review organic spintronicsrsta.royalsocietypublishing.org/content/roypta/369/1948/3054.full.pdf ·...

Phil. Trans. R. Soc. A (2011) 369, 3054–3068doi:10.1098/rsta.2011.0155

REVIEW

Organic spintronicsBY I. BERGENTI*, V. DEDIU, M. PREZIOSO AND A. RIMINUCCI

Institute of Nanostructured Materials, ISMN-CNR, Via P. Gobetti101 40129 Bologna, Italy

Organic semiconductors are emerging materials in the field of spintronics. Successfulachievements include their use as a tunnel barrier in magnetoresistive tunnelling devicesand as a medium for spin-polarized current in transport devices. In this paper, wegive an overview of the basic concepts of spin transport in organic semiconductors andpresent the results obtained in the field, highlighting the open questions that have tobe addressed in order to improve devices performance and reproducibility. The mostchallenging perspectives will be discussed and a possible evolution of organic spin devicesfeaturing multi-functional operation is presented.

Keywords: magnetic thin film; spin injection; spin transport; organic semiconductors

1. Introduction

The advancement of technology based on organic semiconductors (OSCs) overthe past decade has been spectacular. Organic light-emitting diodes (OLEDs)have already become available to consumers, and organic photovoltaic (OPV)devices are about to enter the market with new applications. New fieldssuch as electric memories for non-volatile data storage and organic field effecttransistors (OFETs), with an established reproducibility in research laboratories,are on the verge of reaching component production. Generally, OSCs offer manyfundamental advantages, including low cost, lightweight, large area coverage,relatively easy engineering of molecular properties and mechanical flexibilitycompatible with plastic substrates.

Challenging opportunities for OSCs have recently emerged in the field ofspintronics, mainly owing to the weakness of the spin-relaxation mechanisms[1]. Spin–orbit coupling (SOC) is very small in most OSCs—carbon has a lowatomic number (Z) and the strength of the spin–orbit interaction is proportionalto Z4 [2]. Furthermore, the hyperfine interaction is expected to be too weak, giventhe p-conjugated nature of transport molecular orbitals and the absence of themagnetic moment in the 12C nuclei [3]. As a consequence, the spin polarizationof the carriers is expected to be retained for very long times, allowing spins to be

*Author for correspondence ([email protected]).

One contribution of 12 to a Theme Issue ‘New directions in spintronics’.

This journal is © 2011 The Royal Society3054

on May 15, 2018http://rsta.royalsocietypublishing.org/Downloaded from

mailto:[email protected]

http://rsta.royalsocietypublishing.org/

Review. Organic spintronics 3055

manipulated. Furthermore, the use of OSCs opens the way to a new generationof spintronic devices with multi-functional behaviour, enabling the sensing, evensimultaneously, of magnetic, electrical, optical and chemical stimuli.

Spin injection in OSCs was initially demonstrated in a series of pioneeringstudies [4,5] on the magnetoresistance (MR) of spin valve (SV) devices withOSCs layers approximately 100–200 nm thick. More recently, this effect wasadditionally confirmed by a direct observation of the spin polarization, at tensof nanometres from the injection interface, involving two powerful techniques:muon spin rotation [6] and two-photon photoemission (2PPE) spectroscopy [7].Also the use of OSCs in tunnelling devices has shown considerable achievements,namely, room temperature operation [8,9] and very high MR values (up to 600%[10]) at low temperatures.

In spite of many fundamental achievements, OSCs-based spintronic devicesshow a wide scatter in device performance, indicating that a significant numberof limiting factors of both technological and fundamental nature have yet to beidentified and controlled. OSCs-based spintronic devices usually include organiccharge/spin-conducting spacer and inorganic ferromagnetic (FM, spin-polarized)electrodes. The performance of such devices obviously depends on the propertiesof the materials used as the organic spacer and magnetic injecting electrodes,but it is also critically influenced by the interfacial properties. Understanding thecomplex electric and magnetic phenomena taking place at these hybrid interfacesis one of the keys for the definition and realization of the potential applicationsof OSCs spintronic devices. In this paper, we will present a brief introductionof the basic properties of OSCs materials followed by the major achievementsobtained in organic spintronics. At the end of the paper, a short overview ofthe main outstanding issues as well as of the most challenging perspectives willbe given.

2. General properties of p-conjugated organic semiconductors

While in inorganic semiconductors the covalent bonds, together with structuresymmetry, promote significant carrier delocalization and the establishment ofa well-defined band structure, OSCs combine a strong intramolecular covalentcarbon–carbon bonding with a weak van der Waals interaction between molecules.This combination of different interactions generates materials with opticalproperties that are very similar to those of their constituent molecules, whereastheir transport properties are governed by the intermolecular interactions.

Most of the properties of OSCs derive from the so-called conjugation: thealternation of single and double bonds in the molecule. pz -orbitals of sp2-hybridized C-atoms in the molecules cause the formation of pseudo-bands, namelyp-bands, which consist of the delocalized electrons over the single molecule.The highest energy electron orbital (analogous to the valence band) is calledthe highest occupied molecular orbital (HOMO), and the lowest energy availableunfilled orbital (analogous to the conduction band) is called the lowest unoccupiedmolecular orbital (LUMO).

p-conjugated OSCs can be grouped into two classes: oligomers and polymers.These two cases differ in their physical and chemical properties and also in theway they are processed. OSCs oligomers are small molecules with an average

Phil. Trans. R. Soc. A (2011)



3056 I. Bergenti et al.

size close to 1 nm, which can be grown in either crystal or thin film form byhigh- or ultra-high-vacuum (UHV) molecular beam deposition. A few of the mainrepresentatives of these kinds of materials are the classes of oligothiophenes (4T,6T and others), metal chelates (Alq3, Coq3 and others), metal–phtalocyanines(CuPc, ZnPc and others), acenes (tetracene, pentacene and others), rubreneand many more. Polymers, on the other hand, are long chains formed froma given monomer, and usually feature connection of oligo-segments of variouslengths, leading to a random distribution of conjugation lengths. Polymer filmsare strongly disordered and totally different processing technologies have to beapplied, including ink-jet printing, spin coating, drop casting and other veryinexpensive methods. However, these growth techniques generally preclude thepossibility of real interface control, as the electrode surface is not as clean as inUHV conditions.

