By section:
1. Introduction
2. Photospheric Isotope Ratios
3. Solar Wind Isotopic Ratios
3.1 He
3.2 Ne
3.3 Mg
4. Isotopic Ratios of Solar Flare Ions
5. Meteorite Isotope Anomalies
5.1 CAIs
5.1.1 Linear isotope anomalies
5.1.2 Non-linear isotope anomalies
Mg, Ca, Ti, Cr, Ba anomalies
5.2 Oxygen isotope variations
5.2.1 General Significance of O Isotope Variations
5.2.2 Interstellar Component Model
Predictions of solar composition from the 2-component
model
Multicomponent nucleosynthetic models
5.2.3 Mass-independent O isotope anomalies
5.3 Noble Gas Isotopes in
Meteorites
5.4 N Isotopes in Meteorites
6. Solar Wind Abundances from Lunar Samples
7. Noble Gas, Nitrogen and C Isotopes; Implications for
Planetary Atmospheres
8. Lunar N Isotopes
9. Higher Energy Solar Particles
This Document reviews what is known about this major issue. In general one does not expect qualitatively different isotopic compositions for elements in solar matter than those found on Earth, e.g. all 18O instead of predominantly 16O as on Earth. (There are very interesting exceptions in H, He, Li, and B; these are considered separately in Document B for H and He, and in Document F for Li and B). Although surprises are always possible, the limited isotopic data available for solar matter (see below) supports the expected qualitative similarity. Consequently, it would be irresponsible to design a solar wind sampling/analysis strategy that depended on finding major (factor of 2 or larger) differences. However, as discussed below, it is plausible to expect ubiquitous small differences in the relative abundances of the isotopes of a given element, (i.e. in the ratio of the number of atoms of two isotopes). As discussed in Document H it is reasonable to assume that either there is no isotopic fractionation between the solar wind and bulk solar matter from the photosphere or that any small corrections can be reliably made.
Theoretically, the input materials to the solar system from stellar nucleosynthesis are expected to be isotopically very heterogeneous because different stars and different parts of a single star are predicted to contain vastly different proportions of the isotopes of a given element. This expected heterogeneity has been confirmed by large variations in the isotopic compositions of intact interstellar grains recovered from meteorites (see e.g. the review by Anders and Zinner, 1993). Qualitatively, this enormous heterogeneity has been mixed away; however, exceptions (such as the direct observations of interstellar grains) to perfect mixing exist, all of which have major implications. The challenge is to recognize and quantify these exceptions. It is worth emphasizing that one does not have to have percentage wise large (say >10%) variations to draw important conclusions. What is important is the accuracy to which one can measure differences, not their magnitude.
In natural terrestrial and lunar matter, isotopic abundance variations arise from (a) radioactivity, (b) isotopic fractionation accompanying physical or chemical processes (evaporation, diffusion, precipitation, etc.), and (c) nuclear reactions (galactic cosmic rays, or to a lesser extent, solar flare ions for lunar samples). These processes have been intensively studied and are relatively well understood. Although the level of precision varies greatly for the various elements, the overall conclusion is: Among terrestrial materials there is no evidence for isotopic variations which cannot be accounted for by the above processes and which would represent "residual" isotopic heterogeneities from the isotopically diverse input materials to the solar system. For example there have been thousands of measurements of terrestrial 86Sr/88Sr, and no variations in this isotopic ratio have been reported beyond what can be accounted for by mass fractionation in thermal ionization sources (circa 1%). Recent observations of 20Ne/22Ne and 129Xe variations in volcanic rocks are possibly the first indications of residual isotopic heterogeneities among terrestrial materials.
Considering only data for non-volatile elements among lunar and terrestrial materials, there is no evidence for systematic isotopic differences between terrestrial and lunar matter which would represent "residual" isotopic heterogeneities. As discussed below, the situation is more complicated for volatile elements in lunar materials because these are predominantly from non-lunar sources.
The overall isotopic homogeneity of lunar and terrestrial material for non-volatile elements represents thorough mixing by a combination of processes: (a) during or subsequent to the formation of the Earth and Moon, (b) in the solar nebula prior to the formation of the Earth or (c) in the interstellar medium prior to incorporation into the solar nebula. This mechanism involves destruction of grains, isotopic homogenization in the gas phase, and then recondensation into a true interstellar grain. The survival of some interstellar grains in meteorites shows that process (c) alone did not produce total isotopic homogenization. Isotope anomalies in meteoritic materials, discussed below, show that the mixing associated with (b), although good, was not perfect. For the Earth and Moon, the isotopic heterogeneity that escaped (c) and (b) has apparently almost entirely been eliminated by (a). Establishing the relation of solar matter to that of meteoritic or terrestrial/lunar matter is a major goal of a solar wind sample return mission (compare Document D).
We now proceed to a summary of the actual data on which the above
overview is based.
As discussed in Document B, except for the lightest elements (H, He, Li, B), strong arguments can be made that the material in the outer convection zone of the Sun preserves the original solar chemical and isotopic composition. It is plausible, and generally accepted, that this composition also applies to the solar nebula out of which the planets formed. This major assumption can be tested by a solar wind sample return, as discussed specifically in Document C.
The
photosphere represents the portion of the outer convection zone accessible
to observations, but determination of isotopic abundances from photospheric
observations is difficult, and little information is available. CO
molecular observations from sunspot umbra by Hall (1973) set upper limits
on any differences in the photospheric and terrestrial isotopic abundances
of C and O: 13C/12C,
15%; 18O/16O, 35%;
and 17O/16O, factor of 2.5.
A thorough high resolution study of hundreds of CO lines by Harris et
al.
(1987) gives 12C/13 = 84±5 and
18O/16 = 440±50. Analogous
observations of MgH limit differences in 25Mg/24Mg
and 26Mg/24Mg to less
than around 20% (Boyer et al., 1971). This level of precision is not
sufficient for planetary science purposes.
Data are available for He, Ne, and Mg.
