Sie sind auf Seite 1von 13

www.advmat.

de

PROGRESS REPORT

The Materials Science of Functional Oxide Thin Films


By Mark G. Blamire,* Judith L. MacManus-Driscoll, Neil D. Mathur, and Zoe H. Barber This article is dedicated to Jan Evetts, who founded the Device Materials Group

Research in the area of functional oxides has progressed from study of their basic chemistry and structure to the point at which an enormous range of desirable properties are being explored for potential applications. The primary limitation on exploitation is the difculty of achieving sufciently precise control of the properties because of the range of possible defects in such materials and the remarkably strong effect of such defects on the properties. This review outlines the reasons underlying this sensitivity and recent results that demonstrate the levels of control which are now possible.

1. Introduction
1.1. Functional Oxides at Cambridge University This paper is published as part of this special issue to mark the 800th Anniversary of the founding of the University of Cambridge. The study of functional oxides is a worldwide activity, but this paper highlights work within Cambridge and our own Device Materials Group in the Department of Materials Science. Functional oxides have been actively researched in Cambridge over the last century. William Bragg undertook the formative work on determining crystal structures, starting in 1911 as a PhD student in Cambridge. Some of the early minerals he studied were oxides, such as calcite (CaCO3) and spinel (MgAl2O4). In the period from the late 1920s to late 50s, work continued on the physics of crystals, including many oxides, with W. A. Wooster. Two classic texts emerged from his time, A Textbook of Crystal Physics[1] and Experimental Crystal Physics,[2] and later Diffuse X-Ray Reections from Crystals.[3] During the Second World War, together with his wife, Wooster set up a private lab in Cambridge. Here, they undertook the rst hydrothermal synthesis of quartz in Britain. In the mid 1940s, a very prominent lady, Helen Megaw, came to the fore. While at Philips Laboratories, she performed the rst
[*] Prof. M. G. Blamire, Prof. J. L. MacManus-Driscoll, Dr. N. D. Mathur, Z. H. Barber Department of Materials Science and Metallurgy University of Cambridge Pembroke Street, Cambridge CB2 3QZ (UK) E-mail: mb52@cam.ac.uk

out for you on a choose!. During the 1970s, in Cambridge, Glazer himself acheived major contributions to the understanding of crystallography of perovskites. With Megaw, he determined the structures and phase transitions of perovskites, such as NaNbO3,[7,8] and then went on to investigate the now widely applied piezoelectric PbZrxTi1 xO3 (PZT).[9] He also discovered the 23 types of tilted octahedra in perovskites, which are critically important in understanding functional properties, most notably ferroelectric and magnetic properties. He developed a notation that has now been widely and internationally adopted.[10] Glazers group was also one of the rst in the world to grow single crystals of PZT.[11] During the 1980s and beyond, with the realization of the enormous potential of oxides for applications, as well as continuing studies on basic physics and crystallography, research has expanded rapidly into the area of oxide device materials, following on from growth of the worlds rst high-temperature superconductor (HTS) thin lm in 1987 by workers in our research group.[12]

determination of the crystal structure of BaTiO3[4] and, having joined Cambridge University, wrote the denitive Ferroelectricity in Crystals.[5] As noted by her last student, Michael Glazer, this was the rst book on the subject, was for a long time just about the only book on the subject, and became almost a bible for those working in ferroelectricity.[6] He also commented that she was the only person I have met who could imagine a crystal structure in her mind and turn it around, and then draw it piece of paper from any direction you cared to

1.2. Functional Oxides for Devices The past 20 years have seen an explosion of interest in the electronic and magnetic properties of functional oxides. The wide range of physical properties that can appear in oxides, such as those based on the perovskite unit cell, arises to a large extent because of the extreme sensitivity of electron interactions to apparently minor structural changes, such as bond angles and lengths. This fact alone demonstrates that a free electron description of the properties is unlikely to be sufcient even for the highly conducting oxides, and in general these materials

DOI: 10.1002/adma.200900947

Adv. Mater. 2009, 21, 38273839

2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

3827

www.advmat.de

are classied as having strongly correlated electronic and magnetic states. Stemming from the discovery of hightemperature superconductivity in perovskite-related crystal structures[13] such as YBa2Cu3O7 x (YBCO), the study of complex oxides became one of the most active elds in materials science and condensed-matter physics. Indeed, there has been an interesting reversion to the study of multiferroic and other phenomena in apparently simple ABO3 perovskites, which are still not understood at a fundamental level. The application of functional oxides in thin-lm form, that is, beyond their use in bulk form, is still rather limited. However, the use of transparent indium tin oxide (ITO) electrodes in displays and solar cells, ferroelectric (Pb,Zr)TiO3 (PZT) lms in ferroelectric random-access memory (FeRAM), binary oxide dielectrics in mobile communications and gate dielectrics, and superconducting YBCO in microwave lters are notable exceptions. The most commonly studied oxide structures are perovskites, but rutile, pyrochlore, and anatase structures are also of interest for the broad range of functionalities achievable. The main current limitation to device applications of complex oxides is the control of both defects and composition, which enormously impact properties. Both interface and bulk properties must be carefully tuned, since, in many cases, the surface property is the critical one, fpr example, in ZnO varistors, BaTiO3 thermistors, and SnO2 CO gas sensors, where the surface space charge is modied by the electron donating adsorbed gaseous species. The basic science in relation to defects in oxides has been intensively studied for many decades but has not been applied widely to materials growth for devices. The general area connecting fundamental science to device applications presents the interdisciplinary eld of materials science with enormous opportunities and challenges, and has been a strong contributing factor in the setting up of a large number of materials-research centers across the world in the last quarter century. The purpose of this Progress Report is to summarise recent developments in the materials science of oxide thin lms, with special reference to control of defects at interfaces. The structure of the paper is as follows: we will provide an introduction to the reasons underlying the importance of defects in oxides, will then discuss the emergent classes of oxides that have driven forward the eld over the past decade, and then, nally, review what we believe are the outstanding issues to be addressed by the materials-science community.

PROGRESS REPORT

The authors of the paper work in the Device Materials Group (DMG) in the Materials Science Department of Cambridge University. DMG has at its focus the study and development of novel functional materials with particular emphasis on thin lm heterostructures of oxides and metals. The group was established by the late Prof. Jan Evetts in the 1960s and now consists of over forty research staff. apparently simple SrTiO3.[14,15] Indeed, the strongly localized properties of oxides mean that surfaces and interfaces (even within structurally perfect materials) have properties that can differ strongly from the bulk. It is possible to model the effect of particular defects within a single crystal with great accuracy, and this is becoming feasible for interfaces as well.[16] In practice, however, multiple defects of unknown type are present in oxides, however ideal, and so in many cases a pragmatic, and to some extent empirical, approach is all that can be applied. The role of materials science is ultimately to transform the exciting properties measured in the laboratory into systems that can be applied. However, far more than has been the case with metallic and semiconducting materials, understanding and controlling the defects in complex oxides is essential to unraveling the basic science underlying the properties.

2.1. The Control of Defect Concentrations Point-defect formation energies in oxides are wide-ranging. Typically, the oxides with higher melting temperatures have higher formation energies. ZnO is a notable exception, with an energy of oxygen vacancy formation as low as 2.5 eV. On the other hand, Al2O3 and spinels have defect-formation energies up to 30 eV.[17] Hence, based on defect thermodynamic arguments alone, for the same growth temperature, ZnO lms will have signicantly higher point defect concentrations than Al2O3 lms. Compared to the conventional semiconductor Si, for which defect formation energies are around 24 eV,[18] binary-oxide defect-formation energies are typically 510 eV, and so for the same growth temperature there are proportionately more defects in the Si case. However, annealing protocols are well advanced for semiconductors, hence by careful annealing below the growth temperature, defects recombine because the entropy contribution promoting their stability diminishes with decreasing temperature.[19] As well as the inuence of thermal excitation on defect formation, oxidation/reduction processes of oxide lms render them susceptible to point defects in the form of metal/oxygen stoichiometry ratio changes. Control of oxygen during lm growth rather than after growth (by post-annealing) is the most effective way of controlling the cation/anion ratio. Since kinetics tend to be insufcient to allow long-range cation mobility, post-annealing is most suitable for adjusting only oxygen stoichiometry. Many perovskites exhibit oxygen nonstoichiometry and, since some functional properties depend strongly on metaloxygen bond lengths,[20] interfacial strain presents additional problems in oxide devices over conventional semiconductor devices for which strain is detrimental, primarily because it produces dislocations and associated recombination centers.[21]

2. Control of Function in Oxides


If subtle structural changes within single crystals can dramatically affect the properties, it is obvious that the situation in real materials, which may be strained and defective, can be exceptionally complex. By way of comparison, so-called band-gap engineering in conventional semiconductors, which exploits the lattice-mismatch-induced strain that arises during heteroepitaxial growth, can be modeled (at least to rst order) in terms of the distortion of the Brillouin zone and the repositioning of the Fermi surface. Both of these directly affect the carrier density, mobility, and effective mass, but do not transform the semiconductor into a superconductor or a ferroelectric, although both of these phenomena are possible, for example in the

3828

2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Mater. 2009, 21, 38273839

www.advmat.de

PROGRESS REPORT

Figure 1. Oxygen stoichiometry, 3 d, versus logarithm of oxygen partial pressure for La0.7Sr0.3MnO3 d, at different temperatures. Adapted with permission from [24]. Copyright 2000 Elsevier.

ux,[27] thereby allowing high currents to be carried; and in DMS, they are necessary to provide charge carriers and to mediate ferromagnetism.[2830] In ferroelectrics, oxygen defects pin domain walls and, in the presence of electric elds, migrate to interfaces, leading to fatigue.[31] In HTS, whilst grain boundaries strongly depress critical currents,[32,33] dislocation cores act as very effective pinning centers,[34] so that thin lms generally show much higher critical current densities than single crystals. In thin-lm HTS devices, where high current-carrying capability is not the paramount requirement, but where atomic-scale interfacial properties are key, variability of oxygen at strained interfaces leads to irreproducible junction properties.[35] In oxides for optoelectronic applications, oxygen defect states within the bandgap lead to lower wavelength emission. This has been studied widely in ZnO, where, in addition to ultraviolet emission at the 3.37 eV bandgap, green emission results from shallow oxygen defect states.[36]