The electrical properties of OSCs are usually described by two main processes:charge injection and charge transport. The charge injection process from ametal into an OSCs is conceptually different from the case of metals/inorganicsemiconductors, mainly because of the vanishingly low density of intrinsiccarriers—about 1010 cm−3 in OSCs. The external electrodes provide the electricalcarriers responsible for the transport as the OSCs molecules can easilyaccommodate an extra charge, although this charge modifies considerably themolecular spatial conformation and causes strong polaronic effects. The carrierspropagate via a random site-to-site hopping between pseudo-localized states,while the external electric field defines the drift direction. Transport significantlydepends on the degree of order of the molecules as well as on the density ofchemical and/or structural defects. Puzzlingly, in OSCs, the impurities do notinduce additional carriers as in inorganic semiconductors, but mainly generatetrap states. Energetically distributed shallow (about/below 0.1 eV) and deep(about 0.5 eV, but even up to 1 eV) traps are known to characterize the electricalproperties (especially the mobility) of many OSCs, their role being stronger inlow-order materials such as Alq3 and polymers [11]. This dependence explainswhy, over recent decades, the experimental characterization of the transportproperties in organic thin films or crystals was mainly affected by sample quality.As a result of the hopping transport, the mobility of charge carriers (electronsand holes) in OSCs is several orders of magnitude lower than in inorganicsemiconductors. At room temperature, electron and hole mobility in amorphousOSCs are in the range of m ∼10−6 to 10−2 cm2 Vs−1. High-purity single crystals ofthe OSCs rubrene reach mh ∼10 cm2 Vs−1. For comparison, mobility in high-puritySi crystals reaches m ∼103 cm2 Vs−1.

3. Spintronics with organic semiconductors

Aspects related to spin-dependent phenomena in inorganic materials, eithersemiconductors or metals, have been discussed over recent decades and mainlyrely on the detection of a spin-polarized current in electronic devices. Thisapproach corresponds to the most common exploitation of spintronic effects indevices, and it is widely accepted as the standard method for proving the abilityto control spin polarization. The prototypical spintronic device is the SV, inwhich two ferromagnetic electrodes of different coercivity are decoupled by a





LSMO

(a)

(b)

Hext

Co

6T

0

–1

–2

–3.0 –1.5 0magnetic field (k Oe)

1.5 3.0

DR/R

(%

)

Figure 1. (a) Schematic drawing of a vertical inorganic/organic SV with LSMO as the bottomelectrode and cobalt as the top one. (b) The response in a 6 T-based SV is obtained by applying abias voltage to the metallic FM electrodes (−0.1 V) and the current is measured as a function ofthe in-plane external magnetic field. (Online version in colour.)

spacer. The resistance of such a device can be switched between two differentstates corresponding to parallel and anti-parallel magnetization of the FM layers(figure 1). The MR is defined as the ratio between the resistance difference ofthe two states and the resistance at zero applied magnetic field. If a thin layerof insulator is chosen as the spacer, we deal with tunnelling magnetoresistanceand the device is characterized by a zero residence time of charges/spins in thebarrier. When the spacer is thicker, we deal with injection devices in which thereis a net flow of spin-polarized carriers directly into the electronic levels of theOSCs. We will discuss this difference in the following two subsections, presentingfirst the case of injection of spin-polarized carriers into the electronic levels ofOSCs and then the tunnelling case.

(a) Organic injection devices

Injection devices are characterized by a finite lifetime of the generated spinpolarization inside the organic material. As mentioned above, the OSCs arecharacterized by long spin-relaxation times: these times range from 10−6 to10−3 s [1,12] and exceed by orders of magnitude the spin-relaxation times in





most inorganic materials (10−9–10−12 s). It is widely expected that OSCs-basedspin-injection devices can offer appealing opportunities for spin manipulation,although it must be pointed out that the understanding of the physics which liesbehind and beyond such manipulation is almost completely missing.

The first report on spintronic effects in organic spintronic devices was publishedin 2002. Lateral devices were lithographically fabricated and consisted of twoFM planar electrodes placed at 100–500 nm from each other and connectedafterwards by an OSCs thin film. The FM electrodes were made from manganiteLa0.7Sr0.3MnO3 (LSMO), well known for its colossal magnetoresistive behaviourand half-metallic properties (nearly 100% spin polarization at 0K). Sexithiophene(6T), a rigid conjugated oligomer rod, was chosen for the spin-transport channel.A strong MR was recorded up to room temperature across 100 and 200 nm of6T, while no MR was found for longer channels. The results were explained asbeing a consequence of the conservation of the spin polarization of the injectedcarriers. Using the time-of-flight approach, a spin-relaxation time of the orderof 1 ms was deduced [4]. Nevertheless, this experiment could not be consideredas the definitive demonstration of the spin injection owing to the fact that nocomparison between parallel and anti-parallel magnetization of electrodes wasavailable.

Soon after this experiment, a device with output characteristics similar tothe known inorganic SVs was presented [5]. In this case, vertical devices werefabricated with LSMO as the bottom electrode, a Co thin film as the topone, and 100–200 nm of tris(8-hydroxyquinoline)aluminium(III) (Alq3) as theconducting spacer. Alq3 was well known for its applications in OLEDs, whereit acts as an efficient emitter. Its use in SVs initially originated from the ideathat spintronics effects could be detected optically. A strong MR was detectedat low temperatures, which vanished at above 210–240 K. These results havebeen repeatedly confirmed by a number of research groups worldwide for thesematerials and similar geometry and thicknesses [13,14]. An improvement in deviceoperation was obtained by careful control of the ferromagnet/organic interfacethrough the insertion of an insulating tunnel barrier between the Alq3 and thetop Co electrode, allowing room-temperature operation [15].

