2 Noble Gases in the Solar System

Transcription

1 2 Noble Gases in the Solar System Rainer Wieler Isotope Geology, ETH Zürich Sonneggstrasse 5 CH-8092 Zürich, Switzerland wieler@2erdw.ethz.ch INTRODUCTION This chapter provides an overview of available noble gas data for solar system bodies apart from the Earth, Mars, and asteroids. Besides the Sun, the Moon, and the giant planets, we will also discuss data for the tenuous atmospheres of Mercury and the Moon, comets, interplanetary dust particles and elementary particles in the interplanetary medium and beyond. In addition, we summarize the scarce data base for the Venusian atmosphere. The extensive meteorite data from Mars and asteroidal sources are discussed in chapters in this volume by Ott (2002), Swindle (2002a,b) and Wieler (2002). Data from the Venusian and Martian atmospheres are discussed in more detail in chapters by Pepin and Porcelli (2002) and Swindle (2002b). Where appropriate, we will also present some data for other highly volatile elements such as H or N. The solar system formed from a molecular cloud fragment traditionally called the solar nebula that was rather well mixed. Therefore, isotopic abundances in almost all available solar system materials are very similar to each other, and elemental abundances in primitive meteorites are also similar to the values in the Sun. The major exceptions to this rule are the noble gases. Because they are chemically inert and volatile, they are very strongly depleted in solid matter. As a consequence, numerous noble gas components can be recognized throughout the solar system which are not necessarily related to the composition of the bulk nebula. Still, one major question in cosmochemistry is to what extent planetary bodies contain reservoirs that reflect the noble gas composition in the nebula or the presolar cloud. To discuss this, we first need a proxy for the nebula composition. The obvious choice is the bulk Sun, except for He, which has been produced in the Sun throughout its history. We will discuss the large database that allows us to infer noble gas abundances and isotopic composition of the present Sun as well as the protosun. This will include data from samples from the lunar dust surface (regolith) and meteorites from asteroidal surfaces which represent an archive of the ancient solar wind, the particle radiation emitted by the Sun. Giant planets are the next choice for deducing the nebula composition, because they formed either from collapsing subnebulae or by gravitational attraction of gas from the nebula onto their early solid cores. Indeed, the 3 He/ 4 He ratio determined by the probe descending during the Galileo mission into Jupiter s atmosphere is presently considered to be the best value for the composition of protosolar He, before deuterium was converted to 3 He in the early Sun. ANALYTICAL TECHNIQUES Table 1 summarizes the techniques used to measure noble gases. By far the most important is mass spectrometry. Mass spectrometers in space are used, e.g., for solar wind and solar energetic particle measurements or atmospheric analyses on Venus, Moon, Mars and Jupiter, while mass spectrometers in the laboratory allow us to analyze extraterrestrial samples available on Earth, i.e., lunar samples, meteorites, interplanetary dust or solar corpuscular radiation trapped by foils exposed in space. Of course, /02/ $10.00 DOI: /rmg

2 22 Wieler Techniques used for specified object Sun UV spectroscopy (flares, corona, emerging active regions) Space-borne mass spectrometry (solar wind, SW; solar energetic particles, SEP) Mass spectrometry in laboratory (solar wind collector foils etc.) Mass spectrometry, laboratory (SW/SEP trapped in lunar regolith and meteorites) Helioseismology & solar models Table 1. Analytical techniques. Remarks and selected references No useful noble gas lines in photospheric spectra. Feldman and Widing 1990; Feldman 1998; Widing Kallenbach et al. 1997; Leske et al. 1999; Gloeckler and Geiss 1998b; review by Wimmer-Schweingruber Apollo SW Composition experiment, MIR space station; upcoming Genesis mission. Geiss et al Lunar regolith archive of ancient solar wind. Pepin et al. 1970a,b; Wieler et al. 1986; Marti Solar He abundance. Christensen-Dalsgaard 1998; Gough Mercury UV-Spectroscopy (Mariner 10) He, Ar in atmosphere (exosphere). Stern 1999b. Venus Mass spectrometry, gas chromatography (Pioneer Venus, Venera 11-14) Moon Ion- and neutral mass spectrometry UV spectroscopy Atmospheric analyses. Hoffman et al. 1980a,b; Donahue and Pollack 1983; Moroz He and Ar in lunar exosphere on ground, 36 Ar/ 40 Ar. Stern 1999b. Upper limit for Kr, Xe from orbit. Feldman and Morrison Ar-Ar ages, exposure ages, solar wind etc. Turner 1970; Culler et al. 2000; Wieler Mass spectrometry, laboratory (Apollo & Luna samples, lunar meteorites) Mars Mass spectrometry (Viking) Atmosphere. Nier and McElroy 1977; Owen et al EUV spectroscopy He in atmosphere. Krasnopolsky et al Mass spectrometry, laboratory (Martian meteorites) Asteroids Mass spectrometry, laboratory (meteorites) Jupiter and other giant planets Mass spectrometry (Galileo probe) Martian atmosphere, crust, mantle. Bogard et al. 1984; Becker and Pepin 1984; Swindle 2002b. Ott 2002; Swindle 2002a. He-Ar, Xe elemental and isotopic abundance in outer Jovian atmosphere. Mahaffy et al In-situ interferometry (Galileo probe) He abundance in outer Jovian atmosphere. v. Zahn et al Radio occultation & infrared He abundance in giant planet outer atmospheres. Conrath et spectroscopy (Voyager) al. 1991; Conrath and Gautier Comets EUV spectroscopy Ar abundance, upper limits for He, Ne. Stern 1999a. in-situ dust collection (see also IDPs) Stardust mission CI chondritic meteorites? Lodders and Osborne 1999; Ehrenfreund et al Interplanetary dust particles, micrometeorites Mass spectrometry, laboratory Dust from asteroids & comets, collected in near Earth space, ice, & sediments. Nier and Schlutter 1992; Olinger et al Heliospheric ions Mass spectrometry, space Sources: Sun, Galaxy, local interstellar cloud. Mewaldt et al. 2001a; Binns et al. 2001; Gloeckler and Geiss 1998b.

