0008_Bell_III_J.F._Thermally_Altered_JGR_1993

JOURNAL OF GEOPHYSICAL RESEARCH, VOL 98, NO E2, PAGES 3373-3385, FEBRUARY 25, 1993 ThermallyAlteredPalagoniticTephra: A SpectralandProcessAnalogto the Soil and Dust of Mars JAMESF BELL III • PlanetaryGeosciences Division,Universityof ttawaii at Manoa,Honolulu RICHARD V MORRIS PlanetaryScienceBranch,NASAJohnsonSpaceCenter,Houston,Texas JOHN B ADAMS Departmentof GeologicalSciences, Universityof Washington, Seattle We studiedsixpalagonitic soilsamples (PH-1throughPH-6) whichwerecollectedat 30-cmintervalsfroma lavaslabonMaunaKea,Hawaii Thesamples represent an alteration sequence caused by heatingduring emplacement of moltenlavaovera preexisting tephracone.Theyarebothspectral andweathering/alteration process analogs to theMartiansurface Techniques employed includedvisibleandnear-IRspectroscopy, M6ssbauer spectroscopy, andmagnetic analysis.Thetwosamples collected frombeyond90 cmfromtheslab (PH-5andPH-6)didnotseethetransient heating eventandarecomposed of coarse-grained glassy basaltic (hawaiitic) particles, someof whichhavebeenmoderately palagonitized sinceformation of theconein thelate Pleistocene Thefoursamples closest totheslab(PH-1through PH-4)havebeenstrongly altered in response to heating duringitsemplacement; theirironoxidemineralogy is dominated bynanophase ferricoxide.The sample adjacent totheslab(PH-1),whichwouldhaveseenthehighest temperatures, hasa factorof lessH20 andcontains crystalline hematiteandmagnetite in additionto nanophase ferricoxide.Lesseramounts of magnetite,but not hematite,are presentin samples30 cm (PH-2) and 60 cm (PH-3) from the slab.The reflectivity spectra of samples PH-2through PH-4aresimilarto spectra of palagonites reported by other workers tobegoodMarsvisibletonear-IRspectral analogs Thereflectivity spectrum of PH-1is anevenbetter spectral analog to Marsin thatit exhibits absorption features indicative of bothnanophase andcrystalline ferric oxides (similar toseveral newMartiantelescopic datasets).Oursampling sitemayalsobea process analog for Mars,in thatheating episodes byvolcanism and/orimpactcrateting couldproduce crystalline ferricoxides from poorlycrystalline palagonitic matedhal Thus,localized thermal alteration events mayprovide a volumetrically important mechanism forthepalagonitization ofbasaltic glassandtheproduction of crystalline ferricoxides on Mars INTRODUCTION Thereis considerable debateon the mineralogyandmodeof formation of the red, wind-blown soils and dust on Mars Most of the data currentlyused to assesstheseparametershave been providedby ground-based telescopic spectral observations (see,for example,Singeret al [1979]andBellet al [1990a])andViking lander geochemicalmeasurements [e.g., Toulminet al., 1977] BecausetheMartianspectrum in thevisibleto near-infrared (VIS- NIR, 0.4-1.1Ixm)is redandrelativelyfeatureless andtheViking XRF dataprovideevidencefor aniron-richclaylikecomposition, reinterpretation of previousdataby Morris et al [1989] andMorris andLatter [ 1990]haveled to the identificationof crystallinehematite (ct-Fe203)on Mars but only as a minor accessory phasein a matrix of a muchmore abundantnanocrystalline component Four principal environments for the production of poorly crystallinepalagonitelikematerialson Mars by the weatheringor alteration of primary igneous rocks have been proposed: hydrothermal alteration induced by impact, volcanism, or geothermal gradients [e.g., Newsom, 1980; Allen et al., 1982; Clifford and Carr, 1991], subpermafrost magtnatic intrusion [Soderblom and Wenner, 1978; Allen et al., 1981], subaerial intrusion above the permafrostlayer [Berkley and Drake, 1981], dominated by poorlycrystalline