Magnetoelectric Effects in Manganites

MISSING IMAGE

Material Information

Title:
Magnetoelectric Effects in Manganites
Physical Description:
1 online resource (101 p.)
Language:
english
Creator:
Jeen,Hyoung Jeen
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Physics
Committee Chair:
Biswas, Amlan
Committee Members:
Dorsey, Alan T
Lee, Yoonseok
Hebard, Arthur F
Norton, David P

Subjects

Subjects / Keywords:
cer -- cmr -- magnetoelectric -- manganites -- multiferroics -- pld
Physics -- Dissertations, Academic -- UF
Genre:
Physics thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Research on manganites has been conducted for more than half century. Recent discoveries of colossal responses to external fields such as colossal magnetoresistance effects and correlation among spin, orbital, and lattice in phase separated manganites and multiferroic manganites have motivated me to understand these materials. The main purpose of this dissertation is to understand magnetoelectric effects in phase separated (La$_{1-y}$Pr$_{y}$)$_{1-x}$Ca$_{x}$MnO$_{3}$ (LPCMO) thin films and multiferroic BiMnO$_{3}$ (BMO) thin films. First, high quality phase separated manganite thin films have been successfully grown. To grow the high quality manganite thin films, extensive effort was devoted to fine tuning of oxygen pressure, temperature, and laser fluence during film growth. As-grown films were characterized with various ex-situ techniques: magnetization measurements, transport measurements, x-ray diffraction, atomic force microscopy, and/or transmission electron microscopy to remove the effects of impurities and unwanted strains except substrate induced strain. Second, three major results were obtained in high quality phase separated LPCMO thin films. These results are based on the dynamic nature of phases in LPCMO. 1) LPCMO thin films showed single domain to multi-domain transition during cooling. This transition can be tuned by substrate stress induced in-plane magnetic anisotropy. 2) Evidence for the origin of colossal electroresistance (CER) effect has been observed. The CER is triggered by dielectrophoresis, or movements of ferromagnetic metallic (FMM) phase, which is manifested in anisotropic transport properties in microfabricated LPCMO cross structures. This fluidic nature of the FMM phase in LPCMO under high electric fields lead to exotic magnetoelectric effects. 3) Electric field effects on magnetotransport properties have been observed. This phenomena can also be tuned by the combined effect of substrate strain and current flow. This combined effect of electric and magnetic fields and strain at the interface of LPCMO suggest new ways to control magnetism (magnetotransport) with electric fields. Third, impurity-free and epitaxial BMO thin films were grown. These films showed ferromagnetism and ferroelectricity at low temperature. Magnetoelectric effects are discussed, especially magnetization change due to electric fields.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Hyoung Jeen Jeen.
Thesis:
Thesis (Ph.D.)--University of Florida, 2011.
Local:
Adviser: Biswas, Amlan.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-02-29

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2011
System ID:
UFE0042870:00001


This item is only available as the following downloads:


Full Text

PAGE 1

MAGNETOELECTRICEFFECTSINMANGANITESByHYOUNGJEENJEENADISSERTATIONPRESENTEDTOTHEGRADUATESCHOOLOFTHEUNIVERSITYOFFLORIDAINPARTIALFULFILLMENTOFTHEREQUIREMENTSFORTHEDEGREEOFDOCTOROFPHILOSOPHYUNIVERSITYOFFLORIDA2011

PAGE 2

c2011HyoungJeenJeen 2

PAGE 3

Tomyfamily 3

PAGE 4

ACKNOWLEDGMENTS AfterIstartedmygraduateschoolinUniversityofFlorida(UF),Ihavealotofchancestogethelpsandadvicesfrommanypeople.IfeellikeIhavelotsofdebtsfromUF.Ithinkthebestwaytopaybackthesehelpsistospreadmyknowledgeandideatoanybody,whowantstolearn.Igreatlyappreciatedmyadvisor,Dr.AmlanBiswas.Heistheadvisornotonlyformyresearchbutalsoformylife.Ialsoappreciatedtocommitteemembers:Dr.ArthurF.Hebard,Dr.AlanDorsey,Dr.YoonseokLee,andDr.DavidNorton.Dr.Hebardalwaysencouragesmetoworkcreativelyinmanganiteresearch.Heiswillingtoanswerallmyquestionsandsometimesgavemesomequestionsformetoanswers.Dr.AlanDorseyiswillingtokeephimselfascommitteemember.Heisalwayseagertodiscusswithme,evenifhisscheduleisbusy.Dr.YoonseokLeeisthementor,evenifhewasnotmydegreecommitteebefore2008.Hehasclearlyansweredtoallmyquestionsonanytopic.Hesometimesdoesnothesitatetopointoutmybadhabits/attitudes,soIcanxthose.Dr.DavidNorton'sencouragement,especiallyinaprofessionalconference,mademecondent.Iappreciatedmyformercommitteemembers:Dr.Hagen,Dr.Long,andDr.Petkova.IamgratefultostaffsinPhysicsdepartment:Dareleneforallthepaperworksandadvices,GregandJohnforconstantsupplyofLiquidHelium,CharlesandBobforgreathelpinlabteaching,TimforwellmaintainingourNewPhysicsBuilding,Marcformachineshopwork,andPamforallthepaperworks.IalsothankstoLaurie,Billie,andJanet.Ialsothankstomylabmates,basementfriends,allotherfriendsImetinGainesville.Ithankyoutostaffs(Dr.Craciun,Dr.Bourne,Dr.Siebein,andBillLewis)inmajoranalyticalinstrumentationcenterandnanoscienceinstituteformedicalandengineeringtechnology,UF.IalsoappreciatedtoDr.Sang-WookCheong(Rutgers)andDr.Ho-NyungLee(OakRidgeNationalLab)abouttheiradvicesonmanagniteresearch. 4

PAGE 5

Amongmycollegeprofessors,IspeciallythanktoDr.Kyung-SooYi,Dr.YunchulChung,Dr.Ho-SoonYang,andDr.Jai-SeokAhntotheirconstantadvices.ThisresearchhasbeensupportedfromNSFDMR-0804452.IamalsogratefulforgraduatestudyabroadprogramfromKoreanscienceandengineeringfoundation(KRF-2005-215-C00043)andgraduatestudentcounciltravelgrantsandcollegeofliberalartsandsciencesandphysicsdepartmenttravelawards. 5

PAGE 6

TABLEOFCONTENTS page ACKNOWLEDGMENTS .................................. 4 LISTOFTABLES ...................................... 8 LISTOFFIGURES ..................................... 9 ABSTRACT ......................................... 11 CHAPTER 1INTRODUCTION ................................... 13 1.1GeneralIntroduction .............................. 13 1.2MagnetoelectricEffects ............................ 14 1.3BasicsofManganites ............................. 15 1.3.1StructureofManganites ........................ 15 1.3.2ElectronicCongurationofMnIons .................. 16 1.4HoleDopedmanganites ............................ 18 1.4.1TransportMechanisminMixedValenceManganites. ........ 18 1.4.2PhaseCoexistenceinHoleDopedManganitesand(La,Pr,Ca)MnO3(LPCMO) ................................ 19 1.5MultiferroismandBiMnO3(BMO) ....................... 21 2EXPERIMENTALMETHODS ............................ 24 2.1ThinFilmGrowth ................................ 24 2.1.1SubstrateSurfaceTreatment ..................... 24 2.1.2PulsedLaserDeposition ........................ 25 2.2TransportMeasurements ........................... 27 2.2.1ConstantCurrentSourcedMeasurements .............. 27 2.2.2ConstantVoltageSourcedMeasurements .............. 29 2.3MagnetizationMeasurements ......................... 30 2.4SurfaceandStructuralPropertyMeasurements ............... 31 3EFFECTSOFOXYGENPRESSUREONPHYSICALPROPERTIESOFLPCMOTHINFILMS ..................................... 33 3.1ExperimentalDetails .............................. 34 3.2ResultsandDiscussion ............................ 35 3.2.1LPCMOonUntreatedNdGaO3(NGO) ................ 35 3.2.2LPCMOonTreatedNGO ....................... 39 3.3Summary .................................... 40 6

PAGE 7

4SINGLEDOMAINTOMULTI-DOMAINTRANSITIONDUETOIN-PLANEMAGNETICANISOTROPYINPHASESEPARATEDLPCMOTHINFILMS .......... 42 4.1ExperimentalDetails .............................. 43 4.2ResultsandDiscussion ............................ 45 4.2.1StructureandTransport ........................ 45 4.2.2MagneticProperties .......................... 47 4.2.3VariationofCoerciveFieldwithTemperature ............. 49 4.2.4StrainInducedMagneticAnisotropy ................. 54 4.3Summary .................................... 58 5ANISOTROPICTRANSPORTINLPCMOTHINFILMS ............. 60 5.1LPCMO(y=0.6)ThinFilms ......................... 61 5.2ElectricFieldInducedAnisotropicTransport ................. 64 5.3EffectsofElectricFieldsandStrainsonMagnetotransport ......... 69 5.3.1CurrentEffectsonMagnetotransport ................. 72 5.3.2StrainEffectsonMagnetotransport .................. 73 5.4Summary .................................... 76 6OPTIMIZATIONOFBMOTHINFILMSANDPHYSICALCHARACTERIZATION 77 6.1OptimizationofBMO .............................. 78 6.2StructuralStudy ................................ 86 6.3MagnetoelectricEffects ............................ 89 6.4Summary .................................... 90 7CONCLUSIONANDFUTUREWORK ....................... 92 REFERENCES ....................................... 94 BIOGRAPHICALSKETCH ................................ 101 7

PAGE 8

LISTOFTABLES Table page 2-1Oxidesubstrates ................................... 24 2-2Filmgrowthconditionofmanganitethinlms ................... 27 5-1IdealcompositionandratioofLPCMO ....................... 63 5-2CompositionofLPCMOatthemiddleofthethinlms .............. 63 5-3CompositionofLPCMOneartheinterface ..................... 64 8

PAGE 9

LISTOFFIGURES Figure page 1-1Perovskitestructure ................................. 16 1-2EnergydiagramofamanganeseioninMnO6octahedra ............. 17 1-3Trendsof(T)inLPCMOwithvaryingPrcontents ................ 19 1-4Ferroelectricityduetoemptyd-shell ........................ 22 2-1Atomicforcemicroscopy(AFM)imagesoftreatedsubstrates .......... 25 2-2PLDset-upandplumes ............................... 26 2-3Schematicdiagramsoftransportmeasurements ................. 28 2-4Magnetoelectriceffectmeasurementset-up. .................... 31 3-1AFMimagesofLPCMOthinlmsonuntreatedsubstrates ............ 34 3-2(T),Max.TCR,andTPofLPCMOthinlmsonuntreatedsubstrates ..... 35 3-3M(H)ofLPCMOthinlmsonuntreatedsubstrates ................ 37 3-4XRDpatternsofLPCMOthinlmsonuntreatedsubstratesandNGO. ..... 37 3-5ActivationenergyofLPCMOthinlms ....................... 38 3-6AFMimagesofLPCMOthinlmsontreatedsubstrates ............. 39 3-7(T),Max.TCR,andTPofLPCMOthinlmsontreatedsubstrates ...... 40 4-1(T)andAFMimagesofLPCMOthinlms .................... 46 4-2XRDresultsofLPCMOthinlms .......................... 47 4-3M(T)ofLPCMOthinlms ............................. 48 4-4In-planeZFCmagnetizationhysteresisloopsofLPCMOthinlmonSLGO .. 49 4-5In-planeZFCmagnetizationhysteresisloopsofLPCMOthinlmonNGO ... 51 4-6Hc(T)andMr(T)ofLPCMOthinlms ....................... 52 4-7M(H)oftheLPCMOthinlmonNGOandatomicstructureofNGO ...... 53 4-8M(H)withHalonghardaxis,Ha(T),Hc(T),andKu(T)ofLPCMOthinlmsonNGO ........................................ 56 4-9AngledependentHc(T)andH)]TJ /F3 7.97 Tf 6.59 0 Td[(1c(cos())ofLPCMOthinlmonNGO ..... 57 5-1SEMimageofLPCMOandTEMimageofLPCMO(300 ........... 61 9

PAGE 10

5-2TEMimageofLPCMO(600),SADofLPCMOandNGO,andFFTofLPCMO 62 5-3EDXandXRRofLPCMO .............................. 63 5-4ProposedmodelsforCEReffect .......................... 65 5-5OpticalmicroscopyimageofLPCMOandsequentialtransportmeasurementssetup ......................................... 66 5-6ZeroeldR(T)incrossstructureandCEReffects ................ 67 5-7IsothermalR(t)measurements ........................... 69 5-8IsothermalsequentialR(V)measurements .................... 70 5-9ZeroeldR(T)ofLPCMOandschematicdiagramsofdifferentMR ....... 71 5-10CurrenteffectsonmagnetotransportandMER .................. 72 5-11R(H)inthreedifferentmagneticelddirectionsat60K .............. 73 5-12dR/dTofR(T)under8TandthreedifferentAMRasafunctionoftemperature 74 6-1ThemonoclinicunitcellofBiMnO3along(111)and(203)directions ...... 79 6-2)]TJ /F6 11.955 Tf 11.18 0 Td[(2diffractionofBMOthinlmsonSTOgrownindifferentoxygenpressureordifferenttemperature ............................... 80 6-3)]TJ /F6 11.955 Tf 11.96 0 Td[(2diffractionofBMOthinlmsonSTOindifferentcoolingrate ....... 81 6-4M(T)ofBMOthinlms ................................ 82 6-5M(H)ofBMOthinlms ................................ 83 6-6FerroelectricityofaBiMnO3thinlmonSrTiO3 .................. 85 6-7)]TJ /F6 11.955 Tf 12.46 0 Td[(2diffractionpatternandGIXDofaBMOthinlmonSTO.Inset:!scanof(111)BMOpeak. ................................. 87 6-8(a)Polegureoff211gSTOandf110gBMO(b)ReciprocalspacemapoftheBMOthinlmonSTO. ................................ 88 6-9X-rayreectivityofBMOthinlmsonSLGOandSTO.Inset:AnAFMimageoftheBMOonSLGO. ................................ 89 6-10)]TJ /F6 11.955 Tf 9.72 0 Td[(2andGIXDofBMOthinlmsonSTOandSLGO.-scanoftheBMOlmandSLGO ...................................... 89 6-11M(T),R(T),andmagnetoelectriceffectsofBMOonSTO ............. 91 10

PAGE 11

AbstractofDissertationPresentedtotheGraduateSchooloftheUniversityofFloridainPartialFulllmentoftheRequirementsfortheDegreeofDoctorofPhilosophyMAGNETOELECTRICEFFECTSINMANGANITESByHyoungJeenJeenAugust2011Chair:AmlanBiswasMajor:PhysicsResearchonmanganiteshasbeenconductedformorethanhalfcentury.Recentdiscoveriesofcolossalresponsestoexternaleldssuchascolossalmagnetoresistanceeffectsandcorrelationamongspin,orbital,andlatticeinphaseseparatedmanganitesandmultiferroicmanganiteshavemotivatedmetounderstandthesematerials.Themainpurposeofthisdissertationistounderstandmagnetoelectriceffectsinphaseseparated(La1)]TJ /F8 7.97 Tf 6.58 0 Td[(yPry)1)]TJ /F8 7.97 Tf 6.58 0 Td[(xCaxMnO3(LPCMO)thinlmsandmultiferroicBiMnO3(BMO)thinlms.First,highqualityphaseseparatedmanganitethinlmshavebeensuccessfullygrown.Togrowthehighqualitymanganitethinlms,extensiveeffortwasdevotedtonetuningofoxygenpressure,temperature,andlaseruenceduringlmgrowth.As-grownlmswerecharacterizedwithvariousex-situtechniques:magnetizationmeasurements,transportmeasurements,x-raydiffraction,atomicforcemicroscopy,and/ortransmissionelectronmicroscopytoremovetheeffectsofimpuritiesandunwantedstrainsexceptsubstrateinducedstrain.Second,threemajorresultswereobtainedinhighqualityphaseseparatedLPCMOthinlms.TheseresultsarebasedonthedynamicnatureofphasesinLPCMO.1)LPCMOthinlmsshowedsingledomaintomulti-domaintransitionduringcooling.Thistransitioncanbetunedbysubstratestressinducedin-planemagneticanisotropy.2)Evidencefortheoriginofcolossalelectroresistance(CER)effecthasbeenobserved.TheCERistriggeredbydielectrophoresis,ormovementsofferromagneticmetallic(FMM)phase,whichismanifestedinanisotropic 11

PAGE 12

transportpropertiesinmicrofabricatedLPCMOcrossstructures.ThisuidicnatureoftheFMMphaseinLPCMOunderhighelectriceldsleadtoexoticmagnetoelectriceffects.3)Electriceldeffectsonmagnetotransportpropertieshavebeenobserved.Thisphenomenacanalsobetunedbythecombinedeffectofsubstratestrainandcurrentow.ThiscombinedeffectofelectricandmagneticeldsandstrainattheinterfaceofLPCMOsuggestnewwaystocontrolmagnetism(magnetotransport)withelectricelds.Third,impurity-freeandepitaxialBMOthinlmsweregrown.Theselmsshowedferromagnetismandferroelectricityatlowtemperature.Magnetoelectriceffectsarediscussed,especiallymagnetizationchangeduetoelectricelds. 12

PAGE 13

CHAPTER1INTRODUCTION 1.1GeneralIntroductionPerovskitetypemanganeseoxides(manganites)haveinducedrenewedinterestsforthelasttwodecadesduetotheircolossalresponsestoexternaleldssuchascolossalmagneto-resistance(CMR),colossalelectro-resistance(CER),photo-inducedmetal-insulatortransition,andcolossalpiezo-resistance(CPR)effects.[ 1 5 ]Theseresponsesarerelatedtothecorrelationamonglattice,spin,andorbitals.Thiscorrelationcreatesmultiplegroundstates,viz.ferromagneticinsulator,ferromagneticmetal,antiferromagneticchargeorderedinsulator,andparamagneticinsulatorasafunctionofchemicaldoping,temperature,magneticelds,electricelds,andstrainelds.Currently,theoriginofthesecolossalbehaviorshasbeenexplainedbyphasecompetition(orphaseseparation)amongdifferentphaseswithsimilarfreeenergies[ 6 ].Inotherwords,controlofphaseswithexternalvariablesmakesitpossibletotune/modifythecolossalbehaviors.Thispointattractsscientiststotrytounderstandandmakeuseofthesepropertiesforpossibledeviceapplications,e.g.bolometers,aresistivememory,andamagneticmemory[ 7 8 ].Recently,anotheraspectinmanganites,multiferroisminmanganites,hasbeenhighlightedduetotheresurgenceofinterestinmagnetoelectriceffects,wheremagnetic(electric)polarizationiscontrolledbyelectric(magnetic)elds[ 9 11 ].Thisresurgenceisnotsurprising,sincewetakeadvantageofmagnetismandelectricityinoureverydaylife.Forexample,computerharddrivesusegiantmagneto-resistanceeffect,whereelectronictransportacrossferromagnetic(FM)metal/non-magneticmetal/FMmetaltrilayersdependsonrelativemagnetizationdirectionsofFMlayers.Also,ferroelectricrandomaccessmemorymakesuseofferroelectricpolarizationtocreatenon-volatilestates.However,thesedevicescanonlyperformbinaryoperations,sincejusttwodegreesoffreedomexist.Iftheferromagnetisminamaterialcanbecontrollednotonly 13

PAGE 14

bymagneticeldsbutalsobyelectricelds,thatmaterialcanproducefourdifferentstates[ 12 ].Thispropertywillbeadvantageoustominimizedevicesizeandincreasedegreesofintegration.Ithasbeenpredictedthatthemagnetoelectriccouplingwilldependonbothmagneticandelectricsusceptibilities[ 9 ].Thus,astrongcandidateforsuchmagnetoelectriceffectisamaterialthatisbothferromagneticandferroelectricmaterial,whichisatypeofmultiferroics[ 13 14 ].Inthisdissertation,twomanganites,viz.phaseseparated(La1)]TJ /F8 7.97 Tf 6.59 0 Td[(yPry)0.67Ca0.33MnO3andmultiferroicBiMnO3,wereinvestigatedtondmagnetoelectriceffects.Inthefollowingsections,magnetoelectriceffects,basicsofmanganites,phaseseparation,andmultiferroismaredescribed. 1.2MagnetoelectricEffectsMagnetoelectriceffectisdenedasmagnetization(electricpolarization)changebyelectric(magnetic)elds.Thiseffectcanbederivedfromderivationofthefreeenergyofmaterials.Freeenergyofamaterialcanbeexpandedasafunctionofelectriceld(E)andmagneticeld(H)[ 9 15 ]. F(E,H)=0ij 2EiEj+0ij 2HiHj+ijEiHj+,(1) Pi(E,H)=PSi+0ijEj+ijHj+,(1) Mi(E,H)=MSi+0ijEj+ijHj+,(1)Firstderivativeofthisfreeenergy(F)toelectriceld(E)istheelectricpolarization(Pi).Itszerothordertermisthespontaneouspolarization,whichiscommonlyfoundinferroelectricmaterials.Betweenthetworstorderterms,one,wherethepolarizationislinearlydependentonthemagneticeld,iscalledthelinearmagnetoelectriceffect.Intherstderivativeofthefreeenergytomagneticeld(H),theterm,wherethemagneticpolarization(Mi)islinearlydependentontheelectriceld,isalsocalledthemagnetoelectriceffectterm.Theconstant()iscalledaslinearmagnetoelectric 14

PAGE 15

couplingconstant[ 9 ].Itisnotedthathigherordermagnetoelectriceffectsexist.However,inthisdissertationmagnetoelectriceffectwillrefertotheterm,.Sincemagnetoelectriceffectislimitedbymagneticandelectricsusceptibilities,itisexpectedthatsimultaneouslyferromagneticandferroelectricmaterialswillbethebestoptiontogetthemaximalmagnetoelectriceffects[ 9 ].However,themagnetoelectriceffectinmultiferroicmaterialsisnotverylarge.ItwasonlyreportedsmallmagnetodielectriceffectinbulkBiMnO3[ 16 ].Differentoriginsinferroelectricityandferromagnetismwouldreducethemagnetoelectriccoupling[ 17 ].Here,exoticmagnetoelectriceffectsinphaseseparatedmanganiteswillbeintroduced,whichisrelatedtochangeintheirdielectricconstantnearthemetal-insulatortransition[ 9 ]and/orcontrolofmagneticphasesbyelectriceld.Theword'exotic'meansitdoesnotincludesmagnetoelectriccouplingintheusualsenseasdescribedearlier,evenifitismanifestedfrommagnetizationchangesunderhighelectriceldsinsomesystem.Forexample,thiseffectisrelatedtotheelectriceldinducedinsulatortometaltransition(colossalelectroresistance),sincemanganiteisasystem,wherecharge,spin,andlatticearecorrelated[ 2 ].PossiblemagnetoelectriceffectsinphaseseparatedLPCMOthinlmwereinvestigatedwhichcouldalsorevealtheoriginofcolossalelectroresistanceeffects. 1.3BasicsofManganites 1.3.1StructureofManganitesThemanganiteshavethechemicalformulaABO3,wheretheAsiteisusuallyoccupiedbyrareearthionslikeLanthanum(La),Praseodymium(Pr),andNeodymium(Nd)andtheBsiteisoccupiedbyaManganeseion(Mn).Thecompoundusuallyformsintheperovskitestructure(Figure1-1).WhentheA-sitecationbecomessmalle.g.forthecationYttrium(Y),manganitescanhaveadifferentcrystalstructure,e.g.ahexagonalstructure.Inperovskitestructure,rareearthionsareatthecenterofcube,Mnionsareatthecornerofthecube,andoxygenionsarelocatedinbetweenMnions. 15