Most of the LSMO/Alq3/Co devices with and without tunnel barriers haveshown negative MR, that is, the resistance decreased when the magnetizationof the two electrodes was switched from parallel to anti-parallel. The origin ofthis inversion is still the subject of discussion. Spin-resonant effects related tothe matching of the narrow LUMO level (less than 0.1 eV) conducting spin-up(spin-down) carriers in OSCs with peaks of the spin-down (spin-up) DOS in thecollecting electrodes have been tentatively proposed previously [15]. Recently,Barraud et al. [10] advanced a model explaining the sign inversion by spin-selective filtering at interfaces with strong binding between OSCs molecules andthe electronic state in the electrode.

Electrical detection of spin injection and transport has been pursued in a num-ber of other OSCs: a-NPD (N ,N -bis 1-naphthalenyl-N ,N -bis phenylbenzidiane)and CVB (4,4′-(bis(9-ethyl-3-carbazovinylene)-1,1′-biphenyl) [16] were demons-trated to behave in a very similar way to Alq3 with LSMO and Co magneticelectrodes. The same electrodes have been used to investigate spin injection intetraphenylporphyrin; however, the detected MR in this case was attributed tothe tunnelling effects [14].





All the experiments described above were performed using UHV conditions forthe OSCs growth. Spin-coated polymers, such as regio-random and regio-regularpoly(3-hexylthiophene) (RRaP3HT and RRP3HT, respectively), have also shownnoticeable MR [17,18]. In these reports, the SV effect was positive and a smallvalue was recorded even at room temperature.

In addition to these magnetoresistive approaches, two powerful techniques havebeen recently exploited to confirm, in a direct way, the spin injection and spintransport into an organic layer; namely, low-energy muon spin rotation (mSR)[6] and time-resolved 2PPE [7]. mSR measurements were carried out duringdevice operation in a bilayered organic structure of Alq3 and NPD sandwichedbetween two FM contacts; a clear correlation between spin polarization loss in theorganic material and the magnetoresistive signal was detected. The spin diffusionlength was estimated to be close to 30 nm for Alq3. The second experimentused a spectroscopic technique based on the analysis of energy dispersion andspin polarization of photoemitted electrons obtained by an appropriate laserexcitation. The experiment provided direct information about the efficiency ofspin injection. Cinchetti et al. [7] presented results on phthalocyanine layers grownon cobalt thin films, demonstrating very high spin-injection efficiency with a spindiffusion length close to 100 nm.

In spite of these well-established results, some effects related to the permanenceof spin-polarized carriers inside the OSCs, such as the Hanle effect, have not beenyet detected. While this is one of the most intriguing outstanding issues, a clearunderstanding of whether the Hanle effect can be detected in a system withincoherent hopping transport of carriers is still debated. Alternative methodssuccessfully applied in inorganic semiconductor spintronics for testing the spintransport, such as optical generation of spin-polarized carriers and studies onspin-relaxation and spin-coherence times, cannot be transferred to OSCs ina straightforward way owing to the weak SOC making the optical methodunfeasible.

(b) Organic tunnelling devices

Similar to inorganic spintronics, devices involving spin-polarized tunnellingacross an organic barrier show much higher MR, having already achievedvalues attractive for sensor applications. Initially, tunnelling was demonstratedin non-traditional complex geometries such as nanopores with a self-assembledmonolayer of octanethiol barrier [19] or devices with Langmuir–Blodgettmonolayer barriers [20]. Such devices generally have shown MR effects, but theseeffects were quite irreproducible, clearly indicating a strong influence by a numberof parameters that are difficult to control.

Surprisingly, the utilization of patterning technologies more akin to inorganicspintronics has immediately led to significant breakthroughs. Sub-millimetre-sized SVs obtained with standard shadow masking showed 6 per cent MR at roomtemperature involving hybrid organic/inorganic Alq3/Al2O3 and rubrene/Al2O3barriers [8,21]. In these studies, the spin polarization of carriers was accuratelymeasured with the Meservey–Tedrow method [22]. Patterning by UV lithographyallowed tunnel junctions to be produced with areas ranging from 5 × 5 to 50 ×150 mm2 involving pinned CoFeB/MgO (exchange-biased) and free Co magneticelectrodes and Alq3 barriers [9]. Up to 16 per cent tunnelling MR at room





temperature and a significant yield of reproducibly working devices have beenachieved in this case. Following this series, the most original patterning has beenrecently proposed by Barraud and co-workers [10]. An Alq3-based spin-tunnellingdevice fabricated via a nanoindentation technique featured a cross section ofonly a few tens of square nanometres. The device performance of the resultingstructures was impressive, showing a MR reaching 300 per cent at 2K—a newrecord for organic devices and one that compares well with the best inorganictunnel junctions.

The experiments reported above clearly show the possibility of achieving spin-polarized tunnelling across organic or organic–inorganic hybrid barriers and ofreaching significant MR values [23], in principle, competitive with inorganicspintronic field. Moreover, recently, a 600 per cent MR was reported in atunnelling device with a metal–phthalocyanine barrier (P. Seneor, R. Mattana& E. Beaurepaire 2010, private communication). In our opinion, the tunnellingresults briefly described here indicate that OSCs may well be considered asinteresting candidates for substituting highly performing inorganic tunnel barriersin some selected cases, such as flexible sheet applications.

