Icarus 212 (2011) 24–41

Contents lists available at ScienceDirect

Icarus

journal homepage: www.elsevier .com/ locate/ icarus

On the chemistry of mantle and magmatic volatiles on Mercury

Mikhail Yu. Zolotov ⇑School of Earth and Space Exploration, Arizona State University, Tempe, AZ 85287-1404, USA

a r t i c l e i n f o a b s t r a c t

Article history:Received 21 November 2009Revised 8 November 2010Accepted 19 December 2010Available online 25 December 2010

Keywords:Mercury, InteriorMercury, SurfaceVolcanismCosmochemistry

0019-1035/$ - see front matter � 2010 Elsevier Inc. Adoi:10.1016/j.icarus.2010.12.014

⇑ Fax: +1 480 965 8102.E-mail address: zolotov@asu.edu

The surface of Mercury contains ancient volcanic features and signs of pyroclastic activity related tounknown magmatic volatiles. Here, the nature of possible magmatic volatiles (H, S, C, Cl, and N) is dis-cussed in the contexts of formation and evolution of the planet, composition and redox state of its mantle,solubility in silicate melts, chemical mechanisms of magma degassing, and thermochemical equilibria inmagma and volcanic gases. The low-FeO content in surface materials (<6 wt%) evaluated with remotesensing methods corresponds to less than 2.3 fO2 log units below the iron–wüstite buffer. These redoxconditions imply restricted involvement of hydrous species in nebular and accretion processes, and a lim-ited loss of S, Cl, and N during formation and evolution of the planet. Reduced conditions correspond tohigh solubilities of these elements in magma and do not favor degassing. Major degassing and pyroclasticactivity would require oxidation of melts in near-surface conditions. Low-pressure oxidation of graphitein moderately oxidized magmas causes formation of low-solubility CO. Decompression of reduced N-sat-urated melts involves oxidation of nitride melt complexes and could cause N2 degassing. Putative assim-ilation of oxide (FeO, TiO2, and SiO2) rich rocks in magma chambers could have caused major degassingthrough oxidation of graphite and S-, Cl- and N-bearing melt complexes. However, crustal rocks maynever have been oxidized, and sulfides, graphite, chlorides, and nitrides could remain in crystallizedrocks. Chemical equilibrium models show that N2, CO, S2, CS2, S2Cl, Cl, Cl2, and COS could be amongthe most abundant volcanic gases on Mercury. Though, these speciation models are restricted by uncer-tain redox conditions, unknown volatile content in magma, and the adopted chemical degassingmechanism.

� 2010 Elsevier Inc. All rights reserved.

1. Introduction: Volcanic features and composition of surfacematerials

Morphological and color analyses of Mariner 10 images suggesta volcanic origin of some surface features on Mercury (Strom et al.,1975; Basaltic Volcanism, 1981; Spudis and Guest, 1988; Robinsonand Lucey, 1997). The presence of volcanic rocks is confirmed bythe initial MESSENGER results (Head et al., 2008, 2009a; Robinsonet al., 2008; Murchie et al., 2008; Prockter et al., 2010). The volca-nic features demonstrate mostly effusive lunar-like activity with-out stratovolcanoes or large sheet structures (Head et al., 2008,2009a). However, several features could be interpreted as pyroclas-tic deposits (Head et al., 2008, 2009a; Murchie et al., 2008; Robin-son et al., 2008; Kerber et al., 2009; Prockter et al., 2010).

Spectroscopic data in visible, near- and mid-infrared rangesindicate the compositional diversity of the Mercury’s surface(e.g., Sprague et al., 2009; Denevi et al., 2009; Vernazza et al.,2010). Suggested volcanic plains and pyroclastic deposits arebrighter and redder than underlying rocks and impact ejectae

ll rights reserved.

(Robinson and Lucey, 1997; Murchie et al., 2008; Robinson et al.,2008; Ernst et al., 2010). The ground mid-infrared emission spectrasuggest silicate surface materials with spatially variable SiO2 con-tent and dominant mafic to ultramafic compositions (Boyntonet al., 2007). These interpretations are consistent with high heightsof lobate flow fronts indicating high-volume eruptions of maficlava similar to terrestrial flood-basalts (Wilson and Head, 2008).Some regions (e.g., the inner part of Caloris basin) could be repre-sented by evolved igneous rocks with intermediate compositions,as suggested from mid-infrared spectra (Sprague et al., 2002,2009; Boynton et al., 2007). Intermediate compositions may repre-sent rocks formed through upward fractional crystallization of aputative magma ocean (Brown and Elkins-Tanton, 2009) orthrough differentiation of impact melts and basaltic magma inchambers. Although magma chambers are suggested from someimages (Head et al., 2009b), there are no morphological signs ofvolcanic activity associated with evolved magmas (Head et al.,2008, 2009a).

Visible, near- and mid-infrared spectra of Mercury’s surface areconsistent with the presence of Mg-rich olivine and pyroxene, Mg-and Ca-clinopyroxene, plagioclase feldspars, silicate glasses, andsulfides (Sprague et al., 1995, 2002, 2009; Warell et al., 2006,

M.Yu. Zolotov / Icarus 212 (2011) 24–41 25

2010; Boynton et al., 2007; Vernazza et al., 2010), though interpre-tations are model-dependent and surface mineralogy remainsuncertain. The lack of spectral features in visible and near-infraredranges can be accounted for by the presence of low-albedo opaquephase(s), which could represent sulfides and/or Ti-bearing oxides(Rava and Hapke, 1987; Denevi and Robinson, 2008; Deneviet al., 2009; Vernazza et al., 2010). The detection of Mg in the exo-sphere of Mercury (McClintock et al., 2009) agrees with the identi-fication of Mg-silicates by ground-based spectroscopy, althoughMg sulfides could be present as well (Sprague et al., 1995). Thepresence of Ca and alkalis in the exosphere (e.g., McClintocket al., 2009) also agrees with the identification of Ca-pyroxenesand feldspars in surface materials. Surface spectra of Mercury innear- and mid-infrared are distinct from lunar spectra, probablyindicating a fundamental difference in abundance and compositionof ferrous silicates (Vernazza et al., 2010).

Although the abundance and speciation of iron in surface mate-rials remains uncertain, the majority of remote-sensing data indi-cate a low-FeO content. The lack of 1 lm absorption in most ofMercury’s spectra, the color of the surface, and microwave emissiv-ity data indicate an FeO-depleted composition (<3–6 wt% FeO)(e.g., McCord and Clark, 1979; Vilas, 1988; Blewett et al., 1997;Robinson and Taylor, 2001; Boynton et al., 2007). Mid-infraredground emission spectra do not reveal FeO-rich compositions(Sprague et al., 2002, 2009; Boynton et al., 2007; Vernazza et al.,2010). These ground observations are consistent with the interpre-tations of Mariner 10 color data (Robinson and Lucey, 1997; Deneviand Robinson, 2008). Initial MESSENGER results also do not indi-cate a presence of abundant surface ferrosilicates (Solomon et al.,2008; McClintock et al., 2008; Robinson et al., 2008).

The interpretation of MESSENGER neutron spectrometer dataobtained during three flybys indicates elevated (lunar-mare type)concentrations of Fe and Ti, though phase composition cannot beobtained (Lawrence et al., 2010). These authors suggest an ilmenitecontent of 7–18 wt% in surface materials. Lawrence et al. (2010)also proposed early crystallization of ilmenite in oxidized lunar-type magmas (cf., Usselman and Lofgren, 1976) that may lead tolate crystallization of FeO-poor pyroxenes. Alternatively, a pres-ence of Fe in surface sulfides (e.g., FeS, Sprague et al., 2009; Vern-azza et al., 2010) could also be consistent with the low-FeO contentin silicates (Blewett et al., 2009). In the latter case, Ti could be pres-ent in less oxidized minerals than ilmenite.

The presence of Fe-bearing oxides (magnetite, ulvospinel,ilmenite) in Mercury’s rocks seems unlikely because their forma-tion requires oxidized conditions that are inconsistent with FeO-poor silicates (Riner et al., 2010, this paper). In reduced (FeO-poor)magmatic conditions, Ti could be present in Mg-rich geikielite–ilmenite solid solution (MgTiO3–FeTiO3) (Riner et al., 2010), rutile(TiO2), perovskite (CaTiO3), or even sulfides. Although ilmenitehas been suggested as an opaque spectrally neutral phase to ac-count for low reflectance in visible and near-infrared ranges (Ravaand Hapke, 1987; Robinson et al., 2008; Denevi et al., 2009; Vern-azza et al., 2010), it does not provide a good match with near-infra-red emission spectra (Sprague et al., 2009). In contrast, rutile couldbe present in surface materials based on interpretation of mid-infrared emission spectra (Sprague et al., 2009) and near-infraredreflectance spectra (Vernazza et al., 2010). Titanium-rich sulfidesare present in enstatite meteorites (Keil, 1969) that are depletedin FeO-bearing minerals (Keil, 1968). The presence of spectral fea-tures of enstatite in Mercury’s mid-infrared spectra (e.g., Spragueet al., 2009) does not exclude the occurrence of Ti sulfides inFeO-depleted surface materials. This paper (Section 5) shows thatrutile, Ti2O3, and some Ti sulfides could form in FeO-poor mag-matic conditions. It is possible that a majority of Fe and Ti is in re-duced forms that reflect redox conditions during magmaticcrystallization.

The identity of the volatiles that caused the pyroclastic erup-tions is enigmatic. Kerber et al. (2009) showed that abundant mag-matic gases (up to �1.3 wt% depending on molecular mass) areneeded to form the large pyroclastic deposit at Caloris basin. Theysuggested Earth-like volatiles (H2O, CO2, CO, H2S) along with abun-dant S gases to explain the eruptions. This paper discusses the ele-mental abundances of H, S, C, Cl, and N in Mercury’s silicateinterior, the composition of possible magmatic volatiles, andchemical aspects of degassing. Then, thermochemical calculationsare used to estimate volcanic gas speciation at appropriate bulkcompositions, redox conditions (fO2), temperatures (T), and pres-sures (P).

2. Volatiles in the interior

Observational data on Mercury’s volatiles are limited to exo-spheric Na, K, and possibly, S- and H-bearing ionized moleculesformed through sputtering of surface materials (Potter and Mor-gan, 1985, 1986; Zurbuchen et al., 2008). The abundance and spe-ciation of volatiles in the interior would reflect the compositionand redox state of parent planetesimals, and specifics of accretion,impact re-processing, differentiation, and initial degassing. Thepathways of formation and evolution of Mercury remain poorlyconstrained, although diverse scenarios have been suggested, asreviewed by Solomon (2003), Taylor and Scott (2003), Benz et al.(2008), and Breuer et al. (2008). Most of these models are aimedat explaining a high core/mantle mass ratio and the observedFeO deficiency in surface materials. In this section, we discuss vol-atiles in the context of formation and evolution pathways.

2.1. Bulk composition in the framework of formation and evolutionscenarios

The high density of Mercury (5.43 g/cm3) implies a high Fe/Siratio and a significant fractionation of materials that originatedwith solar Fe/Si ratio. The fractionation could have been associatedwith nebular condensation under specific conditions, chemical gra-dients in the nebula, differential re-condensation related to chon-drule-forming and impact events, aerodynamic separation ofparticles, and preferential evaporation and/or impact removal ofsilicates. The latter two processes could have involved km-sizeplanetesimals, Moon- to Mars-sized planetary embryos, and Mer-cury itself.

Thermodynamic models for condensation of the solar nebulargas (e.g., Grossman, 1972; Grossman and Larimer, 1974; Ebel,2006) demonstrate that high temperature condensates (�1300–1450 K) have an elevated metal/silicate ratio compared to low-Tcondensates. Preferential accumulation these condensates inhigh-T inner nebula could have accounted for a high Fe/Si ratioof planetesimals that formed Mercury. High gas (H2) pressure inthe innermost nebula would also facilitate early high-T condensa-tion of Fe-metal compared to Mg-silicates (Grossman, 1972; Lewis,1972). Subsequent chondrule-forming events caused by gas shockwaves (Desch and Connoly, 2002) could have favored metal con-densation at high gas densities over formation of silicates (includ-ing Fe-silicates) through re-condensation. This process could havebeen more efficient in the innermost nebula because of high pres-sure. High-T removal of silicates through evaporation (cf. Fegleyand Cameron, 1987) from separate chondrule melts and/or nebulardust could have also contributed to a high Fe/silicate ratio on Mer-cury. Extremely reduced conditions, which may reflect elevated C/O and H2/H2O ratios in the inner nebula (e.g., Ciesla and Cuzzi,2006), also support preferential formation of Feo-rich condensates(Larimer and Batholomay, 1979), and account for the low-FeO con-tent in accreted materials. Note that condensation of nebular gas

26 M.Yu. Zolotov / Icarus 212 (2011) 24–41

with solar abundances also leads to reduced condensates withminute FeO content in silicates (e.g., Ebel, 2006; Fedkin and Gross-man, 2006; Grossman et al., 2008).

High T–P conditions in the inner solar nebula may imply a lowamount of condensed volatiles (at least H, S, and alkalis). However,elevated C/O and H2/H2O ratios in the inner nebula could have ledto accumulation of S, C, and N in reduced solids (e.g., Larimer andBatholomay, 1979). If Mercury’s primary materials were formed atC/O > �0.8, the volatiles could have been condensed in phaseswhich occur in enstatite chondrites (Mg, Ca, Cr and Ti sulfides, Cspecies, and nitrides; Keil, 1968, 1989; Brearley and Jones, 1998)(e.g., Larimer, 1975; Larimer and Batholomay, 1979; Grossmanet al., 2008). This notion and the low FeO in enstatite meteoritesled to the suggestion of significant contribution from enstatitechondrite (EC) material to the bulk composition of Mercury (Was-son, 1988; Sprague et al., 1995). Enstatite chondrites are isotopi-cally identical to Earth and could be similar to the dominantconstituent of our planet (Javoy et al., 2010). A deficiency of H2Ogas in the inner nebula caused by diffusion of water vapor towardthe water condensation front (Stevenson and Lunine, 1988; Cyret al., 1998; Ciesla and Cuzzi, 2006) and/or evaporation of con-densed C-bearing materials that migrated toward the Sun (Larimerand Batholomay, 1979; Lodders and Fegley, 1993) could accountfor even greater contribution of the EC-like material to the compo-sition of Mercury. Condensation models show that a 85% depletionof nebular H2O gas content would lead to formation of mineralsfound in enstatite chondrites (Hutson and Ruzicka, 2000; Paseket al., 2005). Mn–Cr isotopic systematics (Shukolyukov and Lugm-air, 2004) and xenon isotopes (Jee-Yon et al., 2009) in enstatitemeteorites also indicate their formation in the hot inner solar neb-ula sunward of the asteroid belt. However, the Fe/Si ratio in EC issimilar to solar abundances and a metal–silicate separation is re-quired to form Mercury from reduced condensates, which may ormay not contain minerals found in enstatite chondrites.

Disruptive collisions of planetesimals and differentiated plane-tary embryos (e.g., Wetherill, 1986; Asphaug et al., 2006) wouldprovide an efficient mechanism for metal–silicate separation.Numerical simulations of Benz et al. (1988, 2007) demonstrate thatthe density of Mercury could be accounted for by stripping of itssilicate mantle by major impact(s). Dynamical models show thatmajor collisions were more frequent within �2 AU from the Sun,and that iron meteorites could have formed through impact dis-ruptions of differentiated bodies in the terrestrial planet region(Bottke et al., 2006). High relative velocities of planetesimals andplanetary embryos at low AU should have led to strong collisions(e.g., Wetherill, 1986). Rapid accretion of planetesimals at lowAU also favored their fast differentiation (through decay of 26Al,Grimm and McSween, 1993) and made metal–silicate separationmore efficient. In addition to silicates, some volatiles could havebeen lost through decompression and boiling of silicate fluids incollisions (Melosh et al., 2004; Asphaug et al., 2006) followed byrapid quenching before re-condensation of volatiles.