PAGE 16

Figure1-1. Perovskitestructure[ 18 ] TheseMnionsaresurroundedbysixoxygenionsforminganMnO6octahedron.ThismeanselectronicenergylevelsofMnionsarealsogovernedbycoulombinteractionfromoxygenions,whichisknownasthecrystaleldeffect. 1.3.2ElectronicCongurationofMnIonsWhenanMnionisinfreespace,theenergyof3dorbitalsaredegenerate.However,whentheMnionissurroundedbysixoxygenions(twoineachaxis),theseoxygensprovideacrystaleldsothatthed-orbitalswillbesplitintotwoenergylevels:t2glevelconsistsofthreed-orbitals(dxy,dyz,anddzx),whichhavelowoverlapwith2porbitalsofoxygen,whileeglevelconsistoftwod-orbitals(dx2)]TJ /F8 7.97 Tf 6.59 0 Td[(y2andd3z2)]TJ /F8 7.97 Tf 6.58 0 Td[(r2)whichhavehigherenergythanthet2gorbitalsduetohigheroverlapwiththeoxygen2porbitals.Thiseffectisknownascrystaleldsplitting(Figure1-2).WhenMnionshavevalencyof3+,byHundruleoneelectronisineglevelwhilethreeelectronsareinthet2glevelwithsamespindirection.Jahn-Tellereffecttakesplaceintheoctahedrontominimizeenergybydistortingtheoctahedron.Thisdistortioncausesmoreenergysplittinginbothegandt2glevels.Iftheelongationisalongthez-axis,anelectronineglevelwillbeind3z2)]TJ /F8 7.97 Tf 6.58 0 Td[(r2ratherthandx2)]TJ /F8 7.97 Tf 6.59 0 Td[(y2duetolesscoulombinteractionfromoxygenions.Degreesof 16

PAGE 17

Figure1-2. EnergydiagramofamanganeseioninMnO6octahedra.Manganeseorbitalisred,andoxygenorbitalsaregrey. Jahn-TellerdistortioncanbecontrolledbysizeofAsitecationandchemicaldopingintheidealcaseofdefect-freesamples.WhenA-siteisdopedwithdivalentionssuchascalcium,strontium,orbarium,valencyofMnionsareinbetween3+and4+.LotsofinterestingphenomenahavebeenobservedduetothisvariationinvalencyoftheMn-ions.Forexample,eventhoughitsparentcompounds,LaMnO3andCaMnO3areA-typeandG-typeantiferromagneticinsulators,La1)]TJ /F8 7.97 Tf 6.59 0 Td[(xCaxMnO3incertaindopinglevelshowsmetallicbehaviorandferromagnetismatlowtemperature[ 19 ].Toexplainthistransportphenomenainholedopedmanganites,twomicroscopicmechanismswillbeintroduced,viz.doubleexchangeinteractionatlowtemperatureandpolaronformationathightemperature. 17

PAGE 18

1.4HoleDopedmanganites 1.4.1TransportMechanisminMixedValenceManganites.Transportpropertyandmagneticpropertyinhole-dopedmanganitesarecloselyrelated.Doubleexchangeinteractioncanqualitativelyexplaintransportpropertyofhole-dopemanganites.Thisinteractionisthehoppingofegelectronsinmanganitestoemptyegorbitalviaoxygenions.Atransferredelectronshouldhavethesamespindirectionasthet2gelectronsinMn4+,sinceemptyegorbitalstatesarestronglygovernedbyt2gelectronsbyHund'scoupling.Spinalignmentandtransfermechanismmeanthatferromagneticandmetallicbehaviorscometogether.Followingequation,proposedbyZener,explainsincreaseofconductivity(),whentemperature(T)islowerthanaCurietemperature(TC)[ 20 21 ]. xe2 ahTC T,(1)whereaisalatticeconstant,xisholeconcentration,andhisPlanck'sconstant.DoubleexchangemechanisminitiallyexplainslotsofpuzzlesofCMRbehaviorandlowtemperatureconductionphenomena.However,hightemperatureresistancebehavior,especiallyabovemagneticCurietemperature,cannotbeexplainedsolelybydoubleexchangemechanism.Anotherimportanteffectistheelectron-phononcouplingorpolaronformation.Apolaronisdenedascombinationoftheelectronanditsstraineldinaninsulator[ 22 ].LocalizationofchargecarriersoccursduetoJahn-TellerdistortinofMn3+sitesinmixedvalencemanganites[ 21 ].Polaronformationexplainsthehightemperatureresistivityofmixedvalencemanganites,toacertainextent.Twopolaronmodelsareusuallyusedforexplaininghightemperaturebehaviorinmanganites.Thesmallpolaronmodelshowsfollowingrelationinresistivity()asafunctionoftemperature(T),(T). (T)/TExpEA kBT,(1) 18

PAGE 19

Figure1-3. SchematicplotofresistivityasafunctionoftemperatureinLPCMOwithdifferentPrcontents. whereEAisanactivationenergy,andkBisBoltzmann'sconstant.Themodelpredicts=1foranadiabaticcaseinwhichchargecarriermotionisfasterthanmotionoflatticeand=1.5foranon-adiabaticcase.Thehightemperaturebehaviorofmixedvalencemanganitesaregenerallywellexplainedbythissmallpolaronmodel[ 21 23 ].Anotherproposedmodelisthevariablerangehopping,whichpredictsisproportionaltoExp(1 T)0.25intransportmeasurements[ 24 ]. 1.4.2PhaseCoexistenceinHoleDopedManganitesand(La,Pr,Ca)MnO3(LPCMO)Theoptimallydopedmanganitesusuallyshowmetal-insulatortransition(MIT)with/withoutexternalelds.TheMITbehaviorhasbeeninterpretedasaresultofelectronic,magnetic,andpossiblystructuralphaseevolution.CoexistenceofdifferentphaseshasbeenobservedwithvarioustechniquesnearMITregionandbelowMITtemperature.[ 8 25 27 ].Theoretically,ithasbeenshownthatquencheddisorderand 19

PAGE 20

long-rangestraineldscancreatenano-ormicro-meterscalephaseseparation[ 6 28 29 ].Electronicphasecompetitionsuchascompetitionbetweendoubleexchangeinteractionandpolaronformationinthelongrangecoulombrepulsioncanexplainnano-meter-sizedphaseseparation[ 30 ].Thisideahasbeenre-consideredafterdiscoveryofmicro-meterscalephaseseparationinphaseseparatedmanganites,(La,Pr,Ca)MnO3system.Asalternatives,acombinationofquencheddisorderandstraineldduetodifferenceinstructurebetweeninsulatorandmetalprovidesareasonabletheoreticalbasisforsuchmicrometerscalephaseseparation.LPCMOisderivedfromLa0.67Ca0.33MnO3(LCMO)andPr0.67Ca0.33MnO3(PCMO).LCMOexhibitsferromagnetismandmetallicbehaviorbelowabout250K,whilePCMOshowsantiferromagnetismandinsulatingbehavior,sincePrionintheA-sitecreatesmorepronouncedJahn-Tellerdistortion.Additionally,PCMOshowschargeorderingamongMn3+andMn4+ionsbelowabout200K.LPCMOcanbethoughtasacombinationofthesetwocompounds.ForacertainratiobetweenLaandPrions,evenbelowinsulator-metaltransition,insulatingphasesandferromagneticmetallicphasecoexist(Figure1-3).Coexistencewasveriedbydifferentmeasurementssuchasdarkeldtransmissionelectronmicroscopy,magneticforcemicroscopy(MFM),transportmeasurement,magnetizationmeasurement,andrecentlyLorentzmicroscopy.[ 25 27 ]ThepossibilityofchemicalinhomogeneityhasbeenruledoutbycompositionanalysisanddepinnedmagneticregioninMFMimages.MFMresultssupportpercolativetransportresultingininsulator-metaltransition,sincetheyshowedgrowingmagneticregionsduringtheinsulator-metaltransition.TheMFMresultsarestrongevidenceofspatialcorrelationbetweenmetallicphaseandferromagneticphase.Recently,detailedmagneticdomainstructurehasbeenrevealedbyLorentzmicroscopy.Innarrowtemperaturerange,singleferromagneticdomainsbreakdowntomulti-domainferromagneticregions.Thismicroscopicresulthasrecentlybeenconrmedbymacroscopicmagnetizationmeasurementsaspartofthisthesiswork[ 8 ]. 20

PAGE 21

1.5MultiferroismandBiMnO3(BMO)Multiferroismwasinitiallydescribedascoexistenceofmultipleferroicproperties,viz.ferroelectricity,ferromagnetism,andferroelastisity.ArepresentativemultiferroicmaterialisBiMnO3(BMO).Currently,thedenitionofmultiferroismincludesantiferromagnetism,sincesomemanganites,e.g.YMnO3,InMnO3,TbMn2O5,andBiFeO3(BFO)showedantiferromagnetismbut(tunable)ferroelectricityundermagneticeldsandthusamagnetoelectriceffect.Thesemultiferroicmaterialscanbeclassiedintermsoforiginofferroelectricity,sinceallthemagneticorderingoriginatesfromsuperexchangeinteraction[ 17 31 ].ThersttypeofmaterialsareonesinwhichferroelectricityoriginatesfromA-sitecation,i.e.BFOhaslonepairelectronsinBiions.Anothertypeofmaterialsareonesinwhichferroelectricityisinducedbymagneticfrustration,e.g.spiralmagneticorderingandcollinearspinordering.Thisiscalledasimproperferroelectricity,sincethisisinducedbymorecomplexlatticedistortionand/orsomeotherorderingwhichmentionedabove[ 17 ].InthersttypeofmaterialsthelonepairoftheA-sitecationdistortsthecrystalstructureandresultsinanon-centrosymmetricstructurethusbreakinginversionsymmetryandallowingferroelectricitytoexist.Amongthesematerials,ndingferromagneticandferroelectricoxidesisdifcult,sinceferromagnetismandferroelectricityaremutuallyexclusive.Ferromagnetismintransitionmetaloxidesrequirespartiallylledd-electrons,sincestrongHundcouplingalignselectronspinsinitsownd-shell.Thetypeofsuperexchangeinteractiondetermineslongrangemagneticorders,viz.antiferromagnetismorferromagnetism.However,intransitionmetaloxides,e.g.BaTiO3andPbTiO3,mechanismofferroelectricityrequiresemptyd-shellinB-sitecations.Theemptyd-shellspontaneouslymakescovalentbondswithelectronpairsofanoxygenatom,namelyp-dhybridization[ 22 ](Figure1-4).Thiscreatescooperativeatomicdisplacements[ 15 ].However,thishybridizationcouldnothappeninthesystemwithpartiallylledd-shellofB-sitecationsduetocoulombrepulsionbetweenelectrons.Theserequirementsseemtoindicatethatcoexistenceofmagneticorderingand 21

PAGE 22

Figure1-4. (a)Ferroelectricityduetoemptyd-shellinB-sitecation(b)itsprojection[ 18 ]. ferroelectricityisnotpossible.However,thereareveryfewmultiferroicmaterialssuchasBiMnO3[ 32 ].BiMnO3ispredictedtohaveferromagnetismbelow105Kandferroelectricitybelow450K[ 33 ].Theoriginofferroelectricityhasbeenbelievedaslone-pairinBiions,whiletheoriginofferromagnetismisduetosuperexchangeinteractionsamongMnions.RecentexperimentalandtheoreticalresultsindicatedcentrosymmetryofBiMnO3crystalstructureeveninlowtemperature[ 34 35 ].Instead,ithasbeenclaimedthatferroelectricityinBiMnO3canbeinducedbypeculiarorbitalordering.Thisorbitalorderingcreateslongrangeantiferromagnetic(AFM)interactions[ 36 38 ].ThisAFMinteractiontendstobreakinversionsymmetry.Experimentallyandtheoretically,thisAFMinteractioncanbestabilizedbyoxygencontent,defects,substratestrains,andpressure[ 36 39 41 ].OneofthereasonscontributingtothiscontroversyisthedifcultyofsynthesizingBMOeitherinbulkorthinlmform.BMOisameta-stablephase,soitsgrowthrequireshighpressuresynthesis.Also,Biionisvolatile,soit'seasytocreatecationvacancy.ItisalsoreportedthatoxygenstochiometryiscriticaltoexplainferromagnetisminbulkBMO.Raoetal.proposedthatoxygendecientsamplesaremoreferromagnetic,whilestochiometricBMOhasantiferromagnetictransitiontemperatureataround30K[ 39 ].However,Beliketal.proposeditishardtomakeoxygendecientBMOsample.Theyshowedthat,whenBMOisoxygenrich, 22

PAGE 23

antiferromagnetismisinduced[ 42 ].BMOthinlmshavebeenstudiedextensivelybyseveralresearchgroups,sincethedevelopmentofthinlmdepositiontechniquesmakesitpossibletostabilizedthedesiredchemicalphaseviasubstratestrain[ 43 ].Foradecade,researchersusedcompressivesubstratestrainandBismuthrichtargettostabilizeBMOphase.However,veryfewgroupshavesuccessfullydemonstratedferromagnetismandferroelectricityinBMO.ComparedtobulkBMOresearch,researchonphysicalpropertiesofBMOthinlmsisinveryearlystage.Thisdissertationisorganizedasfollows:inChapter2,experimentaltechniquesareintroduced.GrowthofhighqualityandepitaxialLPCMOthinlmsisdiscussedinChapter3.Chapter4describestheobservationofsingledomaintomultidomaintransitioninLPCMOthinlmsduetosubstratestressinducedin-planemagnetoanisotropy.Anisotropic(magneto)transportpropertiesinLPCMOtocapturetheoriginofCEReffectsandmagnetoelectriceffectsaredescribedinChapter5.Chapter6outlinesthegrowthofimpurity-freeandepitaxialBMOthinlmsandtheirmagnetoelectriceffects.ConclusionandfutureworkaregiveninChapter7. 23

PAGE 24

CHAPTER2EXPERIMENTALMETHODS 2.1ThinFilmGrowth 2.1.1SubstrateSurfaceTreatmentForthinlmgrowth,fourdifferentepi-polishedoxidesubstrateshavebeenused,viz.(110)NdGaO3(NGO),(001)SrTiO3(STO),(001)LaAlO3(LAO),and(001)SrLaGaO4(SLGO)fromCrysTecGmbh(Table2-1).NGOandSLGOsubstratesaregrownbyCzochralskimethod,whileSTOgrownbyVerneuillmethod.LAOsubstratesaregrownbybothmethods.STOandNGOsubstrateswerethermallyandchemicallytreatedtogetatomicterraces.Theadvantagesofgettingatomicterracesareinprovidingasinglyterminatedsurfaceandmakingitsimplertounderstandkineticsofthinlmgrowth.Itisnotedthatcommerciallyavailablesubstratesareusuallycutwithsmallangletosurfacenormal.ThismiscutanglecreatesmixedsurfaceterminationofeachsubstratesuchasSrOandTiO2onanSTOsubstrate. Table2-1. Oxidesubstrates SubstrateCrystalstructureDirectionLatticeparameters(A) NdGaO3(110)Orthorhombic(110)3.863(110)3.863(001)3.854SrTiO3(001)Cubic(001)3.905SrLaGaO3(100)Tetragonal(100)and(010)3.842(001)12.680 Here,treatmentmethodsforNGOandSTOsubstrateswillbebrieyintroduced.Thesemethodsarebasedonpublishedresults[ 44 46 ].First,NGOsubstratesweresonicatedinacetoneandethanolfor5minuteseach.Then,NGOsubstrateswerethermallytreatedinairfor30minutesat950C.Thetemperaturewasrampedupat10C/min,whiletemperaturewasrampeddowngraduallyto550Cat12C/minandthenallowedtocooldowntoroomtemperatureafterswitchingoffthefurnace.Second,STOsubstratesweresonicatedinethanolfor5minutesandinacetonefor5minutes.STOsubstratesweresoakedindeionizedwaterat55-70Cfor30minutes 24

PAGE 25

Figure2-1. Atomicforcemicroscopyimagesoftreatedsubstrates:(a)NdGaO3,(b)SrTiO3,and(c)LaAlO3 inanultrasonicbathtohydroxylatetopmostSrOlayertoSr(OH)2[ 47 ].Subsequently,theseSTOsubstratesweresonicatedinmixtureofhydrochloricacid(HCl)andnitricacid(HNO3)(ratio:3:1)for12minutes.Afterthischemicaletching,thesesubstrateswereimmediatelysoakedindeionizedwaterfor30secondsanddriedwithcompressedairtostopchemicaletching.Then,thesesubstrateswerethermallytreatedat1000-1050Cfor30minutesinair.ThesetreatedNGOandSTOsubstratesconsistentlyshowedwideterracesandunit-cellheightofsteps(Figure2-1).LAOandSLGOsubstrateswerealsotreated.LAOwasetchedinhydrochloricacidfor5minandtreatedfor30minutesin900C.Itshowedlocalterracesandunit-cellheightofsteps.SLGOsubstratesweredirectlythermallytreatedat840C,sinceSLGOreactswithwater.AFMimageindicatedroughersurfacewithnoterracesandsteps. 2.1.2PulsedLaserDepositionManganitethinlmsweregrownusingpulsedlaserdeposition(PLD)technique.ThePLDtechniquehasbeenusedextensivelyfortwodecadesowningtoeaseof 25

PAGE 26

Figure2-2. (a)PulsedLaserdepositionset-up,(b)PlumeofLPCMO,(c)PlumeofBiMnO3. handlingandlowcostofmaintenance.Itiswellsuitedforgrowingmulticomponentlmslikecupratesuperconductorsandmanganites,sinceitisanon-equilibriumprocess[ 48 ].Recently,atomiclevelcontrolofdepositioninPLDhascreatednewavenuetostudyinterfacephysics,e.g.interfaceconductionandinterfacesuperconductivityinbetweentwoinsulators.APLDsystemusesashortlaserpulse(about25nsforKrFlaser)tovaporizethesurfaceofpolycrystallinetargetmaterials,andtheresultantforwarddirectedplumetransportstargetcomponentstoaheatedsubstratewithoutlossofstoichiometryinvariousgasenvironments(Figure2-2).AKrF248nmpulsedlaserwasusedandthedepositiontakesplaceinahighvacuum(HV)chamberwhichincludesacarouselof6targetsallowinggrowthofmulti-layerswithoutexposureofthesampletoair.Thecontrolvariablesaresubstratetemperature,O2partialpressure,laseruence,pulserepetitionrate,andtarget-substratedistance.Wevariedthesubstratetemperature,O2partialpressure,andlaseruencefortheoptimizationprocess.Optimizationofthinlmgrowthconditionshasbeencompletedforthefollowingcompositions:(La1)]TJ /F8 7.97 Tf 6.59 0 Td[(yPry)0.67Ca0.33MnO3(LPCMO)(y=0.5and0.6),La0.67Sr0.33MnO3,La0.67Ca0.33MnO3,andBiMnO3(Table 26

PAGE 27

2-2).Repetitionratewas5Hz,andgrowthratesare30A/minforBiMnO3and25A/minforothermanganitelms.Atomicforcemicroscopy,transportmeasurements,magnetizationmeasurements,x-raydiffraction,andtransmissionelectronmicroscopyhavebeenconductedtoconrmtheoptimizationprocess. Table2-2. Filmgrowthconditionofmanganitethinlms CompositionSubstrateTemperatureO2PartialpressureLaseruence(C)(mTorr)(J/cm2) LPCMO(y=0.5)(110)NdGaO37801500.5LPCMO(y=0.6)(110)NdGaO37801300.5La0.67Ca0.33MnO3(110)NdGaO37802000.5La0.67Sr0.33MnO3(001)SrTiO3700400.5BiMnO3(001)SrTiO3634351.2 2.2TransportMeasurementsTransportmeasurementswereperformedinanAmericanMagnetics,Inc.(AMI)9teslasuperconductingmagnetsystemwithJanisSuperVaritempvariabletemperatureinsert.Thetemperaturecanbevariedfrom1.5Kto300KandaLakeshore332temperaturecontrollerwasusedfortemperaturecontrol.ThemagneticeldiscontrolledbyanAMI420powersupplyprogrammerandmodel4Q05100PSfour-quadrantpowersupply.Twodifferenttypesoftransportmeasurementshavebeenemployedtomeasureresistanceofmanganitethinlms. 2.2.1ConstantCurrentSourcedMeasurementsCurrentsourcemeasurementsutilizeaKeithley2000voltmeterandaKeithley220currentsource(Figure2-3(a)).Thinlmsampleswerecuttohavedimensionsofabout1mmby5mm.Fourequallydistantindiumdotsweresolderedalongthelongerside.Thesedotswereconnectedtofourpinconnectorsthroughgoldwires.Fourpinconnectorswereconnectedtothecontrolpanelwithcopperwires.Determinationofresistance,basedontheOhm'slaw,canbemadebymeasuringvoltagedropacrosstheinnertwoindiumdots.Insomecases,dcpolarityreversalmethodhasbeenemployedtoremoveoffsetandthermoelectricvoltages[ 49 ].Theschemeistoipthecurrent 27

PAGE 28

Figure2-3. Schematicdiagramsoftransportmeasurementset-upfor(a)currentsourcedfourprobemethodand(b)voltagesourcedtwoprobemethod.(c)ResistivityasafunctionoftemperatureofLPCMO(y=0.5)takenbyconstantcurrentsourcedmeasurements. directioninaspecicmanner,viz.positive,negative,andpositive. V1,+=Vsample+Vt,const,(2) V2,)]TJ /F6 11.955 Tf 10.41 1.79 Td[(=)]TJ /F7 11.955 Tf 9.3 0 Td[(Vsample+Vt,const+Vt,linear,(2) V3,+=Vsample+Vt,const+2Vt,linear,(2) Vsample=V1,+)]TJ /F6 11.955 Tf 11.95 0 Td[(2V2,)]TJ /F6 11.955 Tf 9.74 1.79 Td[(+V3,+ 4,(2) Rsample=Vsample Iapplied,(2)whereV1,+,V2,)]TJ /F1 11.955 Tf 7.08 1.79 Td[(,andV3,+aremeasuredvoltages,Vsampleisavoltageacrossthesample,Vt,constisaconstantthermoelectricvoltage,Vt,linearisalinearlyvarying 28

