sign in sign up
<<
Home , Meteoritics

Chambers: Meteoritic Diversity and Formation 487

Meteoritic Diversity and Planetesimal Formation

John Chambers Carnegie Institution of Washington

Meteorite parent bodies vary widely in composition and structure even though they all formed in the same protoplanetary nebula over a limited range of time and space. Nebula material under- went substantial mixing at an early stage leading to approximate isotopic homogeneity in meteor- ites. Turbulence and gas drag transported gas, dust, and solid bodies smaller than 1 km over large radial distances. After Jupiter formed, bodies up to 100 km in size moving on eccentric orbits experienced rapid inward drifting due to gas drag. The parent bodies of most primitive meteor- ites required at least 2–3 m.y. to form. This provided time for diverse components (various types of , refractory inclusions, and metal grains) to form and evolve due to nebular heating events and interactions with nebular gas. Parent bodies acquired different mix- tures of each component depending on when and where they accreted. Much information about where parent bodies formed has been erased by the unknown process that removed the vast majority of the primordial solid material from the belt. However, it appears that time of formation was more important than location in determining the composition of parent bodies.

1. INTRODUCTION close analogs of the Sun’s protoplanetary nebula. Circum- stellar disks are heated by radiation from the central star, and their parent bodies are a highly diverse and also by the release of gravitational energy as material group of objects. Different meteorite classes have a range of moves inward through the disk and accretes onto the star. It elemental compositions and O-isotope abundances. Mete- is unclear what mechanism causes material to flow through orite parent bodies have undergone varying degrees of ther- the disk, but it is generally assumed to involve turbulence mal and aqueous alteration, and experienced different colli- (density waves and vortices may also play a role). In very sional histories. Some meteorites formed under extremely young disks, temperatures at the midplane may be high reducing conditions while others are highly oxidized. Me- enough to vaporize common silicates as far as 3 AU from teorite components range from ones that are dominated by the star (Boss, 1993; Woolum and Cassen, 1999). However, highly refractory minerals to ones that are rich in this high-temperature phase is short-lived. Disks older than and have never experienced high temperatures. Short-lived a few hundred thousand years probably have midplane tem- isotopes were clearly present when some meteorite com- peratures of 200–750 K at 1 AU from the star and lower ponents formed but not others. still at 2.5 AU (Woolum and Cassen, 1999). Temperatures Despite this broad diversity, meteorite parent bodies all may even be low enough for water ice to condense through- formed in the Sun’s protoplanetary nebula over a limited out this region (Chiang and Goldreich, 1999). range of time and space. Astronomical observations, theo- If the Sun’s protoplanetary nebula was hot and turbu- retical studies, and some characteristics of meteorites them- lent early in its history, it is plausible that material would selves suggest these parent bodies formed in a nebula that have been mixed and homogenized. There is some evidence was turbulent and characterized by widespread mixing of that this happened. The Mg-isotope ratios observed on Earth, material. Gas, fine dust, millimeter-sized particles, boulders, the Moon, Mars, and a group of meteorites sug- kilometer-sized , and even large were gest they formed in a nebula where Mg isotopes had been probably transported over radial distances of an astronomi- homogenized (Norman et al., 2004). Similarly, the mix of cal unit (AU) or more. Thus, it is something of a paradox that Fe isotopes in several groups of chondritic (primitive) and meteorite parent bodies came to be so different from one achondritic (melted) meteorites can be derived from a uni- another. In this chapter, I will discuss the evidence for wide- form reservoir by mass-dependent fractionation (Zhu et al., spread mixing due to a number of physical processes, and 2001), i.e., purely as a result of physical processes that de- try to reconcile this with the existence of the diverse group pend on mass differences between isotopes. of meteorites we see today. The situation is more complicated for O, an element that was abundant in both solid and gaseous phases in the solar 2. NEBULAR MIXING AND nebula. Oxygen-isotope ratios vary significantly from one HETEROGENEITY meteorite group to another, and these differences cannot be explained by mass-dependent fractionation alone. Some of Many young stars lie at the center of disks containing the variation seen in carbonaceous can be attrib- gas and dust, surrounded by more tenuous envelopes of ma- uted to aqueous alteration in the parent bodies of these terial (Beckwith et al., 1990). These disks are thought to be meteorites (Clayton, 2004). However, several O-isotopic

487 488 Meteorites and the Early Solar System II reservoirs are still required to explain differences between tion of volatile elements implies that solids were preferen- the groups. Calcium-aluminium-rich inclusions tially retained in the nebula while gas accreted onto the Sun. (CAIs) (meteoritic components that probably formed at a Recent theoretical models suggest that centimeter- to meter- very early stage) lie along a single mixing line in O-isotope sized solid objects were actually highly mobile in the nebula phase space with a common 16O-rich end member (Clayton, (Stepinski and Valageas, 1997). It appears that solid parti- 2004). This suggests that the nebula was initially uniformly cles could have been lost at a faster rate than the gas, unless rich in 16O, with the variations seen in primitive meteorites solids rapidly aggregated to form bodies that were on the arising at a later stage. order of 1 km in size. A small fraction of grains in primitive meteorites con- Primitive meteorites are mostly made up of : tain highly unusual isotopic ratios in many elements (Zinner, rounded, millimeter-sized particles that are believed to have 1998; Nittler, 2003). These isotopic mixtures cannot be formed by flash heating in the nebula. Different types of explained in terms of mass fractionation. Apparently, these chondrule exhibit different elemental depletion patterns. “presolar” grains formed outside the solar nebula and Type-I (FeO-poor) chondrules tend to be substantially de- avoided homogenization with the other material in the so- pleted in moderately volatile elements, while type-II (FeO- lar system (Nittler, 2003). rich) chondrules are not (Zanda, 2004). If elemental abun- Many CAIs once contained short-lived isotopes such as dance patterns are the result of condensation in a hot neb- 41Ca, 26Al, 10Be, and 53Mn. The inferred initial amounts of ula, we would infer that the precursors of type-I chondrules 41Ca and 26Al are often correlated (Sahijpal et al., 1998), formed closer to the Sun than precursors of the type-II chon- suggesting they came from the same source, possibly a drules. However, this conclusion is hard to reconcile with nearby supernova. The 10Be was distributed in a different the traditional viewpoint that meteorites, manner (Marhas et al., 2002), and may have been produced which contain many type-II chondrules, come from the in- in the solar nebula when material lying close to the Sun was ner asteroid belt, while carbonaceous chondrites, composed irradiated by solar particles (McKeegan et al., 2000). [See largely of type-I chondrules, are thought to come from the Desch et al. (2004) for an alternative source for 10Be.] outer asteroid belt. Calcium-aluminum-rich inclusions typically formed with An alternative interpretation is that the elemental deple- 26Al/27Al ratios close to a “canonical” value of 4.5 × 10–5 tions seen in chondritic meteorites were caused by partial (MacPherson et al., 1995; Huss et al., 2001). This suggests evaporation of chondrules during the events that melted that these particles formed in a region of the nebula that them. This idea remains controversial because chondrules was isotopically uniform in Al. A few CAIs (the “FUN” show few signs of having undergone mass-dependent frac- kind), and some inclusions in CH chondrites, contained little tionation in elements such as K (Humayun and Clayton, or no 26Al when they formed (Krot et al., 2002). FUN inclu- 1995). In general, one would expect to see some mass frac- sions contain anomalous amounts of 50Ti and other super- tionation during partial evaporation. However, this need not nova-generated isotopes that were not produced via radio- be the case if the chondrule-heating events took place in active decay (Clayton et al., 1988). It is possible that FUN regions with high partial pressures of the rock-forming ele- inclusions predate other CAIs and formed before the nebula ments (Alexander et al., 2000; Alexander, 2004). Thus, we was homogenized. can understand the lack of strong mass-dependent fractiona- The decay product of 53Mn appears to be present in solar tion if chondrules formed in regions where the solid-to-gas system bodies