morphology, chemistry, and spectral properties of hawaiian rock …scott/workshop_reading/... ·...

Morphology, chemistry, and spectral properties of Hawaiian rock

coatings and implications for Mars

Michelle E. Minitti,1 Catherine M. Weitz,2 Melissa D. Lane,2 and Janice L. Bishop3

Received 5 October 2006; revised 22 November 2006; accepted 12 December 2006; published 30 May 2007.

[1] We studied coatings on five glass-rich basalts from the Big Island of Hawaii. Thecoatings are characterized by complex morphologies and their thicknesses range from 3 to80 mm. Coating chemistries are predominantly hydrated silica with minor amounts ofFe-, Ti-, and S-bearing materials. Visible, near infrared, and thermal infrared spectra of thecoatings demonstrate that coatings as thin as 3 mm mask the spectral character of thesubstrate basalt. The Fe-, Ti-, and S-bearing components control the VNIR coating spectra.The hydrated silica component of the coatings dominates the thermal infrared coatingspectra. The coating chemistries are consistent with leaching and/or dissolution of basaltglass or tephra in aqueous, acidic, and oxidizing conditions and subsequent precipitationof insoluble phases. Similar alteration conditions are thought to have occurred onMars, making formation of such coatings on Martian lithologies feasible. Given thecapabilities of various Mars missions, if coatings like those of this study were present onMartian lithologies, their chemical and/or spectral signatures would be detectable.Chemical and spectral data from thin (�10 mm), Fe- and Ti-bearing coatings are consistentwith phases capable of explaining the high-SiO2 component of Type 2 surfaces previouslyidentified in the Thermal Emission Spectrometer data set and are potentially consistentwith spectral data from the Mars Exploration Rovers.

Citation: Minitti, M. E., C. M. Weitz, M. D. Lane, and J. L. Bishop (2007), Morphology, chemistry, and spectral properties of

Hawaiian rock coatings and implications for Mars, J. Geophys. Res., 112, E05015, doi:10.1029/2006JE002839.

1. Introduction

[2] Remote sensing studies are an important tool for theidentification of primary lithologies which provide insightinto the fundamental processes that shape planetary surfa-ces. Classification of primary lithologies through remotesensing depends on direct access to those lithologies,access that is potentially hindered by the presence ofsecondary lithologies. Coatings are a known class ofsecondary materials that influence remote sensing studiesof Mars. At visible and near infrared (VNIR) wavelengths,Martian dark regions exhibit evidence of thin coatingsexpressed as negative near infrared spectral slopes [Singerand McCord, 1979]. Coatings also potentially influencethermal infrared spectral data. Thermal Emission Spec-trometer (TES) data from wind streaks are consistent withthe signature of thin (tens of micrometers) dust coatings[Johnson et al., 2002]. SiO2 [Kraft et al., 2003] andphyllosilicate coatings [Wyatt and McSween, 2002] havebeen proposed as alternatives to the SiO2-rich glasscomponent identified by spectral deconvolution of theType 2 lithology established by TES data [Bandfield et al.,

2000]. The presence of either coating material would alterthe classification of the Type 2 lithology previously classi-fied as andesite to basaltic andesite. Michalski et al. [2005]also explored alternatives to the SiO2-rich component inthe Type 2 spectrum by linking crystal-chemical characte-ristics of various SiO2-rich materials and clays to theirthermal infrared signatures. Michalski et al. found thatclays could not yield the thermal infrared signature ofthe SiO2-rich component of the Type 2 spectrum and thatmore highly polymerized materials (tectosilicates) wererequired to account for the SiO2-rich component spectralsignature. Such materials included primary high-SiO2

glasses (obsidian), zeolites, allophane, and Fe- and/orAl-bearing opal. It is unclear if such materials are compa-tible with available VNIR data. Palagonitic coatings [e.g.,Morris et al., 2003] also have been suggested as analternative to the SiO2-rich glass component and havethe advantage that they possess proper thermal infraredand VNIR spectral character.[3] The presence and influence of coatings is supported

by lander-based studies of Martian surface lithologies. Dustcoatings influenced the analyses of rocks at the MarsPathfinder landing site [e.g., McSween et al., 1999] andboth the Mars Exploration Rover (MER) landing sites [e.g.,Bell et al., 2004; Christensen et al., 2004]. Additionally, aclear, homogeneous material that created relatively smoothrock surfaces was indicated by the photometric propertiesof the gray rock surfaces at the Mars Pathfinder site[Johnson et al., 1999]. Excess oxygen in Mars Pathfinder

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 112, E05015, doi:10.1029/2006JE002839, 2007ClickHere

for

FullArticle

1Center for Meteorite Studies, School of Earth and Space Exploration,Arizona State University, Tempe, Arizona, USA.

2Planetary Science Institute, Tucson, Arizona, USA.3SETI Institute/NASA-ARC, Mountain View, California, USA.

Copyright 2007 by the American Geophysical Union.0148-0227/07/2006JE002839$09.00

E05015 1 of 24

http://dx.doi.org/10.1029//2006JE002839

Alpha Proton X-Ray Spectrometer (MP-APXS) data [Foleyet al., 2003] and, possibly, elevated SiO2 concentrationsare also consistent with the presence of hydrated SiO2

coatings. Direct chemical analyses obtained by the MERSpirit at Gusev Crater reveal the presence of coatingsand/or rinds on basalt rocks [Gellert et al., 2004; McSweenet al., 2004; Haskin, 2005; Hurowitz et al., 2006]. Chem-ical and spectral analyses obtained by the OpportunityMER at Meridiani Planum also offer evidence of coatingsand/or rinds on several outcrops along the rover’s traverse[Learner et al., 2006; Squyres et al., 2006; Jolliff et al.,2006]. The variety of potential coatings and their impor-tance to understanding surface composition of Mars makesthe study of potential Mars coating analogs a valuableexercise.[4] Hawaiian rocks have long served as effective analogs

for Martian lithologies [e.g., Singer, 1980, 1982; Morris etal., 2000; Bishop et al., 2002]. Coatings are also a commonsecondary material on Hawaiian basalts [e.g., Hay andJones, 1972; Farr and Adams, 1984; Curtiss et al., 1985]and are capable of affecting remote sensing data sets [e.g.,Kahle et al., 1988; Crisp et al., 1990; Abrams et al., 1991].The goal of this study is to investigate a collection of rockcoatings taken from different locations at the Hawaii Vol-canoes National Park (Figure 1) as a means of studyingfeasible Martian coating analogs. Thorough study of theHawaiian coatings establishes their chemistry, morphology,distribution, and spectral properties at VNIR and thermalinfrared wavelengths. This study also serves to reveal therelationship between the coatings, their substrate rocks, andformation environments. Finally, the results of the study areapplied to Mars to determine the detectability and possible

appearance of coatings in both in situ and remote sensingstudies of Martian surface materials.

2. Analytical Procedure

2.1. Scanning Electron Microscope and ElectronMicroprobe Measurements

[5] In order to perform detailed microscopic analysis ofthe coatings, we created thin sections of centimeter-sizedchips of each of the five Hawaiian samples. We focused onsample surfaces that contained coating materials represen-tative of the coated surfaces as a whole. Each sample chipwas coated in epoxy prior to cutting, grinding, and polishingin order to avoid chipping or removal of the coating duringthe sectioning process. The curation temperature of theepoxy used on the samples and thin sections was keptbelow 80�C to avoid potential water loss from the samples.[6] We determined coating morphology, thickness, dis-

tribution, and qualitative chemistry with the JEOL 845scanning electron microscope (SEM) at Arizona StateUniversity. Backscattered electron (BSE) imaging permit-ted assessment of coating morphology and measurement ofcoating thickness. We measured qualitative chemistryusing both spot analyses and elemental x-ray mapsobtained from the energy dispersive spectroscopy (EDS)system on the SEM. X-ray maps of Si, Al, Fe, Mg, Ca,Na, K, Ti, and S were also utilized to examine potentialrelationships between the coatings and substrate basalts.For the spot analyses, data were collected over 10-sintervals of time. Each x-ray map is 256 � 256 pixelswith pixel size dependent on the magnification settingfor each image. All SEM measurements were made at

Figure 1. Satellite imagery of the collection sites for the five samples of this study (MIO, MIY, MUO,KWY, and KW).

E05015 MINITTI ET AL.: HAWAIIAN ROCK COATINGS

2 of 24

E05015

an accelerating voltage of 15 kV with a focused beam(�1–2 mm spot) at current levels between 0.2 and 0.6 nA.[7] To clarify the nature of and elemental distribution in

the coating materials, we obtained quantitative chemicaldata from the coatings and their respective substrate basaltsusing the JEOL JXA-8600 electron probe at Arizona StateUniversity. Analyses were made with a focused, 10 nAbeam at an accelerating voltage of 15 kV.

2.2. Visible and Near Infrared Spectroscopy

[8] Visible and near infrared (VNIR) spectra of naturalcoated surfaces and sample interiors exposed by cutting toproduce planar surfaces were obtained in the ReflectanceExperiment Laboratory (RELAB) facility at Brown Uni-versity. Reflectance spectra between 0.32 and 2.55 mm wereobtained at 10�nm sampling intervals from the bidirectionalspectrometer with 30� incidence and 0� emergence anglesas in past experiments [e.g., Pieters, 1983]. Coating spectrawere collected from focused (1-mm diameter) spots in orderto isolate the spectral properties of each unique area of acoating and interior spectra were collected from broad(�1-cm diameter) areas.

2.3. Thermal Infrared Spectroscopy

[9] Thermal infrared (TIR) spectra of unprepared coatedsurfaces and sample interiors exposed by cutting wereobtained in the thermal emission spectroscopy laboratoryof the Mars Space Flight Facility at Arizona State Univer-sity. Emission spectra were collected from broad (�1-cmdiameter) spots on the samples between 200 and 2000 cm�1

at 2 cm�1 spectral sampling. Collection of TIR spectra fromsample chips generally led to high-quality spectra with goodspectral contrast. Consistent collection of both TIR andVNIR spectra from natural surfaces of samples facilitatescomparison of the data sets with each other and permitshigher fidelity comparisons to planetary surface spectra.

3. Data Analyses

3.1. Description of Samples

[10] Each of the five different coatings of this study,collected at a variety of vents and flows on the Big Islandof Hawaii, has a distinctive color, thickness, chemistry,morphology, and spectral character (Figure 2). All thesamples were collected in August 1999 within a 1-week

Figure 2. Coated surfaces of the five Hawaiian samples of this study. Colored portions of the coatingsdescribed in the text are designated with arrows and labels and black portions of the coatings are theuncolored regions of the coatings.


3 of 24

E05015

interval. Annual temperatures for the sample locationsaverage between 21 and 27�C. Two samples were collectedin the Mauna Iki flow field along the Kilauea SW rift zone(labeled MIY and MIO for the yellow and orange coatings,respectively), located about 10 km from Kilauea caldera(Figure 1). The rift zone consists of a linear fissure with

numerous spatter cones. The MIY and MIO samples werefrom the dry climate (rainfall �125 cm/yr) of the Ka’uDesert. Both MIY and MIO were collected from a 1920flow. Another sample with an orange coating was collectedfrom a lava channel at Mauna Ulu (MUO), about 10 kmsoutheast of Kilauea caldera (Figure 1). Mauna Ulu lasterupted between 1969 and 1974 and the sample representslava erupted in 1973. Compared to the Mauna Iki samples,MUO is located in a wetter environment (rainfall �190–200 cm/yr). Finally, two samples with white-and-yellow(KWY) and white (KW) coatings were acquired from 1982flows within 100 m south of the degassing Halemaumauvent at Kilauea. On KWY, the white-and-yellow coatingswere commonly seen on the lava surfaces around the vent,with the yellow coating dominating closer to and downwindof the vent. Outgassing from the active Halemaumau senta constant stream of gases across the surfaces of thesetwo samples. Rainfall is similar to that for MUO, butwater vapor emitted by Halemaumau added to the totalamount of water deposited on the KW and KWY samples.

Table 1. Average Substrate Basalt Glass Compositions Obtained

From Electron Probe Analyses

MIO MIY MUO KWY KW

SiO2 49.62 49.37 49.84 50.82 50.24Al2O3 14.39 14.43 13.79 14.55 14.49MgO 6.67 6.59 8.34 6.89 6.66FeO 10.88 10.94 11.17 10.64 10.62CaO 11.58 11.55 11.12 11.25 11.65Na2O 2.54 2.58 2.39 2.77 2.51K2O 0.50 0.51 0.41 0.49 0.43TiO2 2.80 2.86 2.35 2.70 2.47MnO 0.18 0.16 0.16 0.15 0.17P2O5 0.30 0.31 0.24 0.26 0.24Total 99.46 99.31 99.80 100.51 99.48

Figure 3. BSE images of the four coating morphologies present on the surface of MIO. Arrows point tothe coating morphology in each image; substrate glass is labeled (gl). (a) Morphology 1: thick, marble-textured Fe-bearing SiO2; light-toned layers are enriched in Fe and Ti. (b) Morphology 2: moderatelythick, dense Fe-bearing SiO2 with Fe- and Ti-enriched layers (light-toned layers). (c) Morphology 3: thin,lacy-textured coating containing mineral fragments held in a SiO2 matrix. (d) Morphology 4: thin,marble-textured SiO2 intermittently capped with a Fe- and Ti-enriched layer (bright, outermost materialon coating).


4 of 24

E05015

Despite having distinctive coatings, the substrate basaltsof all five samples are similar in composition. Each basaltis predominantly glassy (85–95%) with phenocrysts ofolivine, augitic pyroxene, anorthite, and rare chromite. Thecompositions of the glass matrices of the basalts arealso similar (Table 1).

3.2. Mauna Iki Orange

3.2.1. Morphology and Chemistry[11] The Mauna Iki Orange (MIO) sample was collected

adjacent to the Mauna Iki fissure vent in a small area wherelava blocks on the order of 5–10 cm were breaking apartalong the surface. MIO represents one of these blocks thatwas loose on the surface (Figure 2). Inspection of the fissurevent indicates that much of the lava within a meter of thefissure is red in color. The sample we collected was locateda few meters away from the fissure and these red rocks in anarea where the rocks appeared loose and orange in color.The sample has an orange and brown coating whose colorand texture varies across the sample surface. The uncoloredareas interspersed among the orange and brown areas on thecoated surface are not uncoated; rather, they have a greasyluster that, when chipped off, differs from the vitreousappearance of the substrate basalt. The orange coating hasa ceramic-like appearance (earthy luster with a mattetexture, like unglazed porcelain) while the brown coatinghas a granular texture. The orange and brown coatingmaterials are either intermixed with or deposited on theuncolored lustrous coating material. The sample is vesicularbut the top surface has fewer vesicles visible apparentlybecause the coating has filled the holes to create a smooth,flat surface.[12] Variation of the coating appearance across the sur-

face is echoed in variations in coating morphologies at themicroscale. MIO exhibits four different types of coatingmorphologies, the widest variety of coating morphologiesobserved in any of the five samples. Morphology 1 consistsof Fe-bearing SiO2 with a marble-like texture that containsthin layers of Fe- and Ti-rich material (Figure 3a).This morphology represents the thickest occurrence ofcoating material on MIO, ranging from 20–80 mm thick.

