Planetary systems, including our






Full text


Cosmochemical evidence for astrophysical processes

during the formation of our solar system

Glenn J. MacPhersona,1and Alan Bossb


Department of Mineral Sciences, National Museum of Natural History, Smithsonian Institution, Washington, DC 20560; and


Department of Terrestrial Magnetism, Carnegie Institution of Washington, Washington, DC 20015

Edited by Mark H. Thiemens, University of California, San Diego, La Jolla, CA, and approved October 25, 2011 (received for review June 22, 2011)

Through the laboratory study of ancient solar system materials such as meteorites and comet dust, we can recognize evidence for the same star-formation processes in our own solar system as those that we can observe now through telescopes in nearby star-forming regions. High temperature grains formed in the innermost region of the solar system ended up much farther out in the solar system, not only the asteroid

belt but even in the comet accretion region, suggesting a huge and efficient process of mass transport. Bi-polar outflows, turbulent

diffusion, and marginal gravitational instability are the likely mechanisms for this transport. The presence of short-lived radionuclides in the

early solar system, especially60Fe,26Al, and41Ca, requires a nearby supernova shortly before our solar system was formed, suggesting that

the Sun was formed in a massive star-forming region similar to Orion or Carina. Solar system formation may have been“triggered”by

ionizing radiation originating from massive O and B stars at the center of an expanding HII bubble, one of which may have later provided the supernova source for the short-lived radionuclides. Alternatively, a supernova shock wave may have simultaneously triggered the collapse and injected the short-lived radionuclides. Because the Sun formed in a region where many other stars were forming more or less

contemporaneously, the bi-polar outflows from all such stars enriched the local region in interstellar silicate and oxide dust. This may

explain several observed anomalies in the meteorite record: a near absence ofdetectable(no extreme isotopic properties) presolar silicate

grains and a dichotomy in the isotope record between26Al and nucleosynthetic (nonradiogenic) anomalies.



solar system origin


lanetary systems, including our

own solar system, arise as a natural byproduct of star formation out of interstellar molecular clouds. In detail, many questions remain, and for which there are two complementary ap-proaches. Astronomical observations of young protostellar objects show the pro-cess as is it happening essentially now, but the great distances involved limit our ability to resolvefine structure within pro-toplanetary disks. Laboratory studies of material from our own solar system, in particular of meteorites, comets, and in-terplanetary dust that are preserved more or less untouched since the birth of our solar system 4.568 Ga ago, yield chemical and physical clues to the large-scale pro-cesses and conditions extant at that time. These highly precise analytical measure-ments thus yield direct links between astronomical observations of large-scale processes in newly forming stellar systems and the same processes that occurred long ago in our solar system.

The Earth and other evolved large bodies in the solar system long ago lost all physical traces of the primordial grains from which they accreted. Very small bodies such as asteroids and comets, how-ever, largely escaped planetary heating and melting and therefore do retain their original accretionary materials in varying degrees of preservation. Fortunately, comets plus collisions in the asteroid belt and elsewhere deliver some of this material into the inner solar system, where it falls to Earth in the form of meteorites,

micro-meteorites, and interplanetary dust. The likely significance of this material was recognized long ago. Goldschmidt (1) and Suess and Urey (2) compiled“cosmic” abundance tables for the elements, based in part on chemical analyses of chondrite meteorites. Suess and Urey (2) and later authors (e.g., ref. 3) compared the mete-oritic abundances with measured abun-dances in the solar photosphere and found that, for most condensable elements (and other than gases), the match between CI chondrites in particular and the Sun was very close. As analytical techniques have improved, the match has also improved to the point at which most elements agree within approximately±10% (reviewed in ref. 4). (Ref. 5 presents a recent review of chondrite classification, terminology, and properties.) However, although chondrites preserve grains from the earli-est history of our solar system, most of those grains have not remained entirely pristine. CI and CM chondrites, for ex-ample, are composed primarily of hydrous phyllosilicates that probably formed by aqueous alteration of primary grains within asteroidal parent bodies. Possibly more pristine early solar system materials come from interplanetary dust particles (IDPs), which include both comet and asteroidal dust. Finally, we now have un-doubted comet dust grains collected by the National Aeronautics and Space Admin-istration’s Stardust mission, which ren-dezvoused with Comet Wild 2 in 2004, collected dust particles from the comet’s coma, and returned them to Earth in 2006.

Collectively, the chondrites, IDPs, and stardust grains give us a window to the ancient past, when the solar system con-sisted only of a rotating dust and gas cloud surrounding the proto-Sun. Moreover, these materials provide information about different parts of the early solar system: Some chondrite grains formed very near the infant Sun and later accreted into parent bodies within the asteroid belt, whereas comets are thought to have accreted at the very fringes of the solar system. Surprisingly, the comet dust and IDPs and chondrites share many compo-nents in common, and this fact alone is (as detailed later) critical to understanding major physical transport processes within the infant solar system.