2.3. Interfaces, Strain, and Defects The best way to understand both the complexity and propensity of equilibrium defect formation in oxides is via Brouwer diagrams, which plot defect concentrations of the various species versus partial pressure of the species of interest.[22] For example, LaMnO3 is widely assumed to be stoichiometric but in fact, while it is relatively oxygen-stoichiometric, the Brouwer diagram for the compound reveals that it can withstand high cation nonstoichiometry in both La and Mn.[23] Figure 1 shows that for La0.7Sr0.3MnO3 d, to achieve oxygen nonstoichiometry, oxygen partial pressures of less than 107 atm (1 atm 101 325 Pa)are required at 1000 8C, and less than 1010 atm at 600 8C. On the other hand, oxygen stoichiometries of 3 d, where d is positive (which, in effect, means the oxygen is stoichiometric but La and Mn vacancies are present) occurs very readily at all temperatures for oxygen partial pressures greater than 107 atm.[24] Holes are associated with the cation vacancies but, since the use of manganites in spintronic devices relies on them being conducting, such defects are not necessarily detrimental to performance. However, in oxide ferroelectrics such as Ti-based perovskites and BiFeO3, they are critical, since leakage currents need to be minimized. In addition to controlling equilibrium defects by adjusting growth temperature and post-annealing treatments, the growth kinetics and growth mode both play important roles in xing the concentration of nonequilibrium defects. In contrast to high-rate, high-supersaturation growth of oxides by standard pulsed laser deposition (PLD), which results in island growth and also rather defective materials, low-rate growth has been demonstrated using reective high-energy electron diffraction (RHEED) to produce a layer-by-layer morphology, rather perfect lms, and very sharp interfaces.[25,26] Structural distortion through ion substitutions and interfacial strain through heteroepitaxy and lattice mismatch in thin-lm systems provide powerful means of materials-property control; for example, it has been demonstrated that imposition of a biaxial strain in a manganite thin lm through lattice mismatch with a substrate can lead to similar changes in magnetic and transport properties as caused by cation substitution in the bulk.[37] The tremendous sensitivity of the properties of oxides to defects and strain is greatly amplied at interfaces. This is crucial for applications: the effects underlying semiconductor electronics are all interfacial, and include the pn junction, two-dimensional electron gases (2-DEGs), and the electric-eld effect. In conventional semiconductors, the realization of applications has been assisted by the robustness of the properties and the capability for growing atomically perfect heterostructures. In oxides, surfaces and interfaces have an even greater tendency than the bulk to develop nonstoichiometry, strain, or structural defects, with a correspondingly greater modication of properties. Strain control of functional properties at interfaces is probably best exemplied in oxide ferroelectrics. In both SrTiO3 and BaTiO3[15,38] thin lms of thickness below around 50 nm, huge changes in the ferroelectric TC, remanent polarization, and structural phase-transition temperature occur through use of substrate-controlled, coherent in-plane biaxial strain. Specialized substrates such as GdScO3, and (LaAlO3)0.29 (SrAl0.5Ta0.5O3)0.71 (LSAT) are used in order to tune the strain level and corresponding property-enhancement level. Figure 2 demonstrates that the TC of BaTiO3 can be enhanced to nearly 700 8C. Multilayering is another route to control strain and properties: in strained BaTiO3/SrTiO3/CaTiO3 multilayers just a few unit-cells thick, enhanced polarization is induced in the BaTiO3 layer.[39] Emerging oxide electronic applications, where defects play a strong role because of the necessity to create an atomically perfect interface between two different oxides, include transparent eld effect transistors (FETs)[40] for which ZnO is a promising system.[41] Junction diodes, homoepitaxial pn junctions formed of the same parent compound, are ideal, since they are nominally free from strain and hence less prone to defects. However, there

2.2. Benecial Defects Whilst in most functional oxide systems defects are detrimental to functionality, two example systems where they are benecial are HTSs and dilute magnetic semiconductors (DMSs): in HTS conductors, defects are necessary to provide pinning of magnetic

Adv. Mater. 2009, 21, 38273839

2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

3829

www.advmat.de

PROGRESS REPORT

1200 1000 Temperature (K) 800 600 400 200 -1.5 -1 -0.5 0 0.5 in-plane strain (%) Paraelectric 1 1.5 Paraelectric

Transition

Figure 2. Expected ferroelectric phase diagram of (001) BaTiO3 under biaxial in-plane strain. Adapted with permission from [38]. Copyright 2004 American Association for the Advancement of Science. The data points show the observed properties of strained lms grown on different substrates.

respectively.[47,48] Whilst this is a powerful method for patterning surfaces through selective photochemical growth, there are many other potential applications for the formation of highly oxidizing or reducing surfaces. Another area for oxide lms that has huge but as yet unrealized potential is in magnetism, notably in magnetic data storage and possibly for the next-generation spin FETs, as discussed in Section 4.2.1. There are numerous magnetic oxides with above-room-temperature ferromagnetism,[49] but control of surfaces and interface defects are again of paramount importance. Near-perfect interfaces are required in oxides because the localized character of the electrons means that it is easier to weaken ferromagnetic exchange and transport properties. If a strain or chemically induced interfacial barrier layer of sub-nanometer dimensions is present at the interface, as occurs for example in the structurally and chemically dissimilar LSMO/ Fe3O4 bilayers, there is magnetic decoupling.[50] In fact, the true origin of such interfacial barriers in magnetic structures at the (sub)-nanometer is not well understood. Sophisticated synchrotron-spectroscopic techniques are required to fully characterize the atomic-scale interface defects.

are no oxide systems that can be readily doped to give both n- and p-type behavior with high carrier concentration and high mobility. Again, ZnO is currently the best option for a homogeneous system. However, high-performance p-type ZnO is still somewhat a holy grail, because when ZnO is acceptor-doped, n-type defects (most notably oxygen vacancies) form, leading to selfcompensation. Hence, special tricks need to be played to kinetically subdue the n-type defects.[42] Unfortunately, reduced kinetics also means poorer crystallinity and reduced hole mobility. The atomic-layer deposition (ALD, see Section 3.2.3) chemical growth technique holds the most promise for creating high crystallinity and low defect densities[43,44] and will be a method to watch for creation of complex oxides in the next decade. Particularly topical are interfacial effects in oxide 2-DEGs, ZnO/ZnMgxO1 x[45] and STO/LAO; in the former, exceptionally high mobilities and the rst observation of the quantum Hall effect represent a genuine breakthrough in the application of oxides to high-performance electronics. The underlying issues are discussed in Section 4.2.2, but it is clear that in both cases the interfaces must be highly perfect for the formation of a conducting 2-DEG layer. In the STO/LAO system, controversy over the understanding of the nature of the charge-transfer effect at the interface has been somewhat clouded by the uncontrolled contribution of charge carriers from oxygen vacancies.[14] This still remains an important area for detailed investigation. Most ferroelectric applications rely upon the properties of surfaces and interfaces. Although there has been much research, including the application of very-high-resolution surface techniques, such as scanning surface potential microscopy, electrostatic force microscopy, and piezoresponse force microscopy, many questions remain.[46] It has been demonstrated that surface photoexcited reactions on ferroelectrics can be dened by a pattern of poled domains. When a solution of metal salt is placed over the surface and irradiated, faces oriented in opposite directions, or , may show reductive or oxidative behavior,

3. Thin-Film Deposition
Thin-lm deposition lies at the heart of modern devicefabrication processes. For semiconductor materials, the intolerance to defects has led to the gradual renement of molecular beam epitaxy (MBE) and chemical vapor deposition (CVD), which can produce lms of exceptional purity and structural perfection. Oxides, because of their greater structural complexity, lower surface mobilities, and propensity for defect formation, represent a signicantly increased challenge. However, signicant investment in the development of new growth and characterization techniques has enabled high-quality oxide growth to become feasible at the research scale.

3.1. The Requirement for Epitaxy Epitaxial growth implies a registration between the substrate and lm crystal structures, such that as well as leading to a single out-of-plane growth orientation, there is also an in-plane texture that leads to the formation of a nominally single-crystal lm. In practice, it is rare for the crystalline quality of an epitaxial thin lm to approach that of a single crystal: epitaxial lms tend to contain a high density of dislocations and, possibly, low-angle grain boundaries. The degree of epitaxy is best assessed through detailed X-ray diffraction and analysis both of the spread of the in-plane lattice direction (mosaicity) and the out-of plane alignment (rocking curve). High-quality epitaxial oxide lms may show rocking curve widths well below 0.18. Although there is no single universal reason for requiring epitaxy in functional oxides, it is generally accepted that for most materials and applications epitaxy is desirable. In magnetic conductors such as the manganites, the highly localized nature of the conduction electrons render the materials extremely sensitive to local disorder, such as at grain boundaries. It was, for exmaple, proved fairly shortly after the discovery of colossal magnetore-