He. Precise 4He/3He ratios were obtained from Al foils exposed to the solar wind during Apollo (SWC experiment; Geiss et al., 1972). The ISEE-3 mass/charge spectrometer provided essentially continuous coverage of 4He/3 He over a 3-4 year period (1978-1982) with the interesting result that, although hourly averages of this ratio show order of magnitude variations, the longterm average from Aug. 1978 through Dec. 1981 (2050±200) is within errors of the average of the Apollo foil results (2350±120) which represents about 5 days total exposure in a period around solar maximum . Here the ± represents the variability of the data, independent of systematic error estimates. The ISEE-3 ratio for speeds greater than 450 km/sec (1900) is within errors the same as the average for all speeds. Consistent results have been obtained with the SWICS instrument on Ulysses (Bodmer et al., 1995). Based on daily averages, the mean for roughly 500 days during 1992 and 1993 is 2290±200, in good agreement with the ISEE-3 and the Apollo data. Thus, despite short-term variability, there appears to be a well-defined long-term average value for 4He/3He. During 1992-1993 Ulysses alternately sampled low-speed "interstream" solar wind and high-speed solar wind from polar coronal holes; however, no systematic variation with solar wind speed was observed for these different solar wind regimes.
Unlike other elements, the solar He isotopic composition will differ from that of the solar nebula because all solar D has been destroyed by thermonuclear reactions producing 3He, greatly enhancing the 3He abundance (Geiss and Reeves, 1972). (This complete destruction appears to be unique to D). Terrestrial planet atmospheres do not retain He and are dominated by 4He produced by radioactivity, so no comparisons are possible here. Gas-rich meteorites (see below) have significant noble gas contributions from unprocessed interstellar materials. As expected, the 4He/3He for Jupiter from the Galileo probe data is 9000 (±10%), much higher than the solar wind ratios (Niemann et al., 1996). The 9000 value is the best available estimate of the nebular 4He/3He, although if liquid He core formation has occurred on Jupiter, this might produce a significant 3He enrichment in the atmosphere.
Ne. Although data from WIND and SOHO instruments should be available soon, only the Apollo foil data are available at present. Figure A1 is a correlation plot of the He and Ne isotopic data from the Apollo solar wind foils. Each measurement is an average over 1 to 44 hours. There are variations in 4He/3He outside of errors, but there is no evidence for any variation in 20Ne/22Ne with an upper limit of daily variations of about ±2% (Geiss et al., 1972). As noted above, even with the relatively small amount of time averaging, the Apollo 4He/3He average converges to that of the ISEE-3 and Ulysses. Any variation in 20Ne/22Ne e.g., due to isotope fractionation during solar wind acceleration, would be expected to correlate with 4He/3He but no Ne isotope variations are observed.
Mg. Spacecraft mass spectrometers on SOHO and ACE should be capable of measuring isotopic ratios to better than ±10% (one sigma) for the major (>0.01-0.02 in fractional isotopic abundance) isotopes for Ne, Mg, Si, S, Ar, Ca, and Fe. Initial data from an equivalent instrument on WIND (Bochsler et al., 1996) show 25Mg/24Mg and 26Mg/24Mg ratios within ±5% ( 1 sigma) of the terrestrial ratios. Data from the SOHO instrument will be better, and errors will also decrease when better instrument calibrations are carried out. These and 22Ne/20Ne are favorable cases, and here it may be possible to approach the ±1% (2 sigma) requirements for sample return isotopic precision. However, in these favorable cases with a returned sample, it will be possible to greatly exceed the general ±1% precision requirement. In general, with a returned sample, it should be possible to exceed the isotopic precision of in-situ instruments by at least an order of magnitude for any element.
Beyond 20Ne/22Ne and the Mg isotopes, the SOHO and ACE instruments will not have sufficient mass resolution to provide CNO isotopic abundances of adequate precision for planetary science purposes. For the SWIMS instrument (Hamilton et al., 1990) FWHM mass resolution is adequate (their Figure 5); however, it is the non-Gaussian tails of the peaks which determine the ability to measure accurately isotope ratios requiring large dynamic ranges. Their Figure 4 shows that tails are present on the high mass sides of peaks which are worth a few percent of the peak intensity at one mass unit above the main peak for the mass 10-20 range. The 17O/16O to be measured will be about 4 x 10-4, so the signal to noise at mass 17 for SWIMS will be of the order 1/25; a mass 17 peak will not be visible in the SWIMS spectrum. For planetary science purposes the 17O/16O ratio needs to be measured to an accuracy of 0.1%, i.e. it is important to distinguish clearly 3.70 from 3.71 x 10-4. Analogous problems exist at mass 15 for the measurement of 15N/14N. Other measurements made difficult because of very large neighboring peaks include important isotopes such as 19F, 21Ne, 41K, 55Mn, 53Cr, 57Fe, 59Co, and 61Ni.
An alternative, but simpler, argument which cuts to the heart of the CNO isotope issue is that there is no time-of-flight mass spectrometer which, in the laboratory, can measure CNO abundances with 0.1% precision. Such measurements are always done with magnetic sector instruments, although good time-of-flight instruments are available. It is unlikely that a spacecraft time-of-flight instrument will provide the required precision.
The general issue of solar wind isotope fractionation is considered
separately in
Document H.
The case of Ne is especially
interesting in that for the last decade it was believed that there were
systematic differences between the Ne isotopic compositions for the solar
wind and for solar flare ions. Five independent measurements from the
Apollo solar wind foils provide a precise and well-defined
20Ne/22Ne=13.7 (Fig. A1). For the large flare of
Sept. 1978, Mewaldt and Stone (1989)
report 20Ne/22Ne=9.7±1.8. A relatively large
correction factor of 1.28 for
charge/mass fractionation has been applied in obtaining this solar flare
ratio. In contrast, two large flares in October and November of 1992
showed higher ratios with the more intense November flare giving
20Ne/22Ne=13.9±1.7 after correction for
charge/mass fractionation by
a factor of 1.11 (Selesnick et al., 1993). At this point there is
no way
to tell whether the 1978 or the 1992 flares, or both, are anomalous. This
is to be contrasted with the uniformity of the solar wind Ne isotope data.
Mg isotopic data for the 1978 flare show no differences greater than
around 20% for both 25Mg/24Mg
and 26Mg/24Mg (Mewaldt et al., 1981)
relative to the terrestrial Mg isotopic composition. The presence of
observed flare to flare Ne isotopic variations and the existence of
ubiquitous charge/mass fractionation (
Document B) appear to make solar
flare ions a less reliable source of solar isotopic compositions than the
solar wind.
This is a fascinating, but complex, set of data. We only summarize the most relevant points for the science objectives of a solar wind sample return. General reviews are, e.g. Podosek (1978), Thiemens (1988), Clayton et al., (1985), Lee (1988), Harper (1993), and Thiemens (1996).
