: simulation of space weathering by the solar wind papers/loeffler space weathering by … ·...

Irradiation of olivine by 4 keV He+: Simulation of space weathering

by the solar wind

M. J. Loeffler,1 C. A. Dukes,1 and R. A. Baragiola1

Received 8 August 2008; revised 22 December 2008; accepted 5 January 2009; published 10 March 2009.

[1] We have studied the effects of 4 keV He+ ion irradiation on olivine whilemeasuring, in situ, changes in the near-infrared (NIR) reflectance and in the chemicalcomposition of the surface. The observed changes in reflectance are reddening and theattenuation of the Fe-3d absorption bands in the NIR. Spectral reddening ofirradiated olivine powder (<45 mm) correlates with the amount of metallic iron formedby ion impact, consistent with the idea that space-weathering effects in the reflectanceof olivine-bearing S type asteroids are due to the formation of metallic iron. Themetallization rate of the powder is about half that of a sectioned rock of olivine, whichwe propose is a consequence of redeposition of sputtered material. The NIRspectral changes observed in ion irradiation experiments are similar to those observedin our previous experiments on vapor redeposition, indicating that different spaceweathering mechanisms can lead to similar final effects on reflectance. Finally, weestimate that at 1 AU the spectral reddening caused by the solar wind saturatesapproximately 2 orders of magnitude faster than comparable reddening caused bymicrometeorite impacts.

Citation: Loeffler, M. J., C. A. Dukes, and R. A. Baragiola (2009), Irradiation of olivine by 4 keV He+: Simulation of space

weathering by the solar wind, J. Geophys. Res., 114, E03003, doi:10.1029/2008JE003249.

1. Introduction

[2] Space weathering is the physical and chemical alter-ation of surfaces of airless bodies exposed to the spaceenvironment [Hapke, 2001; Clark et al., 2002; Chapman,2004]. There they are subject to bombardment by ions fromthe solar wind, magnetosphere ions, energetic electrons,cosmic rays, ultraviolet photons and fast interplanetary dust(micrometeorites), with the relative flux of each of thoseprojectiles depending on the location of the surface in thesolar system. Any of these impacts can produce chemicalalterations by breaking molecular bonds, synthesizing newmolecules or by changing the elemental composition via ionimplantation and preferential sputtering. Micrometeoritesmay, in addition, physically alter the particle size distribu-tion through impacts that melt and reform surface particles,sinter them, break them into smaller pieces, or bring largersubsurface particles to the surface, mixing the soil. Changesin grain size alter the path of the light in the solid leading tochanges in extinction and, therefore, reflectance. Many ofthese alterations caused by space weathering can be inferredfrom the spectral reflectance of the planetary body and,in some cases, seen directly on returned samples, such aslunar soil [Pillinger, 1979; Keller and McKay, 1993;Christoffersen et al., 1996] and interplanetary dust [Bradley,1994].

[3] Most laboratory simulations of space weatheringphenomena on the Moon, Mercury and asteroids havefocused on what are believed to be the two most importanttypes of weathering agents: solar wind ions and micro-meteorites. The solar wind, consisting of mostly hydrogenand helium ions (>99%), has a broad velocity distributionwith a mean value of 468 km/sec (�1 keV/amu) [Gosling,2007]. For this reason, most labratory studies have beendone with 1 keV H and 4 keV He.[4] Several keV ion irradiation experiments on minerals

have been published on the simulation of solar wind impacton mineral surfaces [Nash, 1967; Hapke, 1973; Dukes et al.,1999; Davoisne et al., 2008]. Hapke [1973] showed that H+

irradiation of loose mineral powders, analogous to theMoon regolith, visibly darkens the irradiated surface, whichhe attributed to the formation of metallic iron in thesputtered vapor and their redeposition on nearby grains.Dukes et al. [1999] showed, by in situ surface analysis, thation irradiation of olivine can reduce its Fe2+ into metalliciron, Fe0, near the surface and that the reduction is moreefficient for He+ than for H+. These experiments were doneon cleaved rock surfaces, showing that metallization did notrequire redeposition of Fe atoms on nearby grain surfaces.Other experiments with higher-energy ions (60–400 keV H,He, Ar), more characteristic of high-velocity shocks foundin the interstellar media [Bringa et al., 2007], were shown todarken and redden the reflectance spectra of mineral rocksand powders between 0.25 and 2.5 mm [Brunetto andStrazzulla, 2005; Strazzulla et al., 2005; Brunetto et al.,2006b].

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 114, E03003, doi:10.1029/2008JE003249, 2009ClickHere

for

FullArticle

1Laboratory for Atomic and Surface Physics, University of Virginia,Charlottesville, Virginia, USA.

Copyright 2009 by the American Geophysical Union.0148-0227/09/2008JE003249$09.00

E03003 1 of 13

http://dx.doi.org/10.1029/2008JE003249

[5] The correlation between optical effects and metalliciron content that was observed in lunar soil has beensimulated in the laboratory. Adding metallic iron to silicategels having featureless spectra causes the reddening ofspectral slopes and darkening in the visible-NIR [Noble etal., 2007], consistent with the effects seen on lunar rocks.[6] The second source of space weathering, micromete-

orite impact, has been simulated in the laboratory usingpulsed laser beams with instantaneous power densitiessimilar to those of typical impact by micron-sized meteor-ites [Moroz et al., 1996; Sasaki et al., 2001; Brunetto et al.,2006a; Loeffler et al., 2008a, 2008b]. Those studies showchanges in the reflectance spectra of silicates that can belarger than those seen in ion irradiation experiments. How-ever, to compare these two effects, one needs to considerrelative fluxes of the impactors. The relative role of solarwind versus micrometeorites has been a subject of contro-versy for several decades. Recently, it has been argued thatthe inverse dependence of the degree of reddening ofasteroids with their distance from the Sun implies that thesolar wind is likely the primary mechanism for spaceweathering [Marchi et al., 2006]. However, this inversedependence is tenuous because of the large scatter of thedata, and may be affected by the differences in compositionwith distance to the Sun. Thus, even if this dependence isrelated to space weathering, it does not preclude theimportance of micrometeorite impacts, as closeness to theSun not only implies larger ion fluxes but also micro-meteorites at higher fluxes and velocities [Cintala, 1992].[7] To further the understanding of the role of ion

irradiation on space weathering, we performed experimentson olivine, a common mineral found on many asteroids[Gaffey et al., 1993; Clark et al., 2002] that has been shownto be susceptible to space weathering [Moroz et al., 1996;Sasaki et al., 2001; Brunetto et al., 2006a; Loeffler et al.,2008a, 2008b]. Specifically, in these experiments, we irra-diated sectioned (‘‘flat’’) rock fragments and powders ofSan Carlos olivine with 4 keV He+ ions. A unique aspect ofexperiments is that during irradiation, we monitor the NIRspectra and the surface chemical composition of olivine insitu, e.g., without exposing the samples to the atmosphere.This ability will enable us to directly determine whetherthere is a correlation between irradiation-induced changes inthe reflectance spectra and chemical composition of olivine.