3 Noble Gases in the Solar System 23 laboratory analyses will usually yield data of higher precision than in situ measurements. Conversely, the latter offer the advantage of high temporal and spatial resolution, allowing one to study, e.g., solar wind emitted from different regions or during different regimes, correlations between noble gases and other elements, or short-term variations of fluxes or composition. Noble gases are intrinsically difficult to detect by spectroscopy. For example, solar photospheric spectra, which form the basis for solar abundance values of most elements, do not contain lines from noble gases (except for He, but this line cannot be used for abundance determinations). Yet, ultraviolet spectroscopy is the only or the major source of information on noble gas abundances in the atmospheres of Mercury and comets. In the Extreme Ultraviolet (EUV), photon energies exceed bond energies of molecules and the first ionization potential of all elements except F, He, and Ne, so that only these elements are visible in this part of the spectrum (Krasnopolsky et al. 1997). Other techniques can be used to determine the abundance of He where this element is a major constituent. Studies of solar oscillations (helioseismology) allow a precise determination of the He abundance in the solar interior, and the interferometer on the Galileo probe yielded a precise value for the refractive index and hence the He abundance in the upper atmosphere of Jupiter (see respective sections of this chapter). THE SUN Solar noble gas abundances Determining elemental and isotopic compositions of the elements in the Sun is a fundamental task in cosmochemistry and abundance tables are updated at regular intervals (Anders and Grevesse 1989; Palme and Beer 1993; Grevesse and Sauval 1998). For most elements, the compilations are based on photospheric spectroscopy on the one hand and meteorite data on the other. This reflects the very remarkable fact that in a very rare class of meteorites, the CI chondrites, relative elemental abundances are identical within uncertainties to photospheric values for essentially all but the most volatile elements, e.g., H, N, O, and the noble gases (a further exception is Li, which is destroyed in the sun, Grevesse and Sauval 1998). Anders and Grevesse (1989) define solar abundances as the best estimates obtained from photospheric analyses, if necessary augmented by data from solar wind, solar energetic particles, or astronomical observations, but not from meteorites. Solar system abundances are defined as best estimates for the entire solar system, and are mostly based on the CI chondrite data. Because photospheric and CI chondrite data agree so well with each other, the distinction between solar and solar system is often not made in practice, and actually, cosmochemists usually use the solar abundances obtained from CI chondrites, due to their higher precision. Since the elemental abundances in the sun are also remarkably similar to those in many other stars, the term cosmic abundances is also sometimes used for the former, although this appears to overestimate the role of our place in the universe. For noble gases, the situation is different. Meteoritic noble gas abundances are by many orders of magnitude lower than any likely values in the primordial solar nebula. In addition, because no noble gas lines that could be used for abundance determinations are visible in photospheric spectra, among the noble gases only the photospheric abundances of Ne and Ar have been measured spectroscopically in active regions and the corona. This and other data sources that need to be considered for noble gases are explained in the following two subsections. Furthermore, all deuterium in the nascent Sun has been converted to 3 He, raising the 3 He/ 4 He ratio of the present-day sun by a factor of ~2.5-3 above the value of the proto-sun. Both isotopes of helium are also produced in the

4 24 Wieler present Sun, as intermediate ( 3 He) and end product ( 4 He), respectively, of hydrogen fusion. For these reasons, solar abundance compilations adopt best estimates of the protosolar 3 He/ 4 He ratio, as well as the protosolar 4 He abundance. Elemental abundance values of noble gases in the sun recommended here are listed in Table 2, and isotopic compositions are given in Tables 3-5. In the next two subsections, we comment on these adopted values. Note that in Tables 7, 11, and 13 (found later in this chapter) elemental abundances in various reservoirs are stated relative to solar abundances. These are always the values as reported in the original references, i.e., they have not been renormalized to the solar abundances used here. Elemental abundances. (A) Helium. Solar models are able to deduce accurate, although somewhat model-dependent, protosolar He values by fits to the observed present-day luminosity (e.g., Christensen-Dalsgaard 1998, see Helium in the sun section). The value stated in the body of Table 2 in three different notations is from the compilation by Grevesse and Sauval (1998), two other values for the initial He mass fraction are given in the Table caption (see also next section). In summary, the models provide the protosolar He abundance with an uncertainty of only a few percent. (B) Neon. We adopt the Ne abundance recommended by Holweger (2001), which is based on EUV spectroscopy of emerging active regions (Widing 1997). Holweger (2001) points out that emerging flux events most likely permit direct observation of unfractionated photospheric material, and prefers this data source over the solar particle measurements or extra-solar data used by Anders and Grevesse (1989). Holweger (2001) modified the original Widing (1997) Ne abundance value of [8.08] in the astrophysical notation (see Table 2) by also using ogygen (besides Mg) as a reference element combined with updated photospheric abundances of the reference elements. The Holweger (2001) Ne abundance is ~20% lower than the Widing (1997) value (adopted by Grevesse and Sauval 1998) and 23% lower than the Anders and Grevesse value, but all these estimates agree within stated uncertainties with each other. (C) Ar. Grevesse and Sauval (1998) discuss two recent solar Ar abundance values. One is from coronal spectroscopy, yielding a photospheric Ar/Ca abundance ratio of 1.31±0.30, corresponding to an Ar abundance of [6.47±0.10] (Young et al. 1997). The second value is from solar energetic particles, yielding an Ar abundance of [6.39±0.027] (Reames 1998, value slightly updated and with a lower uncertainty than that stated in Grevesse and Sauval 1998). Grevesse and Sauval (1998) prefer the SEP value due to its lower uncertainty. In Table 2 we adopted the mean of both values, however, with a more conservative uncertainty of [0.1] dec (~25%) that would allow for some fractionation between coronal and photospheric Ar abundance. The Ar value in Table 2 is about 35% lower than that recommended by Anders and Grevesse (1989). (D) Kr and Xe. The only available direct data of Kr and Xe in the Sun are from implanted solar wind (SW) atoms in lunar samples and gas-rich meteorites. These data are not very useful for abundance determinations, because both elements are enriched in the solar corpuscular radiation, as is discussed in the Moon section. The most reliable way to estimate the Kr and Xe abundances in the Sun is by interpolating concentrations of suitable isotopes in CI chondrites. Best suited are isotopes of elements close to Kr and Xe in the periodic system which are largely produced by s-process nucleosynthesis. Table 2 lists the values given by Palme and Beer (1993). Uncertainties of the s-process calculations are estimated to be 5-10%, to which another similar uncertainty has to be added for the abundances of the elements used for interpolation. The Kr and Xe estimates in Table 2 differ by ~15-20% from those adopted by Anders and Grevesse (1989), which