3830

2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Mater. 2009, 21, 38273839

www.advmat.de

PROGRESS REPORT

sistance (CMR) in doped lanthanum manganites that grain boundaries show a high resistance and a strong associated magnetoresistance that were associated with the spin disorder at the interface.[51] The complexity of the unit cell in many oxides creates a tendency to form antisite defects of various types. In Fe3O4, stacking faults, or antisite domain boundaries, couple antiferromagnetically as a consequence of the ferrimagnetic spin structure;[52] since the conduction is highly spin-polarized, such defects induce substantial spin scattering. More generally, many oxides show a signicant anisotropy (magnetic or transport) in their properties, and so epitaxial growth along particular directions can be used to optimize properties. The sensitivity to strain of many oxides means that epitaxy provides a further tool to control the electronic and magnetic properties. There are many examples of systems that, in the bulk, show purely insulating behavior but that in epitaxially strained thin-lm form show strong metalinsulator transitions. It is possible therefore to tune the Curie temperature in ferromagnetic systems mediated by double-exchange[53] directly via the lattice mismatch.[54,55]

monitoring allows the lm grower to monitor the termination of a lm structure;[60] the incorporation of annealing stages following the termination of a complete layer has been shown to maintain layer-by-layer growth over a greater range of lm thicknesses. PLD used in conjunction with RHEED is often referred to as laser-MBE,[57] and has been applied to a wide variety of magnetic oxides and heterostructures. 3.2.2. Molecular Beam Epitaxy (MBE) The effectiveness of MBE lies in its ability to control independently the incident atomic uxes at the growing lm. This requires some form of feedback control, usually rate monitoring, together with RHEED. The ability to monitor the surface termination, together with rate control of individual cations has enabled several groups to perfect the epitaxial growth of oxide compounds that are unstable in the bulk and create atomically precise heterostructures.[6164] Underlying oxide MBE is the requirement to deliver sufcient oxygen to the growing lm to stabilize the crystal structure while maintaining the cleanliness and long molecular mean-free-path necessary for MBE. 3.2.3. Atomic Layer Deposition (ALD)

3.2. Growth Techniques 3.2.1. Pulsed-Laser Deposition The widespread development of techniques for the deposition of high-quality thin lms of multicomponent oxides emerged in parallel with research on oxide HTSs such as YBa2Cu3O7 d. It was rapidly appreciated that the principle requirements for thin-lm growth of the oxide superconducting materials were: accurate stoichiometry control, optimized oxygen thermodynamics, and high structural perfection. Although early work on sputtering demonstrated the feasibility of HTS lm growth,[12] problems of stoichiometry control associated with selective resputtering of the more volatile components led to the widespread adoption of PLD for a wide variety of functional oxides.[26,5658] In PLD, ultrashort pulses of ultraviolet light are used to ablate material from a target (typically a disk of a few centimeters diameter). The short pulse duration ensures that generalized heating of the target is low, and thus that thermal evaporation, which would tend to select for the most volatile elements, is minimized. Since also the plasma plume associated with the ablation process contains only relatively low-energy species, there is negligible resputtering of the growing lm. Taken together, these features imply that in PLD the stoichiometry of the target (at least of the cation species) is usually preserved in the growing lm. In general, reasonably high-quality epitaxial growth can be achieved simply by maintaining an optimum gas pressure and substrate temperature, and ensuring that a compatible stoichiometric ux of material arrives at the substrate. PLD can achieve this without additional in situ monitoring techniques, which is a principal reason for the widespread popularity of the technique. The development of RHEED systems compatible with the high background oxygen pressures required for the growth of many oxides[59] has enabled in situ monitoring of the surface crystallinity in PLD and other growth processes. RHEED

The ALD technique is based on two self-limiting reactions between gas-phase precursor molecules and a solid surface. It thus differs from standard CVD, where there is mixing and thus reaction of precursor gases prior to the substrate. Hence, ALD lends itself to more precisely grown lms due to the ability to control the order in which gases arrive. There are single ALD steps: deposition is the outcome of sequential chemical reaction between a reactive precursor absorbed in monolayers at the substrate surface. There are four basic steps to the process. The surface is exposed to a rst gas (A) that reacts with the sites on the initial surface. This reaction stops when all of the sites are terminated. Excess reactants are then purged from the system. Next, the surface is exposed to a second gas (B), which reacts with the surface sites that result from the (A) reaction. Again, the system is purged after the reaction stops. If the second surface reaction returns the surface back to the previous termination, then atomic-layer controlled growth can be achieved using an alternating ABAB reaction sequence. The main advantages of the technique are: the lm thickness is dependent only on the number of cycles; it gives simple but extremely accurate thickness control with large batch capability; it produces highly homogeneous (atomic-level control of materials composition) and conformal lms; it does not require lineof-sight for deposition, and so very high aspect-ratio geometries or porous structures (including tubes and trenches) can be easily coated with extremely conformal lms; and there is no build-up in corners or blockage of the trenches. Hence, it is ideal for growth of nanostructures (such as in nanotemplates). Films are grown at low temperatures (100300 -C) compatible with many polymers, and multilayer structures can be fabricated in a continuous process. Although ALD is now applied in large scale, and has been instrumental in enabling the switch to directly deposited high-dielectric-constant gate oxides,[65] the challenge is to render it applicable to growth of complex oxides. There is already good promise for, for example, PZT[66] and manganites.[67]

Adv. Mater. 2009, 21, 38273839

2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

3831

www.advmat.de

PROGRESS REPORT

4. Functional Oxide Systems of Current Interest


The widespread interest in functional oxide thin lms stems largely from developments associated with the discovery and optimization of HTSs. Once it was discovered that the complex unit cells of these materials were not an insuperable obstacle to high-quality thin-lm growth, many other systems were quickly explored in thin-lm form. The remainder of this progress report deals with the currently topical areas of functional oxide lm research.

4.1. Materials 4.1.1. Multiferroics and Magnetoelectrics Magnetoelectric coupling describes the interconversion of magnetic and electrical signals using materials. Direct coupling permits a change of electrical polarization due to an applied magnetic eld, as demonstrated in the seminal work on TbMnO3[68] and TbMn2O5.[69] Converse coupling is more rare and describes the reverse scenario, where an applied electric eld produces a change in magnetization, for example in Cr2O3.[70,71] Separately, magnetoelectric coupling may be classied as direct or strain mediated, where the word direct is used again but this time to describe direct rather than strain-mediated interactions between magnetic and electrical stimuli and responses. Clearly, this double use of direct is unfortunate. But more unfortunate is the trivial yet prevalent interchange of multiferroic and magnetoelectric: the former refers to ground states and the latter refers to coupling. This distinction is more than just pedagogical, as magnetoelectric coupling is found in nonmultiferroic materials, such as the piezoelectric paramagnet NiSO4 6H2O[72] and the antiferromagnet Cr2O3.[73] However, it remains to be seen whether multiferroic systems with no magnetoelectric coupling exist. Therefore, it is best to consider that the two elds overlap.[74] Both direct and converse coupling have technological potential, for example for magnetically tunable lters and for electrically written magnetic data storage. However, there are substantial materials issues to be addressed in terms of the availability of suitable materials systems that can operate at room temperature and their sensitivity to defects. Recent materials developments include the demonstration that magnetic order can be electrically switched in the ferroelectric antiferromagnet BiFeO3;[75,76] the use of an interface that is both planar and epitaxial to mediate strain coupling between a ferromagnet and a ferroelectric,[77,78] the demonstration of a composite with vertical nanocolumns of ferroelectric and ferromagnetic phases.[79] Device developments that have the potential for applications include the demonstration of four-state memory using multiferroic tunnel barriers,[80] the electrical control of magnetization via exchange coupling/strain/heating from juxtaposed layers of BiFeO3[81] or YMnO3,[82] insight into the role of exural strain in magnetoelectric bilayers for microwavefrequency applications,[83] the report of giant a.c. magnetoelectric effects in a laminate without exploiting magnetostriction,[84] and the observation that standard multilayer capacitors serendipitously show both direct[85] and converse[86] magnetoelectric effects.

The vast majority of the rather small number of multiferroic oxides only display more than one ferroic order at low temperature. For example, the archetypical single-phase multiferroic BiMnO3,[87] although a high-temperature ferroelectric, only becomes ferromagnetically ordered at 105 K, setting the effective range over which the electrical control of nonvolatile magnetic order can be observed well below this value. For this reason, the majority of research aimed at demonstrating potential applications has focused on BiFeO3, which is both electrically and (anti)ferromagnetically ordered well above room temperature.[88,89] BiFeO3 is a ferroelectric with a high saturation polarization; over the past ve years, it has become clear that obtaining high polarization and remanence in thin lms requires careful defect optimization, principally through the control of oxygen vacancies that donate carriers and cause electrical leakage.[9092] Through the phenomenon of exchange coupling,[93] in which the ordering of a ferromagnet and an antiferromagnet are interfacially coupled, it is possible to use antiferromagnets to control the reversal of ferromagnetic layers in spin electronic devices. Thus, provided magnetoelectric coupling in BiFeO3 can be used to reverse the surface spins in an exchange-coupled bilayer, it should be possible to achieve the electric switching of a magnetic layer that is widely seen as the most obvious application of multiferroics. Aspects of this behavior have been observed in BiFeO3 (and other systems),[75,76,90,94] but there is not a complete understanding of the internal magnetic behavior of exchangecoupled systems, particularly those in which the magnetic

Figure 3. Magnetic hysteresis loop for a giant magnetoresistive spin-valve deposited on a BiFeO3 layer. The shifted lower half of the loop indicates the bias induced in the Ni Fe layer adjacent to the BiFeO3. Adapted with permission from [90].

