Tokura

Interactions among electronic spins, charges, and orbitalsaccount for a rich variety of patterns in some oxides, and-with the advent of new crystal-growth technologies-mayform the basis for a new type of electronics.Yoshinori Tokuras with any other quantum particle, an electron exhibitsAwavelike and particlelike characteristics. Which as-pect predominates in a solid depends on how an electroninteracts with its neighbors. According to the Bloch theo-rem, for instance, an electron placed in a periodic latticebehaves like an extended plane wave. However, when thenumber of free electrons in a solid becomes comparable tothe number of the constituent atoms and the mutual elcc-tron-elect.ron interaction becomes strong, electrons maylose their mobility.The dual nature is most apparent in correlated-elec-tron systems, such as the transition-metal oxides in whichelectron interactions strongly determine electronic prop-erties. In the transition-metal ions, for example, d elec-trons experience competing forces: Coulombic repulsiontends to localize individual electrons at atomic lattice sites,while hybridization with the oxygenp electron states tendsto delocalize the electrons. The subtle balance makes manyof the transition-met.al oxides excellent resources forstudying and taking advantage of the metal-insulatortransition that can accompany dramatic changes in a sys-tem’s electronic properties.An electron in a solid has three attributes that det.er-mine its behavior: charge c-e), spin (S = i1/2), and orbitalsymmetry. One can imagine an orbital, which representsthe electron’s probability-density distribution, as theshape of an electron cloud in a solid. The charge, spin, andorbital degrees of freedom-and their coupled dynamics-can produce complex phases such as liquid-like, crystal-like, and liquid-crystal-like states of electrons, and phe-nomena such as electronic phase separation and patternformation.‘.”The correlation of electrons in a solid produces a richvariety of states, typically through the interplay betweenmagnetism and electrical conductance. That interplay hasitself been a long-standing research topic among condensedmatter physicists. But since the discovery of copper-oxidehigh-temperature superconductors in 1966, a more generalinterest in the Mott transition-the metal-insulator tran-sition in a correlated-electron system-has been emerging.:’Yoshinori Tokura is a professor of applied physics at the Univer-sity of Tokyo in Japan and a director of the Correlated ElectronResearch Center in Tsukuba, Japan.

The high-T‘ copper oxides are composedof CuO, sheets that are separated fromeach other by ionic “blocking layers.”Although it has one conduction electron(or hole) per Cu site, each CuO, sheetis originally insulating because of thelarge electron correlation. That behav-ior is typical of the Mott insulator state,in which all the conduction electrons

are tied to the atomic sites. The superconducting stateemerges when holes from the blocking layers dope the CuOllayers in a way that alters the number of conduction elec-trons and triggers the Mott transition. Researchers believethat the strong antiferromagnetic correlation, which origi-nates in the Mott-insulating CuO, sheets and persists intothe metallic state, is most responsible for the mechanism ofhigh-q. superconductivity.WIdespread interest incopper oxides and other corre-lated-electron systems during the past 17 years led to a re-discovery of the so-called colossal magnetoresistancetCMR) phenomenon,:-”which is a gigantic decrease of re-sistance induced by application of a magnetic field. Thisarticle features some examples of the dramatic phasechanges in CMR manganites and other transition-metaloxides that arise from a close interplay among electronproperties and their affect on the lattice. Researchers cannow control the electronic and magnetic phases of corre-lated-electron materials in unconventional ways, in somecases with ultrafast response times. Such newly won con-trol offers hope that correlated-electron systems may pro-vide a basis for novel electronics.

Spins, charges, and orbitalsThe quantum mechanical wavefunction of an electronadopts various shapes when bound to an atomic nucleusby the Coulombic force (see the box on page 51). In theMott-insulating state of a crystal, the d electrons are al-most entirely localized on the atomic sites, which makesthe spin and orbital degrees of freedom combine to produceversatile ordering patterns. Prototypical cases for per-ovskite oxides are shown in figure 1 for LaVO 1 (t,, electronsystem) and LaMnO, (c, electron system ).