Morphology 2 resembles the first morphology, but has adenser (fewer visible pores and fractures) Fe-bearing SiO2

matrix and more Fe- and Ti-bearing layers (Figure 3b). Thismorphology tends to occur as thinner coatings, rangingfrom 20–25 mm. Morphology 3 consists of 1–5 mm-sizedfragments of various minerals (pyroxene, feldspar, andoxides) held together by a SiO2-rich matrix in an opennetwork yielding a lace-like texture (Figure 3c). Morphol-ogy 4 consists of a 2–8 mm thick Fe-bearing SiO2 layeroccasionally capped by a thin (<1 mm) Fe- and Ti-rich layer(Figure 3d). Morphology 4, with or without the Fe- andTi-rich capping layer, is found across much of the coatedsurface of MIO. Overall, very little of the centimeter-longsection studied of the coated surface of MIO is free of oneof the above coatings. The coatings do not grade into oneanother but in some areas, Morphology 4 is found betweenone of the other coating morphologies and the substratebasalt. X-ray maps reveal that the substrate basalt remainsconstant in composition up to the basalt-coating interface.[13] Visual inspection of the fragment used to create the

studied thin section allows some correlation of the coatingmorphologies observed at the macroscopic and microscopicscales. Morphologies 1 and 2 (Figures 3a and 3b) corre-spond to the portions of the coated surface of MIO that havethe orange, ceramic-like coating material. Morphology 3(Figure 3c) corresponds to the granular brown coating.Morphology 4 without the Fe- and Ti-rich capping layermost likely corresponds to the uncolored portions of thecoating. At the microscopic scale, uncapped Morphology 4is observed underneath the thicker coating morphologiesand, at the macroscopic scale, the uncolored coating isobserved under the orange and brown coatings. It is notclear which manifestation of the coating corresponds toMorphology 4 with the Fe- and Ti-rich capping layer. It ispossible that capped Morphology 4 also represents theuncolored portions of the coating and/or represents weaklycolored areas of the coating interspersed between occur-rences of the orange and brown coatings.3.2.2. Electron Probe Results[14] The analyses of the SiO2-rich matrices of the coat-

ings have low oxide totals (Table 2). While the marble-liketextures of coating Morphologies 1 and 4 (Figures 3a and3d) suggest that porosity may contribute to the low totals, theless porous version of the orange coating (Morphology 2,Figure 3b) also has a low total (Table 2). Thus, the lowtotals most likely reveal that the coatings are water-bearingSiO2-rich coatings, making the coatings opaline. We do nothave mineralogical evidence specifically identifying a par-ticular form of opal, but instead use the term opaline todescribe the nature of the coating matrices (water-bearingSiO2). The bright layers and caps exhibited in the orangecoatings (Figure 3) are too narrow to sample alone with theelectron beam, but they have higher concentrations of Feand Ti than the matrix SiO2 and they are more Fe-rich thanTi-rich (Table 3). The Fe- and Ti-bearing cap on the thinSiO2 coating has subequal amounts of Fe and Ti (Table 3).Overall, analyses of the SiO2-rich matrices of all the coatingmorphologies contain minor (1–5 wt.%) amounts of Fe andTi and very low concentrations (<1 wt.%) of all elements.3.2.3. Spectral Properties[15] In the VNIR, the spectrum of an orange coated

surface has high reflectivity and exhibits several distinct

Table 2. Representative Coating Compositions Obtained From

Electron Probe Analyses: SiO2 Matrices

MIOa MIOb MIOc MIYd KWYe KWYf KWg KWh

SiO2 62.95 62.76 62.54 68.15 73.66 55.04 74.04 60.46Al2O3 0.13 0.23 0.88 0.78 0.24 0.11 0.26 0.22MgO 0.41 0.12 0.24 0.18 0.07 0.12 0.03 0.34FeO 4.38 2.75 1.24 0.39 0.67 0.47 0.28 0.54CaO 0.52 0.10 1.29 0.35 0.10 0.05 0.08 0.34Na2O 0.05 0.05 0.71 0.43 0.00 0.06 0.08 0.13K2O 0.10 0.13 0.91 0.35 0.03 0.08 0.05 0.17TiO2 2.61 1.15 0.73 0.39 2.49 0.43 0.12 0.29MnO 0.04 0.01 0.02 0.00 0.01 0.04 0.00 0.03P2O5 0.13 0.12 0.05 0.00 0.05 0.00 0.00 0.09Total 71.31 67.41 68.62 71.02 77.31 56.40 74.93 62.62

aMIO Morphology 1.bMIO Morphology 2.cMIO Morphology 4.dMIY matrix.eKWY marble-textured coating.fKWY non-porous coating.gKW marble-textured coating.hKW non-porous coating.


5 of 24

E05015

features (Figure 4). The spectrum has a major absorptionband at <0.7 mm and minor features, expressed as inflec-tions, at 0.48, �0.66, and �0.85 mm. Bands also occur near1.0, 1.45, 1.93, and 2.3 mm. These absorptions are over-printed on a negative spectral slope extending from thereflectance peak at 0.74 mm to longer wavelengths. Suchnegative spectral slopes are characteristic of the VNIRspectra of coated basalts [Singer and McCord, 1979; Fischerand Pieters, 1993]. The spectrum of an uncolored area on thecoated surface of MIO clearly demonstrates that whileuncolored, it is not uncoated. Like the spectrum of theorange coating, the uncolored surface spectrum has a nega-tive continuum slope and absorptions at 0.48 and 1.93 mm(Figure 4), but it is less reflective at all wavelengths. Thespectral character of a region of brown coating was notassessed because it was not recognized as a potentiallydistinct coating type until after the spectral data werecollected. The interior spectrum of MIO, taken after cuttingthe sample to expose an uncoated surface, has a reflectivitycomparable to the uncolored coating spectrum and clearlydiffers from the spectra of the coatings (Figure 4). The

interior spectrum exhibits strong absorptions near 1 and2.25 mm and weaker absorptions near 0.5 and 1.3 mm.[16] In the TIR, a coating spectrum obtained from a

surface exhibiting a combination of the orange, brown,and uncolored coatings displays two strong absorptionsnear 1100 and 470 cm�1, a weaker absorption near1250 cm�1 that manifests itself as a shoulder on the1100 cm�1 feature and two weak features near 900 and780 cm�1 (Figure 5). The spectrum of the coated surfaceof MIO significantly differs from the interior spectrum,which has a single, boxy absorption between 800 and1200 cm�1 and a single, broad absorption at <600 cm�1

Table 3. Representative Coating Compositions Obtained From

Electron Probe Analyses: Fe- and Ti-Enriched Layers

MIOa MIOb MIYc MUOd KWYe KWf

SiO2 33.36 33.44 57.39 56.87 53.42 42.00Al2O3 0.57 0.19 0.80 11.94 0.07 0.06MgO 0.10 0.05 0.21 7.15 0.10 0.03FeO 43.18 3.29 5.33 9.14 2.72 1.96CaO 0.15 0.31 0.29 8.97 0.73 0.10Na2O 0.45 0.93 0.41 2.62 0.13 0.02K2O 0.26 0.79 0.41 0.40 0.06 0.15TiO2 4.28 4.75 12.42 2.31 14.31 38.05MnO 0.02 0.02 0.00 0.14 0.00 0.00P2O5 1.79 0.61 0.77 0.40 0.62 1.20Total 84.17 44.41 78.04 99.94 72.16 83.58

aBright layers in Morphologies 1 and 2.bCapping layer on Morphology 4.cCapping layer.dOverlapping coating and substrate analysis.eCapping layer on marble-textured coating.fCapping layer on marble-textured coating.

Figure 4. VNIR spectra from the orange (black solid) anduncolored (black dashed) portions of the MIO coating andfrom the interior (gray) of MIO. Interior spectrum is offset�0.04 along the reflectance axis for clarity.

Figure 5. Thermal infrared spectra of coated samplesurfaces as seen in Figure 2. Some spectra are offset alongthe emissivity axis for clarity (KWY, �0.04; MIY, �0.08;KW, �0.12; MIO, �0.14). Sharp features at �670 cm�1 aredue to incompletely removed CO2 vapor in the samplechamber.

Figure 6. Thermal infrared spectra of four of the sampleinteriors from cut surfaces; no interior spectrum wasobtained from KW. Some spectra are offset along theemissivity axis for clarity (KWY, �0.02; MUO, �0.03;MIY, �0.06). Sharp features at �670 cm�1 are due toincompletely removed CO2 vapor in the sample chamber.


6 of 24

E05015

(Figure 6). The low spectral contrast of the interiorspectrum is likely due to the porous nature of the MIObasalt and the scattering of energy [e.g., Lyon, 1964;Aronson et al., 1966; Hunt and Vincent, 1968; Salisbury andWald, 1992; Moersch and Christensen, 1995; Mustard andHays, 1997; Lane, 1999].

3.3. Mauna Iki Yellow

3.3.1. Morphology and Chemistry[17] The Mauna Iki Yellow (MIY) sample has a far less

variable coating than the MIO sample. The colored portionsof the coating are orange-yellow and exhibit a ceramic-liketexture that is uniform across the dense basalt substrate. Thepatchy occurrences of the colored coating are found almostexclusively in low-lying areas on the sample surface. Theuncolored areas of MIY, like MIO, are coated with amaterial that has greasy luster that differs from the colorand luster of the underlying basalt. The colored coating,where present, always overlies the uncolored, greasy lustercoating. The rock is relatively dense compared to MIO andis likely a piece of a lava flow rather than spatter. Olivinegrains, 1 mm in size, are visible within the interior of therock but not on the coated surface.[18] The bimodal nature of the MIY coating is con-

firmed at the microscale. One coating morphology simplyconsists of a marble-textured SiO2-rich coating that rangesfrom 5–7 mm thick (Figure 7a). The second coatingmorphology is also marble-textured SiO2, but is cappedby an Fe- and Ti-bearing layer <1 mm thick (Figure 7b).Both coatings lie in direct contact with the substrate basaltwhich exhibits no chemical gradient leading up to thecoating-basalt interface. Visual inspection of the samplechip used to make the studied thin section appears to linkthe colored portions of the coating to the Fe- and Ti-capped coating morphology and the uncolored portions tothe uncapped SiO2 morphology.3.3.2. Electron Probe Results[19] The MIY coatings exhibit low oxide totals, which

are again likely due to combination of porosity and water(Table 2). Unlike MIO, the opaline matrix of the coatings

has low concentrations (<1 wt.%) of Fe and Ti, similar tothe concentrations of the other elements in the coatings.The bright caps present on some of the SiO2 coatings areenriched in Fe and Ti relative to the underlying opalinematrix, with Ti present in equal or greater amounts than Fe(Table 3).3.3.3. Spectral Properties[20] In the VNIR, the orange-yellow coating has a high

reflectivity with a reflectivity maximum at 0.65 mm and anabsorption edge to the short wavelength side of the reflec-tivity maximum that has an inflection at 0.48 mm (Figure 8).At wavelengths longer than 0.65 mm, the spectrum has anegative slope which flattens out near a minor inflection at1.9 mm. The uncolored coating spectrum has very lowreflectivity and is nearly featureless although it also exhibitsa negative spectral slope (Figure 8). The spectra of thecoatings significantly differ from the spectrum of the MIYinterior (Figure 8). The interior spectrum exhibits a broadhump between �0.45 and 0.75 mm, strong absorptions near1.0 and 2.3 mm and a weaker absorption near 1.3 mm.

Figure 7. BSE images of coating types present on MIY. Arrows point to the coating type in each image;substrate glass is labeled (gl). (a) Thin, marble-textured SiO2. (b) Thin, marble-textured SiO2 with Fe-and Ti-enriched capping layer (bright material at outermost edge of coating).

Figure 8. VNIR spectra from the orange-yellow (blacksolid) and uncolored (black dashed) portions of the MIYcoating and from the interior (gray) of MIY.


7 of 24

E05015

[21] In the TIR, the coating spectrum has strong absorp-tions near 1100 and 450 cm�1 and a weak absorption near1250 cm�1 (Figure 5). The small inflection at 667 cm�1 isdue to a small amount of CO2 gas in the sample chamber.The interior spectrum has a broad, squared-off shapebetween 800 and 1200 cm�1 and three weak absorptionsbetween 250 and 550 cm�1 (Figure 6). The high spectralcontrast of the interior spectrum results from the densenature of the MIY sample and lack of substantial scatteringof the emitted energy.

3.4. Mauna Ulu Orange

3.4.1. Morphology and Chemistry[22] The Mauna Ulu Orange (MUO) sample has a dis-

continuous, powdery orange coating over a porous basaltsubstrate that can be wiped off the rock surface unlike thecoatings on the other four samples that are cemented ontotheir substrate basalts. The sample was collected from asmall surface lava channel that appeared entirely covered bythe powdery orange coating. Distribution of the orangecoating is not controlled by sample topography, thoughtopographic lows have higher concentrations of the orangematerial. The uncolored areas of the coated surface corre-spond to a dull, apparently altered version of the substratebasalt. Unlike the orange coating, the uncolored coating isaffixed to the substrate basalt. The colored coating is alwaysfound on top of the uncolored coating.[23] SEM analyses reveal that MUO is coated with a very

thin (2–3 mm), marble-textured, SiO2-enriched (relative tothe substrate basalt) layer that is nearly ubiquitous acrossthe surface of the studied sample (Figure 9). The outermostportion of the SiO2-enriched coating is discontinuouslycapped by Fe- and Ti-enriched material. Through visual

inspection of the thin-sectioned MUO sample, it was notpossible to definitively link macroscale and microscalecoating morphology. It is likely, though, that the ubiquitousSiO2 coating corresponds to the uncolored coating and thatthe Fe- and Ti-enriched regions on the coating correspond tothe powdery, colored portions of the coating. Beneath thethin, SiO2 coating, the substrate basalt exhibits no chemicalvariation at or near the basalt-coating interface.3.4.2. Electron Probe Results[24] The thinness of the coating required us to retrieve

data from an analysis point overlapping the coating andsubstrate basalt. These analyses provide support for thequalitative chemical analyses from the SEM because theyillustrate that the coating is enriched in SiO2 relative to thesubstrate basalt (Tables 1 and 3). Enrichments of Fe and Tiat the outermost portion of the coating apparent in theelemental maps from the SEM (Figure 9) do not appear inthe electron probe data most likely due to the very small sizeof the enriched zones relative to the analysis spot size.3.4.3. Spectral Properties[25] In the VNIR, the orange coating spectrum resembles

the colored coating spectrum of MIY, but with lowerreflectivity (Figure 10). The spectrum has a reflectivitymaximum at 0.62 mm, a strong absorption band at wave-lengths <0.6 mm and a minor inflection at 0.48 mm. Atwavelengths longer than 0.62 mm, the spectrum has anegative slope that flattens out at wavelengths longer than1.9 mm. The uncolored coating spectrum is, in effect, aless-reflective version of the colored coating spectrum(Figure 10). It has a broad reflectivity maximum at shorterwavelengths (0.56 mm) and reaches a flat continuum slopeat shorter wavelengths (�1.2 mm) than the colored coatingspectrum, but shares the 0.48 mm inflection found in thecolored coating spectrum. The interior spectrum is darkwith broad features centered near 1.1 and 1.9 mm and aweak feature at <0.64 mm (Figure 10).[26] In the TIR, the coated surface spectrum has three

dominant absorptions at 1100, �880, and �470 cm�1 andone minor absorption near 1250 cm�1 (Figure 5). Theinterior spectrum exhibits a single, U-shaped absorption

Figure 9. BSE image of the coating present on MUO.Arrow points to the coating; substrate glass is labeled (gl).Coating is very thin, marble-textured SiO2 with discon-tinuously distributed enrichments of Fe and Ti (brightermaterial at outermost edge of coating).

Figure 10. VNIR spectra from the orange (black solid)and uncolored (black dashed) portions of the MUO coatingand from the interior (gray) of MUO. Interior spectrum isoffset �0.01 along the reflectance axis for clarity.


8 of 24

E05015

between 1200 and 800 cm�1 and a single, broad absorptionat wave numbers smaller than �600 cm�1 (Figure 6).