3832

2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Mater. 2009, 21, 38273839

www.advmat.de

PROGRESS REPORT

ordering is established well above room temperature.[95] Moreover, strain could play a role. Figure 3 shows exchange bias induced in a permalloy (Ni80Fe20, Py) layer deposited on BiFeO3. This bias is an indicator that sufcient reorganization of the BiFeO3 surface spins has taken place during eld annealing that a nonzero interaction exists with the Py spins. In conventional exchange-bias systems, such eld annealing takes place above the Neel temperature of the antiferromagnet, and so the ordering on cooling can be driven by the surface-exchange interaction with the ferromagnet. Even in such conventional systems, it is now known that defects play a crucial role in determining the magnitude of the exchange bias.[96] In unconventional systems (such as BiFeO3), which cannot be annealed above TN without inducing interfacial reaction or diffusion, defects play an even stronger role, because only local changes in the antiferromagnetic order can be driven by the exchange interaction.[95] 4.1.2. High-Spin-Polarization Oxides As device dimensions continue to shrink, conventional semiconductor technology is approaching its fundamental physical limits. Further performance improvements beyond the end of the semiconductor road-map appear neither technically nor economically feasible. Spin-electronic (spintronic) devices offer a replacement technology in which information is encoded in the form of spin as well as charge. Most spintronic device concepts rely on a source of spin-polarized currents, and so there is strong interest in materials that have the largest possible polarizations. 4.1.2.1. Manganites: The eld of manganites became popular in 1994 as the discovery of CMR[97] fell during the decade between the discovery of giant magnetoresistance (GMR)[98,99] and its application in disk-drive read heads. Commercial exploitation of low-eld magnetoresistance[51,100] in manganites did not develop, and the phenomenon of magnetic and electronic phase separation became the dominant research theme as complex patterns on a range of length scales were observed using modern imaging techniques[101] such as scanning tunneling microscopy (STM),[102] magnetic force microscopy (MFM),[103] and transmission electron microscopy (TEM).[104] Phase-separation phenomena in manganites remain interesting because the complex patterns are difcult to predict. Models based on strain inhomogeneities[105] are certainly reasonable, and can be linked with experimental data.[106,107] However, the role of strain remains confusing, and samples with nominally the same chemical formula can behave quite differently if prepared as single crystals, polycrystals, or thin lms. A particularly elegant experiment demonstrated this point directly by showing that the ground state of La0.5Ca0.5MnO3 could be dramatically changed by altering the grain size.[108] More generally, the coupled order parameters produce rich phase diagrams, and it should be noted that this richness is achieved even with just two (magnetic and charge) order parameters,[109] that is, ignoring strain. Manganites also remain important as sources of highly spin-polarized electrons for spin electronics. Some of the multiferroicmagnetoelectric experiments described above use manganites for this reason,[80] or simply because manganites are ferromagnets that show reasonable magnetostriction and can be epitaxially integrated with ferroelectrics despite a 34% lattice

mismatch.[77,78] Other recent developments that exploit the high spin polarization of manganites include the transformation of spin information into large electrical signals using a carbon nanotube with manganite electrodes,[110] the use of an organic channel (Alq3) to achieve a qualitatively similar result that is quantitatively weaker but persists at room temperature,[111] and magnetic tunnel junctions that exploit interfacial engineering to try to control charge transfer between the tunnel barrier and the manganite electrodes.[112,113] This latter point about chemically controlling manganite interfaces is particularly important, as manganite surfaces differ strongly from the bulk.[114,115] The applications potential of manganites remains unclear. Apart from the use of manganites in spintronics, as discussed above, there is some key basic science yet to emergein particular an answer to the controversy[116] concerning the nature of the charge ordered phase. This phase is one of two predominant ordered phases seen in manganites, the other being the ferromagnetic metallic phase exploited in spintronics. However, its basic nature remains unknown, as the model of an ordered pattern of Mn3R and Mn4R is overly simplistic. Ultimately, the answers may only come when experimental techniques develop further. The most immediate excitement with manganites could be their ability to display electrically controlled memory effects for resistive random-access memory (RRAM).[117] 4.1.2.2. Double Perovskites: Relatively few highly spin-polarized oxides have Curie temperatures above room temperature, and those that do tend to be structurally more complex than simple doped perovskites. Double perovskites, such as the prototypical Sr2FeMnO3 (SFMO), have relatively high TCs,[118] and band structure calculations suggest high spin polarizations.[119] The primary problem is that although SFMO is a pure compound, and some of the problems encountered in the manganites, such as localized strain and charge-ordering, are irrelevant; obtaining the necessary long-range ordering of the B-site cations (Fe and Mo in the case of SFMO) extremely difcult. Sr2FeMoO6 is a ferrimagnet with a Curie temperature of 430 K and a predicted saturation magnetization of 4 mB per formula unit. In practice, it has proved difcult to obtain the theoretical magnetization values from any of the double perovskites. These materials have the same tendency of simple perovskites for nonstoichiometry and properties that are strongly perturbed by grain boundaries.[120,121] It is now widely accepted that this is a consequence of the antisite disorder that is possible within the double perovskite unit cell.[122] These compounds share the basic ABO3 perovskite unit cell with the manganites and other ferromagnetic oxides, such as SrRuO3; however, the B-site cations form an ordered superlattice that doubles the dimensions of the unit cell. If this order is absent, then the material can be considered more correctly as a solid solution between two simple perovskite compounds AB1O3 and AB2O3. In partially disordered Sr2FeMoO6, a fraction of Fe ions will be nearest neighbors, and so can couple antiferromagnetically via the superexchange interaction; models for the magnetic properties of disordered double perovskites predict a total moment that decreases linearly with disorder.[123] This linear dependence has been widely observed experimentally[122,124] although there is some uncertainty as to the precise scaling factor. Methods of controlling the order have been developed,[125] but double perovskites are

Adv. Mater. 2009, 21, 38273839

2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

3833

www.advmat.de

regarded as difcult materials with no immediate prospect of application. The doubled unit cell can also potentially induce antiphase boundaries similar to those reported in Fe3O4.[126] These boundaries can be expected to induce a similar reduction in the total moment to uncorrelated disorder, but will have a signicant effect on conductivity and magnetoresistance, particularly in thin lms.[124] A sub-500 K TC is probably still marginal for device operation. However, it has been shown that Sr2CrReO6 has a Curie temperature close to 630 K,[127] and so, provided it can also be shown to be highly spin polarized and can be grown in thin-lm form, there is clear potential for device applications of the double perovskites. The growth of high-quality thin lms of the double perovskites such as Sr2FeMoO6 and Sr2CrWO6 has proved difcult for a number of reasons: the materials have complex chemical phase diagrams and the compounds are part of a wide solid-solution region, they require reducing conditions for growth to control the B site cation valence states and oxygen stoichiometry.[128,129] This is further complicated by the tendency to antisite disorder discussed above. To some extent, these problems can be controlled by appropriate oxidation conditions; ultrasonic spray pyrolysis has been shown to be an economical and exible route to thin-lm growth.[130] Using PLD with sufciently high deposition temperatures, it is possible to form well-ordered high-quality epitaxial lms with smooth surfaces.[131,132] However, the high deposition temperatures and low background oxygen pressure required for optimal growth create a high density of oxygen vacancies in the SrTiO3 substrates, which provide an optimum lattice match; this renders the substrates highly conductive, and so prevents accurate resistivity or device measurements.[14] Although there is reason to expect good device performance from the double perovskites, the complexities of lm growth have so far prevented the fabrication of heterostructure devices that might be expected to show large room-temperature TMR values. Bibes et al.[133] have fabricated Sr2FeMoO6/SrTiO3/Co junctions that show TMR values of around 60% at low temperatures, which is interpreted in terms of a high spin polarization for Sr2FeMoO6. Rager et al.[130] have reported a polarization value by point contact Andreev reection of 60%. 4.1.2.3. Other Highly Polarized Oxides: Fe3O4 and CrO2 are the simplest potential candidate compounds. The very high TC $858 K of Fe3O4 has made it a prime target of research, but it has become clear that, as in the double perovskites, obtaining the required long-range order is exceptionally difcult, so that even in highly metallic thin lms high values of spin polarization have not been achieved.[134] CrO2, in contrast, is highly spin polarized, and arguably has some claims to be the closest approximation to a half-metallic material.[135137] It has been used for this purpose in a number of important experiments,[138,139] but commercial applications appear unlikely because of the comparatively low TC (390 K). 4.1.3. Dilute Magnetic Oxides While spintronics promises lower switching energies and faster speeds, a major limitation on its development as a viable technology is the lack of room-temperature magnetic semicon-

ductor materials that would permit efcient spin transfer into active devices. This has been the driver underlying the explosion of interest in dilute magnetic oxides (DMOs) with properties tunable between insulating and metallic. There are a considerable number of oxides that show similar ferromagnetic behavior when doped, including TiO2 d,[140,141] SnO2 d,[142] In2O3 d[143] Cu2O,[144] and (LaSr)TiO3.[145] By far the most attention has been directed to ZnO doped with transition-metal (TM) ions. ZnO has various advantages: it grows easily on sapphire, it may be prepared by different routes, its band gap of $3.4 eV is convenient for optical applications, and it is already used for a number of applications utilizing its optical and piezoelectric properties. Taken together, there is a large body of evidence that shows that the defect bands in oxides, particularly ZnO, are highly spin polarizable. Pure ZnO has been found to be ferromagnetic when grown marginally oxygen decient,[146] when doped with carbon,[147] Ga,[148] and with most of the rst-series TM ions,[149] although most work has been done with cobalt doping. Magnetic circular dichroism (MCD) observed at the band-gap energy demonstrates that the conduction band is spin polarized; this has been seen for Ti-, V-, Mn-, and Co-doped ZnO.[150] Substantial tunnel magnetoresistance (TMR), which provides a more direct demonstration of spin-polarization, has been reported in ZnO,[151153] as well as TiO2[141] and Co-LaSrTiO3,[145] and an anomalous Hall effect (AHE) has been seen in heavily oxygen-decient TiO2 d.[140] In a classic dilute magnetic semiconductor,[154] such as GaMnAs,[155] the magnetism comes from the local moments, on the Mn ions, that are coupled magnetically, by the RudermanKittelKasuyaYoshida (RKKY) interaction via the mobile holes. In this case, the divalent Mn ions on the Ga3R site generate both the magnetism and the carriers. It is now understood that the DMOs are different for several fundamental reasons: divalent TM ions replace divalent Zn, and defects or codopants are necessary to provide carriers. It is well known that strain, bulk, and surface defects (particularly acute in thin-lm heterostructures with large surface and interface areas) have a strong inuence on the functional properties of oxides, and this is particularly so for magnetism: this renders the properties of DMOs much more sensitive to defects, and hence to the exact growth conditions, than GaMnAs. There is considerable controversy over the origin of the magnetism in doped ZnO and other oxides. It seems clear that the controversy has its origin in the lack of control of defect levels in ZnO and DMOs in general. We take the view that there has simply been insufcient attention paid to the materials control, and a serious effort is now required to address this. 4.2. Oxide Interfaces 4.2.1. Tunnel Devices Probably the simplest possible application of an oxide is as a tunnel barrier. Here, apparently, the only materials requirement is that the oxide should be insulating with a suitable band-gap of perhaps 25 eV. Solving the Schrodinger equation to model tunneling through a rectangular potential barrier is universally taught as an elementary example of quantum mechanics. At this

