
and a spin lifetime as a
function of x or Bext
KR, peak KR, and
a spin lifetime as a
function of x, Bext & E
Spin accumulation
• Figure A: Twodimensional images
of spin density ns.
• Figure B: Twodimensional images
of reflectivity.
• These pictures are
direct observation of
the spin-Hall effect.
Hanle effect of unstrained
sample
A0
L S 
2
1
Larmor precession frequency;
g B Bext
L 

s
Electron spin lifetime;
Electron g factor;
g
Bohr magneton;  B
The peak Kerr rotation;
A0
Kerr rotation in strained sample
Schematics of the strained InxGa1-xAs sample
Kerr rotation, peak KR,
Bint and R as a
function of x or Bext
Hanle effect of strained sample
A1
A1

Kerr rotation ;
2
[( s )  1] [( s ) 2  1]
Kerr rotation peak ;
A1
Precession frequency;
g B ( Bext  Bint )
 

g B ( Bext  Bint )
 

The spin polarization is maximum
At Bext   Bint when Bext cancels out
Bint .
Result of Y. K. Kato’ s experiment
• The band structure that is the origin of intrinsic
one is broken by straining of the sample. This
experiment is caused, they think, by extrinsic
one.
• If the effect originated in the lattice, then it
should depend on the strain direction.
2.2. J. Wunderlich’s
experiment
Method of detection ;
Electroluminescence of LED
Strong point ;
It can detect magnetic field at very small spot.
The crystal used at this experiment is very clean crystal.
It has possibility of intrinsic spin-Hall effect.
Weak pint ;
It can’t detect spatial map of spin .
p-n junction LED device
Schematic cross section of the device
Conduction and valence band profiles
Schematic of sample and data
SEM image of the SHE LED device
b: Polarization along z axis measured with
active LED 1.
c: Polarization with fixed Ip for LED 1 or LED 2.
Result of J. Wunderlich’s
experiment
• Hitach researchers believe their
experiment to be too low in signal to
account for the polarization they observed.
The spin-Hall effect in their sample is, they
believe,intrinsic in origin.
Summary
• Using the Kerr rotational microscope, the extrinsic SpinHall effect was observed in a n-GaAs thin layer. (Kato’s
experiment)
• In a strained InGaAs device, no clear evidence of the
strain effect on the Spin-Hall effect was observed. Thus,
the intrinsic Spin-Hall effect is not observed in this
sample. (Kato’s experiment)
• Using electrolumimescemce from LED made of
2DHG/2DEG junction, spin polarization is found, which
depends on direction of the electric field across 2DHG.
This result suggest the intrinsic Spin-Hall effect.
(Wunderlich’s experiment)