In LaVO,, for example, the spins order in the z direc-tion as ferromagnetic chains, with all spins aligned paral-lel, but in the my plane as antiferromagnets, with neigh-boring spins antiparallel. That spin configuration inducesan ordered state among the orbitals: d,, and d,, occupy al-ternate lattice sites in every x, y, and z direction. Thespin-orbital ordering makes the electronic structurehighly anisotropic despite the material’s nearly cubic sym-metry. On the other hand, in YVO,, a similar perovskitewith a larger lattice distortion, the staggered spin and or-bital order are just the reverse of the LaVO, case.’

In the manganites, the Jahn-Teller effect, a local de-0 2003 American lnstltute of Physics, S-0031-9228-0307-030-3

50 July 2003 Physics Today

say-is usually much smaller.Consequently, minute pertur-Nickelatc (NilCuprate (Cu)Manganite (Mn)CR&, MO,bations in an input signal canx = ‘I;3x = ‘I8x = ‘Iacontrol a quickly switching

electronic phase output.

The most dramatic exam-ples of such phase tuningoccur in materials in which thetwocompetingelectronicphases form a bicritical point.The manganite Pr,, .),(Ca, ,Sr,),, ,,MnO $,is a case inpoint.“‘The insulating charge-orbital ordered state picturedin figure 2d and the ferromag-netic metallic state competewith each other. Used as a con-trol parameter, the calcium-Ystrontium composition ratioXR,A A4 0subtly alters the lattice distor-tion. The relative stability of

the two phases can

. .be criti-Figure 2. Charges order as stripes In ‘1 v,lrlrty oi dlrrt tionstally tuned, because that distortion governs the cf-electron2nd p,lttcm\\ in holes-doprd two-dImension‘~l met,ll-oxdchopping interaction.,4Abicritical point forms when the crit-shcrts of I,~ycrd pcmvskitc m,~tr~rl,ll. (a) The crystal druc-ical temperatures of the charge-orbital ordering CT,.,,) andturc of thcsc c~x,mplcs. K ,~nd A drc r,3rr-mrth and ,~lkal~ ne-the ferromagnetic transition CT,,.,,) coincide. The phase

cnrth ions, respec tivcly. M drca the trJn\\ition metal Ions NI,change in either direction occurs as an abrupt first-orderCu, ,lnci Mn pictured III the series. (b) In their most stabletransition, and the high-temperature phase above T,.,, andform, the tlopcd holes (hluc circ It+) ,lntl spins (arrow\\) tom

?‘,:>, is subject to gigantic phase fluctuation between the

stripes running along tht‘ dingon,ll in L+ ,Sr,NiO, (x = I/G).competing states. (See figure 3.)(c) In LJ, ,Ha,CuO, ix = l/ri), the stripe5 3rc horimntdl, dc-The colossal magnetoresistancc may be viewed as a

plrted a\\ Oluc met,~llic h,lnd\\. (d) A chc~ckd)oard clur#z-hallmark of the bicritical point. In CMR, a magnetic field

ordering pstttarn tlistingulsho the h,lli-doped st,lte of

aligns the spins on Mn sites, and that alignment influencesLa, \\Sr , ,MnO, (x = I/L), hut the orbIt‘ll\\ order in stripe\he conduction of electrons through adjacent oxygen atoms.that run ,~long the di,~~oml. Bec,~usc of the di,~gonalThis effect transforms the material from a charge-or-orl)it,ll 5tripr dircctlon, the terrom,~jinrtit /igmf: 5pindered/orbital-ordered (CO-00) insulator into a ferromag-c hdlns appear coupled antift~rrom;~~n~~tl~~~lly with thenetic metal, with a dramatic change in resistivity. (See thencighboriny: spin c-hdlns.article by Xeil Mathur and Peter Littlewood in PWSICSTODAY, January 2003, page 25.1 In addition, in what istermed the magnetochromism effect, the magnetic-field-induced transition can also change the optical reflectivity

mechanism of high-temperature superconductivity re-spectrum of the material on an energy scale up to 3 eV.mains controversial, the stripe’s liquid-crystal-like char-The presence of a random potential causes dramatic-and sometimes profound-effects on the material elec-acter undoubtedly determines some of the copper-oxide

tronic properties and may change the phase diagram inmagnetic and electronic properties when the material is

such a bicritical region. That random potential in the tran-underdoped.’