PROGRESS REPORT
3834

2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Mater. 2009, 21, 38273839

www.advmat.de

PROGRESS REPORT

extreme, tunneling is a simple process in which the resistance imposed by a tunnel current is determined only by the barrier height and electron (effective) mass. However, in reality, tunneling is critically inuenced by intrinsic factors such as the complex band structure of the electrodes and the barrier, and by extrinsic issues including barrier defects, interface states, and nonstoichiometry. The challenge for materials science is to link the electrical response of a tunneling system to the device microstructure. Paradoxically, in view of the inherent complexity of superconductivity, it is most feasible to do this for superconducting devices. This is because, as discovered by Giaever and Zeller,[156] the long-range coherence (and for simple metallic superconductors, isotropy) of the superconducting state averages over the intrinsic band structure and defects in the electrodes so that the current-versus-voltage characteristic of a superconductor/ insulator/superconductor (SIS) junction can be related directly to the density of states (DoS) of the superconducting condensate and associated excitations. The energy gap in this DoS means that, at low temperatures for voltages less than the sum of the electrode energy gaps, single-stage tunneling is forbidden. The residual subgap current therefore provides direct information on conduction processes via barrier defects relative to single-step tunneling (which dominates for high voltages). Despite these considerable simplications, directly relating the tunnel conductance to barrier/interface defects has rarely been unambiguously achieved because of the difculty of obtaining detailed microstructural information at 12 nm length scales in buried layers. Magnetic tunnel junctions have been a major technological success story; read-heads based on tunnel magnetoresistance (TMR) are found in nearly every modern disk drive, and the performance, in terms of their MR, is constantly being improved. Much higher TMR values have been measured in oxide heterostructures, but it is worth reviewing the situation that applies in the case of simple metal electrodes. The simplest system to analyze is tunneling between two isotropic, homogeneously magnetized polycrystalline ferromagnets through an amorphous tunnel barrier. Here, the tunneling DoS for majority and minority electrons on each side of the barrier is an average over the whole Fermi surface, and so the spin polarization can be validly quoted as a single parameter. This approach applies well to simple metal systems, and is the justication of the Julliere model:[157] even here, however, dening what determines the tunneling spin polarization is complicated. Recently, attention has focused on epitaxial TMR devices, such as (100) Fe/MgO/Fe, that exploit particular features of the electrode and barrier Fermi surfaces.[158] Here, the effective barrier height in the parallel state is lower than in the antiparallel state, so that the barrier has a strongly spin-ltering effect. Nevertheless, the experimentally observed TMR values are much lower than those predicted theoretically[159] (although still higher than for simple amorphous barriers). The issues are believed to arise predominantly from oxygen vacancies in the MgO, but there are obvious issues of partial oxidation of the ferromagnet at the interface that also affect the behavior. For a variety of reasons, principally because many magnetic oxides are probably half-metallic, or at least highly spin polarized, but also including chemical and structural compat-

ibility, oxide heterostructure TMR devices have been of great interest over the past decade. The archetypal system is LSMO/ STO/LSMO,[160] in which there is little change in lattice parameter at the interfaces and high oxygen pressures can be used during growth to minimize the possibilities of oxygen defects in the barrier. It has become increasingly clear that the properties of such systems are exceptionally sensitive to subtle structural changes, such as bond angles and lengths induced by heteroeptiaxy. Various renements, including the use of even better lattice-matched barriers, have yielded TMR values well above 1000% at low temperatures.[113,161] The situation with other nominally half-metallic materials (such as Fe3O4, which has a much higher Curie temperature), is generally poorer, with low TMR values even at low temperatures.[162] Spin-polarized tunnel devices are also being developed as sources of spin-polarized carriers for semiconductor spintronics. For this, instead of tunneling from a ferromagnetic electrode, it is possible instead to use a spin-polarizing barrier;[163] this is perhaps a pragmatic choice, given the greater number of oxide magnetic insulators compared with metallic oxide ferromagnets. In such materials, the barrier heights for the two spinpolarization signs are exchange-split, and so the transmission coefcient for one spin sign can be orders of magnitude higher than for the other. Recent work using the ferromagnetic semiconductors EuO and EuS has produced encouraging results,[164] although limited by the low TC of these materials. Preliminary results using higher TC spinels, such as NiFe2O4[165] and multiferroics,[80] show some effect, although, again, so far only at low temperatures. These results present a signicant challenge for materials understanding at ultrashort length scales. The electrical properties of crystalline systems can only be fully understood through the use of detailed band-structure calculations of the heterostructure system;[159] however, the incorporation of chemical and microstructural defects is beyond current techniques, even if these could be unambiguously identied in experiment. Given that the technological drive remains for high TMR and spin injection at room temperature, it seems that progress in the short term can only be made by careful comparative experiments that exploit the understanding of stoichiometry and defects at the bulk level, with the aim of minimizing extrinsic factors. 4.2.2. Two-Dimensional Electron Gases Along with multiferroics, the most active area of oxide research has been the study of interfacial two-dimensional gases in perovskite bilayers. The eld was initiated by the study of bilayers of LaAlO3 and SrTiO3. Both these materials are insulators with a relatively wide band-gap, yet when grown as a bilayer, Ohtomo and Hwang[166] found a conducting interfacial layer with high mobility and carrier density. Indeed, as they acknowledged in the paper reporting these ndings, these values seemed unreasonably high. In principle, the physics underlying this behavior is rather simple. In the (001) growth direction, the bilayer consists of alternating SrO and TiO2 layers that switch to LaO and AlO2 at the interface. Within the bulk, the former two planes are uncharged, while the layers within LaAlO3 are alternately positively and negatively charged. Without compensation, this charge alterna-

Adv. Mater. 2009, 21, 38273839

2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

3835

www.advmat.de

tion would lead to the rather grandly termed polar catastrophe, in that the internal electric eld would create a linearly increasing surface potential as the LaAlO3 layer becomes thicker. This potential can be eliminated by the transfer of charge density, half that of the LaO plane, to the interface.[167] At least for cases of negative charge transfer (where the interface is formed by TiO2/ LaO), this charge is mobile, and forms a conducting layer at the interface with a large potential mobility because of the low level of defects in this interface. It is assumed that the charge transfer results in a Ti3/Ti4 mixed valence that may have magnetic[168,169] as well as electrical properties. The reality is clearly rather more complicated. The initial samples were grown at rather low oxygen pressures, and there is now good evidence that this results in a substantial oxygen deciency in the SrTiO3 substrate. Herranz et al.[170] demonstrated this directly by measuring Shubnikov de Haas oscillations, indicative of three-dimensional metallicity, in samples grown under these reducing conditions, and Basletic et al.[171] have directly mapped the spread of the conductivity into the substrate with decreasing oxygen pressure by using scanning probe microscopy of a sample cross-section. These observations presumably account for the unphysically large charge densities and mobilities originally reported from such systems. The solution is to grow samples at higher oxygen pressures and to anneal so as to ensure full oxygenation of the material. A number of groups have now performed systematic experiments in which most of the parameters of the structure have been varied.[169] Samples grown at high enough oxygen pressure show a dependence of the conductivity on the LAO thickness that is consistent with the basic charge-distribution model in that up to four unit cells the structures remain insulating, but beyond this the surface potentials are sufcient to drive charge distribution, and the interface becomes conducting.[172] However, even within this charge-distribution regime, there remains a strong dependence of the conductivity on the oxygen pressure. This is an exciting and emergent eld in which superconductivity[173] and its control by electric eld[174,175] have been demonstrated.

PROGRESS REPORT

4.2.3. Vertical Nanocomposites A novel method to control strain (and hence also defects) is through formation of vertical nanocomposite epitaxial structures.[79,176] Figure 4 shows a spontaneously ordered nanocheckerboard in a thin lm. This was achieved in the BiFeO3 (B, magnetoelectric and ferroelectric)/Sm2O3 (S, dielectric) system, where the atomic concentrations of each phase were 1:1. In these structures, strain can be controlled throughout the thickness of relatively thick epitaxial lms (>500 nm), and the strain in the lower-modulus phase is controlled by that in the higher-modulus phase. This is very different to the situation on normal bilayer lms, because strain relaxation occurs in lms of >50 nm thickness. Here, the vertical column widths are so small ($10 nm) that the strain cannot fully relax. We have achieved much lower dielectric losses in BiFeO3 when it is coupled to Sm2O3 than when it is in the form of a pure lm.[177] It appears that the compression induced by the Sm2O3 in BiFeO3 reduces oxygen deciency and charge-carrier concentration, thus reducing the leakage.

Acknowledgements
Thanks are given to Prof. M. Carpenter for helpful discussions and to Prof. H. Wang, Texas A and M, for providing the planar TEM image of Figure 4. Funding is acknowledged from the EU grants Marie Curie EXT-014156 NanoFen and Novel Nanoscale Devices based on functional Oxide Interfaces (NANOXIDE) and various EPSRC grants. This article is part of the Special Issue celebrating the 800th Anniversary of the University of Cambridge. Received: March 19, 2009 Revised: May 8, 2009 Published online: August 5, 2009

Figure 4. Planar transmission electron microscopy image showing the surface of a self-organized vertical nanocheckerboard structure of BiFeO3 and Sm2O3.