Undoubtedly the catalyst for the modern revolution in cosmochemistry was the fall, in February 1969, of the Allende mete-orite in the northern desert of Mexico. In part, this is because of Allende being a rare type of meteorite, a CV3 carbonaceous chondrite. Even more remarkable were its size (more than 2 tons recovered) and the timing of the fall. Grossman (ref. 6, p. 559) characterized the situation aptly: “At a time when many of the world’s geochemical laboratories were in close

Author contributions: G.J.M. and A.B. performed research and wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

1To whom correspondence should be addressed. E-mail:


communication with each other and were perfecting their analytical techniques to derive the maximum information from a minimum of sample in preparation for the return of thefirst lunar rocks only six months thence, there were suddenly sev-eral tons of a single carbonaceous chon-drite available for study.”However, the singularity of Allende does not stop there. Like all chondrites, Allende and the other CV3 chondrites are aggregates of pre-planetary grains of diverse character (Fig. 1). The difference is in the grain size of some of these components. Immedi-ately upon looking at a broken or cut face of Allende, one’s eye is drawn to the prominent centimeter-sized white clasts (Fig. 1). Composed mainly of the oxides and silicates of aluminum, calcium, mag-nesium, and titanium, these clasts are referred to as refractory inclusions or (more commonly, and herein) calcium-aluminum–rich inclusions (CAIs). All chondrite varieties contain such objects, but in every other chondrite type, they are both rare and very small (<1 mm). In the CV3 chondrites, CAIs attain sizes as large as 2 cm. Had Allende been any type of chondrite other than a CV3, not only would the (rare and much smaller) CAIs have been missed, they could not have been analyzed easily with the types of in-strumentation available at the time. The last piece of the Allende miracle again relates to timing. Although the size and peculiar nature of the CAIs in Allende quickly caught people’s attention in 1969, interest was greatly raised because of other recent events. Larry Grossman was just completing his PhD thesis at Yale University, doing complete thermody-namic equilibrium calculations for the ex-pected condensation of solids from a hot gas of solar composition (published in ref. 7). Although not thefirst to attempt such calculations, his were more compre-hensive in terms of bulk chemistry and were thefirst to take specifically into ac-count the changing composition of the gas as condensation proceeds. He showed that thefirst (i.e., highest temperature) major

minerals to condense are oxides and sili-cates of calcium, aluminum, magnesium, and titanium. In other words, people in 1969 realized that the Allende CAIs po-tentially contain thefirst solid materials to have formed in our solar system. Allende provided us with just the right samples at just the right time. And because the Apollo samples were 6 mo yet to come, the new laboratories needed something to do. Within a few years of Allende’s fall, two singular discoveries based on Allende CAIs would change the course of cosmo-chemistry. Thefirst was the discovery (8) of excesses of16O (relative to17O and

18O) in Allende CAIs. The second was the

discovery (9) of excess26Mg that resulted from the in situ decay of26Al at the time of CAI crystallization. There quickly fol-lowed the discovery in a rare few CAIs [referred to as FUN, for Fractionation and Unidentified Nuclear isotopic effects (10)] of intrinsic nuclear anomalies that could be traced back directly to presolar nucleosynthesis (e.g., ref. 11). Thus, the direct connection wasfirst established between the analysis of solar system objects and astrophysical processes.

The present work is not and cannot be a comprehensive review. Rather, our aim is to highlight how cosmochemical studies have provided compelling evidence for some large-scale astrophysical processes during the formation of our solar system. At the end, we will address some remaining unanswered puzzles and speculate on one possible solution. We will touch only pe-ripherally on the subject of the origin of nuclear anomalies and stellar nucleosyn-thesis, as this topic is covered elsewhere in this special issue of PNAS.

Triggered Star Formation in the Birth of Our Solar System

Star formation requires that portions of a interstellar molecular cloud cross a criti-cal threshold of mass vs. size in order for them to become gravitationally bound and begin collapsing. In low-mass clouds like Taurus–Auriga, star formation arises as a result of turbulence-generated local densityfluctuations, followed by loss of magneticfield support, cloud collapse and fragmentation, and continued gravita-tional collapse of the individual fragments. In giant molecular clouds like Orion, star formation can also be“triggered”by the effects of massive OB stars within the clouds. Numerically, most low-mass (i.e., Sun-like) stars form in giant molecular clouds (e.g., ref. 12). However, statistical considerations aside, there is direct cos-mochemical evidence that our own Sun formed within a giant molecular cloud and that its formation was, indeed, triggered. The evidence primarily comes from the isotopic compositions of meteoritic com-ponents, including CAIs and chondrules.

The idea behind the eventual discovery began with Urey (13), who predicted that the short-lived radionuclide26Al might have existed in the earliest solar system and provided the heat source nec-essary for early planetary melting and differentiation. At that time, neither the instrumentation nor the proper samples existed to test the idea. Schramm et al. (14) revived the idea and measured a number of meteoritic samples in the search for the daughter product of26Al decay,26Mg, but again their analytical techniques were not sufficient to resolve any likely signature. Moreover, as it turned out, they still were not looking at the proper samples. Following the fall of Al-lende and the recognition of the CAIs as possible solar nebular condensates, several laboratories began in earnest the search for extinct26Al. Finally, Lee et al. (9)

demonstrated unambiguous evidence for the in situ decay of26Al in an Allende CAI at a level much greater than the steady-state level throughout the interstellar me-dium. This result has been repeated and greatly amplified since then (reviewed in ref. 15; see example in Fig. 2) (16). The Lee et al.finding (9) of course raised the immediate question: Where did the excess

26Al come from? With a half-life of 0.71

Myr, it must have been made no more than 1 to 2 Myr before the time of CAI formation. This led Cameron and Truran

Fig. 1. A cut surface of the Allende meteorite.

Fig. 2. (A) A cut surface of the Leoville meteorite, showing a centimeter-sized CAI for which the magnesium isotope data (B) clearly indicate the in situ decay of26Al at the time that the inclusion formed 4.56 billion years ago. The abundance ra-tio at the time of formara-tion is given by the slope of the correlation line:26Al/27Al = (5.27±0.17)× 10−5. Data are from ref. 16.






