Accretion of Jupiter's atmosphere from a supernova-contaminated molecular cloud
aa r X i v : . [ a s t r o - ph . E P ] D ec Accretion of Jupiter’s Atmosphere from aSupernova-Contaminated Molecular CloudSubmitted to Icarus 14-Oct-2008Revised 24-Nov-2009Accepted 9-Feb-2010
Henry B. ThroopSouthwest Research Institute1050 Walnut St, Ste 300, Boulder, CO 80302 [email protected]
John BallyCenter for Astrophysics and Space AstronomyUniversity of Colorado, BoulderUCB 389, Boulder, CO 80309-0389Received ; accepted 2 –
ABSTRACT
If Jupiter and the Sun both formed directly from the same well-mixed proto-solar nebula, then their atmospheric compositions should be similar. However,direct sampling of Jupiter’s troposphere indicates that it is enriched in elementssuch as C, N, S, Ar, Kr, and Xe by 2–6 × relative to the Sun (Wong et al. e.g. Trinquier et al.
1. Introduction
The Solar System’s composition reflects that of the initial cloud core from which itformed. On small bodies such as the terrestrial planets, subsequent thermal and chemicalprocesses have altered the composition. However, for the Solar System’s two largest bodies– the Sun and Jupiter (mass 1 M J ≈ M ⊙ ) – the atmospheric composition is relativelystable against change since the time of formation. In particular, the relative atmospheric 3 –abundances of gases carbon (C), nitrogen (N), sulfur (S), argon (Ar), krypton (Kr), andxenon (Xe) are believed to be fixed at the time of formation 4.5 billion years ago, andnot affected substantially by subsequent evolution. The noble gases are stable against lossdue to Jupiter’s high escape velocity, and stable against sinking by condensation becausethey don’t combine with other atomic species. Elements C, N, and S combine into largermolecules such as CH , NH , and H S, but are believed to remain well mixed in Jupiter’stroposphere without settling to the interior.Jupiter’s composition at 5–20 bars was directly measured by the Galileo descent probe’smass spectrometer in 1995. Initial results (Atreya et al. ∼ ± × relative to the then-current solar abundances ofAnders and Grevesse (1989, hereafter AG89). More recent work has better determined theSolar abundances, leading to revised Jupiter:Solar abundances. Somewhat surprisingly,these new results actually increased the scatter between the elements, to the broaderrange of 4 ± × (Wong et al. . ± . × solar (Grevesse et al. . ± . × (Lodders 2008)(hereafter Lod08). All enrichments in this paper are in terms of numberdensity relative to hydrogen, normalized to the Sun; i.e. , ( n i /n H ) / ( n i /n H ) sun .
2. Previous models
Several previous models have been proposed to address Jupiter’s ‘metallicity problem.’All are based on condensation of the volatile elements into ices, which can then beconcentrated relative to hydrogen. The ‘amorphous ice model’ (Owen and Encrenaz 2006;Owen et al. et al. et al. et al. <
20 Kat 5 AU at several Myr (Hersant et al. et al. (2007) measures20–50 K at 50 AU, and much warmer inward, and Dutrey et al. (2003) measures 50 K at100 AU. Models by Chiang and Goldreich (1997) predict 75 K at 5 AU and 55 K at 10 AU;work by Dullemond and Dominik (2005) assumes 90 K at 5 AU and 30 K at 50 AU; andthe model of (Jonkheid et al. et al. et al. et al. e.g. , 10–50 AU) where temperatures were lower.Radial migration subsequently brought them to 5 AU. For instance, Alibert et al. (2005)explores the possibility that Jupiter formed in its entirety at 10–15 AU via the clathratemodel, and then migrated inward. Guillot and Hueso (2006) use photo-evaporation andviscous migration of the entire nebula to explain Jupiter’s enrichment. In this work, UVflux from internal and/or external stars heats and preferentially removes hydrogen fromthe inner proto-planetary disk. These planetesimals then migrate inward where they heatand go into forming Jupiter. This model is appealing, and consistent with the idea thatthe young solar system may have experienced the effects of nearby massive stars ( e.g. ,Tachibana and Huss 2003; Throop et al. et al. et al. et al. × in all species, fitting the original Galileodata very well. However, no SCIPs have ever been identified in the Solar System: comets,for instance, are depleted substantially in N and are of decidedly non-SCIP composition(Iro et al. × ). All of themodels attempt to fit the roughly 3 ± × enhancement of the original Galileo values, anddo not try to fit the newer 4 ± × values; Owen and Encrenaz (2006) dismisses the revisedvalues as being possibly due to systematic errors. None of the models fit the high Ar valueof 5 . ± . × of GAS07, though the 2 . ± . coldest condensation temperature ( ∼
35 K into clathrates, and 20 K as a solid)of all the species measured, so its high GAS07 value would be a particular challenge to allthe condensation models.Finally, Lodders (2004) proposed an entirely different model: that Jupiter formed fromcarbon-rich (rather than ice-rich) planetesimals. However, this scenario depends on the 7 –assumption that the Galileo probe sampled a typical region of Jupiter’s atmosphere, andmost work supports the opposite view that the probe hit an anomalously dry spot.