The metal-rich (60–70 vol.%) CB carbonaceous chondrites (e.g.,Weisberg et al., 2001; Krot et al., 2003) could have formed throughcollisions of planetary embryos (Campbell et al., 2002; Rubin et al.,2003; Krot et al., 2005). CB chondrites are strongly depleted inmoderately volatile lithophile elements (Na, K, Mn, Zn, Se, etc.)and have high abundances of lithophile refractory elements (Weis-berg et al., 2001; Krot et al., 2003). They have low-FeO contents insilicates (<4 mol.%), and the high Feo/silicate ratio makes themplausible analogs for Mercury’s bulk composition (Taylor and Scott,2003). Several other metal-rich meteorites (CH chondrites, palla-sites, ureilites, mesosiderites, and IVA iron meteorites) with signsof rapid cooling could have formed through major collisions as well(Keil et al., 1994; Yang et al., 2007; Scott, 2007). Signs of a major,early impact in the Shallowater enstatite achondrite (aubrite) (Keil,

1989; Keil et al., 1994) demonstrate that extremely reduced plane-tesimals were also involved in these processes. Independent of theorigin of metal-rich planetesimals, an aerodynamic separation ofdenser and/or larger planetesimals could have led to an elevatedFeo/silicate ratio in the high-P inner nebula and may have contrib-uted to the formation of metal-rich Mercury (Weidenschilling,1978).

Our working hypothesis is that Mercury formed from reduced,high-T, high-P, anhydrous nebular condensates that were affectedby removal of silicates in major collisions occurring before and/orafter formation of the planet. Although enstatite chondrites andCB-like chondrites provide insights about reduced nebular solidsand metal–silicate fractionation through impacts, it is possible thatnone of these groups of meteorites represent the composition ofMercury. In particular, FeO-poor and Feo-rich condensates of thesolar gas composition could be among Mercury’s protoliths, thoughthey do not contain minerals found in enstatite chondrites. Like-wise, the presence of hydrated carbonaceous clasts in CB chon-drites (Bonal et al., 2010) and their oxygen isotopic composition(Clayton and Mayeda, 1999; Rubin et al., 2003; Ivanova et al.,2008) indicate their formation in the main asteroid belt.

2.2. Hydrogen and water

Hydrogen-bearing ionized molecules H2Oþ; H2O; OHþ� �

andH2S+ have been tentatively identified in Mercury’s exosphere byMESSENGER (Zurbuchen et al., 2008). These molecules could haveresulted from sputtering of surface materials, which may includewater ice (Zurbuchen et al., 2008) and/or protons implanted withthe solar wind. Water ice (Slade et al., 1992) may be associatedwith radar-bright craters in the north polar region (Harmon andSlade, 1992; Harmon et al., 2001) and could represent moleculestrapped after cometary/meteoritic impacts (Moses et al., 1999).

Although endogenic sources of Mercury’s H are not excluded,the interior of Mercury is likely to be depleted in H and H2O. Thelow-FeO content in surface silicates and suggested FeO-poor man-tle (Robinson and Taylor, 2001) are indicators of water scarcity thatalso implies reduced conditions. Water is the major oxidizing agentin the solar nebula and inside asteroids and planetary bodies (Le-wis and Prinn, 1984; Dreibus and Wänke, 1989; McSween and Lab-otka, 1993; Arculus and Delano, 1987; Kasting et al., 1993; Cieslaand Cuzzi, 2006), and reduced rocks often reflect processes that oc-curred under deficiency of water. The low-FeO content in surfacesilicates suggests a small (if any) contribution from processes suchas aqueous alteration of parent planetesimals, accretion of icy andhydrated planetesimals, and Feo–H2O interaction in the interior.

Outward diffusion of nebular water vapor and a limited pene-tration of H2O-bearing planetesimals into the inner solar nebulacould have created H2O-deficient conditions (e.g., Stevenson andLunine, 1988; Ciesla and Cuzzi, 2006). Together with high temper-ature, these conditions would prevent gas–solid type hydration(Lewis, 1972, 1988). A metal–silicate separation through major col-lisions (Benz et al., 1988, 2007) implies a loss of H and water(Asphaug et al., 2006). Even if some water had accreted on Mer-cury, it would have been consumed in oxidation reactions withFeo-metal (e.g., Dreibus and Wänke, 1989), elemental C (Hollowayand Jakobsson, 1986; Kadik et al., 2006) and other reduced solids(Fe3P, MgS, etc.) leading to the formation and escape of H2 andCO. If Mercury’s mantle formed through upward crystallization ofa magma ocean (Brown and Elkins-Tanton, 2009), water mighthave been expelled toward the surface and partially degassed atlow-P conditions. Low solubility of water in Feo-metal saturatedmagmas (Kadik et al., 2004 and refs. therein) would also favoredwater loss. Finally, possible impact removal of the upper part ofthe mantle (Benz et al., 1988, 2007) would have led to major lossof water and H.

Temperature, K1200 1300 1400 1500 1600 1700 1800

log

fO2,

bar

-18

-16

-14

-12

-10

-8

IW-3

IW-6

IW

IRI

IW: 2Fe0.947O = 1.894Feo + O2IRI: 2FeTiO3 = 2TiO2 + 2Feo + O2ITOI: 2FeTiO3 = Ti2O3 + 2Feo + 1.5O2RTO: 4TiO2 = 2Ti2O3 + O2

RTO

ITOI

CB/CH

M-A

EC2 wt% FeO6 wt% FeO

Fig. 1. Oxygen fugacity at reduced conditions that corresponds to some mineralbuffers and suggested analogs for Mercury’s mantle melts. The dotted curves referto conditions of 3 and 6 log fO2 units below the iron–wüstite buffer (IW, Eq. (1) inTable 1). IRI, ilmenite–rutile–iron buffer (Eq. (2)); ITOI, ilmenite–titanium oxide–iron buffer (Eq. (3)); RTO, rutile–titanium oxide buffer (Eq. (4)). The gray curvesrepresent melts with 2 wt% and 6 wt% FeO based on the Ariskin et al. (1993) meltmodel for 1500–1700 K. The curves marked CB/CH and M-A correspond to partialmelts of CB/CH chondrites and Mercury’s mantle composition from Morgan andAnders (1980), respectively. The circle symbol refers to a partial melt of an enstatitechondrite.

M.Yu. Zolotov / Icarus 212 (2011) 24–41 27

The discovery of H-bearing glasses (4–46 ppm H2O; Saal et al.,2008) and phosphates in lunar basalts (Boyce et al., 2010) impliesthat some water either survived a Moon-forming collision or wastrapped after the event. The estimated pre-eruptive water contentin lunar glasses (�700 ppm; Saal et al., 2008) could be consideredas an upper limit for water content in Mercury’s rocks becausemelt is enriched in water compared to mantle source rocks, and be-cause of the possible water-bearing nature of proto-Earth.

2.3. Mantle oxidation state

Oxygen fugacity (fO2) of lunar mare basalts (18–21 wt% FeO;Basaltic Volcanism, 1981) is 0 to �2 log units below the iron–wüs-tite buffer (IW, Eq. (1) in Table 1) (Wadhwa, 2008). The lesser FeOcontent on Mercury compared to lunar values implies more re-duced conditions. For 2–6 wt% FeO in silicate rocks, Malavergneet al. (2010) estimated Mercury’s mantle fO2 as 3–6 log units belowIW (IW�3 to IW�6) (Fig. 1). They assume ideal melt in which FeOactivity is equal to concentration in equilibrium with Feo-metal.

Consideration of the non-ideality of silicate melts saturatedwith Feo-metal using empirical equations from Ariskin et al.(1993) leads to more definitive fO2 evaluations. These equationsdescribe relationships between fO2, FeO molar content, and con-centrations of other major oxides in melts. For plausible magmaformed through partial melting of Mercury’s mantle compositions

Table 1Chemical equilibria and their equilibrium constants.

Reaction Log K at 1 bar

298 K 1500 K 1700 K

(1) 2Fe0.947O (wüstite) = 1.894Feo + O2(g); IW buffer �85.8 �11.6 �9.53(2) 2FeTiO3 (ilmenite) = 2TiO2 (rutile) + 2Feo + O2(g); IRI buffer �93.6 �12.9 �10.7(3) 4/3FeTiO3 = 2/3Ti2O3 + 4/3Feo + O2(g); ITOI buffer �103 �14.1 �11.6(4) 4TiO2 = 2Ti2O3 + O2(g); RTO buffer �121 �16.4 �13.2(5) 2MgS + O2(g) = 2MgO + S2(g) 65.8 11.9 10.3(6) 2CaS + O2(g) = 2CaO + S2(g) 32.6 5.90 5.1(7) 2MnS + O2(g) = 2MnO + S2(g) 36.6 6.42 –(8) TiS2 + O2(g) = TiO2 + S2(g) 72.5 14.1 –(9) TiS2 + 0.75O2(g) = 0.5Ti2O3 + S2(g) 42.2 10.0 –(10) 2TiS + 1.5O2(g) = Ti2O3 + S2(g) 139 23.6 –(11) 2MgS + 2SiO2 + O2(g) = 2MgSiO3 + S2(g) 77.3 14.1 12.4(12) 2MgS + SiO2 + O2(g) = Mg2SiO4 + S2(g) 76.2 13.8 12.0(13) 2CaS + 2SiO2 + O2(g) = 2CaSiO3 + S2(g) 62.9 12.1 10.6(14) 2CaS + SiO2 + O2(g) = Ca2SiO4 + S2(g) 55.4 10.8 9.57(15) 2MnS + 2SiO2 + O2(g) = 2MnSiO3 + S2(g) 45.5 8.03 6.93(16) 2MnS + SiO2 + O2(g) = Mn2SiO4 + S2(g) 45.1 7.18 6.66(17) 2TiS2 = 2TiS + S2(g) �54.3 �3.69 –(18) 2FeS = 2Feo + S2(g) �50.0 �4.8a –(19) MgS + FeO = MgO + Feo + 0.5S2(g) �11.1 �0.08 0.26(20) MgS + FeTiO3 = MgO + TiO2 + Feo + 0.5S2(g) �13.9 �0.52 �0.23(21) TiS2 + 2FeTiO3 = 3TiO2 + 2Feo + S2(g) �21.1 1.14 –(22) 1.5MgS + FeTiO3 = 0.5Ti2O3 + 1.5MgO + Feo + 0.75S2(g) �27.7 �1.65 �0.96(23) MgS + 2TiO2 = Ti2O3 + MgO + 0.5S2(g) �27.7 �2.73 �1.46(24) Mg2SiO4 + SiO2 = 2MgSiO3 1.16 0.31 0.36(25) SiO2 = Sio + O2(g) �150 �22.4 �18.7(26) Co + FeO = Feo + CO(g) �20.0 2.45 3.09(27) Co + 0.5O2(g) = CO(g) 24.0 8.47 7.99(28) Co + FeTiO3 = Feo + TiO2 + CO(g) �22.8 2.0 2.62(29) Co + 0.667FeTiO3 = 0.667Feo + 0.333Ti2O3 + CO �27.4 1.43 2.21(30) Co + 2TiO2 = Ti2O3 + CO �39.6 0.27 1.39(31) Co + SiO2 = Sio + 0.5O2(g) + CO(g) �126 �14.0 �10.7(32) TiN + O2(g) = TiO2 + 0.5N2(g) 102 16.6 14.2(33) 2TiN + 1.5O2(g) = Ti2O3 + N2(g) 143 25.1 21.7(34) 2NH3(g) = N2(g) + 3H2(g) �5.75 8.41 8.87(35) H2O(g) = 0.5O2(g) + H2(g) �40.0 �5.73 �4.67

Notes: The equilibrium constants are calculated using thermodynamic properties of corresponding solids, liquids, and gases mainly from Gurvich et al. (1989–1994), Pankratz(1987), Robie and Hemingway (1995), and Barin (2004). Although O2 gas is not present in magmas, it is used to formally quantify fO2 at chemical equilibria. In reality, some ofthe species (FeO, MgS, SiO2, etc.) could be components of magma, a phase or a solid solution. In all equations, Co, Feo, TiO2, FeTiO3, and MgSiO3 represent graphite, iron metal,rutile, ilmenite, and enstatite, respectively.

a Extrapolated from T < 1400 K.

28 M.Yu. Zolotov / Icarus 212 (2011) 24–41

(Taylor and Scott, 2003), concentrations of 2–6 wt% FeO result infO2 of IW�2.3 to IW�3.4 at 1600–1700 K (Figs. 1 and 2). Specifi-cally, the redox state of a partial melt of the Indarch (EH4) enstatitechondrite at 1700 K with 0.25 wt% FeO (McCoy et al., 1999; Taylorand Scott, 2003) corresponds to IW�5.3 (log fO2 = �14.9). Melt de-rived from a mantle composition consisting of the silicate compo-nents of metal-rich CB/CH-like chondrites (Krot et al., 2001) with3.2 wt% FeO (Taylor and Scott, 2003) corresponds to IW�3(log fO2 = �13.5) at 1600 K, and IW�3.1 (log fO2 = �12.6) at1700 K. Similar results are obtained for melt with 3.7 wt.% FeO(Taylor and Scott, 2003) derived from the mercurian mantle com-position of Morgan and Anders (1980): IW�2.9 (log fO2 = �13.4) at1600 K, and IW�3 (log fO2 = �12.5) at 1700 K.

These estimates show that fO2 in Mercury’s magmas could beclose to the conditions determined by mineral equilibria that in-clude Ti(IV) and Ti(III) oxides (ITOI and RTO buffers, Figs. 1 and2). The conditions of the IRI buffer (Eq. (2) in Table 1) correspondto high FeO melt contents that are not consistent with remotesensing observations of surface materials (Fig. 2) (except, possibly,Lawrence et al., 2010, interpretation). In other words, the presenceof ilmenite contradicts low FeO concentration in silicates, and pureilmenite may not be present in FeO-poor Mercury’s rocks (Rineret al., 2010).

The possibility of crystallization of ilmenite before pyroxene at�IW�0.5 (Usselman and Lofgren, 1976) could be reconciled withthe low-FeO content in surface silicates, and the lunar-type oxi-dized mantle cannot be excluded (Lawrence et al., 2010). However,

1700 K

0 5 10 15 20

log

fO2,

bar

-18

-16

-14

-12

FeO mole %

1600 K

1500 K

IW-3

IW-4

a b

IW-2

Fig. 2. Oxygen fugacity as a function FeO concentration in silicate melts. The data are obtaet al., 1993). The curves refer to magmas formed through partial melting of plausible Mecurves represent partial melts of an enstatite chondrite (McCoy et al., 1999), averagedMercury’s mantle (Morgan and Anders, 1980), respectively. In (a), thin dashed curves shoIn (b), the data are presented relative to the IW buffer for low-FeO contents. The dottedwt% FeO in the melt derived from metal-rich chondrites.

Table 2Abundances of elements in some chondrites and aubrites.

S, wt%

CI carbonaceous chondrites 5.41H ordinary chondrites 2.0L ordinary chondrites 2.2LL ordinary chondrites 2.1EH enstatite chondrites 5.6EL enstatite chondrites 3.1Aubrites (enstatite achondrites) 0.24–0.86a

CB chondrites 0.52e

Notes: Chondritic data are from Lodders and Fegley (1998).a Gibson et al. (1985).b Grady et al. (1986).c Garrison et al. (2000), without Shallowater aubrite which contains �400 ppm Cl.d Krot et al. (2003) and Ivanova et al. (2008).e Rubin et al. (2003) and Ivanova et al. (2008).

the redox conditions of �IW�0.5 imply a significant oxidation ofthe innermost solar nebula through incorporation of water-bearingplanetesimals migrated from larger heliocentric distances (Cieslaand Cuzzi, 2006; Grossman et al., 2008). Although this migrationcould explain elevated oxidation states of some materials (e.g.,Type II chondrules with high FeO content) formed in the asteroidbelt, these planetesimals are unlikely to have survived high-T con-ditions in the innermost nebula. Considering limited contributionfrom water-bearing bodies to the composition of Mercury (O’Brienet al., 2006), a reduced mantle seems likely. Metal–silicate frac-tionation through collisions, gas drag, and igneous differentiation(see Section 1) would not have changed the reduced nature ofmaterials that formed Mercury.

