Organic Semiconductors for Spintronic Applications V. Alek DEDIU

Institute for Nanostructered Materials CNR, BOLOGNA, ITALY

Organic Semiconductors for Spintronic Applications

V. Alek DEDIU

Italian School of Magnetism, Pavia 2012

Outline - Motivation - General notions on Organic SC - Brief introduction to injection and transport in OSC - Main achievements in Organic Spintronics > Spin Injection > New (multifunctional) devices - Conclusions: Problems and Possibilities

Most organic semiconductors are characterized by very weak spin scattering strength:

-  low Spin-Orbit Coupling due to low Z values ( SOC ∼ Z4 ) -

(McClure 1952 J Chem Phys)

OS in Spintronics - MOTIVATION

long spin relaxation times 10-6 – 10-3 sec -> transport of spin polarized signals to 102 – 103 nm)

even for very low mobility materials

Technological advantages: Easy to grow, low sensitivity to impurities

Nature Mater. 8, 707 (2009)

What is perhaps the most attractive aspect:

Stable and easily controllable interfaces with many inorganic materials – tuning of the spin injection ability via interface engineering

- backed by an enormous variety of molecules -

Tailoring interface spin selectivity

OSC in Spintronics - MOTIVATION

Chart representing the spin diffusion length lS as a function of the corresponding spin diffusion time, for various spintronics materials. The organic semiconductors cluster in the top-left corner. They have a long spin lifetime but, due to the typically low mobilities, short spin diffusion lengths. [S. Sanvito et al. Nat Mater 8, 963 (2009)].

Intriguing complementarity of Organic and Inorganic materials

Added: Alq3 – our data

Flexible PV cell (Konarka)

Organic field-effect transistors

Sony and Samsung 55-inch OLED tv

2010 - Sony Develops a "Rollable“ OTFT-driven OLED Display

Charges in OSC: intramolecular

Carbon - C - configuration: sp2-hybridised orbitals form a triangle within a plane and the pz-o r b i t a l s a r e i n t h e p l a n e perpendicular to it.

Example: in a benzene rings the

π -bonds become delocalized

http://www.orgworld.de

Small molecules vs polymers

Pentacene

S

S S S

S

S

α -6T

S S

S

S

α -4T

Organic Semiconductors

π-conjugated Molecules (oligomers)

a-quartertiophene

α-sexithiophene

Tris(8-hydroxyquinoline)aluminium (Alq3)

Rubrene

C60

Intermolecular interactions: van der Waals (vdW) – WEAK (10-100 meV) vdW: either (permanent dipole) x (induced dipole) or instantaneous (induced dipole) x (induced dipole)

Charges in OSC: intermolecular

⊗ This leads to: - strong carrier localization - very narrow “bands” < 0.1 eV - soft mechanical properties

Injection-Transport Properties

Charge injection: vanishingly low density of intrinsic carriers – about 1012 cm-3. The electrodes provide carriers: molecules can easily accommodate an extra charge

_

+

V cathode

anode

LUMO

HOMO

van der Waals intermolecular

Carrier injection into OSC is best described in terms of thermally and field assisted charge tunnelling across the inorganic/ organic interface followed by carrier diffusion into the bulk of OSC

_

+

V cathode

anode

LUMO

HOMO

van der Waals intermolecular

Arkhipov et al. JAP84, 848 (1998)

Injection-Transport Properties

The current J can either be injection limited or space charge limited (SCLC). The injection limited current cannot be expressed by an unique formula and has to be analyzed case by case- i.e. for any given interfacing – VERY IMPORTANT PROPERTY: THICKNESS INDEPENDENT – interface resistance dominates

_

+

V cathode

anode

LUMO

HOMO

van der Waals intermolecular

Injection-Transport Properties

The current J can either be injection limited or space charge limited (SCLC). The SCLC current is characterized by a strong thickness (d) dependence. It requires at least one Ohmic contact (unlimited injection efficiency):

JSCLC = µ(E) x V2/d3

_

+

V cathode

anode

LUMO

HOMO

van der Waals intermolecular

Injection-Transport Properties

The Diffusion plays a significant role in disordered low mobility OSC materials

JX = qnµEX + qD(dn/dx)

_

+

V cathode

anode

LUMO

HOMO

van der Waals intermolecular

Diffusion Drift

Considering carriers time of flight and hence spin relaxation one has to analyze carefully the diffusion time

Injection-Transport Properties

Charge Transport Properties: intermolecular

Tunneling vs Injection (devices)

Speaking about Electrical Injection of the Spin Polarization we can evidence two conceptually different approaches: -  Tunneling of SP carriers THROUGH OSC barriers (TMR) -  Injection of SP carriers INTO the OSC electronic states (GMR) Tunneling devices provide apparently higher MR signals: sensors, magnetic switching elements, … Injection devices provide the possibility to implement spintronic effects into OLEDs, OFETs, allow spin manipulation, … Tunneling-like injection into organics generated some confusion and sometimes these two processes are confounded. Sometimes tunneling was claimed in materials and thickness well known for light emission.