4. Spin scattering in organic semiconductors

Spin-relaxation times in OSCs are extremely long—about 10−6 to 10−3 s [12].This is because of the weakness of the two main spin-scattering processes thatwork in OSCs, namely the SOC and the hyperfine interaction. Their role in themagnetotransport in OSCs spintronic devices has been studied at the conceptuallevel, but a clear quantitative link to the properties of any given OSCs materialhas yet to be established. The hyperfine scattering was proposed to act via theprecession of the spins of the injected (polarized) carriers that by hopping frommolecule to molecule encounter a randomly oriented mean nuclear spin field owingto hydrogen bound to the carbon rings [24]. It is worth noting that the hyperfinefield, even if it averages to zero, is inherently not negligible because of the localizednature of electron states in OSCs, which reduces the number of nuclear spinssampled by the carrier. The precession is actually a continuous effect, and theloss of spin polarization depends mainly on the average value of the hyperfine fieldand residential time of the carrier on a given molecule. The former can be definedfrom various resonance techniques and is of the order of millitesla [25], whereasthe latter parameter is known for very few materials [26] and is expected to coverthe range of 10−12 to 10−9 s or even shorter times [27]. This means that onaverage the charge carrier (using the gyromagnetic ratio of the free electron, i.e.28 GHzT−1) precesses for 10−2 to 10−5 of a full rotation while on the molecule,which causes little spin scattering.

The dynamics via which SOC can modify the spin state of an electron aredifferent. For significant coupling strength the spin of the carrier is no longera good quantum number and should be described by a linear combination ofspin-up and spin-down eigenstates. SOC then defines the spin-flip probability fororbital momentum scattering events. In inorganic semiconductors, the main spinscattering mechanisms (Elliott–Yafet, Dyakonov–Perel, Bir–Aronov–Pikus) arerelated to SOC and a clear qualitative and quantitative understanding of theseeffects has been achieved.





Although a straightforward transfer of this knowledge to the OSCs caseis difficult given their completely different behaviour as far as transport isconcerned (band conductivity versus random incoherent hopping), the firsttentative analyses of MR data on the basis of SOC scattering have been recentlyproposed [28,29].

The SOC in OSCs is considerably enhanced in metal–organic moleculesproviding, for example, zero-field splitting (ZFS) values of about 25 meV(200 cm−1) in Os(bpy)3 and other materials that contain heavy metals [30]. Onthe other hand, organic molecules without metals are characterized be very lowZFS of about 0.1 cm−1. Nevertheless, it is still unclear whether SOC is causingany significant spin scattering even in the widely used metal–quinolines. Tosummarize, the scattering strength of both SOC and hyperfine field-scatteringeffects has yet to be identified for different organic materials, and the dominatingmechanism in any material has yet to be established. This is a strong argumentfor a quantitative description of spin transport in OSCs.

In addition to this, interactions of spins with paramagnetic impurities or thedisturbing effect of trapping charges could give important hints into the physicsof spin propagation in OSCs.

5. Multi-functionality in organic semiconductor spin devices

One of the most interesting aspects of organic spintronics is its inherent multi-functionality. This term refers to the ability of a device or material to respond totwo or more independent stimuli by giving an output or by acquiring a given state.For example, OSCs are known to respond to electrical stimuli in the sense definedabove, which gives rise to resistive memory effects [31]. In spintronic devices,the electrodes respond to the application of a magnetic field by orienting theirmagnetizations. Both effects can be put to use in an organic spintronic device [32].

The coexistence of two independent phenomena (namely electrical andmagnetic hysteresis) in the same device or material is very intriguing, but it couldsimply emerge from the juxtaposition of two utterly disconnected systems. At thispoint, one might ask whether and how ‘cross talk’ between the two can exist, andthis is precisely what we observed in our devices [33]. In our vertical SVs, namelyLSMO/Alq3/AlOx/Co, the MR is controlled by the resistive state of the device,as shown in figure 2: we can turn off (inset (a)) and on (inset (b)) the SV effect byputting the device in the high- and low-resistance state, respectively. This doesnot happen in a trivial way. For example, the effect cannot be explained by a sim-ple change in the background resistance. The effect is bipolar: the SV effect can bedestroyed by the application of sufficiently large negative voltage, and it can be re-stored by the application of a sufficiently large positive voltage. The MR is mea-sured at the standard bias voltage of −100 mV [15], i.e. by injecting electrons fromthe LSMO (Co electrode is grounded). Interestingly, we can also smoothly varythe MR amount by stopping somewhere in the negative differential resistance reg-ion (no. 5 in figure 2) and putting the device into an intermediate resistive state.Now the SV effect is only reduced (inset (c)). The same applies for the positivebranch of the I–V curve, where we can also smoothly turn on the giant magneto-resistance (GMR) effect. Indeed, we found that the pre-application of increasinglyhigher biases correspond to an increasing recovery of the initial GMR effect.





–3 3–2 2–1 10

0.75

0.80

0.85

0.95

0.90

1.00

–1.5 –1.0 1.51.0–0.5 0.50–10

–5

5

0

10

15

20cu

rren

t (m

A)

bias (V)

1

2

3

4

5

6

(a)

(b)

(c)

field (k Oe)

–3 3–2 2–1 10

field (k Oe)

–3 3–2 2–1 10

field (k Oe)

R/R

max

(ar

b. u

nit)

0.75

0.80

0.85

0.95

0.90

1.00

R/R

max

(ar

b. u

nit)

0.75

0.80

0.85

0.95

0.90

1.00

R/R

max

(ar

b. u

nit)

Figure 2. An archetypal bistable I–V curve. Numbers identify different regions, as explained in thetext. The insets show the MR curves of the device after being pre-biased at the voltages indicatedby the arrows.

A number of physical mechanisms have been proposed to explain the electricalswitching in organic-based devices [31]. Based on a previous model by Rozenberget al. [34], we developed a phenomenological model that includes the MR effect.This model relies on an occupation-dependent conductance of different domainsinside the device. Carriers have to go through such domains to cross the device.In SVs, such domains can be represented by charge-trapping domains, localizednear the LSMO–Alq3 interface (owing to the threshold asymmetries). Such trapscan be responsible for the reduction in the hopping sites available for transport(in this way, they are changing the conductance of such domains), leading toa reduction in the charge mobility which would be reflected by an incrementof resistance and a loss of spin-polarized current, which is exactly what we areobserving in our multi-functional devices.

