supplementary information - nature research · thorsten schmitte, laurent c. dudac, matthew r....

In the format provided by the authors and unedited.

© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

SUPPLEMENTARY INFORMATIONDOI: 10.1038/NCHEM.2923

NATURE CHEMISTRY | www.nature.com/naturechemistry 1

1

SupplementaryInformation

Oxygenredoxchemistrywithoutexcessalkali-metalionsinNa2/3[Mg0.28Mn0.72]O2

UrmimalaMaitraa†,RobertA.Housea†,JamesW.Somervillea,NuriaTapia-Ruiza,JuanLozanoa,NiccolóGuerrinia,RongHaoa,KunLuoa,LiyuJina,MiguelA.Pérez-Osorioa,FelixMasselc,DavidM.Pickupd,SilviaRamosd,XingyeLue,DanielE.McNallye,AlanV.Chadwickd,FelicianoGiustinoa,ThorstenSchmitte,LaurentC.Dudac,MatthewR.Robertsa,PeterG.Bruceab*

†jointfirstauthoraDepartmentofMaterials,UniversityofOxford,ParksRoad,OxfordOX13PH,UK.bDepartmentofChemistry,UniversityofOxford,ParksRoad,OxfordOX13PH,UK.cDepartmentofPhysicsandAstronomy,DivisionofMolecularandCondensedMatterPhysics,UppsalaUniversity,Box516,S-75120Uppsala,Sweden.dSchoolofPhysicalSciences,UniversityofKent,Canterbury,KentCT27NH,UK.eSwissLightSource,PSI,5232Villigen,Switzerland.

Tableofcontents

MethodsSection----------------------------------------------------------------------------------------------------------2

SupplementaryFigure1-----------------------------------------------------------------------------------------------4

SupplementaryTable1------------------------------------------------------------------------------------------------5










SupplementaryFigure10----------------------------------------------------------------------------------------------15




SupplementaryFigure14---------------------------------------------------------------------------------------------19

SupplementaryFigure15---------------------------------------------------------------------------------------------20


References-----------------------------------------------------------------------------------------------------------------21

http://dx.doi.org/10.1038/nchem.2923

2

MethodsSection

PowderX-raydiffraction(PXRD)patternswererecordedona9KWRigakuSmartlabdiffractometerusingCuKα1radiation(λ=1.54051Å).ThePXRDdatawereanalysedusingtheRietveldrefinementmethodasimplementedintheGSASsoftwaresuitewiththeEXPGUIsoftwareinterface.SamplesweremountedincustomizedholderswithaKaptonwindowtoavoidanyexposuretotheatmosphereandmeasurementswerecarriedoutinreflectionmode.In-situXRDwascarriedoutinanin-situcellwithanX-raytransparentberylliumwindow.AverythinAlfilmwasplacedbetweentheBewindowandthecathodetopreventBeoxidationathighpotentials(around4.5VvsNa+/Na).(Furtherinformationaboutthecelldesigncanbeobtainedatwww.rigaku.com).ThecellwascontrolledbyaBiologicMPGpotentiostat.

ElectrochemistryElectrodeswerepreparedbymixing80wt%activematerial,10wt%SuperPcarbonand10wt%polytetrafluoroethylene(PTFE)binder(inamortarpestleandrollingoutthinfreestandingfilmsofthemixture.FortestingtheeffectofaddingNa2CO3intotheelectrodesadditionalelectrodeswere prepared containing 70 wt% active material, 10 wt% Super P carbon and 10 wt%polytetrafluoroethylene(PTFE)binderand10wt%Na2CO3.ElectrochemicaltestingwascarriedoutincoincellswithaNametal-diskastheanodeanda1MNaPF6(AlfaAesar,≥99.0%) inbatterygradepropylene carbonate (PC) electrolyte (BASF Selectilyte). The PC was distilled using a packed bedcolumnanddriedforseveraldaysoverfreshlyactivatedmolecularsieves(type4Å)priortomakingup the electrolyte. NaPF6 (Alfa)was dried at 60◦C under vacuumbefore preparing the electrolytesolution.Galvanostaticcharge-dischargewascarriedoutinCR2032coincellsusingaMaccorSeries4000atarateof10mAg-1.