2.4. Sulfur, carbon, chlorine, nitrogen, and alkalis

2.4.1. SulfurEstimations of Mercury’s bulk S content strongly depend on the

formation scenario. Accretion of Mercury from EC-like materialswould lead to solar abundances of S (Table 2). In addition to FeS,Mg-, Ca- and Cr-sulfides are the likely products of nebular pro-cesses in H2O-depleted inner nebula (Cyr et al., 1999; Paseket al., 2005; Grossman et al., 2008). Conversely, formation of Mer-cury from metal-rich products of planetary embryo collisions couldlead to an S-depleted planet, considering the low S content in CBchondrites (Table 2). However, independent of the formation sce-nario and redox state of accreted solids, Mercury’s silicate mantle

0 1 2 3 4 5 6

log

fO2 r

elat

ive

to IW

-6

-5

-4

-3

-2

-1

FeO mole %

1500 K

1700 K

FeO wt %1 2 765

RTO, 1700 K

RTO, 1500 K

ITOI, 1700 K

ITOI, 1500 K

IRI, 1700 K IRI, 1500 K

ined with empirical equations for silicate systems saturated with Feo-metal (Ariskinrcury’s mantle compositions from Taylor and Scott (2003). Solid, dashed, and dottedsilicate components of metal-rich CB/CH like chondrites (Krot et al., 2001), and aw fO2 values relative the IW buffer at specific temperatures and melt compositions.

lines refer to fO2 buffers depicted in Fig. 1. The upper horizontal axis corresponds to

C, wt% Cl, ppm N, ppm

3.45 700 31800.21 140 480.25 270 430.31 200 700.39 570 4200.43 230 2400.02–0.14b 1–31c 10–100b

0.37e – 25–106d

mol

e ra

tio

0.1

1

10

EC gas

a

M.Yu. Zolotov / Icarus 212 (2011) 24–41 29

may only contain accessory sulfides because they would likelyaccumulate in the core (cf., Riner et al., 2008). Besides FeS, thisaccumulation could have affected Mg-, and Ca-sulfides, if theywere present (McCoy et al., 1999). The effect of subsequent majorcollisions (Benz et al., 1988, 2007) on concentration of S in themantle is not clear.

C/S Mole ratio

0.1 1 10

Cl/S

0.001

0.01Sun

CI

KilaueaEL

EH

C/S mole ratio0.1 1 10

N/S

mol

e ra

tio

0.001

0.01

0.1

1

10

Sun

CI

ELEH

20

CB

LLLH

Mercury's mantle?

LLL

H

Mercury's mantle?

EC gasb

Fig. 3. Normalized elemental abundances in chondritic materials, gases releasedfrom heated enstatite chondrites (EC gas, Muenow et al., 1992), and Kilauea 1918summit eruption (Symonds et al., 1994). Data for carbonaceous (CI, CB), ordinary (H,L, LL) and enstatite chondrites (EH, EL) are from Table 2, and solar photosphere dataare from (Asplund et al., 2005). The gray ovals show possible elemental ratios inMercury’s mantle and magmas.

2.4.2. CarbonThe solar abundance of C is an order of magnitude higher than

the abundance in CI carbonaceous chondrites and is about two or-ders of magnitude higher than in ordinary, enstatite, and CB chon-drites (Table 2; Asplund et al., 2005; Lodders and Fegley, 1998).This corresponds to a decrease in C content in the main asteroidbelt where carbonaceous bodies become less abundant towardthe Sun (Gradie et al., 1989). The limited formation of C-bearingsolids in the inner solar nebula could be related to formation ofCO from condensed organic species transported from higher helio-centric distances (AU > �5) (Cuzzi et al., 2003). Reduced nebulaconditions have not cased accumulation of abundant graphite inenstatite chondrites, though these meteorites are slightly enrichedin C compared to ordinary chondrites (Table 2). Nebular condensa-tion models of Lewis (1988) do not predict C solids formed at highT–P conditions at the heliocentric distance of Mercury. It followsthat solids condensed in the vicinity of the Sun could be C depleted.Nebular condensates (and Mercury) could be C-rich only if con-densed at higher C/O ratios compared to enstatite chondrites, thatis at C/O > �0.83 when high-T condensates become rich in graph-ite, TiC, and SiC (e.g., Larimer and Batholomay, 1979; Grossmanet al., 2008). Although such enrichment could be caused by mas-sive infall of C-rich solids into the inner solar nebula (Loddersand Fegley, 1993) no meteorite rich in these phases is found. Dy-namic models show that only a limited fraction of carbonaceousplanetesimals formed at AU > 2 could have accreted on Mercury(O’Brien et al., 2006). Major collisions of planetary embryos mayhave no major effect on C abundance, as implied from similar Cabundance in CB and ordinary chondrites (Table 2). Although C/Sratio in Mercury could vary from that in enstatite and carbona-ceous chondrites (Fig. 3), lower ratios similar to EC seem probable.

During differentiation, accreted organic carbon could have beenlost as CO and hydrocarbon gases formed through thermal process-ing, oxidation, and graphitization of organics. However, in anhy-drous high-P igneous systems, the majority of C could have notdegassed because of precipitation of low-solubility graphite (seeSection 4). Independent of the origin of Mercury and bulk C con-tent, a major portion of accreted C could be sequestered in the me-tal core. Although reducing conditions in the mantle may imply astable existence of native C species, Mercury’s silicate rocks andmagmas may not be enriched in C compared to other terrestrialplanets.

2.4.3. ChlorineThe abundance of Cl in Mercury’s mantle rocks and magmas is

hard to constrain because of the unknown bulk planetary contentand limited solubility data for dry reduced melts. On the one hand,a low Cl content in ordinary chondrites compared to CI chondrites(Table 2) and solar abundances implies Cl loss in high-T processesin the inner nebula and thermally processed planetesimals. Thesecondary nature of all Cl phases in ordinary and carbonaceouschondrites also suggests a low-T (<200 K) condensation of Cl (asHCl hydrates, Zolotov and Mironenko, 2007) which may accountfor Cl depletion in the inner Solar System. On the other hand, re-duced conditions in the inner nebula could have favored high-Tformation of halides consistent with elevated Cl abundances inEH enstatite chondrites (Table 2; Rubin and Choi, 2009). A defi-ciency of water vapor and elevated P–T conditions in the inner neb-

ula could have led to formation of metal halides, as exemplified byreactions

2HClðgÞ þ Na2Oðin gas or solidsÞ ! 2NaClþH2OðgÞ

HClðgÞ þ NaAlSi3O8ðsolidÞ ! NaClþ 0:5Al2O3 þ 3SiO2

þ 0:5H2OðgÞ

These interactions of HCl(g) could have led to initial Cl abundanceson Mercury as high as in CI chondrites and the Sun.

Highly metamorphosed EL6 chondrites are depleted in Cl com-pared to EH4 chondrites (Rubin and Choi, 2009). One can suggestsome loss of Cl during thermal/impact metamorphism after accre-tion of Mercury and/or planetary embryos. Chlorine and alkali ele-ments could have been accumulated in the upper horizons ofMercury through magmatic differentiation possibly involving up-ward crystallization of a magma ocean (Brown and Elkins-Tanton,2009). Subsequently, these elements could have been removedthrough impact stripping. Despite the uncertainties, a Cl/S mole ra-tio of �0.01 characterizes many solar system objects (Fig. 3a) andcould be used a proxy for Mercury’s mantle rocks and magmas.

2.4.4. NitrogenChondrites are strongly depleted in N compared to cometary

materials (Jessberger, 1999; Bockelée-Morvan et al., 2004) and so-lar abundances (Asplund et al., 2005). This indicates a major devol-atilization of N-bearing compounds (organics, CHON particles, andNH3 hydrate) in the inner solar nebula. The position of Mercury inthe Solar System implies a limited delivery of N in organic-bearingplanetesimals and planetary embryos (O’Brien et al., 2006). How-ever, Mercury’s rocks may not be depleted in N compared to otherterrestrial planets because of possible accretion of refractorynitrides (TiN, Si3N4) and N-bearing metal alloys condensed fromreduced nebular and/or impact-generated gases. In enstatite chon-drites, the presence of reduced N solids (Keil, 1968, 1989; Rubinand Choi, 2009) accounts for significantly larger N content com-pared to ordinary and some carbonaceous (CB, CV, CO) chondrites

30 M.Yu. Zolotov / Icarus 212 (2011) 24–41

(Table 2; Lodders and Fegley, 1998). The association of N with themetal phase and its high-T (�1000 �C) release from a CB/CH mete-orite (Ivanova et al., 2008) may indicate trapping of degassed Nthrough condensation of metal from impact gases (Sugiura et al.,1999). It follows that major collisions involving differentiatedplanetary embryos and Mercury may not have resulted in a majorN loss. Abundance and speciation of N in enstatite and CB/CH chon-drites could be used as a proxy for Mercury’s initial materials (Ta-ble 2 and Fig. 3b).

The high stability of nitrides during metamorphism at extre-mely reduced conditions (Petaev and Khodakovsky, 1986; Fogelet al., 1989) could have prevented major loss of N before the igne-ous differentiation of Mercury and planetary embryos. The highstability of nitride complexes in reduced, H-free and high-P mag-mas (see Section 4) could also have limited N degassing from aputative magma ocean, as suggested for the terrestrial counterpart(Libourel et al., 2003). The presence of N in meteoritic Feo alloy alsoimplies some sequestering of N into the metal cores on planetaryembryos and Mercury (cf., Fegley, 1983).

2.4.5. AlkalisAlkalis could have not been abundant in high-T nebular conden-

sates and alkali oxides may compose only a few tenths of weightpercent of materials accreted on Mercury (Lewis, 1988). Alkaliscould have been depleted through high-T evaporation in the innernebula (through chondrule formation and/or collisions), afteraccretion (Fegley and Cameron, 1987), or stripped by impacts to-gether with a feldspathic crust (Benz et al., 1988; Lewis, 1988).Although the presence of Na and K in the exosphere (Potter andMorgan, 1985, 1986; Domingue et al., 2007; McClintock et al.,2009) is consistent with feldspar-bearing surface rocks, it doesnot imply an alkali-rich mantle. Enhanced K emission over the Cal-oris basin (Sprague et al., 1990) may indicate a low degree of melt-ing rather than alkali enrichment in the mantle. Note that theapparent lack of crustal subduction implies limited recycling ofalkalis and other incompatible elements in igneous processes. Adeficiency of alkalis in the mantle also implies elevated liquidustemperatures.

3. Temperature and composition of mantle magmas

Temperature and composition of magmas mainly depend on theuncertain mineralogy and volatile content in Mercury’s mantle.Petrologic experiments demonstrate that dry mantle rocks meltat temperatures up to several hundred degrees higher than H2O-rich mantle substrates (e.g., Green, 1975; Righter et al., 2006).Dry lavas on the Moon and Io are hotter than Earth’s counterparts(Basaltic Volcanism, 1981; McEwen et al., 1998; Keszthelyi et al.,2007) and a deficiency of H2O on Mercury would contribute to ele-vated melt temperatures affecting the composition of silicate melt.

Numerical modeling of partial melting using plausible mantlecompositions of Mercury demonstrates the formation of mafic toultramafic melts with low-FeO contents (Taylor and Scott, 2003;Section 2.3). Experimental melting of the Indarch (EH4) enstatitechondrite (McCoy et al., 1999) suggests magma temperatures of1670–1720 K at �20–50% partial melting. The high melt tempera-ture reflects low abundances of FeO, alkalis, and water. The basalticnormative composition of those melts (�34 vol% plagioclase, �42%enstatite, �15% diopside, and �6% sulfides) has been suggested asa proxy for Mercury’s lavas (Burbine et al., 2002). Note that the useof enstatite achondrites (aubrites, Watters and Prinz, 1979) ascompositional analogs for Mercury’s igneous rocks is limited bytheir residual nature. Aubrites do not reveal a clear relationshipwith enstatite chondrites (Keil, 1989) and could have formed afterremoval of basaltic melts (e.g., Burbine et al., 2002). Despite forma-tion of HED achondrites from anhydrous reduced magmas, they

have high FeO/MgO molar ratios (�0.3–2, Mayne et al., 2009; McS-ween et al., 2010) and may not be considered as close analogs forMercury’s igneous melts and cumulates.

Low-FeO content in Mercury’s surface silicates also implies aFeO-poor mantle (Robinson and Taylor, 2001) that would producehigh-T melts. A mantle rich in refractory oxides is consistent withthe non-global appearance of volcanic features and spectral signsof Mg- and Ca-rich silicates in surface materials (see Section 1).The refractory mineralogy may represent accreted Fe-depleted sil-icates (enstatite, forsterite, diopside, and feldspars) and remainedin the mantle after segregation of Feo–FeS melts into the core.Alternatively, the mantle could have formed through upward crys-tallization of a magma ocean leading to a FeO-rich upper layerthough fractional crystallization (Brown and Elkins-Tanton,2009). Later on, this layer could have been stripped by impact(s)(Benz et al., 1988) or even sunk into the lower mantle (Brownand Elkins-Tanton, 2009). The putative evaporation of an earlymantle also results in refractory rocks (Cameron, 1985; Fegleyand Cameron, 1987).

The interpretation of MESSENGER neutron spectrometer datasuggest Mercury’s rocks similar to lunar mare basalts from Luna24 and Luna 16 samples (Lawrence et al., 2010). These data areinconsistent with the melt models for FeO-poor protoliths (Taylorand Scott, 2003; Burbine et al., 2002). Although neutron spectrom-eter data would revolutionize our views on Mercury’s magmatism,they need to be confirmed with independent orbital or in situmeasurements.

4. Solubility of S, C, Cl, and N in reduced anhydrous magmas

The specifics of degassing during eruptions and the abundancesof volatiles in magma are affected by their solubility, which is non-linearly dependent on T, P, and fO2. Reduced melts could be en-riched in S, Cl, and N because of the formation of non-volatile meltcomplexes without oxygen. However, some volatiles could beundersaturated because of limited abundances in mantle sourceregions.

Petrological and material science experiments performed in sil-icate melts at fO2 below IW�2 (Fincham and Richardson, 1954; Fo-gel et al., 1996; McCoy et al., 1999; Siebert et al., 2004; Berthetet al., 2009) reveal S solubility over an order of magnitude higherthan at more oxidizing conditions (Mavrogenes and O’Neill,1999; Holzheid and Grove, 2002) (Fig. 4a). The highest S solubilityhas been reported by Malavergne et al. (2007) (up to 10 wt% at10 kbar and IW�7) and Fogel et al. (1996) (7.8 wt% in CaS-satu-rated enstatite chondrite melts at 1 bar and 1673 K). In contrastto more oxidized systems, S solubility is higher in FeO-poor meltsand could be caused by formation of metal (Mg, Ca, Mn, Ti, etc.)sulfide melt complexes (Fogel et al., 1996; McCoy et al., 1999; Fo-gel, 2005). The anomalous S solubility in reduced melts is illus-trated by the presence of 0.8–2.5 wt% S in basaltic glasses inaubrites, which are the most S-rich samples among achondritesand silicate planetary rocks (Fogel, 2005). These data show that Scontent in Mercury’s igneous materials would indicate redox con-ditions in magma and mantle source regions.

In anhydrous reduced magmas, the solubility of C is mainly con-trolled by the equilibria of C-bearing solids (e.g., graphite) with CO(Eqs. (26) and (27)). Low fO2 and temperature and high pressure fa-vor the stability of graphite. The low solubilities of CO, CO2 (Panet al., 1991) and graphite, which precipitates, account for the lowsolubility of C in magma (<�0.1 wt%, Holloway and Jakobsson,1986; Kadik et al., 2004). Despite possible formation of SiC meltcomplexes (Kadik et al., 2004), high-P (1–5 GPa) solubility of Cin graphite-saturated systems decreases with decrease of fO2

owing to precipitation of C-bearing solids (graphite, diamond,carbides, metal alloys) (Holloway et al., 1992; Kadik et al., 2004;

log fO2 relative to IW-8 -6 -4 -2 0 2

log

N S

olub

ility,

wt%

-6

-5

-4

-3

-2

-1

0

1 bar, 1700 K

10-30 kbar1673-1973 K

25 kbar20 kbar15 kbar10 kbar1673K

log fO2

relatively to IW

-8 -6 -4 -2 0 2

S So

lubi

lity,

wt%

0.1

1

10 9-27 kbar data: Holzheild and Grove, 2000 Malavergne et al., 2007 Berthet et al., 2009

1 bar data: McCoy et al., 1999

a

b

1 bar, 1573 K

3 kbar1 kbar1523 K

Fig. 4. Experimental data on solubilities of S (a) and N (b) in reduced silicate melts.The plot (a) is a modified version of a figure from Berthet et al. (2009) andrepresents data for temperatures �1500–2300 K and pressures from 1 bar to15 kbar. In (b), the circle dotted symbols represent data for basaltic FeO-freemagmas at �1700 K and 1 bar from Libourel et al. (2003). The solubility data atIW+0.9 and 1 bar is from Miyazaki et al. (2004). The high-P data for Na2O–SiO2

melts from Shilobreeva et al. (1994), Mysen et al. (2008), and Roskosz et al. (2006)are presented by squares, triangles, and horizontal dashed lines, respectively.Oxidation conditions in (Roskosz et al., 2006) are not well justified. The conditionsof the IW buffer correspond to Eq. (1). The plots illustrate that oxidation of magmaat fO2 < IW�2 would cause degassing of S and N. If high pressure favors N solubilityat fO2 < IW�1, N2 decompression could cause major N degassing.