[1] W. A. Wooster, A Text-Book on Crystal Physics, Cambridge University Press, 1938, Cambridge, UK. [2] W. A. Wooster, Experimental Crystal Physics, Clarendon Press, Oxford 1957. [3] W. A. Wooster, Diffuse X-Ray Reections from Crystals, Clarendon Press, Oxford 1962. [4] H. D. Megaw, Nature 1945, 155, 484. [5] H. D. Megaw, Ferroelectricity in Crystals, Methuen, London, 1957. [6] http://cwp.library.ucla.edu/Phase2/Megaw,-Helen@851234567.html (Accessed on March 23, 2009). [7] A. M. Glazer, H. D. Megaw, Philos. Mag. 1972, 25, 1119. [8] A. M. Glazer, H. D. Megaw, Acta Crystallogr, A 1973, A29, 489. [9] R. Clarke, A. M. Glazer, Ferroelectrics 1976, 12, 207. [10] A. M. Glazer, Acta Crystallogr, Sect. B: Struct. Sci. 1972, 28, 3384. [11] R. Clarke, R. W. Whatmore, A. M. Glazer, Ferroelectrics 1976, 13, 497. [12] R. E. Somekh, M. G. Blamire, Z. H. Barber, K. Butler, J. H. James, G. W. Morris, E. J. Tomlinson, A. P. Schwarzenberger, W. M. Stobbs, J. E. Evetts, Nature 1987, 326, 857. [13] J. G. Bednorz, K. A. Muller, Z. Phys. B 1986, 64, 189. [14] J. N. Eckstein, Nat. Mater. 2007, 6, 473. [15] J. H. Haeni, P. Irvin, W. Chang, R. Uecker, P. Reiche, Y. L. Li, S. Choudhury, W. Tian, M. E. Hawley, B. Craigo, A. K. Tagantsev, X. Q. Pan, S. K. Streiffer, L. Q. Chen, S. W. Kirchoefer, J. Levy, D. G. Schlom, Nature 2004, 430, 758. [16] J. P. Velev, P. A. Dowben, E. Y. Tsymbal, S. J. Jenkins, A. N. Caruso, Surf. Sci. Rep. 2008, 63, 400.

3836

2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Mater. 2009, 21, 38273839

www.advmat.de

PROGRESS REPORT

[17] Y. M. Chiang, D. P. Birnie, W. D. Kingery, Physical Ceramics: Principles for Ceramic Science and Engineering, John Wiley & Sons Inc, New York 1997. [18] S. Goedecker, T. Deutsch, L. Billard, Phys. Rev. Lett. 2002, 88, 235501. [19] R. Falster, V. V. Voronkov, Mater. Sci. Eng, B 2000, 73, 87. [20] P. G. Radaelli, M. Marezio, H. Y. Hwang, S. W. Cheong, B. Batlogg, Phys. Rev. B 1996, 54, 8992. [21] H. Hartono, C. B. Soh, S. Y. Chow, S. J. Chua, E. A. Fitzgerald, Appl. Phys. Lett. 2007, 90, 171917. [22] R. J. D. Tilley, Principles and Applications of Chemical Defects, CRC Press, 1998, Cheltenham: Thornes. [23] R. J. D. Tilley, Defects in Solids, Wiley publishers, Weinheim, Germany 2008. [24] J. Mizusaki, Solid State Ionics 2000, 129, 163. [25] D. H. A. Blank, G. Koster, G. Rijnders, E. van Setten, P. Slycke, H. Rogalla, J. Cryst. Growth 2000, 211, 98. [26] H. M. Christen, G. Eres, J. Phys.: Condens. Mater 2008, 20, 264005. [27] S. R. Foltyn, L. Civale, J. L. MacManus-Driscoll, Q. X. Jia, B. Maiorov, H. Wang, M. Maley, Nat. Mater. 2007, 6, 631. [28] E. Z. Liu, Y. He, J. Z. Jiang, Appl. Phys. Lett. 2008, 93, 132506. [29] A. Rothschild, W. Menesklou, H. L. Tuller, E. Ivers-Tiffee, Chem. Mater. 2006, 18, 3651. [30] N. Khare, M. J. Kappers, M. Wei, M. G. Blamire, J. L. MacManus-Driscoll, Adv. Mater. 2006, 18, 1449. [31] M. Brazier, S. Mansour, M. McElfresh, Appl. Phys. Lett. 1999, 74, 4032. [32] P. Chaudhari, J. Mannhart, D. Dimos, C. C. Tsuei, J. Chi, M. M. Oprysko, M. Scheuermann, Phys. Rev. Lett. 1988, 60, 1653. [33] J. H. Durrell, M. J. Hogg, F. Kahlmann, Z. H. Barber, M. G. Blamire, J. E. Evetts, Phys. Rev. Lett. 2003, 90, 247006. [34] J. Mannhart, D. Anselmetti, J. G. Bednorz, A. Catana, C. Gerber, K. A. Muller, D. G. Schlom, Z. Phys. B 1992, 86, 177. [35] K. Char, L. Antognazza, T. H. Geballe, Appl. Phys. Lett. 1994, 65, 904. [36] B. X. Lin, Z. X. Fu, Y. B. Jia, Appl. Phys. Lett. 2001, 79, 943. [37] P. K. Muduli, S. K. Bose, R. C. Budhani, J. Phys.: Condens. Matter 2007, 19, 226204. [38] K. J. Choi, M. Biegalski, Y. L. Li, A. Sharan, J. Schubert, R. Uecker, P. Reiche, Y. B. Chen, X. Q. Pan, V. Gopalan, L. Q. Chen, D. G. Schlom, C. B. Eom, Science 2004, 306, 1005. [39] H. N. Lee, H. M. Christen, M. F. Chisholm, C. M. Rouleau, D. H. Lowndes, Nature 2005, 433, 395. [40] K. Nomura, H. Ohta, A. Takagi, T. Kamiya, M. Hirano, H. Hosono, Nature 2004, 432, 488. [41] R. L. Hoffman, B. J. Norris, J. F. Wager, Appl. Phys. Lett. 2003, 82, 733. [42] U. Ozgur, Y. I. Alivov, C. Liu, A. Teke, M. A. Reshchikov, S. Dogan, V. Avrutin, S. J. Cho, H. Morkoc, J. Appl. Phys. 2005, 98, 041301. [43] L. Dunlop, A. Kursumovic, J. L. MacManus-Driscoll, Appl. Phys. Lett. 2008, 93, 172111. [44] C. Lee, Y. Park, J. Lim, J. Korean Phys. Soc. 2007, 50, 590. [45] A. Tsukazaki, A. Ohtomo, T. Kita, Y. Ohno, H. Ohno, M. Kawasaki, Science 2007, 315, 1388. [46] S. V. Kalinin, D. A. Bonnell, Phys. Rev. B 2001, 63, 125411. [47] J. L. Giocondi, G. S. Rohrer, J. Phys. Chem. B 2001, 105, 8275. [48] P. M. Jones, D. E. Gallardo, S. Dunn, Chem. Mater. 2008, 20, 5901. [49] M. Bibes, A. Barthelemy, IEEE Trans. Electron Devices 2007, 54, 1003. [50] M. P. Singh, B. Carvello, L. Ranno, Appl. Phys. Lett. 2006, 89, 022504. [51] H. Y. Hwang, S. W. Cheong, N. P. Ong, B. Batlogg, Phys. Rev. Lett. 1996, 77, 2041. [52] W. Eerenstein, T. T. M. Palstra, T. Hibma, S. Celotto, Phys. Rev. B 2002, 66, 201101. [53] P. G. de Gennes, Phys. Rev. 1960, 118, 141. [54] M. H. Jo, N. D. Mathur, J. E. Evetts, M. G. Blamire, M. Bibes, J. Fontcuberta, Appl. Phys. Lett. 1999, 75, 3689. [55] M. G. Blamire, B. S. Teo, J. H. Durrell, N. D. Mathur, Z. H. Barber, J. L. M. Driscoll, L. F. Cohen, J. E. Evetts, J. Magn. Magn. Mater. 1999, 191, 359.

