REMOTERAMANDETECTIONOFNATURALROCKS

AFINALREPORTSUBMITTEDTOTHEDEPARTMENTOFGEOLOGYANDGEOPHYSICS,UNIVERSITYOFHAWAI’I

ATMĀNOA,INPARTIALFULFILLMENTOFTHEREQUIREMENTSFORTHEDEGREEOF

MASTEROFSCIENCEIN

GEOLOGYANDGEOPHYSICS

August2017

ByGenesisBerlanga

Advisor:

Dr.AnupamK.Misra

RemoteRamanDetectionofNaturalRocksGENESISBERLANGA,1,2*TAYROE.ACOSTA-MAEDA,1SHIVK.SHARMA,1JOHNN.PORTER,1PRZEMYSLAWDERA,1HANNAHSHELTON,1,2G.JEFFREYTAYLOR,1ANUPAMK.MISRA11Hawai’iInstituteofGeophysicsandPlanetology,UniversityofHawai’i,POST#602,1680East-WestRoad,Honolulu,Hawaii96822,USA2DepartmentofGeologyandGeophysics,UniversityofHawai’i,POST#702,1680East-WestRoad,Honolulu,Hawaii96822,USA*Correspondingauthor:genesis.berlanga@gmail.com

ReceivedXXMonthXXXX;revisedXXMonth,XXXX;acceptedXXMonthXXXX;postedXXMonthXXXX(Doc.IDXXXXX);publishedXXMonthXXXX

We report remote Raman spectra of natural igneous, metamorphic, and sedimentary rock samples at astandoffdistanceof5m.HighqualityremoteRamanspectraofunpreparedrocksarenecessaryforaccurateandrealisticanalysisoffutureRamanmeasurementsonplanetarysurfacessuchasMars.OurresultsdisplaytheabilityofaportableCompactRemoteRaman+LIBS+FluorescenceSystem(CRRLFS) toeffectivelydetectand isolate various light and dark-colored mineral phases in natural rocks. The CRRLFS easily detectedplagioclase and potassium feldspar end members, quartz, and calcite in rocks with high fluorescencebackgrounds.Intermediatefeldsparsandquartz,whenfoundinrockswithcomplexmineralogies,exhibitedband shifts and broadening in the 504-510 cm-1 and 600-1200 cm-1 regions. A good approximation ofintermediateplagioclase feldsparswaspossiblebyusingspectralshapeandassigningotherminorRamanpeaks inaddition to the504-510cm-1peaks.Detectionofolivineandpyroxene inmaficrocksallowed forcompositionalcharacterization.OCIScodes:(120.0120)Instrumentation,measurement,andmetrology;(220.0220)Opticaldesignandfabrication;(280.0280)Remotesensingandsensors;(290.5860)Scattering,Raman;(300.6450)Spectroscopy,Raman;(300.6500)Spectroscopy,time-resolved.

http://dx.doi.org/10.1364/AO.99.099999

1.INTRODUCTIONRamanspectroscopy isanotableactive laserspectroscopy

technique that provides molecular and crystal arrangementinformationaboutthetargetmaterialenablingidentificationofbothorganicsandinorganics.Thetechniqueinvolvesshininglightfromamonochromaticlaser,suchasa532nmlaser,onasampleandcollectingtheresultingscatteredlight.Afractionofthescatteredlightisshiftedinenergyfromthelaserfrequencydue to interactions between the incident electromagneticwaves and the vibrational modes of the molecules in thesample. This inelastic shift called the Raman effect occurseffectivelyinstantaneouslyatpicosecondtimescalesof10-12s[1],[2].ARamanspectrumisproducedwithbandpositionsatfrequencies corresponding to the energy levels of differentvibrationalmodes,univocallyidentifyingatargetsample[2].NASA’sMarsScienceLaboratory (MSL)missionpaved the

way for active laser spectroscopy on a planetary surface.CuriosityRover’shighlysuccessfulChemCaminstrumentsuitecontinues to acquire laser-induced breakdown spectroscopy

(LIBS) measurements from standoff distances providingquantitativeandqualitativeelemental compositionsofmajorandminormineralsonthesurfaceofMars[3].Thissuccessfullaser operation on Mars has opened opportunities forcomplementarylaserbasedtechniquessuchasremoteRamanspectroscopy, to provide additional mineralogical analyses.Raman spectroscopy can identify mineral structure,distinguishbetweenmineralpolymorphs suchas calcite andaragonite, and the accompanying fluorescence signal canidentifybiosignatures[4].Acombinedapproachofusingtime-resolved Raman along with LIBS and fluorescencespectroscopy in one compact system, as seen in the futureMars 2020 SuperCam instrument, will help in identifyingvarious hydrated and biological minerals which will help inNASA’s goal for search of life during planetary surfaceexplorations[4],[5].With this in mind, it is not surprising that time-resolved

remote Raman spectroscopy is increasingly highlighted as aviablemethodforplanetarysurfacemolecularanalysis[4],[6]–

[16]. This technique extends traditional micro-Ramantechniquestostandoffdistancesofuptohundredsofmeters,requiresnosamplepreparation,causesminimaltonosampledamage, and can be used under daylight conditions; savingtimeandincreasingthenumberofaccessibletargets[7],[17].The University of Hawai‘i (UH) has developed a CompactRemote Raman+LIBS+Fluorescence System (CRRLFS) that iscapable of Raman, LIBS, and fluorescence measurementsunder daytime conditions from standoff distances of severalmeters[11].TheCRRLFSshowninFigure1isoneofseveralin-house remote Raman systems developed for chemical andmolecular analysis of liquid, gases, and solids, at distancesbetween1.5-430m[18]–[25].Studies by other laboratory groups have attempted to

analyze various pure minerals, powders, and a selection ofrocks using standoff and micro-Raman spectroscopy forplanetary science applications. Micro-Raman systems use averysmalllaserspotsizeintheorderof10µmindiameterforsampleanalysis.Whenstudyingrockswithmultipleminerals,thesesystemsnormallygivespectraofmineralsthatfallwithinthesmalllaserspotsize.Hence,itisdifficulttoidentifyarockwith a singlemicro-Ramanmeasurement. To overcome thisproblem point-counting approaches are employed duringwhichmultiple spectra aremeasured at variousparts of therocktodeterminethepresenceofallmineralphases.InremoteRamansystemsalaserspotsizeofseveralmillimetersisusedand the Raman spectra obtained show the presence of allmineralphasesfallingwithinamuchlargerlaserspotsize.Thistechniqueprovidesbulkrockmineralogy.Previous micro-Raman studies have focused on point-