Nonlinear anomalies.
A
major type of isotopic variation relative to terrestrial material
("anomalies") is that referred to as "nuclear",
"unknown nuclear" or "non-linear" which, in
English, means those isotopic variations which cannot be accounted
for by processes of radioactivity, mass dependent chemical/physical
isotope fractionation, or recent energetic particle bombardment.
In considering these anomalies, it is important to recognize that
chondritic meteorites are complex mixtures of materials
("components")
with independent origins, thus order of magnitude isotopic variations
can exist in micron and sub-micron sized interstellar materials
within a chondrite, but if these contain a small fraction of the
inventory of a given element, they have a negligible effect on
the bulk meteorite isotopic composition, which in turn reflects
an average of the components, essentially all of which formed
within the solar system.
Ca-Al-rich
inclusions (CAIs) (MacPherson et al., 1988) found in carbonaceous
chondrites have been a rich source of isotopic anomalies. These
are high temperature materials with characteristic enrichments
of highly non-volatile (refractory) elements, e.g. Ca and Al,
with respect to less refractory elements such as Fe and Si. CAIs
formed in the solar nebula, apparently very early, by processes
that are still matters of debate. Many of the observed isotopic
variations in CAIs appear to reflect much larger degrees of physical
mass fractionation effects than are present in terrestrial materials,
probably due to high temperature distillation processes (e.g.
Clayton et al., 1985; Papanastassiou and Brigham, 1989;
Ireland et al., 1992; Clayton, 1993).
5.1.1 Linear Mass Fractionation Anomalies in CAIs. In contrast to the "non-linear" anomalies discussed above, some CAI samples show "linear mass fractionation", characterized by isotopic variations relative to terrestrial matter that systematically increase, approximately linearly, with the difference in mass of two isotopes, e.g., the fractional difference for 18O/16O is approximately twice that for 17O/16O. The linear anomalies are only found in CAIs. There is no evidence for these between bulk meteorite, terrestrial, and lunar matter. An interesting complication/exception is that there are some gas phase molecular chemical processes in which "mass independent" isotopic fractionations are observed (e.g. Thiemens, 1996), whereas one would have expected to see linear mass fractionations. (In the present context, these are also considered non-linear.) These possibly play a role in producing O isotope variations (see below).
Analytically, linear mass fractionation
variations are more difficult to detect because all laboratory
mass spectrometric data must allow for instrumental mass fractionation
effects which are also approximately linear. Thus, the application
of instrumental corrections can remove any natural as well as
instrumental mass fractionation! The mass fractionation anomalies
in CAIs have been observed because they are much larger than can
be explained by measurement to measurement variations in
instrumental fractionation. "Dynamic" gas mass spectrometric
techniques (in common use for H,C,O isotopic analyses) rapidly
alternate samples and standards, automatically correcting for
mass fractionation variations, but such techniques require far
too much sample to be used for a solar wind sample return. For
thermal ionization mass spectrometry, special "double spike"
techniques can be used which independently determine instrumental
and natural mass fractionations (e.g. Eugster et al., 1969,
Niederer et al., 1985). In favorable cases, e.g. Ca, these
might be applicable to a returned solar wind sample. In general
our solar wind isotopic precision requirement of 1% two sigma
can be met for linear mass fractionation isotopic variations because
corrections for instrumental variations are significantly less
than 1%/amu.
5.1.2 Non-linear anomalies in CAIs,
i.e. those which show no systematic mass dependence, have been
reported for O, Mg, Si, Ca, Ti, Cr, Fe, Ni, Zn, Sr, Zr, Ba, Nd,
Sm, and Dy (see e.g. Harper, 1993 for references). An important
point is that the CAI non-linear anomalies are relatively small,
parts in 103
to 104,
except for O where anomalies ranging up to 4% in
18O/16O
have been found (e.g. Clayton, 1993). There is no doubt that the
CAIs formed in the solar system (some of them are cm-sized igneous
rocks). In principle there could be unprocessed relict interstellar
grains included in CAIs, but no unambiguous example of these has
been discovered despite serious searches. The origin of the non-linear
CAI anomalies is possibly related to the fact that the CAIs were
among the first-formed materials in the solar system (possibly
the first-formed material). Thus, CAIs may have formed
from solar nebula material at a time when the solar nebula was
isotopically more heterogeneous than the present day inner solar
system, i.e at a time when mixing was less complete than it is
today. The precise time of separation of the Sun and planetary
material relative to CAI formation is unknown, but isotopic heterogeneities
could have been present which remain as systematic isotopic anomalies
between solar and terrestrial matter (Document
D).
Mg, Ca, Cr, Ti, and Ba Systematic non-linear isotopic differences in bulk (>0.1 g) samples among various types of planetary materials are only well-documented for noble gases, N, O, Cr, and Ti. There are also preliminary reports of systematic differences in 137Ba/138Ba at the part in 105 level between terrestrial and bulk chondritic samples (Harper et al., 1992). Anomalies in O, noble gases, and N are discussed below. The Cr anomalies (part in 104 differences) are for 53Cr and are between some (but possibly not all) meteorites and terrestrial/lunar material (e.g. Birck and Allegre, 1985, 1988; Harper and Wiesmann, 1992; Lugmair et al., 1996). For Ti, anomalies are for 50Ti and are between 3 and 12 parts in 104 between chondrites and terrestrial materials (Niederer et al., 1985; Niemeyer, 1988).
Our measurement objective priorities for a solar wind sample return give special emphasis to Mg, Ca, Ti, Cr, and Ba (Objective 7 of Feasibility Study Proposal Table 1-2) relative to a general survey of solar-terrestrial isotopic differences (Objective 12). The rationale for Objective 7 was to select elements for which there is a greater than random chance for seeing isotopic differences between the Earth and Sun, i.e. to cherry pick from the general isotopic survey. Ti and Cr satisfy this requirement because of the systematic, but small, isotopic differences among bulk meteorite samples and the Earth. To some extent there is no difference between Ti/Cr and O except that the O isotopic anomalies are larger, better documented, and more famous. Also, beyond generic isotopic heterogeneity in the solar nebula, it is likely that the details differ for why O and Ti variations have survived.