5 Noble Gases in the Solar System 25 Table 2. Elemental abundances of noble gases in the Sun and the solar corpuscular radiation He * ( 4 He) Ne ( 20 Ne) Ar ( 36 Ar) Kr ( 84 Kr) Xe ( 132 Xe) 4 He/ 20 Ne 20 Ne/ 36 Ar 36 Ar/ 84 Kr 84 Kr/ 132 Xe Sun refs. 1-3 (Si=10 6 ) refs ±0.02 (log- scale) a (10.99) ref. 1 (NHe/NH) ref. 1 (Y0) b 0.275±0.010 (2.73±0.13) (2.80±0.48) 10 (7.51±1.94) 10 (2.73±.13) 10 9 (2.61±.45) ±0.069 (7.97±0.07) (6.35±1.64) ±8 (32±5) 6.43±0.10 (6.36±0.10) ±0.06 (3.05±0.06) 4.1±0.9 (1.09±0.25) 2.16±0.09 (1.58±0.09) 1050±190 41± ±600 29±8 Solar wind 4 650±50 47± Solar energetic particles 5 gradual events Si=10 6 ; [O=10 3 ] impuls. events [O=10 3 ] ( ) ( ) [56,000] [ 20 Ne=142] [46,000] [ 20 Ne=388] ( ) [ 36 Ar=2.8] 400±25 51±4 1: (He & Ar): Grevesse and Sauval 1998, for Ar see text; other recent protosolar He values from solar modeling: Y 0 = (Christensen-Dalsgaard 1998), and Y 0 = (Bahcall et al., 2001). 2 (Ne): Holweger : (Kr & Xe): Palme and Beer All Si-normalized values assume H/Si=27900 (Anders and Grevesse 1989). 4: Murer et al. 1997; Wieler and Baur 1995; Kr and Xe abundances show secular changes. 5: Miller 1998; Reames 1998; Abundances in [ ] are reported relative to O as in orig. ref., because Si is enhanced in SEP due to FIP-effect. Miller 1998: gradual events 4 He/H ~0.1. Reames et al. 1994: impulsive events 4 He/H = *: protosolar He abundance (before onset of hydrogen burning) a: The astronomic abundance scale is logarithmic, with the hydrogen abundance log N H =12 b: The protosolar He (He 0 ) mass fraction Y 0 is defined as Y 0 = m(he 0 )/(m(h 0 )+m(he 0 )+m(heavy elements))

6 26 Wieler are based on earlier s-process calculations and CI abundances. In summary, whereas the He abundance in the (proto)sun is known today with high precision, mainly thanks to solar modeling, the values for Ne-Xe are probably still uncertain by some 20%. Isotopic abundances. Our knowledge of the isotopic composition of noble gases in the sun so far is exclusively based on analyses of the solar wind and solar energetic particles. These are discussed in detail below in this section and in the Moon section. We will also discuss estimates of isotopic fractionations between the sun and the solar corpuscular radiation and conclude that such fractionations are small but probably not entirely negligible. The He isotopic composition in the outer convective zone of the sun given in Table 3 has been corrected for various estimated fractionation effects. Also the solar Ne and Ar isotopic compositions in Table 4 (below) have been very slightly corrected, whereas the Kr and Xe compositions in Table 5 (below) are measured solar wind values. Alternative reservoirs that might conceivably contain isotopically unfractionated solar noble gases, notably the atmosphere of Jupiter (or the other Giant planets), have not been analyzed so far with the necessary precision to yield values of equal quality as the solar wind data (Giant Planets section). However, the 3 He/ 4 He in Jupiter s atmosphere is the currently most widely used estimate for the protosolar He isotopic composition (prior to deuterium-burning) and is thus given in Table 3. Helium in the Sun During its main-sequence lifetime over the past ~4.6 Gyr, the Sun has been generating its energy by the conversion of hydrogen into 4 He in its core where temperatures are high enough that atomic nuclei can overcome the repulsive force due to their electric charge. Hydrogen fusion as well as further stellar nucleosynthesis reactions are discussed in, e.g., Clayton (1968) and Rolfs and Rodney (1988). Briefly, the net reaction 4 H 4 He + 2 e ν produces an energy of MeV, part of which will be carried away by the neutrinos. As the probability of a direct fusion of four protons is essentially zero, the conversion of hydrogen into helium occurs in various chains of two-particle interactions. The most important ones are the three p-p chains (cf. Rolfs and Rodney 1988, Fig. 6.13). In all these reaction chains also 3 He is produced as an intermediate product and in two of them also Be, Li, and B nuclei are produced and destroyed again. In stars like our Sun, which also contain sizeable amounts of elements heavier than H and He, hydrogen fusion also occurs by cyclic reactions involving heavier nuclei, the most important one being the CNO bi-cycle (Rolfs and Rodney 1988, Fig. 6.22). The C, N, and O nuclei involved act as catalysts only, and the net reaction is again the conversion of four protons into a 4 He nucleus. The details of these nucleosynthesis processes are major ingredients for solar models. Although they are conceptually simple (Gough 1998), the so-called standard solar models (e.g., Bahcall and Pinsonneault 1995; Bahcall et al. 2001) are thought to describe the Sun surprisingly well. As noted above, solar modeling provides a means for accurately inferring the He abundance of the Sun. The initial abundance of He, given as the mass fraction Y 0, is a free parameter in solar modeling. Because the luminosity of the Sun depends sensitively on the helium abundance, the value Y 0 can be chosen such that the model yields the correct present-day luminosity. The total initial mass fraction of the elements heavier than He (Z 0 ) is then obtained by adjusting the model to the correct present-day abundance ratio of heavy elements and hydrogen in the photosphere (Z s /X s = ). Christensen-Dalsgaard (1998) obtains Y 0 = and Z 0 = , Bahcall et al. (2001) report for their standard solar model Y 0 = Figure 1 shows the hydrogen abundance in the present-day Sun as a function of distance from the solar center. The