counting approaches to accurately characterize bulk rockmineralogies [26]–[29], hand-held Raman studies illustratedthe possibility of effective Ramanmeasurements in the field[8], [30], and remoteRaman studies using fiber-coupled [4],[9], [22], [23], [31]–[33] and directly coupled optics andspectrograph performed bulk sample analysis withcentimeter-sized target areas at standoff distances [34].PreviousstudiesbytheUniversityofHawai‘iatMānoaRamanresearchgrouphavedemonstratedstandoffRamandetectioncapability for a variety of minerals, rocks, and powderchemicals using fiber-coupled optics and spectrograph [35].Our directly coupled system design showed significantimprovement inremoteRamandetectioncapabilityover thefiber-coupled system [36]. These directly coupled systemsweredevelopedinobliquegeometriesfrom2-50m[21],[37],andalsoincoaxialgeometriesfrom1.5-430m[11],[18]–[20],[24],[25],[36].Theadvantagesofthecoaxialsystemovertheoblique systems are large sampling depth and easy systemalignment. Previouswork using coaxial systemswasmostlyfocused on obtaining remote Raman spectra of variousminerals such as feldspars, hydratedminerals, dark coloredminerals, various organic and inorganic chemicals, water,frost, ice, dry ice etc. However, to interpret in situ remoteRaman measurements from rocks and soils on planetarysurfaces such as Mars, standoff Raman measurements ofnatural, unprepared rock samples are needed. Thesemeasurements will be helpful in evaluating future remote

RamandatafromMarsusingtheSuperCaminstrumentontheMars2020mission.

Fig.1.CRRLFSschematic(above)andactualsystem(below).

The CRRLFS has been able to successfully acquire highquality Raman spectra of various light and dark minerals,water,water-ice,CO2ice,andorganicandinorganicchemicalsatdistancesofupto50mwitha10sintegrationtime[37].Inthis study several igneous, sedimentary, and metamorphicrockswere surveyed using the CRRLFS.We report accurateandunivocalidentificationofvariousmineralphasesinnaturalrocksamplesusingtheCRRLFSfroma5mstandoffdistanceusinga30sintegrationtime.X-raydiffraction(XRD)analyseswere obtained for all samples to establish compositionalgroundtruthforcomparisonwiththestandoffRamanspectra.

2.METHODOLOGY

A.SamplesTwenty four rock samples were selected from Ward’s

Collection of ClassicNorthAmericanRocks 45-7250,Ward’sRock Forming Mineral Student Set 45 H 0355, and handsamples from the Raman Spectroscopy Laboratory at theUniversityofHawai‘iatMānoa.Mostrocksmeasuredbetween2-8 cm in length at their longest dimensions. Alkalic granite(Quincy, MA, USA), nepheline syenite (Bancroft, Ontario,Canada), tonalite (SanDiegoCounty, CA,USA), biotite gneiss(Uxbridge, MA, USA), grey sandstone (Berea, OH, USA), redsandstone(Potsdam,NY,USA),quartzconglomerate(FremontCounty, CO, USA), quartzite (Baraboo, WI, USA), pyroxenite(harzburgite) (Stillwater Complex, MT, USA), and dunite(olivine peridotite) (Balsam, NC, USA) were chosen due tohavingmorethanonemineralcomponent.Pinkmarble(Tate,GA, USA), dolomite marble (Essex County, NJ, USA), calcite

(unknown locality), and siderite carbonatite (Iron Hill, CO,USA)werechosenduetotheirsystematicchemicaldifferences.Albite (Bancroft, Ontario, Canada), oligoclase (unknownlocality),andesine(Keystone,SD,USA),labradorite(unknownlocality),bytownite(CrystalBay,MS,USA),anorthite(Miyaka-jima, Tokyo, Japan), orthoclase (Gothic, CO, USA),microcline(Madawaska,Ontario,Canada),andmilkyquartz(Dekalb,NY,USA) were chosen for use as reference minerals to identifyindividual components in multiminerallic rocks. Thesereference minerals include quartz, the entire plagioclasefeldspargroup,andpartofthealkalifeldspargroup.Several igneous rocks with varied mineral compositions

wereanalyzed.Nephelinesyeniteisanigneousintrusiverockcomposed primarily of nepheline and potassium feldspar.Tonalite (quartz diorite with a quartz content ~20% of therock) is an igneous intrusive rock containing quartz, biotite,plagioclase feldspars, and orthoclase feldspars. Nephelinereacts with quartz to produce potassium feldspars such asorthoclase. Alkalic granite is an igneous rock generallycomposedofquartz,potassiumfeldspars,andminormicasandhornblende.Pyroxenite(harzburgite)iscomposedprimarilyofpyroxene, and someminor clinopyroxene or orthopyroxene.Dunite(olivineperidotite)containsover90%olivine.Twometamorphic rocks were studied: Biotite gneiss is a

highgrademetamorphicrockwithbandeddarkbiotitemicaandlighterfeldsparandquartz.Quartziteiscomposedofpurequartz sandstone heated and metamorphosed throughtectonic orogenic compression. Three sandstones were alsomeasured: Grey sandstone from Berea, OH is historicallycomposed of quartz, feldspar, and organic detritus typical ofargillaceous sandstones.Red sandstone fromPotsdam,NY ispredominantly quartz with accessory hematite impurities.Quartz conglomerate from Fremont County, CO consists ofmostlyquartzandplagioclasecementedbylimestone.Igneousandclaymineralswereusedasreferencestandards

for rockcomponent identification.Pinkanddolomitemarbleareprimarilycomposedofcalcitethedifferencebeingthatpinkmarble has Fe cations and dolomitemarble hasMg cations.Sideriteisacarbonatitethathasironinsteadofcalcium.