in amounts that correlate with distance from ratio was higher than the nebula as a whole. the Sun (Lugmair and Shukolyukov, 2001). This may indi- In most chondrites, the refractory lithophile (rock-loving) cate that 53Mn was heterogeneously distributed in the early elements are present in proportions similar to those seen in solar nebula. Computer simulations of the injection of su- the Sun (Krot et al., 2004). This suggests these elements pernova-generated material into a protoplanetary nebula were distributed quite evenly in the nebula. However, there predict an uneven distribution of material, at least initially is some variation, particularly in the Mg/Si and Al/Si ratios (Vanhala and Boss, 2002), although it is strange that 26Al of chondritic meteorites (Drake and Righter, 2002). Differ- was later homogenized while 53Mn was not. This may indi- ences in the Al/Si ratio may be explained by the fact that cate that 53Mn, like 10Be, came from a different source than some chondrites contain more Al-rich CAIs than others, al- 26Al. though the correlation is not perfect (Krot et al., 2004). The Many primitive meteorites are depleted in moderately different Mg/Si ratios are reflected in the different amounts volatile elements, compared to the Sun, in a way that corre- of olivine and pyroxene seen in various chondrules. These lates with each element’s volatility (Palme et al., 1988; Alex- differences are hard to explain unless chemical reactions ander et al., 2001). This correlation may be due to incom- took place between chondrules and nebular gas (Scott and plete condensation of the more volatile elements during the Krot, 2004). If we abandon the idea that chondrules were early hot phase of the nebula (Cassen, 1996). If this inter- closed systems, then a chondrule’s composition could have pretation is correct, the various components in meteorites changed each time it was heated, due to partial evaporation inherited their elemental depletion patterns from precursors and acquisition of new material (Alexander, 1996; Sears et that formed at an early stage, and subsequent events did not al., 1996; Scott and Krot, 2004). The relative volatility of alter these patterns. In this model, the progressive deple- Mg and Si would have fluctuated according to local neb- Chambers: Meteoritic Diversity and Planetesimal Formation 489

ula conditions such as the dust/gas ratio (Alexander, 1996, Since chondrules are a major component of primitive mete- 2004). Thus, differences in the Mg/Si ratio of primitive me- orites, there must have been many heating events. The pres- teorites may be explained in terms of the different histories ence of relict fragments of type-I chondrules in type-II chon- of their components, provided that these components re- drules and vice versa (Jones, 1996) also argues for multiple mained in the nebula for an extended period of time. heating episodes. In a turbulent nebula, the timescale required to transport Many meteorites contain compound chondrules formed gas throughout the inner solar system would have been quite by collisions between chondrules while they were still hot. short. Theoretical models predict that water could have been These were unlikely to have formed in large numbers un- removed from the entire inner nebula in ~105 yr by “cold less the concentration of solid material was enhanced in the trapping”: a combination of turbulent diffusion of nebula chondrule-forming regions. Models for the chemical evolu- gas and condensation of water ice near the “snow line” (Ste- tion of chondrules during heating events also infer that chon- venson and Lunine, 1988). Nebula models suggest the snow drules formed in regions with dust-to-gas ratios much higher line was located somewhere between 4 and 7 AU from the than the nebula as a whole (Alexander, 2004). Compound Sun during the first few hundred thousand years, when disk chondrules often have nonporphyritic (i.e., fully melted) mass accretion rates were high (Boss, 1996). Later, when textures, suggesting they achieved very high peak temper- the nebula cooled, the snow line moved closer to the Sun. atures. This can be understood if chondrules were melted Many primitive asteroids apparently experienced aqueous by passage through a shock wave in the nebula (Desch and alteration as a result of reactions between water ice and dry Connolly, 2002). Shock waves are not the only possible rock. This implies the asteroid belt was never completely mechanism for producing chondrules, but such an origin dry. However, the abundances of highly volatile materials does appear to explain many of their properties (Chiang, could have been distinctly nonsolar for much of the time 2002). The origin of these shock waves remains unclear, that meteorite parent bodies were forming, due to cold trap- although computer models of disks thought to be similar ping and other processes. Changes in the composition of to the Sun’s protoplanetary nebula show that shock waves the nebula would have led to differences in the composi- can occur (Boss, 2000). tions of parent bodies that formed at different times. Chondrules typically have fine-grained rims of volatile- rich dust, similar in composition to the matrix material that 3. FORMATION OF METEORITE fills the gaps between the chondrules. The rim and matrix COMPONENTS material contains interstellar organic compounds and cir- cumstellar dust grains that apparently have not been heated Calcium-aluminum-rich inclusions from a variety of me- above ~700 K since their formation (Ott, 1993). This means teorite classes have broadly similar primary mineralogies, the chondrules encountered dusty nebular regions that had similar O-isotope abundances, and similar inferred initial escaped high-temperature events before the chondrules were 26Al/27Al ratios (MacPherson, 2004). These properties sug- incorporated into larger bodies. It has been estimated that gest that many CAIs formed under similar nebula condi- chondrules required 100–1000 yr to accrete their rims in a tions over an interval of no more than a few hundred thou- dusty environment (Cuzzi, 2004). Iron and Si are distrib- sand years. It is currently unclear where CAIs formed. The uted in a complementary way in chondrules and matrix of highly refractory nature of CAIs suggests they formed in a CM chondrites (Wood, 1985). Some elements also have region where the ambient temperature was high, otherwise complementary abundances in the chondrules and matrix the volatile elements could have condensed onto them. The of CV chondrites (Palme et al., 1992). The most obvious former presence of 10Be in CAIs suggests they formed close way to interpret these correlations is if chondrule precursors to the Sun, although this interpretation remains a matter of and matrix material coexisted in a low-temperature region debate (e.g., Desch et al., 2004). of the nebula. A subset of this material was heated to form Some ferromagnesian and Al-rich chondrules in carbon- chondrules while the remainder escaped heating and was aceous and ordinary chondrites contained 26Al when they accreted subsequently. formed (Mostefaoui et al., 2002; Kurahashi et al., 2004). Amoeboid olivine inclusions (AOAs) are minor compo- However, the inferred 26Al/27Al ratios were substantially nents seen in many chondrites, and these have compositions lower than those of CAIs. If this difference is due to the intermediate between CAIs and ferromagnesian chondrules. decay of 26Al from an initially uniform reservoir, it means The initial 26Al abundances of AOAs suggest they formed most chondrules formed about 1–3 m.y. after CAIs. This before most chondrules and roughly 0.1–0.5 m.y. after CAIs interpretation is supported by high-precision measurements (Itoh et al., 2002). Amoeboid olivine inclusions have 16O- using the U-Pb isotope system to estimate the age of chon- rich compositions similar to CAIs (Itoh et al., 2002), sug- drules in a CR chondrite (Amelin et al., 2002). The textures gesting they formed in the same nebular environment, and of chondrules suggest they were heated to peak tempera- indeed some AOAs are observed to contain relict CAIs tures of around 1500–1850 K (Hewins and Connolly, 1996), (MacPherson, 2004). and then cooled in a matter of hours. The short cooling From these observations, we can construct a plausible timescale strongly suggests the heated regions were much sequence of events for the formation of meteorite compo- smaller than the nebula itself (Desch and Connolly, 2002). nents. FUN inclusions formed first, the other CAIs followed 490 Meteorites and the Early Solar System II soon afterward, AOAs formed a little later, and chondrules formed last of all, about 1–3 m.y. after CAIs. Chondrules, and possibly CAIs, were formed in multiple events, with each generation having a somewhat different composition. All these components underwent some spatial mixing be- fore incorporation into parent bodies. There were probably some exceptions to this sequence of events. At least one CAI appears to contain a relict chon- drule (Itoh and Yurimoto, 2003), which suggests this chon- drule formed before the CAI. In addition, CH chondrites and some CB chondrites do not fit comfortably into the chronological scheme outlined above. These meteorites contain zoned metal grains that appear to have condensed directly