M.Yu. Zolotov / Icarus 212 (2011) 24–41 31

Hirschmann and Withers, 2008). The solubility of C also decreaseswith increasing pressure (at least in 1–5 GPa range; e.g., Hirsch-mann and Withers, 2008). However, low-P near-surface conditionsfavor formation of CO through oxidation of solid C species, if theyare present in magma (see Section 5).

The behavior of Cl in reduced anhydrous magmas is not wellknown (e.g., Carroll and Webster, 1994; Aiuppa et al., 2009). Chlo-rine could be associated with alkali metals (especially in alkalinemagmas, Kogarko, 1974), and Ca, Mg, and Fe that may form meltcomplexes with a salt-like ionic bonding (Carroll and Webster,1994; Evans et al., 2008). In S-rich melt systems, Cl–S complexesmay also form by analogy with diverse Cl–S species expected inmagmatic gases (Section 6). These O-free associations imply ele-vated solubility of Cl in melts saturated with respect to Feo-metal.Although high Cl solubility is consistent with some experiments(Nicholis and Rutherford, 2009), target research remains to beperformed.

The solubility of N in silicate melts increases dramatically atfO2 < IW�1 (Mulfinger, 1966; Libourel et al., 2003) (Fig. 4b). Fogel(1994) reported up to 0.5 wt% N solubility in synthetic enstatitechondrite melts at IW�10, 1773 K, and 1 bar, which is �5 ordersof magnitude higher than at oxidizing conditions. For FeO-freebasaltic melts, Libourel et al. (2003) determined �0.2 wt% N solu-bility at IW�8, 1700 K, and 1 bar. The increase in N solubility be-low IW�1 in H-free melts is explained by the formation ofnitride complexes, such as Si3N4 (Ito and Fruehan, 1988; Libourelet al., 2003; Miyazaki et al., 2004). Experiments performed atfO2 > IW show that high pressure favors the stability of various N

complexes (SiON, Roskosz et al., 2006; NHþ2 ; NH�2 ; NH3, Mysenet al., 2008). These data reveal N solubility as high as 0.7–1.6wt% at pressures of 20–30 kbar (Fig. 4b), which may representmantle source regions 190–280 km below Mercury’s surface.Although the effect of pressure on N solubility in reduced(<IW�1) melts has not been explored experimentally, the sugges-tion is solubility will increase with pressure (Libourel et al., 2003).The high solubility of N in reduced and/or high-P melts implies asignificant N enrichment in magmas compared to mantle sourcerocks, which may also be N-rich.

5. Chemical pathways of magma degassing

Changes in pressure, composition or redox state of magma mayaffect speciation of volatiles in the melt and cause their separationin the gas phase. Here we consider degassing mechanisms of indi-vidual volatiles and discuss devolatilization of plausible Mercury’smagmas with different redox states.

5.1. Sulfur

Pressure does not have a definitive effect on S solubility (Ber-thet et al., 2009; Fig. 4a) and decompression during ascent of re-duced magma would not cause major degassing. A decrease insolubility would cause formation of sulfide liquid(s) (McCoyet al., 1999; Berthet et al., 2009) rather than degassing. Major Sdegassing from synthetic silicate and enstatite chondrite meltshas not been reported at 1 bar (Fogel et al., 1996; McCoy et al.,1999; Fincham and Richardson, 1954). Partial melting of enstatitechondrite samples in a vacuum at 1200–1600 K (Muenow et al.,1992) did not produce much S gases compared to C-, N-, and Cl-species (see Section 6). A high S content in glasses in aubrites (Fo-gel, 2005) also does not indicate S degassing at suggested low-Pconditions on a parent body (cf., Wilson and Keil, 1991). Relativelyoxidized (IW to IW�2) lunar mare basalts do not reveal signs ofpyroclastic activity related to S. These observations are consistentwith low fS2 values above sulfides that coexist with metal oxides,silicates, and Feo-metal (Figs. 5 and 6). Although decompositionof sulfide melt complexes through near-surface decompression ofreduced magmas could be responsible for generation of some S-bearing gases, the degassing could be inefficient compared to moreoxidized magmas on Io (Zolotov and Fegley, 1999). Decompressionof reduced magma (fO2 < IW) is unlikely to produce �1.3 wt% S gasneeded to form a large pyroclastic deposit at Caloris (Kerber et al.,2009).

Oxidation of S-saturated magma provides a much more power-ful degassing mechanism than plain decompression. An increase infO2 from IW�6 to IW�2 lessens S solubility from 3–8 wt% to <0.3wt% (Fig. 4a). The decrease in solubility is accounted for by oxida-tion of metal sulfide melt complexes, as exemplified by net reac-tions (5)–(10) in Table 1. The lower stability of metal sulfidecomplexes at higher fO2 is consistent with correspondingly higherfS2 values in sulfide–oxide and sulfide–silicate equilibria (Eqs. (5)–(16)) (Figs. 5 and 6). These equilibria show that temperature lesseffect on sulfide stability (fS2) than fO2.

Although sulfide–oxide and sulfide–silicate equilibria calcu-lated at unity activities cannot be used to determine exact fS2 val-ues over melts, equilibrium fS2 numbers may reveal the relativestability of metal sulfide complexes with respect to oxidation. Tita-nium sulfides are the least stable among considered sulfides. Tita-nium could be in an oxide form at fO2 > �IW�7 at 1500 K (Fig. 6a).MgS is more stable than TiS2 but it is much less stable than Ca andMn sulfides (Fig. 5). At much reduced conditions (e.g., at �IW�6)several metal sulfides are stable at fS2 controlled by FeS–Feo equi-librium (Eq. (18); dashed curves in Fig. 5), while at higher fO2 thesulfides become less stable compared to FeS. In the case of magma

IW bufferlo

g fS

2

-15

-10

-5

0

5

IW-3

log

fS2

-15

-10

-5

0

5

IW-6

Temperature, K

log

fS2

-15

-10

-5

0

5

MgS-MgO

MnS-MnOCaS-CaO

CaS-CaO

CaS-CaOMnS-MnO

MnS-MnO

MgS-MgO

MgS-MgO

a

c

e

FeS-FeO

TiS2-TiO2

TiS2-Ti2O3

TiS2-TiO2

FeS-Feo

FeS-Feo

FeS-Feo

IW buffer

IW-3

IW-6

Temperature, K1000 1200 1400 1600 1800 1000 1200 1400 1600 1800

MgS-Fo

MnS-RhdCaS-Woll

b

d

f

FeS-Feo

FeS-Feo

FeS-Feo

MgS-En

MnS-TphCaS-CaOl

MnS-TphMnS-Rhd

CaS-CaOlCaS-Woll

MgS-En MgS-Fo

MnS-Tph

MnS-RhdCaS-CaOl

CaS-Woll

MgS-En MgS-Fo

Fig. 5. Fugacity of S2 in sulfide–oxide (a, c, e; Eqs. (5)–(10)) and sulfide–silicate (b, d, f; Eqs. (11)–(16) equilibria at 1 bar total pressure and different fO2 relative to the IWbuffer. The activities of sulfides, oxides, and silicates are chosen to be unity. The dashed curves refer to Eq. (18) that may control fS2 in magma. En, Fo, Woll, CaOl, Rhd, and Tphare referred to enstatite, forsterite, wollastonite, rhodonite, and tephroite, correspondingly. These equilibria provide estimates for relative fS2 over different sulfides in silicatemelts. Higher fS2 values correspond to lower stability of sulfides and sulfide complexes. Decompression and/or oxidation of melts would favor sulfide to oxide/silicateconversion and formation of S-bearing gases.

32 M.Yu. Zolotov / Icarus 212 (2011) 24–41

oxidation, decomposition of Ti and Mg sulfide melt complexescould be a major reason for S degassing from initially reducedmelts.

Oxidation of magma on Mercury could potentially be driven byassimilation of oxides from crustal rocks. Several equilibrium oxideassemblages correspond to the redox conditions from IW to IW�6(Fig. 1, Eqs. (1)–(4)) and the assimilation of corresponding O-bear-ing phases would increase the fO2 of magma. Although oxidationthrough melting of FeO-bearing silicates is limited by their possiblescarcity, assimilation of putative ilmenite-rich rocks suggested in(Denevi et al., 2009; Lawrence et al., 2010) could oxidize magma.The relatively oxidizing conditions of ilmenite-bearing assem-blages are expressed by the fO2 values of the IRI buffer (Eq. (2),Figs. 1 and 6). The pathways of sulfide oxidation are illustratedby reactions (20) and (21), were FeO is the oxidizing agent. Thecorresponding sulfide-oxide equilibria at 1500 K give fS2 = 0.09and 14 bar, respectively, providing unity activity of non-gaseousspecies (Table 1, Fig. 6a and b). Magmas could also be oxidizedby Ti(IV) assimilated from ilmenite and/or rutile leading to the for-mation of Ti(III) species (see Fig. 1 and Eqs. (3) and (4)). If oxidationis driven by ilmenite’s Ti(IV) and Fe(II) (e.g., Eq. (22)), magma couldbe oxidized beyond �IW�3. In this case, the oxidation limit is setby the ITOI buffer (Eq. (3), Fig. 1). Assimilation of only Ti(IV) phases(rutile) could only cause oxidation of very reduced magmas towardIW�3–IW�5, depending on temperature (Fig. 1), and the oxidationlimit is placed by the RTO buffer (Eq. (4)). Oxidation of magma andsulfide complexes by Fe(II) from ilmenite (Eqs. (20) and (21)) ismore efficient than reactions involving Ti(IV) reduction to Ti(III)(Eqs. (22) and (23)). This is illustrated by higher fS2 values in sul-fide-oxide equilibria that correspond to the IRI buffer (Eq. (2)) com-pared to buffers that include Ti(IV) reduction (Eqs. (3) and (4))(Fig. 6a). However, the possible lack of FeO-rich ilmenite in Mer-cury’s rocks (Riner et al., 2010) limits the oxidation potential of thismineral.

Some decomposition of sulfide melt complexes may also befacilitated by assimilation of SiO2-rich (e.g., intermediate) rocks.If this occurred, an increase in SiO2 activity (a) would increase fS2

in sulfide-silicate equilibria (Eqs. (11)–(16), Fig. 6c). However, ifaSiO2 is controlled by the forsterite-enstatite equilibrium (Eq.(24)) it may not increase beyond �0.4–0.5. Without forsterite, anincrease in aSiO2 from �0.4–0.5 to �1 increases fS2 by only an or-der of magnitude. The small increase in aSiO2 also limits the oxida-tion of magma by SiO2 that leads to formation of elemental Si (Eq.(25), Fig. 7). In addition, these processes could be limited by a lowabundance of evolved rocks.

The discussed S degassing through oxidation could be presentedby net reaction

2S2�ðmeltÞ þ O2 ! S2ðgÞ þ 2O2�ðmeltÞ

(Fogel, 2005). Removal of S2 gas would reduce magma, which willincrease stability of sulfide melt complexes. The reduction may lim-it degassing of S-bearing species even from long-standing shallowmagma chambers that are vulnerable for gas escape through a frac-tured crust.

5.2. Carbon

Formation of CO through graphite oxidation by FeO in melt (Eq.(26)) has been suggested to explain pyroclastic eruptions on theMoon (e.g., Sato, 1976; Fogel and Rutherford, 1995; Nicholis andRutherford, 2009), parent asteroid(s) of ureilites (Warren and Kal-lemeyn, 1992), and was mentioned with respect to Mercury’s vol-canism (Kerber et al., 2009). Degassing could occur when partialpressure of CO exceeds total pressure. At elevated FeO activities,graphite oxidation is favorable at lower pressure and could occurduring magma ascent and in shallow magma chambers.

Elevated fO2 and temperature correspond to higher fCO in thegraphite-CO equilibrium (Eq. (27), Fig. 8) and favor degassing.

log fO2

relative to IW

-6 -4 -2 0

log

fS2, b

ar

-5

-4

-3

-2

-1

01500 K

TiS2

TiO2TiS

Ti2O3FeS-Feo

log fO2

relative to IW

-6 -4 -2 0

log

fS2, b

ar

-6

-4

-2

0

2

MgS

MgO1700K

log f O2 relative to IW-6 -4 -2 0

log

fS2, b

ar

-6

-4

-2

0

2

MgSiO3

aSiO2 = 1

aSiO2 = 0.01aSiO2 = 0.1

a

b

1700 K

aSiO2 = 0.44

2000K

1400K

c

IRI

IRI,

1400

K

IRI,

2000

KIR

I

ITO

I

RTO

RTO

, 140

0 K

ITO

I, 14

00 K

RTO

, 200

0 K

ITO

I, 20

00 K

2MgS + O2 = 2MgO + S2

2MgS + 2SiO2 + O2 = 2MgSiO3 + S2

ITO

I

RTO

MgS

Fig. 6. Phase equilibria among sulfides and oxides of Ti and Mg at a function of fO2

at 1 bar total pressure. Vertical dotted lines show conditions of redox buffers atcorresponding temperatures. In (a), the lines represent Eqs. (4), (8)–(10), and (17) atunity activities of sulfides and oxides at 1500 K. The dashed line corresponds to theFeS–Feo equilibrium (Eq. (18)) that may control fS2. In (b), fugacity of S2 is shown forEq. (5) at unity activities of MgS and MgO. In (c), equilibrium conditions of Eq. (11)are shown for 1700 K at different activities of silica (aSiO2) and unity activities ofMgS and MgSiO3.The value of aSiO2 = 0.44 corresponds to the forsterite–enstatiteequilibrium (Eq. (24)) that may control silica activity in the melt. Stability of MgS islower at higher fO2 and aSiO2 values.

log f O2 relative to IW-6 -5 -4 -3 -2 -1 0

log

aSi

-14

-12

-10

-8

-6

-4SiO2 = Si + O21700 K

1400 K aSiO2 = 1

0.44

aSiO2 = 1

0.50.1

0.1

Fig. 7. Activity of Si as a function of fO2 for the conditions of Eq. (25) for 1400 K and1700 K at 1 bar total pressure. The values at lines refer to activity (a) of SiO2. Solidlines correspond to aSiO2 controlled by Eq. (24). The plot shows that aSiO2 has muchlower effect on aSi than fO2 and temperature.

C(graphite) + 0.5O2(g) = CO(g)IW

1200 1300 1400 1500 1600 1700 1800

log

fCO

, bar

-2

0

2

4

IW-3

IW-6

Temperature, K

IRI

RTOITOI

2 wt% FeO

6 wt% FeO

EC

CB/CH

Fig. 8. Fugacity of CO in the equilibrium with graphite (Eq. (27)) at different fO2

values. The curves correspond to oxidation states controlled by mineral buffersdepicted in Fig. 1 and certain log fO2 values below IW. Curves named IRI, ITOI, andRTO also correspond to Eqs. (28)–(30), respectively. The gray curves show fCO at fO2

that correspond to 2 wt% and 6 wt% FeO in magma saturated with Feo-metal (seeSection 2.3). The bold curve marked CB/CH corresponds to fCO at fO2 of the partialmelt of CB/CH chondrites. The circle EC symbol shows fCO at fO2 of the partial meltof an enstatite chondrite. The activity of graphite is unity. The graph illustrates thepossibility of graphite oxidation by Fe(II) and Ti(IV) oxides and corresponding COdegassing at low lithostatic pressure (cf. Sato, 1976; Fogel and Rutherford, 1995).

M.Yu. Zolotov / Icarus 212 (2011) 24–41 33

However, at fO2 < IW�3, equilibrium fCO does not exceed �102 barat T < 1800 K. This is consistent with the evaluation of fCO in equi-librium with graphite at fO2 that corresponds to the specified low-FeO content in non-ideal magmas saturated in Feo-metal (see Sec-tion 2.3). Graphite-bearing magmas with 2 wt% FeO could produce12–27 bar CO at 1500–1700 K, respectively, and magmas with 6wt% FeO give 36–85 bar CO (Fig. 8). At fO2 that corresponds tothe partial melt of CB/CH chondrites with 3.2 wt% FeO (Section 2.3)partial pressure of CO is 29–49 bar at 1600–1700 K. However, a re-duced partial melt with 0.25 wt% FeO formed from the Indarchenstatite chondrite (McCoy et al., 1999; Taylor and Scott, 2003)could provide only 3.5 bar CO at 1700 K (Fig. 8). The latter valueis consistent with a violent CO degassing in a vacuum from a par-tially melted enstatite chondrite (Muenow et al., 1992), and with adeficiency of C in aubrites compared to enstatite chondrites (Ta-ble 2). Overall, these evaluations imply a possibility for CO degas-sing through low-P oxidation of graphite only near Mercury’ssurface (e.g. from shallow lava flows). The evaluated low fCO num-bers could not be responsible for a major pyroclastic activity, ex-cept lunar-type oxidized magmas.