B.InstrumentationandsoftwareTheCRRLFSincludesalaser,collectionoptics,spectrometer,

and an intensified charge-coupled device (ICCD) camera.Instrumentschematicsareshown inFigure1.Theexcitationsource and collection optics are arranged in a coaxialconfiguration. The instrument contains a small 532 nm Q-switchedfrequency-doubledNd:YAGlasersourcefromBigSkyUltraLasersoperatingat20Hz(20mJ/pulse),anadjustable10xbeamexpander,a45°,532nmlasermirrorfromEdmundOptics (λ/4optical flatness), a7.62 cmdiametermirror lens(Bower) used as a collection telescope (f500mm at F/8), aSemrock 532 nm long pass filter, a 50 µm slit, a Holoplexstacked volume phase transmission grating from KaiserOpticalSystems,Inc.coveringthespectralrangefrom534-611nmand609-696nm,andacooled,miniatureSyntronicsICCD(1408x1044pixels).Thecompactspectrographis10cmlongx8.2cmwidex5.2cmtall.Thesystem isexternally triggered

using a BNC Model 575 pulse-delay generator and customsoftwarecontrolsdarkimageandRamanacquisitions.Agooddescription of the instrument details, gating, and designparameterscanbefoundinGasdaetal.(2015)[11].Micro-Raman measurements using μm-sized laser spots

requirescanandpointcountingtechniquestoprovidewholerockanalysis.ByusingstandoffRamanwebenefitfromlargersizebeamdiameters(5-10mm)capableofwholerockanalysisthat reduce measurement times to seconds and at mostminutes. Time-resolved Raman capabilities down tonanosecond timescales have enabled full Ramanmeasurements while eliminating daylight backgroundillumination and minimizing long-lived mineralphosphorescence. By using a directly coupled systemdesignweincreasesensitivitybyavoidingadditionalscatteringwithinthefiberopticcableandlossesattheopticalconnections.Thiscombinationof systemcomponents andarchitectures allowsfor a compact system to strike a fine balance between anadequateSNRandfastdataacquisitiontimes.We verified our standoff Raman spectra through micro-

Raman spectra using a Renishaw inVia confocal Ramanmicroscope using 514.5 nm laser light. Sample bulkcompositionwasdeterminedthroughhighresolutionpowderX-ray diffraction (XRD) analysis using a Bruker D8 Advanceinstrument operating with a copper Kα X-ray source and aLynxIXElineararraydetector.Crystallinephaseidentificationwas determined through Diffrac.eva and JADE softwarepackages using a search and match technique with theincluded Crystallography Open Database (COD). QualitativeRietveldanalysiswasconductedusingPDIndexersoftwaretodeterminecrystalstructuresandthechemicalcompositionofthesamplemineralcomponents.

C.MethodsStandoffRamanspectrawereacquired for30s (600 laser

pulses)withalaserspotsizeof5mmandalaserpulseenergyof20mJ/pulse.Themeasurements foreachtargetareaonarock were taken three times for verification of systemreproducibility. The Raman signal appears as a two tracespectralimageontheICCDuponpassingthroughtheslitanddiffraction gratings [11]. The two diffraction gratings arestackedandangledslightlydifferentlyfromeachothersothatone tracecontains the lower frequenciesandtheother tracecontainsthehigherfrequencies.Eachtraceisverticallybinnedacross all frequencies and the resulting sums (one for eachpixelcolumn)areplottedintoaspectrum.Spectraweredarksubtractedtoremoveelectronicnoise.Wavelengthcalibrationwas conducted using a neon lamp, sulfur and cyclohexaneRamanspectra,andbariteLIBSspectraasstandards.FortheXRDmeasurements,rocksampleswerepowderedandsieveddownto125μm.Thesamplesweremeasuredfor15minuteseach while on a flat plate rotating stage. Bruker’s DiffracsoftwareprocessedtheXRDimageandplottedthespectruminCPS(counts/s).XRDspectraweredarksubtracted,baselinecorrected as needed, and Kα lines were removed. The 2θvalueswere carefully calibratedusing aNIST-supplied (SRM1976)sintered corundum standard. Micro-Raman

measurementswere typically takenwith10sexposure timeand 10 software accumulations with a 50xmicroscope lensand 2 mW laser power. Micro-Raman spectra were darksubtracted and wavelength calibration was based on Neemissionspectrumandadjustedusingasilicawaferstandard.

3.RESULTSANDDISCUSSION

A.RemoteRamanSpectraThe CRRLFS is capable of fast data acquisition yielding

qualitydatawithhighsignaltonoiseratio.WepresentremoteRaman spectra of several igneous, sedimentary, andmetamorphic rocks at 30 s integration times in order toobserve the subtler Raman peaks. In most cases, the samepeaks are visible at 1 s integration times for the lighter andmixed light and dark sample targets. Darker sample targetsrequired the full 30 s integration time due to the increasedmineralopacityandresultingdecreasedlaserpenetration.

Fig.2.a)RemoteRamanspectraofcarbonatesata5mdistanceovera30sintegrationtime.Spectrawereverticallyshiftedforviewingclarity.b)Analyzedsamplesfromlefttoright:dolomite,pinkmarble,calcite,siderite.Lasertargetwithinredcircle.

Cation Mineral(formula) ObservedRamanPeaks(cm-1)

Ca Calcite(CaCO3) 159.4 283.8 712.4 1086 1435 1749

Ca-Fe PinkMarble(CaFe(CO3)2) 159.4 286 712.4 1086 1437 1749

Fe Siderite(FeCO3) 177.7 301.6 724.7 1095 1442 —

CaMgDolomiteMarble

(CaMg(CO3)2)177.7 301.6 724.7 1097 1442 1760

Symmetry Eg Eg Eg A1g Eg Eg

Table1.TabulatedRamanpeaksforCa,Fe,andMgbearingcarbonates.

Calcite (CaCO3), pink marble (CaFe(CO3)2), sideritecarbonatite(FeCO3carbonatewithaMg:Feratioof~0.25),anddolomite marble (CaMg(CO3)2) were analyzed at a 5 mdistance over 30 s integration times. They each displayed

strongcarbonateRamanpeaksat1085.95,1086.01,1095.45,and1097.34cm-1,respectively,showinganincreasingRamanshift from pure calcite to iron-bearing carbonates tomagnesium-bearing carbonates as shown in Figure 2. ThissameprogressionwasalsovisibleinthesubtlerRamanpeakstabulated in Table 1. Raman active modes correspond tosymmetrymodesEgorA1g.TheRamanpeaksat~1086cm-1areattributedtotheA1gstretchingC-Omode.Ramanpeaksintherangeof700-900cm-1correspondtobendingmodesofthecarbonateionswiththelow700speakscorrespondingtoin-planedeformationoftheplanarCO3units.Frequenciesabove1000cm-1correspondtothesymmetricalsinglemodeA1gandasymmetricalEgstretchingImodesofthecarbonateion.TheRamanpeaksbetween150and180cm-1correspondtoout-of-phasetranslationsofthecarbonate ionsparallel tothe(111)planeoftherhombohedralcellwhilethepeaksinthe280sandlow300s correspond to in-phase librations of the carbonateions around the twofold symmetry axes [38]. A#x symmetrymodescorrespondtonondegeneratesinglemodeswithonlyone set of atomic displacements. A g subscript refers tosymmetric vibrations while a u subscript refers toantisymmetric vibrations. Eg symmetry modes involve twosetsofatomdisplacements.