Spin Injection in OSC – where we are? Brief overview

T6

LSMO w=70-500 nm

V. Dediu, C. Taliani et al. Sol. St. Comm.122 (2002),181 Patent US6325914

Large negative magnetoresistance measured

Advantage: NO short circuits!

Problem: not possible (at least not at all easy) to reach AP configuration.

First Organic SP Injection device

Spin relaxation length LS ∼ 100 nm

Spin relaxation time τ ∼ 10-6 s

Z. H. Xiong, V. Vardeny et al. Nature 427, 821 (2004)

Univ. of Utah, Valy Vardeny group:

La0.7Sr0.3MnO3/Alq3(130nm)/Co

Inverse spin valve effect

Demonstration of the Spin Valve effect

The Spin Valve effects were registered in the ±1 V interval, up to 240 K

Supercon. Sci. Technol. 8, 160 (1995) Phys. Stat. Sol. 215, 1443 (1999)

La0.7Sr0.3MnO3 manganite

ISMN-CNR-Bologna

d = 12 nm

Z scale: 3 nm

Rq= 0,23 nm

Rpv= 1,27 nm

-3 -2 -1 0 1 2 3

-2

-1

0

1

2 10 nm

ΦK (

mde

g)

H (mT)

In col. with R. Sessoli group

La0.7Sr0.3MnO3/Alq3(130nm)/Co, both electrodes and Alq3, was a lucky choice

It becomes very fast the most used device in organic spintronics in an attempt to understand the physics rather than tsting new materials (or discovering new properties of old ones) Although the second approach becomes also vital and important. It generates new devices (see below)

Further developments

Alq3 3 - 250 nm

Co 15 nm LiF, AlOx 2 nm

Vertical Spin Valves: interface engineering

Long channels – injection!

La0.7Sr0.3MnO3 10-15 nm

NdGaO3 / SrTiO3

PRB78, 115203 (2008)

Evolution of LSMO-Alq3-Al2O3-Co spin valves at 100 K

paper in preparation, 2010

22 %

-1000 -500 0 500 1000260k

280k

300k

320k

340k

360k

380k

Res

ista

nce(Ω)

Field (Oe)

2007 data Rough Alq3 rms 5-10 nm

2008 data Smooth Alq3 rms 1 nm

2009-2010 optimized interfaces

arXiv:cond-mat/0701603

Room T GMR in 100nm Alq3 --------------- about 1-2 %

Various OSC in spintronics

Is the SPIN INJECTION in OSC recently fully demonstrated? MR alone is not a sufficient prove. No LEDs (actually OLEDs), not yet Hanle not yet non-local detection BUT: Two –photon photoemission spectroscopy (checking right injection) Muon Spin Rotation in a vertical injection device (injection/transport)

Spin Injection

Further developments

Cinchetti et al. Nature Mater.8,115(2009)

Drew et al. Nature Mater.8,109(2009)

Cinchetti et al. Nature Mater.8,115(2009)

What about OLEDs

No circularly polarized light is expected, but an efficiency enhancement Unless one works on triplet emission: E Shikoh, E Nakagawa and A Fujiwara, JoP 200 (2010) 062027

NON polarised electrodes

e(↑)p(↓)+e(↑)p(↑)+e(↓)p(↑)+e(↓)p(↓) =

1/2(S+T)+T+1/2(S+T)+T=S+3T = 25%

ONE polarised electrode

e(↑)+p(↑),p(↓) ->e(↑)p(↓) +e(↑)p(↑) =

1/2(S+T)+T=1/2S+3/2T = 25%

TWO polarised electrodes

e(↑)+p(↓) -> e(↑)p(↓) = 1/2(S+T) = 50%

e(↑)+p(↑) -> e(↑)p (↑) = T = 100%

Spin Polarized injection in Organic LEDs

TPD (70 nm)