or evenamorphous minerals similar and static gas-solidweathering [Gooding and Keil, 1978] The to certainterrestrialpalagonites[Toulminet al., 1977;Soderblom relative importanceof each of thesemechanismsin altering the and Wenner, 1978; Evans and Adams, 1979; Allen et al., 1981; composition and mineralogy of Martian surface materials is Singer,1982] Recenttelescopic observations at high spectral determinedby a numberof factors,includingthe contemporaneous resolutionby Bell et al [1990a,b]andSingeret al [1990] and stateof theMartian climateandhydrosphere andthe amountof time that the process operates Remote sensing data and in situ 1NowatSpace Sciences Division, NASAAmesResearch Center, measurements of Mars soils, rocks, and dust (the end productsof Moffett Field, Califomia theseprocesses) can providepossibleways to distinguishbetween theseenvironments.For example,detectionof metastableFe-rich Copyright 1993by theAmerican Geophysical Union phyllosilicatemineralson Mars may indicatethepredominance of a gas-solidweathering environment,at least in recent times [e.g., Papernumber92JE02367 0148-0227/93/92JE-02367505.00 Goodingand Keil, 1978] Detectionof hematiticor iron sulfate-rich there is a general consensusthat the near-surfacematerials are 3373 3374 BELL ET AL.: THERMALLY ALTERED TEPHRA AS A MARS SOIL ANALOG unitscombinedwith othergeologiccluesmay indicatethepresence the summitof Mauna Kea Wolfe et al [1992] have mappedthis of spatiallydistincthydrothermal alterationzones(i.e., thegossans 300-m-diameter,40-m-high cone as a part of the Laupahoehoe of Burns [1988]) Volcanic series (hawaiitic composition), with an age of late As partof an ongoingstudyof potentialMars surfaceanalog Pleistocene.Unlike the summit region of Mauna Kea, Puu materials,we collectedsix samplesof palagoniticsoil at 30-cm Huluhulu receives moderate amountsof annual rainfall (= 100 inter•alsfrom a lava slabon MaunaKea, Hawaii.Thesesamples cm/yr) andis denselyvegetatedin placeswith grass,trees,andlow represent an alteration sequencecaused by heating during brush.The westernsideof the conehas beenpartiallyquarried, emplacement of moltenlava over a preexistingtephracone Our exposingin crosssection a steeply dipping (55ø) 15- to 20-m goalsare (1) to examinechangesin themineralogyof our samples elevationlava slabthat apparentlybakedthe preexistingtephrato as a functionof distancefrom the heat source;(2) to examine to forma 1.5-m-widezoneof alteration(Figure1) The regionwithin whatdegreethermallyalteredpalagonitictephrais a Mars surface 0.5 m of the slab is bright reddish-orangein color and has mineralogicalandprocessanalogmaterial;and (3) to examinethe presumablyundergonethe highestdegree of alteration/oxidation validityof theMartianalteration/weathering schemes proposed [Evanset al., 1981] Less altered tephraoccursin the region above.inlight of thisandotherterrestrialanalogstudiesandother between0.5 and 1.5 m from the slab,wherethe colorgradesfrom currentlyavailableMars surfacecompositionaland mineralogic reddish-orangeto dark tan-brown.The tephra 1.5 m and farther information.We acquireda varietyof data,includingvisibleto from the slab is coarseand black and has apparentlynot been near-infrared(V!S-NIR) reflectancespectra,M6ssbauerspectra, affectedby theheatingevent.We obtainedsamplesof thetephraat perpendicular to the saturation magnetization, particle size, and water and Fe six 30-cm intervalsalongtwo separatetransects slab concentrations SAMPLES AND GEOLOGIC SETTING Our sampleswere obtained in 1987 and 1990 from the Puu Huluhulu cinder cone on the island of Hawaii This cone is located in the Humuula district at 2040 m elevation on the Saddle