Cosmochemistry: Understanding the Solar Systemthrough analysis of extraterrestrial materialsGlenn J. MacPhersona,1 and Mark H. Thiemensb,1aDepartment of Mineral Sciences, National Museum of Natural History, Smithsonian Institution, Washington, DC 20560-0119; andbDepartment of Chemistry and Biochemistry, University of California at San Diego, La Jolla, CA 92093-0356

Cosmochemistry is the chemical analysis of extraterrestrial materials. This term generally is taken to mean laboratory analysis, which is thecosmochemistry gold standard because of the ability for repeated analysis under highly controlled conditions using the most advancedinstrumentation unhindered by limitations in power, space, or environment. Over the past 40 y, advances in technology have enabledtelescopic and spacecraft instruments to provide important data that significantly complement the laboratory data. In this special edition,recent advances in the state of the art of cosmochemistry are presented, which range from instrumental analysis of meteorites totheoretical–computational and astronomical observations.

The term cosmochemistry is mod-ern in its origin, but the sciencecan be traced back more than 200y to the time when meteorites

were beginning to be recognized as origi-nating from outside of the Earth. Thisrecognition was based, in part, on chemicalanalyses of aeroliths that revealed them tobe similar to one another but differentchemically from terrestrial rocks (a concisehistorical account is given by Marvin in ref.1). For the first time, scientists realizedthat they were analyzing pieces of extra-terrestrial matter. The modern science wastruly born in the 20th century owing toseveral notable circumstances. The firstwas the development of isotope chemistryand its essential tool, the isotope massspectrometer. Although cosmochemistry isone of the most interdisciplinary of scien-ces, it has been the determination of theisotopic properties of extraterrestrialmaterials and their components that hasprovided the most exciting advances in thefield and the most exacting constraints ontheory. The second circumstance was theinitiation of the Apollo program in theUnited States in the early 1960s. Becausethe chief science goal of Apollo was todeliver rocks from the surface of the Moonto Earth (Fig. 1), National Aeronauticsand Space Administration (NASA) in-vested major funding in US laboratories tobuild state of the art facilities for theanalysis of the precious cargo to come.Various types of MS featured prominentlyin many of these new laboratories facili-tating isotope measurements of elementsacross the periodic chart. The ensuinganalyses of the Apollo samples were someof the most precise made to that time,allowing the use of a minimum of sample.The third circumstance was remarkableand fortuitous: the fall of the Allendemeteorite in early 1969 (Fig. 2), severalmonths before the Apollo 11 Moon land-ing. The timing was propitious, because allof the laboratories that NASA had gone tosuch lengths to fund were complete andwaiting somewhat idly for the arrival of the

Moon rocks. Allende is a rare type ofmeteorite known as a carbonaceouschondrite that preserves, in near pristineform, material left over from the very birthof our Solar System. Because more than 2tons of material were recovered in theform of thousands of stones weighing upto many kilograms each (2), material wasquickly and widely distributed to the newlyestablished laboratories for study. Isotopicstudies of the conspicuous white fragments(calcium-aluminum-rich inclusions; Fig. 3)from Allende revealed compositionsthat must have derived from dying starsshortly before the Solar System’s birthand provided physical evidence of theprocesses that formed the Solar System.Somewhat earlier, studies of noble gases inchondrites (3, 4) revealed the presence ofisotope anomalies: deviations in composi-tion that are not explainable by normalphysical chemical processes that partitionisotopes and that seemed to be presolar in

nature. These anomalies were postulatedto reside in actual presolar grains (i.e.,interstellar grains), the physical isolationand identification of which did not comeuntil more than a decade later. The pres-ence of presolar isotopic signatures meantthat meteorites suddenly had astronomicalsignificance relating to star formation anddeath. Cosmochemistry had graduatedfrom planetary science to a branch ofastronomy.Extraterrestrial materials available for

laboratory study come from many differentSolar System bodies, and not all arrive asmeteorites. Most meteorites come fromthe asteroid belt by collisions that send

Fig. 1. Astronaut Harrison Schmitt at Tracy’s Rock during the Apollo 17 lunar mission. NASA photo-graph AS17-140-21496. (Reproduced with permission from NASA Johnson Space Center.)

Author contributions: G.J.M. and M.H.T. wrote the paper.

The authors declare no conflict of interest.1To whom correspondence may be addressed. E-mail:macphers@si.edu or mthiemens@ucsd.edu.

19130–19134 | PNAS | November 29, 2011 | vol. 108 | no. 48 www.pnas.org/cgi/doi/10.1073/pnas.1111493108

Dow

nloa

ded

by g

uest

on

June

22,

202

0

fragments of asteroids into the inner SolarSystem. There are two fundamental kindsof asteroidal meteorites: chondrites, whichare, by far, the most common and areaggregates of solar nebular dust grains, andplanetary meteorites, mainly achondritesand irons, which are derived from largerasteroids that have undergone partial tocomplete planetary differentiation pro-cesses such as core formation and crustalevolution. Not all meteorites come fromasteroids; a small fraction are known to belunar in origin (by comparison with Apolloand Luna samples), and some likelycome from Mars. Micrometeorites (cosmicdust particles or interplanetary dust) aresimilar to but not identical with meteoritesand partly derive from comets. WithApollo and Luna, we received extrater-restrial materials from manned and roboticspacecraft sent specifically to return sam-ples to Earth. Such sample return missionsare expensive and therefore, rare, butthey provide one kind of information thatcannot be obtained from meteorites orcosmic dust: context. We know exactlyfrom where the samples come. Roboticmissions have delivered samples from theMoon (Luna; USSR), the solar wind(Genesis; United States), a comet (Star-dust; United States), and an asteroid(Hayabusa; Japan).Laboratory studies of extraterrestrial

materials over the past 50 y have led toa number of truly remarkable discoveries.