[56] A. Gupta, R. Gross, E. Olsson, A. Segmuller, G. Koren, C. C. Tsuei, Phys. Rev. Lett. 1990, 64, 3191. [57] H. Koinuma, M. Kawasaki, T. Itoh, A. Ohtomo, M. Murakami, Z. W. Jin, Y. Matsumoto, Phys. C 2000, 335, 245. [58] Z. H. Barber, J. Mater. Chem. 2006, 16, 334. [59] G. Rijnders, G. Koster, D. H. A. Blank, H. Rogalla, Appl. Phys. Lett. 1997, 70, 1888. [60] G. Koster, G. J. H. M. Rijnders, D. H. A. Blank, H. Rogalla, Appl. Phys. Lett. 1999, 74, 3729. [61] R. A. McKee, F. J. Walker, M. F. Chisholm, Phys. Rev. Lett. 1998, 81, 3014. [62] I. Bozovic, V. Matijasevic, Mater. Sci. Forum 2000, 352, 1. [63] D. G. Schlom, J. H. Haeni, J. Lettieri, C. D. Theis, W. Tian, J. C. Jiang, X. Q. Pan, Mater. Sci. Eng, B 2001, 87, 282. [64] D. G. Schlom, L. Q. Chen, X. Q. Pan, A. Schmehl, M. A. Zurbuchen, J. Am. Ceram. Soc. 2008, 91, 2429. [65] P. D. Kirsch, M. A. Quevedo-Lopez, H. J. Li, Y. Senzaki, J. J. Peterson, S. C. Song, S. A. Krishnan, N. Moumen, J. Barnett, G. Bersuker, P. Y. Hung, B. H. Lee, T. Lafford, Q. Wang, D. Gay, J. G. Ekerdt, J. Appl. Phys. 2006, 99, 023508. [66] T. Watanabe, S. Hoffmann-Eifert, C. S. Hwang, R. Waser, J. Electrochem. Soc. 2008, 155, D715. [67] O. Nilsen, E. Rauwel, H. Fjellvag, A. Kjekshus, J. Mater. Chem. 2007, 17, 1466. [68] T. Kimura, T. Goto, H. Shintani, K. Ishizaka, T. Arima, Y. Tokura, Nature 2003, 426, 55. [69] N. Hur, S. Park, P. A. Sharma, J. S. Ahn, S. Guha, S. W. Cheong, Nature 2004, 429, 392. [70] D. N. Astrov, Sov. Phys. JETP 1960, 11, 708. [71] V. J. Folen, G. T. Rado, E. W. Stalder, Phys. Rev. Lett. 1961, 6, 607. [72] S. L. Hou, N. Bloembergen, Phys. Rev. 1965, 138, A1218. [73] P. Borisov, A. Hochstrat, X. Chen, W. Kleemann, C. Binek, Phys. Rev. Lett. 2005, 94, 117203. [74] W. Eerenstein, N. D. Mathur, J. F. Scott, Nature 2006, 442, 759. [75] D. Lebeugle, D. Colson, A. Forget, M. Viret, A. M. Bataille, A. Gukasov, Phys. Rev. Lett. 2008, 100, 227602. [76] S. Lee, T. Choi, W. Ratcliff, R. Erwin, S. W. Cheong, V. Kiryukhin, Phys. Rev. B 2008, 78, 100101. [77] W. Eerenstein, M. Wiora, J. L. Prieto, J. F. Scott, N. D. Mathur, Nat. Mater. 2007, 6, 348. [78] C. Thiele, K. Dorr, O. Bilani, J. Rodel, L. Schultz, Phys. Rev. B 2007, 75, 054408. [79] J. L. Macmanus-Driscoll, P. Zerrer, H. Y. Wang, H. Yang, J. Yoon, A. Fouchet, R. Yu, M. G. Blamire, Q. X. Jia, Nat. Mater. 2008, 7, 314. [80] M. Gajek, M. Bibes, S. Fusil, K. Bouzehouane, J. Fontcuberta, A. BarthAlAmy, A. Fert, Nat. Mater. 2007, 6, 296. [81] Y. H. Chu, L. W. Martin, M. B. Holcomb, M. Gajek, S. J. Han, Q. He, N. Balke, C. H. Yang, D. Lee, W. Hu, Q. Zhan, P. L. Yang, guez, A. Scholl, S. X. Wang, R. Ramesh, Nat. Mater. A. Fraile-Rodr 2008, 7, 478. [82] V. Laukhin, V. Skumryev, X. Marti, D. Hrabovsky, F. Sanchez, a-Cuenca, C. Ferrater, M. Varela, U. Luders, J. F. Bobo, M. V. Garc J. Fontcuberta, Phys. Rev. Lett. 2006, 97, 227201. [83] V. M. Petrov, G. Srinivasan, A. Galkina, J. Appl. Phys. 2008, 104, 113910. [84] Z. P. Xing, J. F. Li, D. Viehland, Appl. Phys. Lett. 2008, 93, 013505. [85] C. Israel, N. D. Mathur, J. F. Scott, Nat. Mater. 2008, 7, 93. [86] C. Israel, S. Kar-Narayan, N. D. Mathur, Appl. Phys. Lett. 2008, 93, 173501. [87] W. Eerenstein, F. D. Morrison, J. F. Scott, N. D. Mathur, Appl. Phys. Lett. 2005, 87, 101906. [88] W. Eerenstein, F. D. Morrison, J. Dho, M. G. Blamire, J. F. Scott, N. D. Mathur, Science 2005, 307, 1203. [89] J. Wang, J. B. Neaton, H. Zheng, V. Nagarajan, S. B. Ogale, B. Liu, D. Viehland, V. Vaithyanathan, D. G. Schlom, U. V. Waghmare, N. A. Spaldin, K. M. Rabe, M. Wuttig, R. Ramesh, Science 2003, 299, 1719.

Adv. Mater. 2009, 21, 38273839

2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

3837

www.advmat.de

PROGRESS REPORT

[90] J. Dho, X. Qi, J. L. MacManus-Driscoll, M. G. Blamire, H. Kim, Adv. Mater. 2006, 18, 1445. [91] X. Qi, M. Wei, Y. Lin, Q. Jia, D. Zhi, J. Dho, M. G. Blamire, J. L. MacManus-Driscoll, Appl. Phys. Lett. 2005, 86, 071913. [92] X. Qi, J. Dho, R. Tomov, M. G. Blamire, J. L. MacManus-Driscoll, Appl. Phys. Lett. 2005, 86, 062903. [93] J. Nogues, I. K. Schuller, J. Magn. Magn. Mater. 1999, 192, 203. [94] N. A. Spaldin, R. Ramesh, MRS Bull. 2008, 33, 1047. [95] M. G. Blamire, IEEE Trans. Magn. 2008, 44, 1946. [96] U. Nowak, K. D. Usadel, J. Keller, P. Miltenyi, B. Beschoten, G. Guntherodt, Phys. Rev. B 2002, 66, 014430. [97] S. Jin, T. H. Tiefel, M. McCormack, R. A. Fastnacht, R. Ramesh, L. H. Chen, Science 1994, 264, 413. [98] P. Grunberg, R. Schreiber, Y. Pang, M. B. Brodsky, H. Sowers, Phys. Rev. Lett. 1986, 57, 2442. [99] M. N. Baibich, J. M. Broto, A. Fert, F. N. Vandau, F. Petroff, P. Eitenne, G. Creuzet, A. Friederich, J. Chazelas, Phys. Rev. Lett. 1988, 61, 2472. [100] N. D. Mathur, S. P. Isaac, G. Burnell, B. S. Teo, L. F. Cohen, J. L. MacManus-Driscoll, J. E. Evetts, M. G. Blamire, Nature 1997, 387, 266. [101] N. Mathur, P. Littlewood, Phys. Today 2003, 56, 25. [102] C. Renner, G. Aeppli, B. G. Kim, Y. A. Soh, S. W. Cheong, Nature 2002, 416, 518. [103] L. Zhang, C. Israel, A. Biswas, R. L. Greene, A. de Lozanne, Science 2002, 298, 805. [104] J. C. Loudon, N. D. Mathur, P. A. Midgley, Nature 2002, 420, 797. [105] K. H. Ahn, T. Lookman, A. R. Bishop, Nature 2004, 428, 401. [106] W. D. Wu, C. Israel, N. Hur, S. Park, S. W. Cheong, A. De Lozanne, Nat. Mater. 2006, 5, 881. [107] Y. A. Soh, P. G. Evans, Z. Cai, B. Lai, C. Y. Kim, G. Aeppli, N. D. Mathur, M. G. Blamire, E. D. Isaacs, J. Appl. Phys. 2002, 91, 7742. [108] P. Levy, F. Parisi, G. Polla, D. Vega, G. Leyva, H. Lanza, R. S. Freitas, L. Ghivelder, Phys. Rev. B 2000, 62, 6437. [109] G. C. Milward, M. J. Calderon, P. B. Littlewood, Nature 2005, 433, 607. [110] L. E. Hueso, J. M. Pruneda, V. Ferrari, G. Burnell, J. P. Valdes-Herrera, B. D. Simons, P. B. Littlewood, E. Artacho, A. Fert, N. D. Mathur, Nature 2007, 445, 410. [111] V. Dediu, L. E. Hueso, I. Bergenti, A. Riminucci, F. Borgatti, P. Graziosi, C. Newby, F. Casoli, M. P. De Jong, C. Taliani, Y. Zhan, Phys. Rev. B 2008, 78, 115203. [112] Y. Ishii, H. Yamada, H. Sato, H. Akoh, Y. Ogawa, M. Kawasaki, Y. Tokura, Appl. Phys. Lett. 2006, 89, 042509. [113] M. H. Jo, N. D. Mathur, N. K. Todd, M. G. Blamire, Phys. Rev. B 2000, 61, R14905. [114] J. H. Park, E. Vescovo, H. J. Kim, C. Kwon, R. Ramesh, T. Venkatesan, Phys. Rev. Lett. 1998, 81, 1953. [115] M. J. Calderon, L. Brey, F. Guinea, Phys. Rev. B 1999, 60, 6698. [116] J. C. Loudon, S. Cox, A. J. Williams, J. P. Atteld, P. B. Littlewood, P. A. Midgley, N. D. Mathur, Phys. Rev. Lett. 2005, 94, 097202. [117] M. Quintero, P. Levy, A. G. Leyva, M. J. Rozenberg, Phys. Rev. Lett. 2007, 98, 116601. [118] D. Serrate, J. M. De Teresa, M. R. Ibarra, J. Phys.: Condens. Mater 2007, 19, 023201. [119] H. T. Jeng, G. Y. Guo, Phys. Rev. B 2003, 67, 094438. [120] A. Sharma, A. Berenov, J. Rager, W. Branford, Y. Bugoslavsky, L. F. Cohen, J. L. MacManus-Driscoll, Appl. Phys. Lett. 2003, 83, 2384. [121] J. L. MacManus-Driscoll, A. Sharma, Y. Bugoslavsky, W. Branford, L. F. Cohen, M. Wei, Adv. Mater. 2006, 18, 900. [122] L. Balcells, J. Navarro, M. Bibes, A. Roig, B. Martinez, J. Fontcuberta, Appl. Phys. Lett. 2001, 78, 781. [123] J. L. Alonso, L. A. Fernandez, F. Guinea, F. Lesmes, V. Martin-Mayor, Phys. Rev. B 2003, 67, 214423. [124] A. Venimadhav, M. E. Vickers, M. G. Blamire, Solid State Commun. 2004, 130, 631.