2. Experimental Methods

2.1. Sample Preparation and Analysis

[8] Olivine samples used in these experiments werenatural San Carlos olivine (Mg1.8Fe0.2SiO4) from PennMinerals (Scranton, PA, USA) and from Arizona Gemsand Crystals (Tucson, Az, USA). We also used syntheticforsterite (99.4% Mg2SiO4) from Reade Advanced Materi-als (Riverside, RI, USA) with the percent concentration ofimpurities listed as: Al2O3 (0.03), Fe2O3 (0.02), TiO2 (0.09),Na2O (0.11), CaO (0.22), and K2O (0.12). To prepareolivine powder samples for use, we separated the largegrains from rock inclusions, crushed them with agate mortarand pestle and used a sieve to obtain a grain size <45 mm.Chemical analysis, using X-ray photoelectron spectroscopy(XPS), showed no discernable difference between theuncrushed grains and the powder. To obtain the sample size

needed for XPS measurements (�6–8 mm), we pressed thepowder to �3000 psi inside an aluminum ring (innerdiameter �10 mm), placed between two stainless steeldisks. In addition, we also performed experiments on asolid piece of olivine, since we were interested in learningwhether changes induced by irradiation depend on physicalproperties of the sample, such as grain size, as has beennoted previously [Hapke, 2001, and references therein]. Toobtain flat specimen, we cut sections of olivine using adiamond blade, and removed the cutting fluid from thesurface with successive acetone, isopropanol, and HPLCwater baths.[9] The samples were mounted on a copper holder and

transferred into an oil free ion-pumped ultra-high vacuum(UHV) chamber with a base pressure between 5 and 10 �10�10 Torr (Figure 1). Surface analyses of the samples wereperformed before, during and after irradiation using XPSand near infrared reflectance (NIR) spectroscopy.[10] XPS is a widely used surface analysis method, which

gives both elemental abundance and chemical information.In this technique, X-rays eject photoelectrons from a samplesurface with a kinetic energy (Ek) given by Ek = Ex-ray �EB � f, where the X-ray energy, Ex-ray, is 1486.6 eV (Al-K),EB is the electron binding energy with respect to the Fermilevel, and 8 is the work function of the spectrometer. Theenergy distribution (spectra) of photoelectrons is given asnumber of counts versus binding energy. The spectra show acontinuous background on which photoelectron and Augerpeaks are superimposed. The escape depth of the electronsanalyzed by XPS depends on their kinetic energy; it is�5 nmfor the Fe-2p and O-1s, and �8 nm for Si-2p, and Mg-2pelectrons in olivine [Seah and Dench, 1979]. The relativearea of the photoelectron peaks is used together with sensi-tivity factors to obtain elemental composition, whereas theprecise value of the electron binding energy gives informa-tion on the chemical state of the element [Moulder et al.,1992].[11] XPS analysis was performed in a Physical Electron-

ics 560 system, equipped with a double-pass, cylindricalmirror electron energy analyzer (CMA) [Dukes et al., 1999;Loeffler et al., 2008b]. The spectrometer was operated infixed transmission mode at 200 eV pass energy for surveyspectra and 40 eV for high-resolution spectra, providingenergy resolutions of 3.2 eV and 0.6 eV, respectively. Theultimate energy resolution of the XPS spectra is limited bythe width of the Al-K X-ray line, 0.85 eV FWHM. Surveyspectra were measured in �5 min, while the more detailedspectra took �60 min.[12] After taking an XPS spectrum, we moved the sample

within the vacuum chamber to the focus of our Thermo-Nicolet Nexus 670 Fourier Transform Infrared Spectrometer(FTIR). The NIR spectrum was measured in the 15000–4000 cm�1 region (0.66–2.5 mm) at a spectral resolution setto 16 cm�1; each spectrum was an average of 800 scans,taking �10 min. Light from a halogen lamp was incidentnormal to the sample and the reflected light was collected at17 degrees with mirrors and focused onto an InGas detector.To obtain the absolute bidirectional reflectance of thesamples, the raw spectrum of the olivine was divided bythat of a PTFE (Teflon) powder (Avocado Research Chem-icals), which has been calibrated against an absolute reflec-

E03003 LOEFFLER ET AL.: IRRADIATION OF OLIVINE BY 4 KEV HE+

2 of 13

E03003

tance standard: spectralon (Labsphere certified reflectancestandard).

2.2. Ion Irradiation

[13] Irradiations were performed with a differentiallypumped, low-energy ion gun. During irradiation the cham-ber pressure remained in the mid 10�8 Torr range, where themean free path of residual gas in the chamber is low enoughthat it does not affect the trajectory of the projectiles or ofthe sputtered particles. We used He gas of purity better than99.99%; the lack of significant amounts of implantedcontaminants was verified in the XPS analyses. Four keVHe ions, which simulate the solar wind, were rastereduniformly over the sample surface analyzed by XPS atnormal incidence. The ion current density was measuredbefore and after each experiment with a Faraday cup to be�1.4 � 1013 ions/cm/sec; precise to within �5%. Duringirradiation, we used an electron flood gun, which emits low-energy (<1 eV) electrons, to prevent electrostatic chargingof the sample, which could ultimately deflect the ion beamaway from the sample.

3. Results

3.1. Fluence Dependence of Near-Infrared Reflectance

[14] Figure 2 (top) shows the in situ NIR spectra of a<45 mm powder of synthetic forsterite and olivine afterdifferent fluences of ion irradiation. The spectrum of for-sterite, which contains �10 times less iron than San Carlosolivine, is featureless in this region and changes onlyslightly during irradiation. The olivine spectrum has theabsorption features characteristic of polycrystalline olivine,because of electronic excitation of the outer shell (3d)electrons of Fe2+ [Burns, 1993]. The spectrum of theunirradiated sample is similar to spectra reported in previousstudies on San Carlos olivine [Yamada et al., 1999; Sasakiet al., 2001; Brunetto et al., 2006a].

[15] During irradiation, there are clear changes in thespectral slope, termed ‘‘reddening,’’ generally consistentwith previous experiments that use comparable ion fluences[Hapke, 2001]. We define reddening, r, from reflectancevalues R outside the main absorptions as: r = (R2.0mm �R0.7mm)/1.3mm, where 1.3 mm is the difference in wave-length. We note that after subtracting a fitted continuum (seebelow in Figure 8), we find that there is no shift in the bandcenter. Even after smoothing each spectrum and taking thederivative, we find that the shift in the �1 mm absorptionband is at most 0.001 mm as a result of irradiation. Thesespectral alterations can be seen more clearly in Figure 2(bottom), where we have divided the spectrum of theirradiated sample by the spectrum of the unirradiatedsample. Interestingly, this change in spectral shape is similarto that reported previously for an olivine powder samplecoated by pulsed-laser deposition of olivine [Loeffler et al.,2008a], laser ablation of olivine [Brunetto et al., 2007],olivine impacted by higher-energy ions [Strazzulla et al.,2005; Brunetto et al., 2006b]. Furthermore, the reddening isalso similar to many asteroid spectra, including, but notlimited to that of Eros [Clark et al., 2002].[16] The similarity of the weathered and unweathered San

Carlos olivine reflectance curves in Figure 2 show thedifficulty that would be encountered when estimating thedegree of space weathering in an actual asteroid spectrum.The small differences can be emphasized by differentiatingthe spectrum. This is shown in Figure 3, where it is quiteclear that reddening occurs essentially below 1.07 mm, andthat the degree of weathering can be ascertained by takingthe ratios of the dips at 0.77 and 0.90 mm to the peaks at1.11 and 1.4 mm. For this reason we propose to usederivative spectra to analyze weathering from asteroids.

3.2. Chemical Analysis of the Surface Using X-RayPhotoelectron Spectroscopy

[17] In addition to measuring the NIR spectrum of olivineduring irradiation we also monitored chemical changes in

Figure 1. Experimental setup.