PAGE 29

thermoelectricvoltageasafunctionoftime,Icurrentisanappliedcurrent,andRsampleisthesampleresistance.Fromtheequationsabove,resistanceofthesamplecanbedetermined.Thedcpolaritymethodisusefultoobservepureresistancestateofmanganitethinlms,sincecurrentisalwaysoffunlessmeasurementsareexecuted.Thiswillallowminimizationofelectriceldeffects[ 3 ].Wecheckedeffectsofcontactresistanceonresistanceofsamples.WemeasuredresistanceofaLPCMO(y=0.5)thinlmonNGOsubstratebytwowiresensingandfourwiresensing(Figure2-3(c)).Overallwecouldnotndnoticeableeffectsofcontactresistances.However,inlowtemperatureregion(below60K),weobservedresistivitygetsnoisyintwowiresensingandevenfourwiresensinginlowcurrent(5nA).Thisnoisybehaviorcanberesolvedtousethehighercurrent(I=20nA),sincedetectionvoltagegetslarger. 2.2.2ConstantVoltageSourcedMeasurementsVoltagesourcedtransportmeasurementshavebeendonebyputtingaknownresistorinserieswithathinlmsample(Figure2-3(b))[ 50 ].Thismethodissuitableformeasuringtheresistanceofhighlyresistivematerials,oncethesampleresistanceishigherthan100Mandcomparabletotheinputimpedanceofthevoltmeter.Aknownseriesresistorof1Matroomtemperaturehasbeenused.Thevoltagedropacrossthesamplecanthenbecalculatedbysubtractingtheappliedvoltagefromthemeasuredvoltageacrosstheknownresistor.Thecurrentinthisseriescircuitcanbecalculatedbydividingthemeasuredvoltagebyknownresistancevalue(1.001M).Fromthesetwocalculatedvalues,resistanceofthesamplecanbecalculated.Thismeasurementmakesitpossibletomeasureresistancesupto100G.Also,thismethodcanpreventpossiblecurrentburst[ 2 ].Fordcmeasurements,weusedKeithley2400sourcemeterasvoltagesourceandKeithley2000voltmeterforvoltagesensing.Foracmeasurements,anSR-830lock-inamplierisusedbothasavoltagesourceandvoltagesensor.A100kasstandardresistorhasbeenused.Itisnotedthatweused 29

PAGE 30

lowfrequencyacvoltage(about5.6Hzandlessthan1Vrms)toavoidcapacitiveeffectsandpossibleinterferenceinsimultaneousmeasurementswithdcsource.Weselectsourcefrequencyatthelowesttemperature,whichmaximizesin-phasecomponentofvoltagesensing,sinceweareinterestedinresistancemeasurements. 2.3MagnetizationMeasurementsMagneticpropertymeasurementsweredoneusingaQuantumDesign5Tsuperconductingquantuminterferencedevice(SQUID).Magneticeldcanbeappliedupto5T,andthetemperaturecanbevariedfrom2Kto300K.Silverpaints,usedtoattachthesubstratetothePLDchamberheater,wasremovedafterdepositionfromthebacksideofthethinlmsbygrindingonsandpaper,thensonicatedinethanol.ItisnotedthatweusedplastictweezerstoavoidpossiblebackgroundsignalsintheSQUIDmagnetometerduetosamplecontamination.Ithasbeenobservedthatgrindingwithstainlesssteeltweezerscreatessignicantmagneticsignals.Thesampleismountedontheinsideofastrawandsecuredbyadditionalinnerstraws.Typicalmagnetizationmeasurementsaremagnetizationasafunctionoftemperature(M(T))inxedappliedeldandmagnetizationasafunctionofmagneticeldsatxedtemperatures(M(H)).AllthemeasurementscanbeprogrammedintheMultiVusoftwarefromQuantumDesign.Magnetoelectricmeasurementsandmagnetizationchangeinelectriceldshavebeenmeasuredusinganelectriceldprobe,whichprovideselectricalconnectionpoints,sothatvoltagecanbeappliedfromaKeithley2400(Figure2-4).Toapplyelectriceld,twomethodshavebeenused.First,twocopperwiresweresolderedontwosidesofthinlms.Thetotalelectriceldacrossthethinlmsislowsuchas0.1kV/cm,when50Visappliedandtheseparationoftwoconnectionsis0.5cm(typicalsizeofthinlms).Second,interdigitalelectrodeswithalternatingseparationbetween10mand80mwereusedinordertoapplyhigherelectricelds.TheelectrodepatternwasmadewithphotolithographyprocessinNanoResearchFacilities,UniversityofFlorida.Depositionof50-nm-thickAu-PdalloyswasdoneintheMajorAnalyticalInstrumentation 30

PAGE 31

Figure2-4. Schematicdiagramofmagnetoelectriceffectmeasurementsetup.50-nm-thickPd-Auinterdigitalelectrodesusedona60-nm-thickBiMnO3thinlm. Center(MAIC),UniversityofFlorida.Afterdeposition,anentirelmwassonicatedinacetoneforlift-offprocess.Asanalstep,twocopperwiresweresolderedwithindiumtotheelectriceldprobe.Twentysix(26)pairsof20mspacingand27pairsof90mspacingweremade.Theexpectedelectriceldsare25kV/cmfor20mspacingand-5.55kV/cmfor90mspacingfor50V. 2.4SurfaceandStructuralPropertyMeasurementsSurfacemorphologyofvariousmanganitesisstudiedusingaDigitalInstrumentmultimodescanningprobemicroscope(SPM)withanextendermodule.Surfacemorphologywasimagedwithnon-contactatomicforcemicroscopy(AFM)mode.ThisSPMalsoprovidesotherimagingmodessuchascontactAFMforimagingsurfacemorphology,Kelvinforcemicroscopyformeasuringsurfacepotential,andelectricforcemicroscopyformeasurerelativesurfacepotential.Maximumscansizeis15m15m.ForstudyingtheinterfacebetweenLPCMO(y=0.6)andNGO,weuseddualbeamfocusedionbeam(FIB)millingandhighresolutiontransmissionelectronmicroscopy 31

PAGE 32

(HRTEM)inMAIC.TheTEMsampleofaLPCMOthinlmwasacquiredasaresultofFIBtrainingbyDr.GeraldBourne.TheHRTEMwasconductedbyDr.KerrySiebein.Forstructuralstudyandphaseidentication,conventional-2x-raydiffraction(XRD)hasbeenemployed.PhilipsAPD3720inMAIChasbeenused.ThisisapowderXRD,butwithprecisemounting,itispossibletoscanthelatticeparameteralongsurfacenormal.Also,itissuitableforidentifyingchemicalphasesofdepositedmaterials.ThisinstrumenthasbeenextensivelyusedfortheoptimizationofgrowthconditionofBiMnO3thinlms,sinceBiionsneartothesurfacecanescapeduringthecoolingprocess.Additionally,highresolutionx-raydiffractionofBiMnO3thinlmswasconductedbyDr.ValentinCraiciuninMAIC. 32

PAGE 33

CHAPTER3EFFECTSOFOXYGENPRESSUREONPHYSICALPROPERTIESOFLPCMOTHINFILMSManganiteshaveshowncolossalresponsestoexternalelds,sincetheirstructure,transport,andmagnetismarecloselyrelated[ 1 6 25 ].Thinimsofmanganitesexhibituniquepropertiesassociatedwithsubstrateinducedstrains[ 51 53 ].Twodirectionsofresearchinmanganitethinlmshavebeenpursued.Oneistogetclosetothephysicalpropertiesofbulkmaterials[ 54 ].Lotsofeffortstoreproducebulkphysicalpropertiesinmanganitethinlmshavebeenmadebypostannealinginanoxygenrichenvironmentorbygrowinglmsinhighlyenergetic,reducingconditions[ 55 56 ].Anotherdirectionistodiscoverphysicalpropertiesuniquetointhinlmsandnotfoundinbulkformsofthesamecompounds,viz.anisotropictransport,substrateinducedmetal-insulatortransition,in-planemagneticanisotropy,andinsulator-metaltransitioninmanganitesuperlattices[ 51 53 57 58 ].Theseuniquepropertiesrequireneoptimizationofthinlmgrowthconditions,sothatwecanonlyexploreintrinsicpropertiesfrominterrelationamongstructure,transport,andmagnetisminmanganites.Thus,thecreationofatomicallyat,epitaxial,andstoichiometricmanganitethinlmsisanessentialsteptosearchforsuchuniquephysicalproperties.Weshowthatatomicallyat,epitaxial,andstoichiometric(La0.5Pr0.5)0.67Ca0.33MnO3(LPCMO)thinlmson(110)NdGaO3(NGO)substratescanbeoptimizedwithnecontrolofoxygenpressurewithoutconventionalpost-annealing.Thisoptimizationwasconrmedbyvariousphysicalpropertymeasurements,viz.surface,transport,structural,andmagneticmeasurements.TheLPCMOthinlms,grownunderoptimaloxygenpressure,haveshownatomicsteps,thehighestmaximumtemperaturecoefcientofresistivity(max.TCR)andpeaktemperature(TP),andferromagneticstateatlowtemperatures.WeproposethatoxygenpressureaffectsthevalencyofMnionduetooxygenvacancy,cationvacancy,andcreation/reductionofgrainboundaries[ 55 59 60 ]. 33

PAGE 34

Figure3-1. 1m1mAFMimagesofLPCMOthinlmsonas-receivedNGOsubstratesgrownunder(a)100mTorr,(b)125mTorr,(c)150mTorr,and(d)175mTorr. 3.1ExperimentalDetailsLPCMOthinlmsweregrownontwodifferentsubstrates,viz.as-receivedepi-polished(110)NGOsubstratesfromCrysTecandsubsequentlythermallytreated(110)NGOsubstrates.Thedetailedprocedurecanbefoundatsection2.1.1.DigitalInstrumentsMultimodescanningprobemicroscopewasusedtostudythesurfacemorphologyofthelmsinthetappingmode.Atomicforcemicroscope(AFM)imagesofthetreatedsubstratesconsistentlyshowedatomicsteps(about0.4nm)andlowmiscutangles(about0.05).WedepositedLPCMOthinlmsusingpulsedlaserdeposition(KrF,=248nm).Thesubstratetemperaturewas780C,andlaseruencewaskeptat0.5J/cm2,whileoxygenpressurewaspurposelyvariedineachdepositionfrom75mTorrto200mTorr.Thethicknessofthethinlmswaskeptat30nmbycontrollingthedepositiontime,unlessspeciedotherwise.Duringcool-downat20C/min,theoxygenpressurewaskeptat440Torr.TransportmeasurementswereperformedinaheliumcryostatequippedwithaJanisvariabletemperatureinsert.DCpolarityreversalmethodwasusedwithfourprobewiringandlowcurrent(5nA)tominimizecurrenteffectonmanganitesthinlms[ 3 ].Additionally,magnetizationmeasurementand)]TJ /F6 11.955 Tf 12.56 0 Td[(2x-raydiffraction(XRD)wereperformedusingaQuantumdesign5TSQUIDmagnetometerinB20,NPBandPhilipsAPD3720inMAIC,UniversityofFlorida. 34

PAGE 35

Figure3-2. Resistivityasafunctionoftemperature((T))oftheLPCMOthinlms.Inset:Max.TCRandTPasafunctionofoxygenpressure(PO2).Alldottedlinesareguidedforeyes. 3.2ResultsandDiscussion 3.2.1LPCMOonUntreatedNdGaO3(NGO)InLPCMOthinlmsgrownunderdifferentoxygenpressures,weobserveddifferencesinsurfacemorphology(Figure3-1(a)-(d)).Weobservedverysmoothsurfaceswithr.m.s.roughnessof0.5nminlmsgrownunder100mTorrorloweroxygenpressures.However,wecouldnotndanycharacteristicfeatureslikeislandsorsteps.Increasingoxygenpressureto150mTorr,wefoundatomicstepsduetostepowgrowthmodeinLPCMOlms.Itsstepheight(about0.4nm)isclosetotheheightofanLPCMOunitcell.In175mTorrorhigheroxygenpressures,thinlmsfolloweda2Dislandgrowthmode.Ther.m.s.roughnessisstilllessthan0.5nm.TheselmsaredifferentfromLPCMOthinlmsshowing3Dislandsgrowthmode,whichweregrownunderhighlaseruence(1.00.2J/cm2)andoxygenpressure(420mTorr)withgoodtransportproperties[ 5 ].Intheoxygenpressurerangesweinvestigated,interlayermasstransportofLPCMOwaswellexecuted,sincether.m.s.roughnessofallsampleswas 35

PAGE 36

abouttheunitcellheightofLPCMO[ 61 ].Thesurfacemorphologiesintheselmswerecorrelatedwithtransportproperties.Thelmswithatomicstepsshowedthehighestmax.TCRandTP.Max.TCRandTPwerelowerinthinlmswhichweregrownunderoxygenpressuresdivergingfromtheoptimaloxygenpressure(150mTorr).Bothmax.TCRandTPhavedomeshapedfeaturesabouttheoptimaloxygenpressure(InsetofFigure3-2).Whentheoxygenpressurewasbelow100mTorrduringdeposition,thegrownlmsshowinsulatingbehaviordowntothelowesttemperaturewemeasured.ThesetrendsindicateachangeoftheMn4+/Mn3+ratio[ 59 ].Thedomeshapedfeaturesinmax.TCRandTPasafunctionofoxygenpressurecanbeexplainedbyachangeinthevalencyofMnion,sinceasimilarchangeinvalencyhasbeenobservedinotherholedopedmanganitesataroundoptimaldopingofdivalentcations[ 62 ].Forexample,La1)]TJ /F8 7.97 Tf 6.59 0 Td[(xCaxMnO3hasitshighestmagneticTCatx=3/8withadomeshapedphaseboundary[ 19 ].Aroundtheoptimaldoping,domeshapedphaseboundarywasachievedbydopingwithdivalentCaion,whichcontrolsthevalencyofMnion.Inourcase,thischangeinthevalencycausedbyvaryingoxygencontentmaycreatethedomeshapeofmax.TCRandTPabouttheoptimaloxygenpressure.Atlessthan150mTorr,theaveragevalencyoftheMnionistowards3+from3.67+duetooxygenvacancy.Howeverathigherthan150mTorr,thevalencyofMnionisshiftedtoward4+duetocationvacancyratherthanoxygenabundance,sinceoxygenrichmanganitesarehardtogrow[ 59 60 ].Thiscationvacancyresultsfromcationoff-stoichiometry,i.e.changeofratiobetweenMnionandotherheavycations(LaandPrions)[ 56 63 ].IfthisoriginatedfromchangeofratiobetweenLaionandPrion,monotonicshiftofTPwouldbeobservedratherthandomeshapeTP[ 25 ].Also,grainboundaryeffectshouldbeincludedtoexplainlowmax.TCR,lowTP,andhigherresistivityatlowtemperatureofthelmgrowninoxygenrichcondition,sincethislmshowsgranularmorphology(Figure3-1(d))[ 64 ]. 36

PAGE 37

Figure3-3. Magneticmoments(lmandsubstrate)asafunctionofmagneticeld,along[1 10]directionofNGO,ofLPCMOthinlmsonas-receivedNGOsubstrates.Inset:theirmagnetizationhysteresisloopsoftheLPCMOthinlms. Figure3-4. )]TJ /F6 11.955 Tf 11.96 0 Td[(2x-raydiffractionpatternsoftheLPCMOthinlmsonuntreatedsubstratesandaNGOsubstrate,InsetshowsactivationenergiesofLPCMOthinlmsonas-receivedNGOsubstrate(opensquare)andontreatedNGOsubstrate(lledcircle). 37

PAGE 38

Figure3-5. (a)ln(/T)asafunctionof1000/TofLPCMOthinlmsandttingwithadiabaticsmallpolaronmodel.(b)ln()asafunctionofT)]TJ /F3 7.97 Tf 6.59 0 Td[(0.25tochecktherelevanceofthevariablerangehoppingmodelintheLPCMOthinlms. WeperformedXRDmeasurementsandmagnetizationmeasurementsat50K,sinceincreaseoflatticeconstantandreductionofmagnetizationaretypicallyseeninoxygendecientbutoptimallydopedmanganites[ 59 60 ].Weobservedpronouncedreductionofremnantmagnetizationanddisappearanceofmagnetichysteresisinthelmgrownunder75mTorr,aftersubtractingparamagneticNGOsignalsbylineartting(Figure3-3)[ 8 ].Also,weobservedadiffuseandbroadXRDpeaksinthesamelm,whiletheXRDpatternsofLPCMOthinlmseitherwithatomicstepsor2DislandswereoverlappedwiththeNGOsubstrate,sinceNGOsubstrateisnearlylatticematchedtoLPCMO(Figure3-4).ItwasreportedthatoxygendecientmanganitelmslowerthevalencyofMniontoward3+,sothatthedoubleexchangemechanismishampered[ 59 ].WealsocalculatedactivationenergiesofLPCMOthinlmsfromtransportdata.Theadiabaticsmallpolaronmodelandthevariablerangehoppingmodelwereemployed[ 21 ].Figure3-5clearlyshowsthetransportofLPCMOthinlmscanwellttedwiththeadiabaticsmallpolaronmodel[ 23 ].Theactivationenergydoesnotshowahugevariationamongthethinlms(insetofFigure3-4).Theactivationenergyisin133.50.5meVforthelmsshowingmetal-insulatortransition.Insulatinglms 38

PAGE 39

Figure3-6. 1m1mAFMimagesofLPCMOthinlmsontreatedNGOsubstratesgrownunder(a)100mTorr,(b)125mTorr,(c)150mTorr,and(d)175mTorr.(e)TreatedNGOsubstratefor(b). grownunder100mTorrhaveahigheractivationenergy(144meV).Hence,oxygenvacanciesmayincreasetheactivationenergy.Combinedwiththehampereddoubleexchangeinteraction,thishigheractivationenergymayleadtoastrongerinsulatorandnoinsulator-metaltransitioninoxygendecientsamples.Fromthemeasurementsandanalysisabove,itisnotedtheLPCMOthinlmwithatomicstepscanbedifferentiatedfromtheLPCMOthinlmwith2Dislands,sincethelmwithatomicsstepshaslowercoerciveeldsandsharperchangesofmagneticmomentsneartocoerciveelds(Figure3-3).Thesereectthegrainboundaryeffects,sincethelmsfollowislandgrowth[ 65 ]. 3.2.2LPCMOonTreatedNGOWealsoconrmedsimilartrendsinsurfacemorphology,transport,XRD,andmagnetizationmeasurementsinLPCMOthinlmsontreatedNGOsubstrates.ItisnotedthattheatomicstepsonLPCMOthinlmsfollowedtheatomicstepsontreatedNGOsubstrates(Figure3-6(b)and(e)).ThisindicatesthatstepedgesontreatedNGOprovidedenergeticallyfavorablesitesforthinlmgrowth[ 61 ].Thetrendsofmax.TCRandTParesimilarinbothLPCMOthinlmsontreatedandas-receivedsubstrates(InsetofFigure3-7).However,theoptimaloxygenpressureforstepowgrowthmode 39

PAGE 40

Figure3-7. (T)oftheLPCMOthinlms.Inset:Max.TCRandTPasafunctionofPO2.Adottedlineisguidedforeyes. isshifteddownto125mTorrforthetreatedsubstrates.Thewindowsforoptimaloxygenpressurearerathernarrow,sincewegotinsulatingthinlmsgrownunder100mTorr.TheactivationenergyofLPCMOshowingmetalinsulatortransitionisabout135.51.5meV.LPCMOgrownunder100mTorrhasanactivationenergyofabout140meV(InsetofFigure3-4).WealsoobservedadiffuseandlowerangledLPCMOpeakinXRDandlowermagnetizationininsulatingLPCMOthinlmsgrownunder75mTorr.Whentheoptimallmsarecompared,thebestlmontreatedsubstrateshowednoresistivityupturnatlowtemperatureunlikethebestlmontheas-receivedsubstrate.Hencethelmonthetreatedsubstrateismoremetallicatlowtemperature.WiththefactthatthemetallicphaseinLPCMOthinlmsonNGOisfavored,itreectsLPCMOthinlmsontreatedsubstratesconformbetterthanLPCMOthinlmsonuntreatedsubstrates[ 3 ]. 3.3SummaryWesuccessfullygrewatomicallyat,epitaxial,andstoichiometricLPCMOthinlmsbytuningoxygenpressure.SurfacemorphologyandtransportmeasurementsarecriticalmeasurementstooptimizethisLPCMOthinlm[ 63 ],sincetransportand 40

PAGE 41

magnetizationmeasurementsarenotenoughtocharacterizethethinlms.Disorderinthinlmsisreducedwhentreatedsubstrateswithsinglyterminatedatomicallysmoothsubstrateareused. 41

PAGE 42

CHAPTER4SINGLEDOMAINTOMULTI-DOMAINTRANSITIONDUETOIN-PLANEMAGNETICANISOTROPYINPHASESEPARATEDLPCMOTHINFILMS1Thecouplingbetweenstructure,transport,andmagnetisminhole-dopedmanganitesleadstophenomenasuchascolossalmagnetoresistance(CMR),colossalelectroresistance(CER),photo-inducedmetal-insulatortransition,andcolossalpiezoresistance(CPR)[ 1 5 ].Whilethesepropertiescouldleadtofutureapplicationsindevicessuchasbolometersandcryogenicmemories,manganitesarealreadyprovidingauniqueinsightintotheeffectofcompetingphasesonthephysicalpropertiesofmaterials[ 7 66 ].ItisnowwidelyacceptedthatphenomenasuchasCMRareconsequencesofthecompetitionamongdifferentphaseswithsimilarfreeenergy.Suchcompetitionleadstophasecoexistenceamongthreedistinctphases,viz.cubicferromagneticmetallic(FMM),pseudo-tetragonal(morepreciselyorthorhombic)antiferromagneticchargeorderedinsulating(AFM-COI),andpseudo-cubicparamagneticinsulating(PMI)phases,inmaterialssuchas(La1)]TJ /F8 7.97 Tf 6.58 0 Td[(yPry)1)]TJ /F8 7.97 Tf 6.59 0 Td[(xCaxMnO3[ 6 25 26 ].Inadditiontowell-knowneffectssuchasCMRandCER,thecoexistenceofthethreemagneticphasesalsoleadstophenomenasuchastemperaturedependentmagneticdomaintransitionandellipsoidalgrowthoftheFMMphase,whichhavebeenobservedusingLorentzmicroscopyinverynarrowtemperatureranges[ 27 ].Duetothesamecouplingbetweencrystalstructure,transport,andmagnetism,manganitethinlmshaveshownpropertiesdistinctfrombulkbehaviorsuchassubstratestraininducedmetal-insulatortransitionandanisotropictransportduetostraineldsfromsubstrates[ 51 53 ].Theeffectofstrainonthetransportpropertiesofmanganiteshasbeenwidelystudiedanditisacceptedthatmultiphasecoexistenceandpercolationplayasignicantrole[ 3 25 26 ].However,theeffectofstrainonthemagnetismofphaseseparatedmanganitesismoresubtleand 1E-printedwithpermissionfromHyoungjeenJeenandAmlanBiswas,Phys.Rev.B,83,064408,(2011).Copyright2011,AmericanPhysicalSociety. 42