sition-metal compound systems comes from temperature-Even more complex features in the doped manganites

arise when orbital and lattice degrees of freedom are addedindependent static (or “quenched”) disorder, including

local lattice distortion and a random mixture of doped-im-to the mix of spin and charge ordering. Figure 2d illus-purity ions sitting in the perovskite M lattice positions,trates manganite-type configurations for the x = ‘/l hole-and from grain boundaries in the polycrystalline ceramicdoping levels.” The kind of spin, charge, and orbital or-dering displayed is ubiquitous in such highly doped,material. Quenched disorder separates the CO-00 and

nonmetallic manganese oxide perovskites. A checkerboardFM regions into various sizes ranging from nanometer to

micron scales6 An external magnetic field can change thepattern depicts the charge ordering. A combination of the

antiferromagnetic interaction between electrons localizedvolume of the respective phases.

The random potential sometimes enhances the com-to t,, states on neighboring metal atoms and a ferromag-netic interaction created by the electrons hopping alongpetition between the two phases and consequently sup-the lobe of cg states determines the spin ordering. Orbitalpresses the long-range orders. The random potentialordering regulates the anisotropic yX electron hopping. Asarises, for instance, in cases with rare-earth R and alka-a result, ferromagnetic zigzag chains form diagonally inline-earth A ions of relatively different ionic radii. In such

a compound, the correlation in charge and orbital order-the ground state.

ing increases down to T,, (as low as 50 K), without any

Control of the electronic phasetrace of long-range order or phase segregation; just aboveThe energy scale of the dominant interactions that deter-the critical temperature, the external magnetic field sud-mine the electronic phases in correlated-electron systemsdenly suppresses the charge-orbital correlation and pro-is on the order of an electron volt. But the energy differ-duces the FM state, typical of the CMR effect. The exis-ence that distinguishes each of the phases-metallic ver-tence of quenched disorder, therefore, is a prerequisite notsus insulating, or ferromagnetic versus antiferromagnetic,only for the CMR effect but also for any use of the mag-abCd24i-52 July 2003 Physics Today

http://www.physicstoday.org20 K0I___24 6 8 10MAGNETIC FIELD (Ti

_L_:_.Li._-i- . . . . ..12APPLIED VOLTAGE 0’)1uuu-ZUUi0TIME (/a)

Figure 4. Metal-insulator phase control in the colossal-magnetoresistance compound Pr, .Ca~, ,MnO). (a) A small increase rnan external magnetic field will lower the resistivlty sharply, depending on temperature. The material also exhibits hysteresisin the resistivity (by increasing ancl then decreasing the magnetic field). (Adapted tram ref. 5.) (b) A huge abrupt resistance