from the hot early nebula, cooled rapidly, and sub- sequently avoided reheating (Meibom et al., 2000). Most chondrules in CH meteorites are nonporphyritic, highly depleted in volatiles, and lack fine-grained dust rims (Krot et al., 2002), suggesting they also formed in a very hot Fig. 1. Lifetime of spherical particles against drift due to gas nebula. Calcium-aluminum-rich inclusions and chondrules drag, at 2.5 AU in the solar nebula, as a function of particle size. in CH chondrites seem to have a close relationship, and they Gas density is 10–9 g/cm3 at 1 AU, varying as 1/r2. Adapted from may have shared a common origin (Krot et al., 2002). It Weidenschilling (1977). seems likely that these meteorites and their components formed at a time and/or place quite different from other primitive meteorites. CO were probably enhanced in the inner nebula rather than depleted by cold trapping. Later, once planetesimals formed 4. RADIAL REDISTRIBUTION OF in large numbers near the snow line, the inner nebula became SMALL SOLIDS depleted in volatile materials (Cuzzi and Weidenschilling, 2006). Thus, we would expect meteorite parent bodies that In most models for protoplanetary disks, gas pressure formed at different times to have different chemistries and decreases with radial distance from the central star. As a oxidation states. result, the gas is partially supported against the star’s gravi- In the absence of turbulence, a 1-mm-diameter CAI tational pull and revolves around the star more slowly than would have drifted from the asteroid belt to a point where a solid body moving on a circular orbit. Solid bodies in the it would vaporize in around ~105 yr. This means that es- Sun’s protoplanetary nebula would have experienced “gas sentially all CAIs would have been lost by the time the drag” that gradually removed angular momentum from their majority of chondrules formed 1–3 m.y. later. The fact that orbit so that they spiralled inward toward the Sun (Weiden- chondrules and CAIs are seen in the same meteorites sug- schilling, 1977). Small particles drifted inward at a rate gests the nebula was turbulent for much of the first 3 m.y. determined by their terminal velocity, which increased with in order to preserve some CAIs for this length of time. particle size. Large particles drifted at a rate that depended Millimeter-sized particles would have coupled strongly to on the ratio of their surface area to volume, which decreased turbulent eddies in the nebula at the level of turbulence with particle size. Drift rates were highest for meter-sized envisioned in most nebular models (Cuzzi et al., 2001). particles (Weidenschilling, 1977). These particles would Theoretical models show that although most CAIs would have drifted 1 AU every few hundred years in a nonturbu- continue to drift inward in a turbulent nebula, some would lent nebula (see Fig. 1). also diffuse outward, thereby extending their lifetime by If a significant fraction of solid material was present in several million years (Cuzzi et al., 2003). These models meter-sized bodies at any one time, gas drag would have suggest that large type-B CAIs, seen in CV chondrites, been the dominant process causing radial redistribution of could only be retained if they formed in a region close to the material in the nebula (Stepinski and Valageas, 1997). Drift- asteroid belt, or were deposited there by an x-wind (see be- ing particles may not have fallen all the way into the Sun. low). Smaller type-A CAIs, seen in most chondrite classes, Instead, small solids would have evaporated once they would have diffused into the asteroid belt and survived for reached a region that was hot enough (Supulver and Lin, several million years even if they initially formed close to 2000). Ices evaporated near the snow line, while silicate- the Sun (Cuzzi et al., 2003). rich bodies evaporated closer to the Sun. It is plausible that It is conceivable that some CAIs were incorporated in solid material drifted inward faster than turbulent diffusion an early generation of planetesimals shortly after the CAIs could act to redistribute the gas, so that the vapor phase of formed. These planetesimals could have survived for sev- each condensible material became enhanced near its con- eral million years, safe from the effects of gas drag. Later densation distance (Cuzzi and Weidenschilling, 2006). At catastrophic collisions could have returned CAIs to the neb- early times, therefore, volatile materials such as water and ula, allowing them to be accreted along with newly formed Chambers: Meteoritic Diversity and Planetesimal Formation 491

chondrules to form new planetesimals (Lugmair and Shu- size distribution. The mean size differs from group to group kolyukov, 2001). Currently, there is no evidence to support but is generally in the range 0.1–1 mm (Scott and Krot, this hypothesis. We lack examples of primitive meteorites 2004). There are two ways to interpret these observations. composed primarily of CAIs, while the known Either compact millimeter-sized particles had to be abundant appear to have had chondritic precursors (Meibom and in the nebula in order to trigger the formation of primitive Clark, 1999) rather than precursors made of CAI-like ma- planetesimals, or compact millimeter-sized particles hap- terial. If CAI-rich planetesimals did form, they are now rare pened to be abundant in the nebula when conditions became or absent in the asteroid belt. Instead, it appears that most right for these planetesimals to form. In the former case, CAIs remained in the nebula for several million years be- chondrules were the perpetrators, while in the latter case they fore they and the chondrules coalesced to form planetesi- were merely bystanders caught up in events around them. mals. Some CAIs show signs of multiple episodes of modi- There is growing evidence from radioactive isotope dat- fication in a nebular setting (Hsu et al., 2000; MacPherson ing that chondrule formation extended over several million and Davis, 1993), which lends weight to this interpretation. years (see Fig. 2), and that chondrules in a given meteorite Nebular gas pressure may have varied nonmonotonically have a wide range of ages (Amelin et al., 2002; Mostefaoui with distance from the Sun. In this case, drag effects would et al., 2002; Kurahashi et al., 2004). Many primitive mete- have collected small solids at each local pressure maximum orite parent bodies apparently required 2–3 m.y. to form (Haghighipour and Boss, 2003), and these locations would even though chondrules had existed in the nebula for at least become preferred sites for planetesimal formation. Even in 1 m.y. at that point. Compact CAIs, with mechanical and a nebula with a smooth pressure gradient, drift rates due aerodynamic properties similar to chondrules, had been to gas drag would have varied with distance from the Sun. present for even longer. This suggests that the formation of Recent studies of individual and collective drift rates for millimeter-sized particles did not trigger planetesimal for- millimeter-sized particles find that the number of particles mation. Unless CAIs and early generations of chondrules in the inner nebula probably increased over time as mate- remained uniformly spaced in the nebula, it seems unlikely rial drifted inward from the outer nebula (Youdin and Shu, that planetesimal formation had to wait for the global num- 2002; Youdin and Chiang, 2004) [but see comments by ber density of millimeter-sized particles to reach a certain Weidenschilling (2003)]. This radial concentration may have level. CI chondrites, which are composed almost entirely of been a necessary prerequisite for the formation of larger fine-grained matrix material, formed at a time and place bodies by gravitational instability (see next section). where chondrules were essentially absent, which also sug- Solid particles may have been redistributed in another gests that chondrule production and planetesimal formation way. Young, actively accreting stars typically have colli- were separate processes. mated jets of gas that arise near the inner edge of the disk The formation times of achondritic asteroids are diffi- and travel away from the star perpendicular to the disk cult to pin down, although the iron IIIAB prob- plane. This “x-wind” may be strong enough to launch milli- ably formed in about ~2 m.y. (Sugiura and Hoshino, 2003). meter-sized particles on trajectories that carry them outward for several astronomical units before they fall back into the disk (Shu et al., 1996). It is possible that a significant frac- tion of the material contained in meteorites drifted close to the Sun and was transported back into the asteroid belt via an x-wind, possibly several times, during the early active phase of the nebula. However, the compositions of most chondrules makes it hard to understand how they could have formed in an x-wind (Hutchison, 2002). In addition, it is unclear why the CI class of chondrites failed to incorpo- rate chondrules if these particles were distributed through- out the nebula by an x-wind. It seems likely that the x-wind mechanism was not effective during the time when most chondrules formed. Equally, recycling by an x-wind was presumably not the only process that acted to preserve CAIs until chondrules and parent bodies formed.