Manganite based OLEDs

OLED off

OLED on

SPECTROMETER

SrTiO3 (0.5 mm)

La0.7Sr0.3MnO3

Alq3 (70 nm)

I

LiF (2 nm) Al

J. Appl. Phys. 93, 7682 (2003) J. Lumin. 110, 384 (2004) Org. Electron. 5, 309 (2004) (Bologna group) PRB 70, 085203 (2004) (IBM, Zurich)

In spite of many attempts, many OLEDs and much LIGHT: :::::::::::::::::::::::::::::::::::::::::::::::: no convincing SP effects detected

Fundamental (?) limitation

Need to increase maximal voltage for high density spin injection Possible solution (?): polarizing carriers injected by efficient non magnetic injectors Will SP injection in organics be ever applicable to OPTOELECTRONICS?

ISMN-Bologna

Signal Intensity

Voltage

Spin injection Light emission

Light emission

∼1 V ? ∼2 V ?

New Devices from Organic Spintronics

Magnetically Enhanced Memristor (MEM)

What is Memristor?

Memristors are resistors with memory

Memristor was formalized as the fourth basic electrical element by Chua in 1971

Chua, IEEE Trans. Circuit Theory, 507,18 (1971)

V

q i

ϕ

iRv =Normal Resistor

( )( )vGi

iqMvϕ=

=

Memristor

M= Memristance G= Memductance

Resistive Memories are Memristors

Pinched I-V Hysteresis Loop

At least two different non-volatile Resistive states

Chua L. , Appl. Phys. A, 102, 765 (2011)

First Experimental Memristor

9/28/11

Strukov, Nature, 453, 80 (2008)

Formally found in 2008 (bistable sytems are known from 60s)

TiO2 resistors. Memristance due To oxygen migration

‘Stateful’ logics from memristors

In 2010 Stateful Logics scheme was published: With a Memristor only crossbar array it is possible to compute and store data at the same device

Memristors are good candidates for future computing and memory applications

Borghetti et Al., Nature, 464, 873 (2010)

Alq3 3 - 250 nm

Co 15 nm LiF, AlOx 2 nm

Electric bistability spintronic device

Long channels – injection!

La0.7Sr0.3MnO3 10-15 nm

NdGaO3 / SrTiO3

Our devices: Bistability in I-V curves

-1.5 -1.0 -0.5 0.0 0.5 1.0 1.510-910-810-710-610-510-410-310-210-1

100K 150K 175K 200K 225K 250K 275K

Cur

rent

(A)

Bias (V)

-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5-10

-5

0

5

10

15

20

Cur

rent

[mA

]

Voltage [V]

1

2

3

4

5

6

Vth+

Vth-

MR can be switched OFF..

-3k -2k -1k 0 1k 2k 3k

3.8M

4.0M

4.2M

4.4M

4.6M

Res

ista

nce

(Ω)

Field (Oe)

100K -100mV

-3k -2k -1k 0 1k 2k 3k

280k

300k

320k

340k

360k

380k

Res

ista

nce

(Ω)

Field (Oe)

100K -100mV SV 22%

0%

..and ON -3k -2k -1k 0 1k 2k 3k

580k

600k

620k

640k

660k

680k

Res

ista

nce

(Ω)

Field (Oe)

100K -100mV SV 11.2%

Applying 2.5V

BEFORE measure

All measurement taken at -100mV

11%

22%

M. Prezioso et al. Adv Mat. (2011) , 23, 1371

Applying -1.5V BEFORE measure

It is possible to recover original MR

-3k -2k -1k 0 1k 2k 3k580k

600k

620k

640k

660k

680k

Res

ista

nce

(Ω)

Field (Oe)

-3.5k -3.0k -2.5k -2.0k -1.5k -1.0k -500.0 0.0 500.0 1.0k 1.5k 2.0k 2.5k 3.0k 3.5k

440k

460k

480k

500k

520k

540k

560k

580k

Res

ista

nce

(Ω)

Field (Oe)

2.5V

3.5V

3V

0%

11.2%

18.8% 21%

-3k -2k -1k 0 1k 2k 3k

3.8M

4.0M

4.2M

4.4M

4.6M

Res

ista

nce

(Ω)

Field (Oe)

-3k -2k 0 2k 3k420k

440k

460k

480k

500k

520k

540k

Res

ista

nce

(Ω)