Fig3.a)Biotitegneiss,tonalite,nephelinesyenite,alkaligranite,remoteRamanspectraata5mdistanceover30s integrationtime.Ramanspectraofcalcite,quartz,nepheline,andorthoclasearedisplayedforreference.Spectrawerescaledandverticallyshiftedforviewingclarity.Biotitegneissscaledx5forbrightspot,x10forintermediateanddarkspots. Nepheline syenite dark spot scaled x4. Tonalite and alkalicgraniteunscaled.Calcitescaledx0.2.Quartzscaledx2.Nephelinescaled

x8.Orthoclasescaledx2.b)Analyzedsamplesfromlefttoright:Row1:biotitegneisslightspot,tonalite,nephelinesyenite,alkalicgranite;Row2:biotitegneissintermediateanddarkspot,calcite,quartz,nepheline,orthoclase.Lasertargetwithinredcircle.

Figure 3 shows remote Raman spectra of light, dark, androughlyequallightanddark(banded)surfaceareasinbiotitegneiss(K(Mg,Fe2+3)(Al,Fe3+)Si3O10(OH,F)2)andalkaligranite,and light and dark colored areas in tonalite, and nephelinesyenite((Na,K)AlSiO4)ata5mstandoffdistanceovera30sintegration time. These measurement targets were chosenbasedonthesurfacegrainsvisibletothenakedeye,butdonotrepresentthefullinternalcrystalvolumeandresultinginternallight penetration contribution to the Raman signal. Lighterareas have stronger Raman spectra for each sample group,darker areas have weaker spectra, and mixed areas haveintermediateintensityspectraforeachsamplegroup.Forbiotitegneiss,thelighterphasesaremadeupofquartz

and plagioclase mineral components as seen by the Ramanpeaksat290,414,465,479,508,and571cm-1.Thesecorrelatewith α-quartz and the potassium feldspar orthoclase(KAlSi3O8).Biotitepeaksarevisibleat287,410,andthebroadbandsat750-807cm-1[23].Thelighterphaseshaveagreaterratioofquartztoplagioclasefeldsparwhilethemixedbandedbiotite gneiss has a slightly greater ratio of plagioclase toquartz. Both spectra exhibit moderately high levels of fastorganic/bio-fluorescence background. The dark phases haveno fast organic/bio-fluorescence background and favor agreaterquartzoverplagioclasefeldsparratio.Alpha quartz typically produces a distinct Raman peak at

465 cm-1 as a product of symmetric stretching of Si-O-Si.Smallerquartzmodesbetween200and400cm-1correspondtolatticedistortions.Plagioclasefeldsparsexhibitadoubletat~475-510cm-1.Theobserved508cm-1peakinbiotitegneissinitially indicates the presence of low albite but furtherexaminationof thespectral shapeandsmallerRamanbands768 and 807 cm-1, suggest an orthoclase K-feldsparinterpretation(alsoconfirmedbypowderXRDanalysesofthesame rock). When the 475 cm-1 peak is present alongsidequartz,the475cm-1peakcompeteswiththequartz465cm-1peak.This is aproductof symmetricT-O-T stretchingof thefour-memberedT-O4 tetrahedral ring (T=SiorAl) [6], [23],[28], [39]. Variations in T-O-T bond angles determine thepositionofthisRamanpeakduetothemotionoftheoxygenatoms in the four-membered ring breathing modeperpendiculartotheT-Tline.Theobserved571cm-1peakisdeterminedbythedegreeofAl-Siordering[6],[28]andcloselycorrelates with the orthoclase feldspar 568 cm-1 peak. AsdemonstratedbytheRamanspectra,thebiotitegneisssamplehasacombinationofquartz,albite,orthoclase,andbiotite.The quartz phases found in our biotite gneiss sample are

mostlywhiteascompared to thedarkerplagioclase feldsparphaseswhichappearlighttodarkgrey.Biotitephasesappearblack.Therearetwointerpretationsfortheunexpectedquartzpreference in the darker phases’ quartz/plagioclase ratio: a)quartzhasamoreorderedandlargerinternalcrystalvolumethan the plagioclase (or other) crystal volumes resulting ingreater internal light penetration contributing to the Raman

signal, or b) a darker mafic mineral with weaker Ramansignatures,suchasbiotite is intergrownwiththeplagioclase,making the plagioclase more opaque thus skewing thequartz/plagioclase ratio to prefer quartz. Overall spectralshapeandbroadRamanbandsmay indicate thepresenceofdarkermaficminerals.Light and dark spots were measured in tonalite (quartz

dioritewithaquartzcontent5-20%oftherock)asshowninFigure 3. Moderate levels of fast organic/bio-fluorescencebackgroundarevisible.QuartzRamanpeaksareidentifiedat198, 466, and 803 cm-1. Plagioclase feldspar peaks areidentified at 287, 484, 509, and 564 cm-1 with a 479 cm-1feldsparpeakanda466cm-1quartzpeakasashoulderofthe484 cm-1 plagioclase peak [6], [23], [28], [39]. Intermediateplagioclase feldspars ranging frombytownite—oligoclasearerepresented by these Raman peaks, and possible K-feldsparmicroclinemaybeattributedtothe803and1122cm-1peaks.Thesamplehasagreaterplagioclasetoquartzcomponentasdemonstratedbytheintensityoftheplagioclase479cm-1andquartz466cm-1Ramanpeaks[6],[28].Muchbroaderregionsat700-850cm-1andat1050cm-1areattributedtobiotite[23].Powder XRD measurements support the presence ofintermediateplagioclaseandesine,K-feldsparmicrocline,andquartz.Ramanmeasurementsadditionallyidentifybiotite.

Fig4.a)RemoteRamanspectraofplagioclaseandalkalifeldsparsolidseries at a 5mdistance over a 30 s integration time. Spectrawerescaled and vertically shifted for viewing clarity. Anorthite scaled x5.

Bytownite scaled x1.5. Labradorite scaled x4. Andesine scaled x10.Oligoclase scaled x1.1. Albite scaled x0.2. Orthoclase scaled x0.1.Microclinescaledx0.35.b)Analyzedsamplesfromlefttoright:Row1:anorthite, bytownite, labradorite, andesine;Row2: oligoclase, albite,orthoclase,microcline.Lasertargetwithinredcircle.