53Cr isotope anomalies (Lugmair et al., 1996) appear to vary systematically with distance of formation from the Sun and might represent a late addition of interstellar (supernova?) material to the solar nebula. The formation of the solar system occurred in our parent molecular cloud, and a close supernova is not a ridiculously implausible idea (Foster and Boss, 1996). The supernova ejecta might contain significant amounts of radioactive 53Mn which otherwise did not exist in the solar nebula and which would decay to 53Cr. The general idea of a late supernova injection has often been discussed but is difficult to constrain (e.g.,Wasserburg and Papanastassiou, 1982, Fahey et al., 1987). If sufficient quantities of 53Mn were non-uniformly distributed, spatially varying 53Cr anomalies would be impressed on the solar system. The apparent trend of Lugmair et al., predicts that the solar 53Cr/52Cr ratio should be less than the terrestrial value. Alternatively, if the apparent radial trend is accidental and late supernova injections are insignificant, then the 53Cr anomalies probably reflect chemical fractionation of Mn and Cr during the formation of the Earth from a 53Mn-rich nebula, in which case a higher 53Cr/52Cr ratio is predicted for the Sun.
Anomalies in Mg due to a similar heterogeneous solar system distribution of 26Al have been discussed, analogous to those proposed for 53Mn. The presence of 26Al is well documented in a variety of chondritic materials (MacPherson et al., 1995; Russell et al., 1996). However, there is no evidence for bulk sample 26Mg/24Mg variations produced by 26Al decay analogous to those for 53Cr. Assuming that there are heterogeneities in the distribution of 26Al /27Al and 53Mn / 55Mn in the solar nebula, the difference between Mg and Cr is not obviously explained by the 6 times longer half-life of 53Mn because the issue is whether the heterogeneities are mixed away, not when the mixing occurs. In the case of a homogeneous distribution of 26Al /27Al and 53Mn /55Mn in the early solar nebula, there is a window of time such that mixing and chemical fractionation can occur after 26Al is decayed but 53Mn is still alive leaving only Cr isotopic differences among planetary materials. Cr is a more favorable case than Mg in that for CI chondritic elemental abundances, 53Mn/53Cr is 8 times larger than 26Al/26Mg.
Ca, Ti, Cr, and Ba are elements which, in meteoritic jargon, show "negative isotope anomalies" in some meteoritic materials, although not bulk samples. In English, this means that an extra amount of some isotope was added to the Earth but was omitted from some special meteoritic inclusion. There are major differences in implications between negative and positive anomalies because, in principle, only a few grams of anomalous material need be floating around the solar nebula and captured by the inclusion to produce positive anomalies. For negative anomalies, many tons of anomalous material had to be added to the Earth and missed by the inclusion. This might represent major compositional gradients in the nebula or late additions of a large amount of interstellar material to the nebula. For both these reasons, these elements are the first choices to look for significant residual isotopic differences between the Earth and Sun.
Given the relatively small numbers
of atoms/cm2
of Ca, Ti, Cr, and Ba in a solar wind sample return, measurement
of part in 103-104
isotopic differences is challenging, although not impossible.
The relatively large amounts of Mg in solar composition make a
high precision measurement more feasible for this element. It
is possible, however, that the differences observed among meteorites
and the Earth are the tip of the iceberg and that variations in
the 0.1-1% range exist for solar matter with respect to either
meteorites or the Earth.
Systematic differences in the O isotopic composition of planetary and meteoritic materials have been documented over the past two decades by R.N. Clayton and co-workers (Clayton et al., 1985; Clayton, 1993). It is surprising that the most abundant element in meteorites, O, shows variations among all inner solar system materials and shows variations larger than all other rock-forming elements. Even excluding CAI materials (see above), the total range in 18O/16O between different types of meteoritic materials and the Earth is 1.4%, at least an order of magnitude larger than bulk sample anomalies for Ti or Cr. As there is no reason to exclude the CAIs from this comparison, the total range is around 4.9%. Combining data for both 17O/16O and 18O/16O on a two-isotope correlation plot permits resolution of families of solar system materials, an example of which is shown in Figure A2.
5.2.1 General Significance of O Isotope
Variations. Figure A2 shows
that the solar system has a resolvable O isotopic structure with
each part of the solar system appearing to have distinct proportions
of the three O isotopes. The cause of the variations is unknown
at present. An important goal of 21st century planetary science
will be to discover and interpret the solar -system-wide structure
in O isotopes. Understanding solar nebular evolution is inextricably
linked to understanding the O isotope variations because they
are a -- possibly unique-- surviving fossil record of the nebular
materials and/or processes that has escaped obliteration in subsequent
planetary evolution. In fact it is widely accepted that the origin
of the solar system cannot be understood without understanding
the origins of the O isotope variations. As the major mass reservoir
in the solar system, the Sun is obviously important, and the solar
O isotopic composition plays a key role in any interpretation.
For these reasons measurement of O isotopic composition is the
highest priority measurement objective for our solar wind sample
return.
5.2.2 Interstellar Component Model. A plausible (and popular) interpretation of the O anomalies is that they reflect incomplete mixing of nucleosynthetic components. Although probably oversimplified, it cannot be ruled out that there could be as few as two components, one richer in 16O and both having approximately the same 17O/18O ratio (Clayton, 1993) The 16O-rich component could be the average of the nebular dust and the 16O-poor component be a relatively well-mixed nebular gas. Individual samples are interpreted as representing various degrees of O exchange between solids and gas. [This model is certainly an oversimplification in that interstellar Al2O3 grains with large variations in 17O/18O (e.g. Nittler, 1996; Hoppe et al., 1995) have been found, and, somewhat surprisingly, no such grain has been found with enrichments in 16O, as required by the two-component model. However, there is also no question but that the combination of (a) survival probability in the solar nebula and (b) the present relatively-crude techniques for extracting interstellar grains results in a highly biased sample of the interstellar grain population extant in the solar nebula.] To explain why all the data on Figure A2 do not lie exactly on a single mixing line, the extreme two-component model must postulate that chemical/physical linear mass fractionation is superimposed on the products of component mixing. [In terms of Figure A2 this moves point off the basic "slope 1" mixing line in terms of relative enrichments (constant 17O/18O) onto a flatter "slope 1/2" line characteristic of linear mass fractionation] This model can explain why O isotope anomalies are much larger than for other rock-forming elements: O is unique in being a major element in both the gas and solid phases in the solar nebula (Clayton, 1993). At the time of formation of the various components in chondrites the O isotopic heterogeneities among the solid phases may have been much less (or more effectively averaged out) than between the gas and solid phases, giving rise to effectively 2 components--solid and gas. At the earlier period of CAI formation both intra-solid and solid-gas phase isotopic heterogeneities may have been larger. Oxygen may be the only element capable of sensing the larger gas-solid heterogeneities, explaining why larger isotopic anomalies are found in oxygen than other rock-forming elements.