5. PLANETESIMALS AND Fig. 2. Formation times of refractory inclusions and chondrules PARENT BODIES in various primitive meteorites based on the U-Pb (CR chondrule and CAIs) and Al-Mg (all except CR chondrule) isotope systems. Data taken from Amelin et al. (2002); Guan et al. (2002); Chondrules are a major component in almost all primi- Hutcheon et al. (2000); Itoh et al. (2002); Mostefaoui et al. (2002); tive meteorites. Carbonaceous chondrites commonly contain and Kurahashi et al. (2004). Horizontal lines show age error bars. 50% chondrules by volume, while ordinary and enstatite For the CO3 and LL3 chondrites, the errors are larger than the chondrites contain up to 80% (Scott and Krot, 2004). Chon- separation of neighboring data points but smaller than the total drules within each chondrite group typically have a narrow spread in chondrule ages. 492 Meteorites and the Early Solar System II

Models for the thermal evolution of Vesta suggest it took until the present day. Numerical accretion models suggest roughly 2.9 m.y. to accrete (Ghosh and McSween, 1998). this delay was relatively short, however. Once planetesimals Thus, at least some differentiated bodies apparently formed appeared in large numbers, only 105–106 yr were required long after chondrules first appeared. This reinforces the to form bodies as large as Ceres (Wetherill and Stewart, notion that planetesimal formation did not coincide with the 1993). In general, planetesimals would have undergone less first appearance of chondrules. It is also possible that many radial migration than smaller objects, so the differences achondritic asteroids formed before chondrules, in which between meteorite parent bodies probably reflect differences case the two processes were entirely separate. in the populations of planetesimals that formed them. How- Several models for the origin of planetesimals are be- ever, there is some evidence for heterogeneity within par- ing pursued, although none of these models is fully devel- ent bodies. For example, the abundance of CAIs in different oped at present. Calculations suggest that pairwise collisions meteorites from the same ordinary chondrite or enstatite between small particles could have formed meter-sized bod- chondrite group can vary by an order of magnitude (Kimura ies on timescales of ~104 yr in the absence of turbulence, et al., 2000, 2002). These differences presumably arose provided that particle-sticking mechanisms were effective because the planetesimals that aggregated to form a given (Weidenschilling, 1980). Aggregation may have continued parent body were not all the same. Similarly, the bimodal in this way until kilometer-sized planetesimals were pro- olivine abundance seen in L chondrites (Wood, 1996) may duced, at which point gravity would aid further accretion. reflect differences among the parent body’s component plan- However, as particles grew in size, collisions became more etesimals. likely to cause disruption rather than accretion, especially if The different O-isotopic compositions of chondrite the nebula was turbulent (Cuzzi and Weidenschilling, 2006). groups have been interpreted to mean that each meteorite It is possible that pairwise accretion stalled at sizes much parent body formed in a different region of the nebula, and less than 1 km. that mixing between these regions was minimal during If the level of turbulence was very low, solid material parent-body formation (e.g., Wood, 1996). The decreasing would have collected near the nebular midplane, and this amount of aqueous alteration seen in carbonaceous, ordi- region could have become gravitationally unstable, forming nary, and enstatite chondrites, respectively, is usually attrib- planetesimals directly (Goldreich and Ward, 1973). How- uted to formation at distances progressively closer to the ever, even with low levels of turbulence, gravitional insta- Sun, where nebula temperatures were higher and water ice bility requires substantial enhancement in the local dust-to- was less likely to condense. This notion of narrow, isolated gas ratio (Sekiya, 1998). This might be achieved either by accretion zones receives some support from the preserva- radial drift and concentration of particles as described in the tion of different 52Cr/53Cr ratios in bulk samples from differ- previous section, or if gas is lost from the nebula at a greater ent meteorites (Lugmair and Shukolyukov, 1998), and also rate than solids (Youdin and Shu, 2002). the observed radial zoning of the asteroid belt with respect A third model for planetesimal formation invokes “tur- to asteroid spectral class, although the different spectral bulent concentration” of small solids in stagnant regions in a classes overlap substantially (Gradie and Tedesco, 1982). turbulent nebula. This could have produced very high con- However, this picture is almost certainly too simplistic. centrations of particles with particular aerodynamic prop- The preservation of CAIs apparently requires that the nebula erties, up to a factor of 105 higher than normal (Cuzzi et al., remained turbulent for several million years. As a result, 2001). High particle concentrations presumably would have different meteorite components should have become thor- aided further growth by one of the mechanisms described oughly mixed together. Most types of chondrule and refrac- above. Encouragingly, models for turbulent concentration tory inclusion are seen in most meteorite groups, but the predict chondrule size distributions similar to those seen in proportions are different in each case. Some of these dif- primitive meteorites (Cuzzi et al., 2001). ferences may be attributed to the aerodynamic processes These models for planetesimal formation share a com- that caused the widespread size sorting of small solids seen mon property: In each case the growth of large bodies in chondrites. This may explain why large, type-B CAIs are depends on local conditions in the nebula. In particular, seen only in CV chondrites. This group of chondrites also the dust-to-gas ratio and the level of turbulence seem to be tends to contain large chondrules and large type-A CAIs key factors. The long delay between the formation of CAIs (Scott and Krot, 2004). If all type-B CAIs were large, we and the appearance of many parent bodies suggests that a might convince ourselves that they would tend to appear threshhold effect was at work. It is plausible that planetesi- only in those meteorites with large components. mal formation was delayed until turbulence fell below a cer- More perplexing are the different numbers of CAIs seen tain level, or until the local solid-to-gas ratio was enhanced in different meteorite groups. Several carbonaceous chon- to a certain point (Youdin and Shu, 2002). Once suitable drite groups contain 4–13% refractory inclusions by volume conditions arose, the formation of planetesimals could have (Scott et al., 1996; Scott and Krot, 2004). In ordinary and proceeded rapidly. enstatite chondrites, CAIs are present in only trace amounts, There was presumably a time lag between the appear- usually ~0.1% (Scott and Krot, 2004). This is hard to ex- ance of kilometer-sized planetesimals and the formation of plain if meteorite parent bodies formed at similar times in meteorite parent bodies that were large enough to survive a well-mixed nebula, even if they formed at different loca- Chambers: Meteoritic Diversity and Planetesimal Formation 493