Lightanddarksportsweremeasuredinnephelinesyenite((Na,K)AlSiO4) as shown in Figure 3. The sample exhibitsmoderately high levels of fast organic/bio-fluorescencebackground, particularly prominent in the lighter targets.Nepheline, plagioclase feldspar, quartz, and calcite aredistinguishable in the nepheline syenite Raman spectrum.Nepheline in nepheline syenite is confirmed by the Ramanpeaksat397and992 cm-1 as verified throughanalysesof apurenephelinesample.Onlynephelineisreadilyidentifiableinthedarktargetanalyzed.OthermineralcomponentsidentifiedinthelightertargetwereacalciteRamanpeakattributedtotheA1gstretchingC-Omodeat1087cm-1[38],anα-quartzpeakat467cm-1productofsymmetricstretchingofSi-O-Si,aweakerα-quartzpeakat205cm-1correspondingtolatticedistortions[23], [39], and a low albite 508 cm-1 peak product ofcorresponding T-O-T bond angles [6], [28]. Powder XRDmeasurementsconfirmthepresenceofnepheline,albite,andquartz.ThelackofanyK-feldspardetectionindicatesthatthenephelinemineral component did not reactwith the quartzcomponentduringmeltcrystallization[40].Figure3alsoshowsalkalicgraniteremoteRamanspectraof

dark, light,andbothlightanddarkspotswithminimaltonofast organic/bio-fluorescence. Quartz is seen as a majorcomponentwithRamanpeaksat211,357,466,and1156cm-1.Feldsparsignaturesarevisibleas509,268,285,and807cm-1andtheshiftedquartzbandfrom465to466cm-1supportsthepresenceoffeldspars.The509cm-1Ramanpeakindicatesthepresence of an intermediate plagioclase feldspar such aslabradorite,andesine,oroligoclase,butthe268and285cm-1doublet,thebroad1156cm-1peakapproaching1139cm-1,andtheoverallspectralshapeintheregionindicatesthepresenceofmicrocline.Possiblebiotitepeaksarevisibleat403cm-1andthe broad bands at 750—800 cm-1 [23]. Lighter phases aredominated by quartz and darker more opaque phases aredominatedby feldsparsorothermaficminerals in thisalkaligranitesample.PowderXRDmeasurementscorrelatewiththeRamandataindicatingthepresenceofquartzandmicrocline,althoughbiotitewasnotdetected.Figure4showsremoteRamanspectraofthefullplagioclase

feldsparsseries(An0-100)andpartialpotassiumfeldsparseries.Raman identification of albite (NaAlSi3O8), oligoclase(Na,Ca)(Al,Si)AlSi2O8, andesine (NaAlSi3O8-CaAl2Si2O8),labradorite (Ca,Na)Al(Al,Si)Si2O8), bytownite(NaSi,CaAl)AlSi2O8),andanorthite(CaAl2Si2O8)canbebroadlyclassified into three groups: endmember anorthite,endmember albite, and intermediate plagioclases. Anorthitehasamajorpeakat504cm-1.Albitehasamajorpeakat508cm-1and483cm-1.Theintermediateplagioclasefeldsparshavemajorpeaksat510cm-1and478—481cm-1,withbytownitehavingamajorpeakat506cm-1and483cm-1.ThesepeaksareaproductofsymmetricT-O-Tstretchingofthefour-memberedT-O4tetrahedralringwhereT=SiorAl.TheRamanpeakat

293 cm-1 (cage-shear modes) decreases with increasinganorthite content.Thespectral shapeof thesingleordouble760—830 cm-1 peaks (deformationmodes of tetrahedral) isuseful as adiagnostic for identifying feldspars [6], [28], [41].The peak at 1099 cm-1 is characteristic of vas (Si-O-Al)antisymmetricstretchingmodesofframeworksilicates.ThesebandsshifttowardslowerfrequencywithAltoSisubstitutions[6],[41],[42].Ahighlevelofbackgroundfluorescenceisvisibleinthelighterphases,moreconsistentwithsodium-richalbiteor oligoclase while little to no background fluorescence isvisibleinthedarkerphases,moreconsistentwithpotassium-richorthoclase.Internalreflectionscancontributetoastrongerfluorescence signal in lighter crystals [43] and absorption ofphotonscansignificantlyreducebothRamanandfluorescencesignals in the darker materials. For the partial potassiumfeldsparseries:monoclinicorthoclase (KAlSi3O8)and triclinicmicrocline (KAlSi3O8), the major Raman peaks at 475—513cm-1shiftupwithincreasingcrystalgeometry.Figure 5 shows remote Raman spectra of grey and red

sandstone, quartz conglomerate, and quartzite. Organics insedimentary rocks contribute to a fast fluorescence baselineobserved in the four samples as the upward slopingfluorescence background towards the higher wavenumbers.This is consistent with biological detritus component ofsandstonesandothersedimentaryrocksonEarth[44].IftheMartian surface materials are sterile, fast organic/bio-fluorescence should not be present [45]–[47]. Quartzite isdisplayed for comparison. It has a relatively flat baselinecontainingminimal tono fastorganic/bio-fluorescencemoretypicalofmetamorphicrocks.GreysandstoneexhibitsstrongquartzandplagioclaseRamanpeaksat212and468cm-1andred sandstone exhibits both quartz and plagioclase peaks at468cm-1and178and203cm-1[6],[28].Quartzconglomerateisvisiblewithclearquartzpeaksat466and204cm-1.Subtlerbroad bands at 513, 177, and 263 cm-1 hint to K-feldsparcomponents. Quartzite exhibits quartz Raman peaks at 465and205cm-1.

Fig 5. a) Grey sandstone, quartz conglomerate, red sandstone, andquartziteremoteRamanspectraata5mdistanceover30sintegrationtime. Spectrawerevertically shifted for viewing clarity. b)Analyzed

samples from left to right: grey sandstone, red sandstone, quartzconglomerate,quartzite.Lasertargetwithinredcircle

Fig.6.DuniteandpyroxeniteremoteRamanspectraata5mdistanceovera120sintegrationtime.Themainolivinepeaksarevisibleat824,and 854 cm-1. Pyroxene was scaled x10 for viewing clarity. Photoinsert:Analyzedsamples fromleft toright:pyroxenite,dunite.Lasertargetwithinredcircle.