3 of 13

E03003

the surface with X-ray photoelectron spectroscopy. Wemonitored O, C, Fe, Mg, and Si and list the measuredelectron binding energies in Table 1. Electrostatic chargingof the surface by photoelectron emission and changes in thework function of the spectrometer may cause the peakposition to shift between irradiations, and thus the spectrawere compared after the following procedure. First, thespectra were all energy shifted so that the peak position ofSi-2p oxide coincided at a particular binding energy; thissmall shift was always <0.5 eV. Next, the absolute bindingenergy scale was obtained by using the standard energy of284.8 eV [Powell et al., 1979; Dukes et al., 1999] for the 1speak of the adventitious carbon (present in the unirradiatedspectra because of exposure to atmospheric contaminants).The binding energies measured for San Carlos olivine are ingood agreement with previously published values of olivine[Dukes et al., 1999], forsterite [Hochella and Brown, 1988],fayalite [Schott and Berner, 1983], and Mg1.2Fe0.8SiO4 [Yinet al., 1971].

[18] The photoelectron peak due to adventitious carbonquickly disappears to the noise level after a fluence of�1017 ions/cm2. However, there is a weak low-energyshoulder at 283.8 eV, which remained throughout theexperiment indicating that it is intrinsic to the olivinesample, as expected from terrestrial olivine.[19] Besides the removal of atmospheric contaminants at

low fluences, the other major change observed with XPS isin the Fe-2p photoelectron region. The spectra are shown asa function of fluence in Figure 4. The initial Fe-2p3/2 peakposition at 711 eV indicates the presence of Fe3+ in additionto the intrinsic Fe2+ in olivine (at 709.1 eV). These resultsare in agreement with previous experiments [Schott andBerner, 1983; Dukes et al., 1999] and are attributed tooxidation of surface iron. During bombardment, the surfaceferric iron (Fe3+) was removed and reduced to Fe2+ and Fe0

(at 706.7 eV), because of the preferential loss of oxygenfrom the surface, detected by XPS, via the breaking of Fe-Obonds.[20] No significant change in the binding energy of Si or

Mg core electrons was observed under ion bombardment(Figure 5). The different behavior of Si and Mg compared toFe is expected on the basis of the higher strength of the Si-Oand Mg-O bonds. Finally, through the irradiation series, thepeak position of the O-1s transition remained at 531.2 ±0.2 eV, indicative of the intact SiO4 silicate tetrahedra.[21] Previous in situ XPS studies of irradiation of olivine

are in general agreement with our results: ion irradiation ofolivine forms metallic iron in the sample. The notableexception is the result by Davoisne et al. [2008], whichshows the iron in olivine oxidized toward Fe3+ rather thanreduced to Fe0 when the flux of He ions was below 6 �1014 ions/cm2/sec, but reduced to Fe0 at higher fluxes. Wespeculate that this effect was caused by some minorcontamination, e.g., an oxygen containing species in thebackground gas. In our experiments, where irradiationswere done at much lower pressures and with He+ flux 40times lower than their threshold flux of Davoisne et al.

Figure 2. (top) Absolute reflectance spectra of <45 mmforsterite powder before (a) and after (b) irradiation with2.3 � 1018 ions/cm2. The lower curves are the absolutereflectance spectra of <45 mm olivine powder as a functionof 4 keV He+ fluence of 0, 1.5, 5.3, and 25 (1017 ions/cm2)from top to bottom at 700 nm. (bottom) Reflectance of<45 mm olivine powder divided by that from a virginsample; the fluences are 3, 5.3, and 25 (1017 ions/cm2) fromtop to bottom at 700 nm.

Figure 3. Derivative of the reflectance of a virgin sample(v) and a sample irradiated with 4 keV He+ to 3.1 � 1018

ions/cm2 (i), together with their difference (D). Note thatthe effect of irradiation appears below 1.07 mm.


4 of 13

E03003

[2008], we found that the iron reduced toward Fe0 and didnot form additional Fe3+. Furthermore, increasing the ionflux by a factor of six yielded the same results.

3.3. Comparison of Reddening and Metallization

[22] To compare our XPS data with spectral reddeningdata, we quantified the fraction of surface Fe that is metallicby taking peak areas after deconvolving the Fe 2p3/2photoelectron region (Figure 4) into oxide (Fe3+ and Fe2+)and metallic (Fe0) components using the known positions ofeach component and a Shirley baseline fit [Shirley, 1972]. InFigure 6, we show the fluence dependence of metallizationof iron (based on the XPS data) and the fluence dependenceof spectral reddening. From this plot, it is clear that thereddening is correlated with the increase of metallic iron inthe sample. This result agrees with previous observationsthat lunar soils containing metallic iron nanoparticles hadstrong red slopes in the visible (380–850 nm) reflectance[Keller et al., 1998b].[23] We note that the rate of metallization of the <45 mm

powder olivine sample is �2 times slower than for a flat

olivine section (Figure 7). That the reduction of iron occursin the flat surface, where redeposition is unimportant, willbe discussed below.

3.4. Attenuation of the Fe2+ Absorption Band andDarkening

[24] In addition to the reddening, we also quantified thearea of the 1 mm absorption band by converting thespectrum to optical depth units, taking the natural logarithmof the reflectance, and subtracting a polynomial baseline(Figure 8). This method is similar to the one used in ourprevious works on ices [Loeffler et al., 2006]. Figure 8shows that, as the sample is irradiated, the area of the 1 mmabsorption band decreases by �11% after 3 � 1018 He+/cm2. One can see from Figure 2 that darkening by ionirradiation is at most 2% at 2.0 mm.

3.5. Reoxidation Upon Exposure to Atmosphere

[25] These studies are the first in which ion-inducedchemical and reflectance changes in olivine have beenmeasured simultaneously in situ. We also addressed whetherin situ measurements were necessary, by studying the effectof exposing samples to air before analysis. After He+

irradiation, we removed a sample from vacuum, exposedit to atmosphere and remeasured its reflectance and XPSspectra. The NIR reflectance did not change, within 3%,after exposure to atmosphere for �10 min. However, theamount of metallic iron on the surface decreased signifi-

Table 1. Electron Binding Energies in San Carlos Olivinea

Olivine Constituents Binding Energy (eV)

Fe3+ (2p3/2) 711.4Fe2+ (2p3/2) 709.1Fe0 (2p3/2) 706.7O (1s) 531.4C (1s) 283.8Si (2p) 101.8Mg (2p) 50.1

aThe binding energies reported here are the experimental values shiftedas described in the text. The value listed for C (1s) is for the intrinsic carbonleft at high fluences. Adventitious carbon (Eb = 284.8 eV) is used to fix theabsolute binding energy scale.

Figure 4. XPS spectra of the Fe-2p iron region duringirradiation of <45 mm olivine powder with 4 keV He+ ions.Fluence from bottom to top (1017 ions/cm2) are 0, 3, 5.3,7.8, 11, 15, 20, 25, and 31. The spectra have been displacedvertically for clarity.

Figure 5. XPS spectra of the Mg-2p and Si-2p regionsbefore (0) and after (i) irradiation of a <45 mm olivinepowder with 4 keV He+ ions to a fluence of 3.1 � 1018 ions/cm2. Notice the constancy of the peak positions, in contrastwith those of Fe-2p in Figure 4. The increase in intensityafter irradiation is due to the removal of an impurityoverlayer. In the top spectra one also notices the reductionof iron in the position of the weak Fe-3p photoelectron line.