Multi-functionality can be fundamental for the development of futureintegrated memory–logic devices. In particular, the electrical control of the MRis considered to be of fundamental importance. In the International TechnologyRoadmap for Semiconductors (2009), the importance of the spin–metal oxidesemiconductor field effect transistor was pointed out; this is a three-terminaldevice in which the gate can be used to turn on and off the SV effect between thedrain and the source electrodes. In our devices, we achieve the same goal with atwo-terminal device. For this reason, they are of great interest in the developmentof future electronics.





OSC

s

(a) (b)

FM botto

mco

ntac

t

FM top

cont

act

OSC

s

Figure 3. Schematic drawing of an inorganic–organic interface. The deposition of OSCs onto theFM electrode gives rise to a well-defined interface (a), whereas the deposition of a FM layer ontoOSCs is associated with interdiffusion of the metal particles and interaction of metal atoms withthe organic molecules (b). (Online version in colour.)

6. Interfaces in organic spintronic devices

A common feature of all spintronic devices (both tunnelling and injection)involving OSCs is the hybrid organic–inorganic (HOI) interface sustaining aspin-transfer effect. Similar to inorganic spintronic devices, the interface qualityis crucial, and this quality demands considerable efforts and indeed considerableinvestments. As an additional remark, in vertical devices, the FM top contactis deposited on the soft organic layer and it is still unclear to what extentthis results in well-defined contact interfaces [35,36]. The latter is of greatimportance for reproducible device performance (figure 3). But, on the otherhand, hybrid interfaces seem to possess intriguing and rich properties of stillnot fully acknowledged importance for the whole field of spintronics.

Initially, the investigation of these interfaces pursued a kind of passive approachconcentrating mainly on the quantification of the electronic barriers (interfaceenergetics) while interfacing OSCs with new materials never used previouslyin OLED or OPV applications (manganite [37]; Co [38]). Further and moreadvanced investigations, involving the magnetic properties of HOI interfacestogether with a few recent MR results and theoretical models have unexpectedlyrevealed conceptually new properties at such interfaces, mainly related to aconsiderable spin-selection ability—an ability strongly amplified by its tuningpossibility. Thus, a strongly spin-polarized state was found to be inducedinside the HOMO–LUMO gap at the Fe/Alq3 interface supposed to promotea nearly 100 per cent filtering and leaving open only the channel of minorityiron spins [39]. Spin-polarized scanning tunnelling microscopy characterizationshave shown that only one phthalocyanine molecule, H2Pc, placed on the Fesurface is able to invert the metal spin polarization [40]. Along with this, thedetected inversion of the MR between the LSMO/Alq3/Co tunnel (positiveSV–MR) and injection (negative SV–MR) devices promoted the developmentof a more general model accounting for the spin-selective injection at theHOI interface on the basis of spin-dependent broadening of the conductingelectronic levels of the first monolayer of OSCs in contact with a magnetic





electrode [10,41]. This rapidly developing field of organic spintronics gainedsignificant attention and was recently named ‘spinterface’ [12]. It is the authors’opinion that these unusual (interface) properties place, for the first time,organic spintronics in a position that is truly competitive with that of inorganicspintronics [10,41].

7. Outstanding issues and future perspectives

We have presented above some important achievements in the young field oforganic spintronics, concentrating mainly on the electrical operation of variousdevices. Along with these achievements, the field continues to face importantunsolved issues of both theoretical and experimental aspects. Let us discuss themost pressing problems and propose a tentative analysis and possible solutions.

There is no explicit model for the (electrical) spin injection in OSCs. Theconductivity mismatch considerations [42] seem not to be easily applicable to thehybrid interfaces that OSCs form with inorganic injectors. The main reason forthis is the very low density of intrinsic carriers in organic materials: in OSCs,the spin carriers are provided by the extra charges injected from the electrodes,with the organic material acting as a capacitor–resistance system rather than asa simple resistor. For this reason, it is also difficult to detect the effects of spinaccumulation [43], which are known in inorganic materials. Spin accumulation inOSCs must be accompanied by charge accumulation, the latter offering strongrepulsion forces. To conclude on injection issues, there seems to be no specialrestriction on spin injection into OSCs, yet there is also no clear evidence thatthe injection proceeds without the loss of spin polarization.

Some mechanisms were proposed recently to account for the spin scatteringinside OSCs: the hyperfine interaction [44] and SOC [28] have been convincinglybrought into the centre of discussions as the main candidates. Nonetheless,considerable additional experimental and theoretical efforts are required toprovide a clear quantitative definition of the parameters describing the strength ofthese two mechanisms in various OSCs. Surprisingly, the role of phonons for spintransport in OSCs remains absolutely unclear despite the strong role of polaronsin these materials.

In organic spintronics, one of the most soundly established results is that GMRpeaks at low biases (e.g. 1 mV) and decreases as the magnitude of the voltage isincreased; the GMR vanishes at about 1V. A naive model based on drift diffusionfails to explain this fact, because an increase in drift should cause an increase inthe MR, not a decrease. Less pressing, but still significant, is the fact that toaccount for the observed GMR, at 1 mV in a 100 nm thick organic layer witha mobility of 10−5 cm2 Vs−1, one would need a spin lifetime of 10 ms. Yet, nocomprehensive model has been proposed that accounts for these facts.

One of the hallmarks of spin transport in inorganic semiconductors is thepresence of the Hanle effect—which is due to the spin precession about an appliedmagnetic field [45]—which in most experiments is applied perpendicular to thespin. In an SV configuration, for high-mobility materials such as Si, the precessioncauses the MR to oscillate as a function of the applied magnetic field. The numberof oscillations within a given time increases linearly with the applied magneticfield. As the number of oscillations increases, they also decrease in intensity





because the spins progressively lose their phase coherence. With a sufficientlylarge number of oscillations, the spins are completely de-phased, that is, thespin polarization is completely lost. The number of oscillations can be madegreater by increasing the time the spins spend in the semiconductor (slow driftor slow diffusion). In this regime, even a small perpendicular magnetic field cancompletely depolarize the spins; if q is the angle between the spin and the appliedmagnetic field, then MR (q) = MR(0) × cos2 q [46].