change (and hysteresis) also accompany voltage changes applied to a circuit connecting the sample in series with a 1 -MI1 re-sistor. (Adapted from ref. 11.) (c) A S-ns laser pulse creates a resistance change of almost 8 orders of magnitude while a volt-age is applied between gold electrodes. The change in surface reflectance at the local metal-insulator transition induced bythe laser pulse shows up as a white streak (circled) between the 0.2-mm distance separating the electrodes (see Inset picture).(Adapted from ref. 12.)to about 200 K, despite Curie temperatures between 330 and370 K.” The widespread speculation is that this seeminglyrapid fade-out of the spin polarization arises from a changein the magnetism of LaSrMnO, in the insulating interface.Efforts to control the interface electronic states of the corre-lated-electron perovskites are now under way in the Corre-lated Electron Research Center in Tsukuba, Japan.The ordered double perovskite family of materials-for example, the Sr&IB,B,O, group, where B, = Fe or Cr andB, = MO or Re-are well suited as TMR-junction devicesbecause, as it turns out, they exhibit robust and high-TVhalf-metallic properties. The transition-metal elements B,and B,> alternately occupy the perovskite lattice sites inthose materials and form a rock-salt pattern. Their ferro-magnetic transition temperature is high, near 420 K forSr,FeMoO,, and as high as 615 K for Sr,CrReO,. Those ad-vantages place perovskites high on the list of possible fu-ture spintronic materials.What about the related possibility of controlling elec-tric current by varying the d-electron orbital states? Onemight label that technology “orbitronics,” analogous as itis to spintronics. As an effect that exploits an orbital cor-relation, CMR may be considered an example. Conductionoccurs through a process known as the double-exchangeinteract:ion. The electrons on each transition-metal ion actlike strongly coupled ferromagnets, so the angle betweenlocal spin moments on adjacent sites determines the elec-tron hopping from site to site. Controlling the magneticfield intensity diminishes the orbital correlation and en-hances electron hopping, and thus becomes tantamount tocontrolling electrical conduction.

One can also regulate the electrical conduction bychanging the orbital shape on a lattice site. For example,when orbitals have d,u ,1 symmetry and are ordered in thenearly cubic perovskite lattice, the charge dynamics are54 July 2003 Physics Today

highly anisotropic: Conduction is entirely confined withinthe xy plane.The notion of ultrafast switching of the orbital stateusing light irradiation or electric field pulses is key in or-bitronics. Because an orbital’s rod or planar shape repre-sents the electron’s probability-density distribution, theorbital naturally couples to the electric field through itsanisotropic polarizability, much as rod- or planar-shapedliquid-crystal molecules respond to an electric field differ-ently by virtue of their distinct polarizabilities. The orbitalwave, or “orbiton,” represents the dynamical response ofthe orbital to an external field, just as a magnon repre-sents a spin wave in the magnetically ordered state. Thatis, once the field excites an orbital at a particular latticesite, the orbital propagates in the solid through the inter-actions between orbitals at different transition-metal ions.Eiji Saitoh and colleagues, using Raman spectroscopy, re-cently observed’”such an orbiton mode in LaMnO:,.Switching frequency is what makes orbital states advan-tageous over spin as a control parameter. Orbiton fre-quencies are typically faster by a factor of lo4 than spin-precession frequencies, and can reach 100 THz.As an example of the modulation of the orbital state,figure 5 gives time-resolved snapshots of a photo-excitedLa,,.,Sr,,,MnO,+ crystal surface,‘” which shows the or-bital-charge-ordered state below To, = 220 K. Because ofthe orbital ordering, an originally tetragonal compound ex-hibits an optical anisotropy in the xy plane. Polarizationmicroscopy, which measures the polarization change be-tween incident and reflected light, provides the contrastbetween orbital-ordered domains.Making materials: The frontier

Researchers are exploring a broader range of electronicmaterials and properties beyond the high-T, and CMR-re-http:Jlwww.physicstoday.orgFigure 5. (a) Charge-orbital ordering patterns on themanganite (MnO,) sheet in hole-doped La,,,;%, ,MnO,.The colored arrows distinguish identical (blue) and al-ternating (red) orbitals; charges are identical along bothdiagonals. (b) A polarization microscope image of this

material shows contrast between chargearbital or-dered states (bright) and and their domain boundaries(dark) at 77 K. The slight residual strain introducedduring crystal growth accounts for the periodic domainstructure. (c) When a pulsed laser, having an energy of1.5 eV and a duration of 100 is in this experiment, ir-radiates the surface, the electronic transition of elec-trons across the optical gap destroys the orbital orderwithin 200 is. That photomelting of the orbital statecreates the dark contrast on the left. The disorderedstate persists for about 10 ns because of the stabilizinginfluence of the lattice deformation and then disap-pears through thermal diffusion to its original fully or-bital-ordered state. (Adapted from ref. 15.)