3. Our model: A contaminated molecular cloud
We propose an entirely different solution to the problem. Rather than forming Jupiterand the Sun from identical ‘Solar nebula’ material and invoking condensation or transportwithin the nebula to modify Jupiter’s composition, we propose that the composition ofJupiter and the Sun differ because of intrinsic temporal and spatial variations in theSolar nebula composition as the interstellar medium (ISM) is polluted by massive stellarwinds and supernovae (SNs). We show that this model can explain Jupiter’s chemicalcomposition, easily fits into environmental formation scenarios, and is consistent with otherheterogeneities in both our Solar System and distant star clusters.In the scenario that we propose, multiple stages of star formation occurred withina giant molecular cloud (Fig. 1). The cloud was of average size, 10–20 pc, and withinthe GMC lay several pc-scale molecular clouds. Stars, including the Sun with its disk,began to form in clusters within these clouds. The Sun’s orbit through the cluster tookit on a long path several pc across. Before any nearby O/B stars turned on, the clusterenvironment remained cool and dark for several Myr. During this time the Sun passedthrough the ISM and gravitationally swept up material onto its disk by Bondi-Hoyleaccretion (Throop and Bally 2008).If the Sun formed alone, then the composition of the ISM might be uniform. But in ourproposed scenario, other clusters nearby formed a few Myr earlier, some with higher-massstars, and the ISM surrounding these clusters soon began to be polluted in heavy elementsproduced by these massive stars. Nucleosynthesis in these stars enriched their ejecta by 8 – ∼ × or more relative to Solar H. The Sun’s evolving orbit took it through these morepolluted regions of the ISM. Material was accreted onto both the Sun and its disk, but thedisk’s large cross-section and low mass made it easier to pollute than the Sun. The disk’smetallicity slowly increased, and Jupiter’s core and atmosphere formed from the disk. Thedisk dispersed within 5–10 Myr, around the same time as the local ISM dispersed and thecluster spread apart, ejecting the Sun as an unbound field star. Jupiter’s final compositionthus represented largely the same material as the Sun, but with the late accretion ofpolluted material reflecting in Jupiter’s enriched composition today.Our proposed model reflects much of the current understanding of star formation inlarge clusters and molecular clouds ( e.g. Bally 2008). Stars do not form in isolation, theISM is not of uniform composition, stars travel on long orbits through their young clusters,and material from the ISM can be accreted onto young stars and disks in the severalMyr after they form. All of these are newly appreciated processes that have not beenincorporated into existing models of Solar System formation. We find Jupiter’s metallicityto be one natural consequence of these processes, and we explore here the issues involvedwith it. In the following sections we examine the detailed chemical constraints on Jupiter’scomposition from such a model ( § § §
4. Chemical constraints
In this section we investigate the detailed chemical enrichment from this ‘pollutedaccretion’ scenario and whether it can explain Jupiter’s current composition.We start by assuming that Jupiter’s atmosphere is composed of a mixture of threedistinct components. The majority of the mass is made of Solar material, whose composition 9 –is well determined by the measurements of GAS07 and AG89. The remainder is thesmall amounts of pollution that come from massive stellar winds and/or supernovae. Thecompositions of these ejecta are distinct ‘fingerprints’ of the stars, determined mostlyby their initial mass and composition. In general, the winds of these stars are enrichedprimarily in light elements (C, N, O, Ne), while the SN ejecta several Myr later containheavier species (S, Ar, Kr, Xe). As one example, the demise of a star with an initial massof 20 M ⊙ can produce 5 M ⊙ of O, 0.5 M ⊙ of C, 0.2 M ⊙ of Si, 0.001 M ⊙ of Al, and 10 − M ⊙ of Fe in its SN phase (Woosley and Heger 2007, hereafter WH07). In our models, weuse the grid of winds and ejecta computed by WH07. These models span the mass range12–40 M ⊙ , and consider a variety of nucleosynthesis rate coefficients and explosion energyparameters, for a total of 66 models.We denote the elemental abundances in the Solar, wind, and SN components as n ⊙ , n w , and n sn ; the Jupiter abundance is n J . It is then possible to calculate model Jupitercompositions by using n J , i = f ⊙ n ⊙ ,i + f w n w , i + f sn n sn , i , (1)where i is the species and f ⊙ + f sn + f w = 1. We use the present-day solar abundance,but the difference between this and the Sun’s primordial composition is insignificant for ourpurposes.Using Eq. 1, we found combinations of solar material and ejecta that would produceJupiter’s measured composition. We searched all 66 = 4356 possible combinations of the66 wind and 66 SN models of WH07, coupled with the single solar abundance. For eachtrial, we computed coefficients that best fit Jupiter. Our routine attempted to fit only thewell measured stable species (C, N, S, Ar, Kr, Xe), and computed results for both