3.5. Kilauea White-Yellow

3.5.1. Morphology and Chemistry[27] KilaueaWhite-Yellow (KWY) has a mixture of white,

yellow, and uncolored coatings and was taken from a veryvesicular lava fragment/pumice near the Halemaumau vent.Distribution of the white and yellow coatings is controlled tosome degree by sample topography in that neither coating sitsatop bulges or ridges along the sample surface. The whitecoating has an opaque, chalk-like appearance and is stratig-raphically the lowest coating morphology. The white coatingis also found lining vesicles interior to the surface of thesample. The uncolored coating has a greasy, translucentappearance and it has a gradational relationship with theyellow coating which has a ceramic-like appearance. Theuncolored coating is commonly, but not exclusively, foundover the topographically high portions of the sample surfaceand the gradation of the uncolored coating to the yellowcoating occurs at the base of the topographical highs. Wherethe yellow and the uncolored coatings have this geometry, theyellow coating always overlies the white coating. It is notclear if the white coating also extends under the uncoloredportions of the coating.[28] At the microscopic level, KWY has two distinct

coating morphologies: a marble-textured, nearly pure SiO2

material commonly capped by a Ti-rich layer (Figure 11a)and a S-bearing SiO2 material with a less porous SiO2

matrix than the former material. The S-bearing SiO2 coatingis unlike coatings observed on the MIO, MIY, and MUOsamples and commonly has minerals such as feldspars andoxides embedded in the matrix (Figure 11b). In someoccurrences, the S-bearing SiO2 matrix contains minoramounts of Fe. The coating morphologies are present eitherdirectly in contact with the substrate basalt or with themarble-textured coating underlying the S-bearing SiO2

material; the S-bearing SiO2 coating is never observed

below the marble-textured coating. Both coating morphol-ogies are between 10 and 70 mm thick. The substrate basaltremains constant in composition up to the basalt-coatinginterfaces with both coating types.[29] On the basis of the combination of observations at

the macroscopic and microscopic scales, the white coatingappears to correspond to the Ti-capped, marble-texturedSiO2 coating. This association is supported by the observa-tions that the white coating is always underneath theuncolored and yellow coatings and the Ti-capped SiO2

coating is found underneath the S-bearing SiO2 coating.This observation necessitates that dense, mineral- and S-bearing SiO2 coating represents both the uncolored andyellow coatings.3.5.2. Electron Probe Results[30] Opaline coatings for both coating morphologies are

indicated by low oxide totals. The dense coatings have totalsas low as the marble-textured coatings, suggesting that thepresence of water is the primary factor in the low totals(Table 2). The SiO2 of the marble-textured coating isaccompanied by minor amounts of FeO (0.2–0.7 wt.%)and TiO2 (0.3–2.5 wt.%), with TiO2 typically in greaterabundance than FeO. All other oxides are at concentrations<1 wt.%. The caps of the marble-textured coatings are alsopredominantly SiO2, FeO, and TiO2, with TiO2 present inmuch greater abundance than FeO (Table 3). The dense SiO2

coating has minor, subequal (0.2–0.5 wt.%) amounts of FeOand TiO2 (Table 2). The S content of the matrix of the densecoating was difficult to quantify with the electron probe dueto the volatile nature of the element. Qualitatively, the SEMx-ray maps of S, Fe, and Ti in the dense coating suggestconcentrations of S > Ti > Fe. In contrast, SEM x-ray datafrom the KWY marble-textured coating and any othercoating morphology on MIO, MIY, or MUO show noindication of the presence of S in these coatings.3.5.3. Spectral Properties[31] Despite different appearances, the three coating

colors on KWY exhibit similar spectral properties in the

Figure 11. BSE images of coating types present on KWY. Arrows point to the coating type in each image;substrate glass is labeled (gl). (a) Moderately thick, marble-textured SiO2 with a Ti-enriched capping layer(bright material at outermost edge of coating). (b) Moderately thick, mineral- and S-bearing SiO2.


9 of 24

E05015

VNIR. The yellow coating spectrum is the most reflectivewith a strong absorption band at wavelengths <0.65 mm, abroad reflectance maximum between 0.65 and 0.75 mmand a negative spectral slope at wavelengths >0.75 mm(Figure 12). Two bands near 1.8 and 1.95 mm are over-printed on the negative spectral slope. Like the coloredMIO, MIY, and MUO coatings, the yellow KWY coatingalso exhibits a minor absorption feature near 0.48 mm. Theprimary differences between the white and yellow coatingsare that the white coating spectrum is darker, has a majorabsorption band at <0.46 mm, and a broad reflectancemaximum from about 0.4 to 0.7 mm (Figure 12). Thewhite coating spectrum shares the minor absorption featurenear 0.48 mm and the bands near 1.8 and 1.95 mm withthe yellow coating. The uncolored coating spectrum isnearly identical to the white coating spectrum but is less

reflective than the white coating (Figure 12). The spectraof all three coating colors differ from that of the interiorspectrum which is dark but exhibits weak absorptions at<0.7 mm, near 1 and 2 mm.[32] In the TIR, a spectrum from a surface containing all

three coating colors exhibits a pronounced feature near1100 cm�1, a smaller shoulder near �1250 cm�1, and afeature near 470 cm�1 (Figure 5). The feature at 667 cm�1

is again due to small amounts of CO2 gas in the samplechamber during sample measurement. This spectrum issimilar to the other glass spectra and suggests that theminor amount of sulfur in the coating does not affect theoverall spectral character of the coating. The interiorspectrum is unlike those of the other samples (Figure 6).It exhibits absorptions near 1100, 900, and 470 cm�1,which resemble the features in the MUO coating spectrum(Figure 5). The low spectral contrast of the interior spec-trum is likely due to the porous nature of the sample.

3.6. Kilauea White

3.6.1. Morphology and Chemistry[33] The Kilauea White (KW) sample has a white coating

accreted over a large fraction of a very porous, glassy pillowlava. The white coating can be divided into two morphol-ogies: an opaque, chalk-like white coating and an opaque,ceramic-like white coating. The distribution of the whitecoatings is not controlled by sample topography, particularlythe chalky coating which appears underneath ledges ofbroken vesicles and inside vesicles interior to the sample.Where the ceramic coating is observed, it is stratigraphicallyabove the chalky coating. The uncolored areas of KW arecoated with a translucent material that has a greasy luster.The relationship between the uncolored and both whitecoatings is gradational, with transitions from uncolored tomilky to chalky or ceramic coatings common across thecoated sample surface.[34] At the microscopic scale, two coating morphologies

are present on KW and both are nearly identical to thosefound on KWY (Figures 11 and 13). One minor difference

Figure 12. VNIR spectra from the yellow (as labeled),white (as labeled) and uncolored (black dashed) portions ofthe KWY coating and from the interior (gray) of KWY.Interior spectrum is offset �0.01 along the reflectance axisfor clarity.

Figure 13. BSE images of coating types present on KW. Arrows point to the coating type in eachimage; substrate glass is labeled (gl). (a) Moderately thick, marble-textured SiO2 with a Ti-enrichedcapping layer (bright material at outermost edge of coating). (b) Moderately thick, mineral- and S-bearingSiO2.


10 of 24

E05015

is that the marble-textured, pure SiO2 coating (Figure 13a)possesses a Ti-rich cap less frequently than the comparablecoating on KWY. The range of KW coating thicknesses is10–35 mm. As with KWY, the S-bearing coating (Figure 13b)of KW is observed overlying the marble-textured SiO2

coating and the opposite is not observed. No chemicalvariation is observed in the substrate basalt underlying bothcoating types at or near the basalt-coating interface. Corre-lations between macroscopic and microscopic observationssuggest that the chalky white coating corresponds to themarble-textured SiO2 coating and the ceramic white coatingcorresponds to the S-bearing coating. Because the unco-lored coating is observed to merge with both the chalky andceramic coatings, it is not clear which microscopic coatingmorphology (or if both morphologies) correspond(s) to theuncolored coating.3.6.2. Electron Probe Results[35] Low oxide totals indicative of opaline coatings are

obtained in analyses of both coating morphologies (Table 2).SiO2 in the marble-textured coating is accompanied byminor amounts of FeO and TiO2 (0.1–0.3 wt.%), smallerthan those observed in the comparable SiO2 coatings on theother samples. The capping layer on the marble-texturedcoating contains FeO and TiO2, but TiO2 exists in muchhigher concentrations relative to FeO (Table 3). The mineral-and S-bearing coating is comparable in composition to theanalogous coating on KWY, with similar small amounts ofFeO and TiO2 (Table 2). S was again difficult to measurequantitatively with the electron probe but qualitative SEMx-ray data suggest that S is present in greater quantity thanFe or Ti. Comparison of the SEM x-ray data of the S- andmineral-bearing coatings of KWY and KW suggest thattheir S contents are similar.3.6.3. Spectral Properties[36] In the VNIR, the spectrum of a white coated surface

(powdery and opaque not differentiated) exhibits a strongabsorption band near 0.3 mm, a reflectance maximum at0.44 mm and a consistently negative spectral slope atwavelengths longer than 0.44 mm (Figure 14). The positionof the reflectance maximum is shifted to shorter wave-lengths relative to the positions of the maxima of the MIO,MIY, MUO, and KWY colored coating spectra (Figure 15).

Another notable characteristic of the white coating spectrumis its lack of a 0.48-mm feature. The uncolored coating hasvery low reflectance and has an absorption band near0.3 mm, a sharp reflectance peak at 0.37 mm, and anegatively sloping continuum at wavelengths longer thanthe reflectance peak (Figure 14). The interior spectrum isthe darkest of all the interior VNIR spectra and is flat andnearly featureless. Subtle detail in the interior spectrumdiscernable when the spectrum is viewed on an extendedreflectance scale (not shown) resembles that of the MUOand KWY interior spectra (Figure 14).[37] In the TIR, a coated surface spectrum resembles

those of the other coating spectra, with strong absorptionsnear 1100 and 470 cm�1, a shoulder near 1250 cm�1 and aweak absorption near 900 cm�1 (Figure 5). The highlyporous nature of the KW basalt precluded obtaining a high-quality TIR spectrum of the sample interior. It is anticipatedthat due to the similar modal mineralogy of the KW basaltto the other four basalts, the spectrum of the KW interiorwould share the same characteristics as the interior spectraof the other basalts.

4. Summary

[38] The five coated Hawaiian basalts described in detailabove and summarized in Table 4 share notable similaritiesand differences. All five samples possess a commonmarble-textured, opaline coating capped by an FeO- and/or TiO2-rich layer. Unique coatings exist on MIO, KWY,and KW. MIO has a dense opaline coating with FeO- andTiO2-bearing layers (Morphology 2, Figure 3b) and amineral-bearing coating (Morphology 3, Figure 3c). KWYand KW have dense, mineral- and S-bearing opalinecoatings that overlie the marble-textured coatings of eachsample (Figures 11 and 13). All of the coating typescontain very little Al2O3 (<1 wt.%), making them unlikeother SiO2-rich coatings or palagonites found on Hawaii(Table 2; Hay and Jones [1972]; Farr and Adams [1984];Curtiss et al. [1985]; Morris et al. [2000]). Despite theirdistinctive colors, the overall VNIR character of thecoatings is similar, with subtle differences only in thespecific wavelength of reflectance maxima and the pres-ence of minor features. The overall TIR character of the

Figure 14. VNIR spectra from the orange (black solid)and uncolored (black dashed) portions of the KW coatingand from the interior (gray) of KW.

Figure 15. Collected VNIR spectra of the colored coatingsfrom each of the five samples.


11 of 24

E05015

coatings is very similar, with differences primarily con-fined to the relative depths of absorption features thatlikely are related to the surface textures and/or porosities(Figure 5). The coating spectra of all five samples differsignificantly from their corresponding interior spectra,illustrating the ability of coatings to mask the spectralcharacter of primary lithologies even when the coatings arethin. The following discussion seeks to interpret thesimilarities and differences between the coatings and placethem in context of the formation mechanisms and environ-ments responsible for formation of the coatings.

5. Discussion

5.1. Coating Formation Conditions

[39] Given the large proportion of glass in the fiveHawaiian samples of this study, understanding the behaviorof glass under alteration conditions that exist in the regionwhere the samples were collected is important in interpretingthe origin of the coatings. Past studies of alteration productsfrom Hawaiian rocks constrain the conditions that likelycontrolled alteration of the rocks of this study. Water isalmost exclusively cited as the altering fluid of Hawaiianlithologies [Macdonald, 1944; Hay and Jones, 1972; Farrand Adams, 1984; Curtiss et al., 1985; Morris et al., 2000].All five samples of this study originate from areas withlevels of rainfall (�120–200 cm/yr) that are sufficient tosupport glass alteration given the similarly wet or drierregions in which glass alteration has been recorded [e.g.,Hay and Jones, 1972; Farr and Adams, 1984; Curtiss et al.,1985]. Another potential source of water is high-temperaturesteam alteration associated with active vents [Macdonald,1944; Morris et al., 2000]. KWY and KW, which originatefrom the degassing Halemaumau vent, could have beensubjected to such steam vent alteration. However, MUOdid not originate near active vents and MIO and MIY occurboth upwind and downwind from vents that were activepostdeposition, making it unlikely that their coatings arerelated to degassing (S.K. Rowland, personal communica-tion). Thus, we assume that all the coatings of this studyformed as meteoric water interacted with the substrate glass.The involvement of water in coating formation is alsoimplicated by the low oxide totals obtained from coatinganalyses (Table 2). We do not expect the coatings to be aproduct of direct deposition from volcanic gases [e.g.,Chemtob et al., 2006] given the consistency of the coatingchemistry among samples despite their varying spacingrelative to active vents.

[40] Precipitation over the regions where the samples ofthis study were collected is consistently acidic in botheruptive and non-eruptive periods of Kilauea, from pH � 4at Mauna Iki to a pH as low �3 at Halemaumau [Hardingand Miller, 1982]. The pH of precipitation in these areas islower than the average island precipitation pH because rainfalling within �10 km of the summit is influenced byvolcanic emissions [Harding and Miller, 1982]. Incorpora-tion of S from active vent gases is cited as a factor increating acidic altering fluids [Macdonald, 1944; Banin etal., 1997; Morris et al., 1996, 2000; Tosca et al., 2004]which lead to unique alteration products. For example, thejarosite in Hawaiian alteration products is interpreted as theproduct of weathering in acidic fluids [Banin et al., 1997;Morris et al., 1996; Golden et al., 1996; Schiffman et al.,2007] and requires a pH < 4 [van Breemen, 1982; Burns,1987; Burns and Fisher, 1990; Morris et al., 2000]. Thepresence of S in the KW and KWY coatings suggests thatvolcanic gases influenced the samples of this study.[41] The products of Hawaiian basalt glass alteration

also provide evidence for oxidizing conditions duringalteration. Hydrolytically altered and palagonitized samplesstudied by Hay and Jones [1972] and Morris et al. [2000]have elevated Fe3+/Fe2+ relative to their precursors. Sam-ples altered in proximity to steam vents also show elevatedFe3+/Fe2+ and elevated concentrations of Fe3+, Ti, and P[Morris et al., 2000]. The Fe3+, Ti, and P enrichments areinterpreted as the result of basalt glass leaching underoxidizing conditions [Morris et al., 2000]. Fe3+-relatedabsorptions in the VNIR spectra of the coatings (seesection 5.6 below) provide evidence of oxidizing condi-tions in the formation of the coatings of this study.

5.2. Coating Formation Mechanisms

[42] To understand the origin of the coatings, it isnecessary to understand the behavior of basaltic glass atambient temperature in aqueous, oxidizing, and acidicconditions. Can such conditions lead to predominantlySiO2 coatings with coexisting Fe- and/or Ti-rich materialsand with virtually no Al? Multiple laboratory studies on thebehavior of basalt glass in water suggest two possible end-member pathways by which conditions of Hawaiian alter-ation can lead to the observed coatings. The first pathwayinvolves leaching the basalt glass of nearly everything butSiO2 and later precipitating Fe- and Ti-bearing secondaryphases from the altering fluid. The second pathway involves

Table 4. Coating Characteristic Summary

MIOa

MIY MUO

KW KWY

M1 M2 M3 M4 Marble Mineral Marble Mineral

Thickness (mm) 20–80 20–25 10 2–8 5–10 3 10–70 10–70 10–35 10–35Color orange orange brown uncolored orange-yellow orange white white white yellowMarble-textured x x x x x xNonporous x x xMineral-bearing x x xMatrixb Fe Fe Fe Fe,Ti S SLayersc Fe Fe Fe, Ti Fe, Ti Fe, Ti Ti Ti

aM1, M2, M3, and M4 refer to morphologies 1, 2, 3, and 4, respectively.bElements in the SiO2 matrix of the coating.cDominant element(s) in internal or capping layers of the coating.