The involvement of ilmenite’s Fe(II) in graphite oxidation (Eq.(28)) would lead to CO degassing at greater depths. At fO2 con-trolled by the IRI buffer (Eq. (2)) at 1700 K, fCO in graphite-bearingmagma could be as high �400 bar (Fig. 8). This means that CO maystart degassing �4 km below the surface (were partial P of CO be-comes equal to lithostatic P). Although possible, oxidation ofgraphite by Ti(IV) (Fogel and Rutherford, 1995) is less efficient thanby FeO, as illustrated by lower fCO values in corresponding equilib-ria (Eqs. (29) and (30); Fig. 8). Nevertheless, high magma temper-ature strongly favors oxidation by TiO2 (Eq. (30)), and a majorassimilation of TiO2 could lead to fCO up to 25–70 bar at 1700–1800 K, which implies a possibility for degassing from the depth<230–650 m.

Oxidation of graphite by the SiO2 in magma (Eq. (31)) men-tioned by Kerber et al. (2009) may also contribute to CO productionat high-T and low-P conditions. This notion is consistent with signsof low-P reduction of SiO2 in Co-, Feo-bearing enstatite chondriticmelts leading to dissolution of Sio in Feo-metal (McCoy et al., 1999).

Degassing and escape of low-solubility CO would favor furthergraphite oxidation by shifting Eq. (27) to the right hand side.Although CO removal makes magma more reduced and decreasesthe potential for oxidation, the extent of reduction could be limited

Temperature, K

log

fN2,

bar

-6

-4

-2

0

2

4

6

IW-3

IW-6

log f O2 relative to IW1000 1200 1400 1600 -6 -5 -4 -3 -2 -1 0

log

fN2

TiN

TiO2

1700K

Ti2O3

a b

ITOI

RTO

Fig. 9. Phase equilibria between osbornite (TiN) and Ti oxides. (a) Equilibriumfugacities of N2 for Eqs. (32) and (33) at fO2 controlled by redox buffers resented byEqs. (3) and (4). At IW�6, TiN coexists with Ti2O3. (b) Stability fields at 1700 K. Thefigure illustrates that decompression and/or oxidation of reduced magma maycause TiN decomposition and nitrogen degassing.

34 M.Yu. Zolotov / Icarus 212 (2011) 24–41

by the C content in magma. Overall, formation of CO on Mercurycould be restricted by the low activities of FeO and TiO2 that woulddecrease along with oxidation of graphite and reduced melt com-plexes of S, Cl, and N. Reduced magmas (fO2 < �IW�3) may nothave enough oxidants to produce 0.55 wt% CO that is needed toform the largest pyroclastic deposit at Caloris (Kerber et al., 2009).

5.3. Chlorine

Degassing of Cl is not favored by redox conditions below IW(Nicholis and Rutherford, 2009) and low pressure (Aiuppa et al.,2009). Even at relatively oxidizing conditions, Cl is a minor compo-nent of Earth’s volcanic gases (Symonds et al., 1994; Oppenheimer,2003; Aiuppa et al., 2009) as well as Io’s (Lellouch et al., 2003; Feg-ley and Zolotov, 2000). Degassing of Cl from reduced magmas onMercury could be limited because of Cl association with Mg, Ca,and S (Section 4). Nevertheless, the release of abundant Cl-bearinggases in a vacuum from enstatite chondrite melts (Muenow et al.,1992) implies the presence of Cl in Mercury’s gases. Chlorine coulddegas as Cl–S species, Cl2, Cl, and alkali chlorides as observed andmodeled for Io and the Moon (Lellouch et al., 2003; Fegley andZolotov, 2000; Schaefer and Fegley, 2005a; Fegley, 1992), and isconfirmed in this paper (Section 6). Putative oxidation of magmaon Mercury would favor decomposition of Cl–metal and Cl–S meltcomplexes, but the Cl gas content would not be sufficient to ex-plain observed signs of pyroclastic activity.

5.4. Nitrogen

A strong decrease in N solubility with decreasing pressure (Ros-kosz et al., 2006; Mysen et al., 2008) and increasing fO2 from IW�8to IW�1 (Libourel et al., 2003) may suggest N degassing throughdecompression and/or oxidation of magma (Fig. 4b). The processescould be expressed by reaction

Si3N4ðin meltÞ þ 3O2 ! 3SiO2ðin meltÞ þ 2N2ðgÞ

(Ito and Fruehan, 1988). Experimental data show that major degas-sing is also expected from decompression of N-saturated magma atfO2 > � IW�1. At IW, up to 1.6 wt% N2 could degas through magmaascent from 25 kbar (�230 km below Mercury’s surface) (Mysenet al., 2008). If N solubility in reduced melts is proportional toP1=2

N2(Libourel et al., 2003), decompression of FeO-depleted basaltic

magma (IW�6, �1700 K) from 1 kbar causes degassing of up to 0.7wt% N2. Similar amounts of degassed N2 are expected if reducedmelts (fO2 < IW�2) have the same high-P (10–20 kbar) solubilityas more oxidized magmas. Note however, that suggested decreasein N solubility with lowering pressure also requires oxidation (seereaction above) and could be inefficient at fO2 < IW�1 withoutassimilation of oxidants. It is unclear if the �0.55 wt% gas withmolecular mass of 28 (CO or N2) needed to explain pyroclastic erup-tions (Kerber et al., 2009) could be obtained without invoking oxi-dation by crustal oxides. Nevertheless, N2 could be the mostabundant volcanic gas if magma oxidation has not occurred, andsome S-, C-, and Cl-species could separate into the N2-rich gas.

Nitrogen degassing from extremely reduced magmas has beendiscussed to account for presumed pyroclastic activity on the par-ent body of aubrites (Wilson and Keil, 1991). The low N content inaubrites compared to enstatite chondrites (Table 2) may suggest aloss of �150–400 ppm N in igneous processes. These numbers donot contradict the amounts of N released through partial meltingof enstatite chondrites in a vacuum (Muenow et al., 1992), throughdegassing mechanism is unknown. In addition to decomposition ofnitride melt complexes, N could be released through near-surfaceoxidation of solid nitrides, if they were present in igneous systemson Mercury and elsewhere. Fig. 9 illustrates that osbornite (TiN)could decompose because of oxidation and/or decompression of

magma. Other nitrides observed in enstatite chondrites (e.g., sino-ite, Si2N2O) could be subjected to high-T and low-P oxidation aswell (cf., Fogel et al., 1989). However, without magma oxidation,osbornite may crystallize upon cooling, as suggested from the re-duced mineralogy of Ti in aubrites (Keil, 1969).

5.5. Degassing chemistry of magmas with different fO2

The preceding discussion shows that fO2 strongly affecting sol-ubility of S, C, Cl, and N in silicate melts. The redox state controlsamounts of volatiles sequestered in melt in magma source regionsand influences melt-gas partitioning in low-P conditions duringmagma ascent. Abundances of volatiles in mantle source regionsand building blocks of the planet are less important for volcanicgas composition.

Relatively oxidized lunar-type magma (�IW to �IW�2) wouldproduce more CO than reduced melts through low-P oxidation ofgraphite in ascending magmas. The production of CO could bethe major driving force for pyroclastic activity. These magmas alsohave the potential for major nitrogen degassing through decom-pression (see Fig. 4b), if melts are saturated with N at depth. Moreoxidized magmas could emit more S because of formation of S2O,SO, and SO2 (see Section 6). Degassing of Cl could be limited be-cause of the increasing solubility with lowering pressure and alow initial content. Therefore, putative ilmenite-rich magmas withfO2 = IW � IW�1 (Lawrence et al., 2010) could emit gases with thefollowing composition: C > N P S > Cl. Magma sourced from deepand/or undepleted mantle regions would deliver and degas moreC and N species.

More reduced magmas (fO2 < �IW�3) could deliver more S, Cl,and N because of higher solubilities of these elements. Fewer Ccould be supplied in magma because of its lesser solubility at lowerfO2. Degassing of C, S, and Cl could be limited because of high sta-bility of graphite and melt complexes of S and Cl. However, thesemagmas may emit more N through decompression. Althoughdegassing of N may account for pyroclastic eruptions of extremelyreduced melts, more experimental efforts are needed to supportthis suggestion. In addition to N2 (see Section 6), correspondingvolcanic gases would contain C, Cl, and S species formed throughlow-P decomposition of graphite and melt complexes in the closeproximity of Mercury’s surface. Degassing experiments with ECmelts (Muenow et al., 1992) cold be considered as proxy for thiscase.

The most efficient degassing is expected through putative oxi-dation of much reduced magmas (fO2 < �IW�5) by crustal oxides.

M.Yu. Zolotov / Icarus 212 (2011) 24–41 35

For example, reduced volatile-rich magmas delivered from core–mantle boundary after the disrupting Caloris impact could be oxi-dized through assimilation of crustal rocks that may (Denevi et al.,2010; Lawrence et al., 2010) or may not contain ilmenite. This pro-cess should strongly increase the amount of formed S–C–N–Clgases and may account for major pyroclastic activity. Below weuse thermochemical equilibrium calculations to evaluate the speci-ation of corresponding volcanic gases.

6. Speciation models for volcanic gases

Ideal gas chemical equilibria in the S–C–O–Cl–N system havebeen calculated at variable T, P, fO2, and assumed bulk composi-tions with the use of thermodynamic properties of gas species(Barin, 2004) and graphite. The calculations were performed bythe Gibbs free energy minimization method (Van Zeggeren andStorey, 1970) with the GEOCHEQ code (Mironenko et al., 2008)modified by the author. The applicability of these models to Mer-cury is limited by the uncertainties in the oxidation state of mag-mas, bulk compositions of volatiles, and their solubilities inmelts. Therefore, we explore the effects of these parameters ongas speciation rather than formulate definite predictions.

First, we evaluated speciation of gases released from heated andmelted enstatite chondrites (Muenow et al., 1992). The gases arecharacterized with the following elemental ratios S:C:Cl:N =1:12.5:1.5:3.5. These ratios are significantly different from bulkcompositions of chondritic materials (Fig. 3) and could reflectlimited volatility of S. The speciation of gases was modeled at1700 K, 1 bar total pressure, and log fO2 = �14.9 (IW�5.3) evalu-ated for the partial melt of the Indarch enstatite chondrite (Sec-tion 2.3). The calculated volume (mole) fractions of major gasesare as follows: CO, 0.788; N2, 0.113; Cl, 0.033; Cl2, 0.030; S2,0.015; CS2, 0.014. Concentrations of other gases can be seen inFig. 10a at 1700 K. The results presented in Fig. 10 also exploreeffects of temperature and fO2 on the speciation of gases with thatbulk composition. Elevated temperatures favor higher concentra-tions of smaller gas molecules: Cl, S2, CS, SCl and S. Condensationof graphite below 1460 K strongly decreases CO abundance andgas becomes rich in N2 and Cl2. More oxidizing conditions than

Temperature, K1000 1200 1400 1600 1800 2000

Gas

Mol

e Fr

actio

n

-5

-4

-3

-2

-1

0

S2

CO

CS2

S

COS

S3

CO2S2Cl

SCl2

SCl

Cl2

S2Cl2

a1 bar, IW-5.3

N2

Cl

S2

S3

S2Cl

SClCS

COS

ClCN

Cl

COClCOCl2

ClCN

graphite

Fig. 10. Speciation of volcanic gases with bulk composition of volatiles released from ens1 bar total pressure. (b), At T = 1700 K and 1 bar. In (a), the vertical dotted line shows uppethe symbols at the X axis show conditions for redox buffers (Fig. 1) and fO2 of partial meand Anders (1980), M–A melts (see Section 2.3).

in enstatite chondrite melts correspond to higher concentrationsof O-bearing gases (CO2, SO2, SO, S2O) and lower contents of C spe-cies (CS2, CS, ClCN) (Fig. 10b). For redox conditions below IW themost abundant volatiles are CO, N2, Cl, Cl2, S2, and CS2. These com-positions are not similar to terrestrial volcanic gases that containH2O, CO2, SO2, H2S, HCl, and H2 (Symonds et al., 1994; Oppenhei-mer, 2003).

Another set of speciation models has been developed for bulkgas compositions typical for chondritic analogs of Mercury andits mantle (Figs. 11–13). The calculations were performed at Cl/Smole ratio of 0.01 that represents enstatite, ordinary, and CI chon-drites, the solar photosphere, and the Kilauea summit 1918 volca-nic gases (Fig. 3a). This value represents a low range of the Cl/Sratio in Earth’s volcanic gases (Aiuppa et al., 2009) and could alsobe considered as a lower limit for Mercury’s gases. The preferredN/S mole ratio of 0.017 characterizes EH enstatite chondrites andis similar to that in EL and CB chondrites (Fig. 3b). The C/S mole ra-tios of 0.19 and 1.9 were chosen to represent EH and CB chondrites,respectively. The latter value also roughly represents CI chondritesand Kilauea gases (Fig. 3).

Elevated temperature favors higher abundances of CO, Cl, S, CS,SCl, SO, and S2O (Fig. 11). Concentrations of CS2, S3–S8, S2Cl, SCl2,and S2Cl2 are higher at lower temperatures. Condensation ofgraphite occurs at temperatures below those suggested for Mer-cury’s typical magmas. At �1700 K, extremely reduced gases(IW�5.3, see Section 2.3) mostly consist of S2, CS2, and CO(Fig. 11a). Carbon monoxide dominates in more oxidized and C-rich gases (Fig. 11b). The latter gases could be associated withmelts formed through partial melting of silicates from metal-rich(CB/CH) chondrites.

High-P gases, which characterize magma before its fragmenta-tion, have elevated concentrations of CS2, S2Cl, and S2Cl2

(Fig. 12). Low pressure corresponds to higher fractions of CO, S2,S, Cl, SO, S2O, and SO2. In C-rich systems, concentrations of C gasescould be controlled by graphite which is stable at elevated pres-sures. Lowering pressure in moderately oxidized systems favorsoxidation of graphite to CO (Fig. 12b and Section 5.2). Decompres-sion of ascending magma could be accompanied by increasing CO/CS2 and S2/CS2 ratios.

log f O2 relative to IW

-6 -4 -2 0

log

Gas

Mol

e Fr

actio

n

-5

-4

-3

-2

-1

0

S2

CO

CS2

S

COS

S3

SO

CO2

S2Cl

COCl

CS

SCl2

SCl

Cl2Cl

SO2

S2O

S2Cl2

RTO

ITOI

IRI IW

N2

EC melt BC, M-A melts

ClCN

S3

1700 K, 1 bar

b

tatite chondrite samples (Muenow et al., 1992). (a), At log fO2 = �14.9 (IW�5.3) andr temperature boundary of graphite stability, which condenses below 1460 K. In (b),lts of enstatite (EC) and CB chondrites, and melts of Mercury’s mantle from Morgan

Temperature, K

Gas

Mol

e Fr

actio

n

-5

-4

-3

-2

-1

0S2

CO

CS2

SS5

COS

S3

SO

CO2

S2Cl

S4 CS

SCl2

SCl

Cl2

S6

S2Cl2

a1 bar, IW-5.3, C/S=0.19

N2

Cl

Temperature, K1000 1200 1400 1600 1800 2000 1000 1200 1400 1600 1800 2000

Gas

Mol

e Fr

actio

n

-5

-4

-3

-2

-1

0

S2

CO

CS2

S

S5

COS

S3

SO

CO2

S2Cl

S4

CSSCl2

SCl

Cl2

Cl

S6

S2O

S2Cl2

b1 bar, IW-3.1, C/S=1.9

S4

N2

S3

Cl

S2Cl

graphite

Fig. 11. Speciation of volcanic gases as a function of temperature at 1 bar total pressure. (a), At a generalized bulk composition and redox conditions of EH enstatitechondrites, S:C:Cl:N = 1:0.19:0.01:0.017, IW�5.3. (b), At bulk composition and redox conditions of CB chondrites, S:C:Cl:N = 1:1.9:0.01:0.017, IW�3.1. In (b), the verticaldotted line shows upper boundary of graphite stability, which condenses below 1203 K.

log Pressure, bar

Gas

Mol

e Fr

actio

n

-5

-4

-3

-2

-1

0S2

CO

CS2

S COS

S3

SO

CO2

S2Cl

S4

SCSCl2

SCl

Cl2

Cl N2

S2Cl2

S5

S6

log Pressure, bar-4 -3 -2 -1 0 1 2 -4 -3 -2 -1 0 1 2

Gas

Mol

e Fr

actio

n

-5

-4

-3

-2

-1

0

S2

CO

CS2

SCOS

S3

SO

CO2

S2Cl

S4

SC

SCl2

SCl

Cl2

ClN2

SO2

S2OCOCl

S2Cl2

ba1700 K, IW-5.3, C/S = 0.19 1700 K, IW-3.1, C/S = 1.9

Fig. 12. Speciation of volcanic gases as a function of total pressure at 1700 K. Bulk gas compositions and redox conditions are referred to EH (a) and CB (b) chondrites (seecapture of Fig. 11). In (b), the vertical dotted line shows pressure (70 bar) of graphite condensation.