5 of 13

E03003

cantly because of reoxidation, as first pointed out by Dukeset al. [1999]. Figure 9 (top) shows the Fe-2p region after weexposed it to atmosphere by placing it beside our vacuumsystem in our air-conditioned laboratory for different timeperiods for the powder and rock previously irradiated with4 keV He+. Our shortest exposure time (10 min) causes thepercentage of surface Fe metal to drop by a factor of two.After long exposures, up to a month, this effect saturates,leaving 15% metallic iron on the surface from an as-irradiated level of 75%. This effect suggests that the lunarsoils which had been exposed to the air and had a weakmetallic peak in XPS, [e.g., Baron et al., 1977] may havebeen completely metallic before handling in the laboratory.[26] Interestingly, there are stronger effects of exposing

olivine to atmosphere when irradiation was done with 4 keVAr+ ions, as shown in Figure 9 (bottom). After 10 min, onlya weak shoulder of metallic iron remained in the Fe-2p XPSspectrum, indicating nearly full reoxidation over the depthof XPS; measurements a few days later show that the entiresurface had reoxidized. This difference in the reoxidationtimes for different ions is attributed to easier oxidation ofthe shallower and more damaged region created by Ar, ascompared to He.

4. Discussion

4.1. Overview of Sample Reflectance

[27] Electromagnetic radiation incident on the surface of agrain can be absorbed, reflected, or transmitted. The degreeto which each occurs depends on the optical constants of thematerial. For samples that contain grains much larger thanthe wavelength of light, the reflected wave will leave in aspecular direction from each point on the surface. Becauseof roughness, the surface normal varies locally causing lightto reflect in a variety of directions. The reflected andtransmitted light may enter another grain, where it can bereflected or transmitted with some loss of intensity becauseof absorption by the sample. In the range of 0.6–2.5 mm,

absorption is typically due to electronic transitions betweenlevels split by the crystal field, color centers, or overtones offundamental molecular vibrations that occur at longer wave-lengths [Hapke, 1993; Clark, 1999]. As mentioned earlier,the olivine absorption in this wavelength range is due toelectronic excitation of the outer shell 3d electrons in theFe2+ ions.[28] If the sample also contains particles of size d, smaller

than the wavelength of light, e.g., iron nanoparticles, thenthe reflectance spectra will be altered by the wavelengthdependence of scattering and absorption. The efficienciesfor scattering and absorption decrease as (d/l)4 and d/l,respectively and, for d/l < 0.3 for iron [Hapke, 1993], thespectral reddening is dominated by absorption. We note that

Figure 6. Comparison of reddening (circles) in the NIR and formation of metallic iron (crosses)measured with XPS on <45 mm olivine powder as a function of 4 keV He+ fluence. See text for definitionof reddening.

Figure 7. Metallization of iron, as described by XPS, as afunction of fluence for the flat olivine rock and <45 mmpowder samples.


6 of 13

E03003

only a small amount of metallic iron is needed to signifi-cantly alter the reflectance. As can be seen in Figure 10, Fehas an extremely low absorption length, compared toolivine (e.g., at a wavelength of 1 mm, 20 nm for Fe[Johnson and Christy, 1974] and 500 microns for olivine[Brunetto et al., 2007]).

4.2. Overview of Irradiation of Silicates

[29] KeV ions entering a solid sample lose energy mainlyas a result of binary collisions with target atoms. Some ofthe collisions displace target atoms from their equilibriumlattice positions and ultimately lead to damage. A smallfraction of the projectiles are reflected, the majority come torest in the solid. The cascade of displaced target atomsproduces atomic mixing in multicomponent samples. Whilethe cascade cools (times of the order of picoseconds)diffusion allows the lattice to heal, but some point defects(vacancies and interstitials) remain, which accumulate toform line defects and voids. A large accummulation ofdamage leads to amorphization. Voids assist in the forma-tion of projectile aggregates, and He atoms are known to

aggregate into bubbles of gas. Futagami et al. [1990] foundthat irradiation of olivine and other minerals with 3.6 keVHe ions produced an accummulation of implanted He with amaximum concentration of 0.5–1 � 1017/cm3 that saturatedabove a fluence of �1018/cm2, as expected from thecompetition with sputtering. The overlap of damage regionswith increasing fluence in olivine irradiated by 4 keV Heleads to amorphization above a fluence of 1016 ions/cm2

[Bradley et al., 1996; Demyk et al., 2001; Carrez et al.,2002], to swelling [Wilson et al., 1996] and surface rough-ness [Wilson et al., 1996; Davoisne et al., 2008].[30] A direct manifestation of the atomic displacements is

sputtering, the ejection of atoms and molecules via colli-sions with the solid, a process known to be an importantsurface alteration mechanism for both rocky and icy bodiesin space that are subject to radiation [Johnson, 1990; Hapke,2001; Baragiola, 2003]. Sputtering results in erosion of thesurface, but because the probability of ejection depends onthe local binding energy of the atoms and on their mass, it isoften (though not always) the case that the chemicalcomposition is altered significantly by differential sputter-

Figure 8. (top) Decrease in the 1 mm band area (circles) asa function of He ion fluence for a <45 mm olivine powder.(bottom) Polynomial fit of the baseline in the olivine opticaldepth spectra (in �ln (R) units) used to measure the bandarea, before (a) and after (b) irradiation with 4 keV He+ to afluence of 3.1 � 1018 ions/cm2.

Figure 9. XPS spectra of olivine rock and <45 mm powdersamples after different times of exposure to the atmosphere.(top) Effects after He+ irradiation. From top to bottom,samples after exposure to atmosphere for 0 min (rock), 0 min(powder), 10 min (rock), 10 min (powder), 100 min(powder), 3 days (powder), and 8 days (powder). (bottom)Effects after Ar+ irradiation. From top to bottom, samplesafter exposure to atmosphere for 0 min (rock), 0 min(powder), 10 min (rock), 10 min (powder), and 2 days(rock).


7 of 13

E03003

ing, i.e., the sputtering yield depends on the type of atom, inaddition to its surface concentration. The depth of thealtered layer is of the order of the penetration range of theions [Kelly, 1989]. In our in situ experiments, we only detectslight oxygen loss correlated with reduction of iron. Previ-ous reports of Mg loss [Bradley et al., 1996; Demyk et al.,2001; Toppani et al., 2006] were likely affected by exposureof the irradiated samples to the air, water or solvents prior toanalysis, as mentioned in our previous reports [Cantando etal., 2009].[31] On very rough surfaces, such as powders, the sput-

tering yield decreases because of the redeposition of someof the sputtered flux. The amount of redeposition dependson the shape of the surface. An experimental report indi-cated that a powder sample lost �10 times less mass than abulk sample [Hapke and Cassidy, 1978], but this was donewith a unidirectional ion beam at extremely high fluences(8 � 1020 H/cm2). Even at fluences one tenth as large,needle structures in the direction of the ion source developfrom the irregular grain surfaces [McDonnell and Flavill,1974]. These extreme structures allow redeposition of alarge fraction of the sputtered ejecta, which strongly darkenthe surface [Wu et al., 2001]. Thus, the darkening ofsurfaces by ions reported by Hapke [2001] is likely com-plicated by trapping of light by structures generated vialarge, unidirectional irradiation. These structures would notexist on a surface in space, which is irradiated at differentangles of incidence as it rotates with respect to the solarwind.