PAGE 43

isstillbeingdebated.Onesuchproblemisthedistinctionbetweenintrinsicmagneticpropertiesandextrinsiceffectsonmagneticpropertiesofmanganitethinlms.Forexample,whenLa0.77Ca0.33MnO3thinlmsweregrownonsinglecrystallineNdGaO3substrates,Mathuretal.concludedthatthein-planemagneticanisotropyoriginatednotfromstressanisotropybutfrommagnetocrystallineanisotropy[ 67 ].Here,wereportthatsubstrateinducedstressplaysanintegralroleindeterminingthemagneticpropertiesofmanganitethinlms.Weobservethatin-planestressanisotropyleadstoanin-planemagneticanisotropyandamagneticdomaintransitionasafunctionoftemperatureinphaseseparated(La1)]TJ /F8 7.97 Tf 6.59 0 Td[(yPry)1)]TJ /F8 7.97 Tf 6.58 0 Td[(xCaxMnO3(x=0.33andy=0.6)thinlmsgrownon(110)NdGaO3substrateswithanisotropicin-planestrain.Ourdatashowthatwhileanisotropicstresshasaprofoundeffectonthemagnetism,thein-planeresistivityofthelmsremainsvirtuallyisotropic.BycomparingourresultsforthelmsonanisotropicNGOtothosegrownonisotropic(001)SrLaGaO4(SLGO)substrates,weconcludethatanisotropicstraincanbeusedtocontrolthemagnetichardnessi.e.thecoerciveeldinamixedphasemanganite.Suchcontrolcouldplayanimportantroleinthedesignofnanomagneticdevices. 4.1ExperimentalDetails(La1)]TJ /F8 7.97 Tf 6.59 0 Td[(yPry)1)]TJ /F8 7.97 Tf 6.59 0 Td[(xCaxMnO3(LPCMO,x=0.33andy=0.6)thinlmsoftwodifferentthicknesses(30nmand20nm)weregrownonorthorhombic(110)NdGaO3(NGO)substratesbypulsedlaserdeposition(PLD)(KrF,=248nm).Thesubstratetemperatureduringgrowthwas780C,O2partialpressurewas130mTorr,laserenergydensitywasabout0.5J/cm2,andgrowthratewaskeptatabout0.4A/s.StepowgrowthhasbeenconsistentlyobservedundertheseconditionsintheLPCMOthinlmsupto60nmthickness.Themagneticresponsereportedinthisworkwasobservedinfourdifferentlms.Thethicknesswascontrolledbydepositiontimeandthenconrmedbyatomicforcemicroscopy[ 68 ].Thelatticemismatchstrainsinthetwoin-planedirectionsofLPCMOthinlmsgrownonNGOsubstrates(LPCMO//NGO)are110=0.49%and 43

PAGE 44

001=0.26%duetodifferentin-planelatticeparametersoftheNGOsubstrates(d110=3.863A,d001=3.854A,dLPCMO=3.844A)[ 69 70 ],where(%)dsubstrate)]TJ /F8 7.97 Tf 6.58 0 Td[(dlm dlm100.Wealsogrew30-nm-thickLPCMOthinlmson(001)tetragonalSrLaGaO4(SLGO)substrates(LPCMO//SLGO)usingthesamegrowthparametersgivenabove.SincethesubstrateinducedstrainonLPCMO//SLGOthinlmsisnegligibleduetowellmatchedin-planelatticeparametersofSLGO(d=3.842A),thelmsgrownonthetwodifferentsubstratescanbeusedtoisolatetheeffectofanisotropicstrainonthemagnetismandtransportofLPCMO.Thestructureofthelmswascharacterizedbyconventional)]TJ /F6 11.955 Tf 12.61 0 Td[(2x-raydiffractionusingaPhilipsAPD3720diffractometer.MagneticpropertiesweremeasuredusingaQuantumDesign5TSuperconductingQuantumInterferenceDevice(SQUID)magnetometer.SinceLPCMOthinlmsshowthermalhysteresisinbothmagneticandtransportproperties,thermaldemagnetizationwasexecutedbyheatingupto150K(atemperaturehigherthanthehystereticregionandthemagneticCurietemperature,TC130K)beforeeachmeasurement.Fielddemagnetizationofthesuperconductingmagnetwasalsoperformedat150K.Formeasuringthemagneticmomentasafunctionoftemperature(M(T)),a100OeeldwasappliedtominimizemagneticeldinducedphasechangeoftheLPCMOlms[ 71 ].TwodifferentmethodswereusedtoremovethebackgroundparamagneticsignalfromtheNGOsubstratesandobtainthemagneticmomentsofLPCMOlms.Therstmethodwasdirectsubtractionforwhichmagneticmomentmeasurementsasafunctionofappliedeld(M(H))werecarriedoutatdifferenttemperaturesforthebareNGOsubstratesbeforelmdeposition.M(T)wasalsomeasuredforthesamesubstratesinaeldof100Oe.AfterdepositionoftheLPCMOlmsonthesubstrates,M(H)andM(T)curveswereacquiredunderthesameconditionsasthemagnetizationmeasurementsofbareNGOsubstrates.TheM(H)andM(T)curvesofthesubstrateswerethensubtractedfromthecorrespondingM(H)andM(T)curvesofthelmplussubstratetoobtainthemagneticmomentsoftheLPCMOthinlms.Thesecondmethodwasbasedonlinearbackgroundtting.The 44

PAGE 45

rawM(H)dataofLPCMOlmshavesignalsbothfromtheferromagneticlmsandtheparamagneticNGOsubstrates.Sincetheparamagneticsignalislinearatlowelds,thebackgroundsignalcanbeobtainedbyttingalinearfunctiontothedatafrom700Oeto2000Oe(i.e.magneticeldsgreaterthanthecoerciveeld),andthensubtractedfromtherawdatatoobtainthemagneticmomentsofLPCMOlms.ThelinearbackgroundttingprocedurecanbeusedonlyfortheM(H)measurements.BackgroundsubtractionwasnotrequiredforLPCMOlmsgrownonnon-magneticSLGOsubstrates.ADigitalInstrumentsMultimodescanningprobemicroscopewasusedinthetappingatomicforcemicroscopemodetocheckthesurfacemorphologyandthicknessofthethinlms[ 68 ].Resistivityasafunctionoftemperature((T))wasmeasuredinaheliumcryostatequippedwithaJanisvariabletemperatureinsert. 4.2ResultsandDiscussion 4.2.1StructureandTransportThesurfacesofourLPCMOthinlmsonbothNGOandSLGOsubstratesaresmoothonanatomicscale(insetsofFigure4-1(a)and4-1(b)).Ther.m.s.roughnessoftheLPCMO//NGOlmsisabout2A.ThelmsonNGOusuallyshowstep-owgrowthmodewithunitcellstepheightasshownforthe20nm-thicklm(insetofFigure4-1(a)).Thestepheightisabout4A,whichiscomparabletothelatticeconstantofbulkLPCMO.Ther.m.s.roughnessofa30-nm-thickLPCMO//SLGOlmisabout4Aandstep-owgrowthwasnotobserved.ResistivitymeasurementsoftheLPCMO//NGOlmsshowsharptransitionsattheinsulator-metaltransitiontemperature(TIM,obtainedwhilecooling)andmetal-insulatortransitiontemperature(TMI,obtainedwhilewarming)whereas,theLPCMO//SLGOthinlmsshowamoregradualtransitionandnarrowerthermalhysteresis(Figure4-1(b))[ 72 ].TheTMIofthe20-nm-thickLPCMO//NGOthinlmwasabout92K,whichis16Khigherthanthatofthe30-nm-thicklm,whiletheTIMarewithin4K(60Kforthe20-nm-thicklmand56Kforthe30-nm-thicklm).ThelargevariationofTMIwiththicknessisrelatedtothestrain-inducedstaticphaseseparated 45

PAGE 46

Figure4-1. (a)(T)ofa30-nm-thick(circle)and20-nm-thick(square)LPCMOlmson(110)NGOsubstrates.Theinsetshowsa2m2mAFMimageofa20-nm-thickLPCMOthinlmgrownonNGO.(b)(T)behaviorofa30-nm-thickLPCMOthinlmonanSLGOsubstrate.Theinsetshowsa2m2mAFMimageofa30-nm-thickLPCMOthinlmonSLGO. (SPS)stateatlowtemperaturesandwearecurrentlyperformingexperimentstostudythiseffectindetail[ 3 ].)]TJ /F6 11.955 Tf 12.29 0 Td[(2x-raydiffractionpatternofthe30-nm-thickLPCMO//NGOlmdidnotshowanyindividualLPCMOpeaksduetothesimilarlatticeparametersofLPCMOand(110)NGOintheout-of-planedirection,whilethex-raydiffractionpatternoftheLPCMO//SLGOlmclearlyshowssharpLPCMOpeaks(Figure4-2).Thus,alltheLPCMOlmsweregrownwithasinglechemicalphaseandwerehighlyorderedalong 46

PAGE 47

Figure4-2. )]TJ /F6 11.955 Tf 11.96 0 Td[(2x-raydiffractionpatternsofthe30-nm-thickLPCMOthinlmgrownonanSLGOsubstrate(thickline)andthe30-nm-thickLPCMOthinlmgrownonanNGOsubstrate(thinline).Theinsetshowsthelowangleregionindetail. perpendiculardirectiontothesubstratesurface.ThelmsaresmoothonanatomicscaleandshowsharpresistivitytransitionsatTIMandTMI. 4.2.2MagneticPropertiesFigure4-3showsM(T)curvesofa30nm-thickLPCMO//NGOthinlm,takenundereldcooling(FC)andeldcooledwarming(FCW)inaeldof100Oeappliedparalleltothe[110]NGOdirection.Wewillshowinsection 4.2.4 thatthe[110]NGOdirectionistheeasyaxisfortheLPCMO//NGOlms.WeusedthedirectsubtractionmethodtogetthemagneticmomentsoftheLPCMO//NGOlmsonly.Duetothebackgroundsubtraction,thezeroeldcooled(ZFC)M(T)dataisaccompaniedbyalargerelativeerrorandhenceisnotshownhere.TheM(T)behaviorissimilartopreviousresultsobtainedforbulkLPCMO(x=0.375andy=0.600)[ 73 ].Theparamagnetictoferromagnetictransitionoccursatapproximately130K.Althoughtheentirelmdoesnotbecomeferromagneticat130Ksincetheselmsshowmultiphasecoexistence,wedenetheTCtobeapproximately130K,atwhichM 47

PAGE 48

Figure4-3. M(T)ofthe30-nm-thickLPCMOthinlmonNGOundereldcooling(FC)(square)andeldcooledwarming(FCW)(circle)runsina100Oeeldalongthe[110]NGOdirection.TheinsetshowsM(T)behaviorofthe30-nm-thickLPCMOlmonSLGOunderzeroeldcooling(ZFC),eldcooling(FC),andeldcooledwarming(FCW)inanin-plane100Oeeld. becomesameasureablenon-zerovalue[ 74 ].WhilethisdenitionofTCisnotrobust,itapproximatelymarksthetemperatureatwhichtheferromagneticregionsarerstnucleated.TheM(T)graphshowsthermalhysteresisintheFCandFCWrunsatsimilartemperatureswhereweobservedhysteresisinthe(T)behavior.Thishysteresisisduetotheuidphaseseparated(FPS)statetransformingintotheglassystaticphaseseparated(SPS)stateatlowtemperatures[ 3 71 ].Themagneticmomentsaturatesbelow30KastheLPCMO//NGOlmenterstheSPSregion.IntheSPSstatetheFMMregionsarefrozeninspaceandhencethemagneticmomentisconstantwithareductionintemperature[ 3 71 ].ThetransitionfromtheFPStoSPSstatealsoleadstoauniquebehaviorofthecoerciveeldasafunctionoftemperaturetobedescribedinthenextsection.M(T)behavioroftheLPCMO//SLGOlmisshownintheinsetofFigure4-3,whichshowsmagneticthermalhysteresisandsaturationofthemagneticmoment(0.6B=Mn)below30K,similartotheLPCMO//NGOlm. 48

PAGE 49

4.2.3VariationofCoerciveFieldwithTemperature Figure4-4. In-planezeroeldcooled(ZFC)magnetizationhysteresisloopsofthe30-nm-thickLPCMOthinlmgrownonSLGOalongthetwoorthogonalin-planedirectionsat50K. Wewillnowdiscusstheeffectofanisotropicstrainonthemagneticpropertiesofphaseseparatedmanganites.Previousstudieshaveshownthatin-planeanisotropicstrainleadstouniaxialmagneticanisotropyintheplaneofthethinlm[ 57 67 ].However,thesestudieswereperformedonpurelyferromagneticmanganites.Inaphaseseparatedmanganite,weexpectthatthemagneticanisotropymayleadtoanomalousbehaviorofthesubmicrometersizedferromagneticregions[ 26 ].Magnetichysteresisloopsofthethinlmsweremeasuredinthezeroeldcooled(ZFC)statetostudythepureunmagnetizedstateateachtemperature.ThemagneticmomentsofLPCMO//SLGOlmscouldbedirectlymeasured,sinceSLGOsubstrateisnon-magnetic.NGOsubstratesareparamagnetic,whichnecessitatesacarefulbackgroundsubtractiontoobtainthemagnetizationofLPCMO//NGOlms.TheM(H)dataweremeasuredalongtwoin-planedirectionsoftheSLGOsubstrate,viz.[100]and[010].Magnetichysteresisloopsinthetwodirectionsshowthatthereisnegligiblemagneticanisotropyatallmeasuredtemperaturesintermsofremanences 49

PAGE 50

andcoerciveeldsasshowninFigure4-4forT=50K.Thesignicanceoftheseisotropichysteresisloopswillbediscussedinsection 4.2.4 .TheM(H)loopsalsoshowthattheremanencesaremuchlowerthanmagneticmomentsat2kOe(Figure4-4),whichsuggestsweakinteractionamongferromagneticregionsintheLPCMO//SLGOlms[ 75 ].DuetothelargeparamagneticsignalsfromtheNGOsubstratesweusedtwodifferentmethodsviz.,directsubtractionandlinearbackgroundttingtogetZFCmagnetichysteresisloopsoftheLPCMO//NGOlms.Themagneticeldisappliedparalleltothe[110]directionoftheNGOsubstrateswhichisalsothemagneticeasyaxisofthelmsonNGO(sectoin 4.2.4 ).AsshowninFigure4-5,theM(H)loopsobtainedusingdirectsubtractionandlinearbackgroundttingshowsimilarresults.Forexample,at50Kdirectsubtractiongivesacoerciveeldof2304Oeandaremanenceof2.40.1B/Mnwhilelinearbackgroundttinggivesacoerciveeldof2305Oeandaremanenceof2.30.1B=Mn.Thetwobackgroundsubtractionmethodsshowsimilarresultsthroughoutthetemperaturerangeinvestigated.ComparedtotheLPCMO//SLGOlms,theLPCMO//NGOlmsshowrectangularM(H)loopsi.e.theirsaturationmagnetizationiscomparabletotheirremanence.Thesharpchangeofthemagneticmomentatthecoerciveeldsupportsdomainwallmotionbyeithernucleationmodelorpinningmodel,andtherectangularshapeofhysteresisloopsimpliesastronglydevelopeduniaxialanisotropyintheLPCMOthinlmsonNGOsubstrates,asdiscussedinsection 4.2.4 [ 75 ].TheTCoftheLPCMO//NGOlmisapproximately130K(Figure4-3).From130Kdownto80K,theobservedcoerciveeldsarelessthantheeldstepsize.WhilesuchsmallHccouldbeduetosmallsingledomain(SD)FMMregionsinamatrixofPMIand/orCOIphases,webelievethattheremanenteldinthesuperconductingmagnetinaSQUIDmagnetometergivesrisetotheobservedhysteresisinthistemperaturerange[ 76 ].Hence,weassumethatHc0andthatthelmisinasuper-paramagneticstatefrom130Kdownto80K.Below80K,theHcrstincreasessharplytoabout 50

PAGE 51

Figure4-5. (a)ZFCmagnetizationhysteresisloopofthe30-nm-thickLPCMOthinlmgrownonNGO,at50Kwiththeeldappliedalong[110]directionoftheNGOsubstrateafterusingdirectsubtraction(DS).TheinsetshowsM(H)behaviorofthesubstratebeforedeposition(circle)andthelmplussubstrateafterdeposition(square).(b)ZFCmagnetizationhysteresisloopofthe30-nm-thickLPCMOthinlmgrownonNGO,at50Kwiththeeldappliedalong[110]directionoftheNGOsubstrateafterusinglinearbackgroundtting.TheinsetshowsM(H)behaviorofLPCMOandNGO(square)andthettedparamagneticbackground(triangle). 300Oeat60Kandthendecreasesgraduallytoabout190Oeat30K(Figure4-6(a)).However,Mrincreasesmonotonicallyasthetemperatureisdecreasedfrom80Kto30K(inset,Figure4-6(a)).ThebehaviorofHcandMrasafunctionoftemperatureshowsthattheFMMregionsgrowinthesametemperaturerangeinwhichthecoerciveeldshowsnon-monotonicbehavior.ThegrowthoftheFMMregionswithloweringtemperatureissupportedbymagneticforcemicroscopy,Lorentzmicroscopy,transportmeasurements,andothermagneticmeasurements[ 25 27 ].Thenon-monotonicbehaviorofHcasafunctionoftemperaturealongwiththemonotonicincreaseofMrstronglysuggestsanSDtomultidomain(MD)transitionasafunctionoftemperature.SDFMMregionsarenucleatedbelowTCandgrowdownto60K.FurthergrowthoftheFMMregionsbelow60KresultsinMDbehaviorandhence,adecreaseinHc.Suchmagneticdomaintransitionistypicallyobservedinferromagneticneparticlesasafunctionofparticlesize[ 77 ].HereweobserveanSDtoMDtransitionwith 51

PAGE 52

decreasingtemperatureduetoanincreaseinsizeofFMMregionsembeddedinanon-ferromagneticAFM-COI/PMImatrix. Figure4-6. (a)Hc(T)fora20-nm-thickLPCMOthinlmgrownonNGOfromZFCM(H)curveswithHalong[110]NGOusinglinearbackgroundtting(invertedtriangle)anda30-nm-thickLPCMOthinlmgrownonNGOfromZFCM(H)curveswithHalong[110]NGOusinglinearbackgroundtting(square)anddirectsubtraction(triangle)and10kOeFCM(H)curveswitheldalong[110]NGOusinglinearbackgroundtting(circle).TheinsetshowsMr(T)forthe30-nm-thickLPCMOthinlmgrownonNGOfromZFCM(H)curveswithHalong[110]NGOusinglinearbackgroundtting.(b)Hc(T)fora30-nm-thickLPCMOlmgrownonSLGOfromZFCM(H)curveswithHappliedalongtwoorthogonalin-planedirections.TheinsetshowsMr(T)forthe30-nm-thickLPCMOthinlmgrownonSLGOfromZFCM(H)curveswithHappliedalongtwoorthogonalin-planedirections. Below30Kdownto5K,thecoerciveeldincreasesagain(Figure4-6(a)).ToexplainthisunexpectedincreaseinHc,wehavetogobacktoFigure4-3,whichshowsthatbelow30KLPCMOisintheSPSstate[ 3 ].Asaresult,thesizeanddistributionoftheferromagneticregionsremainconstantbelow30Kandtheincreaseofcoerciveeldmaybeduetothereductionofthermalenergyinthemultidomainferromagneticregionsofconstantsize.ToconrmthishypothesiswettheHc(T)curvefrom5Kto30KtotheequationHc(T)=Hc0=[1)]TJ /F6 11.955 Tf 12.04 0 Td[((T=b)q],wherebisrelatedtothespinipenergybarrieratzeromagneticeld,Hc0isthecoerciveeldat0K,andqistheexponentoftemperature,sinceithasbeenobservedthatthecoerciveeldsofmultidomainnanoparticlesshowaT2=3dependence[ 78 79 ].Fromtting,itisestimatedthatq 52

PAGE 53

is0.680.02(Hc0360Oeanda=E0=kB88K)whichconrmsourhypothesis.ThemagneticbehaviorofLPCMOinthelowtemperatureregionisalsodifferentfromspinglassmaterialssuchastheCuMnsystem,wheredisplacedhysteresisloopsareobservedafterFC[ 80 ].WhenwemeasuredmagnetichysteresisloopsafterFCto10Kinamagneticeldof10kOe,wecouldnotobserveanysignicantdifferenceinthepositiveandnegativecoerciveelds(datanotshown).WhenFCiscarriedoutinaeldof10kOe(whichishigherthanallthecoerciveeldsmeasuredatdifferenttemperatures),itincreasestheFMMphaseofthethinlmsothatthedomaintransitionsetsinatthehighertemperatureof70K(Figure4-6(a)).IntheLPCMO//SLGOlms,allthemagneticquantities(Hc,Mr)showgradualincreaseasthetemperatureisreduced(Figure4-6(b)).Inparticular,themagneticdomaintransitionobservedinLPCMO//NGOlmswasnotobservedintheLPCMO//SLGOlms.Wethusconcludethatthedomaintransitionisduetosubstrateinducedanisotropicin-planestresscoupledwithphasecoexistenceinLPCMOthinlms.WenowpresentadetaileddescriptionofthemagneticanisotropyinLPCMOthinlms. 4.2.4StrainInducedMagneticAnisotropyWemeasuredM(H)curvesoftheLPCMO//NGOlmsalongtwoperpendicularin-planedirectionsviz.,[110]and[001]directionsoftheNGOsubstratestocheckforpossiblestraininducedmagneticanisotropy.FromFigure4-7(a),itisclearthatthe[001]NGOdirectionisthemagnetichardaxis,whilethe[110]NGOdirectionisthemagneticeasyaxisoftheLPCMO//NGOlms.Asdiscussedearlier,whenthemagneticeldisappliedalongtheeasyaxisthelmsshowmagnetichysteresisloopsbelowTC.Sincewemeasuredapproximatelysquareshapedthinlms(6mm6mm)whicharechemicallyinasinglephase,wecouldneglectthein-planeshapeanisotropyfromthelmgeometry.Also,whenwemeasured(La1)]TJ /F8 7.97 Tf 6.58 0 Td[(yPry)1)]TJ /F8 7.97 Tf 6.59 0 Td[(xCaxMnO3(x=0.33andy=0.5)thinlmson(110)NGOat10K,infullyferromagneticstate,westillobservedanin-planemagneticanisotropysimilartotheanisotropyofferromagneticLa0.77Ca0.33MnO3 53