In terms of this model, one can view
the solar O isotope composition as the average nebula composition
towards which different planetary materials were evolving when
separated from the nebula or the isotopic composition from
which later-formed objects (planets?) formed. It may turn out
that various planetary objects form a time sequence in terms of
O isotopes with the earliest formed materials being the most distant
on Figure A2 from solar
composition and vice versa. This relative
formation sequence might be unobtainable in any other way.
Predictions of Solar Composition from the Two-Component Model. A general prediction of the two-component model is that the solar O isotopic composition should lie on the mixing line in Figure A2. The simplest specific model prediction would (arbitrarily) regard the meteoritic material with the greatest 16O depletion (the Carlisle Lakes chondrites) as totally equilibrated with nebular gas and thus have the solar composition, except for possible mass-dependent fractionation. Gas in equilibrium with CL solid material would be 0.1%/amu enriched in 17O and 18O (Onuma et al., 1972), so that a solar composition of ~21% O from CL-like solids and ~79% O from equilibrated gas (solid/gas proportions from Anders & Grevesse, 1989) would give the composition plotted as SCL on Figure A2. A different two-component model based on O isotope data for different components in the Murchison meteorite (Clayton and Mayeda, 1984) predicts a solar composition as indicated by SM on Figure A2.
Solar predictions based on the two-component model are discussed further by Wiens et al. (1997).
At the other extreme it is quite possible that there was essentially no interaction between gas and solids in the solar nebula. To understand the origin of chodrules (molten silicate droplets which give chondritic meteorites their name) it is generally accepted that enhancement of the ratio of solid to gas relative to the average solar nebula value is required (settling of solids to the nebular midplane?) In the extreme case inner solar system material may have formed entirely from solids, whereas solar O would be a mixture of gas and solids, producing significant solar-terrestrial O isotopic differences with the solar O isotopic composition potentially being anywhere on Figure A2.
Multicomponent Nucleosynthetic Models.
It is likely that the two-component model is oversimplified.
In the 21st century we will presumably know the spatial relations
for O isotopes in the present solar system. In terms of a general
nucleosynthetic component model, this knowledge, coupled with
precise solar O isotope data, will allow models of the distribution
of different classes of interstellar grains and gas in the solar
nebula. If principle, if one knew the O isotopic compositions
of all the different interstellar grains, the O isotopic
composition of each planetary object could be modeled as a mixture
of these grains. The equivalent model for solar O isotopic composition
would specify the proportions of the various stellar inputs to
the solar system, a result of immense importance. More realistically,
even by the end of the 21st century we may not have a reliable
complete inventory of the interstellar materials that went into
the solar system simply because many, conceivably the most abundant,
types of grains were completely destroyed by nebular and planetary
processing. In this, more realistic, case the solar O isotopic
composition is more important because relative to it (as
the solar system average) it will be possible to recognize excesses
or depletions of specific known types of interstellar grain contributions
in different planetary materials and thus gain knowledge of the
distribution of some interstellar materials in the nebula.
5.2.3 "Mass independent" O isotopic anomalies This is a special case of a non-linear anomaly. A viable alternative model to explain the variations shown on Figure A2 invokes relatively recently discovered "mass-independent" isotopic fractionation accompanying some types of gas phase molecular reactions (e.g. Thiemens and Heidenreich, 1983; Thiemens, 1988; Thiemens, 1996). For O these processes primarily change the 16O abundance, leaving the 17O/18O ratio unaffected. To date, three types of reactions show mass-independent O isotope fractionations: (a) recombination reactions of gas phase O molecules, (b) gas phase isotope exchange reactions and (c) thermal dissociation reactions. Although there is some basic understanding of the mechanisms and all of the above reactions are potentially relevant for the solar nebula, detailed models of nebular processes have not been worked out.
In terms of the mass independent molecular
reaction model, knowledge of the solar O isotopic composition
is required to quantitatively model the consequences of such processes.
Specific models can be anticipated in the next few years; however,
some general qualitative predictions are possible. For the case
of recombination reactions the solar composition would be more
enriched in 16O than the most 16O-depleted meteoritic
materials.
This is a distinct prediction from that made by the two-component
mixing model.
Self-shielding isotope fractionation during molecular photodissociation. Differential effects of self shielding are well established in astrophysical CO studies for which observations of 13C16O lines look deeper into molecular clouds than for 12C16O lines for which many clouds are opaque (e.g. Langer, 1991). Nevertheless, whether this is important or not for the solar nebula is dependent on nebular models. Assuming that self shielding effects on molecular dissociation are important, one can plausibly guess the direction of the effect on O isotopic variations in meteoritic materials. For simplicity assume that there initially both solids and gas had a homogeneous O isotopic composition. High energy solar photons can dissociate CO and H2O gas phase molecules. Assuming that self-shielding is important, the 17O- and 18O- bearing molecules will be preferentially dissociated and the radicals produced will be preferentially incorporated into grains, making the grains 17O- and 18O-rich and 16O-poor relative to the initial composition which is now preserved in the Sun. In an extreme model the most 16O-rich meteoritic material (CAI) would be closest to the solar value, essentially the opposite prediction from the nucleosynthetic 2 component models.
It is likely that by the time of return
of a solar wind sample these models will be obsolete; however,
all models will make some prediction about the solar O isotope
composition.