theseand the remaining elements (He, O, Ne, P).Our best fit (‘Model A’) is shown in Figures 2–3. This model finds Jupiter’s composition 10 –to be well described by 87% solar nebula, 9% stellar winds from a 40 M ⊙ star (WH07’smodel s40a28A ), and 4% supernova ejecta from a 20 M ⊙ star (WH07’s s20a37n ). Thetotal contamination is 13% ( i.e. , 0.13 M J ). The fit is excellent at matching the observedquantities of C, S, Ar, and Kr, and the lower limit for O. The largest deviation is for N,where we are slightly below the error bar. The wind predominantly supplies C, N, and Owhile the SN supplies the remaining species. Both stars have high enough mass (and thusshort enough lifetimes) that they can form and explode within the 10 Myr timeframe ofGMCs. We assume the GAS07 and WLA08 values for the Solar and Jovian composition.This model is shown here because it is the best fit; many other combination of differentmass SN and wind ejecta provided much worse fits.A variant of this fit (’Model A2’) is shown in Figures 4–5. In this case we have usedthe latest Lod08 value for Ar, instead of that of GAS07. This model is fit with 78% solarnebula, 8% winds from a 40 M ⊙ star (WH07’s s40a28A ), and 14% SN ejecta from a 15 M ⊙ star (WH07’s s15a34c ). The only difference to the fit is the Ar abundance. Becauselower-mass SNs are less efficient at heavy-element nucleosynthesis, the fit here requires asubstantially larger SN contribution than does Model A (14% vs. 4%). The 15 M ⊙ star hasa lifetime of ∼
11 Myr, putting it on the upper end of individual cloud lifetimes but withinthe timescale of large regions like Orion.Finally, a third fit is shown in Figures 6–7 (‘Model B’). This model differs in that wehave used the abundances of AG89 and AMN03 for n sol and n Jup . Although the newerabundances are probably preferred, using the old ones gives a test of the robustness of ourfits. Also, the nucleosynthetic yields of WH07 start with the AG89 solar abundances, so ina sense this fit is more self-consistent, even though it is based on slightly older data. ModelB requires about 6% total contamination, less than half that required by Model A. ModelB is comprised of 94% solar, 4% stellar winds from a 40 M ⊙ star ( s40a28A ), and 1.5% SN 11 –ejecta from a 25 M ⊙ star ( s25a41d ).All three of our models fit the data well. In all cases the vast majority of Jupiter’smass (78%-95%) comes from the Solar nebula, in agreement with standard models. Andin each, a combination of winds and SN – which one would expect in a realistic cluster –works better than any single component by itself. All three fits are slightly low in N andXe; the remaining species are fit very well.The individual ‘fingerprints’ of the SNs can be seen in Figures 7, 5, and 3. In all themodels, the lighter species (C, N, O) are produced in the winds, while the heavier elementscome from the SNs and are highly dependent on the SN mass. A large difference betweenthe three models is the SN C:N ratio, which is several times higher in Model B than theothers (compare red curves in 5 and 3). The great deal of carbon ejected by the Model BSN allows the total contamination in this model to be about half that in Model A.Detailed yields from our three models are listed in Table 1. This table lists additionalspecies measured by Galileo, but which are not in equilibrium at the entry site andthus not expected to fit: He, Ne, and P. Helium and Ne are believed to combine intoHe-Ne ‘raindrops’ which sink to Jupiter’s interior, and cannot be used as a constraint(Roulston and Stevenson 1995). Helium itself is produced by the Sun so its primordialSolar abundance cannot be directly measured. O was measured at Jupiter but its valueis believed to be anomalously low due to the probe’s dry entry site. Encouragingly, ourmodels predict primordial Jovian abundances for O, Ne, and P similar to those of the otherspecies. The predicted values for O are in the range 2.6–2.9, similar to global O valuesdetermined spectroscopically (WLA08).Several changes could improve the quality of our fits. First, we have used a simplefitting method, assuming contamination by only one wind and one supernova. In realisticstar-forming regions such as Orion there are several dozen stars within a few pc all above 12 –8 M ⊙ that will explode as SN; using multiple stars will increase the ease of fitting Jupiter’scomposition. Second, the SN ejecta models we use are quantized in relatively large massbins (5 M ⊙ ), and all use certain common assumptions for stellar and explosion parameters.The SN ejecta yields are very model dependent; for instance, the models of Young and Fryer(2007) vary in abundance for individual species by 50% or more from those of WH07 forstars of similar mass. The WH07 yields ignore stellar rotation, which may be important(Hirschi et al. σ errorbar. However, Xe condenses at a substantially warmer temperature than any of the othernoble gases (at ∼
55 K into clathrates, it is the easiest to condense), so if our model were toinclude condensation explicitly, it would operate in the direction to correct this deficiency.For now, however, we have intentionally chosen to keep our models simple to demonstratethat good fits are possible even with a very limited set of parameters.