Field (Oe)

Voltage Dependent

-3k -2k -1k 0 1k 2k 3k

440k

460k

480k

500k

520k

540k

560k

580k

Res

ista

nce

(Ω)

Field (Oe)

-3k -2k 0 2k 3k420k

440k

460k

480k

500k

520k

540k

Res

ista

nce

(Ω)

Field (Oe)

3V

18.8%

18.6%

behavior Scheme

Bistability mechanisms in Organics

•  Filamentary // tends to have multiple discrete levels

•  Redox reactions // difficult to imagine a reversible redox

reaction at such interfaces that can give such effect

•  Cobalt inclusions // ruled out by interfacial analyses

•  Conformational modifications // requires really high electric

fields

•  Charge trapping // Trapped charge can lower the mobility by a

sort of “coulomb blockade”

Simple model for the description of both GMR-RM

M. J. Rozenberg et al. PRL (2004), 92, 178302

MEM as Logic Gate

A (E) B(H) “AND”

1 0 0 0 0 0 1 1 1 0 1 0

A

B MR

Patent Pending

The experimental ‘AND’ truth table

The experimental ‘AND’ truth table

Switching can be done repeatedly

0 10 20 30 40 50 60

-3.0µ

-2.0µ

-1.0µ

0.0

1.0µ

Cur

rent

(A)

time (s)

turn on pulse+1.5V

turn off pulse-2V

(different sample different switching biases)

Writing/reading/Erasing Cycles. It can be done many times. The OFF state seems more reproducible

NON VOLATILE

0 200 400 600 800 1000

1E-9

1E-8

1E-7

1E-6

1E-5

- Cur

rent

[A]

Cycle

OFF ON

WRER Cycles +-4V

Electrical bistability Tested up to 14000 Cycles Roff/Ron up to 104

Retention time at least 24h@100K

New Physics (o Chemistry?)

-1500 -1000 -500 0 500 1000 1500

440k

460k

480k

500k

520k

540k

560k

580k

Res

ista

nce

(Ohm

s)

Field (Oe)

100K, V = -100mV

22 %

Inverse spin valve effect

LSMO(10nm)/ Alq3(200nm)/AlOx (2nm)/Co(30nm)

spin-up spin-up

LSMO(10nm)/ Alq3 (2 nm)/Co(30nm)

spin-up spin-up

Direct spin valve effect

C. Barraud, P. Seneor, AD, A. Fert, Nature Physics 6, 615 (2010)

-1500 -1000 -500 0 500 1000 1500

440k

460k

480k

500k

520k

540k

560k

580k

Res

ista

nce

(Ohm

s)

Field (Oe)

LSMO/Alq3 ∼200nm/AlOx/Co R(H) at 100 K

Injection into organic levels: Interface Spin selection “ON”

Direct tunneling allowed: Interface Spin selection “OFF”

Positive (T)MR

Negative (G)MR

C. Barraud, P. Seneor, AD, A. Fert, Nature Physics 6, 615 (2010)

S. Sanvito, News&Views, Nature Physics 6 (2010)

Interface bonding: various hybridization effects

Atodiresei, N. et al. Phys. Rev. Lett. 105, 066601 (2010).

First direct demonstration: Proximity induced SP states

Y. Zhan, M. Fahlman, et al. Adv. Materials 21, 1, 2009

Fe Alq3

The hyperfine scattering model(Bobbert et al. PRL 102, 156604 (2009))

features two important characteristics:

- weak temperature dependence -

- spin transport weakly dependent on mobility -

The latter results from the fact that hyperfine spin scattering is an intramolecular process, while momentum scattering (mobility) is exclusively intermolecular hopping effect

Recently proposed spin scattering mechanism

Recently proposed spin scattering mechanisms

CONCLUSIONS

Organic Spintronics it is still a young science – much to be understood Spin Injection – straightforward demonstration still missing, but many indirect evidences support that Fascinating interface physics/chemistry – many possibilities for spin-tuning at hybrid interfaces Nonetheless many questions still open – see above - new device paradigms are already coming out

Thanks to

Colleagues: I. Bergenti, A. Riminucci, M. Prezioso, P. Graziosi, F. Borgatti, A. Gambardella, R. Cecchini, E. Lunedei EC projects: OFSPIN, HINTS, IFOX MIUR: few FIRB and PRIN projects Many collaborators from differnt EU countries, USA, Japan and Russia