7 Noble Gases in the Solar System 27 Figure 1. Hydrogen model abundance X (by mass) in the Sun as a function of distance from the center (solar surface at r/r = 1). The He abundance profile is essentially the complement of the H profile minus ~2% of heavy elements. Helioseismology provides evidence for gravitational settling of He, which is incorporated in the model represented by the solid line, where the He abundance increases abruptly below the outer convective zone at r/r = 0.7. The dotted line is for a model neglecting He settling. [Used by kind permission of Kluwer Academic Publishers, from Christensen- Dalsgaard (1998) Space Science Reviews, Vol. 85, Fig. 1, p 22.] dotted line is for a model that neglects gravitational settling of helium and heavier elements towards the Sun s center. The helium profile essentially follows this dotted line, although lower by 2% to account for the heavy elements. Near the center, hydrogen is depleted due to nuclear burning. Outwards of about 0.25 solar radii, however, the hydrogen abundance stays almost constant and so therefore does the He abundance. In this model, the He/H ratio at the surface is very close to the initial ratio 4.6 Ga ago. A model that takes into account element settling below the Outer Convective Zone (e.g., Vauclair 1998) is shown by the solid line in Figure 1. In the Outer Convective Zone (OCZ) in the outermost ~30% of the solar radius, mixing leads to a constant H/He ratio, which is above the initial ratio, however. He is therefore depleted near the solar surface by ~10% relative to the value expected without gravitational settling. Helioseismology, the science of probing the Sun s interior by observing solar oscillations, allows a deduction of the He abundance in the outer part of the Sun (e.g., Gough et al. 1996) and provides clear experimental support for a He depletion near the surface. The He abundance in the convective zone determined by helioseismology is close to 0.25 (Christensen-Dalsgaard 1998), clearly lower than the initial value of ~0.275 predicted by standard solar models without He-settling, but in good agreement with models accounting for element settling. It may be worth noting here that the solar neutrino problem, i.e., the unexpectedly low abundance of neutrinos from the solar interior detected on Earth, is no longer suspected to indicate severe flaws in standard solar models (e.g., Bahcall 2000). Noble gases in the solar corpuscular radiation The Sun emits the solar wind (SW), a continuous stream of particles with velocities mostly between ~ km/s (Parker 1997). Ions of higher energy are emitted especially during so-called solar energetic particle (SEP) events (Reames 1998). The

8 28 Wieler solar corpuscular radiation provides a unique opportunity to directly analyze solar matter, either by mass spectrometers in space or by laboratory analyses of irradiated natural or artificial targets. Recent space missions include WIND, SAMPEX, Ulysses, the Solar and Heliosperic Observatory (SOHO), and the Advanced Composition Explorer (ACE) (e.g., Wimmer-Schweingruber 2001, and references therein). Lunar dust, aluminum foils exposed on the Moon during the Apollo missions, and targets on the Genesis mission to be returned to Earth are examples of irradiated samples. In lunar samples and the Apollo foils almost exclusively noble gases have been studied, since most other elements are not scarce enough in the targets themselves to allow detection of a solar contribution. Hence, noble gases in dust samples that were irradiated on the lunar surface up to several billion years ago are very important for the investigation of solar history. Some major results of the lunar sample studies are presented below in this section; more details are provided in the Moon section. Studies of elemental and isotopic composition of the solar corpuscular radiation have to account for the fact that particle selection and acceleration may lead to fractionations. In the solar wind, such effects are minor for isotopic ratios but clearly significant for some elemental ratios, whereas in SEPs severe isotopic effects also occur. Therefore, abundance studies in the solar corpuscular radiation sometimes yield information about fractionation processes rather than directly about solar composition. Helium in the solar wind and the Outer Convective Zone. The 3 He abundance in the present and past solar wind is of great interest. First, it helps us to establish an estimate for the protosolar deuterium abundance (Geiss and Reeves 1972; Geiss 1993; Gloeckler and Geiss 2000), because all D has been burnt to 3 He very early in solar history. Second, 3 He builds up in the course of solar history at intermediate depths in the sun by incomplete hydrogen burning. The 3 He/ 4 He ratio in the solar wind and especially its long-term evolution are therefore very sensitive indicators of mixing of material into the outer convective zone (Bochsler et al. 1990). Table 3 shows 3 He/ 4 He values in the solar wind obtained by different techniques. Although all values agree within ±7% with each other, reported uncertainties are often quite large. The recent SWICS Ulysses value is at the low end of those stated in Table 3. It represents a SW regime associated with coronal holes ( high-speed SW of up to 800 km/s), which is thought to be less fractionated relative to the OCZ (see below) than the interstream-sw with speeds on the order of ~ km/s. Figure 2 shows that the 3 He/ 4 He ratio in this slow-sw is probably some 5-10% higher than in the coronal-hole dominated SW (Bodmer and Bochsler 1998b; Gloeckler and Geiss 2000). The weak dependence of 3 He/ 4 He on the SW regime indicates, however, that overall the solar wind isotopic composition reflects the He composition at the solar surface quite accurately (Bodmer and Bochsler 1998b). These authors estimate the ( 3 He/ 4 He) ratio in coronal-hole SW to be only ~9% higher than the value in the OCZ, which they calculate as (3.75±0.70) Gloeckler and Geiss (2000) arrive at the same value but with a lower uncertainty (Table 3). The OCZ-estimate is also important in assessing the protosolar D/H ratio in the next section. The data from lunar samples will be discussed in more detail in the Moon section. In summary, the He isotopic composition in the solar wind and hence in the OCZ is reasonably well known, but further improvements are nevertheless desirable, e.g., to reduce uncertainties in the value of the protosolar D/H ratio, as is discussed next. The protosolar D/H ratio. The protosolar deuterium abundance is an upper limit to the primordial abundance of this isotope and provides a crucial observational test for the Big Bang cosmological model (Schramm 1993). For example, the primordial deuterium