Operandodifferentialelectrochemicalmassspectrometry(DEMS)analysiswascarriedouttostudythe different gases generated during cell cycling. The set up consisted of a quadrupole massspectrometer(ThermoFischer)equippedwithturbomolecularpump(PfeifferVacuum)andmass-flowcontrollers (Bronkhorst).Twoelectrode typecells (ECC-Std fromEL-CELL)withgas inletandoutletportswereusedfortheoperandomeasurements.ThecellconsistedofNaanode,1MNaPF6inPCelectrolyteandthesamecathodeasdescribedabove.MoredetailsoftheDEMSset-uparegiveninapreviouspublication.1

Mn K-edge X-ray absorption near edge structure (XANES) measurements were undertaken atbeamlineB18at theDiamondLightSource,Harwell,UK.Thebeamline isequippedwithadouble-crystalmonochromator(withtwocrystalsSi(111)andSi(311))andworksintherange2.05–35keV.TheXANESspectrawerecollectedintransmissionmodeandtheintensitiesofboththeincidentandtransmittedX-raybeamsweremeasuredusinggas-filledionisationchambers.Tocorrectforanydriftinmonochromator,Mnmetalfoilwasplacedinfrontofathirdionisationchamber.Foreachsamplethree scans were taken, summed, calibrated, background subtracted, and normalised using theprogramAthena.Mn2O3andMnO2wereusedasreferencesforMn3+andMn4+,respectively.

SoftXASandRIXS TheOK-edge soft x-ray absorption spectra (SXAS) and resonant inelastic x-rayscattering(RIXS)spectrawererecordedattheADRESSbeamlineoftheSwissLightSource,PSI,usingtheSAXESspectrometer.2,3ToobtainSXASspectraattheOK-edge,wesimultaneouslyrecordedthetotalfluorescenceyield(TFY)signalusinganx-raysensitivephotodiodeandthetotalelectronyield(TEY)signalbymeasuringthesampledraincurrent.TFYdatawasusedforanalysisduetothelowersignaltonoiseratiooftheTEYdata,possiblybecauseofconductivityissues.ForrecordingtheO-KSXASaswellastheRIXSspectrathemonochromatorbandwidthwassettoabout40meV.ThetotalresolutionfortheRIXSspectrawasabout55meV.

3

RamanSpectroscopyRamanspectraofthematerialsatdifferentstatesofchargewerecollectedusingaRamanRenishawInViaspectrometerequippedwithadiodelaser(λ=785nm)andalaserpowerof1.5mW. For thesemeasurements all sampleswere sealed between two glass slides under argonatmosphere.

ScanningTransmissionElectronMicroscopy: ADF-STEMandABF-STEMdatawere collectedonanaberration corrected JEOL ARM 200F operated at 200 kV. The convergence semi-angle used was22mrad,andthecollectionsemi-angleswere9.5-20.7mrad(ABF)and69.6-164.8mrad(ADF).Inallcases,setsoffast-acquisitionmulti-frameimageswererecordedandsubsequentlycorrectedfordriftandscandistortionsusingSmartAlign.4

17ONMRmethod: Solution-state 17ONMR experimentswere performed in a BrukerDiff50 probecoupledwitha17Oinsertona400MHzBrukerAvanceIIIspectrometeratthe17OLarmorfrequencyof54.3MHz.Thespectrawererecordedwithzg30pulsesequence;theappliedπ/2pulselengthwas23μs.Allsampleswereloadedin5mmNMRtubessealedwithair-tightcapsinanAr-filledgloveboxand the volumesof the liquid sampleswere kept the same (700μL). The spectrawere externallyreferencedwithwaterat0.0ppm.Theelectrolyteofachargedbatterywascollectedbysoakingitsseparatorinthepristineelectrolytefor24hrsinsideanAr-filledglovebox.

TGA-MS:ThermogravimetricAnalysiswascarriedoutonpowder samplesof~30-40mgquantitiesunderinertAratmosphereusingaNETZSCHJupiterSTA449F3TGA.ThiswascoupledwithaNETZSCHAëolosQMS403Dmassspectrometertoprovideinoperandomassspec.