4.3. Test for the Effect of Redeposition

[32] Hapke and collaborators [Hapke, 2001, and referencestherein] have clearly shown the importance of high-fluenceion irradiation (hence the solar wind) in the darkening ofFe-bearing minerals. Whereas they conclude that metallic

Fe only forms in redeposited coatings, we found that it iscreated also in flat rocks, where redeposition is negligible.[33] Thus, we need to examine in more detail the relative

role of direct alteration of the surface versus redeposition.We find that, although the changes in chemical compositionof the olivine rock and the powder samples are very similar,the formation rate of metallic iron is a factor of two slowerfor the powder, which we postulate to be related toredeposition.[34] To test the magnitude of redeposition that occurred in

our experiments, we analyzed the relative rate of surfacedepletion of adventitious carbon on the flat rock versus thepowder. After sputter-cleaning each surface we exposed it tothe atmosphere for 10 min to allow the deposition ofatmospheric hydrocarbons. Then, we reinserted the samplesinto vacuum and measured with XPS the amount ofadventitious carbon on the surface and its decrease as aresult of He+ irradiation. Since the hydrocarbon content inthe air can change from day to day and with time, werepeated each experiment on different days but foundvariations of less than 10%. Our results for the carbondepletion versus fluence are shown in Figure 11 for flat rockand powder samples. We found that the carbon removal ratefor the powder is about a third lower than that for the flatsurface. Thus, we can conclude that about two thirds of theejecta is redeposited. This rough estimate is similar to the to0.70 predicted by modeling of redeposition [Cassidy andJohnson, 2005]. The difference with the value of 0.9measured by Hapke and Cassidy [1978] is most likelydue to the needle topography that is expected in the experi-ments by these authors, as discussed above.[35] To understand the different conclusions proposed by

us and Hapke [2001, and references therein] on the role ofdirect metallization of a surface, we note that those authors’

Figure 10. Absorption path length for olivine (data adapted from Brunetto et al. [2007]) and iron (datataken from Johnson and Christy [1974]).


8 of 13

E03003

conclusions that only sputtered deposits will contain metal-lic iron were based mainly on experiments where irradiationwas done on a ball situated in an alumina crucible, bothmade of impure alumina (clean alumina did not darken)[Hapke, 2001]. In such geometry, only the top of the balland the crucible were irradiated and only the lower half ofthe ball would have sputtered material deposited onto it.The results showed that the lower half darkened much morethan the top half of the ball, and were not affected by ion-induced topography mentioned above. However, they wereinterpreted to mean that redeposition of the aluminum (onthe bottom of the ball) was more important than directirradiation (of the top of the ball) in altering the opticalreflectance. We argue that this difference is caused simplyby the fact that the thickness of the altered region of thetop of the ball is on the order of the ion range and islimited by sputtering, whereas the thickness of the depositin the bottom of the ball can increase indefinitely withfluence, since it is being sputtered from another surface(the crucible).[36] The fact that metallization does not require redepo-

sition, as evidenced in this and previous work on sectionedsurfaces [Dukes et al., 1999; Davoisne et al., 2008], doesnot imply that irradiation of sectioned surfaces will causesignificant optical effects. The effect on the spectral reflec-tance will be larger on powders, only because scattering ofthe grains causes light to pass through the altered regionsmultiple times, irrespective of whether the main mechanismfor formation of metallic iron is vapor redeposition or direction bombardment.[37] In closing this discussion, we propose that the reason

why the powders metallize at a slower rate than a flatsurface is that there is more more redeposition in the powderand the redeposited material is partially oxidized, as com-monly happens in the sputter deposition of oxides invacuum [Mauvernay et al., 2007].

4.4. Mechanism for Alteration of Reflectance Spectraof Silicates

[38] As mentioned above, Hapke [2001] has shown thatdarkening of visible reflectance spectra in several types ofmineral oxides was correlated with the sputter redepositionof coatings containing metallic iron.[39] Recent experiments in Catania [Brunetto and

Strazzulla, 2005; Strazzulla et al., 2005; Brunetto et al.,2006b] have shown that higher-energy (60–400 keV) heavyions can also cause significant alteration of spectral reflec-tance but at much lower ion fluences than used by Hapke[2001]. Although heavy ions are unimportant in the solarwind, these experiments can help in understanding weath-ering mechanisms. The Catania group proposed that thealteration they observed at lower fluences was caused byamorphization, which would be an additional mechanismindependent of the metallic iron in the sample. However, wesuggest that the lower fluences needed to change thespectral reflectance in these higher-energy heavy ion experi-ments are due to the difference in density of energydeposition and projectile range, not to a separate mecha-nism. The 60–400 keVAr+ ions used by the Catania grouphave �100 times higher stopping power and penetrate 2.5–14 times more than the 1 keV H ions used by Hapke’s group(J. F. Ziegler and J. P. Biersack, Stopping and Range of Ionsin Matter (SRIM)-2006, 2006, available at http://srim.org)and will cause more damage, create more metallic iron, andcause more spectral alteration for the same ion fluence. Asan additional test, we irradiated a sectioned olivine rockwith 4 keV Ar, which has a stopping power �11 timeshigher than 4 keV He used in our experiments and foundthat the metallization rate is �26 times faster than it was forHe. Regarding the role of amorphization, we note the subtlechange of optical properties due to amorphization wouldhave a negligible effect in optical reflectance at near infraredwavelengths. In fact, we detected no reflectance changes atfluences of 1016–1017 ions/cm2 where olivine has alreadybeen amorphized by 4 keV He [Carrez et al., 2002], and noreddening after irradiating iron-free forsteritewith 2� 1018 Heions/cm2 (Figure 2 (top)). It is only at the high fluences ofirradiation of olivine, where formation of metallic iron issignificant (Figure 4), that we observe spectral reflectancechanges. Therefore, we conclude that metallization, ratherthan amorphization, is responsible for the changes in thespectral reflectance.

4.5. Formation of Metallic Iron in Silicates by the SolarWind

[40] During irradiation of olivine, metallic iron is formedin the sample, because oxygen is preferentially removed bysputtering, and radiation enhanced diffusion assists in thenucleation of the reduced iron. Metallic iron can also formin a sputtered deposit because Fe atoms deposit preferen-tially compared to the lighter O atoms, as shown by Hapke[2001].[41] The effect of metallic iron particles on the reflectance

spectra will depend on their size. Lunar soil analysis hasshown spectrally darker soils contain large iron nanopar-ticles (>10 nm), while smaller nanoparticles (<5 nm) aremore prominent in soils that are spectrally reddened [Kelleret al., 1998a]. These results are consistent with more recentlaboratory measurements [Noble et al., 2007] on silicate

Figure 11. Removal of of adventitious carbon from thesurface of an olivine rock and olivine powder by sputtering.The solid lines are double exponentials fit to the data.


9 of 13

E03003

gels, which show that addition of Fe nanoparticles smallerthan 15 nm produce reddening while addition of Fe particleslarger than 40 nm induce darkening of the silicate. There-fore we conclude that since the He bombardment reddensour olivine samples but barely darkens their spectral reflec-tance, the metallic iron formed in our experiments exists invery small precipitates.