[125] X. Z. Liao, A. Sharma, M. Wei, J. L. MacManus-Driscoll, W. Branford, L. F. Cohen, Y. Bugoslavsky, Y. T. Zhu, D. E. Peterson, Y. B. Jiang, H. F. Xu, J. Appl. Phys. 2004, 96, 7747. [126] J. Linden, M. Karppinen, T. Shimada, Y. Yasukawa, H. Yamauchi, Phys. Rev. B 2003, 68, 174415. [127] H. Kato, T. Okuda, Y. Okimoto, Y. Tomioka, Y. Takenoya, A. Ohkubo, M. Kawasaki, Y. Tokura, Appl. Phys. Lett. 2002, 81, 328. [128] J. Rager, M. Zipperle, A. Sharma, J. L. MacManus-Driscoll, J. Am. Ceram. Soc. 2004, 87, 1330. [129] A. Sharma, J. L. MacManus-Driscoll, W. Branford, Y. Bugoslavsky, L. F. Cohen, J. Rager, Appl. Phys. Lett. 2005, 87. [130] J. Rager, A. V. Berenov, L. F. Cohen, W. R. Branford, Y. V. Bugoslavsky, Y. Miyoshi, M. Ardakani, J. L. MacManus-Driscoll, Appl. Phys. Lett. 2002, 81, 112505. [131] J. B. Philipp, P. Majewski, L. Alff, A. Erb, R. Gross, T. Graf, M. S. Brandt, J. Simon, T. Walther, W. Mader, D. Topwal, D. D. Sarma, Phys. Rev. B 2003, 68, 144431. [132] A. Venimadhav, F. Sher, J. P. Atteld, M. G. Blamire, J. Magn. Magn. Mater. 2004, 269, 101. [133] M. Bibes, K. Bouzehouane, A. Barthelemy, M. Besse, S. Fusil, M. Bowen, P. Seneor, J. Carrey, V. Cros, A. Vaures, J. P. Contour, A. Fert, Appl. Phys. Lett. 2003, 83, 2629. [134] M. Fonin, Y. S. Dedkov, R. Pentcheva, U. Rudiger, G. Guntherodt, J. Phys.: Condens. Mater 2007, 19, 315217. [135] H. van Leuken, R. A. Degroot, Phys. Rev. B 1995, 51, 7176. [136] R. J. Soulen, J. M. Byers, M. S. Osofsky, B. Nadgorny, T. Ambrose, S. F. Cheng, P. R. Broussard, C. T. Tanaka, J. Nowak, J. S. Moodera, A. Barry, J. M. D. Coey, Science 1998, 282, 85. [137] J. M. D. Coey, M. Venkatesan, J. Appl. Phys. 2002, 91, 8345. [138] R. S. Keizer, S. T. B. Goennenwein, T. M. Klapwijk, G. Miao, G. Xiao, A. Gupta, Nature 2006, 439, 825. [139] A. Gupta, X. W. Li, G. Xiao, Appl. Phys. Lett. 2001, 78, 1894. [140] H. Toyosaki, T. Fukumura, Y. Yamada, K. Nakajima, T. Chikyow, T. Hasegawa, H. Koinuma, M. Kawasaki, Nat. Mater. 2004, 3, 221. [141] H. Toyosaki, T. Fukumura, K. Ueno, M. Nakano, M. Kawasaki, Jpn. J. Appl. Phys. 2005, 44, L896. [142] J. M. D. Coey, J. Appl. Phys. 2005, 97, 10D313. [143] J. Philip, A. Punnoose, B. I. Kim, K. M. Reddy, S. Layne, J. O. Holmes, B. Satpati, P. R. Leclair, T. S. Santos, J. S. Moodera, Nat. Mater. 2006, 5, 298. [144] M. Wei, N. Braddon, D. Zhi, P. A. Midgley, S. K. Chen, M. G. Blamire, J. L. MacManus-Driscoll, Appl. Phys. Lett. 2005, 86, 072514. [145] G. Herranz, R. Ranchal, M. Bibes, H. Jaffres, E. Jacquet, J. L. Maurice, K. Bouzehouane, F. Wyczisk, E. Tafra, M. Basletic, A. Hamzic, C. Colliex, J. P. Contour, A. Barthelemy, A. Fert, Phys. Rev. Lett. 2006, 96, 027207. [146] N. H. Hong, A. Barla, J. Sakai, N. Q. Huong, Phys. Status Solidi C: Curr. Topics Solid State Phys. 2007, 4, 4461. [147] H. Pan, J. B. Yi, L. Shen, R. Q. Wu, J. H. Yang, J. Y. Lin, Y. P. Feng, J. Ding, L. H. Van, J. H. Yin, Phys. Rev. Lett. 2007, 99, 127201. [148] V. Bhosle, J. Narayan, Appl. Phys. Lett. 2008, 93, 021912. [149] M. Venkatesan, C. B. Fitzgerald, J. G. Lunney, J. M. D. Coey, Phys. Rev. Lett. 2004, 93, 177206. [150] J. R. Neal, A. J. Behan, R. M. Ibrahim, H. J. Blythe, M. Ziese, A. M. Fox, G. A. Gehring, Phys. Rev. Lett. 2006, 96, 197208. [151] Q. Y. Xu, L. Hartmann, S. Q. Zhou, A. McKlich, M. Helm, G. Biehne, H. Hochmuth, M. Lorenz, M. Grundmann, H. Schmidt, Phys. Rev. Lett. 2008, 101, 076601. [152] S. Ramachandran, J. T. Prater, N. Sudhakar, D. Kumar, J. Narayan, Solid State Commun. 2008, 145, 18. [153] C. Song, X. J. Liu, F. Zeng, F. Pan, Appl. Phys. Lett. 2007, 91, 042106. [154] T. Dietl, H. Ohno, F. Matsukura, J. Cibert, D. Ferrand, Science 2000, 287, 1019. [155] H. Ohno, Science 1998, 281, 951. [156] I. Giaever, H. R. Zeller, Phys. Rev. B 1979, 1, 4278.

3838

2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Mater. 2009, 21, 38273839

www.advmat.de

PROGRESS REPORT

[157] M. Julliere, Phys. Lett. A 1975, 54, 225. [158] S. Yuasa, T. Nagahama, A. Fukushima, Y. Suzuki, K. Ando, Nat. Mater. 2004, 3, 868. [159] J. Mathon, A. Umerski, Phys. Rev. B 2001, 63, 220403. [160] J. Z. Sun, Phys. C 2001, 350, 215. [161] M. Bowen, M. Bibes, A. Barthelemy, J. P. Contour, A. Anane, Y. Lemaitre, A. Fert, Appl. Phys. Lett. 2003, 82, 233. [162] X. W. Li, A. Gupta, G. Xiao, W. Qian, V. P. Dravid, Appl. Phys. Lett. 1998, 73, 3282. [163] X. Jiang, R. Wang, R. M. Shelby, R. M. Macfarlane, S. R. Bank, J. S. Harris, S. S. P. Parkin, Phys. Rev. Lett. 2005, 94, 056601. [164] J. S. Moodera, T. S. Santos, T. Nagahama, J. Phys.: Condens. Mater 2007, 19, 165202. [165] U. Luders, A. Barthelemy, M. Bibes, K. Bouzehouane, S. Fusil, E. Jacquet, J. P. Contour, J. F. Bobo, J. Fontcuberta, A. Fert, Adv. Mater. 2006, 18, 1733. [166] A. Ohtomo, H. Y. Hwang, Nature 2004, 427, 423. [167] S. A. Pauli, P. R. Willmott, J. Phys.: Condens. Mater 2008, 20, 264012. [168] K. Janicka, J. P. Velev, E. Y. Tsymbal, J. Appl. Phys. 2007, 103, 07B508. [169] A. Brinkman, M. Huijben, M. Van Zalk, J. Huijben, U. Zeitler, J. C. Maan, G. Van der Wiel, G. Rijnders, D. H. A. Blank, H. Hilgenkamp, Nat. Mater. 2007, 6, 493.

[170] G. Herranz, M. Basletic, M. Bibes, C. Carretero, E. Tafra, E. Jacquet, K. Bouzehouane, C. Deranlot, A. Hamzic, J. M. Broto, A. Barthelemy, A. Fert, Phys. Rev. Lett. 2007, 98, 216803. [171] M. Basletic, J. L. Maurice, C. Carretero, G. Herranz, O. Copie, M. Bibes, E. Jacquet, K. Bouzehouane, S. Fusil, A. Barthelemy, Nat. Mater. 2008, 7, 621. [172] S. Thiel, G. Hammerl, A. Schmehl, C. W. Schneider, J. Mannhart, Science 2006, 313, 1942. [173] N. Reyren, S. Thiel, A. D. Caviglia, L. F. Kourkoutis, G. Hammerl, C. Richter, C. W. Schneider, T. Kopp, A. S. Ruetschi, D. Jaccard, M. Gabay, D. A. Muller, J. M. Triscone, J. Mannhart, Science 2007, 317, 1196. [174] A. D. Caviglia, S. Gariglio, N. Reyren, D. Jaccard, T. Schneider, M. Gabay, S. Thiel, G. Hammerl, J. Mannhart, J.-M. Triscone, Nature 2008, 456, 624. [175] C. Cen, S. Thiel, G. Hammerl, C. W. Schneider, K. E. Andersen, C. S. Hellberg, J. Mannhart, J. Levy, Nat. Mater. 2008, 7, 298. [176] H. Zheng, J. Wang, S. E. Loand, Z. Ma, L. Mohaddes-Ardabili, T. Zhao, L. Salamanca-Riba, S. R. Shinde, S. B. Ogale, F. Bai, D. Viehland, Y. Jia, D. G. Schlom, M. Wuttig, A. Roytburd, R. Ramesh, Science 2004, 303, 661. [177] H. Yang, H. Wang, G. F. Zou, M. Jain, N. A. Suvorova, D. M. Feldmann, P. C. Dowden, R. F. DePaula, J. L. MacManus-Driscoll, A. J. Taylor, Q. X. Jia, Appl. Phys. Lett. 2008, 93, 142904.

Adv. Mater. 2009, 21, 38273839

2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

3839

Das könnte Ihnen auch gefallen