12 of 24

E05015

dissolving the basalt as a whole followed by coatingformation by secondary phase precipitation alone.

5.3. Leaching and Precipitation

[43] In the former scenario, leaching leads to nonstoi-chiometric (relative to the original glass chemistry) removalof cations from the glass as cations are exchangedfor protons from the water [El-Shamy et al., 1972; Bunkeret al., 1983; Berger et al., 1987; Guy and Schott, 1989;Thorseth et al., 1991]. Water at low pH encourages non-stoichiometric dissolution because acidic conditions enhanceion exchange and suppress network dissolution [Bunkeret al., 1983; Furnes, 1984; Guy and Schott, 1989; Thorsethet al., 1991; Stroncik and Schmincke, 2002; Crovisier et al.,2003]. Acidic conditions also lead to high dissolution rates[White, 1984; Gislason and Eugster, 1987; Guy and Schott,1989; Oelkers and Gislason, 2001]. Leaching leads to alayer depleted in univalent and divalent cations andenriched in higher valence cations (for example, Al, Si,and Ti) because lower valence cations are more easilyexchanged [White, 1984]. Retention of Si in a leached layeris observed in multiple studies [Bunker et al., 1983; White,1984; Guy and Schott, 1989; Thorseth et al., 1991; Eicket al., 1996; Tosca et al., 2004] and in the experimentsof El-Shamy et al. [1972] and Banin et al. [1997], Siremains in the altering basalt at pHs of < 7 and 3–4,respectively. Very acidic conditions are cited as the causebehind the transformation of basaltic samples to nearly pureopal in solfatarically altered rocks at Kilauea [Macdonald,1944]. Si is not completely immobile under acidic condi-tions, however. Si is observed in solution in fluids alteringbasalt glass in multiple studies [White, 1984; Guy andSchott, 1989; Oelkers and Gislason, 2001; Eick et al.,1996; Tosca et al., 2004]. To preserve a SiO2-rich leachedlayer, the pH of the system must remain low because theleaching process can transition to a stoichiometric process(i.e., uniform removal of all cations) at pH > 9 [El-Shamy etal., 1972; Crovisier et al., 1987; Thorseth et al., 1991; Dauxet al., 1994; Crovisier et al., 2003; Stroncik and Schmincke,2002]. The transition to stoichiometric dissolution occursbecause slightly basic pH permits breakage of Si–O bonds[El-Shamy et al., 1972; Paul and Zaman, 1978; Scholze,1988]. Rise in pH during basalt glass alteration is observedin closed systems [El-Shamy et al., 1972; Bunker et al.,1983; Crovisier et al., 1987; Gislason and Eugster, 1987;Jercinovic et al., 1990; Thorseth et al., 1991; Daux et al.,1994; Stroncik and Schmincke, 2002; Crovisier et al.,2003]. The pH also rises during basalt glass alteration atmoderately acidic pHs (>2–3) and low water-to-rock ratios(<10; Tosca et al. [2004]; Hurowitz et al. [2006]). Thus, anopen system with constantly renewed acidic fluid, a startingfluid with a very low pH (<2–3) and/or a moderately acidicpH with a high water-to-rock ratio (>100) appears necessaryto preserve SiO2-enriched leached layers.[44] A system with consistently low pH can also explain

loss of Al from the leached layer. A pH of <3 facilitates lossof Al from the glass and retention in the altering fluidbecause Al forms soluble ions at low pH [Paul and Zaman,1978; Jercinovic et al., 1990]. Once Al is lost from thebasalt, a number of factors could prevent its precipitation.Maintenance of low pH (<5) keeps Al in solution andprevents formation of hydrated or hydroxylated Al-bearing

phases [Stumm and Morgan, 1996; Drever, 1997; Hurowitzet al., 2006]. High flow rate, consistent with an opensystem, has been shown to lead to a zero precipitation/release ratio for Al and nonzero precipitation/release ratiosfor other cations, particularly Si [Daux et al., 1997].[45] The balance between the acidic and oxidizing con-

ditions of Hawaiian alteration dictates the behavior of Feduring leaching. Very acidic conditions (pH < 3; Paul andZaman [1978]; Hurowitz et al., 2006]) place Fe in solution,slow its oxidation (pH < �4; Burns [1993a, 1993b]; Stummand Morgan [1996]; Hurowitz et al. [2006]) and, in turn, itsprecipitation. However, slightly less acidic (pH > 3; Burns[1993a, 1993b]; Thorseth et al. [1991]; Eick et al. [1996];Morris et al. [2000]; Crovisier et al. [2003]; Tosca et al.[2004]; Hurowitz et al. [2006]) and oxidizing conditionslead to the retention (or precipitation) of Fe3+ duringleaching. Thus, to create Fe-rich material, conditions mustbe either oxidizing and just acidic enough (pH � 3–4) toretain Fe in the glass or sufficiently acidic (pH < 3) toremove Fe from the glass in preparation for its laterprecipitation into Fe-rich material. In the latter case, a slightincrease in pH would facilitate Fe oxidation and precipita-tion as is observed in multiple studies (pH > 3, Thorseth etal. [1991]; pH 7, Eick et al. [1996]; pH �5.5, Crovisier etal. [2003]; pH 4, Tosca et al. [2004]). Reduction of water/rock of the system, closing of the system and/or reduction inthe system flow rate could facilitate an increase in theoriginal fluid pH as cations are liberated from the glass[El-Shamy et al., 1972; Crovisier et al., 1987; Thorseth etal., 1991; Daux et al., 1994; Stroncik and Schmincke, 2002;Crovisier et al., 2003]. An increase in pH that encouragesFe oxidation and precipitation must be small enough toprevent Al precipitation and SiO2 dissolution. The precip-itated phases are typically amorphous or poorly crystallineferric oxides or ferric oxyhydroxides [e.g., Morris et al.,2000; Crovisier et al., 2003; Tosca et al., 2004].[46] Hawaiian alteration conditions can lead to a variety

of Ti behaviors during leaching. Morris explained Ti enrich-ments in altered volcanic material as the result of Tiretention during steam vent alteration governed by acidic(pH > 3) and oxidizing conditions. Tosca et al. [2004] andEick et al. [1996] observed enrichments in Ti due to Tiretention in leached layers formed at pH � 2–3 and pH 3and 5, respectively. Alternatively, Ti-bearing secondaryproducts found as a result of glass alteration [e.g., Furnesand E1-Anbaawy, 1980; Golden et al., 1996] demonstratethat Ti is not totally immobile during leaching at acidicconditions [Furnes, 1984]. Examples of secondary Tiphases are an amorphous TiO2-rich phase [Furnes andE1-Anbaawy, 1980] or Ti-bearing jarosite [Golden et al.,1996].

5.4. Dissolution and Precipitation

[47] Rather, than trying to separate all cations from Sithrough leaching, a second alteration pathway liberates allelements through total glass dissolution and separates theelements through precipitation. Dissolution-precipitation iscommonly cited as the process behind palagonite formation[Furnes, 1984; Hay and Iijima, 1968; Jercinovic et al.,1990; Daux et al., 1994; Crovisier et al., 2003; Stroncik andSchmincke, 2002]. Dissolution-precipitation must be con-sidered as a coating formation pathway because in studies


13 of 24

E05015

that produced SiO2-enriched leached layers, the thicknessesof the leached layers are much less (50–200 nm, Eick et al.[1996]; 1–2 mm, Berger et al. [1987]; 2–4 mm, Thorseth etal. [1991]; 5 mm, Guy and Schott [1989]; 10 mm, Tosca etal. [2004]) than the coating thicknesses of most of thesamples of this study. Additionally, experimentally pro-duced leached layers are very rarely pure SiO2. Theyconsistently contain Al in significantly greater amounts thanthose measured in our coatings (Table 2). As stated above,there is evidence for nonstoichiometric glass dissolutionwith liberation of network modifying [El-Shamy et al.,1972; Bunker et al., 1983; Berger et al., 1987; Guy andSchott, 1989; Thorseth et al., 1991] and forming [El-Shamyet al., 1972; Furnes and E1-Anbaawy, 1980; Jercinovic etal., 1990; Tosca et al., 2004] cations at acidic conditions.Glass dissolution could also be stoichiometric if alterationtook place in a closed system or in a system with asufficiently low flow rate to permit an increase in theoriginal fluid pH [El-Shamy et al., 1972; Crovisier et al.,1987; Crovisier et al., 2003; Stroncik and Schmincke, 2002;Thorseth et al., 1991; Daux et al., 1994]. Once all cationsare in solution, complementary conditions exist that permitSi, Fe, and Ti precipitation. Tosca et al. [2004] found thatamorphous SiO2 precipitated from altering fluids of pH 0–4,particularly from fluids with pH < 2. Curtiss et al. [1985]modeled the evaporation of acidic fluids (pH = 3.6–5.5)after their interaction with basalt glass and tephra and con-sistently found amorphous SiO2 as a precipitate. Gislasonand Eugster [1987] found that chalcedony-like SiO2 pre-cipitated from higher pH fluids (pH � 7–9), but at highertemperatures (45–65�C) than are likely relevant to thecoatings of this study. In general, the prevalence of precip-itation of secondary SiO2 liberated from the alteration ofbasaltic rocks is described in detail byMcLennan [2003]. Feand Ti precipitation most readily occur at pH >3 [Paul andZaman, 1978; Thorseth et al., 1991; Morris et al., 2000]which permits incorporation of Fe and Ti-rich materials intoor on coprecipitated SiO2. Si, Fe, and Ti precipitation canoccur without Al precipitation if pH remains below �5, asstabilization of Al-bearing hydroxides and phyllosilicates isgenerally suppressed at pH < 5 [Stumm and Morgan, 1996;Drever, 1997; Hurowitz et al., 2006]. Separation of Al fromFe and Ti is also facilitated by elevated flow rate [Dauxet al., 1997].[48] Dissolution-precipitation is often cited as the mecha-

nism behind coating or palagonite formation because of asharp chemical boundary between the alteration product andthe substrate glass [Hay and Iijima, 1968; Hay and Jones,1972; Allen et al., 1981; Farr and Adams, 1984; Furnes,1984; Jercinovic et al., 1990; Crovisier et al., 2003].Microanalytical studies of leached layer-substrate glassboundaries, however, have revealed that the leaching pro-cess does not necessarily leave behind a chemical gradientin the substrate glass that can be detected by SEM or theelectron probe. The chemical gradients associated withlaboratory and natural leaching of ferromagnesian silicatesand glasses with <60 wt.% SiO2 have been observed to existon the nanometer-scale [Smets and Lommen, 1982; Schottand Petit, 1987; Mogk and Locke, 1988; Petit et al., 1990;Eick et al., 1996]. Thus, a lack of chemical gradient in thesubstrate glass below a coating is not diagnostic of theprocess that led to the coating formation.

5.5. Coating Morphology and Chemistry

[49] The morphology and chemistry (Table 4) of thecoatings on the five Hawaiian samples of this study canbe explained by the action of one or both of the abovemechanisms. The leaching mechanism is capable of pro-ducing the marble-textured coatings of MUO and MIY,given their comparable thicknesses (3–10 mm) to leachedlayers observed in other studies [Guy and Schott, 1989;Thorseth et al., 1991; Tosca et al., 2004]. The porosities ofthe MUO and MIY coatings are also consistent withsignificant removal of cations from the original glass [e.g.,Thorseth et al., 1991]. Formation of the coatings via thedissolution/precipitation mechanism cannot be ruled out,however. Whether leaching or total glass dissolution wasresponsible for coating formation, the pH of the system hadto be <3 to hinder retention (or precipitation) of Fe and Ti(and Al) throughout the bulk of the coatings. Thus, the Fe-and Ti-bearing capping layers on MUO and MIY mostlikely result from Fe and Ti precipitation caused by a slightincrease in system pH (3–4; Burns [1987]; Thorseth et al.[1991]; Tosca et al. [2004]; Hurowitz et al. [2006]).[50] If the leaching mechanism were responsible for the

formation of the marble-textured coatings on MIO, KW, andKWY, then sustained, low pH conditions were required toform the thick, nearly pure, hydrous SiO2 coatings. Thiswould imply a unique environment for formation of thecoatings as the literature offers many more examples of thickcoatings formed via the dissolution/precipitation mechanism[Hay and Iijima, 1968; Furnes, 1984; Jercinovic et al., 1990;Daux et al., 1994; Stroncik and Schmincke, 2002; Crovisieret al., 2003]. No matter which mechanism was responsiblefor coating formation, detailed chemistry of the predomi-nantly SiO2 matrix of the coatings and their Fe- and Ti-enriched layers offer insight into chemistry of alteration. Thedearth of Fe and Ti (and Al) throughout the bulk of themarble-textured coating on KW, like that of MIY (Tables 2and 4), suggests that conditions were very acidic (pH < 3)during coating formation. The presence of Fe and Ti in theSiO2 matrix of the marble-textured coatings of MIO andKWY (Tables 2 and 4), however, suggests slightly less acidic(pH > 3) conditions supported the retention or precipitationof these elements in the SiO2 matrix of the coatings. The lackof Al in both MIO and KWY suggest that pH remainedbelow 5 and/or other alteration factors (for example, flowrate) played a role in preventing Al retention or precipitation.The Fe- and Ti-rich layers in or on the coatings (Table 4; Fe-rich layers in MIOMorphologies 1 and 2, the Fe- and Ti-richlayer on MIO Morphology 4, Ti-rich capping layers on KWand KWY) are likely the result of precipitation, but the causeof the varying abundances of Fe and Ti in the layers is notwell understood. We assume that the balance of Fe and Tiprecipitation was controlled by a variety of factors (pH,redox conditions, fluid/rock, flow rate, etc.). Spectral data(see section 5.6 below) offer insight into hosts of Fe and Ti inall the coatings.[51] To explain the mineral-bearing coatings of MIO

(Morphology 3; Figure 3c), KWY (Figure 11b), and KW(Figure 13b), the origin of the minerals and SiO2 encasingthem must be explained. The minerals, which includepyroxene, feldspar and Fe-Ti oxides, most likely originatefrom tephra [Hay and Jones, 1972; Farr and Adams, 1984;Curtiss et al., 1985] deposited on the rocks by eruption


14 of 24

E05015

or wind. The proximity of all sample locations to theHalemaumau vent provides an abundant source of ash ortephra. SiO2 is derived from leaching and/or dissolution ofeither the basalt substrate, as outlined above, or of the tephraitself. SiO2 release during pyroxene and feldspar dissolutionat low pH has been observed in multiple studies [Siegel andPfannkuch, 1984; Casey et al., 1988; Rose, 1991; Knauss etal., 1993; Stillings and Brantley, 1995; Tosca et al., 2004;Hurowitz et al., 2005]. Derivation of SiO2 from tephraalteration and subsequent formation of coatings consistingof the SiO2 and the remnant minerals has been described byHay and Jones [1972], Farr and Adams [1984] and Curtisset al. [1985]. The brown MIO coating (Morphology 3;Figure 3c) appears to capture an early stage of tephra entrap-ment by SiO2. The KWYand KW coatings (Figures 11b and13b) likely represent a later stage of tephra-to-coating transi-tion as they are primarily SiO2 with few, still-unalteredminerals remaining in the SiO2 matrix. An important charac-teristic of these coatings is that they consistently rest onmarble-textured coatings, suggesting that tephra depositionand incorporation into a coating occurred after the process thatled to the marble-textured coatings. The stratigraphy of themarble-textured and mineral-bearing coatings echoes the mor-phology observed by Tosca et al. [2004] with amorphous SiO2

precipitated as a secondary phase over SiO2-rich leachedlayers.[52] The mineral-bearing coatings of KW and KWY also

contain S, which most likely originates from the incorpora-tion of S-bearing volcanic gases into the water altering thebasalts. The Halemaumau vent, in addition to serving as asource of tephra, is a feasible source for such volcanic gases.S from volcanic exhalations dissolves into aqueous fluids andtypically becomes incorporated into alteration products assecondary sulfates [e.g., Burns, 1987; Banin et al., 1997;Tosca et al., 2004]. The VNIR spectral data described belowprovide some insight into the nature of the S in the mineral-bearing KW and KWY coatings. The presence of S in thealtering fluid is consistent with the acidic conditions impli-cated in coating formation for all the samples of this study.However, the lack of S in the coatings of MIO, MIY, orMUO or in the marble-textured coatings of KWY or KWsuggests that different conditions governed formation of themarble-textured and mineral-bearing coatings. Differentformation pathways for the marble-textured and mineral-bearing coatings are suggested by the fact that the lattercontains mineral fragments and that the latter appears tohave consistently formed later than the former. It is possiblethat the two coating types formed from separate fluids orfrom a single fluid whose character (composition, acidity,concentration, etc.) evolved through interaction with thesubstrate basalt and/or the marble-textured coatings to yielddiffering behaviors of S. Given the proximity of the KWandKWY sites to an actively degassing vent, perhaps themineral-bearing coatings did not form via aqueous alterationat all, but from vapor deposition. Deposition of Si and S[Stearns and Vaksvik, 1935; Macdonald, 1944] from volca-nic gases has been observed near active vents. Ash or tephramust have been entrapped by the Si vapors to explain theminerals in the KW and KWY coatings. Alternatively, iffluid conditions were such that S did precipitate during theformation of the marble-textured coatings, post-precipitationremoval of the S-bearing phase is required. For example,

removal of sulfate through a wetting event was suggestedby Curtiss et al. [1985] to explain the lack of secondarysulfates in their SiO2-rich coatings.