SQUID:FCmagnetizationmeasurementsoftheAr-treatedpowderwascarriedoutinQuantumDesignInc.SQUID-VSM.

4

Figure1.PXRDpatternofthepristineNa0.67Mg0.28Mn0.72O2materialrefinedinspacegroupP63/mcm.(RefinedparametersareshownintableS1).Thepinktickmarksindicatetheallowedreflections.The2thetaranges(highlightedinlightgrey)havebeenexcludedfromtherefinementaspeaksintheseregionsarose fromtheair sensitivesampleholder.Theblackcurve is theexperimentaldiffractionpattern,theredcurveisthecalculateddiffractionpatternasobtainedfromRietveldrefinements,thegreencurveisthebackgroundsubtractedandthebluecurveisthedifferenceplot.

5

Table1.RefinedparametersoftheNa0.67Mg0.28Mn0.72O2pristinematerial.

AtomsWycoffpositions

X Y Z Occupancy Uiso

Mg1/Mn1 2b 0 0 0 0.814(5)/0.186(2) 0.015(2)/0.018(4)

Mg2/Mn2 4d 1/3 2/3 0 0.011(10)/0.989(5) 0.016(5)/0.016(2)

O 12k 0.354 0.354 0.08 1 0.015(4)

Na1 6g 0.301 0 ¼ 0.397(4) 0.028(8)

Na2 4c 1/3 2/3 ¼ 0.403(4) 0.032(8)

SpacegroupsP63/mcm a=b=5.0095(3),c=11.218(4),χ2=3.3.Stoichiometry fromrefinementNa0.67Mg0.28Mn0.72O2

6

Table2.StoichiometryatvariouspointsofchargeanddischargeasdeterminedfromICPanalysis

CompositionfromICP

Na(±0.027)

Mg(±0.01)

Mn(±0.026)

Pristine 0.671 0.265 0.734

4.5V 0.175 0.266 0.731

Dis2V 0.784 0.266 0.733

7

Figure2.TemperaturedependenceofmagnetizationunderFCcondition.

Theeffectivemagneticmomentwascalculatedfromequation𝜒" =$%&'$())

* $+*

,-+.= 2.827 𝐶=3.479

BM,whereCistheCurieconstantobtainedfromfitting1/χvsT.Forfitting,χ isconvertedintothecgs units (emu/mol.Oe). The effective spin only magnetic moment for compositionNa0.67+Mg0.28Mn(III)0.11Mn(IV)0.61O2 is calculated to be 𝜇566 = 0.11 4.9 ; + 0.61 3.87 ; =3.434BM.However,consideringthe0.14molesofNaextractedintheregion1beingcontributedbyMn(III)/Mn(IV) redox (consistent with electrochemistry), 𝜇566 = 0.14 4.9 ; + 0.58 3.87 ; =3.470BM.

8

Figure3.TGA-MSofNa0.67Mg0.28Mn0.72O2fromwhichNa2CO3hasbeenremoveusingaheatingstepinArgon.Afterheattreatment,nomasslossisobservedinthiscleanedsamplebelow750̊Catwhichpointasmallamountofoxygenislostfromthematerial.TheTGAexperimentwasperformedona~40 mg sample contained in an alumina crucible which was ramped at 10 °C min-1 from roomtemperatureto800°CinArgon.ThisresultconfirmsthatthesematerialsarealmostfreefromNa2CO3impurities.

0

200

400

600

800Te

mpe

ratu

re (

o C)

96

97

98

99

100

Mas

s C

hang

e (%

)

0 20 40 60 80

0.0

2.0E-10

4.0E-10

6.0E-10

8.0E-10 Oxygen m/z=32

Time (mins)

Ion

Cur

rent

(A

)

Carbon Dioxide m/z=44

9

Figure4.SEMimagesoftheNa0.67Mg0.28Mn0.72O2pristinematerial.Theimagesweretakenatanoperatingvoltageof5keV(ZeissGeminiSEM-500)

10

100

125

150

175

200

0 10 20 30 40 502.60

2.65

2.70

2.75

Cap

acity

(mA

h/g)

Charge Discharge

Avg Discharge voltage

Vol

tage

(V)

Cycle no.