PAGE 54

Figure4-7. (a)M(H)ofthe30-nm-thickLPCMOthinlmgrownonNGOwithHalonghardaxis(circle)andeasyaxis(square)at60K.TheinsetshowsatomicstructureofNGO(d110=3.863Aandd001=3.854A)[ 18 ].(b)(T)measuredalongthehardaxis(circle)andeasyaxis(square).Theinsetshows(T)inalinearscale. thinlmsgrownon(001)NGO[ 3 67 ].Thus,wecouldneglecttheshapeanisotropyduetoelongatedorstripe-likeFMMregions[ 27 81 ].Hence,therearetwopossibleinterpretationsofthismagneticanisotropyviz.,magnetocrystallineanisotropyorstressanisotropy[ 75 ].Murakamietal.suggestedthatmagnetocrystallineanisotropyleadstoadomaintransitioninsinglecrystal(La1)]TJ /F8 7.97 Tf 6.59 0 Td[(yPry)1)]TJ /F8 7.97 Tf 6.59 0 Td[(xCaxMnO3[ 27 ].Hence,magnetocrystallineanisotropycouldalsobethereasonbehindtheobservedin-planemagneticanisotropyinourthinlms.However,stressanisotropycouldalsoplayacriticalroleduetothedifferentin-planelatticeconstantsof(110)NGOsubstrates.Alongthe[001]NGOdirection,thetensilestrainonLPCMOis0.26%andalongthe[110]NGOdirectionthetensilestrainis0.49%(insetofFigure4-7(a)).Themagneticeasyaxiswasobservedtobealongthedirectionwithlargertensilestraini.e.the[110]NGOdirection(Figure4-7(a)).ThisresultisconsistentwithexperimentsonLa0.67Sr0.33MnO3thinlmsgrownunderanisotropictensilestress[ 57 ].Boschkeretal.showedthatthein-planemagnetizationisduetothecompressionoftheMnO6octahedraintheout-of-plane 54

PAGE 55

directioncausedbytensilestrainandtheuniaxialanisotropyisduetodifferenceinin-planetensilestrainscausingfurtherdistortionoftheoctahedra[ 57 ].Throughthismechanismitispossiblethatifasubstratehasdifferentin-planelatticeconstants,themagneticeasyaxiswillbealongthedirectionwithhighertensilestrain.Todistinguishbetweenthesetwopossibleoriginsofmagneticanisotropy,wemeasuredthemagneticanisotropyofLPCMOthinlmsgrownonSLGOsubstrateswhichexertnegligiblestressonLPCMOandthereisnoin-planeanisotropicstress.FromFigure4-4andFigure4-6(b)itisclearthatthemagnetichysteresisloops,M(T),Hc(T),andMr(T)ofthe30nm-thickLPCMO//SLGOthinlmarealmostidenticalforappliedeldsalongthe[100]and[010]directionsoftheSLGOsubstrates,atalltemperatures.SoourobservationisthattheLPCMO//SLGOlmhasnegligiblein-planemagneticanisotropy,whilethereisclearmagneticanisotropyintheLPCMO//NGOlms.Hence,themainreasonformagneticanisotropyintheLPCMOthinlmsistheanisotropicstressexertedbythesubstrate.Whileweobservedclearin-planestraininducedmagneticanisotropy,wedidnotobserveanysignicantin-planeanisotropyinthetransportproperties(Figure4-7(b))(wedidobservealowerresistivityatlowtemperaturesalongthemagnetichardaxisbutthatwasduetoahigherelectriceldinthatdirectionsincethedistancebetweenthevoltageleadswasshorteralongthe[001]direction[ 3 ]).Infact,inalinearscale(insetofFigure4-7(b))the(T)behaviorappearsidenticalinthetwoin-planedirections.TheTIMis60Kinbothdirectionsandatthattemperaturetheresistivityanisotropy(NGO[001])]TJ /F12 7.97 Tf 6.59 0 Td[(NGO[110] NGO[110]100)is15%.Thus,webelievethatanisotropicin-planepropertiesduetosubstrateinducedstraincanbeclearlyobservedinmagneticmeasurements,butnotintransportmeasurements.OurobservationisincontrasttoWardetal.whosuggestedthatthedifferenceinTIMandmaximumobservedalongtwoperpendicularin-planedirectionsofLPCMOthinlmsgrownonNGOsubstrateswasduetoanisotropicstrainandwasmaximizedasmagneticeldwas 55

PAGE 56

Figure4-8. (a)M(H)curvesofthe30-nm-thickLPCMOthinlmgrownonNGOwithHalong[001]NGOobtainedusinglinearbackgroundtting.TheinsetshowsMsasafunctionoftemperaturefortheeasyaxis(circle)andhardaxis(square)directions.(b)Ha(square)andHc(circle)asafunctionoftemperature.TheinsetshowsKuofthe20-nm-thick(unlleddiamond)andthe30-nm-thick(lleddiamond)thinlmsasafunctionoftemperature. lowered[ 53 ].Suchastraininducedresistivityanisotropyhasalsobeenpredictedtheoretically[ 82 ].Itispossiblethatsuchresistivityanisotropycanonlybeclearlyobservedwhentheresisitivitymeasurementsaretakenatthescaleofphaseseparationi.e.inthemicrometerscale,andrequiresfurtherinvestigation.Next,werampedthemagneticeldupto12kOeontheLPCMO//NGOlmsalongthe[001]NGOdirectioni.e.thehardaxis.TheremovalofthebackgroundsignalwasmoredifcultduetothelackofhysteresisintheM(H)curves.Thelinearbackgroundttingmethodwascarriedoutforeldvaluesabove5kOe,sincethemagneticmomentsaturatesatahighereldalongthehardaxisasshowninFigure4-8.Saturationmagnetizationvaluesalongthehardaxisdirectionwereslightlylowerthanthosealongtheeasyaxisatalltemperatures(insetofFigure4-8(a)).SincetheLPCMO//NGOlmsshowstrongin-planeuniaxialanisotropy,wecouldestimatetheuniaxialanisotropicconstant(Ku)asafunctionoftemperatureinthe20nm-thickandthe30nm-thickLPCMOlmsusingsaturationmagnetization(Ms)values,anisotropicelds(Ha)andtherelationKu=MsHa/2(insetofFigure4-8(b))[ 75 ].Asexpected,thethinnerlm 56

PAGE 57

Figure4-9. (a)Hc(T)behaviorofthe30nm-thickLPCMOlmgrownonNGOforthreedifferentvaluesof,whereistheanglebetweentheappliedeldHandtheeasyaxis.(b)H)]TJ /F3 7.97 Tf 6.58 0 Td[(1casafunctionofcos()atdifferenttemperatures.TheinsetshowsstandarderroroftheslopeoftheH)]TJ /F3 7.97 Tf 6.59 0 Td[(1cvs.cos()graphateachtemperature showshigherKuvaluesduetolargersubstrateinducedstraineffects.ThesevaluesarecomparabletothatofLa0.7Ca0.3MnO3thinlmsonNGOsubstrates(3.6105erg/cm3at77K)andhigherthanthatofaLa0.7Sr0.3MnO3thinlmonSrTiO3(8.4104erg/cm3atroomtemperature)[ 67 83 ].Asthetemperaturewasreduced,agradualmonotonicincreaseoftheuniaxialanisotropicconstants,saturationmagnetizationvalues,andanisotropiceldswasobservedunlikethenon-monotonicbehaviorofthecoerciveeldsduetothemagneticdomaintransition.Usingtheuniaxialanisotropicconstantat60Kandexchangestiffnessvalue(A)fromLorentzmicroscopy,wecalculatedthecriticaldiameterfordomaintransitionassumingasphericaldomain[ 27 84 ].Weobtainavalueof85nmforthecriticalradius(rc).Whentheanisotropiceldandcoerciveeldarecompared,differencesnotonlyintrendsfromtemperaturevariationbutalsoinmagnitudescanbeeasilyidentied(Figure4-8(b)).Theanisotropiceldsarelargerbyanorderofmagnitudeormorethanthecoerciveeldsalongtheeasyaxis.ThisdifferenceinmagnitudeimpliesthatStoner&Wohlfarth'smagneticreversalmechanismmaynotbevalid[ 85 86 ].Therearetwomainmechanismsformagneticswitchingbehaviorviz.,Stoner&Wohlfarth's 57

PAGE 58

coherentrotationmodelandanucleation&propagationmechanism[ 75 ].Adirectwaytocheckwhichmodelisapplicableinourcaseisameasurementofcoerciveeldsasafunctionoftheanglebetweenthemagneticeasyaxisofsampleandexternalmagneticelds,sinceeachmodelpredictsaspecicbehavior.Thenucleation&propagationmechanismleadstotheKondorskylaw,(Hc()=Hc(0)/cos())[ 85 87 88 ].Weset[110]NGO(easyaxis)as=0and[001]NGO(hardaxis)as=90.M(H)curvesweretakenattwointermediateangles,=322and=622.Figure4-9(a)showscoerciveeldsasafunctionoftemperatureforthreedifferentangles.Itisclearthatthecoerciveeldincreaseswiththeanglebetweentheeasyaxisandtheappliedmagneticeld.Whentheinverseofcoerciveeldsisplottedasafunctionofcos(),theKondorskylawisobeyedbetween10Kand60K(Figure4-9(b)).Thisbehaviorisalsoobservedinothermanganitesystemsgrownonarticiallymiscutsubstrates[ 89 ].Atthehighertemperaturesof70Kand80K(datanotshownfor80K),theKondorskylawisnotobeyed(insetofFigure4-9(b)).WebelievethatthisdeviationfromKondorskylawmaybeduetothespatialmotionoftheFMM,AFM-COI,andPMIphasesintheFPSstateandpossiblecontributionofotherreversalmechanisms[ 3 ]. 4.3SummaryIn-planeanisotropicstressleadstouniaxialmagneticanisotropyinmanganitethinlms.Wehaveshownthatwhensuchmagneticanisotropyisinducedinphaseseparatedmanganitesitleadstoasingledomaintomultidomaintransitionasafunctionoftemperature.Thedomaintransitionissimilartothatobservedinferromagneticneparticlesasafunctionofparticlesize.Inphaseseparatedmanganitesthesizeoftheferromagneticmetallicregionsembeddedinanon-ferromagnetic(charge-orderedantiferromagneticorparamagnetic)matrixincreasesasthetemperatureisdecreased.Duetothestressinduceduniaxialmagneticanisotropytheincreaseinsizeoftheferromagneticregionsresultsinthetemperaturedependentdomaintransition.Thevariationofthecoerciveeldwithtemperatureisasignatureofthedomaintransition. 58

PAGE 59

Thetemperaturedependenceofthecoerciveeldisafeaturewhichcouldmakeitpossibletousemanganitesascryogenicmagneticmemory,sincemagneticinformationcanbewrittenatatemperaturewithlowcoerciveeldandstoredatalowertemperaturewithhighercoerciveeld. 59

PAGE 60

CHAPTER5ANISOTROPICTRANSPORTINLPCMOTHINFILMSTheinterfaceregionbetweentwoperovskitematerialsisasubjectofrecentinterestsincethisregionhaspropertiesnotobservedinthebulkformofeitheroftheconstituentmaterials[ 90 93 ].Themostwell-knownsystemistheinterfacebetweentheinsulatorsLaAlO3(LAO)andSrTiO3(STO)whichbecomesmetallicandsometimesevensuperconducting,possiblyduetochargetransferbetweenthetwomaterials.Thechargetransferoccurstoavoidapolarcatastropheinultra-thin(fewmonolayers)lmsofthepolarperovskiteLAOgrownonthenon-polarperovskiteSTO[ 91 ].Recentdevelopmentsinpulsedlaserdeposition(PLD)thinlmgrowthtechnologysuchasin-situhighpressurereectionhighenergyelectrondiffraction(RHEED),haveenabledresearcherstoperformsuchexperimentsbygrowingoxidesthinlmsoneatomiclayeratatime[ 94 ].Suchhighlevelofcontrolevenallowsselectionofthetopterminationlayerofthethinlm.Whilewedonothaveanin-situRHEEDinourPLDapparatus,wehavebeenabletoverifythehighqualityofourthinlminterfacesusingpost-depositiontechniquessuchascrosssectiontransmissionelectronmicroscopy(TEM)andx-rayreectivity(XRR).WeareinterestedinstudyingtheinterfacebetweenphaseseparatedLPCMOandNGObecausetheroleofmagneticionssuchasMnandstrainattheinterfaceisstillpoorlyunderstood.Whenwestartedresearchonthisparticulartopic,ourintentwastostudytheinuenceofstrainandmagnetismontheuidandstaticphaseseparationatandneartheinterface.Hence,ourexperimentshavebeenfocusedontwodifferentcategories:(1)effectofanelectriceldonphaseseparationattheinterfaceand(2)thecombinedeffectofelectricandmagneticeldsattheinterface.Wehavethusobservedsignaturesofauniqueversionofthemagnetoelectriceffectattheinterfaceofphaseseparatedmanganites. 60

PAGE 61

5.1LPCMO(y=0.6)ThinFilmsLPCMOthinlms(y=0.6)werecharacterizedusingTEM.CrosssectionalTEMcanprovideimportantinformation,viz.thicknessofathinlm,crystallographicrelationbetweenathinlmandasubstrate,andvariationincompositionofathinlmalonglmnormal.InordertoprepareasampleforcrosssectionalTEM,dualbeamfocusionbeam(FIB)millingwasused.DualbeamFIBusesenergeticGallium(Ga)ionsforfastprocessingandanelectronbeamasimagingprobetominimizedamagebyGaions.CarbonlayerswerecoatedonLPCMOthinlmstoreducecharging.Chargingresultsinfailureofimaging.Ontopofcarboncoating,12m(length)2m(width)1.5m(thickness)ofPlatinum(Pt)wasdepositedintheFIBtoprotectsurfacesofthinlms.AfterdepositingPt,seriesofmillingoperationswithdifferentionbeamcurrentswereconductedtocreateaslabofsampleconsistingoflayersofPt,carbon,LPCMO,andNGO.Theprocessconsistsofcreatingtrenches,thinning,undercutting,furtherthinning,creatingawedge,andnallycuttingthesidewalls.Justbeforenalthinning,thicknessoftheslabisapproximately150nm.Thewedgeshapedslab,whichwasmadewithonesideisthinnerthantheother,increasedthechancetoobserveTEMimagesinelectrontransparentregion(Figure5-1(a)). Figure5-1. (a)ScanningelectronmicroscopeimageofLPCMO(y=0.6)thinlmonNGOpreparedfortransmissionelectronmicroscopy(TEM).(b)TEMimageofLPCMO//NGO(300) 61

PAGE 62

Figure5-2. (a)TEMimageofLPCMO(y=0.6)thinlmonNGO(600k).Anarrowindicatespossibleinterfacebetweenlmandsubstrate.(b)Selectedareadiffraction(SAD)patternofLPCMOandNGO(300).(c)Fastouriertransformation(FFT)ofLPCMOregion. Inlowmagnication(Figure5-1(b)),cleardistinctionofeachlayerwasseen(carbon,LPCMO,andNGO).Also,thicknessoftheLPCMOthinlmisobservedtbeabout25nm.ThethicknessfromAFMmeasurementwas30nm[ 68 ].Inhighmagnication,weobservedclearlatticeimages.DuetoextensivebombardingofGaionsduringTEMsamplepreparation,non-uniformsurfacesand/oramorphouslayerswerecreated,sothatsomeportionsdonotshowclearlatticeimages(Figure5-2(a)).Fromintensityproleoflattices,thelatticeconstantisabout0.4nm.However,clearinterfacehasnotbeenobservedeitherduetoamorphouslayerorcloselatticematchingbetweenthethinlmandthesubstrate.Wealsoperformedselectedareadiffraction(SAD),whichcoveredallareasincludingcarbon,LPCMO,andNGOlayers(Figure5-2(b)).Fastfouriertransformation(FFT)ofLPCMOregionhassamepatternasSADresults(Figure5-2(c)).Thisindicatesthatthelmandthesubstratehavespecicorientationrelations,viz.LPCMO[110]//NGO[110]andLPCMO[001]//NGO[001][ 95 ].Wealsoperformedx-rayenergydispersivespectroscopy(EDX)inTEMapparatus.Thisoptionusesemissionofcharacteristicenergyfromelementsinthesample.Thebeamspotsizeistheoreticallylessthan1nmandpracticallyafewnm.TheEDX 62

PAGE 63

Figure5-3. (a)X-rayenergydispersivespectroscopyonaLPCMOthinlm.Depthproleofmainelementsinthethinlmandthesubstratearedisplayed.(b)X-rayreectivityofa60-nm-thickthinlm.Insetshowsplotsin2()vs.2. inTEMcanthusbeusedasalocalprobeofchemicalcomposition.EDXlinescanshowschemicallysharpinterfacebetweenLPCMOandNGO(Figure5-3(a)).Wealsoperformedspotanalysistocheckforsmallvariationsofchemicalcomposition.Weperformedthisspotanalysisofthreedifferentspots:twospotsinLPCMOregionandaspotinNGOregion.OnespotinLPCMOisthemiddleofLPCMOlayer,andanotherspotinLPCMOisneartheinterfacebetweenLPCMOandNGO.ThespotinNGOisjustbelowtheinterface.InthetwospotsinLPCMO,wedonotdetectsignicantchangeofcomposition(Table5-1,5-2,and5-3).WealsoobservedabundanceofNdandGainNGOregion,sothatitisunlikelythatthereisintermixingattheinterfacebetweenthinlmandsubstrate. Table5-1. IdealcompositionandratioofLPCMO LaPrCaMnLa/MnPr/MnCa/MnPr/LaSum 0.2680.4020.3310.2680.4020.331.51 Table5-2. CompositionofLPCMOatthemiddleofthethinlms LaPrCaMnLa/MnPr/MnCa/MnPr/LaSum 7.169.916.8321.730.330.420.311.381.10 63

PAGE 64

Table5-3. CompositionofLPCMOneartheinterface LaPrCaMnLa/MnPr/MnCa/MnPr/LaSum 6.589.167.0521.750.300.420.321.391.05 Wemeasuredx-rayreectivity(XRR)ofa60-nm-thickLPCMO(y=0.6)thinlm(Figure5-3(c)).XRRcangivelayerthickness,density,androughness[ 96 ].Initially,ouraimwastondthethicknessofourthinlm.XRRresultshowstwodifferenttypesofKiessigfringes,whichindicatesthatourlmisanelectronicallydistinctdoublelayersystem.Recently,itwasreportedthatevenhighqualityLa1)]TJ /F8 7.97 Tf 6.58 0 Td[(xSrxMnO3thinlmscanhaveelectronicallydistinctdoublelayersfromvariousspectroscopicmethods[ 97 ].Tondthethickness,weusedthefollowingformula. 2tq sin2())]TJ /F7 11.955 Tf 11.96 0 Td[(sin2(c)=,(5)wheretisthickness,isaninteger,cisthecriticalanglefortotalexternalreectioninthelayer,isthewavelengthofx-rays(=1.541837A).Weplottedsin2()vs.2,thenappliedleastsquaretting(insetofFigure5-3(c)).Weget49.5nmasthethicknessfromsmalleroscillations.Weestimatethatanapproximately10-nm-thicklayerisformedneartheinterfacebetweenthethinlmandthesubstrateusingthesamemethodonthebroaderoscillations.Thiselectronicallydistinctlayerneartheinterfaceisconsistentwithpolarizedneutronreectivitymeasurementsonourthinlms[ 98 ]. 5.2ElectricFieldInducedAnisotropicTransportNearinsulator-metaltransitiontemperature,reductionofresistivityduetoelectriceldswasreportedintheLPCMOsystem[ 3 99 ].Twopossiblemodelswereproposed:dielectricbreakdownmodelanddielectrophoresismodel[ 100 101 ](Figure5-4).Dielectricbreakdownmodelsuggestsanelectriceldinducedconversionoftheinsulatingphasetometallicphase,whiledielectrophoresismodelproposesanelectriceldinducedrealignmentofthemetallicphasealongtheelectricelddirection.InLPCMO,thedielectricbreakdownmodelpredictsanisotropicdecreaseofresistivity 64

PAGE 65

andincreaseofmagnetization,sincethereductionintheamountoftheinsulatingphasesshouldresultinanincreaseofFMMphase.However,dielectrophoresismodelpredictsnochangeinmagnetizationandanisotropictransport,inwhichresistivityalongelectricelddecreaseswhileresistivityperpendiculartoelectriceldincreases.Wefabricated100m250mcrossstructuresoutof20-nm-thickLPCMOthinlms.Micrometer-scalephaseseparationisseeninLPCMO,soanisotropicresistancemeasurementsneedtoaccountfornon-uniformequipotentials.Onewaytoreducetheeffectofnon-uniformequipotentialsistoperformtheexperimentsonasub-millimetercrossstructure.Bydeningtransverseandlongitudinaldirectionsinthecrossstructure,wecanclearlydetectpossibleanisotropictransport.5-5(a)showsopticalmicroscopyimageofactualcrossstructureofLPCMO.Thisstructurewascreatedbyphotolithographyandwetetching.Also,Figure5-5(b)showsthedenitionoftransversedirection(EGleg)andlongitudinaldirection(FHleg).Alonglongitudinaldirection(FH),highdcvoltagewasapplied.Itisnotedthatbothlegsexperiencedifferentsubstratestresses,sinceeachlegisparallelstoeachcrystallographicdirectionofNGO,e.g.(110)and(001). Figure5-4. Proposedmodelsforcolossalelectroresistance(CER)effect.(a)Statewithnoelectricelds.Underelectricelds,(b)Isotropicconductionduetodielectricbreakdownmodeland(c)Anisotropicconductionduetodielectrophoresismodel. 65