(17) to propose that a nearby supernova not only triggered the collapse of the cloud that would become our solar system, but also seeded it with supernova ejecta that contained26Al (Fig. 3 shows a modern nu-merical simulation of this process; ref. 18). This proposal was countered by an al-ternative idea (19) that local proton irra-diation near our own infant Sun produced the26Al, removing the need for any ex-ternal source. The local-irradiation model has been debated at length, and the gen-eral consensus (not universal) is that it is unlikely for two reasons: (i) the difficulty of coproducing other known short-lived radionuclides (e.g.,41Ca,53Mn) in their

proper abundances relative to26Al (re-viewed in ref. 20); and (ii)26Al is quanti-tatively not correlated with10Be (reviewed

in ref. 21), another short-lived radionu-clide that is produced only by particle ir-radiation. Thus, if all the10Be was made by local irradiation, most of the26Al was

made by some other mechanism. Alumin-ium 26 does correlate strongly with41Ca, however (22)—another very short-lived nuclide—so those two nuclides must be cogenetic. It is not possible for particle irradiation to produce both in their ob-served relative proportions. Thus, plausi-ble production sites for26Al and41Ca are asymptotic giant branch (AGB) stars and supernovas, but AGB stars are highly evolved old Sun-like stars that generally are not located near the sites of stellar birth. Thus, the most likely source of early solar system26Al and41Ca is supernova injection. Because of the very short half-life of41Ca, 105years, that event must have occurred within approximately 1 Myr of the formation of CAIs in our solar system. The second and even stronger in-dication for a nearby supernova is60Fe.

Iron 60 cannot be produced by particle

bombardment because there are no suit-able target elements, and therefore its presence in meteorites is considered an unequivocalfingerprint of a nucleosyn-thetic origin. Shukolyukov and Lugmair (23, 24)first showed evidence for extinct

60Fe in the early solar system, in the form

of Fe/Ni-correlated excesses of60Ni (the daughter product of60Fe) in eucrites (i.e., asteroidal basalts). There also is good evidence (e.g., ref. 25) for extinct60Fe in chondrules, which are the small once-molten silicate droplets that characterize most chondrite meteorites. The60Fe

plausibly originated in either a supernova or an AGB star, but the supernova source can better explain other known short-lived radionuclides such as53Mn (20). A third argument in favor of a supernova source is the advantage that supernova shock waves have over the winds from AGB stars in simultaneously triggering collapse of the presolar cloud and injecting enough freshly synthesized radionuclides into the collapsing cloud to match the in-ferred initial abundances in primitive meteorites (26).

The convincing evidence for a nearby supernova explosion at the time of our solar system’s birth seemingly raises anew one original concern with the Cameron and Truran (17) proposal, namely the huge coincidence required. However, this argument has now been turned upside down (e.g., ref. 12) by the observation that, if our Sun formed in a giant molecular cloud whose evolution was largely con-trolled by early-formed massive O and B stars at its core, a supernova is actually to be expected. Such massive stars have life-times measured in millions rather than billions of years, and they would be ex-pected to explode during the episode of star formation in large stellar“nurseries.”

Note that there is a critical difference between the newer“triggered star forma-tion model”of Hester and Desch (12) versus the one proposed by Cameron and Truran (17) and later investigated by Boss and Keiser (18). In the Hester and Desch scenario (12), it is not the supernova shockwave itself that triggers the ongoing star formation. Rather, it is the outward-moving shock front of the ionized hydro-gen (HII) region, a 10,000-K plasma, created by the intense stellar luminosity, radiation pressure, and stellar winds from the OB stars. In this scenario for our solar system, the supernova was responsible mainly for the late injection of short-lived radionuclides such as26Al and41Ca into the newly formed solar nebula, whereas in the original scenario, the supernova shock not only injects the short-lived radio-nuclides in the presolar molecular cloud core, but it simultaneously triggered col-lapse of the cloud.

Thus, precise measurements of grains within meteorites point to our sun and solar system having formed in a particular astrophysical environment: a giant molec-ular cloud, as opposed to the more quies-cent environment of an isolated region analogous to the Taurus–Auriga cloud. Nebula-Wide Mass Transport of Materials

Chondrules and CAIs from chondritic meteorites formed at very high temper-atures in the solar nebula, in excess of 1,200 °C, and some CAIs formed at tem-peratures approaching 2,000 °C. However, the chondrites themselves formed by rel-atively low-temperature accretion of solid grains in the general region of the asteroid belt. The temperature in the solar nebula dropped off radially away from the very hot center (near the proto-Sun), and at distances out near the asteroid belt the temperature probably was well below 1,000 °C (27). Thus, there has always been a problem reconciling the apparent for-mation temperatures of CAIs and chon-drules with the location of their parent bodies in the asteroid belt. Early ap-proaches to the problem involved explor-ing possible physical mechanisms that could lead to the transport of CAIs inside the disk from the very high temperatures of the inner nebula outward to much larger radial distances from the Sun (26). A somewhat different solution emerged when Shu et al. (28) proposed that CAIs and chondrules actually formed very close to the infant Sun but were expelled by magnetically driven bipolar outflows, up out of the plane of the disk (Fig. 4). Most of the material from these jets escapes into interstellar space, where it is observed in the form of Herbig–Haro objects, but some falls back onto the nebular disk far out from the Sun.