a

Mn3+0

Mn4+blated ones showcased in this article. The recent ad-vance in epitaxial-growth technology- applied totransition-metal oxide thin films, for instance, hasbegun to produce correlated-electron junctions andsuperlattices with well-controlled interface charac-teristics. Magnetic-oxide superlattices-composed of

cferromagnetic, antiferromagnetic, and paramag-netic perovskite layers separated by just a few unitcells-are one example. At the interfaces of thosefilms, the spins, orbital states, and charges aregreatly modified. The interfaces may themselves besubject to external magnetic or electric fields as well.The additional sensitivity between layers highlightsthe importance of exotic materials that possess par-ticular interface properties.

A superlattice prepared from sequential epitax-ial layers made from different building blocks shouldlack any inversion symmetry. That property would

add a different twist on the kinds of designer films re-searchers might use. Choosing one of the layers to be fer-romagnetic, for example, might make the film a polar fer-romagnetwith new,intriguingmagnetoelectronicproperties already built in. One would expect nonlinear ornonreciprocal magneto-optical effects to emerge because ofsimultaneous breaking of space-inversion and time-rever-sal symmetries. Those effects may find applications in op-tical-fiber communication and laser-diode technology. Withsuch technologies and crystal-growth techniques now be-coming available to scientists, the complex phases that canoccur in tailor-made correlated-electron materials makethis research field a fascinating and challenging arena fortesting theories and for spawning new electronics.

Thanks to Naoto Nagaosa, Elbio Dagotto, Peter Littlewood,Ken Miyano, Tsuyoshi Kimura, and Yoichi Okimoto for en-lightening discussion and help in preparing the article.

E',,,,~ E,,,lOO/*m7.8.9.10.11.12.13.14.15.References1. S. A. Kivelson, E. Fradkin, V. J. Emery, Nature 393, 550(1998).2. Y. Tokura, N. Nagaosa, Science 288, 389 r2000).http:Nwww.physlcstoday.org(1998).A. J. Millis. Nature 392, 147 (1998).

Y. Tokura, ed.. Colossal Magnetoresistive Oxides. Gordon &Breach, London (2000).

E. Dagotto, Nanoscale Phase Separation and Colossal Mag-netoresistance: The Physics of Manganites and Related Com-pounds, Springer-Verlag, New York (2002).

H. Sawada. N. Hamada. K. Terakura. K. T. Asada.” Phvs. Rev.B 53, 12742 11996,.C. H. Chen. S.-W. Cheong, A. S. Cooper, Phw. Rev. Lett. 71,2461 (1993).J. M. Tranquada, B. J. Sternliev, J. D. Axe, Y. Nakamura:S. Uchida, Nature 375, 561 (1995).

Y. Tomioka. Y. Tokura, Phys. Rev. B 66, 104416 (20021.A. Asamitsu, Y. Tomioka, H. Kuwahara, Y. Tokura, Nature388, 50 (1997).

K. Miyano, K. T. Tanaka, Y. Tomioka, Y. Tokura, Phys. Rev.Lett. 78. 4257 (1997): M. Fiebie. K. Mivano. Y. Tomioka.Y. Tokura, Science 286, 1925 (1998).”S. Q. Liu, N. J. Wu, A. Ignatiev, Appl. Phys. Lett. 76, 2749(2000).E. Saitoh et al., Nature 410, 180 (2001).

T. Ogasawara, T. Kimura, T. Ishikawa, M. Kuwata-nGonokami, Y. Tokura, Phys. Rev. I3 63,113105 (2001).July 2003 Physics

M. Imda, A. Fujimori, Y. Tokura, Rev. Mod. Phys. 70, 1039Today 55