It has been widely, but not universally (Huss et al., 1996), assumed that meteorites contain noble gases (and N) "trapped" from the gas phase of the solar nebula, but, beyond expected contributions from radioactive decay and galactic cosmic ray nuclear reactions, there are also major contributions from interstellar grains and in some cases from solar wind implanted during surface residence of some of the material on the meteorite parent body. The presumably nebular noble gases are (inappropriately) referred to as planetary noble gases because the abundance ratios of light (He,Ne) to heavy (Kr,Xe) noble gases are much less than solar abundance estimates as is also observed for the terrestrial atmosphere. The noble gases are unique elements in having prominent interstellar grain contributions to bulk meteorite analyses because: (1) the interstellar grains have relatively high noble gas concentrations and (2) the amounts of possible nebula gases are relatively small. Thus, most of the trapped He and Ne arises from the interstellar grains, but most of the Ar, Kr, and Xe might be nebular in origin. A major part, perhaps all, of the presumably nebular gas appears to be adsorbed on elemental carbon. Because this carbon is insoluble in a mixture of HF and HCl acids, this portion of the nebular gas ("Q") is concentrated in such an "acid-insoluble residue". Unfortunately for this purpose, many of the interstellar grains are also concentrated in this residue, complicating independent analysis. However, because Q is lightly bound it is preferentially released by further treatment of the "acid insoluble residue" with an oxidizing acid such as HNO3. There are two ways to get at the isotopic compositions of Q noble gases: (a) by analyzing the gases released when an HF-HCl insoluble residue is treated with HNO3 during "closed system etching" (Wieler et al., 1992) or (b) by difference in samples of an HF-HCl insoluble residue before and after treatment with oxidizing acids (Huss et al., 1996). The inferred noble gas isotopic compositions from these two approaches are close, although analytically significant differences remain.
If the solar nebula were completely homogeneous isotopically,
the meteoritic nebular noble gas isotopic compositions should
be solar. We defer a detailed discussion of the heavy noble gases
(Ar, Kr,Xe) in order to include data on solar noble gases in lunar
samples. Here we only note that the 20Ne/22Ne
ratio of 10.7±
0.2 for Q (Wieler et al., 1992) is distinctly less than the 13.7
for the present day solar wind from the Apollo foil experiments
and distinctly higher than 9.8 for the terrestrial atmosphere.
The Q- solar wind difference is in the right direction to be
due to isotopic fractionation during low temperature adsorption
of nebular Ne onto meteoritic C, but the magnitude is surprisingly
large. Pepin (1991) explains the Q-solar wind difference as a
consequence of isotopic fractionation of solar composition Ne
during loss of a transient atmosphere of a large asteroidal parent
body. Qualitatively consistent fractionations would be expected
for other noble gas isotope ratios. This and other models can
be tested with noble gas isotopic compositions from a solar wind
sample return.
A recent review of these has
been given by Kerridge (1995). Even in bulk samples, a large
range from 15% depletion in 15N / 14N
to a factor of 2 enrichment has been observed. A very important
point made by Kerridge is that, in contrast to the noble gases,
the measured concentrations of N are far too high to explain the
observed isotopic variations by the presence of known interstellar
grains, even though these are known to have large (up to order
of magnitude) variations in 15N/14N.
Variations of the order of 1% in bulk 15N/14N
are likely, however, due to the effects of interstellar grains
in 15N/14N. The large
concentrations of C in the CI and CM chondrites are believed to
have formed in the solar system and have the terrestrial
13C/12C
within about 2%. This C contains relatively large amounts of,
presumably solar nebula, N. The CI and CM meteorites have enrichments
of 2-5% in 15N / 14N compared
to the atmosphere, probably explicable by chemical-physical mass
fractionation effects. An acid insoluble residue from the Dhajala
ordinary chondrite has about 6% lower 15N
/ 14N (Murty, 1996). Although greater information
on the nature of the N components in chondrites is obviously needed,
it is equally clear (see above discussion of Ne) that the 15N
/ 14N in the terrestrial atmosphere is not
the appropriate reference point to interpret the observed isotopic
variations. A directly-measured solar wind 15N
/ 14N would serve this purpose.
6. Solar Wind Abundances from Lunar Samples
Lunar surface materials contain implanted solar wind, but there has been considerable modification by lunar surface processes, especially meteorite impact. The solar contributions can be easily recognized and measured, but the data have proved extremely difficult to interpret. Several papers (e.g., Podosek et al., 1971) have proposed that the isotopic compositions of Xe and Kr in lunar soils have been systematically modified from that of the solar wind. For example there is the exciting possibility that interstellar atoms may be implanted in lunar material (Geiss, 1972). It would be of great importance to recognize these modifications or additions. Somewhat ironically, the important general point is that independent knowledge of contemporary solar wind noble gas, nitrogen, and carbon elemental and isotopic abundances is required in order to interpret the lunar data. A sample return is required to give the isotopic abundances to the required accuracy as well as the elemental abundances of Kr and Xe.
Here the issues of elemental and isotopic abundances are closely intertwined, and we consider both. To obtain highly selected samples that have minimum impact modification of the solar wind record, the mineral ilmenite (FeTiO3) is separated from lunar soils. This approach acknowledges (concedes) that over 99% of lunar material is not suitable for obtaining solar wind composition. Particle track and single grain noble gas analyses of lunar soils have shown that every soil grain has resided on the lunar surface exposed to space regardless of the location from which the soil sample was obtained. Even with ilmenite separates there are complications: (a) effects of galactic cosmic ray spallation nuclear reactions are still present, and these are important for minor isotopes; (b) contributions of higher energy (SEP) solar ions (surprisingly) are important and these appear to have isotopic compositions different from the solar wind; (c) the elemental composition of the solar wind was different in the past; (d) diffusion losses, preferentially of lighter relative to heavier noble gases. Of course, "complications" (b) and (c) are of considerable science interest in their own right and are discussed separately in Secton 9 below. To aid in resolving the solar wind, spallation, and SEP contributions, depth-sensitive analysis technique are used because the undisturbed solar wind is implanted into the outer few hundred Angstroms, the SEP is at greater depths down to around a few microns, and the spallation is uniform over the whole grain. In practice the purest grain size separates are obtained with larger, roughly 100 micron, grain sizes. Two analytical approaches have been generally successful: (1) analyzing stepwise the gases released when an ilmenite mineral separate is exposed to increasingly stronger treatments with HF acid during "closed system etching" (e.g. Benkert et al., 1993). A gentle treatment preferentially extracts gases from grain surfaces. Progressively stronger treatments sample deeper into the grains, although there is no quantitative knowledge of the depths sampled. Closed system etching has the great advantage that extractions take place at room temperature, so there is no diffusive fractionation of elemental abundances during analysis, and elemental data from each temperature step can be unambiguously interpreted. (2) analyzing stepwise the gases released during exposure of ilmenite separates to low pressures of oxygen gas ("combustion") at progressively higher temperatures (up to 600-700 deg. C). At higher temperatures, gases are not extracted from ilmenite due to the formation of oxidized surface layers, but extractions can be continued by vacuum heating ("pyrolysis"). For both these techniques another complication is that 103-104 grains are utilized, each of which has its own history of surface exposure. A complementary technique (Nichols et al., 1994) involves (3) extracting nobles gases by total vaporization of single grains with a laser. Here, individual exposure histories are being measured, but all depth information is lost.