5. Stellar Pollution into the ISM
Now that we have shown that Jupiter can be matched chemically, we investigate therequirements on the environment to support such contamination.As massive stars in a cluster evolve, they perform nucleosynthesis and create heavyelements that pollute the ISM. These are given off in stellar winds (during the stellarlifetime) or SNs (at the end-of-life for stars with M > M ⊙ ). After being injected intothe ISM, these highly enriched ejecta are incorporated into the next generation of stars.The highest-mass stars are the shortest lived: for instance, the 40 M ⊙ stars used in ourfits explode in less than 5 Myr, allowing for a full generation of star formation within the 13 –10 Myr timescales of planet formation and embedded clusters. Stars spend the majority of their lives on the main sequence. During this stage thestellar winds are weak enough that they do not pollute the cluster; for instance, the currentsolar wind is ˙ M ⊙ ≈ × − M ⊙ yr − with v ≈
400 km s − . Red supergiant (RSG)stars, however, have slow massive winds that can easily pollute the cluster. RSGs arenormal post-main-sequence stars of mass 5–15 M ⊙ that have cooled dramatically after theirmain-sequence phase (T > < M ≈ − –10 − M ⊙ yr − and v as low as 10 km s − – ten order of magnitude more loss thanthe present-day Sun (Knapp and Woodhams 1993). The RSG phase lasts for 5–10% of thestellar lifetime, or typically a few 10 –10 yr (Schaller et al. > M ⊙ , overhalf the original stellar mass can be lost during the RSG phase (Garcia-Segura et al. − is only slightly higher than the stellar and gas velocities, sothis material can readily mix with the local ISM. The abundances n w that we consider in § M ⊙ evolve tooslowly ( >
10 Myr) to enter the RSG phase within typical cluster lifetimes. Our Model Aand B fits use winds from stars of 40 M ⊙ with lifetimes ∼ N > a few thousand; an exampleis θ Ori C in the Orion Trapezium core. Smaller 8 M ⊙ stars are produced in clusters of N > a few hundred. 14 –
After passing through the RSG phase, stars with masses > M ⊙ usually end theirlives as SNs. The SN explosion gives off 1–5 M ⊙ of metal-enriched material at speeds of2,000–10,000 km s − . Even at such high speeds, numerical simulations by Ouellette et al. (2007) showed that disks at 1 pc are very resilient to nearby explosions. However, althoughthe disk will absorb most of the intercepted solids such as Fe, most of the SN’s gas isdeflected by the disk. For a 20 M ⊙ explosion at 1 pc, the disk will absorb just 10 − M J ofenriched gas, insufficient for the process described here.But, there are at least two mechanisms by which ejecta may cool and then be accretedonto the disk. First, as the ejecta spreads, it mixes with the ISM until it slows and cools,and this cooler ejecta can be accreted more easily. In order to slow to 10 km s − , 1 M ⊙ of SN ejecta must mix with roughly 1000 M ⊙ of ISM ( i.e. , a 1000:1 mixing ratio). TheSN ejecta can be quite clumpy, with high mixing ratios in some regions and low in others.Rather than mixing uniformly, these clumps appear to be slowed as a unit and preservetheir density, much like a baseball is slowed in the air without fully mixing. Observationsof pc-scale ejecta from various SNs show highly clumpy knots on AU scales or greatermaintained for a year after explosion (Fesen et al. et al. et al. et al. et al. × in molecular clouds (Mesa-Delgado et al. et al. et al. Once enriched material has mixed with the ISM by either of these two methods,accretion onto the disk is straightforward. Throop and Bally (2008) found that the averageISM-to-disk accretion rate for disks in young clusters was 10 − M ⊙ yr − , or ∼ . M J = 10 − MMSN = 10 − M ⊙ ,or the amount delivered in ∼
6. Heterogeneity in the Solar System and star-forming regions
Our model reflects the growing body of evidence that the Solar System did not formfrom a homogeneous cloud in an isolated environment, but rather from a heterogeneousnebula where interactions with its environment played a major role in shaping its evolution.We briefly discuss here four examples of large-scale heterogeneity: two in the terrestrialplanets, one in star-forming regions, and one on galactic scales.The terrestrial bodies have been modified by scores of chemical and physical processessince their formation. Most of these processes act equally on all isotopes of the sameelement, so isotopic differences are not expected in samples formed from a well-mixednebula. However, isotopic variations have indeed been measured in samples from theEarth, Mars, and numerous asteroids. Isotopic differences have been measured for speciesincluding Ba, Cr, S, Ti, Zi, Mb, O, and more (Trinquier et al. et al. Cr/ Cr, and as small as ppm for some others) but are indisputable. UV photochemistryhas been invoked to explain the origin of the O variations (Lyons and Young 2005), butthe remaining heterogeneities have defied explanation by known fractionation processes.Instead, they are consistently thought to be of nucleosynthetic origin, resulting from theincomplete mixing of ejecta from multiple SNs in the material of the young solar nebula(Ranen and Jacobsen 2008; Trinquier et al. et al. et al. (2008) examined the heterogeneity of Al isotopes within Ca-Al inclusions(CAIs) in carbonaceous chondrites. They found differences of > × in primordial Alabundances, leading to the conclusion that there are at least two populations of CAIs: 17 –some which were formed in the presence of Al, and some which were not. This requireseither spatial or temporal variations in the Solar System birth environment. Their preferredinterpretation is that the Al-free CAIs were formed early (possibly during initial collapseof the Solar System), followed by late injection of Al from a nearby massive star. Theysuggested that this massive star could be the same one that injected Fe several Myr laterafter an SN explosion. We caveat this point with mention that a recent study of CAIswithin a single ordinary chondrite did not reproduce the Al variations (Villeneuve et al. × between stars of the same age inthe same subgroup (Cunha et al. O/ O, and the solar value is higher than nearly every molecular cloud or YSOwithin 10 kpc. This enhancement has been proposed to be due to accretion of supernovaejecta and massive stellar winds immediately prior to the proto-Solar nebula’s formation(Young et al. initial solarnebula, Jupiter’s enhancement could be caused by post-collapse contamination of the disk,from Bondi-Hoyle accretion onto the disk over several Myr and several pc, allowing for farmore introduced pollution across the entire molecular cloud. The amount of heterogeneityseen at Jupiter is comparable or smaller than that seen in both star-forming regions (ADFs)and young stars themselves. Because the physical processes of star formation are universal,if the Sun formed in a dense cluster, then the solar system and Jupiter could have naturallyinherited these heterogeneities.