Fig. 3. Theoretical model of a supernova shock wave (top horizontal bar) striking a dense molecular cloud core, inducing dynamical collapse of the cloud, while simultaneously injecting short-lived radio-nuclides (black contours) into the collapsing cloud (modified from ref. 18). The color scales indicate gas density in g/cm3. The cloud is symmetrical around the left border. Models are shown as (A) initial model, (B) after 0.03 Myr, and (C) after 0.1 Myr.


A second piece of the puzzle fell into place with the recognition that all CAIs have very similar oxygen isotopic compo-sitions no matter what kind of chondrite in which they occur. However, the different chondrite classes themselves have distinc-tively different compositions. This implies that all CAIs formed out of the same restricted nebular isotopic reservoir and were later dispersed to their eventual chondrite accretion regions. McKeegan et al. (29) demonstrated that CAIs contained live10Be at the time of their formation. Because this short-lived radio-nuclide forms by particle bombardment and not by stellar nucleosynthesis, the CAIs themselves must have formed very near a source of intense particle bom-bardment. That source likely was the in-fant Sun (refs. 29–31; but see ref. 32 for a dissenting opinion). Finally, the near-identity of CAI oxygen isotopic composi-tions with that of the Sun itself (33) all but confirms the idea that CAIs did form close to the proto-Sun and were physically transported outward into the solar system by some mechanism. However, the mech-anism is still debated. Although bipolar outflow appears to be common among very young stars, the widely cited X-wind model of Shu et al. (28) is only one pro-posed model for producing such outflow. Desch et al. (34) critically examined the details of the X-wind model and con-cluded that it is not consistent with the properties of CAIs and chondrules. For example, temperatures at the X-point were so high that no solids could have existed there at all. Further, the relatively oxidizing conditions at the X-point are inconsistent with the extremely reduced mineral assemblages in CAIs. Finally, the relative abundances of most short-lived radionuclides identified with the earliest

solar system cannot be explained by par-ticle bombardment near the proto-Sun. Desch et al. (34) conclude that alternative mechanisms for producing bipolar outflow are more likely, that CAIs did not form at the X-point but farther out, 0.1 to 0.3 AU from the proto-Sun, and that most of the short-lived radionuclides originated from outside the solar system. Recent as-tronomical evidence also argues against the X-wind model. Extremely high-reso-lution spectral imaging (35) of the jet near the star DG Tau indicates a sharp velocity gradient within the jet, whereas the X-wind model predicts nearly constant velocity. Thus, Agra-Amboage et al. (35) conclude that the X-wind model cannot explain the properties of the DG Tau jet. That work and others (e.g., ref. 36) also specifically note that material in jets may come from a much wider radial distance from the central star than predicted by the X-wind model, perhaps as much as 1 AU (an important factor in preserving dust against evaporation). The preferred model of Agra-Amboage et al. (35) for bipolar outflows is a wind driven by the magnetic structure of the disk, not the central star.

Bipolar outflow is an attractive mass transport mechanism because it is observed to occur in contemporary young stars. However, other transport mechanisms have been proposed that do not involve bipolar outflow. Ciesla (e.g., ref. 37) suggested that particle diffusion within the disk itself led to grains being transported significant distances outward from the Sun over time scales of 0.1 to 1 Myr. Boss (38) showed that a phase of marginal gravitational instability is able to rapidly transport small particles from the inner disk to the outer disk while simultaneously homogenizing most initial isotopic het-erogeneity, over time scales less than 0.001 Myr. It is possible that all three mecha-nisms operated in transporting grains outward from the Sun, but results from the Stardust sample return mission from Comet Wild 2 comet dust may pose a problem for the disk diffusion model. Terminal particles almost exclusively con-sist of anhydrous silicates, chondrule fragments, and even two CAIs, all mate-rials commonly observed in chondritic (i.e., asteroidal) meteorites (39–42). Phyl-losilicates and carbonates are virtually absent (39). The implication is that high-temperature grains from the inner solar system were very efficiently transported not just to the asteroid belt but also well beyond, to the region where Comet Wild 2 accreted. Westphal et al. (43) estimated that between 50% and 65% of the Star-dust grains originated in the inner solar system. Depending on how early the gas giants formed, especially Jupiter, these planets could have posed a formidable obstacle to relatively slow outward

diffu-sive transport of grains past Jupiter’s orbit within the nebula midplane. If Jupiter formed rapidly and cleared away the local disk, and the Stardust grains did form in the inner solar system (but see below), it is more likely that these grains were transported outwards rapidly via bipolar outflow or by a phase of gravitational instability.

Presolar Silicates, the Stardust and FUN CAIs Enigmas, and the Source of Solar System Rocky Matter

Presolar grains in meteorites werefirst suspected to exist in meteorites because of anomalous isotopic compositions of the noble gases in bulk chondrites. Laborious chemical isolation proceduresfinally led to the separation and analysis of actual grains in 1987 (see ref. 44 for a historical review; also see ref. 45). This wasfirst accomplished by dissolving>99% of the meteorites, tracking the isotope anomalies at every step, and when the grains bearing the anomalies werefinally isolated, they turned out to be chemically resistant phases such as diamond, graphite, and SiC. Edward Anders famously described this procedure as burning down the hay-stack tofind the needle. Other presolar phases were later identified, including corundum, spinel, and hibonite. Silicates were conspicuously missing. However, the compositions of Earth and the other rocky bodies require that silicates origi-nally must have dominated the population of solids that went into the making of our solar system. The assumption was that the silicates were less able to survive the high-temperature processing in the early inner solar system, and those that did survive were later destroyed by the ex-treme chemical processing (described earlier) used in the laboratory to isolate presolar grains. Searches for presolar sili-cates required techniques that did not destroy the host meteorite, such as auto-mated in situ isotopic scanning via sec-ondary ion MS. Presolar silicates were eventually found,first in IDPs (46) and subsequently in chondrite meteorites (47), at a level of 300 to 400 ppm relative to the bulk meteorite. The existence of these grains is evidence that presolar silicates were not preferentially destroyed by high temperature processes everywhere in the solar system, and certainly not in the ac-cretion regions where the host IDPs and primitive chondrites formed. How then to explain their low abundances even in such well preserved primitive materials?