9 Noble Gases in the Solar System 29 Table 3. 3 He/ 4 He in the sun (in units of 10-4 ). Protosun (Jupiter) ±0.05 Solar wind and outer convective zone SW OCZ Ulysses-SWICS 2 (coronal-hole SW) 4.08± ±0.70 Ulysses-SWICS ±0.27 ISEE Apollo foils ±0.21 lunar soil 6 last ~100 Myr last ~1000 Myr Solar energetic particles 4.57± gradual events (CME) 7 19±2 impulsive events (flares) 8 (~1-330) 10 3 SEP-He in lunar soil ±0.05 Protosolar deuterium 9 (D+ 3 He)/ 4 He (3.60±0.38) 10-4 D/H (1.94±0.39) : Mahaffy et al. 1998; Jupiter value assumed to be identical to protosolar value (Giant Planets section). 2: Bodmer and Bochsler 1998b. 3: Gloeckler and Geiss : Bochsler : Geiss et al : Benkert et al : Mason et al. 1999, average value of 12 events. 8: Reames et al. 1994; Mason et al : Gloeckler and Geiss 2000, deduced from ( 3 He/ 4 He) OCZ and 3 He/ 4 He in Jupiter (ref. 1, Table 7). Stated uncertainties of SW values are not directly comparable with each other. Uncertainty in (2) mostly due to an assumed systematic uncertainty of detector efficiencies; uncertainties in (6) do not consider possible alterations in samples. Figure 2. 4 He/ 3 He (log-scale) as a function of solar wind speed as determined with the Solar Wind Ion Composition Spectrometer on the Ulysses spacecraft. Contour lines indicate 2, 4, 8, 16, 32, and 64 cases per bin. The scatter is mostly (but probably not entirely) due to statistical fluctuations. A weak increase of 4 He/ 3 He with SW speed is indicated (400 km/s: 4 He/ 3 He = 2350, i.e., 3 He/ 4 He = ; 800 km/s: 4 He/ 3 He = 2570, i.e., 3 He/ 4 He = ). [Used by permission of Springer Verlag, from Bodmer and Bochsler (1998b), Astron. Astrophys., Vol. 337, Fig. 6, p 926.]

10 30 Wieler abundance constrains the baryonic density in the universe. Protosolar D is also important for tracing galactic evolution in the past 4.6 Gyr. To derive the (D + 3 He)/ 4 He ratio in the protosun from the present-day ( 3 He/ 4 He) SW ratio, one needs to consider several effects which may alter the He isotopic composition in the solar wind. First, there is a probable mass fractionation between the SW and the OCZ as discussed in the previous paragraph. Furthermore, the He isotopic composition in the OCZ may itself be slightly different from the protosolar (D + 3 He)/ 4 He due to gravitational settling of He into deeper layers (see the subsection Helium in the Sun and Fig. 1) and a possible minor contribution of 3 He produced during the main sequence life of the Sun and mixed into the convective zone (Geiss 1993; Geiss and Gloeckler 1998). The latter effect is also discussed in the next subsection. Gloeckler and Geiss (2000) estimate that gravitational settling and mixing together lead to a modest enrichment of 3 He in the OCZ by 4±2% and calculate a protosolar (D + 3 He)/ 4 He ratio of (3.60±0.38) 10-4 (Table 3). Assuming for the protosolar 3 He/ 4 He ratio the value of measured in Jupiter (see the Giant Planets section) and protosolar He/H = 0.10, Gloeckler and Geiss (2000) obtain a protosolar (D/H) ratio of (1.94±0.39) 10-5, close to the original estimate of (2.5±0.5) 10-5 by Geiss and Reeves (1972). Alternatively, assuming He in the meteoritic Phase-Q ( 3 He/ 4 He = ) to represent the protosolar composition, Busemann et al. (2001) derive protosolar D/H = (2.4±0.7) It is not the place here to discuss all the major implications of this result, but a historical problem is worth mentioning, which was a major motivation for the Apollo solar wind composition experiment (cf. Geiss 1993). In the 1960s, the D/H ratio in terrestrial oceans ( ) was assumed to represent the cosmic or protosolar deuterium abundance. Based on this assumption the 3 He/ 4 He ratio in the solar wind was expected to be above 10-3, but meteorites thought to contain implanted solar wind noble gases yielded lower values of around These were confirmed by all subsequent studies discussed above, and Geiss and Reeves (1972) and Black (1972) were therefore able to conclude that deuterium is enriched in sea water by almost an order of magnitude relative to the protosolar composition due to low temperature reactions in interstellar clouds (Geiss and Reeves 1981). Further implications of the protosolar 3 He/ 4 He and D/H ratios are discussed by Gloeckler and Geiss (2000). The local interstellar cloud (LIC) represents a galactic sample having experienced chemical evolution for 4.6 Gyr longer than the protosun. D/H in the LIC is lower than the protosolar value whereas 3 He/ 4 He in the LIC is higher than the protosolar value (see also the Elementary particles in interplanetary space section). The direction of both these changes is expected, because stars only destroy deuterium but destroy and produce 3 He. The observed increase of the 3 He abundance is mainly due to production of this isotope in small stars. A secular variation of the 3 He/ 4 He ratio in the solar wind? As already mentioned, lunar dust samples provide a valuable record of the long term evolution of the noble gas composition in the solar wind. One of the most important parameters to be studied is the 3 He/ 4 He ratio, which establishes very stringent constraints on the amounts of material mixed into the outer convective zone from below (Bochsler et al. 1990). This is because a 3 He-rich region gradually evolves at intermediate depths in the solar interior, where temperatures are high enough to produce 3 He in the p-p chains, but too low to further process this isotope. The maximum enrichment after 4.6 Gyr is about a factor of 30 at about 0.3 solar radii (Bochsler et al. 1990). Some non-standard solar models predict that a considerable fraction of this 3 He is mixed into the convective zone and should thus be visible in the solar wind. If so, the 3 He/ 4 He ratio in samples that trapped SW several Ga ago should be significantly lower than in more recently irradiated samples.