PAGE 66

Figure5-5. (a)Opticalmicroscopyimageof100m250mcrossstructureofLPCMOthinlms.(b)Schematicdiagramofthecrossstructure(c)sequentialtransportmeasurementsset-up. Weperformedresistancemeasurementsofthebothlegsonthecrossstructure(EGandFH)usingvoltagesourcedresistancemeasurements.Figure5-6(a)showsR(T)ofbothlegshavesimilarfeatures.Especially,thetransitiontemperaturesaresame.Inordertodetecttransverseresistanceunderhighelectricelds,weusedasmallacvoltagewithlowfrequency.Smalldrivingvoltage(lessthen1Vrms)reducespossibleelectriceldeffectsalongtransversedirection.ACsourcecanalsoreducetheeffectofirregularequipotential.Lowfrequencies(f=5.657Hzand18.647Hz)wereusedtoavoidcapacitiveeffects.Thedc-andac-transportmeasurementsmatcheachother.Atlowvoltage(upto5V),wecouldnotobservecolossalelectroresistance(CER)effects(Figure5-6(b)).TomeasuretheCEReffectathighvoltage,wemeasureddcresistanceasafunctionoftemperaturewith5Vand60ValongEG.Thisshowsclearbreakdownofresistancealongthelongitudinaldirectionnearinsulator-metaltransition(InsetofFigure5-6(b)). 66

PAGE 67

Figure5-6. (a)Resistanceasafunctionoftemperature(R(T))ofbothlegs.(b)R(T)bydcsourceandacsourcealonglegEG.Inset:R(T)alonglegEGwithtwodifferentdcvoltages. Weperformedtwodifferentisothermalsequentialtransportmeasurementsalonglongitudinalandtransversedirections(EGandFH):resistanceasafunctionofelapsedtime(R(t))andresistanceasafunctionofappliedvoltage(R(V))(Figure5-5(c)and5-7(a)).Sequentialmeansweputtimegap(td)betweenlongitudinalandtransversemeasurementstominimizeaninterferenceeffectoftransverseacvoltageonlongitudinalresistancemeasurements.Aftereachisothermaltransportmeasurement,wewarmuptheLPCMOaboveinsulatortometaltransitiontemperature(130K),sincetheelectriceldeffectonLPCMOisirreversible[ 3 ].First,wemeasuredR(t)nearinsulator-metaltransitiontemperature(TIM=65K),sincewecanpossiblymonitordynamicsofeitherdieletrophoreticmotionordielectricbreakdownmotionintimedomain.Allthemeasurementsweredoneat69.5K.Forgure5-7(b),45Vdcwasappliedalonglongitudinaldirection,and1.0Vrms(f=5.657Hz)wasappliedonlyfor1second(t1=0.80s,t2=0.18s,andtd=0.02s)outofeachcycleofthemeasurements(2seconds).Weobservedlongitudinalresistanceshowedsmalleroscillationbehavioratthebeginning.Afterdropofresistanceintransversedirection,hugemodulationoflongitudinalresistancestarts.Thetransverseresistancethenstartstoincreaseagain(gure5-7(b)).Wefurtheroptimizedt1,t2,andtdtominimizeoscillatingfeature,sincetheoscillationoflongitudinalresistanceisonly 67

PAGE 68

observedatwhichacvoltageison(datanotshown).Thisreectsitmaynotbefrommanganitesbutfrommeasurementscheme.Wecouldremovetheoscillationfeatureinfollowingcondition(t1=2s,t2=1s,andtd=0.50s)(gure5-7(c)).25Vdcwasappliedalonglongitudinaldirection,and0.5Vrms(f=5.657Hz)wasappliedalongtransversedirectionat69.5K.Bothtransportdataarequalitativelysimilarexceptoscillatingfeatureingure5-7(b).Ingure5-7(c),breakdownofthetransverseresistanceisshownafterabout1800seconds,whilelongitudinalresistanceincreases.About900secondslater,hugelongitudinalresistancedropwasobserved,whiletransverseresistanceincreases.Therstbreakdownalongtransversedirectionismainlycausedbytransverseacvoltage,whichcreatesnon-uniformelectriceldfordielectrophoresis.Thesecondbreakdownalonglongitudinaldirectionismainlydrivenbydc-voltage.Sincephaseseparatedmanganitesitselfcancreatenon-uniformelectriceld,thedc-voltageassistsdielectrophoresisalonglongitudinaldirection[ 101 ].Infact,wecouldcontrolbreakdowntimebychangingacvoltageanddc-voltage.Thesetwodielectrophoreticmotionsareclearsignaturespatialrealignmentofmetallicphaseinthesample.Itisnotedthatweobserved,whendc-voltageisappliedalongEGdirection,longitudinalbreakdowntimeisshorter(Figure5-7(d)).ThisshorterlongitudinalbreakdowntimehasbeenobservedinseveralLPCMOsamples,whenlongitudinaldirectionismagneticeasyaxis.Alsothemissingoftransversebreakdownduetodielectrophoresis(datanotshown),whendcvoltageisalongmagneticeasyaxis,maybeindicationofelongatedelectricregionsalongtheeasyaxis.Second,weperformedisothermalR(V)measurements(Figure5-8).Wemeasuredbothresistances,whilesweepingdcvoltagealonglongitudinaldirection(EG).Weput0.5sastd.Weobservedtwotothreeordersofmagnitudedropinlongitudinalresistance,whileweonlyobservedequalorlessthananorderofmagnitudedropintransverseresistance(Figure5-8).Toclarifyourobservation,wecalculateanisotropicresistanceratio(ARR(%)REG)]TJ /F8 7.97 Tf 6.59 0 Td[(RFH REG100).ThisARRshowedcleardecreaseof 68

PAGE 69

Figure5-7. (a)Schemeofsequentialmeasurements.Isothermalsequentialresistancesasafunctionofelapsedtime(R(t)).(a)R(t)sat69.5Kunder45Vdcasanappliedvoltageandalternating1.0Vrmsperpendiculartotheappliedvoltage.(b)R(t)sat69.5Kandunder25Vdcandalternating0.5Vrms.(d)R(t)alonglongitudinaldirection.(d)LongitudinalR(t)alongEGandFH.25Vdcwasapplied. resistancealongEG.ThisanisotropictransportinR(V)supportsthedielectrophoresismodel.However,wecouldnotdetectclearupturnoftransverseresistance,whichwastheoreticallypredicted[ 101 ]. 5.3EffectsofElectricFieldsandStrainsonMagnetotransportWehavefoundthatLPCMOthinlmson(110)NGOsubstratesshowstressinducedmagneticanisotropy.ThisanisotropycausessingledomaintomultidomaintransitioninphaseseparatedLPCMOthinlms,butdoesnotcausein-planeanisotropictransportinzeromagneticeld[ 8 ].Also,weobservethatelectriceldsaffecttransportpropertiesoftheseLPCMOthinlmsasdiscussedintheprevioussection.Inthissection,wewillshowwhattheeffectsofstrainandelectriceldsareonmagnetotransportproperties. 69

PAGE 70

Figure5-8. Isothermalsequentialresistanceasafunctionofappliedvoltage(R(V))measurements.(a)R(V)alongappliedelds(EG).(b)R(V)perpendiculartoappliedelds(FH).Insetshowsanisotropicresistanceratioat68Kand69K. Werstfabricatea100m250mcrossstructureofLPCMOfroma30-nm-thickthinlm.Wethencheckedresistanceasafunctionoftemperature(R(T))withoutmagneticelds.IncontrasttoexperimentalresultsfromWardetal.,weobservednodifferenceinR(T)alongtwodifferentdirections:[110]and[001],whicharethemagneticeasyandhardaxes,respectively[ 8 53 ](Figure5-10).TheTIMareabout68Kforbothdirections.WehavedenedTIMasmaximumresistance(dR=dT=0)whilecoolingtocomparethisTIMwithTIMundermagneticelds.TIMundermagneticelds(H=8T)ismoveduptoahighertemperatureof148K,whenthemagneticeldisinout-of-planedirection(Figure5-10(b)andFigure5-11(a)).ThisincreaseofTIMisduetomeltingofinsulatingphasesinmagneticelds[ 1 25 102 ].Whenmagneticeldisinplane,TIMismovedupfurtherinbetween151Kand159K(Figure5-10(c)-(d)andFigure5-11(b)-(c)).SpreadofTIMforin-planemagneticeldswillbeexplainedinthesection 5.3.2 .Westillobservedsmallhysteresisshowing2KdifferenceinTIMandTMIinallcongurationwhichindicatedthattheLPCMOthinlmisstillinphasecoexistencestate.ThepossiblereasonsofincreaseinTIMalongin-planedirectionarethecombinedeffectofdemagnetizationfactor,phasecoexistence,andevolutionofeachphase.Ataglance,thisincreaseofcriticaltemperatureandlower 70

PAGE 71

Figure5-9. (a)ZeroeldR(T)ofLPCMOthinlms.Schematicdiagramsofdifferentmagnetotransport:(b)Magneticelds(H)isoutofplaneandperpendiculartocurrents.(c)Hisin-planebutperpendiculartocurrents.(d)Hisin-planeandparalleltocurrents. breakdownmagneticeldsalongin-planecouldbeexplainedbydemagnetizationfactorfromlmgeometry,sinceactualeldsinsidethesampleismuchhigherthanforin-planecaseforthesameappliedmagneticeld[ 75 ].However,demagnetizationfactorforanout-of-planeeldisnearly1inhomogeneousmagneticthinlms.Thus,itisobviouslynottrueforourcasebecausewedealwithphaseseparatedmanganitethinlmsgrownunderanisotropictensilestrains.Also,itwasreportedthatsubstratestrainscancontrolmagneticeasyaxis.Especially,theLa0.67Sr0.33MnO3isgrownonLaAlO3,easyaxisofthelmisalongout-of-planedirection[ 103 105 ].Inourcase,theLPCMOthinlmsexperiencein-planetensilestrains,sothatwecanruleoutthatpossibilityforthesubstratestrainseffectsonanisotropicmagnetotransport.Webelievethisanisotropicmagnetotransportsareduetomagneticphaseseparationandevolutionofeachphasesintemperature.Breakdownmagneticelds(HB)aredeterminedfromdR=dH0at60K(Figure5-12).Then,wefoundthatout-of-planeHB(1.17T)isfourtimeslargerthanin-planeHB,hard(0.29T)at60K,whileout-of-planeHB(5.23T)isonly1.03timeslarger 71

PAGE 72

Figure5-10. Currenteffectsonmagnetotransport,whencurrent(I)isalongmagneticeasyaxisand(a)magneticeld(H)isoutofplane(Bluelineismagnetoresistanceinhardaxis),(b)Hisalonghardaxis(in-plane),(c)Hisalongeasyaxis(in-plane).(d)Magnetoelectricresistance(MER)of(a),(b),and(c).MagentalineistheMER,whenIisalonghardaxisandHisalongout-of-planedirection. thanin-planeHB,hard(4.03T)at100K,whencurrentisalongmagnetichardaxis.Thisindicates,whenFMMincreasesincooling,demagnetizationfactorfromFMMshapecomesintoplay. 5.3.1CurrentEffectsonMagnetotransportSinceelectriceldeffectswereobserved,wemeasuredmagnetoresistancewithdifferentcurrentsaboutthreedifferentmagneticelddirections(Figure5-10(b)-(d))[ 3 ].Itisnotedthatweobservedahumpafterinsulator-metaltransitionataround80K,andupturnofresistanceatlowtemperature(30K)withlowsensingcurrent(50nA).Thisupturnfeaturecanbeexplained,sincebelow30Kthesampleisenteringthestaticphaseseparation(SPS)region.Whenweapplied5Aandmeasuredmagnetoresistance,theupturnandhumpfeaturesdisappeared.Also,signicant 72

PAGE 73

Figure5-11. R(H)inthreedifferentmagneticelddirectionsat60K.SolidlinesisR(H)incurrentalonghardaxis. decreaseofresistancebelowTIMwasobserved,whichclearlyindicatescurrenteffectsinmagnetoresistance.Magnetoelectricresistance(MER)isdenedasrelativechangeofmagnetoresistancestolowcurrentmagnetoresistance(MER(%)=(RI,HIGH)]TJ /F8 7.97 Tf 6.59 0 Td[(RI,LOW) RI,LOW100).WhenwecalculatedMER,weobserveddistinctincreasesofMERintwotemperatureranges(Figure5-11(d)).Wealsocalculatedtwootherin-planeMER(Figure5-11(d)).In-planeMERisnegligibleaboveferromagneticCurietemperature(TC),butout-of-planeMERlinearlyincreasewithdecreaseoftemperatureaboveTC.Also,atlowtemperature,out-of-planeMERismorepronouncedthanin-planeMER. 5.3.2StrainEffectsonMagnetotransportWehavefoundseveralevidenceofstraineffectsonmagnetotransport,viz.MER(T),R(H),insulator-metaltransitiontemperatureindifferentmagneticelddirection,andanisotropicmagnetoresistance(AMR).WecalculatedMER(T)oftwodifferentcurrentdirections,whilemagneticeldsareparalleltolmnormal(Figure5-11(d)).Themagentalineistheresultfromcurrentsalonghardaxis.ThisclearlyshowslowtemperatureanisotropicMERduetosubstratestrain.WemeasureR(H)forthree 73

PAGE 74

differentmagneticelddirections.Asexpected,breakdownofresistanceisseenathigherelds,whenthemagneticeldisalongout-of-planedirection(Figure5-12).Formagneticeldalongtwoin-planedirections,thebreakdownofresistancerequireshighermagneticeldalongmagneticeasyaxis.Itdoesnotdependonanglebetweencurrentsandmagneticelds.However,asshowninFigure5-12,afterthebreakdown,resistancealongmagneticeasyaxisisalwayslower.ItisnotedFigure5-11(a)showedmagnetoresistance(MR)alongeasyaxishaslowerresistancethanMRalonghardaxis(blueline)inlowtemperature.Thisindicatessubstrateinducedstraineffectsalsocontributetomagnetotransprot.Wegetinsulatortometaltransitiontemperatureontwodifferentcurrentdirections(Figure5-13(a)).IncontrasttoWardetal.,effectsofmagneticelddirectionoutweigheffectsofsubstratestraineffect.TheTIMforIalonghardaxisarealwayslowerthanTIMforIalongeasyaxisby2K.However,thisisequivalenttosweepingtemperaturestep. Figure5-12. (a)dR/dTofR(T)under8T.Threedifferentanisotropicmagnetoresistancesasafunctionoftemperature:(b),(c),and(d). 74

PAGE 75

BeforewediscussAMR,wewilldenethreedifferentAMR,thenwewilldiscussdifferenceinthemandstraineffectsineachAMR.Theresistancesareallnormalizedtothevalueat200K.ThedenitionofROUT?,RIN?,andRIN==canbefoundatFigure5-10(b)-(d). AMRIN==(%)RIN==)]TJ /F7 11.955 Tf 11.95 0 Td[(RIN? Rave100,whereRaveis1 3RIN==+2 3RIN?,(5) AMROUT==(%)RIN==)]TJ /F7 11.955 Tf 11.95 0 Td[(ROUT? Rave100,whereRaveis1 3RIN==+2 3ROUT?,(5) AMROUT?(%)RIN?)]TJ /F7 11.955 Tf 11.95 0 Td[(ROUT? Rave100,whereRaveis1 3RIN?+2 3ROUT?,(5)InallAMRmeasurements,weobservedpeaksataround130K.ItisneartoferromagneticCurietemperature(TC)[ 8 ].WeclearlyseeAMRdoesnotgotozeroaboveTC.Thispeakandnon-zeroAMRaboveTCpreventtoadopttheAMRtheoriesforferromagneticmetals,asitwasarguedinotherholedopedmanganitesystem[ 106 ].InferromagneticmetalsaboveTC,AMRgoestozero.InFigure5-13(b),thesigninAMROUT?canshowtypeofsubstratestrain[ 107 ].ThisAMRparticularlyremovecontributionofLorentzmagnetoresitance(LorentzMR),sothatonlycontributionfrommagnetocrystallineanisotropycanbeinvolved.ItisnotedourcalculationhasoppositesignfromthatofLietal.,sinceweintentionallyswitchtoRIN==-RIN?fromRIN?-RIN==forthisAMRcalculation[ 107 ].ThiswaymakesitpossibletocompareAMROUT?toAMROUT==.Lietal.proposedthinlmswithdifferentsubstratestrains,viz.,compressivestrainandtensilestrains,showdifferentsigninAMROUT?.Inourcase,bothAMROUT?shownegativesign,whichmanifestLPCMOisunderin-planetensilestrainsasexpected.Also,ThedifferenceinmagnitudeofAMROUT?suggestedmoredrasticchangesalongmagneticeasyaxis,whereLPCMOisundermoretensilestrain.ThisaxisalsoprovidesthelongestMn-Odistanceduehighertensilestrains.SinceAMROUT?doesnotincludeLorentzMR,itisinterestingtoseeeffectsofLorentzMR.WemeasuredAMROUT==(Figure5-13(c)).BetweenAMROUT==andAMROUT?,onlyAMROUT==hasnocancellationofLorentzMRbymeasuringRIN==.Alongeasyaxis,theAMROUT==isreducetoabouthalfof 75

PAGE 76

thevalues,whilealonghardaxistheAMRisincreasetwice.ArrowsindicateLorentzMRaffectsdifferentlyindifferentdirections.Wealsomeasuredin-planeAMR(AMRIN==)(Figure5-13(d)).Inthiscase,bothLorentzandmagnetocrystallinecomponentscontributesaswell.Weobservedoppositesignofthisin-planeAMR.Thisshouldbefurtherinvestigated.ItisnotedthesedifferentsignshavebeenobservedinAgdopedLaMnO3thinlm[ 108 ].Infanteetal.explainedthiswiththetheoriesforAMRofferromagneticmetals.He/sheconcludedthesedifferencesareduetotemperaturedependentspin-orbitcouplingandspindependentscatteringeffects.Thisalsocouldbeduetodifferentdegreesofsubstratestraineffects,sinceInfanteetal.grewthinlmson(110)SrTiO3substrates,whichprovideanisotropicin-planestrains. 5.4SummaryProximitytoaninterfaceaffectsphaseseparationinmanganitemainlythroughsubstrateinducedstrain.Electricandmagneticeldsmodulatethephaseseparationespecially,intheuidphaseseparated(FPS)state.Suchmodulationgivesrisetoanalternateformofthemagnetoelectriceffect.TheoreticalsimulationofaFPSstateinelectricandmagneticeldwillleadtoaclearunderstandingofthephenomena. 76

PAGE 77

CHAPTER6OPTIMIZATIONOFBMOTHINFILMSANDPHYSICALCHARACTERIZATION1Multiferroicmaterialsareuniqueinthattheyexhibitbothferromagnetismandferroelectricitysimultaneously.[ 31 ]Suchmaterialsmaybeusedtofabricatedevicessuchasmagnetictunneljunctionswithelectricallytunabletunnelingmagnetoresistanceandmultiplestatememoryelements.[ 12 ]Therecentinterestinmultiferroicsisfueledbothbythepotentialdeviceapplicationsandquestionsabouttheunderlyingphysicalprinciplesleadingtomultiferroism.[ 9 10 13 15 ]Bulkmultiferroicmaterialsarerare,possiblyduetoconictingrequirementsforferromagnetism(FM)andferroelectricity(FE).BiMnO3isperhapsthemostfundamentalmultiferroicandhasbeenreferredtoasthehydrogenatomofmultiferroics.[ 32 ]InBiMnO3(BMO),asinBiFeO3,the6s2lonepairontheBi-ionleadstothedisplacementofthationfromthecentrosymmetricpositionattheA-siteofaperovskiteunitcell.TheresultantdistortionleadstoanFMinteractionbetweentheMn-ionsattheB-siteinBiMnO3.[ 37 38 ]InbulkformBiMnO3hasbeenobservedtobebothFMandFE.[ 33 ]PolycrystallineBiMnO3canbegrownunderhighpressureandwithinaverynarrowrangeofgrowthconditions.WhilethinlmsofBiMnO3havebeengrownbyvariousgroups,fewsuchlmshaveshownmagneticpropertiessimilartobulkBiMnO3andhighenoughresistivitiesi.e.lowleakagecurrentstoallowclearmeasurementofFEproperties.[ 109 111 ]ApossiblereasonforthelowresistivitiesofBiMnO3thinlmsisthesubstrateinducedstrainwhichexacerbatesthegrowthofahighlydistortedperovskitestructure.Additionally,recentelectronandneutrondiffractiondatahavecastdoubtoverthepurportednon-centrosymmetryoftheBiMnO3crystalstructure[ 34 ]andcentrosymmetricstructureshavealsobeenpredictedusingdensityfunctionaltheorycalculations[ 35 ].Sinceanon-centrosymmetriccrystal 1E-printedwithpermissionfromHyoungjeenJeen,GuneetaSingh-Bhalla,PatrickR.Mickel,KristinVoigt,ChelseyMorien,SefaattinTongay,A.F.Hebard,andAmlanBiswas,J.Appl.Phys.109,074104(2011).Copyright2011,TheAmericanInstituteofPhysics. 77

PAGE 78

structureisessentialforferroelectricity,theobservedferroelectricbehaviorofBiMnO3maybeduetostrainand/ororderedoxygenvacancies.[ 112 113 ] 6.1OptimizationofBMOBiMnO3(BMO)hasadistortedperovskite-typestructurewitha=c=3.935A(==91.4)andb=3.989A(=91).[ 114 ]Figure6-1showsthelargermonoclinicunitcellofBMOviewedfromdifferentdirections[ 37 ];wehaveusedthemonoclinicnotationtoindexthex-raydiffractiondataofourthinlms.SincecubicSrTiO3(STO)andtetragonalSrLaGaO4(SLGO)havealatticeparameterof3.905Aand3.844Arespectively,BMOgrowswithan(111)orientationonSTO(001)andSLGO(001)substratesunderacompressivestrainduetoalatticemismatchof1.14%and2.68%respectively.ItisstillunclearwhetherthisstrainisresponsibleforthedifferenceinthemagneticandelectricalpropertiesbetweenthinlmsandpolycrystallineBMO.AsaturationmagnetizationMsatofabout3.6B/Mnat5KandaferromagneticCurietemperature(TC)of105KhasbeenobservedinpolycrystallineBMOalongwithanelectricremnantpolarizationof62nC/cm2at87K,whileinthinlmsanMsatofabout2.2B/MnandaTCofabout100Khasbeenreported.[ 33 109 114 115 ]P)]TJ /F7 11.955 Tf 12.57 0 Td[(EmeasurementsonBMOthinlmshavebeenreportedoccasionally,andaremnantpolarizationofabout16C/cm2hasbeenobserved.[ 111 ]ThestrainmayalsoinuencetheferroelectricdomainwallmotionwhichcoupledwiththelowresistanceofthethinlmshavemadeitachallengetoconrmtheFEnatureofBMOthinlmsleadingtothecontroversialsituationpresentedintheintroduction.Toaddresssuchissues,wehaveoptimizedthegrowthofBMOthinlmsonSTOandSLGO.Wehaveobtainedstoichiometric,epitaxialthinlmsofBMOonSTOwhichhaveahighresistivityatlowtemperaturethusfacilitatingthemeasurementofP)]TJ /F7 11.955 Tf 11.96 0 Td[(Eloopsconrmingthemultiferroicnatureofourlms.TheBMOthinlmsweregrownusingpulsedlaserdeposition(PLD).Anoff-stoichiometric(Bi-rich)targetwithcompositionBi2.4MnO3wasablatedusingaKrFexcimerlaser(=248nm).ThehighBicontentofthetargetallowedustouserelativelyhighsubstrate 78