This“abundance problem”is com-pounded by the results of the Stardust mission, a chief motivation of which was to return samples of solar system matter that had not experienced reprocessing in the inner solar system, such as was experi-enced by grains occurring in chondrites

Fig. 4. The X-wind model for driving bipolar outflow in a young star. In this model, CAIs are envisioned to form near the X-point at Rx, from which they can be launched upward and outward into interstellar space or the outer solar system. The X-point is located less than 0.1 AU away from the star, far inside the current orbit of Mercury. From Shu FH, Shang H, Lee T (1996) Toward an astro-physical theory of chondrules.Science 271:1545–


(e.g., ref. 47). Especially, there was an expectation that comet dust would contain a high fraction of presolar rocky material preserved from the interstellar medium. Thus, much of the returned material was expected to be amorphous and of low-temperature origin. However, as noted earlier, the returned grains instead consist of anhydrous crystalline materials such as olivine and pyroxene with variable but mostly low ferrous iron contents, CAIs (two so far), and chondrule fragments (39– 42). All these grains resemble the kinds of inner solar system grains found in chon-dritic meteorites (41). Although some amorphous grains were found, it is by no means clear that these are primary as op-posed to being products of the high-ve-locity impacts during sample collection (41). Very few presolar grains of any kind have been recognized, based on the crite-rion of isotopic compositions. However, this last statement may well lie at the very heart of the apparent abundance problem: What defines a presolar grain?

All presolar grains identified so far have extreme isotopic compositions that in-dicate the grains were derived from specific freshly nucleosynthesized material, from supernovas or novas or expelled shells of red giant stars (e.g., Fig. 5). However, the solar system in bulk does not have such compositions, and that difference enables identification of the presolar grains. The general explanation for the Stardust sam-ples (e.g., refs. 41, 48) is that most of the grains were formed in the hot inner re-gions of the primitive solar nebula, near the Sun, and then transported outward to the comet-forming region. This imposes a number of singular requirements. First, almost no presolar silicates remained in the outermost parts of the solar system where the comets accreted. Most presolar material was transported inward and pro-cessed in the inner solar system, with very

little escaping such processing. Second, the transport mechanism from the inner solar system to the outer solar system was remarkably efficient (43). And third, the isotopic composition of condensed (i.e., rocky) matter in the solar system is defined entirely by a mixture of supernova plus nova plus red giant grains. This scenario may be consistent with intermittent phases of gravitational instability in the gaseous disk (38).

Individually and together, these three requirements might be met by a very different scenario. If our solar system formed in a giant star-forming region like Orion or Carina where triggered star birth occurred in the walls of the expanding bubble, a significant source of grains would have been the bipolar outflow of neigh-boring young stellar objects formed before (within a few million years at most) or contemporaneously with our own. Al-though grains are not directly observed in outflows from young stellar objects, they are generally accepted to be present be-cause of the depletions of refractory ele-ments from the gaseous component of the jets (e.g., refs. 35, 36). This idea has been considered before (e.g., ref. 49) but has not been examined in any serious way by cosmochemical means. Consider Fig. 6A, which shows a Hubble Space Telescope photo of the Carina Nebula. The two large columns of gas represent enhanced gas density at the edge of the bubble created in the middle of a large molecular cloud by nearby superenergetic O and B stars, and this is the region of triggered star forma-tion. Young protostars are emerging from the tips of each of the two columns, with bipolar outflows clearly visible. Likely there are young protostars within the col-umns that are not visible (postulated in Fig. 6B). The outflow from the protostar at the lower left of Fig. 6Ais even im-pinging on the sides of the column to its right, possibly propelling both gas and grains into that column’s interior. Because star formation is enhanced in the bubble walls, there is likely to be a greatly en-hanced enrichment of outflow grains rel-ative to other kinds of interstellar grains. Moreover, because all the locally forming stars are born out of a common interstellar cloud, the bulk isotopic compositions of all of the other new solar systems would be similar to our own: The outflow-expelled grains would not be isotopically recogniz-able relative to what we think of as nor-mal. The implication of this is that many or even most of the silicates in the most primitive meteorites or IDPs may be pre-solar, but they are not isotopically recog-nizable by current methods. Indeed, Bradley proposed such a model as far back as 1994 (50). Confirming this is an im-portant scientific and technological

chal-lenge for the future, and a test of our model.

Tielens et al. (49) showed a diagram (Fig. 7) in which they attempted to esti-mate the relative contributions to the ISM grain population from different kinds of stellar sources. They adopted an upper

Fig. 5. Oxygen isotopic composition range of presolar oxide and silicate grains Figure courtesy of Dr. A. N. Nguyen; data sources given in ref. 47.