Owing to their low mobility, OSCs are thought to work in this regime. So far,the experiments carried out by several groups could not detect any of the effectsdescribed above in organic devices. The reason for this is still unclear. On the onehand, there might be technological issues and better devices needed to observethe effect. We think that the most effective approach to this problem would bethe investigation of the Hanle effect in planar devices, which would avoid anyproblem arising from short circuits or poor definition of the transport length.

On the other hand, it must be remembered that transport in OSCs isfundamentally different from that of inorganic semiconductors; in addition tothis, spin-scattering mechanisms act in a different way. In this sense, it is possiblethat the Hanle effect cannot be observed for fundamental reasons.

In spite of the enormous body of knowledge existing on the optical propertiesof OSCs, parallel investigations on the MR and well-tuned photoexcitation ofcarriers in the conduction levels or trapped on the defects has not yet beenpursued. In principle, the answers that can be provided by such studies can givea detailed description of the (spin-polarized) carrier energy, nature, density, etc.,i.e. are able to provide perhaps a full quantitative and qualitative description ofthe system.

In order to link spintronics to the important organic applications such asOLEDs and OFETs, it is necessary to overcome some important limitations.Historically, the spin injection in OSCs was initially investigated for its possibleuse in the field of OLEDs [47], the latter being so far the most successfulapplication of OSCs. Light emission from OLEDs results from a radiative decayof a Frenkel exciton, a molecular excited state formed by an electron and a holeinjected from the cathode and anode, respectively [48]. Each exciton follows theFermi statistics and then four spin states are possible: three ‘triplet’ states of totalspin 1 and one ‘singlet’ state of total spin zero. In general, the radiative decay thatleads to fluorescence is allowed only for singlet excitons (owing to vanishingly lowSOC): this naturally limits the OLED efficiency, as not more than 25 per centof the injected carriers are involved in the light emission. The control of the spinpolarization of the injected carriers would thus result in a serious improvementin the singlet-based OLED efficiency and, on the other hand, could provide afascinating mechanism for switching from singlet to triplet device operation [49].It is worth mentioning that efficient triplet emission can be activated by usingOSCs-containing heavy metal ions [50].

Many attempts have been made to manipulate the light output from OLEDsusing spin-polarized currents by substituting conventional electrodes with FMones, but no clear effects have so far been reported. This failure has a simpleexplanation: the available data for spintronic effects in organic materials confirmso far spin injection and transport only for voltages not exceeding approximately1 V, while light emission by OLEDs is usually detected at voltages greaterthan 1 V. Expanding this working interval to a few volts thus represents a





challenging goal. This limitation holds also for the application of spin-polarizedcurrents to OFET devices, whose operating voltages also generally exceed the fewvolts level.

As a final remark, we would like to point out the immense potential of OSCsfor the spintronics field. An almost infinite list of materials, easily modifiableinterfaces and cheap technologies for the growth of films on rigid and flexiblesubstrates make us look optimistically towards the next steps of this young field.

We would like to thank L. E. Hueso for helpful discussions.

References

1 Dediu, V. A., Hueso, L. E., Bergenti, I. & Taliani, C. 2009 Spin routes in organic semiconductors.Nat. Mater. 8, 850. (doi:10.1038/nmat2544)

2 McClure, D. S. 1952 Spin–orbit interaction in aromatic molecules. J. Chem. Phys. 20, 682–686.(doi:10.1063/1.1700516)

3 Günther, H. 1995 NMR spectroscopy: basic principles, concepts, and applications in chemistry,2nd edn. New York, NY: John Wiley & Sons.

4 Dediu, V. A., Murgia, M., Matacotta, F. C., Taliani, C. & Barbanera, S. 2002 Roomtemperature spin polarized injection in organic semiconductor. Solid State Commun. 122,181–184. (doi:10.1016/S0038-1098(02)00090-X)

5 Xiong, Z. H., Wu, D., Vardeny, Z. V. & Shi, J. 2004 Giant magnetoresistance in organic spin-valves. Nature 427, 821–824. (doi:10.1038/nature02325)

6 Drew, A. J. et al. 2009 Direct measurement of the electronic spin diffusion length in a fullyfunctional organic spin valve by low-energy muon spin rotation. Nat. Mater. 8, 109–114.(doi:10.1038/nmat2333)

7 Cinchetti, M., Heimer, K., Wustenberg, J. P., Andreyev, O., Bauer, M., Lach, S., Ziegler,C., Gao, Y. & Aeschlimann, M. 2009 Determination of spin injection and transport in aferromagnet/organic semiconductor heterojunction by two-photon photoemission. Nat. Mater.8, 115–119. (doi:10.1038/nmat2334)

8 Santos, T. S., Lee, J. S., Migdal, P., Lekshmi, I. C., Satpati, B. & Moodera, J. S. 2007Room-temperature tunnel magnetoresistance and spin-polarized tunneling through an organicsemiconductor barrier. Phys. Rev. Lett. 98, 016601. (doi:10.1103/PhysRevLett.98.016601)

9 Szulczewski, G., Tokuc, H., Oguz, K. & Coey, J. M. D. 2009 Magnetoresistance in magnetictunnel junctions with an organic barrier and an MgO spin filter. Appl. Phys. Lett. 95, 202506.(doi:10.1063/1.3264968)

10 Barraud, C. et al. 2010 Unravelling the role of the interface for spin injection into organicsemiconductors. Nat. Phys. 6, 615–620. (doi:10.1038/nphys1688)

11 Nguyen, T. P. 2006 Defect analysis in organic semiconductors. Mater. Sci. Semicond. Process.9, 198–203. (doi:10.1016/j.mssp.2006.01.060)