36 M.Yu. Zolotov / Icarus 212 (2011) 24–41

Redox conditions below IW strongly affect concentrations ofCS2, CS, and O-bearing gases: CO2, CO, SO2, S2O, and SO (Fig. 13).At fO2 < IW�3 major gases are S2, CO, and CS2. Carbon dioxideand SO2 are abundant only for fO2 > �IW�1, which may not repre-sent Mercury. At redox conditions of enstatite chondrite melts, CS2

is the major gas with the mole fraction (�0.1) comparable withconcentrations of CO and/or S2. At fO2 > �IW�3, CS2 fraction is be-low 0.01 level. These models show that oxidation of magma bycrustal oxides (Section 5), could be associated with conversion ofCS2 and CS to O-bearing gases.

The unknown bulk volatile composition is the major constraintfor these models, and additional runs were performed at variableC/S, Cl/S, and N/S ratios at fO2 = IW�4 (Fig. 14). At C/S < 0.4, whichcorresponds to bulk composition of enstatite chondrites, S2 domi-nates over CO (Fig. 14a). At C/S = 2, as in Kilauea and other Earth’svolcanic gases (Symonds et al., 1994; Oppenheimer, 2003), CO ismore abundant than S2 and the gas is rich in CS2 and COS. The lim-ited S volatility from reduced melts, graphite oxidation during

magma assent, or low-P degassing of enstatite chondrite melts(Muenow et al., 1992) would lead to C/S > 2 that results in CO dom-inance (Figs. 10 and 14a). Note that reduced C gases (CO, CS2, andCOS) are significantly less abundant than CO2 at all consideredranges of the C/S ratio.

At Cl/S = 0.01, which represents bulk chondritic materials(Fig. 3), concentrations of Cl gases (Cl, Cl2, S2Cl, etc.) do not exceed1% level. However, the elevated Cl/S gas ratios, as observed byMuenow et al. (1992), may account for the dominance of Cl2 andCl over S-bearing gases (Figs. 10 and 14b). Other models for anhy-drous gases show that the addition of alkalis leads to the formationof alkali halides (NaCl, Na2Cl2, KCl) and Na and K gases, if(Na + K) > Cl (Fegley and Zolotov, 2000; Schaefer and Fegley,2005a). Fluorine could mainly degas as SiF4 (cf. Schaefer and Feg-ley, 2005b) providing an elevated aSi in reduced melts.

An elevated N/S ratio may represent decompression of N-satu-rated magma (Section 5) and low-P degassing of reduced melts re-lated to enstatite chondrites (Figs. 10 and 14c). The increase in N/

log f O2 relative to IW

-6 -4 -2 0

log

Gas

Mol

e Fr

actio

n

-5

-4

-3

-2

-1

0S2

CO

CS2

S

COS

S3SO

CO2

S2Cl

S4CS

SCl2

SCl

Cl2

Cl

SO2

S2O

S2Cl2

RTO

ITOI

IRI IW

1700 K, 1 bar, C/S = 0.19

N2

EC melt BC, MA melts

log f O2 relative to IW

-6 -4 -2 0

log

Gas

Mol

e Fr

actio

n

-5

-4

-3

-2

-1

0

S2

CO

CS2

S

COS

S3

SO

CO2

S2ClCS

SCl2SCl

Cl2

Cl

SO2

S2ORTO

ITOI

IRI IW

1700 K, 1 bar, C/S = 1.9

N2

EC melt BC, M-A melts

ba

Fig. 13. Speciation of volcanic gases as a function of fO2 at 1700 K and 1 bar total pressure. Bulk gas compositions are referred to EH (a) and CB (b) chondrites (see capture ofFig. 11). The symbols at the X axis show conditions for redox buffers (Fig. 1) and fO2 of partial melts of enstatite (EC) and CB chondrites, and melts of Mercury’s mantle fromMorgan and Anders (1980), M–A melts (see Section 2.3).

log C/S Mole Ratio

-2 -1 0 1

log

Gas

Mol

e Fr

actio

n

-5

-4

-3

-2

-1

0S2 CO

CS2

S

COS

S3

CO2

S2Cl

S4

CS

SCl2

SCl

Cl2

Cl

S2O

KilaueaEH

S2Cl2

N2

SO

CB

CI

EL

log Cl/S Mole Ratio

-2 -1 0

log

Gas

Mol

e Fr

actio

n

-5

-4

-3

-2

-1

0S2

CO

CS2

S

COS

S3

CO2

S2Cl

S4

SCl2

SCl

Cl2

Cl

S2O

EC gasEH, Kilauea

S2Cl2

N2

SO

CI

EL Sun

COCl

COCl2

CS

log N/S Mole Ratio

-2 -1 0 1

log

Gas

Mol

e Fr

actio

n

-5

-4

-3

-2

-1

0S2

CO

CS2

S

COS

S3

CO2

S2Cl

S4

SCl2

SCl

Cl2

Cl

S2O

EC gasCB

S2Cl2

N2

SO

CIEL, EH

Sun

CS

cba

Fig. 14. Speciation of volcanic gases as a function of C/S, Cl/S, and N/S ratios at 1700 K, log fO2 = �13.5 bar (IW�4) and 1 bar total pressure. In (a), S:Cl:N = 1:0.01:0.017.Graphite condenses at C/S > 10. In (b), S:C:N = 1:0.37:0.017. The C/S ratio represents EL chondrites. In (c), S:C:Cl = 1:0.37:0.01. The symbols at the X axis show elemental ratiosin chondrites, the solar photosphere, and the Kilauea 1918 volcanic gas, and gases released from enstatite chondrites (EC gas, Muenow et al., 1992).

M.Yu. Zolotov / Icarus 212 (2011) 24–41 37

S ratio corresponds to dilution of the gas with N2, which is the dom-inant N species (Fig. 14c). Volume fraction of ClCN may exceed 10�5

only in reduced C–Cl–N rich gas that could be associated with ensta-tite chondrites (Fig. 10). The fraction of NS does not exceed 10�5, andother considered N gases (CN, C2N2, NO, N, NOCl, etc.) are less abun-dant than 10�9. Even if H is present in Mercury’s magmatic gases,NH3/N2 ratio is 10�5.9–10�11.9 (at H/N = 10�1–10�5, 1700 K and1 bar; see Eq. (34)). At 1700 K and fO2 from IW�3 to IW�6, theH2/H2O mole ratio is 37–1175, respectively (see Eq. (35)). H2S, H2,and HCl could be major H-bearing species in reduced gases.

7. Concluding remarks

The position of Mercury in the Solar System, its high density,and a low-FeO content in silicates imply high-T processing of mate-rials in reduced and anhydrous conditions. These processes could

have involved preferential high-T and high-P condensation in thesolar nebula, evaporation of silicates, and collision-driven losses.High temperature could have restricted accretion of H-bearing spe-cies and favored their lost early in the body’s history, if they ac-creted. However, by analogy with enstatite and CB/CHchondrites, Mercury may not be strongly depleted in S, Cl, and N.These elements could have survived major losses because of stabil-ity of corresponding reduced solids.

The FeO content in Mercury’ surface silicates below �6 wt% cor-responds to fO2 < �IW�2.3 in Feo-metal-saturated magmas. Re-duced conditions favor high solubility of S, N, and Cl in magmasand would cause partitioning of these elements into melts duringmagma formation. However, reducing conditions could have lim-ited volcanic degassing of S, C, Cl, and N.

Although low-P oxidation of graphite during magma ascentleads to formation of CO, this process is much less efficient at

38 M.Yu. Zolotov / Icarus 212 (2011) 24–41

fO2 < IW�2, which may represent Mercury’ mantle melts. Lower-ing N solubility through magma decompression provides a mecha-nism for degassing and pyroclastic activity on Mercury in a widerange of redox conditions. However, an effect of pressure on N sol-ubility below the IW buffer has not been explored experimentallyand low-P oxidation of nitride melt complexes could be inefficient.

Potentially, major magma degassing could have been driven byoxidation of Mercury’s mantle melts through assimilation of crus-tal Fe(II), Ti(IV) species, and SiO2. In this case, oxidation of graphite,and sulfide, chloride, and nitride melt complexes could lead tocombination of C-, S-, Cl-, and N-bearing gases and pyroclasticactivity. This process implies the occurrence of more oxidized crustthan mantle. Although infall of hydrous bodies during the late hea-vy bombardment �4.9 Ga ago (Gomes et al., 2005) could have oxi-dized the crust, the majority of remote-sensing data, except(Lawrence et al., 2010), are consistent with reduced crustal rocks.

Without major oxidation in crustal magma chambers, signifi-cant amounts of FeS (and possibly MgS, CaS, MnS, and Ti-bearingsulfides), graphite, chlorides, and nitrides could remain in crystal-lized igneous rocks. Metal sulfides rather than Ti oxides could bemajor opaque phases consistent with the low-FeO content in sur-face silicates. By analogy with terrestrial komattites (Barnes,2007) and large basaltic magma reservoirs, sulfide liquids couldhave precipitated and formed massive sulfide deposits. However,sulfides trapped in lava crusts and excavated by impacts could beobservable, and the presence of abundant sulfides is consistentwith the low albedo and microwave transparency of surface mate-rials (Sprague et al., 1995). Abundance and speciation of S in sur-face materials could be a key for the mantle redox state and fateof volatiles. Upcoming orbital observations with MESSENGER couldverify these suggestions and constrain the fate of magmaticvolatiles.

Evaluations of volcanic gas speciation are limited by the un-known composition and redox conditions in Mercury’s rocks andmagmas, and uncertain solubilities of Cl and N in reduced melts.Overall, the reduced and dry conditions in Mercury’s magmas im-ply gas speciation that is dissimilar to Earth’s volcanic gases.Chemical equilibrium models show that N2, CO, S2, CS2, S2Cl, Cl,Cl2, and COS could be among the most abundant gases. Relativelyoxidized magmas (�IW to �IW�3) could emit CO, N2, and abun-dant S-bearing gases. More reduced volcanic emanations couldbe rich in N2 and depleted in C-, S-bearing gases.

Acknowledgments

The manuscript is benefited from comments from Steven Leh-ner, Bruce Fegley, and an anonymous reviewer. This work is sup-ported by grants from NSF Planetary Astronomy, NASACosmochemistry, and NASA Planetary Geology and Geophysicsprograms.

References

Aiuppa, A., Baker, D.R., Webster, J.D., 2009. Halogens in volcanic systems. Chem.Geol. 263, 1–18.

Arculus, R.J., Delano, J.W., 1987. Oxidation state of the upper mantle: Presentconditions, evolutions and controls. In: Nixon, P.H. (Ed.), Mantle Xenoliths.Wiley, New York, pp. 589–598.

Ariskin, A.A., Borisov, A.A., Barmina, G.S., 1993. Simulating iron–silicate meltequilibrium in basaltic systems. Geochem. Int. 30, 13–22.

Asphaug, E., Angor, C.B., Williams, Q., 2006. Hit-and-run planetary collisions. Nature438, 155–160.

Asplund, M., Grevesse, N., Sauval, A.J., 2005. The solar chemical composition. ASPConf. Series 336, 25–38.

Barin, I., 2004. Thermochemical Data of Pure Substances. Wiley-VCH.Barnes, S.J., 2007. Cotectic precipitation of olivine and sulfide liquid from komatiite

magma and the origin of komatiite-hosted disseminated nickel sulfidemineralization at Mount Keith and Yakabindie, Western Australia. Econ. Geol.102, 299–304.

Basaltic Volcanism Study Project (BVSP), 1981. Basaltic Volcanism on TerrestrialPlanets. Pergamon Press, New York, 1286pp.

Benz, W., Slattery, W.L., Cameron, A.G.W., 1988. Collisional stripping of Mercury’smantle. Icarus 74, 516–528.

Benz, W., Anic, A., Horner, J., Whitby, J.A., 2007. The origin of Mercury. Space Sci.Rev. 132, 189–202.

Benz, W., Anic, A., Horner, J., Whitby, J.A., 2008. The origin of Mercury. In: Balogh, A.,Ksanfomality, L., von Steiger, R. (Eds.), Mercury. Space Science Series of ISSI.Springer, pp. 7–20.

Berthet, S., Malavergne, V., Righter, K., 2009. Melting of the Indarch meteorite (EH4chondrite) at 1 GPa and variable oxygen fugacity: Implications for earlyplanetary differentiation processes. Geochim. Cosmochim. Acta 73, 6402–6420.

Blewett, D.T., Lucey, P.G., Hawke, B.R., Ling, G.G., Robinson, M.S., 1997. A comparisonof mercurian reflectance and spectral quantities with those of the Moon. Icarus129, 217–231.

Blewett, D.T. et al., 2009. Iron on the surface of Mercury: From contradiction toconvergence? Eos Trans. AGU (Fall Suppl.) 90 (52). Abstract P23D-08.

Bockelée-Morvan, D., Crovister, J., Mumma, M.J., Weaver, H.A., 2004. Thecomposition of cometary volatiles. In: Festou, M.C., Keller, H.U., Weaver, H.A.(Eds.), Comets, vol. II. University of Arizona Press, Tucson, pp. 309–423.

Bonal, L., Huss, G.R., Krot, A.N., Nagashima, K., 2010. Chondritic lithic clasts in theCB/CH-like meteorite Isheyevo: Fragments of previously unsampled parentbodies. Geochim. Cosmochim. Acta 74, 2500–2522.

Bottke, W.F., Nesvorny, D., Grimm, R.E., Morbidelli, A., O’Brien, D.P., 2006. Ironmeteorites as remnants of planetesimals formed in the terrestrial planet region.Nature 439, 821–824.

Boyce, J.W., Liu, Y., Rossman, G.R., Guan, Y., Eiler, J.M., Stolper, E.M., Taylor, L.A.,2010. Lunar apatite with terrestrial volatile abundances. Nature 466, 466–468.

Boynton, W.V., Sprague, A.L., Solomon, S.C., Starr, R.D., Evans, L.G., Feldman, W.C.,Trombka, J.I., Rhodes, E.A., 2007. MESSENGER and the chemistry of Mercury’ssurface. Space Sci. Rev. 131, 85–104.

Brearley, A.J., Jones, R.H., 1998. Chondritic meteorites. In: Papike, J.J. (Ed.), PlanetaryMaterials, Rev. Mineral, vol. 36. Mineral. Soc. Am., Washington, DC, pp. 1–398.

Breuer, D., Hauck II, S.A., Buske, M., Pauer, M., Spohn, T., 2008. Interior evolution ofMercury. In: Balogh, A., Ksanfomality, L., von Steiger, R. (Eds.), Mercury. SpaceScience Series of ISSI. Springer, pp. 21–46.

Brown, S.M., Elkins-Tanton, L.T., 2009. Compositions of Mercury’s earliest crustfrom magma ocean models. Earth Planet. Sci. Lett. 286, 446–455.

Burbine, T.H., McCoy, T.J., Nittler, L.R., Benedix, G.K., Cloutis, E.A., Dickinson, T.L.,2002. Spectra of extremely reduced assemblages: Implications for Mercury.Meteorit. Planet. Sci. 37, 1233–1244.

Cameron, A.G.W., 1985. The partial volatilization of Mercury. Icarus 64, 285–294.Campbell, A.J., Humayun, M., Weisberg, M.K., 2002. Siderophile element constraints

on the formation of metal in the metal-rich chondrites Bencubbin, Weatherford,and Gujba. Geochim. Cosmochim. Acta 66, 647–660.