Fig. 6. A portion of the Carina nebula showing two dense gas columns, each with a young (in-visible) protostar at its tip; both protostars are undergoing bipolar outflow. (A) Note that the outflow from the star at lower left is impinging on the side of the neighboring column. (B) Line drawing ofAshows hypothetical protostars in the hidden interiors of the gas columns, and again with bipolar outflow. Dust from such embedded stars would become fodder for later generations of protostars in the same cloud. Top photo cour-tesy of NASA, ESA, and M. Livio and the Hubble 20th Anniversary Team (STScI).

Fig. 7. Pie diagram showing the postulated fractions of interstellar dust grains that are con-tributed from each of plausible different stellar sources. Young stellar objects contribute as much as 30% according to this model (Discussion). Modified from ref. 49.


limit of 30% for the fraction of interstellar dust might derive from young stellar objects; other authors have argued for a significantly lower fraction of less than 15% (e.g., ref. 51). Independent of what the actual fraction is, in a triggered star-forming region surrounding a large HII region, the fraction could well be higher by a factor of two or more.

Near-contemporaneous stellar outflows from nearby infant stars may provide an explanation for another longstanding puzzle relating to the earliest history of our solar system. When the former presence of short-lived26Al wasfirst demonstrated (9) in an Allende CAI, those authors also analyzed a second CAI at the same time that did NOT once contain26Al at de-tectable levels. This CAI and—as was later shown, a small number of similar objects— contain other unusual isotopic properties, specifically large degrees of mass-de-pendent isotopic fractionation in oxygen and magnesium plus nuclear (not derived from radioactive decay) anomalies in other elements such as calcium and tita-nium (e.g., refs. 11, 52). These rare (in CV3 chondrites) objects were named FUN inclusions in reference to their fraction-ation and unidentified nuclear effects (9). With the advent of secondary ionization MS (i.e., ion microprobe), it was sub-sequently found that much smaller objects (mainly single hibonite crystals) with sim-ilar isotopic properties occur in CM car-bonaceous chondrites. Clayton et al. (53) reviewed (among other things) the data for FUN inclusions and presented a remarkable graph (Fig. 8) that shows a near-perfect noncorrelation between initial26Al/27Al and50Ti, which is a

nu-clear anomaly that can be used to identify FUN inclusions. This noncorrelation re-quires that the early solar system con-tained (at least) two very distinct isotopic reservoirs that did not mix, even when the grains were reheated and, on occasion, melted. Maintaining such distinct reser-voirs in the gas phases is difficult, and is more readily explained if the isotopic sig-natures were preserved entirely in grains.

The existence of FUN inclusions has been a riddle ever since their discovery, and nearly every model for the evolution of the early solar system has difficulty explaining them. For example, their lack of26Al might imply that they formed long after “nor-mal”CAIs. In that case, why do FUN CAIs preserve the nuclear anomalies that clearly are vestiges of nucleosynthetic processes? Alternatively, FUN CAIs might have formed at a much earlier stage of the early solar system than normal CAIs, in which case the26Al itself must have

been injected into the solar system after formation of FUN CAIs but before the formation of normal ones. A test of these hypotheses rests on obtaining a high-precision absolute age of a FUN CAI, a goal that has not yet been achieved be-cause of the great rarity of large FUN CAIs in CV3 chondrites. However, there is a third possibility in the context of the model presented earlier. Other contem-porary infant solar systems near our own, having similar bulk chemistry, would also have produced CAIs as well as other kinds of high-temperature grains near the re-spective young stars. It is quite possible that some of those systems might have had some isotopic differences relative to our own solar system, albeit nowhere near as great as those observed in supernova and nova and red-giant presolar grains. Such systems might, for example, largely have formed before the supernova that gave rise to the26Al in our own system.

The nuclear anomalies seen in FUN in-clusions may reflect heterogeneous iso-topic seeding elsewhere in the cloud by red giant stars or other (much earlier?) supernovas.

Summary and Conclusions

The laboratory study of ancient solar sys-tem materials provides extensive un-derstanding for the astrophysical processes that formed out own solar system, pro-cesses that we can also observe to be oc-curring now in nearby star-forming regions.

Studies of chondrite meteorites indicate that solid grains were evaporated, con-densed, and melted in the hot innermost region of our solar system. Some of this highly processed material ended up much farther out in the solar system, not only the asteroid belt but even in the comet accretion region, suggesting a huge and efficient process of mass transport. Bipolar outflow is one possible mechanism for this transport in view of the fact that it is so common in young stars; turbulent diffusion and marginal gravitational instability may have contributed as well, although the effects of early-formed giant planets on interfering with turbulent diffusion are not yet well understood. The presence of short-lived radionuclides in the early solar sys-tem, such as26Al and41Ca, requires a nearby source of nucleosynthetic mate-rial shortly before solar system formation. The most likely source is a supernova, which is not only possible but probable if the Sun was formed in a massive star-forming region similar to Orion or Carina. In this case the solar system formation was “triggered”either by a supernova shock wave or by ionizing radiation originating from massive O and B stars at the center of an expanding HII bubble. Triggered formation of the solar system in a massive star-forming cloud can explain several observed anomalies in the meteorite re-cord, such as the near absence of detect-able presolar silicate grains and a complete dichotomy in the isotope record whereby

26Al and nucleosynthetic (i.e.,

nonradio-genic) anomalies were present in separate grain populations. The Sun would have formed in the dense border region of the expanding HII bubble, in close proximity to many other newly forming stars. Material expelled from these systems via bipolar outflow would become incorporated into any subsequently formed stars, and it would largely be isotopically indistinguishable from new grains formed in those systems because all originated from the same molecular cloud.