7. Discussion
The model we propose here is a departure from existing models for the solution toJupiter’s metallicity problem. Although historically most Solar System formation modelshave assumed a homogeneous ‘Solar nebula’ composition, our model explicitly assumes theopposite. We incorporate the fact that the composition of the solar nebula can changespatially and temporally. These changes occur during the Solar System’s first 5–10 Myr,as the Sun is traveling through its birth cloud, experiencing the environmental effects ofother stars. The inclusion of this contamination reflects the latest understandings of theenvironmental processes affecting star formation.Our model has three general advantages over the existing amorphous ice and clathratemodels that we describe in §
2. First, it relaxes the very strict low-temperature requirementsfor the formation of Jupiter’s solids. Second, it allows for Jupiter to have formed at its 19 –present location without migration. Third, it explains the fact that different elements areenriched by different amounts ( i.e. , not a uniform 3 × ).Our model is not without problems. Our fits for N are at the edge (or a little beyond)the 1 σ level. Nitrogen is poorly measured in the Galileo probe data, but is a major speciesand an important constraint. We could increase the contribution from stellar winds (whichsupply much of the N), but this would increase the amount of C beyond that observed.Low-mass stars of 4–8 M ⊙ produce large quantities of N sufficient to solve the problem,but these stars have lifetimes of 30 Myr or more before they enter the RSG phase. Theother specie our models are all low in, Xe, has a particularly high condensation temperature(making it easy to condense), and this may in part explain its abundance.Second, the amount of SN contamination required is on the high end of what can besupplied by nearby SNs using direct injection of ejecta into a nearby GMC. Additionalwork is necessary to understand the mixing and concentration mechanisms of this ejecta.However, if the ‘droplet’ model of SN ejecta is correct, then concentrated ejecta can beslowed and cooled over kpc distances, which would solve this problem.Most work suggests that Saturn, Uranus, and Neptune accreted their atmospheresfrom the disk in much the same way Jupiter did. The metallicity enrichments of the outerplanets exceed Jupiter’s – Uranus and Neptune have 30 times the Solar C:H ratio, forinstance – suggesting that they were formed at least in part by the traditional chemicalcondensation series (Lewis 1972). The colder nebula and slower evolution make volatilecondensation easier at greater distances. However, ‘accreted pollution’ of the disk from theISM would affect these planets as well. Because of their different positions in the disk andtheir different formation times, our model cannot predict the enrichment they may receivefrom pollution, except that it could be similar to Jupiter’s. (The outer disk, with its largercross-section, might be more easily contaminated, but this requires more study.) More 20 –detailed analysis of this awaits better knowledge of their bulk atmospheric compositions,which are largely unknown today. Jupiter’s oxygen composition will be probed by Junoin 2017 and will provide a discriminant between many models including our own (whichpredicts a global O:H ratio of ∼ × solar) and that of Lodders 2004 (which assumes ∼ . × solar).Jupiter’s atmosphere may have formed through a variety of processes, of which pollutedaccretion could be only one. For instance, N is among the most difficult species to condensein the various ice models, as it requires a temperature ≤
30 K. Nitrogen, however, is easilysupplied in the cool, slow stellar winds that we study here. It is possible that N wassupplied by these winds, and other species were delivered in part by icy planetesimals froma relatively warm ( ≥
50 K) disk.An advantage of the SN model is that it provides a mechanism for matching not onlythe chemical abundances which we study here, but also isotopic differences. It has alreadybeen shown that SNs may explain the isotopic differences seen in the Solar System’s rockybodies ( § e.g. , Wiens et al. et al. → disk → planet that we describe here hasbroader applications for the formation of extrasolar planets. A strong observed correlationexists between the metallicity of the host star and the existence of extrasolar planets ( e.g. ,Udry et al.