Road betweenMaunaKea andMauna Loa, roughly15 km due southof The originof thesteeplydippinglavaslabis notreadilyapparent becausequarryingdatingback to the 1930shasremovedmuchof thiscap rock in orderto excavatethe cindersbelow.Thus, the age relationshipbetweenthe time of coneformation(late Pleistocene according to Wolfe et al [1992]) and the time of the thermal alteration event is morphologically unclear The cone itself is Fig View of northsideof PuuHuluhuluquarry,lookingeast.Ourmostthermallyalteredsamples (PH-1) wereobtasned froma brightredzoneimmediately belowthe55ødippingblockya'alavaslabthathasflowedoverthecinders.Othersamples wereobtained at 30 cmintervalsalonga lineperpendicular to thelava-cinder contact Thegeologist is roughly1.7m tall for scale BELL ET AL.: THERMALLY ALTERED TEPHRA AS A MARS SOIL ANALOG 3375 intrudedby at leasttwo 30- to 50-cm-widedikesand is surrounded (a) bynumerous recent (1935)andprehistoric (1500-2000 yearold) MaunaLoa pahoehoe and a'a flows(E Wolfe, personal '•1000 communication, 1991' J Lockwood,unpublished data, 1991) Thus, it is possiblethat the steeplava slab could be a Mauna Loa lavaflowthatsloshed uptheside ofthecone orwashighly inflated '"• 600 whenit encountered thecone,andsubsequent sloshbackor deflationleft just the thin slabcoatingthecinders(andcausingthe thermal alteration) (E Wolfe, personalcommunication,1991) However,it seemsunlikelythateitherof theseprocesses couldraise •, 200 J (b) lavato20mabove itsbase level It is more likely that the slab originatedfrom one of the dikes internal tothecone thatbroke outtothesurface, creating a small flank eruption[e.g., Porter, 1972] thatproducedthe meter-thick õ slabandwhichaltered theunderlying cinders Preliminary analyses • •0 of samplesof the slab and the internaldikesshowsthem to be tholeiiticandnearlyidenticalto eachotherin composition (J.F.Bell o I I III et al., manuscriptin preparation,1993) Thesedata supportthe (c) previoussuggestion thatthedikesareMaunaLoa lavasthathave intrudedthispreviouslyexistingMaunaKea cindercone(E Wolfe, unpublished data, 1991) This surprising result islikely caused by the cinder cone's intermediate location between Mauna Kea and Mauna Loa and probably to the interactionof relatively recent MaunaLoa magmawith thecone'spreexistingplumbingsystem.In this scenario,the thermalalterationof the tephrasubjacentto the extrudedslab may have occurredany time up to many tens of thousandsof years after the end of the original Puu Huluhulu (MaunaKea) emption.The mostimportantimplicationfor our study hereis thatthe preexistingtephrawasexposedat the surfacefor a substantialamount of time before emplacementof the lava slab, whichheatedandalteredtheunderlyingtephra MEASUREMENTS AND RESULTS • •0 • •0 • (d) I i I (e) • Sampleswere wet seivedin freonintonine sizefractionsusing 20-, 45-, 90-, 150-, 250-, 500-, and 1000-p.m rhodium-plated -• nickel screensand a 2-mm stainlesssteelwire mesh.Mean particle size calculations[Folk and Ward, 1957] show that the samples PH-3 PH-4 PH-5 PH-{ 3H-1 PH-2 closestto the heat sourceare generally finer-grained than those 40 60 80 100 120 140 2o fartheraway (Figure2a) Presumably,thisresultsfrom thehigher Distance from lieat Source (era) degree of alteration closer to the slab Binocular microscopic examination of our > 90 p.m seive fraction samples reveals Figure2 (a) Mean particlesizevs distancefrom the lava-cindercontact usingthemethodof Folk andWard[ 1957]: qualitativecolor and morphologicchangesbetweenthe samples Meanparticlesizedetermined 3.1 Particle Size Distribution closestto andfarthestfrom the lava slab.The mostthermallyaltered Mz = 0.33 ß (qbl6+ qb50 + qb84),where qbis the standard"phi size" from sedimentary petrologyandthe subscripts refer to cumulativepercentage values.