tions. It seems more likely that the parent bodies formed at Differences in the time of formation of various meteor- different times in a nebula that was continually losing CAIs ite parent bodies may ultimately explain many of their prop- as a result of gas drag. (Chondrules were also lost by gas erties, perhaps along the lines described above. However, drag, but these losses were at least partially offset by new we are still left with a challenging conundrum: If the chon- generations of chondrules.) This suggests the ordinary and drules within a single meteorite group truly have ages span- enstatite chondrites formed later than carbonaceous chon- ning 1 m.y. or more, yet turbulent mixing was effective drites. The very low abundance of CAIs in ordinary and during much of the early history of the nebula, why do the enstatite chondrites could mean these meteorites sample various chondrite groups have distinct elemental abun- some of the last planetesimals to form. This would explain dances and O-isotope ratios? Unfortunately, there is no clear why ordinary-chondrite-like parent bodies are much rarer answer to this problem at present. A possible explanation in the asteroid belt than carbonaceous chondrites (Meibom is that the O-isotopic composition, Mg/Si ratios, and other and Clark, 1999). In addition, there may have been a sub- chondrule properties were easily altered during chondrule- stantial decrease in the level of nebula turbulence at around heating events, while resetting of the Al/Mg isotope system this time, with a corresponding increase in the loss rate of used to age date most chondrules occurred only occasion- CAIs. For this reason, a small difference in planetesimal ally. This possibility will necessarily remain speculative until formation time could have resulted in a large difference in better age data for chondrules become available. the number of CAIs these planetesimals contained. It is interesting to pursue this hypothesis a little further 6. DEPLETION OF THE ASTEROID BELT to see if other meteorite characteristics can be explained in terms of differences in formation time rather than location. Collisions were a necessary part of the formation of In the popular core-accretion model for the formation of meteorite parent bodies and must have involved some mix- giant planets (Pollack et al., 1996), Jupiter formed at about ing of material between different planetesimal populations. 5 AU from the Sun because this is where the snowline was Impact and foreign clasts are seen in many mete- located. However, the snowline could have remained here orite classes (Wilkening, 1973; Bunch, 1988; Lipschutz et al., for no more than a few hundred thousand years, so plane- 1989), so collisions continued to be important after parent tesimal formation at 5 AU must have occurred more rapidly bodies formed, causing some mixing and processing of ma- than in the asteroid belt. It seems likely that the inner nebula terial. In general, foreign clasts make up a small fraction became progressively depleted in water over time as large of the total mass of most meteorites. This suggests that col- numbers of planetesimals formed near 5 AU (Stevenson and lision velocities were large, since relatively little impactor Lunine, 1988). Early generations of planetesimals forming material is incorporated in a target body during a high-speed in the asteroid region would have done so in a water-rich, collision (Scott, 2002). This is presumably why most mete- oxidizing environment, precisely the characteristics needed orite groups have retained a distinct identity rather than be- to explain the chemistry of many carbonaceous chondrites. ing equal mixtures of pieces from many sources. However, At later times, planetesimals would have formed in a reduc- a few meteorites such as Kaidun appear to be composed ing environment, highly depleted in water, yielding ordi- mainly of assorted pieces of other parent bodies (Zolensky nary chondrites, and finally the enstatite chondrites. Con- et al., 1996). ceivably, the enstatite chondrites formed in a region that was Carbonaceous-chondrite clasts are quite common in or- nominally beyond the snowline, yet they remained dry be- dinary chondrites, whereas the opposite is not the case cause almost all the water had already been incorporated (Wasson and Wetherill, 1979; Lipschutz et al., 1989). This into planetesimals lying further from the Sun. can be interpreted to mean that ordinary chondrite parent In this interpretation, the timescale for planetesimal for- bodies were rare in the early asteroid belt (Meibom and mation decreases with distance from the Sun, rather than Clark, 1999). Alternatively, this may indicate that carbo- increasing. This is also true if we adopt the traditional as- naceous chondrites formed in the outer asteroid belt while sumption that enstatite, ordinary, and carbonaceous chon- ordinary chondrites formed closer to the Sun, because col- drites formed at progressively larger distances from the Sun. lision fragments destined to become clasts would have The greater degree of thermal metamorphism seen in ordi- drifted inward due to gas drag. A third interpretation is sim- nary chondrites compared to carbonaceous chondrites is ply that parent bodies accreted at often attributed to the fact that ordinary chondrites formed an early stage, so they were unable to incorporate ordinary first, and so formed with a higher abundance of 26Al, the chondrite fragments since these did not yet exist. Each of primary heat source for asteroid-sized bodies (e.g., Grimm these interpretations could be correct since they are not mu- and McSween, 1993). However, more recent studies suggest tually exclusive. that the degree of thermal metamorphism in carbonaceous There are signs that many large bodies were collisionally chondrites was moderated by hydrothermal convection due disrupted and reassembled early in the solar system. The to the presence of large amounts of water ice (McSween et meteorites apparently formed at depth on a body at al., 2002). In this case, the requirement that ordinary chon- least 200 km in size that was later partially disrupted (Good- drites formed before carbonaceous chondrites no longer rich et al., 2002). The plausibly formed when applies. an impact mixed the molten core of a 200–400-km aster- 494 Meteorites and the Early Solar System II oid with mantle and crustal material from the target and Two dynamical models are being actively pursued at projectile (Scott et al., 2001). Ordinary chondrite breccias present. Each invokes the effects of dynamically unstable exhibit a wide range of cooling rates, which suggests they resonances associated with the giant planets. Asteroids lo- come from different depths on a body at least 200 km in cated in many of the orbital resonances undergo large diameter (McSween et al., 2002). Today, 200-km asteroids changes in their orbits, and are likely to fall into the Sun have collisional lifetimes that are longer than the age of the or become ejected from the solar system on a timescale of solar system, so the collision rate in the early asteroid belt ~106 yr (Gladman et al., 1997). When the giant planets must have been substantially higher than now. This suggests formed, asteroids located in resonances would have been the asteroid belt was also more massive early in its history. perturbed onto highly eccentric orbits. Gas drag is very A high early collision rate would explain why iron me- effective for bodies moving on eccentric orbits. If nebular teorites, derived from the cores of differentiated asteroids, gas was still present in the asteroid belt at this time, many are common, while the corresponding mantle and crustal asteroids would have drifted inward as a result. In addition, material is rarely seen. The cooling rates deduced for most secular resonances probably swept across the asteroid re- iron meteorites suggest they came from small asteroids with gion during the dissipation of the nebula (Nagasawa et