It is significant that 20Ne/22Ne ratios are observed in some extraction steps (techniques 1 and 2) or in a few grains (technique 3) that are close to the 13.7 observed in the Apollo foils. Using the data for other noble gases that are observed along with the 20Ne/22Ne = 13.7 samples, it is possible to extract solar wind noble gas elemental and isotopic abundances from lunar samples (e.g. Pepin, 1991). Note that this bootstrapping approach is only possible because there is independent knowledge of 20Ne/22Ne in the contemporary solar wind. The enormous amount of data selection that has accompanied this extraction is careful, skillful and deserves considerable respect; nevertheless, in itself, the sheer amount of sifting involved is sufficient to justify independent analysis of contemporary solar wind.
Beyond philosophical concerns, there are specific complications associated with solar wind noble gas abundances derived as described above.
(A) Two ilmenite samples which have been intensively studied give different answers for the elemental composition. Ilmenite from lunar soil sample 79035 was exposed to the solar wind roughly 109 yr. ago, whereas ilmenite from 71501 appears to have accumulated solar wind over the last roughly 108 yr. These differences can be explained by time variations in the first ionization potential or first ionization time (FIP/FIT) fractionations (Wieler and Baur, 1995) that are known in the contemporary solar wind (Documents G and H). Thus, it is also possible even the 71501 results are not applicable to the present-day solar wind. Even for a relatively "young" lunar soil such as 71501, most of the solar wind has accumulated 100 My ago and is not a measure of the contemporary solar wind. Independent analyses of contemporary solar wind is required to ecplore the possiblity of changes in the solar wind elemental composition on a 108 yr time scale. The noble gas isotopic compositions appear similar in 71501 and 79035.
A major advantage of the stepwise combustion approach (technique 2) is that N can be measured as well as the noble gases. However, even in the low temperature extraction steps for 71501, which is the optimum data for recent solar wind composition, the N/noble gas ratios are much higher than would be expected from solar abundance estimates (Kerridge et al., 1990). This discrepancy can also be seen in a simple comparison of N/36Ar in bulk soil samples ( e.g. Kerridge, 1993) where, although there is a spread in this ratio, it is always much higher than anticipated from solar abundance estimates. This may mean that (a) the observed N is not from solar wind or (b) there have been large losses of solar wind noble gases relative to N, presumably because N can be chemically bound, whereas the noble gases cannot. Option (a) is discussed further in the section on N isotopes below. Some support for option (b) comes from the relatively large amounts of SEP noble gases relative to those ascribed to the solar wind, which are roughly 100 times larger than estimated from contemporary flux measurements of solar particles at energies higher than those of the solar wind. Extensive loss of solar wind noble gases would be accompanied by significant amount s of elemental and isotopic fractionation, complicating the identification of the ilmenite abundances with the solar values.
(C) Although diffusion loss is not the dominant cause of noble gas isotopic and elemental variations in ilmenite separates (e.g. Becker, 1991; Nichols et al., 1994), there has almost certainly been some loss of He and Ne, possibly Ar. Using only data with 20Ne/22Ne close to 13.7 does not insure that there has been no elemental fractionation due to diffusion loss. Single grain analyses show that 84Kr/132Xe is relatively uniform in ilmenite grains from 71501; moreover, the observed ratio is in good agreement with the average 84Kr/132Xe measured by closed system etching for the same sample. It is very unlikely that this elemental ratio has been affected by diffusion loss; however, there are significant contributions from SEP Xe and Kr to these data and it is not at all certain that the measured 84Kr/132Xe can be ascribed to the solar wind.
Independent knowledge of contemporary solar wind noble gas elemental
and isotopic compositions provides a needed check on whether all
the assumptions made in coaxing solar wind compositional data
from lunar soils are valid.
7. Noble Gas, Nitrogen and C Isotopes; Implications for Planetary Atmospheres.
Here, the issues with isotopic and elemental abundances are again closely intertwined, and we consider both. The origin of the large(roughly 40%) difference in 20Ne/22Ne between solar wind and the terrestrial atmosphere is not known. Many workers think it is the consequence of atmospheric escape. This may leave the meteorite trapped gas compositions unexplained, although Pepin (1991) proposes that the Q noble gases in carbonaceous chondrites have also been fractionated during loss of transient atmospheres in asteroidal-sized meteorite parent bodies. There is no question that knowledge of the other solar wind noble gas elemental and isotopic ratios to compare with Ne is critical to the solution of this major problem.
Research in planetary atmospheres might seem disconnected from solar wind composition, but the two are in fact closely linked. Volatile carbon compounds, nitrogen, and the chemically inert noble gases are the tracers of choice for tracking physical processes. Knowledge of the initial elemental and isotopic abundances of volatile species is imperative for identifying evolutionary mechanisms that might have acted, whether the modeling is forward, from presumed compositions in primordial atmospheres, or backward from known compositions in contemporary atmospheres. For many models, solar wind data directly provide the initial compositions assumed to have been present in unevolved primordial atmospheres. It is worth emphasizing that this is a model assumption. If the atmospheres of the Earth and the other terrestrial planets formed by outgassing of carbonaceous chondrites, then these would provide the appropriate initial isotopic compositions if different from solar. Information that pertains directly to isotopic compositions in the solar nebula environment of planetary accretion is only obtainable from the solar wind. However, data on Ne isotopic composition from terrestrial mantle samples support the assumption of an initial solar composition for the earth's atmosphere prior to loss. Data of adequate precision for all volatiles are only available from returned samples.
Although errors are relatively large at present (at least 10%, 2 sigma), the solar wind Mg isotope data from WIND (Bochsler et al., 1996) show no difference from the terrestrial value. Comparing this with the large Ne isotopic differences implies that large (40%) isotopic differences between solar and terrestrial matter will only be found in volatile elements and gives support to the general interpretation of Ne isotopic fractionation due to atmospheric loss processes.