Carroll, M.R., Webster, J.D., 1994. Solubilities of sulfur, noble gases, nitrogen,chlorine, and fluorine in magmas. In: Carroll, M.R., Holloway, J.R. (Eds.),Volatiles in Magmas. Rev. in Mineralogy, vol. 30. Mineral. Soc. Am., Washington,pp. 231–280.

Ciesla, F.J., Cuzzi, J.N., 2006. The evolution of the water distribution in a viscousprotoplanetary disk. Icarus 181, 178–204.

Clayton, R.N., Mayeda, T.K., 1999. Oxygen isotope studies of carbonaceouschondrites. Geochim. Cosmochim. Acta 63, 2089–2104.

Cuzzi, J.N., Davis, S.S., Dobrovolskis, A.R., 2003. Blowing in the wind. II. Creation andredistribution of refractory inclusions in a turbulent protoplanetary nebula.Icarus 166, 385–402.

Cyr, K.E., Sears, W.D., Lunine, J.I., 1998. Distribution and evolution of water ice in thesolar nebula: Implications for Solar System body formation. Icarus 134, 537–548.

Cyr, K.E., Sharp, C.M., Lunine, J.L., 1999. Effects of the redistribution of water in thesolar nebula on nebular chemistry. J. Geophys. Res. 104, 19003–19014.

Denevi, B.W., Robinson, M.S., 2008. Mercury’s albedo from Mariner 10: Implicationsfor the presence of ferrous iron. Icarus 197, 239–246.

Denevi, B.W. et al., 2009. The evolution of Mercury’s Crust: A global perspectivefrom MESSENGER. Science 324, 613–618. doi:10.1126/science.1172226.

Desch, S.J., Connoly Jr., H.C., 2002. A model of the thermal processing of particles insolar nebula shocks: Application to the cooling rates of chondrules. Meteorit.Planet. Sci. 37, 183–207.

Domingue, D.L., Koehn, P.L., Killen, R.M., Sprague, A.L., Sarantos, M., Cheng, A.F.,Bradley, E.T., McClintock, W.E., 2007. Mercury’s atmosphere: A surface-boundedexosphere. Space Sci. Rev. 131, 161–186.

Dreibus, G., Wänke, H., 1989. Supply and loss of volatile constituents during theaccretion of terrestrial planets. In: Atreya, S.K., Pollack, J.B., Matthews, M.S.(Eds.), Origin and Evolution of Planetary and Satellite Atmospheres. Universityof Arizona Press, Tucson, pp. 268–288.

Ebel, D.S., 2006. Condensation of rocky material in astrophysical environments. In:Lauretta, D.S., McSween, H.Y., Jr. (Eds.), Meteorites and the Early Solar System II.University of Arizona Press, Tucson, pp. 253–277.

Ernst, C.M., Murchie, S.L., Barnouin, O.S., Robinson, M.S., Denevi, B.W., Blewett, D.T.,Head, J.W., Izenberg, N.R., Solomon, S.C., Roberts, J.H., 2010. Exposure ofspectrally distinct material by impact craters on Mercury: Implications forglobal stratigraphy. Icarus 209, 210–223.

Evans, K.A., Mavrogenes, J.A., O’Neill, H.S., Keller, N.S., Jang, L.-Y., 2008. Apreliminary investigation of chlorine XANES in silicate glasses. Geochem.Geophys. Geosyst. 9, Q10003. doi:10.1029/2008GC002157.

Fedkin, A.V., Grossman, L., 2006. The fayalite content of chondritic olivine: Obstacleto understanding the condensation of rocky material. In: Lauretta, D.S.,

M.Yu. Zolotov / Icarus 212 (2011) 24–41 39

McSween, H.Y., Jr. (Eds.), Meteorites and the Early Solar System II. University ofArizona Press, Tucson, pp. 279–294.

Fegley Jr., B., 1983. Primordial retention of nitrogen by the terrestrial planets andmeteorites. J. Geophys. Res. 88 (Suppl.), A853–A868.

Fegley Jr., B., 1992. Lunar volcanic gases: The predicted presence ofchlorofluorocarbon gases. Lunar Planet. Sci. XXIII, 347–348 (abstract).

Fegley Jr., B., Cameron, A.G.W., 1987. A vaporization model for iron/silicatefractionation in the Mercury protoplanet. Earth Planet. Sci. Lett. 82, 207–222.

Fegley Jr., B., Zolotov, M.Yu., 2000. Chemistry of sodium, potassium and chlorine involcanic gases on Io. Icarus 148, 193–210.

Fincham, C.J.B., Richardson, F.D., 1954. The behavior of sulphur in silicate andaluminate melts. Proc. R. Soc. Lond. A 223, 40–62.

Fogel, R.A., 1994. Nitrogen solubility in aubrite and E chondrite melts. Lunar Planet.Sci. XXV, 383–384 (abstract).

Fogel, R.A., 2005. Aubrite basalt vitrophyres: The missing basaltic component andhigh-sulfur silicate melts. Geochim. Cosmochim. Acta 69, 1633–1648.

Fogel, R.A., Rutherford, M.J., 1995. Magmatic volatiles in primitive lunar glasses: I.FTIR and EPMA analyses of Apollo 15 green and yellow glasses and revision ofthe volatile assisted fire-fountain theory. Geochim. Cosmochim. Acta 59, 201–215.

Fogel, R.A., Hess, P.C., Rutherford, M.J., 1989. Intensive parameters of enstatitechondrite metamorphism. Geochim. Cosmochim. Acta 53, 2735–2746.

Fogel, R.A., Weisberg, M.K., Prinz, M., 1996. The solubility of CaS in aubrite silicatemelts. Lunar Planet. Sci. XXVII, 371–372 (abstract).

Garrison, D., Hamlin, S., Bogard, D., 2000. Chlorine abundances in meteorites.Meteorit. Planet. Sci. 35, 419–429.

Gibson, E.K., Moore, C.B., Primus, T.M., Lewis, C.F., 1985. Sulfur in achondriticmeteorites. Meteoritics 20, 503–511.

Gomes, R., Levison, H.F., Tsiganis, K., Morbidelli, A., 2005. Origin of the cataclysmicLate Heavy Bombardment period of the terrestrial planets. Nature 435, 466–469.

Gradie, J., Chapman, C.R., Tedesco, E.F., 1989. Distribution of taxonomic classes andthe compositional structure of the asteroid belt. In: Binzel, R.P., Gehrels, T.,Matthews, M.S. (Eds.), Asteroids, vol. II. University of Arizona Press, Tucson, pp.316–335.

Grady, M.M., Wright, I.P., Carr, L.P., Pillinger, C.T., 1986. Compositional differences inenstatite chondrites based on carbon and nitrogen stable isotopemeasurements. Geochim. Cosmochim. Acta 50, 2799–2813.

Green, D.H., 1975. Genesis of Archean peridotitic magmas and constraints onArchaean geothermal gradients and tectonics. Geology 3, 15–18.

Grimm, R.E., McSween, H.Y., 1993. Heliocentric zoning of the asteroid belt byaluminum-26 heating. Science 259, 653–655.

Grossman, L., 1972. Condensation in the primitive solar nebula. Geochim.Cosmochim. Acta 26, 597–619.

Grossman, L., Larimer, J.W., 1974. Early chemical history of the Solar System. Rev.Geophys. Space Phys. 12, 71–101.

Grossman, L., Beckett, J.R., Fedkin, A.V., Simon, S.B., Ciesla, F.J., 2008. Redoxconditions in the solar nebula: Observational, experimental, and theoreticalconstraints. Rev. Mineral. Geochem. 68, 93–140.

Gurvich, L.V., Veyts, I.V., Alcock, C.B., 1989, 4th ed. Thermodynamic Properties ofIndividual Substances, vol. 3 Hemisphere Publishing, New York.

Harmon, J.K., Slade, M.A., 1992. Radar mapping of Mercury: Full-disk images andpolar anomalies. Science 258, 640–642.

Harmon, J.K., Perillat, P.J., Slade, M.A., 2001. High-resolution radar imaging ofMercury’s north pole. Icarus 149, 1–15.

Head, J.W. et al., 2008. Volcanism on Mercury: Evidence from the first MESSENGERflyby. Science 321, 69–72.

Head, J.W. et al., 2009a. Volcanism on Mercury: Evidence from the first MESSENGERflyby for extrusive and explosive activity and the volcanic origin of plains. EarthPlanet. Sci. Lett. 285, 227–242.

Head, J.W. et al., 2009b. Evidence for intrusive activity on Mercury from the firstMESSENGER flyby. Earth Planet. Sci. Lett. 285, 251–262.

Hirschmann, M.M., Withers, A.C., 2008. Ventilation of CO2 from a reduced mantleand consequences for the early martian greenhouse. Earth Planet. Sci. Lett. 270,147–155.

Holloway, J.R., Jakobsson, S., 1986. Volatile solubilities in magmas: Transportof volatiles from mantles to planet surfaces. J. Geophys. Res. 91, D505–D508.

Holloway, J.R., Pan, V., Gudmundsson, G., 1992. High-pressure fluid-absent meltingexperiments in the presence of graphite: Oxygen fugacity, ferric/ferrous ratioand dissolved CO2. Eur. J. Mineral. 4, 105–114.

Holzheid, A., Grove, T.L., 2002. Sulfur saturation limits in silicate melts and theirimplications for core formation scenarios for terrestrial planets. Am. Mineral.87, 227–237.

Hutson, M., Ruzicka, A., 2000. A multi-step model for the origin of E3 (enstatite)chondrites. Meteorit. Planet. Sci. 35, 601–608.

Ito, K., Fruehan, R.J., 1988. Thermodynamics of nitrogen in CaO–SiO2–Al2O3 slagsand its reaction with Fe–Csat melts. Metall. Trans. B 19B, 419–425.

Ivanova, M.A., Kononkova, N.N., Krot, A.N., Greenwood, R.C., Franchi, I.A.,Verchovsky, A.B., Trieloff, M., Korochantseva, E.V., Brandstätter, F., 2008. TheIsheyevo meteorite: Mineralogy, petrology, bulk chemistry, oxygen, nitrogen,carbon isotopic compositions, and 40Ar–39Ar ages. Meteorit. Planet. Sci. 43,915–940.

Javoy, M. et al., 2010. The chemical composition of the Earth: Enstatite chondritemodels. Earth Planet. Sci. Lett. 293, 259–268.

Jee-Yon, L., Marti, K., Wacker, J.F., 2009. Xe isotopic abundances in enstatitemeteorites and relations to other planetary reservoirs. J. Geophys. Res. 114,E04003. doi:10.1029/2008JE003082.

Jessberger, E.K., 1999. Rocky cometary particulates: Their elemental, isotopic andmineralogical ingredients. Space Sci. Rev. 90, 91–97.

Kadik, A., Pineau, F., Litvin, Y., Jendrzejewski, N., Martinez, I., Javoy, M., 2004.Formation of carbon and hydrogen species in magmas at low oxygen fugacity. J.Petrol. 45, 1297–1310.

Kadik, A., Litvin, Yu.A., Koltashev, V.V., Kryukova, E.B., Plotnichenko, V.G., 2006.Solubility of hydrogen and carbon in reduced magmas of the early Earth’smantle. Geochem. Int. 44, 33–47.

Kasting, J.F., Eggler, D.H., Raeburn, S.P., 1993. Mantle redox evolution and theoxidation state of the Achaean atmosphere. J. Geol. 101, 245–257.

Keil, K., 1968. Mineralogical and chemical relationships among enstatite chondrites.J. Geophys. Res. 73, 6945–6976.

Keil, K., 1969. Titanium distribution in enstatite chondrites and achondrites, and itsbearing on their origin. Earth Planet. Sci. Lett. 7, 243–248.

Keil, K., 1989. Enstatite meteorites and their parent bodies. Meteoritics 24, 195–208.

Keil, K., Haack, H., Scott, E.R.D., 1994. Catastrophic fragmentation of asteroids:Evidence from meteorites. Planet. Space Sci. 42, 1109–1122.

Kerber, L., Head, J.W., Solomon, S.C., Murchie, S.L., Blewett, D.T., Wilson, L., 2009.Explosive volcanic eruptions on Mercury: Eruption conditions, magma volatilecontent, and implications for mantle volatile abundances. Earth Planet. Sci. Lett.285, 263–271.

Keszthelyi, L., Jaeger, W., Milazzo, M., Radebaugh, J., Davies, A.G., Mitchell, K.L.,2007. New estimates for Io eruption temperatures: Implications for the interior.Icarus 192, 491–502.

Kogarko, L.N., 1974. Role of volatiles. In: Sorensen, H. (Ed.), The Alkaline Rocks. JohnWiley, London, pp. 474–487.

Krot, A.N., Meibom, A., Russell, S.S., Alexander, C.M.O’D., Jeffries, T.E., Keil, K., 2001. Anew astrophysical setting for chondrule formation. Science 291, 1776–1779.

Krot, A.N., Keil, K., Goodrich, C.A., Scott, E.R.D., 2003. Classification of meteorites. In:Davis, A.M. (Ed.), Meteorites, Comets and Planets. Treatise on Geochemistry, vol.1. Elsevier, Oxford, pp. 83–128.

Krot, A.N., Amelin, Y., Cassen, P., Meibom, A., 2005. Young chondrules in CBchondrites from a giant impact in the early Solar System. Nature 436, 989–992.

Larimer, J.W., 1975. The effect of C/O ratio on the condensation of planetarymaterial. Geochim. Cosmochim. Acta 39, 389–392.

Larimer, J., Batholomay, M., 1979. The role of carbon and oxygen in cosmic gases –Some applications to the chemistry and mineralogy of enstatite chondrites.Geochim. Cosmochim. Acta 43, 1455–1466.

Lawrence, D.J., Feldman, W.C., Goldsten, J.O., McCoy, T.J., Blewett, D.T., Boynton,W.V., Evans, L.G., Nittler, L.R., Rhodes, E.A., Solomon, S.C., 2010. Identificationand measurement of neutron-absorbing elements on Mercury’s surface. Icarus209, 195–209.

Lellouch, E., Paubert, G., Moses, J.I., Schneider, N.M., Strobel, D.F., 2003. Volcanicallyemitted sodium chloride as a source for Io’s neutral clouds and plasma torus.Nature 421, 45–47.

Lewis, J.S., 1972. Metal–silicate fractionation in Solar System. Earth Planet. Sci. Lett.15, 286–290.

Lewis, J.S., 1988. Origin and composition of Mercury. In: Vilas, F., Chapman, C.R.,Matthews, M.S. (Eds.), Mercury. University of Arizona Press, Tucson, pp. 651–666.

Lewis, J.S., Prinn, R.G., 1984. Planets and Their Atmospheres – Origin and Evolution.Academic Press Inc., Orlando, FL. 480pp.

Libourel, G., Marty, B., Humbert, F., 2003. Nitrogen solubility in basaltic melt. Part I.Effect of oxygen fugacity. Geochim. Cosmochim. Acta 67, 4123–4135.

Lodders, K., Fegley Jr., B., 1993. Lanthanide and actinide chemistry at high C/O ratiosin the solar nebula. Earth Planet. Sci. Lett. 117, 125–145.

Lodders, K., Fegley Jr., B., 1998. The Planetary Scientist’s Companion. OxfordUniversity Press, New York.

Malavergne, V., Berthet, S., Righter, K., 2007. Formation of CaS–MgS in enstatitechondrites and achondrites as a function of redox conditions and temperature:Constraints on their evolution in a planetesimal and in a proto-planet. LunarPlanet. Sci. XXXVIII. Abstract 1737.

Malavergne, V., Toplis, M.J., Berthet, S., Jones, J., 2010. Highly reducing conditionsduring core formation on Mercury: Implications for internal structure and theorigin of a magnetic field. Icarus 206, 199–209.

Mavrogenes, J.A., O’Neill, H.St.C., 1999. The relative effects of pressure, temperatureand oxygen fugacity on the solubility of sulfide in mafic magmas. Geochim.Cosmochim. Acta 63, 1173–1180.

Mayne, R.G., McSween, H.Y., McCoy, T.J., Gale, A., 2009. Petrology of theunbrecciated eucrites. Geochim. Cosmochim. Acta 73, 794–819.

McClintock, W.E. et al., 2008. Spectroscopic observations of Mercury’s surfacereflectance during MESSENGER’s first Mercury flyby. Science 321, 62–65.