4.6. Astronomical Implications

[42] We can scale our results to an effective exposure timein space. The estimates are approximate because fluxes ofsolar wind ions and micrometeorites have changed throughhistory. We consider the case of the Moon; similar estimatescan be done for Mercury and asteroids. For the Moon, weuse the mean number density (n/cm3) and mean velocity(468 km/sec) given by Gosling [2007] at 1 AU and obtain aproton flux of 4.1 � 108 protons/cm2/sec. The incident flux(F0) of He is estimated to be �1.9 � 107 He/cm2/sec,obtained from the product of the proton flux by the averagerelative abundance of helium ions in the solar wind (0.047).However, we note that the cross section being struck by thesolar wind is pR2, while the surface area of a sphere is 4pR2,

so the average flux is actually 1/4 of F0 or 4.8 �106 ions/cm2/sec. Using this value, we find that it takes�13,000 years for an exposed grain to reach a fluencewhere alteration induced by the solar wind saturates (2 �1018 He ions/cm2). The additional effects of solar windprotons can be assumed to scale roughly as the product oftheir flux times the sputtering yield [Johnson and Baragiola,1991], since sputtering is governed by the same type ofcollisions that cause surface damage. Thus, we obtain acharacteristic time for weathering of �5000 years. This doesnot mean that the appearance of all surfacematerial older than�5000 years will be the same, because there is a complexinteraction between the effects caused by the solar wind andmicrometeorites.[43] Calculating the timescales for effects caused by

micrometerorite impacts is non trivial. The impacts arenot uniform [Gault et al., 1974; Horz et al., 1974] andcan modify grains by vaporization and changing theircomposition. In addition, impacts mix the regolith, breaklarge grains, form craters and eject other grains indirectly bythe shock wave induced in the impact. The effect ofchemical alterations on reflectance was studied by Sasakiet al. [2001] using ns pulsed laser irradiation as an analogfor micrometeorite impacts. In comparison with our experi-ments, they reported somewhat more spectral reddening anddarkening after one laser pulse (24 J/cm2), which theyestimated to be equivalent to �108 years exposure at1 A.U. In addition, they assumed that this exposure timemight be a lower limit, since laser irradiation may be moreefficient than dust impacts in altering the sample. Thisassertion is supported by experiments comparing ther-moionic plasma emitted from samples irradiated with pulsedlasers and dust impacts show that similar secondary ionmass spectra and absolute yields could be obtained if thepower density of the laser was �100 times lower than thatof the dust [Kissel and Krueger, 1987].[44] On the other hand, Sasaki et al. [2001] obtained the

timescale of 108 years by estimating that the energy fluxcould be obtained by applying the cumulative flux ofmicrometeorites to 10�12 g particles. However, we notethat it has previously been estimated that >75–80% of thematerial vaporized by impacts at 1 A.U. comes from heavierprojectiles between 10�8 and 10�4 g (10–230 mm particlesassuming a density of 2 g/cm3) [Gault et al., 1972; Cintala,1992], even though their flux is much lower. Thus to obtaina more accurate estimate of timescales for vaporizationcaused by micrometeorite impacts, we need to integratethe differential energy flux over the range of micrometeor-ites responsible for vaporization. The differential flux (f(m))of a micrometeorite is given by f(m) = �dF(m)/dm [Grun etal., 2001], where F is the cumulative flux of particles ofmass larger than m. An approximation for F has been givenby Gault et al. [1972] for 10�13 g < m < 103 g at 1 A.U. Byintegrating the product of the impact energy and �dF/dmover masses between 10�13 g and 103 g, assuming anaverage impact velocity of 13 km/sec [Cintala, 1992], weobtain an energy flux of �1000 J/cm2/106 a.[45] In Figure 12, we replot the ratio of the reflectance

spectra of irradiated samples to that of the virgin samples,shown in Figure 2 together with previous laser ablationstudies [Sasaki et al., 2001; Brunetto et al., 2006a]. We havenormalized all spectra at 2 microns before taking the ratios,

Figure 12. Comparison of our reflectance ratio (Figure 2(bottom)) with that of spectra extrapolated from Sasaki et al.[2001] and Brunetto et al. [2006a], all for San Carlosolivine. We have normalized spectra at 2 mm before takingthe reflectance ratio. The data from top to bottom at 700 nmafter 4 keV He irradiation, a pulse of 12 J/cm2 [Sasaki et al.,2001], a pulse of 24 J/cm2 [Sasaki et al., 2001], and 13–14 pulses of 2 J/cm2 [Brunetto et al., 2006a].


10 of 13

E03003

to compare reddening and not darkening, which can dependon microstructure caused by laser ablation. The NIR spec-trum measured by Sasaki et al. [2001] after one laser pulse(12 J/cm2) is quite similar to our ion irradiation experiment.The energy fluence of 12 J/cm2 is equivalent to �106 yearsof exposure to micrometeorite impacts, taking into accountthe increased efficiency of the laser mentioned above. Thus,we estimate that spectral reddening caused by the solar windsaturates about 2 orders of magnitude faster than estimatedto occur by micrometeorite impacts. However, we note thatif the grain survives for more than a million years, asis likely the case on the surface of the Moon, microme-teorite impacts may become more important as the spec-tral reddening effect is not saturated after one impact (seeFigure 12).[46] The importance of the solar wind is substantiated

also by the fact that amorphous rims in regolith grains havea thickness of the order of the range of solar wind ions(considering slow and fast components) [Christoffersen etal., 1996; Keller and McKay, 1997]. Furthermore, we notethat our experimental results also lead to the conclusion thatreduction of iron in the lunar regolith in the irradiated regioncan explain the formation of submicron Fe particles andtheir effect on optical reflectance, in contrast to the assertionof Hapke [2001] that direct ion impact produces onlynegligible effects. Our conclusion, of course, does notpreclude the contribution of redeposition of micrometeoriteimpact ejecta and, perhaps, redeposition of sputteredparticles.[47] With respect to space weathering effects on asteroids,

our experiments strengthen the idea that, solar wind irradi-ation and the consequent formation of metallic iron alter thereflectance spectra of olivine-bearing S type asteroids. Morespecifically we can apply our results, which show ionirradiation of olivine causes reddening and attenuation ofthe 1 mm absorption band but little darkening of its overallspectrum, to the main belt asteroids 243 Ida (pyroxene rich)and 951 Gaspra (olivine rich). Like in our experiments, thespectra of each of these asteroids show variations in slopeand absorption band depth but minor albedo variationsacross their respective surfaces [Chapman, 1996; Clark etal., 2002]. We speculate that this is a consequence of theirlocation, because although both the solar wind and micro-meteorite fluxes decline as one goes from the Sun to theasteroid belt, the impact speed of the micrometeorites alsodecreases. This will in turn decrease significantly the impactenergy and hence, the amount of material vaporized [Blochet al., 1971; Mandeville and Vedder, 1971; Cintala, 1992].Thus, micrometeorite impacts in the main asteroid belt willmainly act to bring fresh material to the surface, while thesolar wind will be the primary space weathering agent.[48] Closer to the Sun, micrometeorites have higher

fluxes and velocities and thus weathering becomes morecomplicated. As an example, Itokawa and Eros, near-EarthS type asteroids located at similar heliocentric distances,have spectra that exhibit different space weathering trendsand variations across their surfaces [Clark et al., 2002;Murchie et al., 2002; Hiroi et al., 2006; Ishiguro et al.,2007]. These differences likely indicate that other proper-ties, besides ion and micrometeorite impact determine theappearance of those objects. Still, the discovery of weath-ered large boulders on Itokawa [Ishiguro et al., 2007] can

be explained by our results that show metallic iron can beformed by direct ion irradiation of a flat rock of olivine,where redeposition does not occur.

5. Conclusions

[49] Four keV He ions, typical of the solar wind, alter thereflectance spectrum of olivine powders by reddening thespectral slope and attenuating the 1 mm absorption feature, amanner consistent with observations of space weathering.Furthermore, as judged by in situ chemical analysis, spectralreddening corresponds directly to the formation of metalliciron in the outer 50–80 A of the mineral surface. Inaddition, we find that exposing the irradiated surfaces tothe atmosphere causes the metallic Fe to reoxidize, showingthat in situ measurements are necessary to study chemicalchanges occurring during ion bombardment of minerals andthat previous XPS measurements of lunar soils may haveunderestimated the metallic content of the surface.[50] Our results show that metallic iron forms more

slowly on a powder than on a flat rock of olivine, whichwe suggest is due to partial oxidation of redepositedsputtered material. Interestingly, the spectral changes ob-served in our ion irradiation experiments are similar toprevious weathering experiments using laser impact, indi-cating that both processes affect similarly the spectralreflectance of olivine. In addition, we estimate that at1 A.U. the spectral reddening of olivine induced by Hebombardment will saturate 2 orders of magnitude faster thanthe time it takes for similar reddening effects to occur as aresult of micrometeorite impacts.