12 Szulczewski, G., Sanvito, S. & Coey, M. 2009 A spin of their own. Nat. Mater. 8, 693–695.(doi:10.1038/nmat2518)

13 Majumdar, S., Majumdar, H. S., Laiho, R. & Osterbacka, R. 2006 Comparing small moleculesand polymer for future organic spin-valves. J. Alloys Compd. 423, 169–171. (doi:10.1016/j.jallcom.2005.12.104)

14 Xu, W., Szulczewski, G. J., LeClair, P., Navarrete, I., Schad, R., Miao, G., Guo, H. & Gupta,A. 2007 Tunneling magnetoresistance observed in La0.67Sr0.33MnO3/organic molecule/Cojunctions. Appl. Phys. Lett. 90, 072506. (doi:10.1063/1.2435907)

15 Dediu, V. et al. 2008 Room-temperature spintronic effect in Alq3-based hybrid devices. Phys.Rev. B 78, 115203. (doi:10.1103/PhysRevB.78.115203)

16 Wang, F. J., Yang, C. G., Vardeny, Z. V. & Li, X. G. 2007 Spin response inorganic spin valves based on La2/3Sr1/3MnO3 electrodes. Phys. Rev. B 75, 245324.(doi:10.1103/PhysRevB.75.245324)



http://dx.doi.org/doi:10.1038/nmat2544

http://dx.doi.org/doi:10.1063/1.1700516

http://dx.doi.org/doi:10.1016/S0038-1098(02)00090-X

http://dx.doi.org/doi:10.1038/nature02325



http://dx.doi.org/doi:10.1103/PhysRevLett.98.016601


http://dx.doi.org/doi:10.1038/nphys1688

http://dx.doi.org/doi:10.1016/j.mssp.2006.01.060


http://dx.doi.org/doi:10.1016/j.jallcom.2005.12.104

http://dx.doi.org/doi:10.1016/j.jallcom.2005.12.104


http://dx.doi.org/doi:10.1103/PhysRevB.78.115203




17 Majumdar, S., Laiho, R., Laukkanen, P., Vayrynen, I. J., Majumdar, H. S. & Osterbacka, R.2006 Application of regioregular polythiophene in spintronic devices: effect of interface. Appl.Phys. Lett. 89, 122114. (doi:10.1063/1.2356463)

18 Morley, N. A., Rao, A., Dhandapani, D., Gibbs, M. R. J., Grell, M. & Richardson, T. 2008Room temperature organic spintronics. J. Appl. Phys. 103, 07F306. (doi:10.1063/1.2829245)

19 Petta, J. R., Slater, S. K. & Ralph, D. C. 2004 Spin-dependent transport in molecular tunneljunctions. Phys. Rev. Lett. 93, 136601. (doi:10.1103/PhysRevLett.93.136601)

20 Ando, Y., Murai, J., Miyashita, T. & Miyazaki, T. 1998 Spin dependent tunneling in 80NiFe/LBfilm with ferrocene and tris(bipyridine)ruthenium derivatives Co junctions. Thin Solid Films331, 158–164. (doi:10.1016/S0040-6090(98)00913-4)

21 Shim, J. H., Raman, K. V., Park, Y. J., Santos, T. S., Miao, G. X., Satpati, B. & Moodera, J.S. 2008 Large spin diffusion length in an amorphous organic semiconductor. Phys. Rev. Lett.100, 226603. (doi:10.1103/PhysRevLett.100.226603)

22 Tedrow, P. M. & Meservey, R. 1971 Spin-dependent tunneling into ferromagnetic nickel. Phys.Rev. Lett. 26, 192–195. (doi:10.1103/PhysRevLett.26.192)

23 Rocha, A. R., Garcia-Suarez, V. M., Bailey, S. W., Lambert, C. J., Ferrer, J. & Sanvito, S.2005 Towards molecular spintronics. Nat. Mater. 4, 335–339. (doi:10.1038/nmat1349)

24 Bobbert, P. A., Wagemans, W., Oost, F. W. A. v., Koopmans, B. & Wohlgenannt, M. 2009Theory for spin diffusion in disordered organic semiconductors. Phys. Rev. Lett. 102, 156604.

25 Sheng, Y., Nguyen, T. D., Veeraraghavan, G., Mermer, O., Wohlgenannt, M. & Scherf, U.2006 Hyperfine interaction and magnetoresistance in organic semiconductors. Phys. Rev. B 74,045213. (doi:10.1103/PhysRevB.74.045213)

26 Ding, B. F., Yao, Y., Sun, Z. Y., Wu, C. Q., Gao, X. D., Wang, Z. J., Ding, X. M., Choy,W. C. H. & Hou, X. Y. 2010 Magnetic field effects on the electroluminescence of organiclight emitting devices: a tool to indicate the carrier mobility. Appl. Phys. Lett. 97, 163302.(doi:10.1063/1.3505343)

27 Coropceanu, V., Cornil, J., da Silva, D. A., Olivier, Y., Silbey, R. & Bredas, J. L. 2007 Chargetransport in organic semiconductors. Chem. Rev. 107, 926–952. (doi:10.1021/cr050140x)

28 Bandyopadhyay, S. 2010 Dominant spin relaxation mechanism in compound organicsemiconductors. Phys. Rev. B 81, 153202. (doi:10.1103/PhysRevB.81.153202)

29 Yu, Z. G. 2010 Spin-orbit coupling, spin relaxation, and spin diffusion in organic solids. Phys.Rev. Lett. 106, 106602. (doi:10.1103/PhysRevLett.106.106602)

30 Yersin, H. & Strasser, J. 2000 Triplets in metal–organic compounds. Chemical tunability ofrelaxation dynamics. Coord. Chem. Rev. 208, 331–364. (doi:10.1016/S0010-8545(00)00326-X)

31 Scott, J. C. & Bozano, L. D. 2007 Nonvolatile memory elements based on organic materials.Adv. Mater. 19, 1452–1463. (doi:10.1002/adma.200602564)