5.6. Visible/Near Infrared

[53] The absorptions present in the interior spectra of thefive samples are consistent with the presence of basaltic glass,augite, olivine, plagioclase, and chromite in the samples.Broad absorptions in the MIO and MIY spectra (Figures 4and 8) centered near 1.0 and 2.3 mm likely result from Fe2+

electronic excitations in augite. The broad absorption near1.3 mm likely results from olivine, which may also influencethe 1.0-mm band character. The broad absorptions in theMUO, KWY, and KW spectra (Figures 10, 12, and 14)centered near 1.1 and 1.9 mm are likely due to Fe2+ electronicexcitations in the glassy matrix of the basalts, as glass-richbasalts analyzed by Minitti et al. [2002] exhibited absorptionsat similar wavelengths. It is also possible that augite andolivine contribute to the 1.1-mm absorption and that augitecontributes to the 1.9 mm absorption. The shape of all theinterior spectra at <0.7 mm is likely due to the combinedsignature of the basaltic glass and chromite along with thecoarse ‘‘particle size’’ (i.e., bulk, natural-surface samples) ofthe analyzed samples (Figures 4, 8, 10, 12, and 14). The glass-rich basalts analyzed by Minitti et al. [2002] were of coarseparticle size (75–500 mm) and exhibited similar spectralstructure at <0.7 mm. Differences between the interior spectraof each sample (Figures 4, 8, 10, 12, and 14) can be explainedeither by variations in the precise proportions of phases (glass,plagioclase, pyroxene, and oxides) in each sample or byvariations in the proportions of phases present in the analysisspot on each sample.[54] The overall shapes of all the colored coating spectra

are consistent with the spectral character of H2O-bearing(opaline) SiO2, the dominant constituent of the coatings.Opaline SiO2 is capable of accounting for the absorptions at<0.7 mm, the reflectance maxima near 0.7 mm, the negativespectral slopes, and the absorptions near 1.9 mm apparent inmost of the colored coating spectra (Figure 15). OpalineSiO2 would also account for the minor absorption in theMIO and KWY spectra near 1.45 and the minor absorptionin the MIO spectrum near 2.3 mm (Figures 4 and 12). Thebands near 1.45, 1.93, and 2.3 mm are due to overtones andcombinations of H2O or OH vibrations in the opaline SiO2.[55] Other spectral features in the colored coating spectra

appear to result from the Fe-, Ti-, and S-bearing componentsin the coatings. The bands near 0.48, 0.66, and 0.85 mm inthe MIO orange coating spectrum can be accounted for bythe presence of ferric oxides/oxyhydroxides such as hema-tite, goethite, or lepidocrocite (Figure 4; e.g., Burns [1993a,1993b]; Morris et al. [1985]; Bishop et al. [1993]). Ferricoxides/oxyhydroxides would also contribute to the absorp-tion edge at <0.76 mm and the negative spectral slopeexhibited by the colored MIO coating. Ferric oxyhydroxidescould also contribute to the water-related absorption at1.9 mm. Some ferric oxides/oxyhydroxides are consistentwith the orange color of the MIO colored coating and withthe SEM and electron probe data that demonstrate theenrichment of Fe in the bright layers interspersed in thelargely opaline coating. Because the features associatedwith the ferric oxides/oxyhydroxides are weak, the phasesare probably nanophase or minor components of the


15 of 24

E05015

coating. Poorly crystalline ferric oxides/oxyhydroxides areconsistent with the type of phases precipitated out ofleaching fluids [e.g., Tosca et al., 2004]. The broad bandnear 1 mm (Figure 4) is due to Fe2+, possibly from thepyroxene and/or olivine present in the mineral-bearingbrown coating (Morphology 3, Figure 3c). The minorabsorptions near 1.25 and 2.3 mm (Figure 4) are not readilyassigned to a Fe- or Ti-rich phase that might be present inthe coatings but it is possible that they also result fromminerals in the brown coating.[56] The spectra of the colored coatings of MIY and

MUO are similar and exhibit fewer features than the coloredcoating of MIO, making it difficult to determine the phasesthat influence the coatings (Figures 8 and 10). The onlyfeature in both spectra not accounted for by opaline SiO2 isthe minor absorption near 0.48 mm (Figures 8 and 10). Thepresence of Fe3+ in many minerals leads to an absorption atthis wavelength depending on the relative strength of thisabsorption feature compared with the nearby ferric bands[Burns, 1993a, 1993b]. The Fe3+-bearing phases possiblyresponsible for the MIO colored coating absorptions(hematite, goethite, or lepidocrocite) are also capable ofyielding the 0.48-mm absorption [Morris and Lauer, 1990;Burns, 1993a, 1993b]. Either Fe-oxyhydroxide or hematitealso accounts for the particular character of the absorption atwavelengths shorter that 0.7 mm. Lepidocrocite and goethiteare consistent with the orange-yellow to orange color ofthe MIY and MUO coatings and the presence of Fe in thecapping layer of the marble-textured coatings.[57] It is not clear if the signature of the Ti in the capping

layers of MIOMorphology 4, MIYand MUO is expressed inthe colored coating spectra. If Ti is present in an amorphousphase [e.g., Furnes and E1-Anbaawy, 1980], its spectralcharacter may depend on its oxidation state. The spectra ofsilicate glasses with strictly Ti4+ (and no Fe) exhibit a sharpabsorption edge near 0.35 mm but no crystal field bandsbecause Ti4+ lacks the necessary transition electrons [Bell etal., 1976]. Spectra of glasses with a mixture of Ti3+ and Ti4+

have the same 0.35 mm absorption edge and also have acrystal field band near 0.5 mm [Bell et al., 1976]. This0.5-mm band shifts to shorter wavelengths with increasingdegree of oxidation of the glass. A band near 0.49 mm is alsoobserved in the spectra of Ti- and Fe2+-bearing pyroxenesfrom an intervalence charge transfer transition between Ti4+

and Fe2+ [Burns, 1993a, 1993b]. It is possible, then, that anamorphous Ti-bearing phase in the capping layers of theMIO, MIY, and MUO coatings contributes to their observed0.48 mm absorptions. If Ti were present in a mineral such asone of the TiO2 polymorphs (anatase, brookite, and rutile),few of their spectral features would be distinguishable fromthe signatures of opaline SiO2 and/or Fe-bearing phases. TheTiO2 polymorphs exhibit a narrow, shallow absorption near0.34 mm and anatase exhibits a steep absorption edge at�0.39 mm. None of these features is apparent in the MIO,MIY, or MUO coating spectra suggesting that TiO2 is presentin an amorphous phase and/or that the Ti host is notspectrally discernable from the other coating materials.[58] The spectra of the colored coatings of KW and KWY

exhibit complex behavior that can be explained by a numberof factors. A first order observation of both the KW andKWY spectra is that their absorption edges at <0.7 mmoccur at shorter wavelengths than those of the MIO, MIY,

and MUO coatings (Figure 15). This observation suggeststhat Fe has less control over the KW and KWY spectra thanit does over those of MIO, MIY, and MUO. The shorterwavelength absorption edges in the KW and KWY spectracould be due to the presence of S, as the spectrum of pure Shas an absorption edge near 0.40 mm. The absorption edgescould be due to Ti in an amorphous phase or in anatase(sharp absorption edges at 0.35 mm [Bell et al., 1976]) giventhat the capped coatings of KW and KWY have the highestconcentrations of Ti (Table 3). Anatase has been identifiedas a phase in Hawaiian coatings collected in the same areaas those of this study [Deal et al., 2003; Chemtob et al.,2006]; however, that anatase is Fe3+-bearing and proposedto have resulted from vapor deposition [Chemtob et al.,2006]. The KWand KWYabsorption edges could simply beexplained by the spectral dominance of the opaline compo-nent of the coatings relative to the spectral signatures of theminor elements. Opal exhibits a broad absorption edgebetween 0.3 and 0.7 mm and is consistent with the H2O-and OH-related feature at 1.95 mm; we assume that thespectra of the opaline SiO2 in the coatings would alsoexhibit these features.[59] A second order observation of the KW and KWY

spectra is that the KWY spectra have absorptions near0.48 mm whereas the KW spectrum does not (Figures 12and 14). Transitions either in Fe3+ or between Fe2+ andTi4+ can account for the 0.48-mm absorption [Bell et al.,1976; Burns, 1993a, 1993b] in the KWY spectra becausethe KWY coatings contain Fe and Ti. However, the KWcoatings have Fe and Ti, as well. It is possible that theKW spectrum lacks the 0.48-mm feature because the Sand/or opaline components of the coating dominate thespectral signature of the coating thereby masking the0.48-mm feature.[60] The inconsistency among the KW and KWY data

illustrates the complexity of linking the macroscopic,microscopic, and spectral observations from the KW andKWY coatings. The spectral characters of the white andyellow KWY coatings are effectively identical despite theirassociation with two different coating types [white coating:Ti-capped marble-textured coating (Figure 11a); yellowcoating: mineral- and S-bearing coating (Figure 11b)].Similarly for KW, despite identifying two different coatingtypes at the macroscopic and microscopic levels [Ti-capped, marble-textured (Figure 13a); mineral- and S-bearing (Figure 13b)], only one spectral signature wasdetected from the coating. For KW, it is feasible that thespectrum from one of the coating types was not obtainedbecause of the uniform color of the coating, but themicroscopic evidence from KWY suggests that differentcoating spectra should have been identified on KWYbased on color. Perhaps both the S-bearing and marble-textured coatings on KWY have the same signature andthat common components in the coatings (opaline SiO2,Fe, and Ti) dominate the spectral signature. Alternatively,the KWY spectra may simply suggest that the correlationof its macroscopic and microscopic coating morphologieswas incorrect or overly simplistic.

5.7. Thermal Infrared

[61] The TIR interior spectra of most of the samples ofthis study are similar and consistent with derivation from


16 of 24

E05015

the combined spectral signatures of pristine glass, pyrox-ene, olivine, and plagioclase in the basalts (Figure 6). Thebroad, squared-off feature between 1200 and 800 cm�1 andthe minor features between 550 and 350 cm�1 in the MIO,MIY, and MUO spectra resemble features exhibited by themultiple spectra of Minitti et al. [2002] from synthetic,glass-bearing basalts and of Feely and Christensen [1999]and Hamilton and Christensen [2000] of natural, crystallinebasalts. At small wave numbers (i.e., <600 cm�1), theKWY interior spectrum resembles the other interior spectra(Figure 6). Between �1200 and 800 cm�1, however, theKWY spectrum has a bifurcated structure which is consis-tent with the basaltic glass devitrification features identifiedin the spectra of natural Hawaiian basalts by Crisp et al.[1990]. Gradual ordering of basalt glass structure, whichdoes not require the action of water or another alterationagent, initially produces spectral features near 940 and1100 cm�1 [Crisp et al., 1990]; these features correlatewith the features exhibited by the KWY interior (Figure 6).It is also possible that the KWY interior spectrum resultsfrom inclusion of the signature of the coating spectrum dueto patches of the coating in vesicles exposed on theanalyzed interior face. The double feature between 1200and 800 cm�1 is observed by Crisp et al. [1990] in thespectra of basalt glasses partially coated by SiO2 coatingsand by Geotti-Bianchini et al. [1991] and Kraft et al.[2003] in the spectra of basalts coated with thin (<3 mm)opaline coatings.[62] Another notable difference between the interior

spectra is their relative spectral contrasts (Figure 6). Thephysical state of the interior of the samples appears tocontrol the spectral contrast. For example, MIY, whichhas the largest spectral contrast, is a nonporous, massivebasalt whereas MIO and KWY, which have very lowspectral contrasts, are very porous basalts.[63] The major features at 1225, 1100, 780, and 475 cm�1

in the spectra of the coated surfaces (Figure 5) also occur inthe spectra of pure SiO2 glasses or SiO2-rich glasses andcoatings [Almeida, 1992; Kamitsos et al., 1993; Bandfield etal., 2000;Wyatt et al., 2001; Kraft et al., 2003]. The �1225-and 1100-cm�1 bands represent two states of the SiO2

asymmetric stretching mode (n3), the �780-cm�1 bandresults from the symmetric stretching mode (n1), and the�475-cm�1 band represents the rocking (bending) mode[e.g., Lucovsky et al., 1986; Pai et al., 1986; Kirk, 1988;Lange, 1989]. Crisp et al. [1990] observe a 1225-cm�1

shoulder on a well-developed 1100 cm-1 feature related toSiO2 coatings partially covering natural, devitrified basalts.Thus, we conclude that the major features in the thermalinfrared spectra of coated surfaces result from the SiO2

coating and that substrate basalt does not contribute tothe coated surface spectrum, except in the case of MUO(Figure 5). This conclusion is consistent with Kraft et al.[2003] who demonstrated that uniform SiO2 coatings overbasalt as thin as 6 mm are sufficient to completely mask thebasalt signature. The coatings on the MIO, MIY, KWY, andKW samples are between 10 and 80 mm thick.[64] The lone exception, MUO, has a 2–3 mm thick

coating that exhibits a broad absorption near 875 cm�1 inaddition to the major features seen in the other SiO2 coatingspectra (Figure 5). Multiple geometries could account forthe additional MUO coating spectrum feature. Generally,

volume scattering could account for the 875 cm�1 feature asa broad interband feature, but the MUO coating spectrumbetween �2000 and 1300 cm�1 (spectrum not shown above1450 cm�1) does not exhibit evidence for volume scatter-ing. The features in the MUO spectrum also can becorrelated to features exhibited by the spectrum of adevitrified basalt glass partially coated with SiO2 [Crisp etal., 1990]. If the MUO coating spectrum results from thisgeometry, it would explain the resemblance of the spectrumto the KWY interior spectrum (Figure 6). A third possibilityis that the features in the MUO spectrum result from acombination of a pristine (nondevitrified) glass substrateunderlying a thin (<3 mm), uniformly distributed SiO2

coating as observed in the studies of Geotti-Bianchini etal. [1991] and Kraft et al. [2003]. Overall, the MUO coatedsurface spectrum results either from a combination of SiO2

coating and devitrified substrate glass or a nonlinear com-bination of SiO2 coating and pristine substrate glass.[65] There appears to be a general relationship between

coating thickness and the depth of the features at 1250 and1100 cm�1 (Figure 5). MIO has the thickest coating and thedeepest 1250 and 1100 cm�1 features whereas MIY andMUO have thin coatings and shallow features. The corre-lation between coating thickness and spectral feature depthdoes not hold for the 475 cm�1 feature as MUO, KWY, andKW have 475 cm�1 features of roughly the same depth.Also, there does not appear to be a relationship between thecolors of the coatings and the TIR spectral characterbecause the coloring-agent elements are not forming dom-inant structural bonds in the material. This observation isconsistent with the chemical and VNIR data that indicatethe coloring agents in the coatings are minor components.