Olivine-rich dunite ((Mg,Fe)2SiO4) and pyroxenite((Mg,Fe,Ca)xSi2O6)remoteRamanspectraareshowninFigure6.Olivinepeaksarevisibleat824,854,917,and962cm-1.TheRamanspectraofolivineshasacharacteristicdoubletbetween808 and 825 cm-1 as a result of Si-Onb symmetric stretchingbandAg(Si-O)s-str, and 837 and858 cm-1 as a result of Si-Onbasymmetric stretching band Ag(Si-O)a-str. Dunite exhibits adoublet at 824 cm-1 and 854 cm-1 characteristic of forsterite[48], [49]. Chromite (FeCr2O4) is visible at 673 cm-1 [50].AdditionaldunitepeaksassignedaretheAgpeaksat194,305,543,and962cm-1productofT(SiO4),theB1g+B2gpeakat435cm-1,andthev3B3gpeakat917cm-1[51].Pyroxenitedisplaysaharzburgitev16peakat1009cm-1andav11peakat679cm-1.ThispeakisproducedbythesymmetricstretchingvibrationofSi-O bonds (where O is a non-bridging oxygen) in SiO4tetrahedra[52].Thesepeakpositionssuggestironenrichmentconcordant with the ultramafic nature of pyroxenites [53].Additionalminorpeaksat344and396cm-1correspondtothev3andv4bandsrespectively.Thepeaksat196and513cm-1may be attributed to orthoclase ormicrocline but there areinsufficientplagioclasepeakstoestablishaproperidentity.

Fig. 7.Representative10minuteXRDspectrumof amultiminerallicnaturalrockconfirmsthepresenceofquartzandplagioclaseinbiotitegneiss sample. Major peaks have been labeled with correspondinglatticemodes.

SimilartoremoteRaman,powderXRDmeasurementswerecapable of detecting major mineral components inmultiminerallic rocks. These measurements were useful inproviding ground-truth data to support our remote Ramandata.Figure7showsapowderXRDspectrumofbiotitegneiss,confirmingthepresenceofmajorrockcomponentsquartzandorthoclase.ThisXRDspectrumisrepresentativeofthequalityofXRDdataacquiredforallofthesamplesinthisstudy.Raman spectral analysis of the plagioclase and potassium

feldspars,biotitegneiss,tonalite,nephelinesyenite,andalkalicgranitedemonstratethatremoteRamancaneasilydetectend-memberfeldspars(albite,anorthite,microcline/orthoclase)innaturalrocks.Intermediatefeldsparsandquartz,whenfoundinnaturalrockswithcomplexmineralogiesexhibitbandshiftsandbroadeninginthe504-510cm-1and600-1200cm-1region.Agoodapproximationofintermediateplagioclasefeldsparsispossible by using spectral shape and assigning other minorRamanpeaks inadditiontothe504-510cm-1peaks.RemoteRaman can also easily detect calcite, olivine, and pyroxenefacilitating compositional characterization in multiminerallicrocks. Lastly, despite strong organic/bio-fluorescencebackground, quartz and plagioclase were also identified insedimentary rocks. Powder XRD analysis confirmed themineralcomponentidentificationdeterminedfromtheremoteRamanspectra.ThelatticeparametersaretabulatedinTable2.

B.PlanetaryscienceapplicationsRemoteRamananalysisonMarsorotherrockybodieswill

involve spectroscopic analysis in challenging environmentswith less than ideal sample conditions. Complex rockmineralogies, dusty conditions, and the presence of opaquemafic minerals, will introduce challenges to acquiring highqualityremoteRamanspectra.Afirst-orderlookattheremoteRaman spectra of natural rocks with multiple mineralcomponentsprovidesanunderstandingofsystemcapabilities,observedRamanpeakpositions,andthebroaderbandshapesresultingfrommultipleoverlappingRamanbands.

4.CONCLUSIONWehavedemonstratedthecapabilitiesofaCompactRemote

Raman+LIBS+Fluorescence System (CRRLFS) at a standoffdistanceof5mtoanalyzenaturalrocksindaylightconditionsand with high fluorescence backgrounds, without samplepreparation.RemoteRamanspectraofnaturalmultiminerallicrocksresultsinRamansignalmixingfromvariousphasesandit is demonstrated that several mineral phases such ascarbonates, quartz, feldspar, olivine, and pyroxene can beidentified in rocks.Theprovideddataand interpretationarenecessary for accurate analysis of future Ramanmeasurements on planetary surfaces such as on Mars, asexpectedfromtheNASAMars2020rovermission.ExpansionoftheavailablepublishedstandoffRamanspectraofrockswith

complex mineralogies will aid in providing a more realisticassessmentoftheexpectedspectraldatafromtheseandothersurfacerovermissions.

Funding. National Aeronautics and Space Administration(NASA)(NNX13AM98A);TheSuperCamprogramunderMars2020mission.

Acknowledgment.Theauthorswouldliketothank:Mr.MarkWood forwriting thecode to run the ICCDdetectorand thepulse generator and Dr. David Bates for developing customsoftwareusedtoreduceandbatchprocessRamanimagesintospectra..

REFERENCES[1] C.K.Johnson,G.A.Dalickas,S.A.Payne,andR.M.

Hochstrasser,“PicosecondtimescaleRamanprocessesandspectroscopy,”PureAppl.Chem.,vol.57,no.2,pp.195–200,1985.

[2] I.R.LewisandH.G.M.Edwards,HandbookofRamanSpectroscopy:FromtheResearchLaboratorytotheProcessLine,1sted.NewYork:MarcelDekker,2001.

[3] S.Mauriceetal.,“TheChemCaminstrumentsuiteontheMarsScienceLaboratory(MSL)rover:Scienceobjectivesandmastunitdescription,”SpaceSci.Rev.,vol.170,no.1–4,pp.95–166,2012.

[4] R.C.Wiens,S.K.Sharma,J.Thompson,A.Misra,andP.G.Lucey,“Jointanalysesbylaser-inducedbreakdownspectroscopy(LIBS)andRamanspectroscopyatstand-offdistances,”Spectrochim.Acta-PartAMol.Biomol.Spectrosc.,vol.61,no.10,pp.2324–2334,2005.

[5] J.Mustardetal.,“ReportoftheMars2020ScienceDefinitionTeam,”MarsExplor.Progr.Anal.Gr.,no.Cl,pp.155–205,2013.

[6] S.K.Sharma,S.M.Angel,M.Ghosh,H.W.Hubble,andP.G.Lucey,“RemotepulsedlaserRamanspectroscopysystemformineralanalysisonplanetarysurfacesto66meters,”Appl.Spectrosc.,vol.56,no.6,pp.699–705,2002.

[7] S.K.Sharma,“NewtrendsintelescopicremoteRamanspectroscopicinstrumentation,”Spectrochim.Acta-PartAMol.Biomol.Spectrosc.,vol.68,no.4,pp.1008–1022,2007.