11 Noble Gases in the Solar System 31 Data relevant to this problem are shown in Figure 9. This Figure will be discussed in more detail in the Moon section, so here we just note the most salient points. Most importantly, the isotopic composition of SW-He has definitely not changed by more than some 10% per Gyr, as shown by the bulk sample analyses compiled by Geiss (1973) shown as solid circles. This finding rules out non-standard solar models invoking strong internal mixing or high mass-loss after onset of hydrogen burning (Bochsler et al. 1990). These data also were one of the crucial observations allowing Geiss and Reeves (1972) to conclude that the He composition in the outer convective zone and hence the solar wind is close to the protosolar (D + 3 He)/ 4 He ratio (see above). Whether the lunar data actually indicate a small secular increase of the 3 He abundance in the convective zone is still unclear, as is discussed in the Moon section. New analyses, where the solar noble gases from the outermost dust grain layers are selectively sampled (solid squares in Fig. 9), apparently confirm a small secular change, but it seems probable that this is an experimental artifact. If so, corrected data (open squares) indicate an almost constant He isotopic composition in very ancient as well as recently irradiated samples. He in solar energetic particles. SEPs provide another source of information on the composition of the solar corona. SEP events are basically classified into large gradual events and smaller impulsive events (e.g., Reames 1998, 2001). Rise and fall times of gradual events are on the order of days, while those of impulsive events are considerably shorter. SEPs were originally all thought to be accelerated in solar flares (intense abrupt brightenings on the solar surface with radiation emitted over the entire electromagnetic spectrum due to a sudden release of magnetic energy), but it has become clear now that in gradual events solar wind and coronal particles are accelerated at shock waves produced by coronal mass ejections, and that flares are another manifestation of coronal mass ejections (e.g., Gosling 1993). Typical He abundances and isotopic compositions are given in Tables 2 and 3. Large events show a modest, variable enrichment of 3 He, on average by about a factor of 5 over the solar wind value (Mason et al. 1999). In contrast, particle acceleration in impulsive solar flares yields a highly fractionated sample of the corona, e.g., ~250- to 80,000-fold enhancements in the 3 He/ 4 He ratio and enrichments of heavy elements (e.g., Fe/O) of an order of magnitude. These enhancements are believed to be due to resonant wave particle interactions in a solar flare (Reames 1998; Miller 1998). However, on average, impulsive events only contribute a minor fraction to the SEP population. Also, Mason et al. (1999) suggest that the 3 He-enrichment in gradual flares may be due to remnant flare material in the interplanetary medium. Note that in the next subsection and in the Moon section we will discuss a noble gas component in lunar samples also dubbed SEP, as it is thought to be of solar origin and implanted at higher energies than the solar wind. This component seems, however, not to represent the same energy range as SEPs detected in space. Ne-Xe in the solar wind and solar energetic particles. Being rarer than He, the heavier noble gases in the SW are more difficult to analyze directly by mass spectrometers in space. Only recently the first Ne and Ar isotopic data were obtained by the SOHO and WIND missions (Kallenbach et al. 1997; Wimmer-Schweingruber et al. 1998; Weygand et al. 2001). Hence, much of the current knowledge on the heavier noble gases in the SW is based on the analysis of lunar samples and gas-rich meteorites, as is discussed in some detail in the Moon section. Here we compare some of these data with results on the present-day solar wind based on in situ measurements and the Apollo foils. Table 4 shows 20 Ne/ 22 Ne and 36 Ar/ 38 Ar ratios in the solar wind determined by the three techniques. As noted for He already, the stated error bars should be regarded with caution. In particular, the low nominal error bars of the lunar and meteoritic data ignore

12 32 Wieler Table 4. Isotopic abundances of Ne and Ar in the sun and the solar corpuscular radiation. 20 Ne/ 22 Ne 21 Ne/ 22 Ne 36 Ar/ 38 Ar Solar a 13.6± ± ±0.2 Solar wind Apollo foils ± ± ±0.3 SOHO ± ± ±0.65 WIND ±0.7 Moon/meteorites (SW) ± ± ± ±0.07 SEP-component ± ± ±0.05 Solar energetic particles 5 ~5-20 a: Solar wind composition corrected for an assumed fractionation relative to bulk sun (see text). Solar 40 Ar/ 38 Ar = (Anders and Grevesse 1989). 1: Geiss et al. 1972; Cerutti : Kallenbach et al. 1997; Weygand et al. 2001, mean of reported interstream and coronal hole values. 3: Wimmer-Schweingruber et al. 1998; 4: Benkert et al. 1993; Becker et al. 1998; Pepin et al : Leske et al. 1999, possible systematic uncertainties (see the Moon section). It is nevertheless gratifying that all three techniques yield very similar values, which suggests that the solar wind Ne and Ar isotopic composition is known to within a few percent. The lunar data also allow for a precise determination of the abundance of the rare isotope 21 Ne, which is not as well constrained by the other two methods. Anders and Grevesse (1989) adopted for solar Ar the 36 Ar/ 38 Ar value 5.32 of the terrestrial atmosphere. The value measured in the Apollo foils and also the new SOHO value are within their somewhat limited precision consistent with this assumption. However, recent solar wind values derived from lunar and meteoritic samples with a considerably higher precision (Table 4) suggest that the 36 Ar/ 38 Ar ratio in the SW is several percent higher than the air value. As mentioned above, this suggestion relies on the assumption that the lunar and meteoritic values are not severely compromised by systematic errors. The good agreement of the lunar- and meteorite-derived 20 Ne/ 22 Ne ratio with the respective Apollo and SOHO values (Table 4) is a good argument that this is not the case. Two slightly different ( 36 Ar/ 38 Ar) SW values are given in Table 4. The lower one of 5.58 from an acid-etch study of the gas-rich meteorite Kapoeta is preferred by this author, whereas the higher one of 5.77 is adopted in the chapter by Pepin and Porcelli (2002, this volume). If taken at face value, the 36 Ar/ 38 Ar ratio in Jupiter s atmosphere of 5.6±0.25 (Table 7) appears to be another argument that SW-Ar is indeed isotopically slightly different from atmospheric Ar, but we discuss in the Giant Planets section that this argument is less straightforward than it may seem. The Kr and Xe isotopic compositions in the solar wind have not yet been determined directly, neither in space nor with foils. On the other hand, lunar soil samples are well suited to study the isotopic composition of these elements in the solar wind, because the moon contains hardly any indigenous Kr and Xe. The good agreement of the He and Ne compositions deduced from regolithic samples with the values determined in situ or the Apollo solar wind composition experiment gives us confidence that the lunar data for the two heavy noble gases are also reliable. Frequently used recent values are given in Table