32 Hueso, L. E., Bergenti, I., Riminucci, A., Zhan, Y. Q. & Dediu, V. 2007 Multipurpose magneticorganic hybrid devices. Adv. Mater. 19, 2639–2642. (doi:10.1002/adma.200602748)

33 Prezioso, M., Riminucci, A., Bergenti, I., Graziosi, P., Brunel, D. & Dediu, V. A. 2011Electrically programmable magnetoresistance in multifunctional organic based spin-valvedevices. Adv. Mater. 23, 1371–1375. (doi:10.1002/adma.201003974)

34 Rozenberg, M. J., Inoue, I. H. & Sanchez, M. J. 2004 Nonvolatile memory with multilevelswitching: a basic model. Phys. Rev. Lett. 92, 178302. (doi:10.1103/PhysRevLett.92.178302)

35 Borgatti, F. et al. 2010 Understanding the role of tunneling barriers in organic spin valves byhard x-ray photoelectron spectroscopy. Appl. Phys. Lett. 96, 043306. (doi:10.1063/1.3285179)

36 Bergenti, I., Riminucci, A., Arisi, E., Murgia, M., Cavallini, M., Solzi, M., Casoli, F. & Dediu,V. 2007 Magnetic properties of cobalt thin films deposited on soft organic layers. J. Magn.Magn. Mater. 316, E987–E989. (doi:10.1016/j.jmmm.2007.03.165)

37 Zhan, Y. Q., de Jong, M. P., Li, F. H., Dediu, V., Fahlman, M. & Salaneck, W. R. 2008 Energylevel alignment and chemical interaction at Alq(3)/Co interfaces for organic spintronic devices.Phys. Rev. B 78, 045208. (doi:10.1103/PhysRevB.78.045208)

38 Popinciuc, M., Jonkman, H. T. & van Wees, B. J. 2006 Energy level alignment symmetry atCo/pentacene/Co interfaces. J. Appl. Phys. 100, 093714. (doi:10.1063/1.2369651)

39 Zhan, Y. Q., Holmstrom, E., Lizarraga, R., Eriksson, O., Liu, X. J., Li, F. H., Carlegrim,E., Stafström, S. & Fahlman, M. 2010 Efficient spin injection through exchange coupling at






http://dx.doi.org/doi:10.1016/S0040-6090(98)00913-4






http://dx.doi.org/doi:10.1021/cr050140x



http://dx.doi.org/doi:10.1016/S0010-8545(00)00326-X

http://dx.doi.org/doi:10.1002/adma.200602564





http://dx.doi.org/doi:10.1016/j.jmmm.2007.03.165





organic semiconductor/ferromagnet heterojunctions. Adv. Mater. 22, 1626–1630. (doi:10.1002/adma.200903556)

40 Atodiresei, N., Brede, J., Lazicacute, P., Caciuc, V., Hoffmann, G., Wiesendanger, R. & Blügel,S. 2010 Design of the local spin polarization at the organic–ferromagnetic interface. Phys. Rev.Lett. 105, 066601. (doi:10.1103/PhysRevLett.105.066601)

41 Sanvito, S. 2010 Molecular spintronics: the rise of spinterface science. Nat. Phys 6, 562–564.(doi:10.1038/nphys1714)

42 Schmidt, G., Ferrand, D., Molenkamp, L. W., Filip, A. T. & van Wees, B. J. 2000 Fundamentalobstacle for electrical spin injection from a ferromagnetic metal into a diffusive semiconductor.Phys. Rev. B 62, R4790–R4793. (doi:10.1103/PhysRevB.62.R4790)

43 Fert, A. & Jaffres, H. 2001 Conditions for efficient spin injection from a ferromagnetic metalinto a semiconductor. Phys. Rev. B 64, 184420. (doi:10.1103/PhysRevB.64.184420)

44 Bobbert, P. A., Nguyen, T. D., Wagemans, W., van Oost, F. W. A., Koopmans, B. &Wohlgenannt, M. 2010 Spin relaxation and magnetoresistance in disordered organicsemiconductors. Synth. Met. 160, 223–229. (doi:10.1016/j.synthmet.2009.06.002)

45 Huang, B., Monsma, D. J. & Appelbaum, I. 2007 Coherent spin transport through a 350 micronthick silicon wafer. Phys. Rev. Lett. 99, 177209. (doi:10.1103/PhysRevLett.99.177209)

46 Huang, B., Jang, H. J. & Appelbaum, I. 2008 Geometric dephasing-limited Hanle effectin long-distance lateral silicon spin transport devices. Appl. Phys. Lett. 93, 162508.(doi:10.1063/1.3006333)

47 Pulizzi, F. 2009 Why going organic is good. Nat. Mater. 8, 691. (doi:10.1038/nmat2517)48 Friend, R. H. et al. 1999 Electroluminescence in conjugated polymers. Nature 397, 121–128.

(doi:10.1038/16393)49 Bergenti, I., Dediu, V., Murgia, M., Riminucci, A., Ruani, G. & Taliani, C. 2004 Transparent

manganite films as hole injectors for organic light emitting diodes. J. Lumin. 110, 384–388.(doi:10.1016/j.jlumin.2004.08.036)

50 Baldo, M. A., O’Brien, D. F., You, Y., Shoustikov, A., Sibley, S., Thompson, M. E. & Forrest,S. R. 1998 Highly efficient phosphorescent emission from organic electroluminescent devices.Nature 395, 151–154. (doi:10.1038/25954)






http://dx.doi.org/doi:10.1038/nphys1714

http://dx.doi.org/doi:10.1103/PhysRevB.62.R4790


http://dx.doi.org/doi:10.1016/j.synthmet.2009.06.002




http://dx.doi.org/doi:10.1038/16393

http://dx.doi.org/doi:10.1016/j.jlumin.2004.08.036

http://dx.doi.org/doi:10.1038/25954


review organic spintronicsrsta.royalsocietypublishing.org/content/roypta/369/1948/3054.full.pdf ·...

Documents