Figure 5. Capacity (top) and average discharge voltage (bottom) plotted as a function of cyclenumberforNa0.67Mg0.28Mn0.72O2over50cyclesinthevoltagerange2-4.5Vatarateof10mAg-1.Notthatthecapacityfadinghere,~1mAhg-1percycle,issignificantlylowerthanreportedinref19mainmanuscript(Yabucchiet.al.)wherethevoltagerangeforcyclingwaslarger,from1.5Vto4.4V

11

Figure6.PXRDpatternsrecordedinsituusingacellconstructedwithaNa0.67Mg0.28Mn0.72O2cathodeduring the first cycle. The pattern shown after charging to 100mAhg-1 shows a peak for the O2structureandthisincreasesinintensitybytheendofcharge.OndischargethepeaksoftheoriginalP2structureareregainedfully.Thesampledischargedto2V(orange)showsageneralbroadeningofallthepeaksalongwithareductionofthecparameterandexpansionofthea/bparameter,comparedwith the sample discharged to 2.3 V (yellow). The lattice parameter changes are consistent withinsertionofexcessNa(beyondNa0.67).Peakslabelledwitharrowsrepresentreflectionsfromtheinsitucellcomponents.

12

Figure7:ADF-STEMmicrographafter1cycle,whereahighdensityofstackingfaultsalongthe[1-10]-directioncanbeobserved.ThestackingfaultsresultinstreakingofreflectionsinthefastFouriertransformoftheimageintheinset.

13

Figure 8. Operando mass spectrometry data collected during the first cycle ofNa2/3[Mg0.28Mn0.72]O16

2-xO18xwhere x is approximately 1 as determinedby TGA-MS. An identical

traceofCO2isobservedasreportedinFig3.NoevidenceofanyO2orCO2containingO18wasobservedasreported

2.0

2.5

3.0

3.5

4.0

4.5

0 1000

0.0

2.0

m/z = 35 (16O18O) m/z = 36 (18O18O)m/z = 44 (C16O16O) m/z = 46 (C16O18O) m/z = 48 (C18O18O)

Time (mins)

Flux

CO

2 (x1

0-9 m

ol m

in-1)

DEMS Experiment with 18O-labelled material

Volta

ge (V

)

14

Figure9.17ONMRspectraofpristineelectrolytecontaining1MNaPF6inpropylenecarbonate(bluespectrum)andanelectrolytewhichhadbeenextractedfromabattery(with~50at.%17Oenrichedcathode)redspectrumafterchargingto4.5Vvs.Na/Na+.Thetwospectraare identicalTherefore,oxygenhasnotbeenreleasedfromthematerialandcontainedwithintheelectrolyteinanyform.

15

Figure10.TGA-MSdataonas(a)preparedand(b)chargedelectrodes.TheO18labelledO2andCO2tracesshownoevidenceofnewdecompositionproductscontaining18O.

16

Figure 11. Operando mass spectrometry data collected during the first charge for as-preparedNa2/3[Mg0.28Mn0.72]O2i.e.withnoprocedureforremovalofNa2CO3applied(redcurve),materialwith10wt.%ofNa2CO3intentionallyadded(bluecurve)andNa2/3[Mg0.28Mn0.72]O2afterheattreatmentunderAr(blackcurve).QuantitiesgivenareinunitsofmolesofCO2permoleofactivematerial(AM)

17

Figure12.EvolutionoftheXANESMnK-edge(1s-4p). Insetshowsthepre-edge(1s-3d).Figure inbottomrightpanel shows thevariationofMnoxidation state, calculated from thepositionof thecentroidofthepre-edge,withNacontentonchargethendischarge.Thereisnosignificantchangeinthepeakpositionsofthepre-edgeonchargingacrosstheplateauanduntilaround2.3VondischargeimplyingalmostnochangeinMnoxidationstatethroughtheseregions.LinesshownonthebottomrighthandgraphindicatetheoxidationstatesexpectedforthematerialatvariousstatesofchargewhenconsideringMnoxidationandreductionastheonlymechanismofchargecompensationinthismaterial.