13 Noble Gases in the Solar System They are largely based on the first extraction steps of in vacuo etch runs, which should contain essentially the pure and unfractionated SW component, whereas earlier values, e.g., those adopted by Anders and Grevesse (1989), probably reflect the presence of isotopically heavier SEP-Kr and SEP-Xe in bulk sample analyses. Relationships between Kr and Xe in the SW and other planetary reservoirs are discussed in several other chapters in this book (Swindle 2002b; Ott 2002; Pepin and Porcelli 2002). Figure 3 compares various important Xe components, all normalized to the composition of the terrestrial atmosphere. One of the biggest challenges in noble gas cosmochemistry is to explain the relationship between the Xe composition in the Earth s atmosphere and the SW. Atmospheric Xe is depleted in the lighter isotopes by some 4.2% per amu relative to solar Xe (Fig. 3), but a simple mass-dependent fractionation plus additions of radiogenic and fissiogenic isotopes to air-xe would lead to an overabundance of the heaviest two Xe isotopes in the atmosphere. Therefore, in a widely circulated preprint that was to become the by far best-known piece of Samisdat literature in noble gas cosmochemistry, R.O. Pepin and D. Phinney postulated in 1978 that terrestrial atmospheric Xe as well as meteoritic trapped Xe derive from a primordial component different from solar Xe (Pepin 2000; see chapters in this volume by Pepin and Porcelli 2002 and Ott 2002). This component has been dubbed U-Xe (Table 5, Fig. 3). Relative to U-Xe, xenon in the Sun would be enriched by several percent in 134,136 Xe only. This conundrum remains unsolved, but it seems clear that it is not caused by a grossly wrong determination of the heavy isotope abundances of SW-Xe. Atmospheric Kr is probably also slightly fractionated relative to SW-Kr. The composition adopted by Pepin et al. (1995; Table 5) yields a depletion of lighter isotopes of ~7.5 per amu, that of Wieler and Baur (1994) only some 2 per amu, with an uncertainty that would also allow atmospheric Kr to be identical to SW-Kr. More important than this slight discrepancy is the fact that, unlike Xe, atmospheric and solar wind Kr essentially are related to each other by a simple mass dependent isotopic fractionation. The Ne isotopic composition in solar energetic particles has been measured in space since the late 1970s (Dietrich and Simpson 1979; Mewaldt et al. 1979). These first measurements by satellites in (gradual) solar energetic particle events suggested an enrichment of the heavier Ne isotope relative to the SW-Ne composition, but uncertainties were large. Precise data have become available in particular with the ACE mission (Leske et al. 1999, 2001). Unlike in the solar wind, the 20 Ne/ 22 Ne ratio in solar energetic particles is highly variable from event to event, from ~0.4 to ~1.5 times the SW value in the 18 events studied by Leske et al. (2001). This shows that SEPs are less well suited than the SW to determine isotopic abundances in the Sun. The Ne isotopic composition correlates strongly with elemental ratios such as Na/Mg (Fig. 4), suggesting that elemental and isotopic fractionations are both governed by the charge to mass ratio (apart from elemental fractionations due to the FIP effect, see below). The large event-to-event variability makes it difficult to determine the average Ne composition in SEPs (one event may dominate the SEP flux of an entire 11-year solar cycle). Therefore, it is not yet clear whether the average 20 Ne/ 22 Ne in SEPs and the SW differ from each other. Yet, the mean value determined by ACE up to the year 2000 is a few 10% lower than the SW ratio (Leske et al. 2001). This may be important in view of the second trapped solar Ne component (also dubbed SEP) discussed in more detail in the Moon section. In all five noble gases, this component is isotopically heavier than the respective SW values, e.g., 20 Ne/ 22 Ne = 11.2±0.2. Isotopic fractionation Sun solar wind. We noted above that the two He isotopes are expected to be only slightly fractionated between the outer convective zone and the

14 34 Wieler Table 5. Kr and Xe in the solar wind ( 84 Kr 100; 130 Xe 100) 78 Kr 80 Kr 82 Kr 83 Kr 86 Kr Solar wind Ref ± ± ± ± ±.10 Ref ± ± ±.13 terrestrial atm Xe 126 Xe 128 Xe 129 Xe 131 Xe 132 Xe 134 Xe 136 Xe Solar wind ± ± ± ± ± ± ± ±.55 U-Xe ± ± ± ± ± ± ± ±.3 terrestrial atm All values normalized to terrestrial atmospheric Kr and Xe composition by Basford et al : Pepin et al. 1995, average of 14 analyses. 2: Wieler and Baur 1994, in-vacuo etch of lunar soil 71501, 78,80 Kr used for spallation correction. 3: Basford et al : Data of lunar soil by Wieler and Baur 1994, values according to Pepin et al : Pepin Figure 3. Isotopic composition of various Xe components, normalized to 130 Xe and the terrestrial atmospheric composition. Solar wind Xe (SW), U-Xe, and Xe in Jupiter's atmosphere are discussed in the Sun and Jupiter sections (Tables 5 and 7), the meteoritic component Xe-Q (Busemann et al. 2000) is discussed in the chapter by Ott (2002). Xe in Jupiter's atmosphere is probably heavier than terrestrial atmospheric Xe, and might be consistent with either of the other compositions shown. Note that the absolute ordinate position of the Jupiter pattern depends heavily on the choice of the normalisation isotope, as visualized by the error bar on the normalizing isotope 130 Xe (cf. Mahaffy et al. 2000).