[51] Acknowledgments. This work was supported by the NASACosmochemistry and Lunar Advanced Science and Exploration Researchprograms. We would like to thank W. Y. Chang for his technical assistancein the beginning of this project. We would like to thank B. E. Clark and ananonymous referee with their very helpful reviews.

ReferencesBaragiola, R. A. (2003), Water ice on outer solar system surfaces: Basicproperties and radiation effects, Planet. Space Sci., 51, 953 – 961,doi:10.1016/j.pss.2003.05.007.

Baron, R. L., et al. (1977), The surface composition of lunar soil grains: Acomparison of the results of auger and X-ray photoelectron (ESCA)spectroscopy, Earth Planet. Sci. Lett., 37, 263–272, doi:10.1016/0012-821X(77)90172-8.

Bloch, M. R., et al. (1971), Meteorite impact craters, crater simulations, andthe meteoroid flux in the early solar system, paper presented at SecondLunar Science Conference, Lunar and Planet. Sci. Inst., Houston, Tex.

Bradley, J. P. (1994), Chemically anomalous, preaccretionally irradiatedgrains in interplanetary dust from comets, Science, 265, 925 –929,doi:10.1126/science.265.5174.925.

Bradley, J. P., et al. (1996), Radiation processing and the origins of inter-planetary dust, Lunar Planet. Sci., XXVII, 149–150.

Bringa, E. M., et al. (2007), Energetic processing of interstellar silicategrains by cosmic rays, Astrophys. J., 662, 372–378, doi:10.1086/517865.

Brunetto, R., and G. Strazzulla (2005), Elastic collisions in ion irradiationexperiments: A mechanism for space weathering of silicates, Icarus, 179,265–273.

Brunetto, R., et al. (2006a), Space weathering of silicates simulated bynanosecond pulsed UV excimer laser, Icarus, 180, 546 – 554,doi:10.1016/j.icarus.2005.10.016.

Brunetto, R., et al. (2006b), Modeling asteroid surfaces from observationsand irradiation experiments: The case of 832 Karin, Icarus, 184, 327–337, doi:10.1016/j.icarus.2006.05.019.

Brunetto, R., et al. (2007), Optical characterization of laser ablated silicates,Icarus, 191, 381–393, doi:10.1016/j.icarus.2007.04.023.

Burns, R. G. (1993), Mineralogical Applications of Crystal Field Theory,551 pp., Cambridge Univ. Press, New York.


11 of 13

E03003

Cantando, E. D., C. A. Dukes, and R. A. Baragiola (2009), Aqueous dis-solution of olivine enhanced by ion irradiation, J. Geophys. Res., 113,E09011, doi:10.1029/2008JE003119.

Carrez, P., et al. (2002), Low-energy helium ion irradiation-induced amor-phization and chemical changes in olivine: Insights for silicate dust evo-lution in the interstellar medium, Meteorit. Planet. Sci., 37, 1599–1614.

Cassidy, T. A., and R. E. Johnson (2005), Monte Carlo model of sputteringand other ejection processes with a regolith, Icarus, 176, 499–507,doi:10.1016/j.icarus.2005.02.013.

Chapman, C. R. (1996), S-type asteroids, ordinary chondrites, and spaceweathering: The evidence from Galileo’s fly bys of Gaspra and Ida,Meteorit. Planet. Sci., 31, 699–725.

Chapman, C. R. (2004), Space weathering of asteroid surfaces, Annu. Rev.Earth Planet. Sci., 32, 539 – 567, doi:10.1146/annurev.earth.32.101802.120453.

Christoffersen, R., et al. (1996), Microstructure, chemistry, and origin ofgrain rims on ilmenite from the lunar soil finest fraction,Meteorit. Planet.Sci., 31, 835–848.

Cintala, M. J. (1992), Impact-induced thermal effects in the lunar andMercurian regoliths, J. Geophys. Res., 97, 947–973.

Clark, B. E., et al. (2002), Asteroid space weathering and regolith evolu-tion, in Asteroids III, edited by W. F. Bottke Jr. et al., pp. 585–599, Univ.of Ariz. Press, Tucson, Ariz.

Clark, R. N. (1999), Spectroscopy of rocks and minerals, and principles ofspectroscopy, in Manual of Remote Sensing, edited by A. N. Rencz,pp. 3–58, Wiley, New York.

Davoisne, C., et al. (2008), Chemical and morphological evoultion of asilicate surface under low-energy ion irradiation, Astron. Astrophys., 482,541–548, doi:10.1051/0004-6361:20078964.

Demyk, K., et al. (2001), Structural and chemical alteration of crystallineolivine under low energy He+ irradiation, Astron. Astrophys., 368, L38–L41, doi:10.1051/0004-6361:20010208.

Dukes, C. A, et al. (1999), Surface modification of olivine by H+ and He+

bombardment, J. Geophys. Res., 104, 1865–1872, doi:10.1029/98JE02820.Futagami, T., et al. (1990), Helium ion implantation into minerals, EarthPlanet. Sci. Lett., 101, 63–67, doi:10.1016/0012-821X(90)90124-G.

Gaffey, M. J., et al. (1993), Mineralogical variations within the S-typeasteroid class, Icarus, 106, 573–602, doi:10.1006/icar.1993.1194.

Gault, D. E., et al. (1972), Effects of microcratering on the lunar surface,Geochim. Cosmochim. Acta, 3, 2713–2734.

Gault, D. E., et al. (1974), Mixing of the lunar regolith, paper presented atFifth Lunar Conference, Lunar and Planet. Sci. Inst., Houston, Tex.

Gosling, J. T. (2007), The solar wind, in Encyclopedia of the Solar System,2nd ed., edited by L.-A. McFadden, P. R. Weissman, and T. V. Johnson,Academic, Amsterdam.

Grun, E., et al. (2001), Situ measurements of cosmic dust, in InterplanetaryDust, edited by E. Grun et al., pp. 295–346, Springer, New York.

Hapke, B. W. (1973), Darkening of silicate rocks by solar wind sputtering,Moon, 7, 342–355, doi:10.1007/BF00564639.

Hapke, B. W. (1993), Theory of Reflectance and Emittance Spectroscopy,455 pp., Univ. of Cambridge, Cambridge, U. K.

Hapke, B. W. (2001), Space weathering from Mercury to the asteroid belt,J. Geophys. Res., 106, 10,039–10,074, doi:10.1029/2000JE001338.

Hapke, B. W., and W. Cassidy (1978), Is the Moon really as smooth as abilliard ball? Remarks concerning recent models of sputter-fractionationon the lunar surface, Geophys. Res. Lett., 5, 297–300, doi:10.1029/GL005i004p00297.

Hiroi, T., et al. (2006), Developing space weathering on the asteroid 25143Itokawa, Nature, 443, 56–58, doi:10.1038/nature05073.

Hochella, M. F., Jr., and G. E. Brown Jr. (1988), Aspects of silicate surfaceand bulk structure analysis using X-ray photoelectron spectroscopy, Geo-chim. Cosmochim. Acta , 52 , 1641 – 1648, doi:10.1016/0016-7037(88)90232-3.

Horz, F., et al. (1974), Micrometeoroid abrasion of lunar rocks: A MonteCarlo simulation, paper presented at Fifth Lunar Conference, Lunar andPlanet. Sci. Inst., Houston, Tex.