PAGE 79

Figure6-1. ThemonoclinicunitcellofBiMnO3viewed(a)along(111)directionand(b)along(203)direction. temperatures(Ts)andstillgettherightBicontentforstoichiometricBMOlms.Thelmqualitywasextremelysensitivetothecoolingrate,stillsensitivetotheTsandtheoxygenpressure,whileitwasindependentofthegrowthratewithintherangeused.Thelaserenergydensitywaskeptat1.00.2J/cm2.TheoptimumowingoxygenpressureandTswere35mTorrand634C,respectively.Thedepositionratewas0.05nm/s.ThelmswerecooledinanO2atmosphereof680Torratarateof20C/minandhigher(upto40C/min).Here,wepresentresultsfromone60nm-thickBMOthinlms.Weobtainedsimilarresultsfromtheotherthinlmsinthethicknessrangeof30nmto60nm.Figure6-2(a)showsthex-raydiffractiondatafor60-nm-thickBMOthinlmsgrownunderxedoxygenpressure(about40mTorr)butdifferenttemperature.TheinsetshowsthattheBMOmainlygrowswitha(111)orientation.WealsondotherBMOpeakscorrespondingtootherorientationsthan(111)BMOaswellassmallpeakcorrespondingtoMn2O3impurities.Especially,(203)orientationofBMOasexpectedfromthestructureofBMOat2=28.76degrees,sinceitisanotherfaceinperovskite 79

PAGE 80

Figure6-2. )]TJ /F6 11.955 Tf 11.96 0 Td[(2diffractionpatternsofBiMnO3thinlmsonSrTiO3.(a)TemperatureeffectsondiffractionpatternsofBMOinxedpressure(about40mTorr).TheinsetshowstheBMO(111)peakindetail.(b)OxygenpressureeffectsondiffractionpatternsofBMOinxedtemperature(632C). unitcell(Figure6-1andFigure6-2).Wend(111)BMOpeakgetssmallerTsabove638C,whileotherpeaksstillexist.Also,wedonotndoutconsistentchangeofeachpeak.Figure6-2(b)showsthex-raydiffractiondatafor60-nm-thickBMOthinlmsgrownunderxedtemperature(about632C)butdifferentoxygenpressures.Inalloxygenpressureweused,multipleorientationpeaksandimpuritypeakexist.Higheroxygenpressure(45mTorr)loweredsamplequality.Westilldonotndanyconsistentchangeinpeaksduetooxygenpressures.Toremovetheseimpurities,weincreasedthepost-depositionrateofTsupto40C/mininanO2atmosphereof680Torr.ItwasreportedbulkBMOcandecomposedabove600Cfromhightemperaturex-raydiffractionresults[ 116 ].Also,postannealingat460Cin200TorrafterlowtemperaturedepositionpreservesBicontentandcrystallizeBMO[ 117 ].Figure6-3(a)showsthex-raydatafor60-nm-thicklmsgrownusingdifferentcoolingrates,whichconrmthattheimpuritypeakshavebeensuccessfullyremovedusingthemodiedgrowthcondition.TheinconsistentchangeofrelativepeakintensitiesinFigure6-2canbeexplainedbyBire-evaporationafterdeposition.Rapidquenchingtechniqueisthebestwayamongthreeparameters,viz.temperature,oxygenpressure,andcoolingratetoremoveMn2O3 80

PAGE 81

Figure6-3. (a))]TJ /F6 11.955 Tf 11.96 0 Td[(2diffractionpatternsofBMOthinlmsonSTOindifferentcoolingrate.Theinsetshowsimpuritypeaksand(222)BMO.2mby2mAFMimagesofthesurfaceofBMOcooledin(b)20C/min,(c)30C/min,and(d)40C/min. impurityphaseandstabilize(111)orientationofBMOalongout-of-planedirection.Figure6-3(b)-(d)showsthesurfacemorphologyofsampleswithdifferentcoolingrates.Thesurfaceroughnessgetsworseinthisfastcooling.Ther.m.s.roughnessofsamplesis4.7nmfor20C/min,5.5nmfor30C/min,and7.85nmfor40C/min.Allthethinlmsshow3-Dislandgrowthmodewithdifferentr.m.s.roughness.ThechemicalpropertiesofthelmswerecharacterizedwithAugerelectronspectroscopy(AES)usingaPerkin-ElmerPHI660scanningAugermultiprobeinstrument.ToconrmthestoichiometryofthesamplesweperformedAugerelectron 81

PAGE 82

spectroscopy(AES)measurementsat300Kinultrahighvacuum(UHV)conditions.DerivativeAESsurfacespectraweretakenusing5keVprimaryelectronbeamfromkineticenergiesof50eVto1500eVatincidentanglesfrom30to60.Depthprolingwasperformedbytakingsurfacespectrawiththeparametersgivenabovefollowedbyanin-siturepeated3keVAr-ionsputtering.SurfacespectraoftheBMOlmsdisplayedthreemanganese(Mn)peakslocatedat548eV,595eV,645eV,twobismuth(Bi)peaksat106eV,254eVandoneoxygen(O)peakat518eVtogetherwithresiduecarbon(C)peakat273eVwithconcentrationslessthan1%.AftersixsecondsofArsputteringonthesurface,theCpeakdisappearedandBi,MnandOconcentrationsarefoundtobe23.3%,24.1%and52.6%respectivelywithabouta2%error.TheseconcentrationsimplythattheBMOstoichiometryisconsistentwiththemeasuredBMOx-raypeaksfrom)]TJ /F6 11.955 Tf 12.89 0 Td[(2measurements.Moreover,thesensitivityfactorforoxygenisbasedonanMgOmatrixandsincethereisnomatrixparameterintheatomicpercentagecalculations,thiscouldaccountfortheslightlylowerthanstoichiometricoxygenconcentrations. Figure6-4. (a)Magnetizationasafunctionoftemperature(M-T)plotforBMOthinlmsonSTOinanin-planeeldof500Oe.TheinsetshowsrawM-TofthelmsandbareSTO.Thefullsymbolsandopensymbolsarethezeroeldcooledandeldcooleddatarespectively.(b)M-TplotforBMOthinlmsonSrLaGaO4inanin-planeeldof1000Oe.TheinsetshowsrawM-TofthelmsandbareSLGO. 82

PAGE 83

Figure6-5. (a)Magnetizationasafunctionofmagneticeld(M-H)plotforBMOthinlmsonSTOandSLGOat10K.TheinsetshowsrawM-Hofthelms,bareSTO,andbareSLGO.(b)ZoomoftheM-Tplotinlowmagneticelds. ThemagneticpropertiesofBMOarecloselyrelatedtoitsuniquecrystalstructure.BMOissimilartothecompoundLaMnO3(LMO)butduetothe6slonepairtheBiionmovesawayfromthecentrosymmetricpositionattheB-siteofaperovskitestructure.LMOisanA-typeantiferromagnetduetoantiferromagneticallystackedferromagneticlayers.[ 118 ]InBMOthedistortioncausedbytheBi-ionleadstoanFMinteractionbetweenthelayers.[ 37 38 ]Hence,BMOhasanoverallmagneticmomentthathasbeenmeasuredtobeashighas3.6B/Mninpolycrystallinesamples,whichisclosetomaximumpossiblemagnetizationof4B/Mn.[ 114 ]Inthinlmsthemagneticmomentisreducedquitelikelyduetothesubstrateinducedstrain.TheTCinthinlmsisalsolowerthanthevalueofabout105Kobtainedinpolycrystallinesamples.[ 38 114 ]Figure6-4and6-5showtheM)]TJ /F7 11.955 Tf 12.66 0 Td[(TandM)]TJ /F7 11.955 Tf 12.66 0 Td[(Hcurvesof60-nm-thickBMOthinlms.Themagneticeldwasappliedintheplaneofthelmforthemagneticmeasurements.TheM)]TJ /F7 11.955 Tf 11.92 0 Td[(TplotrevealsaTCofabout855K.Asaturationmagnetizationofabout1B/Mnisobtainedat10Kinaeldof50kOe.Figure6-5(b)showsthehysteresisintheM)]TJ /F7 11.955 Tf 11.98 0 Td[(Hplotandacoerciveeldofabout300Oe.ThehysteresisoftheM-Hcurvesandthemagneticmomentbecomenegligibleatabout80KwhichisclosedtotheestimatedTC,conrmingthattheobservedmagnetizationisassociatedwithmagneticorderingwhich 83

PAGE 84

happensatatemperaturelowerthanthecorrespondingTCofbulkBMO[ 33 ](datanotshown).ThereducedmagneticmomentofourthinlmscomparedtobulkBMOisnotduetothepresenceofthenon-magneticimpuritiessinceallthesampleshavesimilarsaturationmagnetizationandcoerciveelds(Figure6-5).Itisnotedmagneticpropertiesareverysensitivetooxygenstoichiometry.Sundaresanetal.suggestedstoichiometricBMOisorthormbicandcantedantiferromagnetic.Itisgoingtoantiferromagneticbelow30K[ 39 ].However,Beliketal.showedoxygendecientBMOishardtobegrownbydirecthighpressureandhightemperaturemethod.TheirstochiometricBMOismonoclinicandferromagnetic[ 42 ].Basedonthemagnetizationresultsfromthereports,ourthinlmisstoichiometricorslightlyoxygenrich(lessthan5%)[ 39 42 ].SincethestructureofBMOiscloselyrelatedtothatofantiferromagneticLMO,thenon-uniformstraindistributioncouldberesponsibleforboththereducedvaluesofTCandsaturationmagnetization.Ithasbeenshownthat3-Dislandgrowthleadstonon-uniformstrainresultinginhighvaluesofstrainsattheislandedges[ 51 119 ].WealsomeasuredtheelectricalpolarizationusinganinterdigitalcapacitancegeometryandaPrecisionLCferroelectrictesterfromRadiantTechnologies.Alowleakagecurrentandhencehighresistivityisarequirementforpolarizationvs.electriceld(P)]TJ /F7 11.955 Tf 12.63 0 Td[(E)measurementsinBMOthinlms.Ouroptimizedthinlmshavearoomtemperatureresistivityofabout10-cm,whichislowerthanvaluesreportedbyothergroups.[ 15 110 ]However,by140K(below140Ktheresistanceistoohightomeasurewithourinstrumentation)theresistivityincreasestoabout1M-cmanditwaspossibletomakedirectpolarizationvs.electriceld(P)]TJ /F7 11.955 Tf 12.75 0 Td[(E)measurementsattemperaturesbelow100Kusinganinterdigitalcapacitancegeometry(Figure6-6(a)).ThecapacitoriscomposedofalternatingV+/V)]TJ /F1 11.955 Tf 12.62 0 Td[(electrodesuniformlyspacedonthelmsurface(Figure6-6(b)).Thisstructureleadstoequipotentialplanesintersectingthelmbetweeneachpairofelectrodes,resultinginacapacitancebetweentheprojectedareasofeachelectrodewithinthelm.Theprojectedareaswerecalculatedusingconformal 84

PAGE 85

Figure6-6. (a)Theinterdigitalcapacitancegeometrydepositedonthelmsurface.(b)SchematicoftheelectrodecongurationfortheP)]TJ /F7 11.955 Tf 11.95 0 Td[(Emeasurements.(c)Remnantpolarizationvs.electriceld(P)]TJ /F7 11.955 Tf 11.95 0 Td[(E)dataofa60nm-thickBMOthinlmonSTOtakenat5K. mappingandequatingthecapacitorthicknesstohalftheelectrodespatialwavelength(e=10m).[ 120 ]Figure6-6(c)showstheremnanthysteresisloopforsamplewith20C/min.at5K(similarresultswereobtainedforthesamplewith40C/min.).Thepolarizationinahysteresisloopiscalculatedbyintegratingthetotaltransferredchargeduringapplicationofabipolartriangularvoltagewaveform.Thispolarizationincludescontributionsfromleakagecurrent,capacitanceandferroelectricdomainswitching.Thepolarizationintheremnanthysteresisloopiscalculatedbyisolatingthetransferredchargefromonlythedomain-switching.Thisisdonebysubtractingtwohysteresisloopsthatareprecededbypolingpulses.Inoneloop,allofthedomainsarepre-switchedsothatnodomain-switchingchargeistransferredduringtheloop,andintheotherloopallthedomainsaresetunswitchedwithchargetransferfromdomain-switchingbeginning 85

PAGE 86

atthecoerciveeld.Theleakagecurrentandcapacitancecontributionsfromthetwoloopscancel,leavingonlytransferredchargefromdomain-switching,orequivalently,theremnantcharge.DividingtheremnantchargebytheprojectedareawendaremnantpolarizationofP23C/cm2at5K,withacoerciveeldEC60kV/cm.TheclearobservationofaferroelectricP)]TJ /F7 11.955 Tf 13.18 0 Td[(EloopinBMOonSTOappearstobeinconictwiththecentrosymmetricstructuresuggestedbyBeliketal..[ 34 ]IfthecrystalstructureofBMOisindeedcentrosymmetric,thenthepossiblereasonsfortheferroelectricbehaviorofBMOthinlmsare:(1)structuraldistortionsduetooxygenvacancies,[ 34 ](2)acentrosymmetrictonon-centrosymmetrictransitionbelowTCi.e.below100K,[ 34 ]and(3)substrateinducedstrain.[ 112 ]Although,theAESmeasurementsonourthinlmsrevealanoxygendeciencywhichcouldleadtotheferroelectricbehavior,wecannotruleouttheroleofsubstrateinducedstrain.Ifthelmisuniformlystrained,thelatticemismatchwhichis-1.14%(compressive),isnotenoughtobreakthecentrosymmetryasshownbyHattetal..[ 112 ]However,ithasbeenshownthatcompressivelatticemismatchstraincouldleadtoanon-unifromstraindistributioninthethinlmduetoislandformationandthestrainattheislandedgescouldfarexceedtheaveragelatticemismatchstrain.[ 51 119 ]Thegrowthmorphologyofourthinlms(insetFigure6-3(b)-(d))suggeststhatsuchnon-uniformstraindistributionisalsoapossiblemechanismfortheappearanceofferroelectricity.Inaddition,sincewemeasuredtheP)]TJ /F7 11.955 Tf 12.53 0 Td[(Eloopsupto85K,theferroelectricbehaviorcouldbeduetoastructuralchangebelowTC. 6.2StructuralStudyHere,weperformeddetailedstructuralstudyofBMOthinlmsonSTOandSLGO.Inprevioussection,weshowedgrowthofimpurity-freeBMOthinlms.Highresolutionx-raydiffractiontechniquecanelucidatestructuralrelationofBMOthinlmsandsubstratesindetail.First,weperformed)]TJ /F6 11.955 Tf 12.4 0 Td[(2X-raydiffractionandgrazingincidencex-raydiffraction(GIXD)ofa60-nm-thickBMOthinlmonSTO(Figure6-7).Wend 86

PAGE 87

(111)BMOorientationandconrmnegligibleimpurityphases.Also,!)]TJ /F6 11.955 Tf 12.78 0 Td[(2rockingcurvehasbeenperformedabout(111)BMOpeak.Duetoroughsurface,wecouldnotndoscillationaroundBMOpeak.Also,weobservedshoulderpeakinbetweenBMOpeakandSTOpeak.WemeasuredseveralrockingcurvesofbareSTOsubstrate,thisshoulderpeakisnotfromBMObutfromSTOsubstrates,whichmaybeduetoimperfectionofcommerciallyavailableSTOsubstrates. Figure6-7. )]TJ /F6 11.955 Tf 11.96 0 Td[(2diffractionpatternandGIXDofaBMOthinlmonSTO.Inset:!scanof(111)BMOpeak. Toshowin-planerelationshipbetweenthethinlmandthesubstrate,wecreatepoleguresofBMOandSTO.2angleisxedat57.78degreescorrespondingto(211)STO,and angleis65-70.Weobservedtwosetsofeightpeaks:f211gandf121g.Polegurefor(110)BMO(cubicnotation)showedfourdistinctpeaks(2=32and =45(Figure6-8(a)).Itisnoted valueis0rightincenterand90attheedgeofthecircles.Sincef110gBMOpeaksfallinbetweenf121gandf211gpeaks.WeconcludeourBMOthinlmsgrownonSTOwithacube-on-cubeorientation.Wealsoperformedmeshscanataround(113)STO(2=81.03)(Figure6-8(b)).TheresultindicatesBMOandSTOhasmatchedlatticeconstantalongin-plane,whileBMO 87

PAGE 88

iselongatedinout-of-planedirection.ThismeshscanshowsepitaxialrelationshipbetweenBMOandSTO.Ratioofin-planelatticeconstanttoout-of-planelatticeconstantshowsabout1%elongationalongout-of-plane.However,intheory,thisratioshouldbeabout4%,ifthevolumeofunitcellisconserved.Thisindicatesvolumemaynotbeconservedundercompressivestrain. Figure6-8. (a)Polegureoff211gSTOandf110gBMO(b)ReciprocalspacemapoftheBMOthinlmonSTO. Toconrmthicknessofourthinlms,x-rayreectivity(XRR)measurementswereperformed(Figure6-9).ABMOthinlmonSLGOshowedoscillationscorrespondingtoabout60nm.ThisoscillationindicatestheBMOthinlmonSLGOhaslowersurfaceroughness.ActuallyinsetofFigure6-9showstheBMOthinlmonSLGOhaslowerroughness.Wemeasured)]TJ /F6 11.955 Tf 12.28 0 Td[(2diffractionshowingdistinctBMOpeaksinaBMOthinlmonSLGO(Figure6-10).GIXDalsoshowednegligibleimpurityphase.WealsoperformedscansofSLGO(105)(2=42.68)andBMO(113)peaks.Thoseshowtwoneighboringpeaksare45awayeachother,whichindicatecubeonsquaregrowthmode.ItisnotedwidthofBMOpeakarewide.Itindicatedin-planemosaicityontheBMOthinlmonSLGOwith!scan. 88

PAGE 89

Figure6-9. X-rayreectivityofBMOthinlmsonSLGOandSTO.Inset:AnAFMimageoftheBMOonSLGO. Figure6-10. (a))]TJ /F6 11.955 Tf 11.96 0 Td[(2patterns(red)andGIXD(magenta)ofaBMOthinlmonSLGOareshown.Also,)]TJ /F6 11.955 Tf 11.95 0 Td[(2patterns(green)andGIXD(blue)ofaBMOthinlmonSTOareshown.(b)-scanofBMO(113)andSLGO(105). 6.3MagnetoelectricEffectsBMOismultiferroic.However,itispredictedlowmagnetoelectriccoupling[ 16 17 ].Here,weusedinterdigitalelectrodestoapplyin-planeelectricelds(upto50kV/cmfor100V).WeusedconventionalphotolithographytoputtheelectrodepatternonBMOthinlms.Wesputtered50-nm-thickPd-AuonBMOthinlms.Weusedlift-offtoremovedregion.Thispatternconsistsoftwodifferentdistantelectrodes:20mand90m.Weused5TSQUIDmagnetometerinB20,NPBwithanelectriceldprobe.Thiselectric 89

PAGE 90

probehaselectricportsatbothendsbyputtingconnectingelectricalconnectionfromsampletoaportandfromotherporttosourcemeter.Weput1MstandardresistorinseriesatoutsideofSQUIDmachine,sothatwecanavoidpossibledamageduetoburstofcurrent[ 2 ](Figure2-4).Tomeasurechangeofmagnetizationunderelectricelds,weapplied1kOeofmagneticelds.Wemeasuredmagnetizationasafunctionoftemperature(M(T))ofaBMOthinlmwith30Vand-30Vrespectively.Weexpect15kV/cmintheshorterelectrodesandabout-3kV/cminthelongerelectrodesbyapply30V.Beforeeachmeasurement,weranadegausssequenceat150K.Electriceldswereappliedbelow145Kfor30Vcase.CurrentincircuitwasmonitoredandusedtocalculateresistanceintheBMOlm.Resistancecouldnotbemeasuredbelow90Kandcurrentwentto0A.Here,weonlydiscussaboutresultsbelow90Kinordertoavoidpossibleartifactduetoleakagecurrent[ 121 ].Difference(M30V-M)]TJ /F3 7.97 Tf 6.59 0 Td[(30V)betweentwoM(T)curvesshowslineardecreaseofmagnetizationdowntolowtemperature.Weobservedachangeofsigninthedifferenceataround60K.Whenwenormalizedthisdifferencetoaverageoftwomagneticmoments(ratio(%)=100(M30V)]TJ /F8 7.97 Tf 6.59 0 Td[(M)]TJ /F11 5.978 Tf 5.76 0 Td[(30V) 0.5(M30V)]TJ /F8 7.97 Tf 6.59 0 Td[(M)]TJ /F11 5.978 Tf 5.76 0 Td[(30V)),weobserveddecreaseoftheratiodowntoabout60K.Blow60K,wegotconstantnegative.Thisisconsistentwithdisappearanceofmagnetoelectriceffectsfromferroelectricmeasurements[ 122 ].Thesystemisunderfreezingbelow50K,sothatmagnetoelectriceffectisreducedbelow50Kfromelectricpolarizationmeasurementundermagneticelds. 6.4SummaryFirst,wehavegrownthinlmsofepitaxialBiMnO3(111)onSrTiO3(001)substrates.Thelmshavethein-planealignmentandstoichiometry.TheferromagneticTCis855Kwithasaturationmagnetizationofabout1B/Mnat10K.Aremnantpolarizationof23C/cm2wasmeasuredat5Kwithacoerciveeldof60kV/cm.Non-uniformstraindistributionmayberesponsiblefortheappearanceofferroelectricityintheseBMOthinlms.Second,wealsogrowBiMnO3(111)onSLGO(001)substrates.These 90

PAGE 91

Figure6-11. (a)Magneticmomentsasafunctionoftemperature(M(T))underelectricelds(30V)ofaBMOthinlmonSTO.Inset:Resistancesafunctionoftemperature(R(T))under30V.(b)Theratioofdifferenceinmagneticmomentstoaveragemagneticmomentsasafunctionoftemperature.Inset:Thedifferenceofmagneticmomentsunder30V lmsshowclearferromagnetism,however,highresolutionXRDresultsindicatethisBiMnO3lmhasmosaicity.Thiscouldbetheoriginofdisappearanceofferroelectricityduetopossiblestrainrelaxationandrandomlyorientatedspontaneouspolarization.Third,weobservedmagnetoelectriceffectsinaBMOthinlmonSTOwithinterdigitalelectrodes.Thiseffectissmallbutshowcleardecreasedownto60K.Below60K,Theeffectissaturated,whichisslightlydifferentresultsfromP(E)measurementsunderhighmagneticelds. 91