6. Summary and Implications for Mars Coatings

[66] This thorough study of five Hawaiian samplesreveals the chemistry, spectral character, and formationconditions of the coatings on the samples. All of thecoatings studied are predominantly H2O-bearing (opaline)SiO2 and contain varying abundances of Fe, Ti, and/or S(Table 4). The chemistry and spectral character of thecoatings are consistent with formation by leaching (partialdissolution)-precipitation and/or dissolution-precipitationmechanisms under aqueous, acidic, and oxidizing condi-tions. Low oxide totals obtained in electron microprobeanalyses and water-related absorption bands in the VNIRspectra of the coatings establish the role of water in coatingformation. The VNIR spectra also almost all contain Fe3+-related absorptions, indicative of oxidizing conditions. Thepresence of S in a subset of the coatings and the lack of Feand Ti outside Fe- and Ti-rich layers in most of the coatingsargue that very acidic (pH < 3) conditions dominatedcoating formation. Furthermore, multiple studies ofHawaiian alteration indicate that aqueous, acidic, and oxi-dizing conditions are feasible, if not likely, for the regionwhere the coatings of this study formed [Macdonald, 1944;Hay and Jones, 1972; Farr and Adams, 1984; Curtiss et al.,1985; Morris et al., 2000]. The basalt glass substrate or, insome cases, tephra deposited on the substrate rocks were thematerials subject to leaching and/or dissolution. Coatingsexist on samples as young as �20 years old, suggesting arelatively rapid timeframe for coating formation. Similarly


17 of 24

E05015

rapid coating formation has been observed on other Ka’udesert samples (�30 years; Curtiss et al. [1985]; Chemtobet al. [2006]). The consistent relationship between thedominant coating chemistry (H2O-bearing SiO2) and thesubstrate chemistry of all five rocks (Table 1) suggests thatlargely opaline coatings could be diagnostic of basalticsubstrate rocks weathered in aqueous, acidic, and oxidizingconditions. In turn, the minor differences in coating che-mistry likely result from minor differences among thespecific formation conditions (for example, pH) of eachsample. The VNIR character of the coatings is dominatedby the minor elements in the coatings, whether theyare incorporated directly in the opaline matrix (S?) orcontained within distinct phases (Fe and Ti). The thermalinfrared character of the coatings is dominated by theopaline component of the coatings. Coating thicknesses(3–80 mm) are sufficient enough to partially or entirelyobscure the spectral character of the substrate basalt at bothVNIR and thermal infrared wavelengths.

6.1. Martian Alteration Conditions

[67] For coatings to form on Mars using pathwayssimilar to those of this study, aqueous, acidic, andoxidizing conditions are required. Sources of water foraltering Martian basaltic rocks exist in varying amountsand over varying timescales on Mars. Definitive evidenceof hydrogen, presumably in the form of water ice, in thecurrent Martian subsurface is provided by the MarsOdyssey Neutron Spectrometer (NS) [Feldman et al.,2004]. NS data reveal that the majority of Mars hashydrogen contents translating to water mass fractions ofat least 2% with significant (20–100%) quantities ofsubsurface water ice at latitudes poleward of 50� northand south. Other potential sources of water are water icedeposits in the midlatitude terrains either in the subsur-face [e.g., Mellon and Jakosky, 1995; Mustard et al.,2001] or on the surface [e.g., Christensen, 2003]. Thesedeposits, likely formed by mobilization of water frompolar water ice deposits to lower latitude terrains duringperiods of high obliquity, could lead to meters-thick waterice deposits [e.g., Jakosky and Carr, 1985; Mellon andJakosky, 1995]. It is feasible that these current or periodicwater ice sources could make water available for weath-ering basaltic rocks. Arguments for water stability underice or snow deposits are provided by Clow [1987], Hecht[2002], and Christensen [2003]. Derivation of water fromregolith-based ice deposits under current conditions iscontroversial [e.g., Mellon and Phillips, 2001] but somestudies suggest that conditions permitting the stability ofwater can exist in limited regions on the current Martiansurface [Haberle et al., 2001; Hecht, 2002]. The persis-tent presence of significant water sources poleward of 50�north and south suggested by NS data and the relativelyshort timescales over which purported midlatitude waterice deposits are replenished (105–106 years; Jakosky andCarr [1985]), suggest that even small quantities of waterfrom these water sources could be expected to lead toweathering over geologic timescales.[68] The results from MER provide strong evidence for the

action of water in the formation and alteration of lithologiesat both landing sites. Chemical and mineralogical data fromGusev Crater rocks are consistent with the creation of

alteration rinds by the interaction of small amounts of waterwith basaltic rocks [Haskin, 2005; Hurowitz et al., 2006].Water is also implicated by the presence of goethite detectedby Mossbauer in Gusev rocks because the hydroxyl OH isstructurally fundamental to goethite and goethite typicallyforms in aqueous environments on Earth [Deer et al., 1966;Cornell and Schwertmann, 1996; Morris et al., 2006].Unusual normative mineralogies, the presence of volatileelements and detection of minerals typically associated withformation in aqueous conditions are stated as evidence ofaqueous alteration of the rocks and soils of the ColumbiaHills [Ming et al., 2006]. The chemistry, mineralogy (par-ticularly sulfates and hematite), sedimentary structures anddiagenetic features of outcrops at Meridiani Planum arguefor the activity of shallow surface and subsurface water[Herkenhoff et al., 2004; Klingelhofer et al., 2004; Riederet al., 2004; Squyres et al., 2004, 2006; Clark et al.,2005; Grotzinger et al., 2005; McLennan et al., 2005].Coatings on a subset of the outcrops at Meridiani are alsohypothesized to have formed through the action of smallamounts of water [Learner et al., 2006; Squyres et al., 2006;Jolliff et al., 2006].[69] Chemical and spectral measurements of Martian

materials provide evidence for oxidizing and acidic con-ditions, perhaps pervasively, on Mars. VNIR spectra fromlight and dark regions on Mars have long been attributed tooxidized materials, from various forms of hematite in thesoils to thin Fe3+-bearing coatings on basaltic rocks[Adams and McCord, 1969; Singer and McCord, 1979;Singer, 1980; Morris et al., 1989, 2000]. The signature ofincreasing alteration in rocks from Gusev Crater is elevatedFe3+/Fetot in Mossbauer data [Morris et al., 2006]. Theexistence of acidic fluids, likely created through the inter-action of acidic volcanic volatiles and Martian water, is usedto explain the high S and Cl contents of Martian soil [Settle,1979; Clark and Van Hart, 1981; Burns, 1987; Banin et al.,1997]. The water that produced alteration rinds of rocks atGusev had to be acidic in order to account for the observedchemical signatures [Hurowitz et al., 2006]. The presence ofFe3+-bearing sulfate in a soil sample at Gusev [Ming et al.,2006; Morris et al., 2006] provides further evidence of theaction of acidic fluids in the Columbia Hills within Gusev.Identification of jarosite at Meridiani confirms that fluidswith pH < 4 [van Breemen, 1982; Burns, 1987; Burns andFisher, 1990; Morris et al., 2000] played a role in theformation of the outcrops across the landing site. OMEGA/Mars Express [Bibring and the OMEGA Team, 2004] hasalso detected sulfates in a variety of locations across Mars[Gendrin et al., 2005], but it is not clear if they requiredacidic fluids to form.

7. Coating Detectability

[70] Given the compelling similarities between Martianalteration conditions and the Hawaiian alteration conditionsthat governed the formation of the coatings of this study, itis important to assess the detectability of such coatings inMartian data sets. This study details the chemical, morpho-logical, and spectral complexity of terrestrial coatings andconstrains the ability of a variety of measurements todelineate the physical and chemical characteristics of suchcoatings. From a chemical standpoint, SiO2-dominated


18 of 24

E05015

coatings tens of micrometers in thickness should be detect-able given the depth sensitivities of both the alpha and x-raymodes of the MER Alpha Particle X-Ray Spectrometer(MER-APXS) [Rieder et al., 2003]. Such coatings wouldalso be apparent in MER-APXS data from analyses ofsurfaces before and after grinding with the Rock AbrasionTool (RAT). For the MPF-APXS instrument, Foley et al.[2003] claim that agreement between the data from thex-ray and alpha modes indicate that the analyzed rocksexhibited no significant heterogeneity in the top tens ofmicrometers. While thicker coatings would likely bedetected by the MPF-APXS and MER-APXS instruments,detection of thin (<5 mm), SiO2-dominated coatings andminor chemical differences in those coatings might bedifficult, depending on details of the depth sensitivitiesand detection limits for different elements. As shown in

this work, differences between the strictly Fe- and Ti-bearing coatings and the S-bearing coatings (Figure 15)would be apparent at VNIR wavelengths (<1 mm) availablefrom previously landed instruments (Imager for Mars Path-finder [Smith et al., 1997]; MER Pancam [Bell et al.,2004]). Data from these wavelengths could not isolate thesignature of the opaline component of the coatings, butlonger wavelength VNIR data such as those available fromcurrent orbiting VNIR instruments (OMEGA; CRISM[Murchie et al., 2007]) could establish the presence ofopaline coatings through the spectral signature of theirwater bands. Minor chemical differences of the coatingswould not be apparent in thermal infrared spectral data, butthe overall opaline nature of coatings could be isolated inthe thermal infrared. The different sensitivities of VNIR andthermal infrared data demonstrate the importance of usingspectral data from a range of wavelengths to study theMartian surface. Finally, the ability of Martian rock mea-surements to discern the weathering environments of coatedrocks depends on the aggregate detectability of coatings inthe multiple data sets discussed above. If opaline coatingswere detected overlying basaltic substrate rocks by inter-preting their chemical and/or spectral signatures, the resultswould suggest the action of water in periodic contact withbasaltic glass, perhaps over many wetting and drying cycles.If minor chemical differences of coatings could be detected,such as S in largely SiO2 coatings, it could provide thefurther constraint of involvement of volcanic gases incoating formation. The measurement of such chemistrycould in turn provide insight into the proximity and/oractivity of a volcanic center.[71] Are coatings like those of this study consistent with

current data from Mars missions? Constraints established byMichalski et al. [2005] on the spectral character of the com-ponent responsible for the high-SiO2 glass signature in theTES surface Type 2 spectrum indicate that the spectrum ofthe component must have an emissivity minimum between�1075 and 1110 cm�1. The emissivity minima of the

Figure 16. IMP spectra from rocks at the Mars Pathfinderlanding site (Maroon, Orange, Gray, Black) compared to thespectrum of the MUO coating. IMP spectra from Murchie etal. [2005].

Table 5. VNIR Spectral Parameters

535-nm BandDepth (�10�3)

754- to 864-nmSlope (�10�5)

754- to 1009-nmSlope (�10�5)

900-nm BandDepth (�10�3) 803:904-nm

CoatingsMIY �15.30 �31.15 �26.05 �7.78 1.13MIO 30.63 �32.00 �22.54 �12.20 1.08MUO �15.33 �11.59 �9.57 �2.09 1.16KWY Yellow �60.73 �16.76 �22.14 �26.56 1.05KW �8.60 �27.79 �25.25 �10.90 1.12

Gusev rocksSol 82 Mazatzal Outcrop 24.83 7.66 3.57 �13.95 0.99Sol 82 Mazatzal Brush 9.81 �0.05 �0.12 �3.53 1.03Sol 55 Humphrey Brush 6.14 �5.74 �4.33 0.73 1.13Sol 55 Humphrey Bottom 34.18 2.11 �8.19 �24.89 1.01Sol 55 Humphrey Top 9.37 �6.27 �9.03 �7.97 1.07Sol 35 Adirondack Coated 7.20 �8.14 �6.66 �11.23 1.04Sol 236 Ebenezer Outcrop 35.80 16.96 5.37 �17.65 0.98Sol 226 Clovis Brush Hole 19.89 0.06 0.14 �10.07 1.01Sol 226 Clovis Outcrop 35.06 17.69 6.18 �22.11 0.97

Meridiani rocksSol 51 Shark Tooth Outcrop 39.24 0.68 4.22 �12.42 1.01Sol 214 Escher Outcrop 18.52 �2.36 �2.71 �5.62 1.05Sol 561 Lemon Rind Outcrop 36.47 4.61 1.85 �21.71 0.99Sol 561 Lemon Rind Brush 26.27 �1.40 5.46 4.68 1.01


19 of 24

E05015

thermal infrared spectra of the coatings of this study fallbetween 1090 and 1100 cm�1 (Figure 5), making thecoatings viable candidates for the high-SiO2 component ofthe Type 2 spectrum. In particular, the spectra of MIY andMUO with minima at 1090 cm�1 provide the closest matchto the spectrum of the obsidian component (emissivityminimum at 1087 cm�1) utilized in the deconvolution ofthe Type 2 spectrum [Bandfield et al., 2000; Michalski et al.,2005].[72] In Mars Pathfinder data, the lack of significant

compositional heterogeneity in the top tens of micrometersof rock analyses [Foley et al., 2003] suggest that the thickestcoatings of this study (for example, MIO, KWY, and KW)cannot explain the MP-APXS data. An MUO-like coatingof only 3 mm thickness could have escaped detection by theMP-APXS; however, its VNIR spectral signature does notprovide a compelling match to IMP data from a variety ofrocks (Murchie et al. [2005]; Figure 16).[73] The fundamental process that led to the coatings of

this study (basalt alteration in aqueous, acidic, and oxidizing

conditions) was cited as the formation mechanism ofalteration rinds observed on Gusev rocks [Hurowitz et al.,2006]. While the basalts at Gusev are crystalline (asopposed to glassy basalts of this study), it is still instructiveto quantitatively compare these Hawaiian coatings andthose proposed on Mars [Hurowitz et al., 2006; Learner etal., 2006; Jolliff et al., 2006] using spectral parameters(band depths, ratios, and slopes) employed by Farrand etal. [2006] to study Pancam data of Gusev rocks (Table 5).NIR slopes (between 754 to 864 and 754 to 1009 nm) areindicators of the positions of a reflectance maximum and apossible long wavelength (>800 nm) absorption. NIR ratio(ratio of reflectances at 800 and 900 nm) is a measure of thepresence and/or strength of an Fe2+- or Fe3+-related absorp-tion near 900 nm. There is significant overlap betweenNIR ratios of Gusev plains rocks and rocks of this study(Figures 17a and 17b) but only MUO and KWY have754-to-864 nm NIR slopes that fall within representativestandard deviations of the Gusev rock data (Figure 17a).Specifically, MUO and KWY most closely correspond to

Figure 17. Comparison of VNIR spectral parameters of the coatings of this study and coated rocks atthe Gusev and Meridiani MER landing sites. Open circles represent rocks from the Gusev Crater plains,black triangles represent rocks from the Columbia Hills in Gusev Crater and gray squares represent rocksfrom Meridiani. Black diamonds represent the Hawaiian samples (HI coatings). (a) A 754- to 864-nmslope versus 800:900-nm ratio. Each Hawaiian sample is labeled; labels apply to Figure 17b. Error barsrepresent the average error over all measured Gusev data [Farrand et al., 2006]. (b) A 754- to 1009-nmslope versus 800:900-nm ratio. Symbols as in Figure 17a. Error bars represent the average error over allmeasured Gusev data [Farrand et al., 2006]. (c) A 535-nm band depth versus 900-nm band depth.Symbols as in Figure 17a.