[8] J.Jehlička,P.Vítek,H.G.M.Edwards,M.Heagraves,andT.Čapoun,“ApplicationofportableRamaninstrumentsforfastandnon-destructivedetectionofmineralsonoutcrops,”Spectrochim.Acta-PartAMol.Biomol.Spectrosc.,vol.73,no.3,pp.410–419,2009.

[9] F.Rull,A.Vegas,A.Sansano,andP.Sobron,“AnalysisofArcticicesbyRemoteRamanSpectroscopy,”Spectrochim.Acta-PartAMol.Biomol.Spectrosc.,vol.80,no.1,pp.148–155,2011.

[10] S.M.Angel,N.R.Gomer,S.K.Sharma,C.McKay,andN.Ames,“RemoteRamanspectroscopyforplanetaryexploration:Areview,”Appl.Spectrosc.,vol.66,no.2,pp.137–150,2012.

[11] P.J.Gasda,T.E.Acosta-Maeda,P.G.Lucey,A.K.Misra,S.K.Sharma,andG.J.Taylor,“Nextgenerationlaser-basedstandoffspectroscopytechniquesforMarsexploration,”Appl.Spectrosc.,vol.69,no.2,pp.173–192,2015.

[12] R.C.Wiensetal.,“THESUPERCAMREMOTERAMANSPECTROMETERFORMARS2020,”LPSCXLVIII,no.Abstract#2600,2017.

[13] N.Tarceaetal.,“RamanSpectroscopy—APowerfulToolforinsituPlanetaryScience,”SpaceSci.Rev.,vol.135,no.1–4,pp.281–292,Mar.2008.

[14] V.Kleinetal.,“RemoteRamanspectroscopyasaprospectivetoolforplanetarysurfaces,”J.RamanSpectrosc.,vol.35,no.6,pp.433–440,Jun.2004.

[15] P.Colomban,“Theon-site/remoteRamananalysiswithmobileinstruments:areviewofdrawbacksandsuccessinculturalheritagestudiesandotherassociatedfields,”J.RamanSpectrosc.,vol.43,no.11,pp.1529–1535,Nov.2012.

[16] J.Blacksberg,G.R.Rossman,andA.Gleckler,“Time-resolvedRamanspectroscopyforinsituplanetarymineralogy.,”Appl.Opt.,vol.49,no.26,pp.4951–4962,2010.

[17] A.K.Misra,S.K.Sharma,P.G.Lucey,R.C.F.Lentz,andC.H.Chio,“Daytimerapiddetectionofmineralsandorganicsfrom50and100mdistancesusingaRemoteRamansystem,”LidarRemoteSens.Environ.Monit.VIII,vol.6681,no.2007,p.66810C–66810C–14,2007.

[18] T.E.Acosta-Maedaetal.,“RemoteRamanmeasurementsofminerals,organics,andinorganicsat430mrange,”Appl.Opt.,vol.55,no.36,p.10283,2016.

[19] A.K.Misra,S.K.Sharma,andP.G.Lucey,“RemoteRamanspectroscopicdetectionofmineralsandorganicsunderilluminatedconditionsfromadistanceof10musingasingle532nmlaserpulse,”Appl.Spectrosc.,vol.60,no.2,pp.223–228,2006.

[20] S.K.Sharma,A.K.Misra,S.M.Clegg,J.E.Barefield,R.C.Wiens,andT.Acosta,“Time-resolvedremoteRamanstudyofmineralsundersupercriticalCO2andhightemperaturesrelevanttoVenusexploration.,”Philos.Trans.A.Math.Phys.Eng.Sci.,vol.368,no.1922,pp.3167–3191,2010.

[21] S.K.Sharmaetal.,“Remote-RamanspectroscopicstudyofmineralsundersupercriticalCO2relevanttoVenusexploration,”Spectrochim.Acta-PartAMol.Biomol.Spectrosc.,vol.80,no.1,pp.75–81,2011.

[22] J.D.Stopar,P.G.Lucey,S.K.Sharma,A.K.Misra,andH.W.Hubble,“AremoteRamansystemforplanetaryexploration:EvaluatingremoteRamanefficiency,”Instruments,Methods,Mission.Astrobiol.Vii,vol.5163,p.99–110\r236,2004.

[23] J.D.Stopar,P.G.Lucey,S.K.Sharma,A.K.Misra,G.J.Taylor,andH.W.Hubble,“Ramanefficienciesofnaturalrocksandminerals:PerformanceofaremoteRamansystemforplanetaryexplorationatadistanceof10meters,”Spectrochim.Acta-PartAMol.Biomol.Spectrosc.,vol.61,no.10,pp.2315–2323,2005.

[24] S.K.Sharma,A.K.Misra,P.G.Lucey,andR.C.F.Lentz,“AcombinedremoteRamanandLIBSinstrumentforcharacterizingmineralswith532nmlaserexcitation,”Spectrochim.Acta-PartAMol.Biomol.Spectrosc.,vol.73,no.3,pp.468–476,2009.

[25] A.K.Misra,S.K.Sharma,T.E.Acosta,J.N.Porter,andD.E.Bates,“Single-pulsestandoffRamandetectionofchemicalsfrom120mdistanceduringdaytime,”Appl.Spectrosc.,vol.66,no.11,pp.1279–1285,2012.

[26] A.Wangetal.,“DevelopmentoftheMarsmicrobeamRamanspectrometer(MMRS),”J.Geophys.Res.,vol.108,no.E1,p.5005,2003.

[27] G.BazalgetteCourrèges-Lacoste,B.Ahlers,andF.R.Pérez,“CombinedRamanspectrometer/laser-inducedbreakdownspectrometerforthenextESAmissiontoMars,”Spectrochim.Acta-PartAMol.Biomol.Spectrosc.,vol.68,no.4,pp.1023–1028,2007.

[28] J.J.Freeman,A.Wang,K.E.Kuebler,B.L.Jolliff,andL.A.Haskin,“Characterizationofnaturalfeldsparsbyramanspectroscopyforfutureplanetaryexploration,”Can.Mineral.,vol.46,no.6,pp.1477–1500,2008.

[29] L.a.Haskin,A.Wang,K.M.Rockow,B.L.Jolliff,R.L.Korotev,andK.M.Viskupic,“Ramanspectroscopyformineralidentificationandquantificafionforinsireplanetarysurfaceanalysis :Apointcountmethod,”J.Geophys.Res.,vol.102,no.97,pp.19293–19306,1997.

[30] I.Martínez-Arkarazo,D.C.Smith,O.Zuloaga,M.A.Olazabal,andJ.M.Madariaga,“EvaluationofthreedifferentmobileRamanmicroscopesemployedtostudydeterioratedcivilbuildingstones,”J.RamanSpectrosc.,vol.39,no.8,pp.1018–1029,Aug.2008.