al., diameters <100 km (Haack and McCoy, 2004) that prob- 2000). Numerical simulations suggest that a combination ably accreted at an early stage while 26Al was still present. of resonance sweeping and gas drag was sufficient to re- The paucity of mantle-derived meteorites suggests that all move most bodies up to 100 km in diameter from the as- the parent bodies have been severely eroded teroid region (Franklin and Lecar, 2000). or catastrophically disrupted, and the mantle material de- The asteroid belt may once have contained bodies larger stroyed. Some core material has survived, in part because it than Ceres. Such objects would have been too massive to be was buried more deeply in the asteroid, but mostly because affected appreciably by gas drag, but they probably would iron have collisional lifetimes 1–2 orders of mag- not have remained in the asteroid region for long. Gravita- nitude longer than stony bodies, according to the cosmic- tional interactions between such large bodies would have ray exposure ages seen in meteorites (Herzog, 2004). caused frequent changes in their orbits. Sooner or later, each The existence of a largely intact basaltic crust on Vesta body would have ended up in an unstable resonance zone, places strong constraints on the collisional evolution of the leading to its removal. Numerical simulations suggest the asteroid belt after this crust formed. Numerical models sug- most likely outcome would be the loss of all large objects gest the mass of the main belt has been within a factor of together with ~99% of objects the size of modern asteroids 5 of its current mass ever since the present regime of high (Wetherill, 1992; Petit et al., 2001). Simulations that take relative velocities was established (Davis et al., 1994). The into account changes in the orbits of the giant planets find high relative velocities in the belt today were almost cer- that the asteroid belt would have approached its current mass tainly caused directly or indirectly by gravitational pertur- on a timescale of ~10 m.y. (Chambers and Wetherill, 2001). bations from Jupiter and Saturn. The giant planets are It is unclear which of these two mechanisms was respon- thought to have formed in less than 10 m.y. (Pollack, 1996), sible for removing most of the original solid material from while thermal models for Vesta suggest its crust formed in the asteroid belt. However, the outcome would have been about 6.6 m.y. (Ghosh and McSween, 1998). It seems likely similar in a number of ways. Most surviving asteroids would that both the mass of material in the asteroid belt and the have undergone substantial radial migration as a result of collision rate were approaching their current values only the clearing process. The degree of radial displacement 10 m.y. after the solar system formed. In contrast, the colli- depends sensitively on the initial orbital parameters of an sion rate may have been higher by several orders of magni- asteroid in each model (Franklin and Lecar, 2000; Cham- tude before Vesta’s crust formed. bers and Wetherill, 2001). As a result, any primordial com- The asteroid belt currently contains about 5 × 10–4 M positional gradient in the asteroid region has been partially of material. Accretion models suggest the region once con- erased. This is reflected in the degree of overlap seen in tained at least 100× as much material in order to form bod- the radial distributions of the major asteroid spectral classes ies such as Vesta within a few million years (Wetherill, (Gradie and Tedesco, 1982). The degree of mixing cannot 1992). A smooth interpolation of the amount of rocky ma- have been too extreme, however, since some radial struc- terial contained in the terrestrial and giant planets suggests ture is apparent today. the asteroid belt once contained ~2 M of solid material. If true, this means the modern belt is depleted by a factor 7. SUMMARY of roughly 4000. While collisional erosion has certainly played a role in removing mass from the asteroid belt, it Understanding the diversity of meteorites is a fundamen- seems highly unlikely that collisions alone could have re- tal problem for planetary science. The solution to this prob- moved the great majority of the primordial material from lem should tell us a great deal about the formation of the this region in only a few million years. For this reason, solar system and planetary systems in general. While we are attention has shifted to dynamical processes that would have not there yet, some aspects of the solution are becoming depleted the belt. apparent. The high mass accretion rates seen in very young Chambers: Meteoritic Diversity and Planetesimal Formation 495 protoplanetary disks and the approximate isotopic homo- ture during nebula evolution. Astrophys. J., 417, 351. geneity of many meteorite components suggests the Sun’s Boss A. P. (1996) Evolution of the solar nebula III. Protoplanetary protoplanetary nebula was hot and turbulent at an early disks undergoing mass accretion. Astrophys. J., 469, 906. stage, leading to substantial mixing of material. Subsequent Boss A. P. (2000) Possible rapid gas giant planet formation in the events mainly acted to introduce heterogeneity. It now seems solar nebula and other protoplanetary disks. Astrophys. J. Lett., 536, L101–L104. likely that interactions between small solids and nebula gas Bunch T. E. and Rajan R. S. (1988) Meteorite regolithic breccias. produced variations in the O-isotope ratios and Mg/Si ratios In Meteorites and the Early Solar System (J. F. Kerridge and and depletions in the moderately volatile elements seen in M. S. Matthews, eds.), pp. 144–164. Univ. of Arizona, Tucson. primitive meteorites. Mixing of different amounts of these Cassen P. (1996) Models for the fractionation of moderately vola- solids led to differences in meteorite bulk compositions. tile elements in the solar nebula. Meteoritics, 31, 793–806. Variations in the amounts of water ice and short-lived radio- Chambers J. E. and Wetherill G. W. (2001) Planets in the asteroid nuclides gave rise to different thermal histories for meteor- belt. Meteoritics, 36, 381–400. ite parent bodies of similar size. Some differences between Chiang E. I. (2002) Chondrules and nebula shocks. Meteoritics, meteorites may be attributed to formation in different parts 37, 151–153. of the nebula. However, it is difficult to draw firm conclu- Chiang E. I. and Goldreich P. (1999) Spectral energy distribution sions here because the process responsible for clearing most of passive T Tauri disks: Inclination. Astrophys. J., 519, 279– 284. of the primordial mass from the asteroid belt also caused Clayton R. N. (2004) Oxygen isotopes in meteorites. In Treatise on widespread radial mixing of the surviving asteroids. The , Vol. 1: Meteorites, , and Planets (A. M. several-million-year spread in the ages of components seen Davis, ed.), pp. 129–142. Elsevier, Oxford. in primitive meteorites makes it unlikely that the proper- Clayton R. N., Hinton R. W., and Davis A. M. (1988) Isotopic ties of meteorites can be explained solely in terms of their variations in the rock forming elements in meteorites. Philos. formation location. Instead, many of the differences seen Trans. R. Soc., 325, 483–501. in meteorites may reflect differences in the time of parent- Cuzzi J. N. (2004) Blowing in the wind III: Accretion of dust rims body formation. Different formation times can explain the by chondrule sized particles in a turbulent protoplanetary neb- varying number of refractory inclusions seen in the chon- ula. Icarus, 168, 484–497. drite groups, as well as the meteorites’ different oxidation Cuzzi J. N. and Weidenschilling S. J. (2006) Particle-gas dynam- states, primordial water