Ishiguro, M., et al. (2007), Global mapping of the degree of space weath-ering on asteroid 25143 Itokawa by Hayabusa/AMICA observations,Meteorit. Planet. Sci., 42, 1791–1800.

Johnson, P. B., and R. W. Christy (1974), Optical constants of transitionmetals: Ti, V, Cr, Mn, Fe, Co, Ni, and Pd, Phys. Rev. B, 9, 5056–5070,doi:10.1103/PhysRevB.9.5056.

Johnson, R. E. (1990), Energetic Charged-Particle Interactions With Atmo-spheres and Surfaces, Springer, Berlin.

Johnson, R. E., and R. Baragiola (1991), Lunar surface - sputtering andsecondary ion mass spectrometry, Geophys. Res. Lett., 18, 2169–2172,doi:10.1029/91GL02095.

Keller, L. P., and D. S. McKay (1993), Discovery of vapor deposits in thelunar regolith, Science, 261, 1305 – 1307, doi:10.1126/science.261.5126.1305.

Keller, L. P., and D. S. McKay (1997), The nature and origin of rims onlunar soil grains, Geochim. Cosmochim. Acta, 61, 2331 – 2341,doi:10.1016/S0016-7037(97)00085-9.

Keller, L. P., et al. (1998a), Space weathering: Reflectance spectroscopyand TEM analysis of individual lunar soil grains, paper presented atLunar and Planetary Institute Conference, Lunar and Planet. Sci. Inst.,Houston, Tex., 1 Mar.

Keller, L. P., et al. (1998b), Surface-correlated nanophase iron metal inlunar soils: Petrography and space weathering effects, paper presentedat Workshop on New Views of the Moon: Integrated Remotely Sensed,Geophysical, and Sample Datasets, Lunar and Planet. Sci. Inst., Houston,Tex.

Kelly, R. (1989), Bombardment-induced compositional change with alloys,oxides, oxysalts, and halides: Part II. The role of segregation, Nucl.Instrum. Methods Phys. Res. Sect. B, 39, 43 –56, doi:10.1016/0168-583X(89)90739-8.

Kissel, J., and F. R. Krueger (1987), Ion formation by impact of fast dustparticles and comparison with related techniques, Appl. Phys. A, 42, 69–85, doi:10.1007/BF00618161.

Loeffler, M. J., et al. (2006), Synthesis of hydrogen peroxide in waterice by ion irradiation, Icarus, 180, 265 – 273, doi:10.1016/j.icarus.2005.08.001.

Loeffler, M. J., et al. (2008a), Laboratory simulations of redeposition ofimpact ejecta on mineral surfaces, Icarus, 196, 285–292, doi:10.1016/j.icarus.2008.02.021.

Loeffler, M. J., et al. (2008b), Laboratory simulations of sulfur depletion atEros, Icarus, 195, 622–629, doi:10.1016/j.icarus.2008.02.002.

Mandeville, J. C., and J. F. Vedder (1971), Microcraters formed in glass bylow density projectiles, Earth Planet. Sci. Lett., 11, 297 – 306,doi:10.1016/0012-821X(71)90183-X.

Marchi, S., et al. (2006), A general spectral slope-exposure relation for S-type Main belt and near-Earth asteroids, Astron. J., 131, 1138–1141,doi:10.1086/498721.

Mauvernay, B., et al. (2007), Elaboration and characterization of Fe1–xOthin films sputter deposited from magnetite target, Thin Solid Films, 515,6532–6536, doi:10.1016/j.tsf.2006.11.131.

McDonnell, J. A. M., and R. P. Flavill (1974), Solar wind sputtering on thelunar surface: Equilibrium crater densities related to past and presentmicroparticle influx rates, Lunar Sci., III, 2441–2449.

Moroz, L. V., et al. (1996), Optical effects of regolith processes on S-asteroids as simulated by laser shots on ordinary chondrite and othermafic materials, Icarus, 122, 366–382.

Moulder, J. F., et al. (1992), Handbook of X-Ray Photoelectron Spectro-scopy, 261 pp., Phys. Electron., Eden Praire, Minn.

Murchie, S., et al. (2002), Color variations on Eros from NEAR multi-spectral imaging, Icarus, 155, 145–168, doi:10.1006/icar.2001.6756.

Nash, D. (1967), Proton irradiation darkening of rock powders: Contam-ination and temperature effects, and applications to solar wind darkeningof the Moon, J. Geophys. Res., 72, 3089 – 3104, doi:10.1029/JZ072i012p03089.

Noble, S. K., et al. (2007), An experimental approach to understanding theoptical effects of space weathering, Icarus, 192, 629–642, doi:10.1016/j.icarus.2007.07.021.

Pillinger, C. T. (1979), Solar-wind exposure effects in lunar soil, Rep. Prog.Phys., 42, 897–961, doi:10.1088/0034-4885/42/5/003.

Powell, C. J., et al. (1979), Results of a joint Auger-ESCA round robinsponsored by ASTM committee E-42 on surface analysis, I, ESCA re-sults, J. Electr. Spectrosc. Related Phenomena, 17, 361 – 403,doi:10.1016/0368-2048(79)80001-8.

Sasaki, S., et al. (2001), Production of iron nanoparticles by laser irradiationin a simulation of lunar-like space weathering, Nature, 410, 555–557,doi:10.1038/35069013.

Schott, J., and R. A. Berner (1983), X-ray photoelectron studies of themechanism of iron silicate dissolution during weathering, Geochim. Cos-mochim. Acta, 47, 2233–2240, doi:10.1016/0016-7037(83)90046-7.

Seah, M. P., and W. A. Dench (1979), Quantitative electron spectroscopy ofsurfaces: A standard data base for electron inelastic mean free path insolids, Surf. Interface Anal., 1, 2–11, doi:10.1002/sia.740010103.

Shirley, D. A. (1972), High-resolution X-ray photoemission spectrum ofthe valence bands of gold, Phys. Rev. B, 5, 4709–4714, doi:10.1103/PhysRevB.5.4709.

Strazzulla, G., et al. (2005), Spectral alteration of the Meteorite Epinal (H5)induced by heavy ion irradiation: A simulation of space weathering ef-fects on near-Earth asteroids, Icarus, 174, 31 – 35, doi:10.1016/j.icarus.2004.09.013.

Toppani, A., et al. (2006), Segregation of Mg, Ca, Al, and Ti in silicatesduring ion irradiation, Lunar Planet. Sci., XXXVII, abstract 2056.

Wilson, C., et al. (1996), Atomic force microscopy of olivine, in AtomicForce Microscopy/Scanning Tunneling Microscopy 3, edited by S. H.Cohen and M. L. Lightbody, pp. 125–134, Springer, New York.


12 of 13

E03003

Wu, C., et al. (2001), Near-unity below-band-gap absorption by microstruc-tured silicon, Appl. Phys. Lett., 78, 1850–1853, doi:10.1063/1.1358846.

Yamada, M., et al. (1999), Simulation of space weathering of planet-form-ing materials: Nanosecond pulse laser irradiation and proton implantationon olivine and pyroxene samples, Earth Planets Space, 51, 1255–1265.

Yin, L. I., et al. (1971), Core binding energy difference between bridgingand nonbridging oxygen atoms in a silicate chain, Science, 173, 633–635, doi:10.1126/science.173.3997.633.

��R. A. Baragiola, C. A. Dukes, and M. J. Loeffler, Laboratory for Atomic

and Surface Physics, University of Virginia, 351 McCormick Road, P. O.Box 400238, Charlottesville, VA 22904, USA. ([email protected])


13 of 13

E03003

: simulation of space weathering by the solar wind papers/loeffler space weathering by … ·...

Documents