15 Noble Gases in the Solar System 35 Figure Ne/ 20 Ne versus Na/Mg in 18 large (gradual) solar energetic particle events measured mostly by the Advanced Composition Explorer (Leske et al. 2001). The Ne isotopic composition varies by almost a factor of 4 between the different events, but correlates well with elemental ratios such as Na/Mg, indicating that both elemental and isotopic fractionations are governed by the ionic charge-to-mass ratio. The average Ne isotopic composition in SEP events may be different from the SW value, which is indicated by the dashed horizontal line. Figure courtesy of R.A. Leske. solar wind, e.g., by ~9% in coronal-hole-dominated (fast) SW. Isotopes with a smaller relative mass difference are expected to show even less fractionation. Potential fractionation mechanisms in the solar wind are discussed by Bochsler (2000), Kallenbach et al. (1998b), and Kallenbach (2001). The 24 Mg/ 26 Mg ratio is expected to be enhanced in the SW by no more than 3% due to inefficient Coulomb coupling (Bodmer and Bochsler 1998a,b) and probably by 1-2% due to gravitational settling in the outer convective zone (Kallenbach et al. 1998b; Bochsler 2000). Precise Mg and Si data from the SOHO, ACE and WIND missions can be used to test these predictions (Kallenbach et al. 1998b; Kallenbach 2001). The SOHO data are shown in Figure 5. At a SW velocity of ~620 km/s the Si and Mg isotopic compositions are identical to the terrestrial or meteoritic values within a systematic uncertainty of the instrument calibration of ~1.5% per amu. At lower velocities, the light isotopes become slightly enriched, by (1.4±1.3)% per amu at 350 km/s (dashed line in Fig. 5 and its 2σ error). So, whereas the systematic uncertainties are still somewhat too large to firmly conclude whether Mg and Si are isotopically fractionated in the bulk solar wind, the results strongly indicate a slightly variable isotopic composition as a function of solar wind speed, with the heavy isotopes being less abundant in the slow wind. Kallenbach (2001) reports isotopic ratios of various elements in the bulk solar wind obtained as average values from various missions. The 24 Mg/ 26 Mg and 28 Si/ 30 Si ratios obtained from WIND and SOHO (R. Kallenbach, pers. comm.) are both larger than the solar system values, by (1±2.2)% and (1±1.4)%, respectively. Ne and Ar in the solar wind can be expected to be similarly fractionated as Mg and Si. This suggests that 20 Ne/ 22 Ne and 36 Ar/ 38 Ar values in the solar wind differ from solar values probably by between 1-3% (R. Kallenbach, pers. comm. 2001). The solar Ne and Ar isotopic compositions adopted here (Table 4) take this Mg-Si-derived fractionation and its uncertainty into account. The SW isotopic compositions used to derive the solar values are those based on lunar and meteorite samples. Table 4 illustrates that the Ne and Ar isotopic compositions of the bulk sun are probably less accurately known than the currently most precise solar wind data may suggest, due to the uncertainty of the fractionation correction sun-solar wind. Interestingly, the solar Ar composition is within uncertainty still equal to the terrestrial atmospheric value or that of primordial meteoritic Ar (Busemann et al. 2000; Ott 2002, this volume), as was already the case for the Anders and Grevesse (1989) value. Therefore, even though we are quite sure today that the atmospheric 36 Ar/ 38 Ar ratio is slightly different from the solar wind value, it is still uncertain whether terrestrial atmospheric Ar is fractionated relative to bulk solar Ar (see also the discussion of Ar in Jupiter in the Giant Planets section).

16 36 Wieler Figure 5. Isotopic compositions of Mg, Si, and Ne as a function of solar wind velocity, as measured by the SOHO mission. The ordinate shows the weighted average variations of the abundances of the heavier isotope of the three elements (per amu), relative to the abundances at 620 km/s. For Mg and Si, the isotopic compositions at 620 km/s are identical to the terrestrial values, although all Mg and Si data may be systematically offset by up to 1.5% per amu due to uncertainties in the instrument calibration. [Used by permission of Kluwer Academic Publishers, from Kallenbach et al. (1998b), Space Science Reviews, Vol. 85, Fig. 5, p 364.] Elemental abundances of noble gases in the solar wind; the FIP effect in Kr and Xe. Elemental abundances of noble gases in the SW are given in Table 2 and discussed in the Moon section, as these data are largely derived from lunar (and meteoritic) regolith samples. Here we briefly mention the particularly interesting case of Kr and Xe. Both Kr/Ar and especially Xe/Ar in lunar samples are systematically higher than the inferred solar ratios. Very probably this is not due to fractionation during or after implantation but rather indicates an enrichment of Xe and Kr in the solar wind source region. Such enrichments are well known for elements with a first ionization potential (FIP) less than about 10 ev and the enrichment factor for Xe of ~4 is similar to that of low-fip elements in the (slow) solar wind (see the Moon section for details). THE GIANT PLANETS The He abundances The hydrogen- and helium-dominated atmospheres of the giant planets remind us strongly of the composition of the Sun. Therefore, the predominant view prior to the Voyager missions was that the He abundance in the atmospheres of the giant planets is identical to the protosolar value, and it was hoped that an accurate determination would be of importance for our understanding of the evolution of the universe. The first He abundance determination in the Jovian atmosphere by infrared spectroscopy onboard the Pioneer 10 and 11 missions (He mole fraction = 0.12±0.06; Orton and Ingersoll 1976) was too inaccurate to test this assumption. However, a few years later the Voyager data revealed a considerable variability in the He abundances of the four giant planets, indicating that Jupiter and, in particular, Saturn are depleted in He in their outer atmospheres. Therefore, rather than being of cosmological significance, an accurate measurement of the He abundance in the atmospheres