abundances, and thermal histories. ics and primary accretion. In Meteorites and the Early Solar System II (D. S. Lauretta and H. Y. McSween Jr., eds.), this volume. Univ. of Arizona, Tucson. Acknowledgments. I would like to thank C. Alexander, A. Cuzzi J. N., Hogan R. C., Paque J. M., and Dobrovolskis A. R. Boss, L. Chambers, L. Nittler, A. Youdin, and two anonymous (2001) Size selective concentration of chondrules and other reviewers for helpful comments and suggestions during the prepa- small particles in protoplanetary nebula turbulence. Astrophys. ration of this manuscript. This work was supported by NASA’s J., 546, 496–508. Origins of Solar Systems and Planetary Geology and Geophysics Cuzzi J. N., Davis S. S., and Dobrovolskis A. R. (2003) Blowing in programs. the wind II. Creation and redistribution of refractory inclusions in a turbulent protoplanetary nebula. Icarus, 166, 385–402. REFERENCES Davis D. R., Ryan E. V., and Farinella P. (1994) Asteroid collisional evolution: Results from current scaling algorithms. Planet. Alexander C. M. O’D. (1996) Recycling and volatile loss in chon- Space Sci., 42, 599–610. drule formation. In Chondrules and the Protoplanetary Disk Desch S. J. and Connolly H. C. (2002) A model of the thermal (R. H. Hewins et al., eds.). pp. 233–242. Cambridge Univ., processing of particles in solar nebula shocks: Application to Cambridge. the cooling rates of chondrules. Meteoritics, 37, 183–207. Alexander C. M. O’D. (2004) Chemical equilibrium and kinetic Desch S. J., Connolly H. C., and Srinivasan G. (2004) An inter- constraints for chondrule and CAI formation conditions. Geo- stellar origin for the beryllium 10 in calcium-rich, aluminum- chim. Cosmochim. Acta, 68, 3943–3969. rich inclusions. Astrophys. J., 602, 528–542. Alexander C. M. O’D., Grossman J. N., Wang J., Zanda B., Drake M. J. and Righter K. (2002) Determining the composition Bourot-Denise M., and Hewins R. H. (2000) The lack of po- of the Earth. Nature, 416, 39–44. tassium isotopic fractionation in Bishunpur chondrules. Mete- Franklin F. A. and Lecar M. (2000) On the transport of bodies oritics, 35, 859–868. within and from the asteroid belt. Meteoritics, 35, 331–340. Alexander C. M. O’D., Boss A. P., and Carlson R. W. (2001) The Ghosh A. and McSween H. Y. (1998) A thermal model for the early evolution of the inner solar system: A meteoritic perspec- differentiation of asteroid 4 Vesta based on radiogenic heating. tive. Science, 293, 64–69. Icarus, 134, 187–206. Amelin Y., Krot A. N., Hutcheon I. D., and Ulyanov A. A. (2002) Gladman B. J., Migliorini F., Morbidelli A., Zappala V., Michel Lead isotopic ages of chondrules and calcium-aluminum rich P., Cellino A., Froeschlé C., Levison H. F., Bailey M., and inclusions. Science, 297, 1678–1683. Duncan M. (1997) Dynamical lifetimes of objects injected into Beckwith S. V. W., Sargent A. I., Chini R. S., and Guesten R. asteroid belt resonances. Science, 277, 197–201. (1990) A survey for circumstellar disks around young stellar Goldreich P. and Ward W. R. (1973) The formation of planetesi- objects. Astron. J., 99, 924–945. mals. Astrophys. J., 183, 1051–1062. Boss A. P. (1993) Evolution of the solar nebula II. Thermal struc- Goodrich C. A., Krot A. N., Scott E. R. D., Taylor G. J., Fioretti 496 Meteorites and the Early Solar System II

A. M., and Keil K. (2002) Formation and evolution of the Yamato 792947, 793408 and 82038: The most primitive H ureilite parent body and its offspring (abstract). In Lunar and chondrites, with abundant refractory inclusions. Meteoritics, Planetary Science XXXIII, Abstract #1379. Lunar and Plan- 37, 1417–1434. etary Institute, Houston (CD-ROM). Krot A. N., Meibom A., Weisberg M. K., and Keil K. (2002) The Gradie J. and Tedesco E. (1982) Compositional structure of the CR chondrite clan: Implications for early solar system pro- asteroid belt. Science, 216, 1405–1407. cesses. Meteoritics, 37, 1451–1490. Grimm R. E. and McSween H. Y. (1993) Heliocentric zoning of Krot A. N., Keil K., Goodrich C. A., Scott E. R. D., and Weisberg the asteroid belt by aluminum-26 heating. Science, 259, 653– M. K. (2004) Classification of meteorites. In Treatise on Geo- 655. chemistry, Vol. 1: Meteorites, Comets, and Planets (A. M. Guan Y., Huss G. R., MacPherson G. J., and Leshin L. A. (2002) Davis, ed.), pp. 83–128. Elsevier, Oxford. Aluminum-magnesium isotopic systematics of aluminum-rich Kurahashi E., Kita N. T., Nagahara H., and Morishita Y. (2004) chondrules in unequilibrated enstatite chondrites (abstract). In Contemporaneous formation of chondrules in the 26Al-26Mg Lunar and Planetary Science XXXIII, Abstract #2034. Lunar system for ordinary and CO chondrites (abstract). In Lunar and and Planetary Institute, Houston (CD-ROM). Planetary Science XXXV, Abstract #1476. Lunar and Planetary Haack H. and McCoy T. J. (2004) Iron and stony iron meteor- Institute, Houston (CD-ROM). ites. In Treatise on Geochemistry, Vol. 1: Meteorites, Comets, Lipschutz M. E., Gaffey M. J., and Pellas P. (1989) Meteoritic and Planets (A. M. Davis, ed.), pp. 325–346. Elsevier, Oxford. parent bodies: Nature, number, size, and relation to present day Haghighipour N. and Boss AP. (2003) On gas drag-induced rapid asteroids. In Asteroids II (R. P. Binzel et al., eds.), pp. 740– migration of solids in a nonuniform solar nebula. Astrophys. 777. Univ. of Arizona, Tucson. J., 598, 1301–1311. Lugmair G. W. and Shukolyukov A. (1998) Early solar system Herzog G. F. (2004) Cosmic-ray exposure ages of meteorites. In timescales according to 53Mn-53Cr systematics. Geochim. Cos- Treatise on Geochemistry, Vol. 1: Meteorites, Comets, and mochim. Acta, 62, 2863–2886. Planets (A. M. Davis, ed.), pp. 347–380. Elsevier, Oxford. Lugmair G. W. and Shukolyukov A. (2001) Early Solar System Hewins R. H. and Connolly H. C. (1996) Peak temperatures of events and timescales. Meteoritics, 36, 1017–1026. flash-melted chondrules. In Chondrules and the Protoplanetary MacPherson G. J. and Davis A. M. (1993) A petrologic and ion Disk (R. H. Hewins et al., eds.), pp. 197–204. Cambridge Univ., microprobe study of a Vigarano type-B refractory inclusion- Cambridge. evolution by multiple stages of alteration and melting. Geo- Hsu W., Wasserburg G. J., and Huss G. R. (2000) High time reso- chim. Cosmochim. Acta, 57, 231–243. lution by use of the 26Al chronometer in the multistage forma- MacPherson G. J. (2004) Calcium-aluminum-rich inclusions in tion of a CAI. Earth Planet. Sci. Lett., 182, 15–29. chondritic meteorites. In Treatise on Geochemistry, Vol. 1: Humayun M. and Clayton R. N. (1995) Potassium isotope cosmo- Meteorites, Comets, and Planets (A. M. Davis, ed.), pp. 201– chemistry: Genetic implications of volatile element depletion. 