18

Figure13.RamanspectraofsampleswithvarioussodiumcontentspreparedthroughchargingthepristineNa0.67Mg0.28Mn0.72O2material.Samplesa-grepresentvariousstatesofchargeasrepresentedin Fig 4(a). Raman spectra of Li2O2, ZnO2,Na2O2 and the as preparedmaterial are also shown forcomparison.AbsenceofO-Ovibrationsindicatenotrue(1.4ÅO-Obondlength)peroxidespeciesareformed.Notethatthestandardswerechosen(alkaliandtransitionmetal)tospantheenvironmentsexpected around any true peroxo species inNa0.67Mg0.28Mn0.72O2. The feature at around 650 cm-1appears to evolve with charging. Peaks at similar frequencies have also been observed in othermaterialswhichexhibittheP2toO2phasetransitiononcharginge.g.P2-typeNa2/3Mn1/2Fe1/2O2.5Asa result of changes in the Raman spectrum associated with the P2/O2 transition and with thedistortionsoftheOcoordinationenvironmentaroundMn,itisunfortunatelynotpossibletoextractunambiguousinformationontheformationofO2

n-peroxo-typespeciesbycomparingthechargedandpristineRamanspectra.

19

Figure 14. One electron energy level diagram representing M-O more covalent vs less covalentbondinginteractions.MoreionicinteractionslikeMn/Mg-OplacestheO2pstatesatrelativelyhighenergies,whicharethenaccessiblewithinthestabilitywindowoftheelectrolyte.MorecovalentM-OinteractionpushestheO2pstatesdowninenergywellabovethevoltagewindowofelectrolyteoxidation.

20

Figure15.SpinresolvedtotalandpartialdensityofstatesforNa2/3[Mg1/3Mn2/3]O2.ThetotaldensityofstatesisrepresentedbytheblackareawhileredandblueareascorrespondtoO2pandMn3dstates, respectively. TheFermienergy is set to0eV, and is arbitrarilyplaced in themiddleof thecalculatedbandgap,asindicatedbythedashedblackline.Withinthevalenceband,theO2pstatesaredominantnearthetopofthevalenceband,whileMn3dstatesarelocatedprimarilyafeweV'sbelow the top of the valence band. This is consistentwith O oxidation occurring prior toMn4+/5+oxidation.

Table3.CalculatedparametersofthesupercellNa16Mg8Mn16O48(2x2x1supercell)

a b c α β γ CalculatedfromDFT

10.238Å 10.238Å 11.271Å 90.000° 90.000° 120.015°

Deviationfromexperiment

2.19% 2.19% 0.47% 0.00% 0.00% 0.01%

21

References

1. Luo,K.etal.Charge-compensationin3d-transition-metal-oxideintercalationcathodesthroughthegenerationoflocalizedelectronholesonoxygen.Nat.Chem.8,684–691(2016).

2. Strocov,V.N.etal.High-resolutionsoftX-raybeamlineADRESSattheSwissLightSourceforresonantinelasticX-rayscatteringandangle-resolvedphotoelectronspectroscopies.J.SynchrotronRadiat.17,631–643(2010).

3. Schmitt,T.etal.High-resolutionresonantinelasticX-rayscatteringwithsoftX-raysattheADRESSbeamlineoftheSwisslightsource:Instrumentaldevelopmentsandscientifichighlights.J.ElectronSpectros.Relat.Phenomena188,38–46(2013).

4. Jones,L.etal.SmartAlign—anewtoolforrobustnon-rigidregistrationofscanningmicroscopedata.AdvancedStructuralandChemicalImaging,8,1-16(2015).

5. MortemardDeBoisse,B.,Carlier,D.,Guignard,M.,Bourgeois,L.&Delmas,C.P2-NaxMn1/2Fe1/2O2phaseusedaspositiveelectrodeinNabatteries:Structuralchangesinducedbytheelectrochemical(De)intercalationprocess.Inorg.Chem.53,11197–11205(2014).

supplementary information - nature research · thorsten schmitte, laurent c. dudac, matthew r....

Documents