(b) Total iron contentof bulk (< mm size) samplesexpressedas weightpercentFe, determined by INAA (c) Bulk saturation magnetization Js versus distance from heat source, corrected for total water and iron tephraparticlesexhibita thickreddish-brown palagonitization rind that flakesoff and accountsfor the greaternumberof fine particles in these samples Many of the cinders have been altered so thoroughlyfrom thisprocessthattheyno longerhaveanyunaltered content.(d) Weightpercentbulk structuralwatercontent(H2O+) versus cores The tephra far from the lava slab showsmany black, distancefrom heat source.(e) Weight percentbulk adsorbedwatercontent (H20-) versusdistancefrom heat source relatively pristine cindersas well as minor amountsof orangeyellow glass.The 90-150 grnsizefractionof all samplesalsoshow numerouswhite and greenparticles,most likely plagioclaseand olivine of the unaltered tephra (hawaiitic) based on variation diagrams [BasalticVolcanismStudyProject, 1981;Frey et al., 1990] 3.2 Fe and Water Concentrations Both adsorbed(here designated H2O-) and structural (here designated H2O+) water abundances were determinedfor these Fe abundanceswere measured using instrumental neutron activationanalysis(INAA) techniques (Figure2b) and wereusedto samplesusing a DuPont 902 moistureevolutionanalyzer(Figures normalizethebulk Jsvalues(Figure2c, seesection3.5) to totaliron 2d and 2e) Adsorbed water content at room temperatureranges content.Abundancesof other components(i.e., Na20, CaO, La) from to wt % for all samples,and thereis a slightincreasein wereusedto identifythecomposition of thelava slab(tholeiitic)and H20- at the finer grain sizes(Figure 3) Structuralwater released 3376 BELLETAL.:THERMALLY ALTERED TEPHRA ASA MARSSOILANALOG v PH-6 H20- • PH-5 x PH-4 + PH-3 APH-2 O PH-1 (a) f < 20 i 20-45 i i 45-90 90-150 i 150-250 i 250-500 i 500-1000 Particle Size Fraction 14 v PH-6 H20+ • PH-5 12 x PH-4 + PH-3 • 10 aPH-2 oPH-1 • o • • (b) Particle Size Fraction Fig.3 Weightpercent bulk(a) adsorbed H20' and(b) structural H20+ asa function of grainsize.ThePH-6samples arefarthest fromtheheatsource;PH-1 samples areclosest.Noticethe strongdecrease in totalwatercontentin the sampleclosestto thelava slab in H2O+ uponheatingabove875øCrangesfrom to 15 wt % and also grains.Figures2d and2e showa slightgeneraldecrease showsa slightincreasewith decreasing grainsize(Figure3) The and H20' with distancefor the samplesfarthestfrom the slab,and samples closestto theheatsourcearedepleted in H20+ by a factor thistrendis consistentwith the generalincreasein grainsize asseen of 2-3 relativeto samplesfartheraway,whichis consistent with a decreasing thermalgradientfromthelava-tephra contact.The H20abundancealsodecreases for the samplesclosestto the slab.This may reflect a mineralogicchangeassociated with this thermal alterationeventbecausetheseotherwisef'mer-grained particlesmight be expectedto containmore adsorbedwaterthanadjacentcoarser in Figure2a 3.3 X Ray Diffraction (XRD) Measurements XRD measurements were performedon a ScintagModel 2000 instrumentusing sample preparation methods similar to those describedby Goldenet al [thisissue].Coarse(500-1000 grn)and BELL ET AL.: THERMALLY ALTEREDTEPHRAASA MARS SOIL ANALOG fine (< 20 Ixm) fractions of the samples closest to (PH-1) and farthestfrom (PH-6) the heat sourcewere analyzed.Initial analysis showsthat the X ray diffraction patremsof all four samplesare dominatedby plagioclase;minorto traceamountsof pyroxeneare alsopresent.Both fine fractionscontainminoramountsof smectite clay andhematite,althoughthereis