20 of 24

E05015

the spectral parameters of Gusev plains rocks (Adirondack,Humphrey). Only MUO continues to be consistent withGusev plains data on plots of 754-to-1009 nm NIR slope(Figure 17b). On both NIR slope plots (Figures 17aand 17b), MUO falls at the least-altered end of the trenddelineated by Gusev data [Farrand et al., 2006]. Thereis little overlap in band ratios and slopes between theHawaiian coatings and the Meridiani rocks, a result consis-tent with the observation that leaching was likely notresponsible for the Meridiani coatings [Learner et al.,2006; Squyres et al., 2006; Jolliff et al., 2006].[74] The rocks of this study have no spectral similarities

to Mars data on a plot of band depths at 535 and 900 nm(Figure 17c). All Mars data fall at positive 535 nm banddepths whereas MUO has a negative 535 nm band depth.Only MIO has a positive 535-band depth, but its otherspectral parameters do not coincide with any of the Marsrock spectral parameters (Figure 17). The spectra them-selves reveal why the Hawaiian-rock spectral parameters donot match those of the Mars rocks. The overlap in the NIRratios of the Hawaiian and Mars rocks is largely a coinci-dence. The Mars data tend to have NIR ratios >1 because ofan absorption near 900 nm [e.g., Farrand et al., 2006]whereas the Hawaiian samples have NIR ratios >1 becauseof the coating-derived, well-developed negative slope atwavelengths >700 nm (Figure 15). The significant offsets ofmost of the Hawaiian-rock data from the Mars-rock data onthe NIR ratio plots (Figures 17a and 17b) are due also to thenegative NIR slope of the Hawaiian-sample spectra. MUOhas NIR slopes comparable to Gusev data due to its lowreflectivity and therefore shallow slope. The position of theFe3+-related absorption in the majority of the Hawaiiansample spectra (�480 nm; Figure 15) produces negative535-nm-band depths. The spectra of most of the Hawaiianrocks steeply increase in reflectance from 480 nm to theirreflectance maxima between 620 and 750 nm (Figure 15).Only MIO has a sufficiently deep feature near 480 nm todeflect its spectrum enough to produce a positive 535-nmband depth. For the Gusev rocks, Farrand et al. [2006]attributes the 535-nm feature attributed to a number offactors including degree of oxidation, the proportion offerric phases, dust and/or hematite. If the 535-nm featurein the Gusev data is due to dust on analyzed rocks, then thecoating spectra of this study would naturally lack thisfeature and overlap in 535-nm band depth would not beexpected. Similarly, the Hawaiian rocks would lack thefeature if the ferric phase responsible for absorption in thiswavelength range is present in lesser quantities than it is inthe Gusev rocks. Thus, if mismatch between the 535-nmband depths of MUO and the Gusev rocks can be explained,it is possible for the MUO coating spectrum to be compat-ible with the Gusev spectral data. Chemically, no significantdifference exists between SiO2 contents of MER-APXSanalyses pre- and postgrinding with the RAT [Gellert etal., 2004]. Furthermore, projections of chemistry on pre-and postgrinding Gusev rocks are consistent with olivinedissolution, which implies Si depletion [Hurowitz et al.,2006], not enrichment as would be expected from thepresence of a SiO2-rich coating. However, because MUO,the most thinly coated Hawaiian sample, provides the mostpromising spectral analog to the Gusev plains data, itspresence could be masked in the MER-APXS data [Rieder

et al., 2003]. Thus, the combined Pancam and MER-APXSdata suggest that if opaline coatings are present on the rocksof the Gusev plains, then they must be thin.

[75] Acknowledgments. The RELAB facility at Brown University isa multiuser, NASA-supported facility (NAG5-3871). MEM thanks Scott K.Rowland for helpful discussions regarding the eruptive history of MaunaIki. MDL expresses her gratitude to Phil Christensen for the use of hisemission spectrometer. The manuscript was significantly improved by theinput of Nicholas Tosca and Tom Farr. Funding for this research wasprovided by NASA’s Mars Global Surveyor Data Analysis Program (MDL)and NASA’s Mars Fundamental Research Program (NNG04BG63G).

ReferencesAbrams, M., E. A. Abbott, and A. B. Kahle (1991), Combined use ofvisible, reflected infrared, and thermal infrared images for mappingHawaiian lava flows, J. Geophys. Res., 96(2), 475–481.

Adams, J. B., and T. B. McCord (1969), Mars: Interpretation of spectralreflectivity or light and dark regions, J. Geophys. Res., 74(20), 4851–4856.

Allen, C. C., J. L. Gooding, M. Jercinovic, and K. Keil (1981), Alteredbasaltic glass: A terrestrial analog to the soil of Mars, Icarus, 45, 347–369.

Almeida, R. M. (1992), Detection of LO modes in glass by infrared reflec-tion spectroscopy at oblique incidence, Phys. Rev. B, 45(1), 161–170.

Aronson, J. R., A. G. Emslie, and H. G. McLinden (1966), Infrared spectrafrom particulate surfaces, Science, 152, 345–346.

Bandfield, J. L., V. E. Hamilton, and P. R. Christensen (2000), A globalview of Martian surface compositions from MGS-TES, Science, 287,1626–1630.

Banin, A., F. X. Han, I. Kan, and A. Cicelsky (1997), Acidic volatiles andthe Mars soil, J. Geophys. Res., 102, 13,341–13,356.

Bell, P. M., H. K. Mao, and R. A. Weeks (1976), Optical spectra andelectron paramagnetic resonance of lunar and synthetic glasses: A studyof the effects of controlled atmosphere, composition, and temperature,Proc. Lunar Sci. Conf., 7, 2543–2559.

Bell, J. F., III, et al. (2004), Pancam multispectral imaging results from theSpirit Rover at Gusev Crater, Science, 305, 800–806.

Berger, G., J. Schott, and M. Loubet (1987), Fundamental processes con-trolling the first stage of alteration of a basalt glass by seawater: Anexperimental study between 200 and 320�C, Earth Planet. Sci. Lett.,84(43), I-445.

Bibring, J.-P., and the OMEGA Team (2004), Eur. Space Agency Spec.Pub., 1240, 37.

Bishop, J. L., C. M. Pieters, and R. G. Burns (1993), Reflectance andMossbauer spectroscopy of ferrihydrite-montmorillonite assemblages asMars soil analog materials, Geochim. Cosmochim. Acta, 57(19), 4583–4595.

Bishop, J. L., S. L. Murchie, C. M. Pieters, and A. P. Zent (2002), A modelfor formation of dust, soil, and rock coatings on Mars: Physical andchemical process on the Martian surface, J. Geophys. Res., 107(E11),5097, doi:10.1029/2001JE001581.

Bunker, B. C., G. W. Arnold, E. K. Beauchamp, and D. E. Day (1983),Mechanisms for alkali leaching in mixed Na-K silicate glasses, J. NonCryst. Solids, 58, 295–322.

Burns, R. G. (1987), Ferric sulfates on Mars, J. Geophys. Res., 92, 570–574.

Burns, R. G. (1993a). Origin of electronic spectra of minerals in the visibleto near-infrared region, in Remote Geochemical Analysis: Elemental andMineralogical Composition, edited by C. M. Pieters and P. A. J. Englertpp. 3–29, Cambridge Univ. Press, New York.

Burns, R. G. (1993b), Rates and mechanisms of chemical weathering offerromagnesian silicate minerals on Mars, Geochim. Cosmochim. Acta,57, 4555–4574.

Burns, R. G., and D. S. Fisher (1990), Iron-sulfur mineralogy of Mars:Magmatic evolution and chemical weathering products, J. Geophys.Res., 95, 14,415–14,421.

Casey, W. H., H. R. Westrich, and G. W. Arnold (1988), Surface chemistryof labradorite feldspar reacted with aqueous solutions at pH = 2, 3, and12, Geochim. Cosmochim. Acta, 52, 2795–2807.

Chemtob, S. M., B. L. Jolliff, and R. E. Arvidson (2006), Si and Ti-richsurface coatings on Hawaiian basalt and implications for remote sensingon Mars, Proc. Lunar Planet. Sci. Conf., Abstract #1443.

Christensen, P. R. (2003), Formation of recent Martian gullies throughmelting of extensive water-rich snow deposits, Nature, 422, 45–48.

Christensen, P. R., et al. (2004), Initial results from the Mini-TES experi-ment in Gusev Crater from the Spirit Rover, Science, 305, 837–842.


21 of 24

E05015

Clark, B. C., and D. C. Van Hart (1981), The salts of Mars, Icarus, 45,370–378.

Clark, B. C., et al. (2005), Chemistry and mineralogy of outcrops atMeridiani Planum, Earth Planet. Sci. Lett., 240(1), 73–94.

Clow, G. D. (1987), Generation of liquid water on Mars through the meltingof a dusty snowpack, Icarus, 72, 95–127.

Cornell, R., and U. Schwertmann (1996), The iron oxides: Structures,properties, reactions, occurrences and uses, 573 pp., VHC, New York.

Crisp, J., A. B. Kahle, and E. A. Abbott (1990), Thermal infrared spectralcharacter of Hawaiian basaltic glasses, J. Geophys. Res., 95(B13),21,657–21,669.

Crovisier, J.-L., J. Honnorez, and J. P. Eberhart (1987), Dissolution ofbasaltic glass in seawater: Mechanism and rate, Geochim. Cosmochim.Acta, 51, 2977–2990.

Crovisier, J.-L., T. Advocat, and J.-L. Dussossoy (2003), Nature and role ofnatural alteration gels formed on the surface of ancient volcanic glasses(Natural analogs of waste containment glasses), J. Nucl. Mater., 321, 91–109.

Curtiss, B., J. B. Adams, and M. S. Ghiorso (1985), Origin, developmentand chemistry of silica-alumina rock coatings from the semi-arid regionsof the island of Hawaii, Geochim. Cosmochim. Acta, 49, 49–56.

Daux, V., J.-L. Crovisier, C. Hemond, and J. C. Petite (1994), Geochemicalevolution of basaltic rocks subjected to weathering: Fate of the majorelements, rare earth elements, and thorium, Geochim. Cosmochim. Acta,58(22), 4941–4954.

Daux, V., C. Guy, T. Advocat, J.-L. Crovisier, and P. Stille, (1997), Kineticaspects of basaltic glass dissolution at 90�C: Role of aqueous silicon andaluminum, Chem. Geol., 142, 109–126.

Deal, K. S., R. E. Arvidson, and B. L. Jolliff (2003), Remote mapping ofthe Ka’u Desert, Hawai’i: Silica in a Mars analog terrain, Proc. LunarPlanet. Sci. Conf., Abstract #1952.

Deer, W. A., R. A. Howie, J. Zussman (1966), An Introduction to the RockForming Minerals, 528 pp., John Wiley, Hoboken, N. J.

Drever, J. I. (1997), The Geochemistry of Natural Waters, 436 pp., PrenticeHall, Inc., Upper Saddle River, N. J.

Eick, M. J., P. R. Grossl, D. C. Golden, D. L. Sparks, and D. W. Ming(1996), Dissolution kinetics of a lunar glass stimulant at 25�C: The effectof pH and organic acids, Geochim. Cosmochim. Acta, 60(1), 157–170.

El-Shamy, T. M., J. Lewins, and R. W. Douglas (1972), The dependence ofthe pH on the decomposition of glasses by aqueous solutions, J. Soc.Glass Technol., 13(3), 81–87.

Farr, T. G., and J. B. Adams (1984), Rock coatings in Hawaii, Geol. Soc.Am. Bull., 95, 1077–1083.

Farrand, W. H., J. F. Bell III, J. R. Johnson, S. W. Squyres, J. Soderblom,and D. W. Ming (2006), Spectral variability among rocks in visible andnear-infrared multispectral Pancam data collected at Gusev crater:Examinations using spectral mixture analysis and related techniques,J. Geophys. Res., 111, E02S15, doi:10.1029/2005JE002495.

Feely, K. C., and P. R. Christensen (1999), Quantitative compositionalanalysis using thermal emission spectroscopy: Application to igneousand metamorphic rocks, J. Geophys. Res., 104(E10), 24,195–24,210.

Feldman, W. C., et al. (2004), Global distribution of near-surface hydrogenon Mars, J. Geophys. Res., 109, E09006, doi:10.1029/2003JE002160.

Fischer, E., and C. M. Pieters (1993), The continuum slope of Mars:Bi-directional reflectance investigations and applications to OlympusMons, Icarus, 102, 185–202.

Foley, C. N., T. Economou, and R. N. Clayton (2003), Final chemicalresults from the Mars Pathfinder alpha proton x-ray spectrometer,J. Geophys. Res., 108(E12), 8096, doi:10.1029/2002JE002019.

Furnes, H. (1984), Chemical changes during progressive subaerial palago-nitization of a subglacial tholeiite hyaloclastite: A microprobe study,Chem. Geol., 43, 271–285.

Furnes, H., and M. I. H. E1-Anbaawy (1980), Chemical changes and authi-genic mineral formation during palagonitization of a basanite hyaloclastite,Gran Canaria, Canary Islands, Neues Jahrb. Mineral Abh., 139, 279–302.

Gellert, R., et al. (2004), Chemistry of rocks and soils in Gusev Crater fromthe Alpha Particle X-ray Spectrometer, Science, 305, 829–832.

Gendrin, A., et al. (2005), Sulfates in Martian layered terrains: The OMEGA/Mars Express view, Science, 307(5715), 1587–1591.

Geotti-Bianchini, F., L. De Riu, G. Galiardi, M. Guglielmi, and C. G.Pantano (1991), New interpretation of the IR reflectance spectra ofSiO2-rich films on soda-lime glass, Glastech. Ber., 64(8), 205–217.

Gislason, S. R., and H. P. Eugster (1987), Meteoric water-basalt interac-tions. I: A laboratory study, Geochim. Cosmochim. Acta, 51, 2827–2840.

Golden, D. C., R. V. Morris, D. W. Ming, J. F. Bell III, S. V. Yang, and D. R.Thompson (1996), Occurrence of a titanium-bearing jarosite fromHawaii, Proc. Lunar Planet. Sci., XXVIIth, 427–428.

Grotzinger, J. P., et al. (2005), Stratigraphy and sedimentology of a dry towet eolian depositional system, Burns formation, Meridiani Planum,Mars, Earth Planet. Sci. Lett., 240(1), 11–72.

Guy, C., and J. Schott (1989), Multisite surface reaction versus transportcontrol during the hydrolysis of a complex oxide, in Kinetic Geo-chemistry, edited by J. Schott and A. C. Lasaga, Chem. Geol., 78,181–204.

Haberle, R. M., C. P. McKay, J. Schaeffer, N. A. Cabrol, E. A. Grin, A. P.Zent, and R. Quinn (2001), On the possibility of liquid water on present-day Mars, J. Geophys. Res., 106(E10), 23,317–23,326.

Hamilton, V. E., and P. R. Christensen (2000), Determining the modalmineralogy of mafic and ultramafic igneous rocks using thermal emissionspectroscopy, J. Geophys. Res., 105(E4), 9717–9733.

Harding, D., and J. M. Miller (1982), The influence on rain chemistry of theHawaiian volcano Kilauea, J. Geophys. Res., 87(C2), 1225–1230.

Haskin, L. A. (2005), Water alteration of rocks and soils from the SpiritRover site, Gusev Crater, Mars, Nature, 436, doi:10.1038/nature03640.

Hay, R. L., and A. Iijima (1968), Nature and origin of palagonite tuffs ofthe Honolulu Group on Oahu, Hawaii, Geol. Soc. Am. Mem., 116,338–376.

Hay, R. L., and B. F. Jones (1972), Weathering of basaltic tephra on theisland of Hawaii, Geol. Soc. Am. Bull., 83, 317–332.

Hecht, M. H. (2002), Metastability of liquid water on Mars, Icarus, 156,373–386.

Herkenhoff, K. E., et al. (2004), Evidence from Opportunity’s microscopicimager for water on Meridiani Planum, Science, 306(5702), 1727–1730.

Hunt, G. R., and R. K. Vincent (1968), The behavior of spectral features inthe infrared emission from particulate surfaces of various grain sizes,J. Geophys. Res., 73(18), 6039–6046.

Hurowitz, J. A., S. M. McLennan, D. H. Lindsley, and M. A. A. Schoonen(2005), Experimental epithermal alteration of synthetic Los Angelesmeteorite: Implications for the origin of Martian soils and identificationof hydrothermal sites on Mars, J. Geophys. Res., 110, E07002,doi:10.1029/2004JE002391.

Hurowitz, J. A., S. M. McLennan, N. J. Tosca, R. E. Arvidson, J. R.Michalski, D. W. Ming, C. Schroder, and S. W. Squyres (2006), In situand experimental evidence for acidic weathering of rocks and soils onMars, J. Geophys. Res., 111, E02S19, doi:10.1029/2005JE002515.