8. Acknowledgments
We thank S. Atreya, W. Bottke, F. Ciesla, A. Heger, T. von Hippel, H. Levison, A.Morbidelli, and M. Wong for useful discussions. Heger and S. Woosley also kindly providedthe ejecta compositional data which we use here. HT and JB graciously acknowledgesupport from NASA Origins grants NNG06GH33G and NNG05GI43G; HT acknowledgessupport from NASA Exobiology grant NNG05GN70G, and JB acknowledges support fromthe University of Colorado Center for Astrobiology, which is supported by the NASAAstrobiology Institute. 22 –Fig. 1.— The proposed ‘polluted accretion’ scenario for Jupiter’s atmosphere. In this model,the Sun and its disk form in a low-metallicity molecular cloud (left). The Sun’s orbit takesit through other regions (right) of higher metallicity, polluted by massive stellar winds andsupernovae. Ongoing Bondi-Hoyle accretion from the ISM delivers this enriched material tothe disk, where it is incorporated into Jupiter’s atmosphere. 23 –Fig. 2.— Elemental abundances of Jupiter from Galileo probe (grey), and fits for our ‘ModelA’ case (red). The y axis plots the elemental number abundances relative to the Solarabundances, normalized to hydrogen. The model consists of a linear combination of 87%solar composition, 9% from stellar winds from a 40 M ⊙ star, and 4% ejecta from an SNof original mass 20 M ⊙ . O is a lower limit because the Galileo probe entered Jupiter ata cloud-free location believed to be anomalously dry. Solar and Jupiter compositions arerevised 2007 values (GAS07, WLA08).Fig. 3.— Sources of individual species in our ‘Model A’ case. The model is a linear combi-nation of elemental abundances from the Sun (green curve), stellar winds (blue curve), andan SN (red curve). Each line is normalized to H at 1.0. The stellar winds produce much ofthe C and N, while the supernova supplies most of the remaining species. 24 –Fig. 4.— Elemental abundances of Jupiter from Galileo probe (grey), and fits for our ‘ModelA2’ case (red). The model consists of a linear combination of 78% solar composition, 8%from stellar winds from a 40 M ⊙ star, and 14% ejecta from an SN of original mass 20 M ⊙ .Abundances are the same as Model A, except the Lod08 argon value is used.Fig. 5.— Sources of individual species in our ‘Model A2’ case. Same as Figure 3, but usingLod08 argon value. 25 –Fig. 6.— Our ‘Model B’ fits. Same as Figure 2, but assuming 1989 solar composition data(AMN03, AG89). The coefficients are Solar (94%), stellar winds from 40 M ⊙ star (4%), anda supernova from a star of original mass 25 M ⊙ (1.5%). The fractional contamination inthis model is 6%, less than half that in Model A. The low predicted value for Xe might beexplained by its particularly high condensation temperature (see text). The broad similarityof the results to the ‘Model A’ case shows that our model is robust against small changes toknowledge of the composition of the Sun and Jupiter.Fig. 7.— Individual components of our ‘Model B’ fits. Same as Figure 3, but using 1989solar abundances (AMN03, AG89). 26 –Table 1: Details of model results, based on the listed ejecta yields from winds and SNcompared with Jupiter. Table lists species observed by Galileo; abundances are in numberdensity relative to H, normalized to Solar. The checked species are those which are stable atJupiter and our model attempts to fit; we predict the unchecked species but do not attemptto fit them. The three models are based on different measurements for the Jupiter:Solarabundances. 27 – REFERENCES
Alibert, Y., O. Mousis, and W. Benz, 2005, On the volatile enrichments and composition ofJupiter.
Astrophys. J. , L145–L148.Anders, E. and N. Grevesse, 1989, Abundances of the elements: meteoritic and solar.
Geochimica et Cosmochimica Acta , 197–214.Atreya, S. K., P. R. Mahaffy, H. B. Niemann, M. H. Wong, and T. C. Owen, 2003,Composition and origin of the atmosphere of Jupiter – an update, and implicationsfor the extrasolar giant planets. Plan. Spac. Sci. , 105–112.Bally, J., 2008, Overview of the Orion complex. In Handbook of Star Forming Regions (ed.B. Reipurth), ASP Conf. Series, 459–480, Astron. Soc. Pac., San Francisco.Bally, J., W. D. Langer, and W. Liu, 1991, Infrared dust and mm-wave CO emission in theOrion region.
Astrophys. J. , 645–663.Chiang, E. I. and P. Goldreich, 1997, Spectral energy distributions of T Tauri stars withpassive circumstellar disks.
Astrophys. J. , 368–376.Cunha, K. and D. L. Lambert, 1994, Chemical evolution of the Orion association. II. Thecarbon, nitrogen, oxygen, silicon, and iron abundances of main-sequence B stars.
Astrophys. J. , 170–191.Cunha, K., V. V. Smith, and D. L. Lambert, 1998, Chemical evolution of the Orionassociation. IV. The oxygen and iron abundances of F and G stars.
Astrophys. J. , 195–205.Cunha, K., V. V. Smith, E. Parizot, and D. L. Lambert, 2000, Light-element abundancepatterns in the Orion association. I. Hubble Space Telescope observations of boronin G dwarfs.
Astrophys. J. , 850–860. 28 –Dauphas, N., B. Marty, and L. Reisberg, 2002, Molybdenum evidence for inheritedplanetary scale isotope heterogeneity of the protosolar nebula.
Astrophys. J. ,640–644.Dullemond, C. P. and C. Dominik, 2005, Dust coagulation in protoplanetary disks: A rapiddepletion of small grains.
Astron. & Astrophys. , 971–986.Dutrey, A., S. Guilloteau, and M. Simon, 2003, The BP Tau disk: a missing link betweenClass II and III objects?