As all atmospheric evolution models have adjustable parameters, they cannot be quantitatively tested if only the 20Ne/22Ne ratio is known. However, if all solar noble gas elemental and isotopic ratios are known from a solar wind sample return, there are then ample data to enable quantitative testing of models. If the difference in solar wind and terrestrial atmosphere 20Ne/22Ne reflects hydrodynamic escape of gases from our atmosphere, then the solar wind 21Ne/22Ne is predicted to be higher than in our atmosphere, readily testable with sample return data. The amount of enrichment is model-dependent, but if hydrodynamic escape is not important, there might not be any enrichment. If sample return data confirms the qualitative predictions of the hydrodynamic escape models, then it would be possible to obtain improved quantitative knowledge of the contributions that nuclear reactions in the Earth make to atmospheric 21Ne.
Precise measurement of the solar wind 129Xe abundance provides
a needed initial value for models of the early evolution of the
terrestrial planets based on 129I radioactive decay.
The possibility
that solar, planetary atmosphere isotopic differences reflect
general nebular heterogeneity rather than atmosphere evolution
can be simply checked by the isotopic comparisons with non-volatile
elements. Even at the present level of precision the solar wind
Mg isotopic data discussed above shows that any solar-terrestrial
isotopic variations for Mg are much less than those for Ne supporting
the escape interpretation for the differences in
20Ne/22Ne.
Perhaps the major unsolved mystery from Apollo is a roughly 20%
increase in the 15N/14N ratio with time for lunar
soil samples
exposed to the solar wind. This trend might represent: (1) a
systematic secular change in the solar wind N isotopic composition
(e.g. Kerridge, 1989), (2) the presence of nonsolar sources of
lunar surface N (lunar interior, Earth's early atmosphere, etc.)
early in the history of the Moon (Bochsler, 1994) or (3) dominance
of higher energy solar particles with low 15N/14N in
the older
lunar samples (Bochsler and Kallenbach, 1994). All models have
important implications; none is completely satisfactory. For
example problems with model (1) are there is no known solar process
that can account for a 20% change in 15N/14N, and
that the N/Xe
ratio in all lunar soil samples is far higher than expected for
the solar wind. But, on the other hand there is a good correlation
of N and 36Ar concentrations in lunar soils which is hard to explain
by models (2) and (3) because the 36Ar is universally accepted
as solar wind derived. All models predict a solar wind
15N/14N
ratio about 20% higher than the terrestrial atmosphere. Kim et al. (1995) propose
that there are changes in the solar wind 15N/14N on a million
year time scale and that the contemporary solar wind 15N /14N
ratio would be only about 2% higher than the terrestrial atmosphere. Thus, some models
will fail for sure, and it
is possible that solar wind sample data would falsify
all models, a common occurrence in planetary missions. Terrestrial
atmospheric hydrodynamic escape models are only compatible with
model (1). Model (3) is
directly tested by measurements of higher energy ions
(Objective
10). If the contemporary solar wind is found to have
15N/14N distinctly above the highest lunar ratio,
this would support
(1) because the youngest lunar samples still correspond to roughly
100 my old solar wind, and alternatives (2) and (3) put the source
of the variability at much older times. The possibility that
mechanism (1) represents solar surface nuclear processes will
be tested with data from Objectives
11, 12, and 18. If
model (1) should prevail, it would mean that solar evolutionary
processes can change isotopic compositions from those in the solar
nebula, complicating one of our major objectives to varying degrees
depending on details (which elements, etc.). However, confirmation
of (1) could fairly be labeled a conceptual breakthrough, because
there is no known solar process which could cause the required
N isotopic evolution. This is why the lunar N isotopic data have
remained a mystery.
9. Higher Energy Solar Particles
The impetus to pursue this objective comes from significant progress in untangling the solar particle data in lunar samples. The literature in this area is very complicated, not for the faint of heart; however, a major conclusion has emerged: There are noble gases, presumably solar in origin, which reside at greater depth in the ilmenite grains than can be accounted for by solar wind implantation. This deeper component ("SEP") of higher energy ions than the solar wind has a different noble gas isotopic composition (Benkert et al., 1993; Wieler and Baur, 1993; Wieler et al., 1996) . For the case of Ne, where the contemporary solar wind 20Ne / 22Ne is known, the difference is well established with the SEP 20Ne / 22Ne being 11.2 ± 0.2 compared to the 13.7 from the Apollo foils. The most interpretable experiments come from the Zurich "closed system etching" approach in which noble gases are released and analyzed based on a progressively stronger acid attack on ilmenite mineral separates. In most experiments initial, relatively gentle, etching releases Ne with a 20Ne / 22Ne ratio approaching 13.7. Later, stronger etch steps releases Ne with the 20Ne / 22Ne approaching 11.2.
In addition to the distinct neon "SEP" isotopic ratios, the existence of distinct high energy He, Ar, and Kr, if not also Xe, has been demonstrated (Benkert et al., 1993; Wieler and Baur, 1994; Pepin et al., 1995). For all five noble gases, the high energy particle isotopic composition is fractionated relative to solar wind by the square of the mass ratios. However, elemental abundance ratios of the high energy component are roughly the same as the solar wind.
As discussed above, there is also now convincing evidence that solar particle elemental abundances, at least He/Ne and Kr/Xe, were somewhat different in the past, although it is not clear that the solar wind, as opposed to higher energy particles, shows the compositional time dependence. There are potentially other inputs to the measured lunar noble gases and N (terrestrial atmosphere/magnetotail, early outgassed lunar volatiles, meteoritic or cometary volatiles, interstellar atoms). From this point of view, knowing present-day solar wind and superthermal elemental and isotopic compositions is required to be able recognize the presence of these other components in the Apollo and future lunar samples.
Given the great analytical sensitivity for noble gases, it will
be possible to measure the elemental and isotopic abundances at
deeper implantation depths than the solar wind ions (<10 keV/amu),
corresponding to the 10-100 and 100-1000 keV/amu ranges.(Proposal
Table E1-2,
Objective 10). The fluxes of these higher energy particles
are much lower, but data on C, N, Mg, and Fe elements and isotopes
may still be possible. The changes in isotopic and elemental
noble gas ratios inferred from the lunar data can be directly
compared with those from a returned solar wind sample.