Jakosky, B. M., and M. H. Carr (1985), Possible precipitation of ice atlow latitudes of Mars during periods of high obliquity, Nature, 315,559–561.

Jercinovic, M. J., K. Keil, M. R. Smith, and R. A. Schmitt (1990), Altera-tion of basaltic glasses from north-central British Columbia, Canada,Geochim. Cosmochim. Acta, 54, 2679–2696.

Johnson, J. R., et al. (1999), Preliminary results on photometric propertiesof materials at the Sagan Memorial Station, Mars, J. Geophys. Res., 104,8809–8830.

Johnson, J. R., P. R. Christensen, and P. G. Lucey (2002), Dust coatings onbasaltic rocks and implications for thermal infrared spectroscopy of Mars,J. Geophys. Res., 107(E6), 5035, doi:10.1029/2000JE001405.

Jolliff, B., A. Knoll, W. Farrand, R. Sullivan, and the Athena Science Team(2006), Rock rinds at Meridiani and surface weathering phenomena, FallAGU.

Kahle, A. B., A. R. Gillespie, E. A. Abbott, M. J. Abrams, R. E. Walker,G. Hoover, and J. P. Lockwood (1988), Relative dating of Hawaiian lavaflows using multispectral thermal infrared images: A new tool for geo-logic mapping of young volcanic terranes, J. Geophys. Res., 93, 15,239–15,251.

Kamitsos, E. I., A. P. Patsis, and G. Kordas (1993), Infrared-reflectancespectra of heat-treated, sol-gel-derived silica, Phys. Rev. B, 48(17),12,499–12,505.

Kirk, C. T. (1988), Quantitative analysis of the effect of disorder-inducedmode coupling on infrared absorption in silica, Phys. Rev. B, 38(2),1255–1273.

Klingelhofer, G., et al. (2004), Jarosite and hematite at Meridiani Planumfrom Opportunity’s Mossbauer spectrometer, Science, 306(5702), 1740–1745.

Knauss, K. G., S. N. Nguyen, and H. C. Weed (1993), Diopside dissolutionkinetics as a function of pH, CO2, temperature, and time, Geochim.Cosmochim. Acta, 57, 285–294.

Kraft, M. D., J. R. Michalski, and T. G. Sharp (2003), Effects of pure silicacoatings on thermal emission spectra of basaltic rocks: Considerations forMartian surface mineralogy, Geophys. Res. Lett., 30(24), 2288,doi:10.1029/2003GL018848.

Lane, M. D. (1999), Midinfrared optical constants of calcite and theirrelationship to particle size effects in thermal emission spectra of granularcalcite, J. Geophys. Res., 104, 14,099–14,108.

Lange, P. (1989), Evidence for disorder-induced vibrational mode couplingin thin amorphous SiO2 films, J. Appl. Phys., 66(1), 201–204.

Learner, Z. A., et al. (2006), Surface coatings at Meridiani Planum, Mars,Proc. Lunar Planet. Sci. Conf., Abstract #2040.

Lucovsky, G., P. D. Richard, D. V. Tsu, S. Y. Lin, and R. J. Markunas(1986), Deposition of silicon dioxide and silicon nitride by remote


22 of 24

E05015

plasma enhanced chemical vapor deposition, J. Vac. Sci. Technol. A Vac.Surf. Films, 4(3), 681–688.

Lyon, R. J. P. (1964), Evaluation of infrared spectrophotometry for compo-sitional analysis of lunar and planetary soils: II. Rough and powderedsurfaces, NASA Conf. Rep., CR-100.

Macdonald, G. A. (1944), Solfataric alteration of rocks at Kilauea volcano,Am. J. Sci., 242(9), 496–505.

McLennan, S. M. (2003), Sedimentary silica on Mars, Geology, 31(4),315–318.

McLennan, S. M., et al. (2005), Provenance and diagenesis of the evaporite-bearing Burns formation, Meridiani Planum, Mars, Earth Planet. Sci. Lett.,240(1), 95–121.

McSween, H. Y., Jr., et al. (1999), Chemical, multispectral, and texturalconstraints on the composition and origin of rocks at the Mars Pathfinderlanding site, J. Geophys. Res., 104, 8679–8715.

McSween, H. Y., Jr., et al. (2004), Basaltic rocks analyzed by the SpiritRover in Gusev Crater, Science, 305, 842–845.

Mellon, M. T., and B. M. Jakosky (1995), The distribution and behavior ofMartian ground ice during past and present epochs, J. Geophys. Res.,100(E6), 11,781–11,799.

Mellon, M. T., and R. J. Phillips (2001), Recent gullies on Mars and thesource of liquid water, J. Geophys. Res., 106, 23,165–23,180.

Michalski, J. R., M. D. Kraft, T. G. Sharp, L. B. Williams, and P. R.Christensen (2005), Mineralogical constraints on the high-silica Martiansurface component observed by TES, Icarus, 174(1), 161–177.

Ming, D. W., et al. (2006), Geochemical and mineralogical indicators foraqueous processes in the Columbia Hills of Gusev crater, Mars, J. Geo-phys. Res., 111, E02S12, doi:10.1029/2005JE002560.

Minitti, M. E., J. F. Mustard, and M. J. Rutherford (2002), Effects of glasscontent and oxidation on the spectra of SNC-like basalts: Applications toMars remote sensing, J. Geophys. Res., 107(E5), 5030, doi:10.1029/2001JE001518.

Moersch, J. E., and P. R. Christensen (1995), Thermal emission from par-ticulate surfaces: A comparison of scattering models with measured spec-tra, J. Geophys. Res., 100, 7465–7477.

Mogk, D. W., and W. W. Locke III (1988), Application of Auger ElectronSpectroscopy (AES) to naturally weathered hornblende, Geochim. Cos-mochim. Acta, 52, 2537–2542.

Morris, R. V., and H. V. Lauer Jr. (1990), Matrix effects for reflectivityspectra of dispersed nanophase (superparamagnetic) hematite with appli-cation to Martian spectral data, J. Geophys. Res., 95, 5101–5109.

Morris, R. V., H. V. Lauer Jr., C. A. Lawson, E. K. Gibson Jr., G. A. Nace,and C. Stewart (1985), Spectral and other physicochemical properties ofsubmicron powders of hematite (a-Fe2O3), maghemite (g-Fe2O3), mag-netite (Fe3O4), goethite (a-FeOOH), and lepidocrocite (g-FeOOH),J. Geophys. Res., 90, 3126–3144.

Morris, R. V., D. G. Agresti, H. V. Lauer Jr., J. A. Newcomb, T. D. Shelfer,and A. V. Murali (1989), Evidence for pigmentary hematite on Marsbased on optical, magnetic, and Mossbauer studies of superparamagnetic(nanocrystalline) hematite, J. Geophys. Res., 94, 2760–2778.

Morris, R. V., D. W. Ming, D. C. Golden, and J. F. Bell III (1996), Anoccurrence of jarositic tephra on Mauna Kea, Hawaii: Implications for theferric mineralogy of the Martian surface, in Mineral Spectroscopy: ATribute to Roger G. Burns, edited by M. D. Dyar, C. McCammon andM. W. Shaefer, Spec. Publ. Geochem. Soc., 5, 327–336.

Morris, R. V., et al. (2000), Mineralogy, composition, and alteration of MarsPathfinder rocks and soils: Evidence from multispectral, elemental, andmagnetic data on terrestrial analogue, SNC meteorite, and Pathfindersamples, J. Geophys. Res., 105, 1757–1817.

Morris, R. V., T. G. Graff, S. A. Mertzman, M. D. Lane, and P. R.Christensen (2003), Palagonitic Mars from rock rinds to dust: Evidencefrom visible, near-IR, and thermal emission spectra of poorly crystallinematerials, Proc. Lunar Planet. Sci. Conf. 34th, abstract 1874.

Morris, R. V., et al. (2006), Mossbauer mineralogy of rock, soil, and dust atGusev crater, Mars: Spirit’s journey through weakly altered olivine basalton the plains and pervasively altered basalt in the Columbia Hills,J. Geophys. Res., 111, E02S13, doi:10.1029/2005JE002584.

Murchie, S., O. Barnouin-Jha, K. Barnouin-Jha, J. L. Bishop, R. V. Morris,J. Johnson, and H. Y. McSween Jr. (2005), New insights into the geologyof the Mars Pathfinder landing site from spectral and morphologic anal-ysis of the 12-color SuperPan Panorama, Icarus, in revision.

Murchie, S. L., et al. (2007), Compact Reconnaissance Imaging Spec-trometer for Mars (CRISM) on Mars Reconnaissance Orbiter (MRO),J. Geophys. Res., doi:10.1029/2006JE002682, in press.

Mustard, J. F., and J. E. Hays (1997), Effects of hyperfine particles onreflectance spectra from 0.3 to 25 mm, Icarus, 125, 145–163.

Mustard, J. F., C. D. Cooper, and M. K. Rifkin (2001), Evidence for recentclimate change on Mars from the identification of youthful near-surfaceground ice, Nature, 412, 411–414.

Oelkers, E. H., and S. R. Gislason (2001), The mechanism, rates and con-sequences of basaltic glass dissolution: I. An experimental study of thedissolution rates of basaltic glass as a function of aqueous Al, Si andoxalic acid concentration at 25 C and pH = 3 and 11, Geochim. Cosmo-chim. Acta, 65(21), 3671–3681.

Pai, P. G., S. S. Chao, Y. Takagi, and G. Lucovsky (1986), Infrared spectro-scopic study of SiOx films produced by plasma enhanced chemical vapordeposition, J. Vac. Sci. Technol. A Vac. Surf. Films, 4(3), 689–694.

Petit, J.-C., J.-C. Dran, and G. Della Mea (1990), Energetic ion-beamanalysis in the earth sciences, Nature, 344(6267), 621–626.

Pieters, C. M. (1983), Strength of mineral absorption features on thetransmitted component of near-infrared reflected light: First results fromRELAB, J. Geophys. Res., 88, 9534–9544.

Paul, A., and M. S. Zaman (1978), The relative influences of Al2O3 andFe2O3 on the chemical durability of silicate glasses at different pH values,J. Mater. Sci., 13, 1499–1502.

Rieder, R., R. Gellert, J. Bruckner, G. Klingelhofer, G. Dreibus, A. Yen, andS. W. Squyres (2003), The new Athena alpha particle X-ray spectrometerfor the Mars Exploration Rovers, J. Geophys. Res., 108(E12), 8066,doi:10.1029/2003JE002150.

Rieder, R., et al. (2004), Chemistry of rocks and soils at Meridiani Planumfrom the alpha particle X-ray spectrometer, Science, 306(5702), 1746–1749.

Rose, N. M. (1991), Dissolution rates of prehnite, epidote, and albite,Geochim. Cosmochim. Acta, 55, 3273–3286.

Salisbury, J. W., and A. E. Wald (1992), The role of volume scattering inreducing spectral contrast of reststrahlen bands in spectra of powderedminerals, Proc. Lunar Planet. Sci. Conf. 23rd, 1205–1206.

Schiffman, P., R. Zierenberg, N. Marks, J. L. Bishop, and M. D. Dyar(2007), Acid fog deposition at Kilauea volcano: A possible mechanismfor the formation of siliceous-sulfate rock coatings on Mars, Geology, (inpress).

Scholze, H. (1988), Glass-water interactions, J. Non Cryst. Solids, 102,1–10.

Schott, J., and J.-C. Petit (1987), New evidence for the mechanisms ofdissolution of silicate minerals, in Aquatic Surface Chemistry, edited byW. Stumm, pp. 293–315, Wiley-Interscience, Hoboken, N. J.

Settle, M. (1979), Formation and deposition of volcanic sulfate aerosols onMars, J. Geophys. Res., 84, 8343–8354.

Siegel, D. I., and H. O. Pfannkuch (1984), Silicate mineral dissolution atpH 4 and near standard temperature and pressure, Geochim. Cosmochim.Acta, 48, 197–201.

Singer, R. B. (1980), The composition of the Martian dark regions:Observations and analysis, Ph.D. thesis, 301 pp., Massachusetts Insti-tute of Technology, Boston.

Singer, R. B. (1982), Spectral evidence for the mineralogy of high-albedosoils and dust on Mars, J. Geophys. Res., 87(B12), 10,159–10,168.

Singer, R. B., and T. B. McCord (1979), Mars: Large scale mixing of brightand dark surface materials and implications for analysis of spectralreflectance, Proc. Lunar Planet. Sci. Conf., 10, 1835–1848.

Smets, B., and T. Lommen (1982), The leaching of sodium aluminosilicateglasses studied by secondary ion mass-spectrometry, Phys. Chem.Glasses, 23(3), 83–87.

Smith, P. H., et al. (1997), Results from the Mars Pathfinder camera,Science, 278, 1758–1765.

Squyres, S. W., et al. (2004), In situ evidence for an ancient aqueousenvironment at Meridiani Planum, Mars, Science, 306(5702), 1709–1714.

Squyres, S. W., et al. (2006), Two years at Meridiani Planum: Results fromthe Opportunity Rover, Science, 313, 1403–1407.

Stearns, H. T., and K. N. Vaksvik (1935), Geology and ground waterresources of the island of Oahu, Hawaii, Hawaii Div. of Hydrography,Bull., 1, 88.

Stillings, L. L., and S. L. Brantley (1995), Feldspar dissolution at 25�C andpH 3: Reaction stoichiometry and the effect of cations, Geochim. Cos-mochim. Acta, 59, 1483–1496.

Stroncik, N. A., and H.-U. Schmincke (2002), Palagonite—A review, Int. J.Earth Sci., 91, 680–697, doi:10.1001/s00531-001-0238-7.

Stumm, W., and J. J. Morgan (1996), Aquatic Chemistry: Chemical Equili-bria and Rates in Natural Waters, 3rd ed., 1022 pp., Wiley-Interscience,Hoboken, N. J.

Thorseth, I. H., H. Furnes, and O. Tumyr (1991), A textural and chemicalstudy of Icelandic palagonite of varied composition and its bearing on themechanism of the glass-palagonite transformation, Geochim. Cosmo-chim. Acta, 55, 731–749.

Tosca, N. J., S. M. McLennan, D. H. Lindsley, and M. A. A. Schoonen(2004), Acid-sulfate weathering of synthetic Martian basalt: The acidfog model revisited, J. Geophys. Res., 109, E05003, doi:10.1029/2003JE002218.


23 of 24

E05015

van Breemen, N. (1982), Genesis, morphology and classification of acidsulfate soils in coastal plains, in Acid Sulfate Weathering, edited by J. A.Kittrick, D. S. Fanning, and L. R. Hossner pp. 95–108, Spec. Pub., 10,Soil Sci. Soc. America, Madison, WI.

White, A. F. (1984), Weathering characteristics of natural glass and influ-ences on associated water chemistry, J. Non Cryst. Solids, 67, 225–244.

Wyatt, M. B., and H. Y. McSween Jr. (2002), Spectral evidence for weath-ered basalt as an alternative to andesite in the northern lowlands of Mars,Nature, 417, 263–266.

Wyatt, M. B., V. E. Hamilton, H. Y. McSween Jr., P. R. Christensen, andL. A. Taylor (2001), Analysis of terrestrial and Martian volcanic com-positions using thermal emission spectroscopy: I. Determination of

mineralogy, chemistry, and classification strategies, J. Geophys. Res.,106, 14,711–14,732.

��J. L. Bishop, SETI Institute/NASA-ARC, 515 N. Whisman Rd.,

Mountain View, CA 94043, USA.M. D. Lane and C. M. Weitz, Planetary Science Institute, 1700 E. Fort

Lowell Rd., Suite 106, Tucson, AZ 85719, USA.M. E. Minitti, Center for Meteorite Studies, School of Earth and Space

Exploration, Arizona State University, Box 871404, Tempe, AZ 85287-1404, USA. ([email protected])


24 of 24

E05015

morphology, chemistry, and spectral properties of hawaiian rock …scott/workshop_reading/... ·...

Documents