Astron. & Astrophys. , 1003–1011.Fesen, R. A., M. C. Hammell, J. Morse, R. A. Chevalier, K. J. Borkowski, M. A. Dopita,C. L. Gerardy, S. S. Lawrence, J. C. Raymond, and S. van den Bergh, 2006, Theexpansion asymmetry and age of the Cassiopeia A supernova remnant.
Astrophys. J. , 283–292.Fesen, R. A., P. A. Hoeflich, A. J. S. Hamilton, M. C. Hammell, C. L. Gerardy, A. M.Khokhlov, and J. C. Wheeler, 2007, The chemical distribtion in a subluminous typeIa supernova: Hubble Space Telescope images of the SN 1885 remnant.
Astrophys.J. , 396–409.Freyer, T., G. Hensler, and H. W. Yorke, 2006, Massive stars and the energy balance ofthe interstellar medium. II. The 35 M ⊙ star and a solution to the ‘missing windproblem’. Astrophys. J. , 262–280.Garcia-Segura, G., N. Langer, and M.-M. Mac Low, 1996, The hydrodynamic evolution ofcircumstellar gas around massive stars. II. The impact of the time sequence O star → RSG → WR star.
Astron. & Astrophys. , 133–146.Gomes, R. S., A. Morbidelli, and H. F. Levison, 2004, Planetary migration in a planetesimaldisk: why did Neptune stop at 30 AU?
Icarus , 492–507. 29 –Grevesse, N., M. Asplund, and A. J. Sauval, 2007, The solar chemical composition.
SpaceSci. Rev. , 105–114.Guillot, T. and R. Hueso, 2006, The composition of Jupiter: sign of a (relatively) lateformation in a chemically evolved protosolar disk.
Month. Not. Royal Astron. Soc. , L47–L51.Hahn, J. M. and R. Malhotra, 1999, Orbital evolution of planets embedded in a planetesimaldisk.
Astron. J. , 3041–3053.Hersant, F., D. Gautier, and J. I. Lunine, 2004, Enrichment in volatiles in the giant planetsof the Solar System.
Plan. Spac. Sci. , 623–641.Hersant, F., D. Gautier, G. Tobie, and J. I. Lunine, 2008, Interpretation of the carbonabundance in Saturn measured by Cassini. Plan. Spac. Sci. , 1103–1111.Hester, J. J., S. J. Desch, K. R. Healy, and L. A. Leshin, 2004, Perspectives: The cradle ofthe solar system. Science , 1116–1117.Hirschi, R., G. Meynet, and A. Maeder, 2005, Yields of rotating stars at solar metallicity.
Astron. & Astrophys. , 1013–1022.Iro, N., D. Gautier, F. Hersant, D. Bockelee-Morvan, and J. I. Lunine, 2003, Aninterpretation of the nitrogen deficiency in comets.
Icarus , 511–532.Johnstone, D. and J. Bally, 2006, Large-area mapping at 850 um. V. Analysis of the clumpdistribution in the Orion A South molecular cloud.
Astrophys. J. , 383–397.Johnstone, D., H. Matthews, and G. F. Mitchell, 2006, Large area mapping at 850 micron.IV. Analysis of the clump distribution in the Orion B South molecular cloud.
Astron.J. , 259–274. 30 –Jonkheid, B., C. P. Dullemond, M. R. Hogerheijde, and E. F. van Dishoeck, 2007, Chemistryand line emission from evolving Herbig Ae disks.
Astron. & Astrophys. , 203–216.Knapp, G. R. and M. Woodhams, 1993, Mass loss from cool supergiant stars.Kroeger, D., G. Hensler, and T. Freyer, 2006, Chemical self-enrichment of HII regions bythe Wolf-Rayet phase of an 85 M ⊙ star. Astron. & Astrophys. , L5–L8.Krot, A. N., K. Nagashima, M. Bizzarro, G. R. Huss, A. M. Davis, B. S. Meyer, andA. A. Ulyanov, 2008, Multiple generations of refractory inclusions in the metal-richcarbonaceous chondrives Acfer 182/214 and Isheyevo.
Astrophys. J. , 713–721.Levison, H. F. and M. J. Duncan, 1997, From the Kuiper Belt to Jupiter-Family Comets:The Spatial Distribution of Ecliptic Comets.
Icarus , 13–32.Lewis, J. S., 1972, Low temperature condensation from the Solar nebula.
Icarus , 241.Lodders, K., 2004, Jupiter formed with more tar than ice. Astrophys. J. , 587–597.Lodders, K., 2008, The solar argon abundance.
Astrophys. J. , 607–611.Lyons, J. R. and E. D. Young, 2005, CO self-shielding as the origin of oxygen isotopeanomalies in the early solar nebula.
Nature , 317–320.Mabry, J. C., A. P. Meshik, C. M. Hohenberg, Y. Marrocchi, O. V. Pravdivtseva,R. C. Wiens, C. Olinger, D. B. Reisenfeld, J. Allton, R. Bastien, K. McNamara,E. Stansbery, and D. S. Burnett, 2007, Refinement and implications of noble gasmeasurements from Genesis.