PAGE 92

CHAPTER7CONCLUSIONANDFUTUREWORKAtomicallyatholedopedmanganitethinlms,especially(La1)]TJ /F8 7.97 Tf 6.58 0 Td[(yPry)1)]TJ /F8 7.97 Tf 6.58 0 Td[(xCaxMnO3(LPCMO)thinlmshavebeensuccessfullygrown.AFMandXRDdatashowthatinoptimaloxygenpressure,LPCMOthinlmsfollowstepowgrowthmodeandareepitaxial.EDXinTEMandhighangleannulardarkeld(HAADF)imaginghaveconrmedclearinterfacebetweensubstrateandLPCMOthinlms.HighresolutionZ-contrastlatticeimagescouldnotbeachievedduetohighelectronnoise.Z-contrastlatticeimageswillgivemoreclearinformationofourLPCMOthinlmssuchasinterfacesharpness,structuralperfectionandchemicalhomogeneity[ 43 ].Singledomaintomulti-domaintransitioninLPCMOhasalsobeenconrmedinthinlmform.However,thetemperaturerangeofthetransitiontodoubledomainfromsingledomainiswiderthanthecaseinbulksamples.Thestrongin-planeanisotropyinthephaseseparatedmaterialmaybetheoriginofthistransition.Peakincoerciveeldmapisshiftedtohighertemperatureinthickerlms,whichsuggestscriticalsingledomainradiusissmallerinthethickerlms.AnisotropictransportpropertiesunderhighelectriceldsinLPCMOthinlmsshowthatCEReffectoriginatesduetodielectrophoresis.Magnetizationmeasurementsunderelectricelds(about0.5kV/cm)donotshowclearevidenceofchangeinmagnetization.MagnetizationmeasurementofarrayedLPCMOthinlmsinhighelectriceldswillbeaninterestingexperimenttoconrmourndingandbeperformedinthenearfuture.ThecombinedeffectofelectricandmagneticeldsonmagnetotransportpropertiesinLPCMOthinlmswerealsoinvestigated.Weobservedthathighcurrents(5A)candecreaseresistancebelowTIMinthepresenceofamagneticeld,whichclearlyindicatescurrenteffectsonmagnetoresistance(MR).ThiscurrenteffectsintheSPSregionindicatestunnelingprocessacrossinsulatingphasesismoreactiveunderhighcurrents.Also,inthislowtemperatureregion,weobservedanisotropic 92

PAGE 93

magnetoelectricresistance(MER)duetosubstratestrains.Inthickerlms,magneticanisotropyislowered.ThisstrainelddependentMERisauniqueversionofthemagnetoelectriceffectattheinterfaceofphaseseparatedmanganites.ThedifferenceinmagnitudeofAMROUT?suggestsmoredrasticchangesalongmagneticeasyaxis,whereLPCMOexperienceshighertensilestrain.AngledependentmagnetotransportmeasurementswillhelptounderstandthisanisotropicAMRandelectriceldeffectonmagnetotransport.Impurity-freeandepitaxialBiMnO3thinlmsweresuccessfullygrownandcharacterizedusingvarioustechniques.Magnetizationdatashowedclearferromagnetismbelowabout85K.Thehighestmagnetizationat10Kwasabout1B/Mn.Ferroelectricitywasalsoobservedinthelowtemperaturerangewithasaturationpolarizationof23C/cm2at5K.Theferroelectrichysteresisloopclosedataround85K.Thesimilarcriticaltemperaturesforferromagnetismandferroelectricitysuggestsacouplingofthetwoorderparameters.Wearecurrentlyoptimizingthethinlmstolowerthesurfaceroughnessthroughthefollowingsteps:In-situpost-annealing,adjustmentofset-temperatureforfastcooling(increasefrom475C),andimmediatestartoffastcooling.InordertoreduceoxygencontentinBMO,itispossiblethatfastcoolingwithoutoxygenquenchingmaybenecessary. 93

PAGE 94

REFERENCES [1] S.Jin,T.H.Tiefel,M.McCormack,R.A.Fastnacht,R.Ramesh,andL.H.Chen, Science 264,413(1994). [2] A.Asamitsu,Y.Tomioka,H.Kuwahara,andY.Tokura, Nature 388,50(1997). [3] T.Dhakal,J.Tosado,andA.Biswas, Phys.Rev.B 75,092404(2007). [4] K.Miyano,T.Tanaka,Y.Tomioka,andY.Tokura, Phys.Rev.Lett. 78,4257(1997). [5] J.Tosado,T.Dhakal,andA.Biswas, J.Phys.:Condens.Mat. 21,192203(2009). [6] K.H.Ahn,T.Lookman,andA.R.Bishop, Nature 428,401(2004). [7] H.T.Yi,T.Choi,andS.W.Cheong, Appl.Phys.Lett. 95,063509(2009). [8] H.JeenandA.Biswas, Phys.Rev.B 83,064408(2011). [9] M.Fiebig, J.Phys.D:Appl.Phys. 38,R123(2005). [10] N.A.SpaldinandM.Fiebig, Science 309,391(2005). [11] R.RameshandN.A.Spaldin, Nat.Mater. 6,21(2007). [12] M.Gajek,M.Bibes,S.Fusil,K.Bouzehouane,J.Fontcuberta,A.Barthelemy,andA.Fert, Nat.Mater. 6,296(2007). [13] C.N.R.RaoandC.R.Serrao, J.Mater.Chem. 17,4931(2007). [14] W.Prellier,M.P.Singh,andP.Murugavel, J.Phys.:Condens.Mat. 17,R803(2005). [15] W.Eerenstein,N.D.Mathur,andJ.F.Scott, Nature 442,759(2006). [16] T.Kimura,S.Kawamoto,I.Yamada,M.Azuma,M.Takano,andY.Tokura, Phys.Rev.B 67,180401(2003). [17] S.-W.CheongandM.Mostovoy, NatMater 6,13(2007). [18] T.C.OzawaandS.J.Kang, J.Appl.Cryst. 37,679(2004). [19] P.Schiffer,A.P.Ramirez,W.Bao,andS.-W.Cheong, Phys.Rev.Lett. 75,3336(1995). [20] C.Zener, Phys.Rev. 82,403(1951). [21] M.B.SalamonandM.Jaime, Rev.Mod.Phys. 73,583(2001). [22] C.Kittel,IntroductiontoSolidStatePhysics(Wiley,NewYork,NY,1995). [23] M.ZieseandC.Srinitiwarawong, Phys.Rev.B 58,11519(Nov1998). 94

PAGE 95

[24] J.M.D.Coey,M.Viret,L.Ranno,andK.Ounadjela, Phys.Rev.Lett. 75,3910(1995). [25] M.Uehara,S.Mori,C.H.Chen,andS.W.Cheong, Nature 399,560(1999). [26] L.Zhang,C.Israel,A.Biswas,R.L.Greene,andA.deLozanne, Science 298,805(2002). [27] Y.Murakami,H.Kasai,J.J.Kim,S.Mamishin,D.Shindo,S.Mori,andA.Tonomura, Nat.Nanotechnol. 5,37(2010). [28] A.Moreo,M.Mayr,A.Feiguin,S.Yunoki,andE.Dagotto, Phys.Rev.Lett. 84,5568(2000). [29] G.Wachtel,D.Orgad,andD.I.Golosov, Phys.Rev.B 78,184433(2008). [30] A.J.Millis,P.B.Littlewood,andB.I.Shraiman, Phys.Rev.Lett. 74,5144(1995). [31] H.Schmid, Ferroelectrics 162,317(1994). [32] N.A.HillandK.M.Rabe, Phys.Rev.B 59,8759(1999). [33] A.M.dosSantos,S.Parashar,A.R.Raju,Y.S.Zhao,A.K.Cheetham,andC.N.R.Rao, SolidStateCommun. 122,49(2002). [34] A.A.Belik,S.Iikubo,T.Yokosawa,K.Kodama,N.Igawa,S.Shamoto,M.Azuma,M.Takano,K.Kimoto,Y.Matsui,andE.Takayama-Muromachi, J.Am.Chem.Soc. 129,971(2007). [35] P.Baettig,R.Seshadri,andN.A.Spaldin, J.Am.Chem.Soc. 129,9854(2007). [36] I.V.SolovyevandZ.V.Pchelkina, NewJ.Phys. 10,073021(2008). [37] T.Atou,H.Chiba,K.Ohoyama,Y.Yamaguchi,andY.Syono, J.SolidStateChem. 145,639(1999). [38] A.M.dosSantos,A.K.Cheetham,T.Atou,Y.Syono,Y.Yamaguchi,K.Ohoyama,H.Chiba,andC.N.R.Rao, Phys.Rev.B 66,064425(2002). [39] A.Sundaresan,R.V.K.Mangalam,A.Iyo,Y.Tanaka,andC.N.R.Rao, J.Mater.Chem. 18,2191(2008). [40] E.Ohshima,Y.Saya,M.Nantoh,andM.Kawai, SolidStateCommun. 116,73(2000). [41] D.P.Kozlenko,A.A.Belik,S.E.Kichanov,I.Mirebeau,D.V.Sheptyakov,T.Strassle,O.L.Makarova,A.V.Belushkin,B.N.Savenko,andE.Takayama-Muromachi, Phys.Rev.B 82,014401(2010). [42] A.A.Belik,T.Kolodiazhnyi,K.Kosuda,andE.Takayama-Muromachi, J.Mater.Chem. 19,1593(2009). 95

PAGE 96

[43] J.H.Lee,L.Fang,E.Vlahos,X.Ke,Y.W.Jung,L.F.Kourkoutis,J.-W.Kim,P.J.Ryan,T.Heeg,M.Roeckerath,V.Goian,M.Bernhagen,R.Uecker,P.C.Hammel,K.M.Rabe,S.Kamba,J.Schubert,J.W.Freeland,D.A.Muller,C.J.Fennie,P.Schiffer,V.Gopalan,E.Johnston-Halperin,andD.G.Schlom, Nature 466,954(2010). [44] V.Leca,Ph.D.thesis,UniversityofTwente(2003). [45] K.M.Rabe,C.H.Ahn,andJ.-M.Triscone,PhysicsofFerroelectrics:AModernPerspective(Springer,Berlin,2007). [46] M.Kareev,S.Prosandeev,J.Liu,C.Gan,A.Kareev,J.W.Freeland,M.Xiao,andJ.Chakhalian, Appl.Phys.Lett. 93,061909(2008). [47] G.Koster,B.L.Kropman,G.J.H.M.Rijnders,D.H.A.Blank,andH.Rogalla, Appl.Phys.Lett. 73,2920(1998). [48] T.Venkatesan,X.D.Wu,A.Inam,andJ.B.Wachtman, Appl.Phys.Lett. 52,1193(1988). [49] Nanotechnologymeasurementhandbook(Keithley,Cleveland,OH,2007). [50] S.H.Yun,T.Dhakal,D.Goswami,G.Singh,A.Herbard,andA.Biswas, J.Appl.Phys. 103,07E317(2008). [51] A.Biswas,M.Rajeswari,R.C.Srivastava,Y.H.Li,T.Venkatesan,R.L.Greene,andA.J.Millis, Phys.Rev.B 61,9665(2000). [52] Y.Suzuki,H.Y.Hwang,S.-W.Cheong,T.Siegrist,R.B.vanDover,A.Asamitsu,andY.Tokura, J.Appl.Phys. 83,7064(1998). [53] T.Z.Ward,J.D.Budai,Z.Gai,J.Z.Tischler,L.Yin,andJ.Shen, Nat.Phys. 5,885(2009). [54] W.S.Choi,Z.Marton,S.Y.Jang,S.J.Moon,B.C.Jeon,J.H.Shin,S.S.A.Seo,T.W.Noh,K.Myung-Whun,H.N.Lee,andY.S.Lee, J.Phys.D:Appl.Phys. 42,165401(2009). [55] W.Prellier,P.Lecoeur,andB.Mercey, J.Phys.:Condens.Matter 13,R915(2001). [56] Z.Marton,S.S.A.Seo,T.Egami,andH.N.Lee, J.Cryst.Growth 312,2923(2010). [57] H.Boschker,M.Mathews,E.P.Houwman,H.Nishikawa,A.Vailionis,G.Koster,G.Rijnders,andD.H.A.Blank, Phys.Rev.B 79,214425(2009). [58] A.Bhattacharya,S.J.May,S.G.E.teVelthuis,M.Warusawithana,X.Zhai,B.Jiang,J.-M.Zuo,M.R.Fitzsimmons,S.D.Bader,andJ.N.Eckstein, Phys.Rev.Lett. 100,257203(2008). 96

PAGE 97

[59] H.L.Ju,J.Gopalakrishnan,J.L.Peng,Q.Li,G.C.Xiong,T.Venkatesan,andR.L.Greene, Phys.Rev.B 51,6143(1995). [60] W.Prellier,M.Rajeswari,T.Venkatesan,andR.L.Greene, Appl.Phys.Lett. 75,1446(1999). [61] H.M.ChristenandG.Eres, J.Phys.:Condens.Matter 20,264005(2008). [62] A.Urushibara,Y.Moritomo,T.Arima,A.Asamitsu,G.Kido,andY.Tokura, Phys.Rev.B 51,14103(1995). [63] P.Orgiani,R.Ciancio,A.Galdi,S.Amoruso,andL.Maritato, Appl.Phys.Lett. 96,032501(2010). [64] H.Y.Hwang,S.-W.Cheong,N.P.Ong,andB.Batlogg, Phys.Rev.Lett. 77,2041(1996). [65] T.Taniyama,M.Yamasaki,andY.Yamazaki, Appl.Phys.Lett. 81,4562(2002). [66] A.Goyal,M.Rajeswari,R.Shreekala,S.E.Loand,S.M.Bhagat,T.Boettcher,C.Kwon,R.Ramesh,andT.Venkatesan, Appl.Phys.Lett. 71,2535(1997). [67] N.D.Mathur,M.H.Jo,J.E.Evetts,andM.G.Blamire, J.Appl.Phys. 89,3388(2001). [68] Forthicknessmeasurement,thethinlmswerepartiallydippedintheetchingsolution(H2O+KI+10%HCl)for10-40seconds.Andthen,thestep,createdbyetchingsolution,weremeasuredwithatomicforcemicroscope. [69] J.A.Collado,C.Frontera,J.L.Garca-Muoz,C.Ritter,M.Brunelli,andM.A.G.Aranda, Chem.Mater. 15,167(2003). [70] L.Vasylechko,L.Akselrud,W.Morgenroth,U.Bismayer,A.Matkovskii,andD.Savytskii, J.AlloysCompd. 297,46(2000). [71] P.A.Sharma,S.B.Kim,T.Y.Koo,S.Guha,andS.-W.Cheong, Phys.Rev.B 71,224416(2005). [72] TIMandTMIweredenedasthemaximumorminmumchangeofresistivityasafunctionoftemperature. [73] L.GhivelderandF.Parisi, Phys.Rev.B 71,184425(2005). [74] Withdirectsubtraction,thezeroeldcooled(ZFC)warmingmangetizationmeasurementshowedanegativeatlowtemperature.However,whentheLPCMOwasatneartoFerromagnetictoParamagnetictransition,theriseanddecreaseofmagneticmomentswereobservedatthesametemperature. [75] IntroductiontoMagneticMaterials,editedbyB.D.CullityandC.D.Graham(Wiley-IEEEPress,Hoboken,NJ,2008). 97

PAGE 98

[76] RemantFieldsinMPMPSuperconductingMagnets,1014-208(QuantumDesign,SanDiego,CA,2001). [77] E.F.KnellerandF.E.Luborsky, J.Appl.Phys. 34,656(1963). [78] R.Das,Ph.D.thesis,UniversityofFlorida(2010). [79] R.Skomski, J.Phys.:Condens.Mat. 15,R841(2003). [80] K.BinderandA.P.Young, Rev.Mod.Phys. 58,801(1986). [81] W.Wu,C.Israel,N.Hur,S.Park,S.-W.Cheong,andA.deLozanne, Nat.Mater. 5,881(2006). [82] S.Dong,S.Yunoki,X.Zhang,C.Sen,J.-M.Liu,andE.Dagotto, Phys.Rev.B 82,035118(2010). [83] Y.Suzuki,H.Y.Hwang,S.-W.Cheong,andR.B.vanDover, Appl.Phys.Lett. 71,140(1997). [84] C.Kittel, Rev.Mod.Phys. 21,541(1949). [85] D.Givord,P.Tenaud,andT.Viadieu, IEEET.Magn. 24,1921(1988). [86] D.Elbaz,D.Givord,S.Hirosawa,F.P.Missell,M.F.Rossignol,andV.Villas-Boas, J.Appl.Phys. 69,5492(1991). [87] E.Kondorsky,J.Phys.(USSR)2,161(1940). [88] S.Reich,S.Shtrikman,andD.Treves, J.Appl.Phys. 36,140(1965). [89] Z.-H.Wang,G.Cristiani,andH.U.Habermeier, Appl.Phys.Lett. 82,3731(2003). [90] A.Ohtomo,D.A.Muller,J.L.Grazul,andH.Y.Hwang, Nature 419,378(2002). [91] A.OhtomoandH.Y.Hwang, Nature 427,423(2004). [92] J.MannhartandD.G.Schlom, Science 327,1607(2010). [93] S.Smadici,P.Abbamonte,A.Bhattacharya,X.Zhai,B.Jiang,A.Rusydi,J.N.Eckstein,S.D.Bader,andJ.-M.Zuo, Phys.Rev.Lett. 99,196404(2007). [94] G.J.H.M.Rijnders,G.Koster,D.H.A.Blank,andH.Rogalla, Appl.Phys.Lett. 70,1888(1997). [95] J.Kobayashi,Y.Tazoh,andS.Miyazawa, J.Cryst.Growth 131,138(1993). [96] U.Pietsch,V.Holy,andT.Baumbach,High-ResolutionX-rayScattering(Springer,Berlin,2004)p.145. [97] J.-S.Lee,D.A.Arena,P.Yu,C.S.Nelson,R.Fan,C.J.Kinane,S.Langridge,M.D.Rossell,R.Ramesh,andC.-C.Kao, Phys.Rev.Lett. 105,257204(2010). 98

PAGE 99

[98] S.SinghandM.Fitzsimmons,(privatecommunication). [99] G.Garbarino,C.Acha,P.Levy,T.Y.Koo,andS.-W.Cheong, Phys.Rev.B 74,100401(2006). [100] P.D.BealeandP.M.Duxbury, Phys.Rev.B 37,2785(1988). [101] S.Dong,H.Zhu,andJ.-M.Liu, Phys.Rev.B 76,132409(2007). [102] Y.Tomioka,A.Asamitsu,H.Kuwahara,Y.Moritomo,andY.Tokura, Phys.Rev.B 53,R1689(1996). [103] F.Tsui,M.C.Smoak,T.K.Nath,andC.B.Eom, Appl.Phys.Lett. 76,2421(2000). [104] J.C.Jiang,E.I.Meletis,andK.I.Gnanasekar, Appl.Phys.Lett. 80,4831(2002). [105] J.Dho,Y.N.Kim,Y.S.Hwang,J.C.Kim,andN.H.Hur, Appl.Phys.Lett. 82,1434(2003). [106] M.Egilmez,M.M.Saber,A.I.Mansour,R.Ma,K.H.Chow,andJ.Jung, Appl.Phys.Lett. 93,182505(2008). [107] Q.Li,H.S.Wang,Y.F.Hu,andE.Wertz, J.Appl.Phys. 87,5573(2000). [108] I.C.Infante,V.Laukhin,F.Sanchez,J.Fontcuberta,O.Melnikov,O.Y.Gorbenko,andA.R.Kaul, J.Appl.Phys. 99,08C502(2006). [109] W.Eerenstein,F.D.Morrison,J.F.Scott,andN.D.Mathur, Appl.Phys.Lett. 87,101906(2005). [110] M.Gajek,M.Bibes,A.Barthelemy,K.Bouzehouane,S.Fusil,M.Varela,J.Fontcuberta,andA.Fert, Phys.Rev.B 72,020406(2005). [111] J.Y.SonandY.-H.Shin, Appl.Phys.Lett. 93,062902(2008). [112] A.J.HattandN.A.Spaldin, Eur.Phys.J.B 71,435(2009). [113] H.Yang,Z.H.Chi,J.L.Jiang,W.J.Feng,J.F.Dai,C.Q.Jin,andR.C.Yu, J.Mater.Sci. 43,3604(2008). [114] H.Chiba,T.Atou,andY.Syono, J.SolidStateChem. 132,139(1997). [115] A.F.M.dosSantos,A.K.Cheetham,W.Tian,X.Q.Pan,Y.F.Jia,N.J.Murphy,J.Lettieri,andD.G.Schlom, Appl.Phys.Lett. 84,91(2004). [116] H.Faqir,H.Chiba,M.Kikuchi,Y.Syono,M.Mansori,P.Satre,andA.Sebaoun, J.SolidStateChem. 142,113(1999). [117] S.Havelia,S.Wang,M.Skowronski,andP.A.Salvador, J.Appl.Phys. 106,123509(2009). 99

PAGE 100

[118] J.B.Goodenough, Phys.Rev. 100,564(1955). [119] Y.ChenandJ.Washburn, Phys.Rev.Lett. 77,4046(1996). [120] R.IgrejaandC.J.Dias, Sens.ActuatorA-Phys. 112,291(2004). [121] W.Eerenstein,F.D.Morrison,F.Sher,J.L.Prieto,J.P.Atteld,J.F.Scott,andN.D.Mathur, Philos.Mag.Lett. 87,249(2007). [122] P.R.MickelandA.F.Hebard,(privatecommunication) 100

PAGE 101

BIOGRAPHICALSKETCH HyoungJeenwasborninPusan(now,Busan),SouthKoreaasasonofKangjaLeeandGwangsooJeen.Beforeheenteredhighschool,hewasmoreinterestedinsports,viz.taekwondo,basketball,andbaseball.Inhighschool,hisinterestsgotshiftedtoChemistryandPhysics.Hewasalsoinuencedbyhisfather,whoisaPhysicsprofessorinPusanNationalUniversity(PNU).HewenttoPNUandmajoredPhysicsashisfatherdid.Afteroneyearinthecollege,hebecameadrillsergeantinKoreanArmyTrainingCenterforhisnationalduty.Aftergettingbackfromtheduty,hehadvariouschancestolearncondensedmatterphysicsandmetexcellentphysicists.Then,hedecidedtostudyabroad.HegotintoDepartmentofPhysics,UniversityofFloridaatFall,2005.HeluckilyjoinedDr.BiswasgroupatJuly,2008.Whileworkinginhisgroup,hehaslearnedmanythingsnotonlyforsciencebutalsoforlifefromDr.Biswasandhisstudents.Healsometmanynicecolleaguesandprofessorsintheschoolandprofessionalconferences,sothathecanexpandhisknowledges. 101