ACKNOWLEDGMENTS.We thank Dr. Ann Nguyen for providing Fig. 5, Drs. John Bradley and Larry Nittler for helpful discussions, and two anonymous reviewers for very constructive suggestions that greatly improved the paper. G.J.M. would like to acknowledge the pioneering work of G. Herbig, who long ago considered the possibility that bi-polar outflow from young stars might be an im-portant source of interstellar dust. This work was supported by National Aeronautics and Space Administration (NASA) Grants NNX11AD43G (to G.J.M.; Cosmochemistry program), NNX09AF62G (to A.P.B.; Origins of Solar Systems), and NASA Astrobiology Institute NNA09DA81A (to A.P.B.).

1. Goldschmidt VM (1938) Geochemische Verteilungsgestze der Elemente IX. Die Mengenverha¨ltnisse der Elemente

und der Atom-Arten. Skrifter Utgitt av Det Norske Vidensk. Akad. Skrifter I. Mat. Naturv. Kl. (Oslo), No. 4.

2. Suess HE, Urey HC (1956) Abundances of the elements.

Rev Mod Phys28:53–74.

Fig. 8. Initial26Al/27Al ratios in early solar system refractory objects (mainly hibonite), plotted against50Ti in the same objects. Data and original diagram are from ref. 53, and the modified dia-gram shown is reproduced from ref. 54. Reprinted from Meteorites, Comet And Planets: Treatise On Geochemistry, Vol 1, MacPherson GJ, Calcium aluminum-rich inclusions in chondritic meteorites, Pages 201-246, Copyright 2003, with permission from Elsevier.


3. Anders E, Ebihara M (1982) Solar-system abundances of

the elements.Geochim Cosmochim Acta46:2363–2380.

4. Palme H, Jones A (2003) Solar system abundances of the

elements. InMeteorites, Comet And Planets: Treatise

On Geochemistry, eds Holland HD, Turekian KK

(Elsevier–Pergamon, Oxford), Vol 1, pp 41–61.

5. Scott ERD, Krot AN (2003) Chondrites and their

components. In Meteorites, Comet And Planets:

Treatise On Geochemistry, eds Holland HD, Turekian KK

(Elsevier–Pergamon, Oxford), Vol 1, pp 143–200.

6. Grossman L (1980) Refractory inclusions in the Allende

meteorite.Annu Rev Earth Planet Sci8:559–608.

7. Grossman L (1972) Condensation in the primitive solar

nebula.Geochim Cosmochim Acta36:597–619.

8. Clayton RN, Grossman L, Mayeda TK (1973) A component of primitive nuclear composition in carbonaceous


9. Lee T, Papanastassiou DA, Wasserburg GJ (1976)

Demonstration of26Mg excess in Allende and evidence

for26Al.Geophys Res Lett3:109112.

10. Wasserburg GJ, Lee T, Papanastassiou DA (1977) Correlated oxygen and magnesium isotopic anomalies

in Allende inclusions: II. Magnesium.Geophys Res Lett


11. McCulloch MT, Wasserburg GJ (1978) Ba and Nd

isotopic anomalies in the Allende meteorite. Ap. J.


12. Hester JJ, Desch SJ (2005) Understanding our origins:

Star formation in HII region environments.Chondrites

and the Protoplanetary Disk, eds Krot AN, Scott ERD, Reipurth B (ASP Conference Series, San Francisco), Vol

341, pp 107–130.

13. Urey HC (1955) The cosmic abundances of potassium, uranium, and thorium and the heat balances of the Earth,

the Moon, and Mars.Proc Natl Acad Sci USA41:127–144.

14. Schramm DN, Tera F, Wasserburg GJ (1970) The isotopic abundance of 26Mg and limits on 26Al in the

early solar system.Earth Planet Sci Lett10:44–59.

15. MacPherson GJ, Davis AM, Zinner EK (1995) The distribution of aluminum-26 in the early Solar System:

A reappraisal.Meteoritics30:365–386.

16. MacPherson GJ, et al. (2010) Early solar nebula condensates with canonical, not supracanonical, initial

26Al/27Al ratios. Astrophysical Journal Letters 711:


17. Cameron AGW, Truran JW (1977) The supernova

trigger for formation of the solar system.Icarus30:


18. Boss AP, Keiser SA (2010) Who pulled the trigger: A supernova or an asymptotic giant branch star?

Astrophys J717:L1–L5.

19. Lee T (1978) A local proton irradiation model for isotopic

anomalies in the solar system.Astrophys J224:217–226.

20. Goswami JN, Vanhala HAT (2000) Extinct radionuclides

and the origin of the solar system.Protostars and

Planets IV, eds Mannings V, Russell SS, Boss AP (Univ

Arizona Press, Tucson), pp 965–994.

21. McKeegan K, Davis AM (2003) InMeteorites, Comet

And Planets: Treatise On Geochemistry, eds Holland HD,

Turekian KK (Elsevier–Pergamon, Oxford), Vol 1, pp


22. Sahijpal S, Goswami JN, Davis AM, Grossman L, Lewis RS (1998) A stellar origin for the short lived

nuclides in the early solar system.Nature391:559–561.

23. Shukolyukov A, Lugmair GW (1993a) Fe-60 in eucrites.

Earth Planet Sci Lett119:159–166.