substantially morehematitein 3377 S2 S1 D3 PH-1 than PH-6 The PH-1 fine fraction also shows evidence for a minor amountof an Fe-Ti spinel phase(s).Both coarsefractions containminoramountsof Fe-Ti spinelphases,andthe PH-6 coarse PH-2 fraction also shows minor amounts of olivine and a trace of smectite clay These results are consistent with the reflectance and M6ssbauerspectradiscussed below PH-3 3.4 M6ssbauerSpectroscopy M6ssbauerspectroscopyis an importantanalytic tool for the characterization andidentificationof theiron-bearingphasespresent in samples.Specific and unique identification of many ferric oxide/oxyhydroxide and ferrous-bearing minerals can be PH-.4 accomplished using laboratorymineralreferencespectrasuchas thoseof Morris et al [1985] andMurad [1988] Roomtemperature 57FeM6ssbauer spectraof oursamples were obtainedon Ranger Scientific spectrometers using the methods describedin detail by Morris et al [1985, 1989] These spectra were fit to theoretical line shapesusing an in-house computer program(JSCF1T) The isomershift (IS) is reportedrelativeto the midpoint of the spectrumof metallic iron foil at 293 K For magnetically split spectra, the quadrupole splitting (QS) and hyperfinefield strength(Bhf)were calculatedfrom 0.5.[(6-5)-(2-1)] and [6-5], respectively,wherethe numbersinsidethe bracketsrefer to sextet line positions numbered from low to high velocity M6ssbauerspectraand parametersfor bulk ( 1.10 mm/s) Thereforethis identification PH-5 PH-6 D1: np-Ox (oct-Fe3+) D2: Px, G D3: -10 FeMS Data (293 K) OI -8 < mm Size Fraction -6 -4 -2 10 Velocity (turn/s) Fig Room temperatureM6ssbauerspectraof bulk (< mm) size fractionof all PH seriessamples.Key to phases:np-Ox, nanophaseiron oxide;p-Fe(3+), paramagnetic Fe3+;Px, pyroxene;G, glass(Fe2+);O1, olivine; Hm, hematite;cd-Mt, cation-deficientmagnetite is tentative.Theratioof theoctahedral to tetrahedral peakareasfor magnetite is = 1, sothatit is strongly cationdeficient[e.g.,Daniels and Rosencwaig,1969] and/orhasTi substitutedat the octahedral TABLE AverageM6ssbauerParameters at 293 K for Iron-Bearing Phasesin PalagoniticSoilsPH-1 ThroughPH-6 Phase IS, mm/s QS, mm/s Bhf, T Ferroux Doublets Olivine 1.13+0.02 2.87+0.03 Pyroxene 0.98+0.02 2.36+0.03 Glass 0.99+0.02 2.05+0.02 Nanophase ironoxide 0.35+0.02 Ferric Doublet 0.69+0.04 Magnetic Sextets Hematite 0.38+0.02 Magnetite(octahedral) 0.66+0.02 Magnetite (tetrahedral) 0.29+0.02 -0.18+0.02 51.5+0.2 -0.08+0.02 -0.03+0.02 46.7+0.2 49.5+0.2 sims[e.g.,Lipkaet al., 1988] ThehematitemayalsocontainTi The doublethavingIS = 0.35 mm/sandQS = 0.69 mm/sresults from octahedrallycoordinatedferric iron The parametersare consistent with thosefor small-particle ironoxides/oxyhydroxides like ferrihydrite[e.g., Murad and Johnson,1987], nanophase (superparamagnetic) hematite[e.g.,Kundiget al., 1966;Morris et al., 1989],andothersuperparamagnetic ferricoxides[e.g.,Murad andJohnson,1987;Murad, 1989].Becausewe cannotidentifythe mineralogyof thesmalloxideparticles,we usethegenericphrase "nanophase ferric oxide (np-Ox)" to refer to iron oxide/oxyhydroxide particleshavingnanoscale( < 20-50 nm) particle dimensions Ferrihydrite,superparamagnetic (at room temperature) particlesof hematiteandgoethite,andnanometer-sized particles of inherently paramagnetic(at room temperature) lepidocrociteare all nanophaseoxidesthat couldcontributeto the ferricdoublet.We believethatcontributions to thedoubletby ferric 3378 BELL ET AL.: THERMALLY ALTERED TEPHRAAS A MARS SOIL ANALOG TABLE RelativePercent Spectral Areasof Phases fromM6ssbauer Datafor < mm,500-1000tam,and < 20 grnSize Fractionsof PalagoniticSoilsPH-1 ThroughPH-6 Sample PH-1 PH-2 PH-3 Size Fraction Olivine