McClintock, W.E., Vervack Jr., R.J., Bradley, E.T., Killen, R.M., Mouawad, N., Sprague,A., Burger, M.H., Solomon, S.C., Izenberg, N.R., 2009. MESSENGER observationsof Mercury’s exosphere: Detection of magnesium and distribution ofconstituents. Science 324, 610–613.

McCord, T.B., Clark, R.N., 1979. The Mercury soil: Presence of Fe2+. J. Geophys. Res.84, 7664–7668.

McCoy, T.J., Dickinson, T.L., Lofgren, G.E., 1999. Partial melting of the Indarch (EH4)meteorite: A textural, chemical and phase relations view of melting and meltmigration. Meteoritics 34, 735–746.

McEwen, A.S. et al., 1998. High-temperature silicate volcanism on Jupiter’s moon Io.Science 281, 87–90.

40 M.Yu. Zolotov / Icarus 212 (2011) 24–41

McSween, H.Y., Labotka, T.C., 1993. Oxidation during metamorphism of the ordinarychondrites. Geochim. Cosmochim. Acta 57, 1105–1114.

McSween, H.Y., Mittlefehldt, D.W., Beck, A.W., Mayne, R.G., McCoy, T.J., 2010. HEDmeteorites and their relationship to the geology of Vesta and the Dawn Mission.Space Sci. Rev. doi:10.1007/s11214-010-9637-.

Melosh, H.J., Cassen, P., Sears, D., Lugmair, G., 2004. Role of planetary impacts inthermal processing of chondrite materials. In: Workshop on Chondrites and theProtoplanetary Disk, Lunar and Planetary Institute, Houston, pp. 125–126.

Mironenko, M.V., Melikhova, T.Yu., Zolotov, M.Yu., Akinfiev, N.N., 2008. GEOCHEQ-M: Program complex for thermodynamic and kinetic modeling of geochemicalprocesses in rock–water–gas systems. Version 2008. Vestnik Otdelenia Nauk oZemle RAN 1 (26). <http://www.scgis.ru/russian/cp1251/h_dgggms/1-2008/informbul-1_2008/mineral-22e.pdf>.

Miyazaki, A., Hiyagon, H., Sugiura, N., Hirose, K., Takahashi, E., 2004. Solubilities ofnitrogen and noble gases in silicate melts under various oxygen fugacities:Implications for the origin and degassing history of nitrogen and noble gases inthe Earth. Geochim. Cosmochim. Acta 68, 387–401.

Morgan, J.W., Anders, E., 1980. Chemical composition of Earth, Venus, and Mercury.Proc. Natl. Acad. Sci. 77, 6973–6977.

Moses, J.I., Rawlins, K., Zahnle, K., Dones, L., 1999. External sources of water forMercury’s putative ice deposits. Icarus 137, 197–221.

Muenow, D.W., Keil, K., Wilson, L., 1992. High-temperature mass spectrometricdegassing of enstatite chondrites: Implications for pyroclastic volcanism on theaubrite parent body. Geochim. Cosmochim. Acta 56, 4267–4280.

Mulfinger, H.O., 1966. Physical and chemical solubility of nitrogen in glass melts. J.Am. Ceram. Soc. 49, 462–467.

Murchie, S.L. et al., 2008. Geology of the Caloris basin, Mercury: A view fromMESSENGER. Science 321, 73–76.

Mysen, B.O., Yamashita, S., Chertkova, N., 2008. Solubility and solution mechanismsof NOH volatiles in silicate melts at high pressure and temperature-aminegroups and hydrogen fugacity. Am. Mineral. 93, 1760–1770.

Nicholis, M.G., Rutherford, M.J., 2009. Graphite oxidation in the Apollo 17 glassmagma: Implications for the generation of the lunar volcanic phase. Geochim.Cosmochim. Acta 73, 5905–5917.

O’Brien, D.P., Morbidelli, A., Levison, H.F., 2006. Terrestrial planet formation withstrong dynamical friction. Icarus 184, 39–58.

Oppenheimer, C., 2003. Volcanic degassing. In: Rudnick, R.L. (Ed.), The Crust.Treatise in Geochemistry, vol. 3. Elsevier. pp. 123–166.

Pan, V., Holloway, J.R., Hervig, R.L., 1991. The pressure and temperature dependenceof carbon dioxide solubility in tholeiitic basalt melts. Geochim. Cosmochim.Acta 55, 1587–1595.

Pankratz, L.B., Mah, A.D., Watson, S.W., 1987. Thermodynamic Properties ofSulfides. US Bureau of Mines Bulletin 689.

Pasek, M.A., Milsom, J.A., Ciesla, F.J., Lauretta, D.S., Sharp, C.M., Lunine, J.I., 2005.Sulfur chemistry with time-varying oxygen abundance during Solar Systemformation. Icarus 175, 1–14.

Petaev, M.V., Khodakovsky, I.L., 1986. Thermodynamic properties and conditions offormation of minerals in enstatite chondrites. In: Saxena, S.K. (Ed.), Chemistryand Physics of Terrestrial Planets. Springer-Verlag, New York e.a., pp. 106–135.

Potter, A.E., Morgan, T.H., 1985. Discovery of sodium in the atmosphere of Mercury.Science 229, 651–653.

Potter, A.E., Morgan, T.H., 1986. Potassium in the atmosphere of Mercury. Icarus 67,336–340.

Prockter, L.M., Ernst, C.M., Denevi, B.W., Chapman, C.R., Head, J.W., Fassett, C.I.,Merline, W.J., Solomon, S.C., Watters, T.R., Strom, R.G., 2010. Evidence for youngvolcanism on Mercury from the third MESSENGER flyby. S cience 329, 668–671.

Rava, B., Hapke, B., 1987. An analysis of the Mariner 10 color ratio map of Mercury.Icarus 71, 397–429.

Righter, K., Drake, M.J., Scott, E., 2006. Compositional relationships betweenmeteorites and terrestrial planets. In: Lauretta, D., McSween, H.Y. (Eds.),Meteorites and the Early Solar System, vol. II. University of Arizona Press,Tucson, pp. 803–828.

Riner, M.A., Bina, C.R., Robinson, M.S., Desch, S.J., 2008. Internal structure ofMercury: Implications of a molten core. J. Geophys. Res. 113, E08013.doi:10.1029/2007JE002993.

Riner, M.A., McCubbin, F.M., Lucey, P.G., Taylor, G.J., Gillis-Davis, J.J., 2010. Mercurysurface composition: Integrating petrologic modeling and remote sensing datato place constraints on FeO abundance. Icarus 209, 301–313.

Robie, R.A., Hemingway, B.S., 1995. Thermodynamic Properties of Minerals andRelated Substances at 298.15 K and 1 Bar (105 Pascals) Pressure and at HigherTemperatures. US Geol. Surv. Bull. 2131. US GPO, Washington, DC.

Robinson, M.S., Lucey, P.G., 1997. Recalibrated Mariner 10 color mosaics:Implications for mercurian volcanism. Science 275, 197–200.

Robinson, M.S., Taylor, G.J., 2001. Ferrous oxide in Mercury’s crust and mantle.Meteorit. Planet. Sci. 36, 841–847.

Robinson, M.S. et al., 2008. Reflectance and color variations on Mercury: Indicatorsof regolith processes and compositional heterogeneity. Science 321, 66–69.

Roskosz, M., Mysen, B.O., Cody, G.D., 2006. Dual speciation of nitrogen in silicatemelts at high pressure and temperature: An experimental study. Geochim.Cosmochim. Acta 70, 2902–2918.

Rubin, A.E., Choi, B.-G., 2009. Origin of halogens and nitrogen in enstatitechondrites. Earth Moon Planets 105, 41–53.

Rubin, A.E., Kallemeyn, G.W., Wasson, J.T., Clayton, R.N., Mayeda, T.K., Grady, M.,Verchovsky, A.B., Eugster, O., Lorenzetti, S., 2003. Formation of metal andsilicate globules in Gujba: A new Bencubbin-like meteorite fall. Geochim.Cosmochim. Acta 67, 3283–3298.

Saal, A.E., Hauri, E.H., Cascio, M.L., van Orman, J.A., Rutherford, M.C., Cooper, R.F.,2008. Volatile content of lunar volcanic glasses and the presence of water in theMoon’s interior. Nature 454, 192–195.

Sato, M., 1976. Oxygen fugacity and other thermochemical parameters of Apollo 17high-Ti basalts and their implications on the reduction mechanism. Proc. LunarSci. Conf. VII, 1323–1344.

Schaefer, L., Fegley Jr., B., 2005a. Alkali and halogen chemistry in volcanic gases onIo. Icarus 173, 454–468.

Schaefer, L., Fegley Jr., B., 2005b. Silicon tetrafluioride on Io. Icarus 179, 252–258.

Scott, E.R.D., 2007. Impact origins for pallasites. Lunar Planet. Sci. XXXVIII. Abstract2284.

Shilobreeva, S., Kadik, A., Matveev, S., Chappyzhnicov, B., 1994. Solubility of N2 inbasalt and albite melt at pressures 1–3 kbar and temperature 1250 �C. In:International Workshop on Noble Gas Geochemistry and Cosmochemistry,Yamada Conf. XXXVIII, pp. 21–22 (abstract).

Shukolyukov, A., Lugmair, G.W., 2004. Manganese–chromium isotope systematicsof enstatite meteorites. Geochim. Cosmochim. Acta 68, 2875–2888.

Siebert, J., Malavergne, V., Guyot, F., Combes, R., Martinez, I., 2004. The behaviour ofsulphur in metal–silicate core segregation experiments under reducingconditions. Phys. Earth Planet. Inter. 144, 421–432.

Slade, M.A., Butler, R.P., Muhleman, D.O., 1992. Mercury radar imaging: Evidence forpolar ice. Science 258, 635–640.

Solomon, S.C., 2003. Mercury: The enigmatic innermost planet. Earth Planet Sci.Lett. 216, 441–455.

Solomon, S.C. et al., 2008. Return to Mercury: A global perspective on MESSENGER’sfirst Mercury flyby. Science 321, 59–62.

Sprague, A.L., Kozlowski, R.W.H., Hunten, D.M., 1990. Caloris Basin – An enhancedsource for potassium in Mercury’s atmosphere. Science 249, 1140–1143.

Sprague, A.L., Hunten, D.M., Lodders, K., 1995. Sulfur at Mercury, elemental at thepoles and sulfides in the regolith. Icarus 118, 211–215.

Sprague, A.L., Emery, J.P., Donaldson, K.L., Russell, R.W., Lynch, D.K., Mazuk, A.L.,2002. Mercury: mid-infrared (3-13.5 lm) observations show heterogeneouscomposition, presence of intermediate and basic soil types, and pyroxene.Meteorit. Planet. Sci. 37, 1255–1268.

Sprague, A.L., Donaldson, H.K.L., Kozlowski, R.W.H., Helbert, J., Maturilli, A., Warell,J.B., Hora, J.L., 2009. Spectral emissivity measurements of Mercury’s surfaceindicate Mg- and Ca-rich mineralogy, K-spar, Na-rich plagioclase, rutile, withpossible perovskite, and garnet. Planet. Space Sci. 57, 364–383.

Spudis, P.D., Guest, J.E., 1988. Stratigraphy and geologic history of Mercury. In: Vilas,F., Chapman, C.R., Matthews, M.S. (Eds.), Mercury. University of Arizona Press,Tucson, pp. 118–164.

Stevenson, D.J., Lunine, J.I., 1988. Rapid formation of Jupiter by diffuse redistributionof water vapor in the solar nebula. Icarus 75, 146–155.

Strom, R.G., Trask, N.J., Guest, J.E., 1975. Tectonism and volcanism on Mercury. J.Geophys. Res. 80, 2478–2507.

Sugiura, N., Zashu, S., Weisberg M.K., Prinz, M., 1999. H, C, and N isotopiccompositions of bencubbinites. Lunar Planet. Sci. XXX. Abstract 1329.

Symonds, R.B., Rose, W.I., Bluth, G.J.S., Gerlach, T.M., 1994. Volcanic-gas studies:Methods, results and applications. In: Carroll, M.R., Holloway, J.R. (Eds.),Volatiles in Magmas. Rev. Mineral, vol. 30. Mineral. Soc. Am., Washington, DC,pp. 1–66.

Taylor, G.J., Scott, E.R.D., 2003. Mercury. In: Davis, A.M. (Ed.), Meteorites, Cometsand Planets, Treatise on Geochemistry, vol. 1. Elsevier, pp. 477–485.

Usselman, T.M., Lofgren, G.E., 1976. The phase relations, textures, and mineralchemistries of high-titanium mare basalts as a function of oxygen fugacity andcooling rate. Proc. Lunar Sci. Conf. 7, 1345–1363.

Van Zeggeren, F., Storey, S.H., 1970. The Computation of Chemical Equilibria.Cambridge Univ. Press, Cambridge.

Vernazza, P., DeMeo, F., Nedelcu, D.A., Birlan, M., Doressoundiram, A., Erard, S.,Volquardsen, E., 2010. Resolved spectroscopy of Mercury in the near-IR withSpeX/IRTF. Icarus 209, 125–137.

Vilas, F., 1988. Surface composition of Mercury from reflectance spectrophotometry.In: Vilas, F., Chapman, C.R., Matthews, M.S. (Eds.), Mercury. University ofArizona Press, Tucson, pp. 59–76.

Wadhwa, M., 2008. Redox conditions on small bodies, the Moon and Mars. Rev.Mineral. Geochem. 68, 493–510.

Warell, J., Sprague, A.L., Emery, J.P., Kozlowski, R.W.H., Long, A., 2006. The 0.7–5.3spectra of Mercury and the Moon: Evidence for high-Ca clinopyroxene onMercury. Icarus 180, 281–291.

Warell, J., Sprague, A., Kozlowski, R., Rothery, D.A., Lewis, N., Helbert, J., Cloutis, E.,2010. Constraints on Mercury’s surface composition from MESSENGER andground-based spectroscopy. Icarus 209, 138–163.

Warren, P.H., Kallemeyn, G.W., 1992. Explosive volcanism and the graphite–oxygenfugacity buffer on the parent asteroid(s) of the ureilite meteorites. Icarus 100,110–126.

Wasson, J.T., 1988. The building stones of the planets. In: Vilas, F., Chapman, C.R.,Matthews, M.S. (Eds.), Mercury. University of Arizona Press, Tucson, pp. 622–650.

Watters, T.R., Prinz, M., 1979. Aubrites: Their origin and relationship to chondrites.Proc. Lunar. Sci. Conf. X, 1073–1093.

Weidenschilling, S.J., 1978. Iron/silicate fractionation and the origin of Mercury.Icarus 35, 99–111.

Weisberg, M.K., Prinz, M., Clayton, R.N., Mayeda, T.K., Sugiura, N., Zashu, S., Ebihara,M., 2001. A new metal-rich chondrite grouplet. Meteorit. Planet. Sci. 36, 401–418.

M.Yu. Zolotov / Icarus 212 (2011) 24–41 41

Wetherill, G.W., 1986. Accumulation of the terrestrial planets and implicationsconcerning lunar origin. In: Hartmann, W.K., Phillips, R.J., Taylor, G.J. (Eds.),Origin of the Moon. Lunar and Planetary Institute, pp. 519–550.

Wilson, L., Head, J.W., 2008. Volcanism on Mercury: A new model for the history ofmagma ascent and eruption. Geophys. Res. Lett. 35, L23205. doi:10.1029/2008GL035620.

Wilson, L., Keil, K., 1991. Consequences of explosive eruptions on small Solar Systembodies – The case of the missing basalts on the aubrite parent body. EarthPlanet. Sci. Lett. 104, 505–512.

Yang, J., Goldstein, J.I., Scott, E.R.D., 2007. Iron meteorite evidence for earlyformation and catastrophic disruption of protoplanets. Nature 446, 888–891.

Zolotov, M.Yu., Fegley Jr., B., 1999. The oxidation state of volcanic gases and interiorof Io. Icarus 141, 40–52.

Zolotov, M.Yu., Mironenko, M.V., 2007. Hydrogen chloride as a source of acidfluids in parent bodies of chondrites. Lunar Planet. Sci. XXXVIII. Abstract2340.

Zurbuchen, T.H., Raines, J.M., Gloeckler, G., Krimigis, S.M., Slavin, J.A., Koehn, P.L.,Killen, R.M., Sprague, A.L., McNutt, R.L., Solomon, S.C., 2008. MESSENGERobservations of the composition of Mercury’s ionized exosphere and plasmaenvironment. S cience 321, 90–92.