24. Shukolyukov A, Lugmair GW (1993b) Live iron-60 in

the early solar system.Science259:1138–1142.

25. Telus M, et al. (2011) Possible heterogeneity of 60Fe in chondrules from primitive ordinary chondrites.

Program and Abstracts from the 42nd Lunar and Planetary Science Conference (Lunar and Planetary Institute, Houston), LPI contribution 1608, p 2559. 26. Boss AP, Keiser SA, Ipatov SI, Myhill EA, Vanhala HAT

(2010) Triggering collapse of the presolar dense cloud core and injecting short-lived radioisotopes with

a shock wave. I. Varied shock speeds. Ap. J.708:


27. Boss AP (1998) Temperatures in protoplanetary disks.

Annu Rev Earth Planet Sci26:53–80.

28. Shu FH, Shang H, Lee T (1996) Toward an astrophysical

theory of chondrules.Science271:1545–1552.

29. McKeegan KD, Chaussidon M, Robert F (2000)

Incorporation of short-lived (10)Be in a

calcium-aluminum-rich inclusion from the Allende meteorite.


30. MacPherson GJ, Huss GR, Davis AM (2003) Extinct10Be

in type A CAIs from CV chondrites. Geochim

Cosmochim Acta67:3165–3179.

31. Goswami JN, Marhas KK, Chaussidon M, Gounelle M, Meyer BS (2005) Origin of short-lived radionuclides

in the early solar system. Chondrites and the

Protoplanetary Disk, eds Krot AN, Scott ERD, Reipurth B (ASP Conference Series, San Francisco), Vol 341, pp


32. Desch SJ, Connolly HC, Jr., Srinivasan G (2004) An interstellar origin for the Beryllium 10 in

calcium-rich, aluminum-rich inclusions. Astrophys J 602:


33. McKeegan KD, et al. (2011) The oxygen isotopic composition of the Sun inferred from captured solar


34. Desch SJ, Morris MA, Connolly HC, Boss Alan P (2010) A

critical examination of the X-wind model for

chondrule and calcium-rich, aluminum-rich inclusion

formation and radionuclide production. Astrophys J


35. Agra-Amboage V, Dougados C, Cabrit S, Reunanen J (2011) Sub-arcsecond [Fe II] spectro-imaging of the DG

Tau jet: Periodic bubbles and a dusty disk wind?Astron


36. Podio L, Medves S, Bacciotti F, Eislöffel J, Ray TP (2009) Physical Structure and Dust Reprocessing in a sample of

HH Jets.Astron Astrophys506:779–788.

37. Ciesla FJ (2010) Residence times of particles in diffusive protoplanetary disk environments. I. Vertical motions.

Astrophys J723:514–529.

38. Boss AP (2008) Mixing in the solar nebula:

Implications for isotopic heterogeneity and

large-scale transport of refractory grains.Earth Planet Sci


39. Zolensky M, et al. (2008) Comparing Wild 2 particles to

chondrites and IDPs.Meteorit Planet Sci43:261–272.

40. Nakamura T, et al. (2008) Chondrulelike objects in

short-period comet 81P/Wild 2.Science321:1664–1667.

41. Ishii HA, et al. (2008) Comparison of comet 81P/Wild 2

dust with interplanetary dust from comets. Science


42. Matzel JEP, et al. (2010) Constraints on the formation age of cometary material from the NASA Stardust


43. Westphal AJ, et al. (2009) Mixing fraction of inner solar

system material in comet 81P/WILD2.Ap. J.694:18–28.

44. Anders E, Zinner E (1993) Interstellar grains in primitive meteorites: diamond, silicon carbide, and graphite.


45. Davis AM (2011) Stardust in meteorites.Proc Natl Acad

Sci USA108:19142–19146.

46. Messenger S, Keller LP, Stadermann FJ, Walker RM, Zinner E (2003) Samples of stars beyond the solar

system: Silicate grains in interplanetary dust.Science


47. Nguyen AN, Nittler LR, Stadermann FJ, Stroud RM,

Alexander CM O’D (2010) Coordinated analyses of

presolar grains in the Allan Hills 77307 and Queen

Elizabeth Range 99177 meteorites.Astrophys J719:


48. Brownlee D, et al. (2006) Comet 81P/Wild 2 under

a microscope.Science314:1711–1716.

49. Tielens AGGM, Waters LBFM, Bernatowicz TJ (2005) Origin and evolution of dust in circumstellar and

in-terstellar environments. Chondrites and the

Proto-planetary Disk, eds Krot AN, Scott ERD, Reipurth B (ASP

Conference Series, San Francisco), Vol 341, pp 605–631.

50. Bradley JP (1994) Chemically anomalous, preac-cretionally irradiated grains in interplanetary dust

from comets.Science265:925–929.

51. Dwek E, Scalo JM (1980) The evolution of refractory

interstellar grains in the solar neighborhood.

Astrophys J239:193–211.

52. Lee T, Papanastassiou DA, Wasserburg GJ (1978) Ca

isotopic anomalies in the Allende meteorite.Astrophys


53. Clayton RN, Hinton RW, Davis AM (1988) Isotopic variations in the rock-forming elements in meteorites.

Philos Trans R Soc Lond325:483–501.

54. MacPherson GJ (2003) Calcium-aluminum-rich

in-clusions in chondritic meteorites.Meteorites, Comet

And Planets: Treatise On Geochemistry, eds Holland HD,

Turekian KK (Elsevier–Pergamon, Oxford), Vol 1, pp