246. Elsevier, Oxford. Geochim. Cosmochim. Acta, 59, 2131–2148. MacPherson G. J., Davis A. M., and Zinner E. K. (1995) The Huss G. R., MacPherson G. J., Wasserburg G. J., Russell S. S., distribution of aluminum-26 in the early solar system — A and Srinivasan G. (2001) 26Al in CAIs and chondrules from reappraisal. Meteoritics, 30, 365. unequilibrated ordinary chondrites. Meteoritics, 36, 975–997. Marhas K. K., Goswami J. N., and Davis A. M. (2002) Short-lived Hutcheon I. D., Krot A. N., and Ulyanov A. A. (2000) 26Al in nuclides in grains from Murchison: Evidence for solar anorthite-rich chondrules in primitive carbonaceous chondrites: system evolution. Science, 298, 2182–2185. Evidence chondrules postdate CAI (abstract). In Lunar and McKeegan K. D., Chaussidon M., and Robert F. (2000) Incorpo- Planetary Science XXXI, Abstract #1869. Lunar and Planetary ration of short-lived 10Be in a calcium-aluminium-rich inclusion Institute, Houston (CD-ROM). from the . Science, 289, 1334–1337. Hutchison R. (2002) Major element fractionation in chondrites by McSween H. Y., Ghosh A., Grimm R. E., Wilson L., and Young distillation in the accretion disk of a T Tauri sun? Meteoritics, E. D. (2002) Thermal evolution models of asteroids. In Aster- 37, 113–124. oids III (W. F. Bottke Jr. et al., eds.), pp. 559–571. Univ. of Itoh S., Rubin A. E., Kojima H., Wasson J. T., and Yurimoto H. Arizona, Tucson. (2002) Amoeboid olivine aggregates and AOA-bearing chon- Meibom A. and Clark B. E. (1999) Evidence for the insignificance drule from Y-81020 CO3.0 chondrite: Distribution of oxygen of ordinary chondritic material in the asteroid belt. Meteorit- and magnesium isotopes (abstract). In Lunar and Planetary ics, 34, 7–24. Science XXXIII, Abstract #1490. Lunar and Planetary Institute, Meibom A., Desch S. J., Krot A. N., Cuzzi J. N., Petaev M., Houston (CD-ROM). Wilson L., and Keil K. (2000) Large-scale thermal events in Itoh S. and Yurimoto H. (2003) Contemporaneous formation of the solar nebula: Evidence from Fe,Ni metal grains in primi- chondrules and refractory inclusions in the early Solar System. tive meteorites. Science, 288, 839–841. Nature, 423, 728–731. Mostefaoui S., Kita N. T., Togashi S., Tachibana S., Nagahara H., Jones R. H. (1996) Relict grains in chondrules: Evidence for and Morishita Y. (2002) The relative formation ages of ferro- chondrule recycling. In Chondrules and the Protoplanetary magnesian chondrules inferred from their initial aluminum-26/ Disk (R. H. Hewins et al., eds.), pp. 163–172. Cambridge Univ., aluminum-27 ratios. Meteoritics & Planet. Sci., 37, 421–438. Cambridge. Nagasawa M., Tanaka H., and Ida S. (2000) Orbital evolution of Kimura M., Lin Y., and Hiyagon H. (2000) Unusually abundant asteroids during depletion of the solar nebula. Astron. J., 119, refractory inclusions and iron-oxide-rich silicates in an EH3 1480–1497. chondrite: Sahara 97159 (abstract). Meteoritics, 35, A87. Nittler L. R. (2003) Presolar stardust in meteorites: Recent ad- Kimura M., Hiyagon H., Palme H., Spettel B., Wolf D., Clayton vances and scientific frontiers. Earth Planet. Sci. Lett., 209, R. N., Mayeda T. K., Sato T., Suzuki A., and Kojima H. (2002) 259–273. Chambers: Meteoritic Diversity and Planetesimal Formation 497

Norman M., McCulloch M., O’Neill H., and Brandon A. (2004) Sugiura N. and Hoshino H. (2003) Mn-Cr chronology of type Magnesium isotopes in the Earth, Moon, Mars and pallasite IIIAB iron meteorites. Meteoritics, 38, 117–144. parent body: High-precision analysis of olivine by laser-abla- Supulver K. and Lin D. N. C. (2000) Formation of icy planetesi- tion multi-collector ICPMS (abstract). In Lunar and Planetary mals in a turbulent solar nebula. Icarus, 146, 525–540. Science XXXV, Abstract #1447. Lunar and Planetary Institute, Vanhala H. A. T. and Boss A. P. (2002) Injection of radioactivities Houston (CD-ROM). into the forming solar system. Astrophys. J., 575, 1144–1150. Ott U. (1993) Interstellar grains in meteorites. Nature, 364, 25–33. Wasson J. T. and Wetherill G. W. (1979) Dynamical, chemical and Palme H., Larimer J. W., and Lipschutz M. E. (1988) Moderately isotopic evidence regarding the formation locations of asteroids volatile elements. In Meteorites and the Early Solar System and meteorites. In Asteroids (T. Gehrels, ed.), pp. 926–974. (J. F. Kerridge and M. S. Matthews, eds.), pp. 436–461. Univ. Univ. of Arizona, Tucson. of Arizona, Tucson. Weidenschilling S. J. (1977) Aerodynamics of small bodies in the Palme H., Spettel B., Kural G., and Zinner E. (1992) Origin of solar nebula. Mon. Not. R. Astron. Soc., 180, 57–70. Allende chondrules (abstract). In Lunar and Planetary Science Weidenschilling S. J. (1980) Dust to planetesimals — Settling and XXIII, pp. 1021–1022. Lunar and Planetary Institute, Houston. coagulation in the solar nebula. Icarus, 44, 172–189. Petit J. M., Morbidelli A., and Chambers J. E. (2001) The primor- Weidenschilling S. J. (2003) Radial drift of particles in the solar dial excitation and clearing of the asteroid belt. Icarus, 153, nebula: Implications for planetesimal formation. Icarus, 165, 338–347. 438–442. Pollack J., Hubickyj O., Bodenheimer P., Lissauer J., Podolak M., Wetherill G. W. (1992) An alternative model for the formation of and Greenzweig Y. (1996) Formation of giant planets by con- the asteroids. Icarus, 100, 307–325. current accretion of solids and gas. Icarus, 124, 62–85. Wetherill G. W. and Stewart G. R. (1993) Formation of planetary Sahijpal S., Goswami J. N., Davis A. M., Lewis R. S., and embryos — Effects of fragmentation, low relative velocity, and Grossman L. (1998) A stellar origin for the short lived nuclides independent variation of eccentricity and inclination. Icarus, in the early solar system. Nature, 391, 559. 106, 190. Scott E. R. D., Love S. G., and Krot A. N. (1996) Formation of Wilkening L. L. (1973) Foreign inclusions in stony meteorites — chondrules and chondrites in the protoplanetary nebula. In I. Carbonaceous chondritic xenoliths in the Kapoeta . Chondrules and the Protoplanetary Disk (R. H. Hewins et al., Geochim. Cosmochim. Acta, 37, 1985–1986. eds.), pp. 87–98. Cambridge Univ., Cambridge. Wood J. A. (1985) Meteoritic constraints on processes in the solar Scott E. R. D., Haack H., and Love S. G. (2001) Formation of nebula. In Protostars and Planets II (D. C. Black and M. S. mesosiderites by fragmentation and reaccretion of a large dif- Matthews, eds.), pp. 687–702. Univ. of Arizona, Tucson. ferentiated asteroid. Meteoritics, 36, 869–891. Wood J. A. (1996) Unresolved issues in the formation of chon- Scott E. R. D. (2002) Meteorite evidence for the accretion and drules and chondrites. In Chondrules and the Protoplanetary collisional evolution of asteroids. In Asteroids III (W. F. Bottke Disk (R. H. Hewins et al., eds.), pp. 55–70. Cambridge Univ., Jr et al., eds.), pp. 697–709. Univ. of Arizona, Tucson. Cambridge. Scott E. R. D. and Krot A. N. (2004) Chondrites and their compo- Woolum D. S. and Cassen P. (1999) Astronomical constraints on nents. In Treatise on Geochemistry, Vol. 1: Meteorites, Comets, nebular temperatures: Implications for planetesimal formation. and Planets (A. M. Davis, ed.), pp. 143–200. Elsevier, Oxford. Meteoritics, 34, 897–907. Sears D. W. G., Huang A., and Benoit P. H. (1996) Open-system Youdin A. N. and Shu F. (2002) Planetesimal formation by gravi- behaviour during chondrule formation. In Chondrules and the tational instability. Astrophys. J., 580, 494–505. Protoplanetary Disk (R. H. Hewins et al., eds.), pp. 221–232. Youdin A. N. and Chiang E. I. (2004) Particle pileups and plane- Cambridge, Cambridge. tesimal formation. Astrophys. J., 601, 1109–1119. Sekiya M. (1998) Quasi-equilibrium density distributions of small Zanda B. (2004) Chondrules. Earth Planet. Sci. Lett., 224, 1–18. dust aggregations in the solar nebula. Icarus, 133, 298–309. Zhu X. K., Guo Y., O’Nions R. K., Young E. D., and Ash R. D. Shu F. H., Shang H., and Lee T. (1996) Toward an astrophysical (2001) Isotopic homogeneity of iron in the early solar nebula. theory of chondrites. Science, 271, 1545–1552. Nature, 412, 311–313. Stepinski T. F. and Valageas P. (1997) Global evolution of solid Zinner E. (1998) Stellar nucleosynthesis and the isotopic compo- matter in turbulent protoplanetary disks II. Development of icy sition of from primitive meteorites. Annu. Rev. planetesimals. Astron. Astrophys., 319, 1007–1019. Earth Planet. Sci., 26, 147–188. Stevenson D. J. and Lunine J. I. (1988) Rapid formation of Jupi- Zolensky M. E., Ivanov A. V., Yang S. V., Mittlefehldt D. W, and ter by diffuse redistribution of water vapour in the solar nebula. Ohsumi K. (1996) The : Mineralogy of an Icarus, 75, 146–155. unusual CM1 lithology. Meteoritics, 31, 484–493.