LPSC , 2412.Mesa-Delgado, A., C. Esteban, and J. Garcia-Rojas, 2008, Small-scale behavior of thephysical conditions and the abundance discrepancy in the Orion nebula. Astrophys.J. , 389–404. 31 –Moeckel, N. and H. Throop, 2009, Bondi-Hoyle-Lyttleton accretion onto a protoplanetarydisk.
Astrophys. J. , 268–277.Ouellette, N., S. J. Desch, and J. J. Hester, 2007, Interaction of supernova ejecta withnearby protoplanetary disks.
Astrophys. J. , 1268–1281.Owen, T. and T. Encrenaz, 2006, Compositional constraints on giant planet formation.
Plan. Spac. Sci. , 1188–1196.Owen, T., P. Mahaffy, H. B. Niemann, S. Atreya, T. Donahue, A. Bar-Nun, and I. de Pater,1999, A low-temperature origin for the planetesimals that formed Jupiter. Nature , 269–270.Ranen, M. C. and S. B. Jacobsen, 2006, Barium isotope heterogeneities in early solarsystem materials: applications to planetary reservoir models.
LPSC , 1832.Ranen, M. C. and S. B. Jacobsen, 2008, Three-component mixing in the protoplanetarydisk: the isotopic evidence. LPSC , 1954.Roulston, M. S. and D. J. Stevenson, 1995, Prediction of neon depletion in Jupiter’satmosphere. EOS , 343.Schaller, G., D. Schaerer, G. Meynet, and A. Maeder, 1992, New grids of stellar modelsfrom 0.8 to 120 Msol at Z=0.020 and Z=0.001. Astron. & Astrophys. Suppl. ,269–331.Stasinska, G., G. Tenorio-Tagle, M. Rodriguez, and W. J. Henney, 2007, Enrichment of theinterstellar medium by metal-rich droplets and the abundance bias in HII regions. Astron. & Astrophys. , 193–204.Tachibana, S. and G. R. Huss, 2003, The initial abundances of Fe in the solar system.
Astron. J. , L41–L44. 32 –Tenorio-Tagle, G., 1996, Interstellar matter hydrodynamics and the dispersal and mixing ofheavy elements.
Astron. J. , 1641–1650.Throop, H. B. and J. Bally, 2008, Tail-end Bondi-Hoyle accretion in young star clusters:Implications for disks, stars, and planets.
Astron. J. , 2380–2397.Throop, H. B., J. Bally, L. W. Esposito, and M. J. McCaughrean, 2001, Evidence for dustgrain growth in young circumstellar disks.
Science , 1686–1689.Trinquier, A., J.-L. Birck, and C. J. Allegre, 2007, Widespread Cr heterogeneity in theinner solar system.
Astrophys. J. , 1179–1185.Tsamis, Y. G. and D. Pequignot, 2005, A photoionization-modelling study of 30 Doradus:the case for small-scale chemical inhomogeneity.
Month. Not. Royal Astron. Soc. , 687–704.Udry, S., D. Fischer, and D. Queloz, 2007, A decade of radial-velocity discoveries in theexoplanet domain. In
Protostars and Planets V , 685–699, U. Ariz. Pr., Tucson.Villeneuve, J., M. Chaussidon, and G. Libourel, 2009, Homogeneous Distribution of 26Alin the Solar System from the Mg Isotopic Composition of Chondrules.
Science ,985–988.Wang, L. and J. Hu, 1994, Blue-shifted oxygen lines and the clumpy ejecta of 1993J.
Nature , 380–382.Watson, A., K. R. Stapelfeldt, K. Wood, and F. Menard, 2007, Multi-wavelength imagingof young stellar object disks: toward an understanding of disk structure and dustgrain evolution. In
Protostars and Planets V , U. Ariz. Pr., Tucson.Weidenschilling, S. J., 1997, The origin of comets in the solar nebula: a unified model.
Icarus , 290–306. 33 –Wiens, R. C., D. S. Burnett, C. M. Hohenberg, A. Meshik, V. Heber, A. Grimberg,R. Wieler, and D. B. Reisenfeld, 2007, Solar and solar-wind composition results fromthe genesis mission.
Space Sci. Rev. , 161–171.Williams, B. J., K. J. Borkowski, S. P. Reynolds, J. C. Raymond, J. A. Morse, W. P.Blair, P. Chavamian, R. Sankrit, S. P. Hendrick, R. Chris Smith, S. Points, andP. F. Winkler, 2008, Ejecta, dust, and synchrotron radiation in B0540-69.3: A moreCrab-like remnant than the Crab.
Astrophys. J. , 1054–1069.Wong, M. H., J. I. Lunine, S. K. Atreya, T. Johnson, P. R. Mahaffy, T. C. Owen, andT. Encrenaz, 2008, Oxygen and other volatiles in the giant planets and theirsatellites. 219–246.Woosley, S. E. and A. Heger, 2007, Nucleosynthesis and remnants in massive stars of solarmetallicity.
Phys. Rep. , 269–283.Young, E. D., M. Gounelle, R. Smith, M. R. Morris, and K. M. Pontoppidan, 2008, Solarsystem oxygen isotope ratios result from pollution by Type II supernovae.