[31] J.Moros,J.A.Lorenzo,P.Lucena,L.M.Tobaria,andJ.Laserna,“SimultaneousRamanSpectroscopy-Laser-InducedBreakdownSpectroscopyforInstantStandoffAnalysisofExplosivesUsingaMobileIntegratedSensorPlatform,”vol.82,no.4,pp.1389–1400,2010.

[32] B.Zachhuber,G.Ramer,A.Hobro,E.t.H.Chrysostom,andB.Lendl,“Stand-offRamanspectroscopy:apowerfultechniqueforqualitativeandquantitativeanalysisofinorganicandorganiccompoundsincludingexplosives,”Anal.Bioanal.Chem.,vol.400,no.8,pp.2439–2447,Jun.2011.

[33] Q.Lin,G.Niu,Q.Wang,Q.Yu,andY.Duan,“CombinedLaser-InducedBreakdownwithRamanSpectroscopy :HistoricalTechnologyDevelopmentandRecentApplicationsCombinedLaser-InducedBreakdownwithRamanSpectroscopy :HistoricalTechnologyDevelopment,”Appl.Spectrosc.Rev.,vol.48:6,no.June2017,pp.487–508,2013.

[34] C.S.Garciaetal.,“RemoteRamansensorsystemfortestingofrocksandminerals,”Proc.SPIE,vol.6538,p.65381I–65381I–9,2007.

[35] S.K.Sharma,P.G.Lucey,M.Ghosh,H.W.Hubble,andK.A.Horton,“Stand-offRamanspectroscopicdetectionofmineralsonplanetarysurfaces,”Spectrochim.Acta-PartAMol.Biomol.Spectrosc.,vol.59,no.10,pp.2391–2407,2003.

[36] A.K.Misra,S.K.Sharma,C.H.Chio,P.G.Lucey,andB.Lienert,“PulsedremoteRamansystemfordaytimemeasurementsofmineralspectra,”Spectrochim.Acta-PartAMol.Biomol.Spectrosc.,vol.61,no.10,pp.2281–2287,2005.

[37] A.K.Misra,S.K.Sharma,T.E.Acosta,andD.E.Bates,“CompactremoteRamanandLIBSsystemfordetectionofminerals,water,icesandatmosphericgasesforplanetaryexploration,”Proc.SPIE-Int.Soc.Opt.Eng.,vol.8032,no.2,p.80320Q–80320Q–12,2011.

[38] M.Prencipe,F.Pascale,C.M.Zicovich-Wilson,V.R.Saunders,R.Orlando,andR.Dovesi,“Thevibrationalspectrumofcalcite(CaCO3):Anabinitioquantum-

mechanicalcalculation,”Phys.Chem.Miner.,vol.31,no.8,pp.559–564,2004.

[39] C.H.Chio,S.K.Sharma,P.G.Lucey,andD.W.Muenow,“Effectsofparticlesizeandlaser-inducedheatingontheRamanspectraofalphaquartzgrains,”Appl.Spectrosc.,vol.57,no.7,pp.774–783,2003.

[40] T.A.Rickard,“TheCrippleCreekVolcano,”Trans.theAmericanInst.Min.Eng.,vol.XXX,pp.367–403,1901.

[41] S.K.Sharma,J.A.Philpotts,andD.W.Matson,“Ringdistributionsinalkali-andalkaline-earthaluminosilicateframeworkglasses-aramanspectroscopicstudy,”J.Non.Cryst.Solids,vol.71,no.1–3,pp.403–410,May1985.

[42] D.W.Matson,S.K.Sharma,andJ.A.Philpotts,“Ramanspectraofsometectosilicatesandofglassesalongtheorthoclase-anorthiteandnepheline-anorthitejoins,”Am.Mineral.,vol.71,pp.694–704,1986.

[43] B.J.Bozlee,A.K.Misra,S.K.Sharma,andM.Ingram,“RemoteRamanandfluorescencestudiesofmineralsamples,”Spectrochim.Acta-PartAMol.Biomol.Spectrosc.,vol.61,no.10,pp.2342–2348,2005.

[44] A.K.Misraetal.,“‘StandoffBiofinder’forfast,noncontact,nondestructive,large-areadetectionofbiologicalmaterialsforplanetaryexploration,”Astrobiology,vol.16,no.9,pp.715–729,2016.

[45] A.Ellery,D.Wynn-Williams,J.Parnell,H.G.M.Edwards,andD.Dickensheets,“TheroleofRamanspectroscopyasanastrobiologicaltoolintheexplorationofMars,”J.RamanSpectrosc.,vol.35,no.6,pp.441–457,Jun.2004.

[46] J.PoppandM.Schmitt,“Ramanspectroscopybreakingterrestrialbarriers!,”J.RamanSpectrosc.,vol.35,no.6,pp.429–432,Jun.2004.

[47] I.B.Hutchinsonetal.,“RamanspectroscopyonMars:identificationofgeologicalandbio-geologicalsignaturesinMartiananaloguesusingminiaturizedRamanspectrometers,”Philos.Trans.R.Soc.,A,vol.372,p.20140204,2014.

[48] T.MouriandM.Enami,“Ramanspectroscopicstudyofolivine-groupminerals,”J.Mineral.Petrol.Sci.,vol.103,pp.100–104,2008.

[49] S.Y.Wang,S.K.Sharma,andT.F.Cooney,“Micro-Ramanandinfraredspectralstudyofforsteriteunderhighpressure,”Am.Mi,vol.78,pp.469–476,1993.

[50] B.Lafuente,R.T.Downs,H.Yang,andN.Stone,Thepowerofdatabases:theRRUFFproject.Berlin,Germany:W.DeGruyter,2015.

[51] B.A.KolesovandC.A.Geiger,“ARamanspectroscopicstudyofFe-Mgolivines,”Phys.Chem.Miner.,vol.31,no.3,pp.142–154,2004.

[52] A.Wang,B.L.Jolliff,andL.A.Haskin,“Ramanspectroscopiccharacterizationofahighlyweatheredbasalt:Igneousmineralogy,alterationproducts,andamicroorganism,”J.Geophys.Res.,vol.104,no.E11,pp.27067–27077,1999.

[53] E.Huang,C.H.Chen,T.Huang,E.H.Lin,andJ.A.Xu,“RamanspectroscopiccharacteristicsofMg-Fe-Capyroxenes,”Am.Mineral.,vol.85,no.3–4,pp.473–479,2000.

Table2:PowderXRDdiffractioncrystallatticeparameters.