Origin and Evolution of the MoonChemical data from the Apollo samplesshowed a remarkable depletion in water,sodium, and other volatile componentscompared with Earth rocks. However,the oxygen isotopes of the lunar samplesare indistinguishable from those isotopeson Earth, although the Solar System as awhole has a huge diversity of isotopiccompositions. Based on these observations,it is now generally thought that Earth’sMoon likely formed as a result of giantcollision between Earth and a Mars-sizedbody only a few tens of millions of years

after Earth’s formation (Fig. 4). After itsformation, the Moon may have beenlargely to completely molten, and theflotation of the mineral feldspar (plagio-clase) on top of this magma ocean gaverise to the lunar highlands. Hence, thehighlands are light-colored anorthosite,and the low-lying plains (Mare) are filledwith dark volcanic basalt. Even beyond thehighland–Mare dichotomy, the Moon is

highly heterogeneous in another and moresubtle way: most of the near side Lunarsurface, centered on the Procellerumbasin, is enriched potassium, rare earthelements, and phosphorous (given theacronym KREEP) relative to the re-mainder of the Moon’s surface (far side).

Meteorites from MarsMost meteorites are on the order of 4.5 Gain age. One small subset of meteoritesstands out in this respect, being in somecases as young as ∼200 my. These mete-orites are all igneous rocks, and some areeven volcanic basalts. They clearly origi-nated on a differentiated planetary bodythat, until very recently, was volcanicallyactive. The most likely culprit was Mars(5). This thought was confirmed in 1983(6) when trapped gases contained in therocks were found to be identical to theMartian atmosphere as measured by theViking spacecraft in 1976. It was proposedthat one Martian meteorite, ALH84001(found in the Antarctic), containedevidence for fossil Martian life (7). Thishighly controversial (but if true, spectacu-lar) idea generated a huge amount of

Fig. 2. Artist’s rendition of the fall of the Allende meteorite over the northern Mexico desert, February8, 1969. Smithsonian Institution image. Don Davis, artist. (Reproduced with permission.)

Fig. 3. Calcium–aluminum-rich inclusions in the Allende meteorite.

MacPherson and Thiemens PNAS | November 29, 2011 | vol. 108 | no. 48 | 19131

SPECIA

LFEATURE:IN

TRODUCTIO

ND

ownl

oade

d by

gue

st o

n Ju

ne 2

2, 2

020

interest in spacecraft exploration of Mars.Although most scientists now think thatALH84001 is not a smoking gun for an-cient Martian life, the 1996 paper byMcKay et al. (7) revolutionized NASA’sMars Exploration Program.

Isotope Anomalies in Calcium–

Aluminum-Rich InclusionsThe analyses of large white clasts (Fig. 3)in the newly fallen Allende meteorite,known as calcium–aluminum-rich in-clusions (CAIs) in reference to theircompositions, quickly led to two remark-able findings. (i) The oxygen isotopiccompositions of the CAIs were like noth-ing seen on Earth, being enriched in 16O

by about 4–5% relative to the other twoisotopes. This finding was interpreted (8)as caused by the presence of a presolarcomponent—possibly tiny grains—preserved in the CAIs. It also led to theapplication of oxygen isotope analysis ofall meteorites, an enormous undertakingthat revealed oxygen isotopes to be vir-tual isotopic maps of the Solar System. Itwas the presence of the oxygen isotopicanomalies that triggered much of the fol-lowing decades of isotopic analysis to findthe source of this anomaly, which afternearly 40 y, has yet to be identified. Everymeteorite class is different, and withinchondrites, every component is different,allowing for systematic studies of meteor-

ites and their evolution. This findingrevolutionized our understanding of theSolar System. (ii) Next followed the dis-covery (9) of excesses of 26Mg that formedby the in situ decay of 26Al, a rare isotopeof aluminum that has a half-life of about0.71 my. The short half-life mandatesthat the 26Al was formed very shortlybefore the Solar System formed, probablyas a result of a nearby supernova. Finally,a few CAIs were found to have isotopicanomalies in other elements that are notthe result of radioactive decay and mustbe traces of nucleosynthetic processeswithin evolved stars that, again, predatedformation of our solar system.

Age of the Earth and Solar SystemNo rocks older than about 4 Ga are pre-served on Earth, and the only samples thathave ages that old are actually singlecrystals (typically zircon). Earth’s history ofplate tectonics and crustal processing andreprocessing has released all traces of theEarth’s first ∼500 my existence. The firstaccurate estimate of the Earth’s age was byPatterson (10), but this estimate was anindirect measurement that relied on me-teorite U-Pb ages to calibrate the leadisotopic composition of the Earth (reviewby Halliday in ref. 11). In contrast, thereturn of the Apollo lunar samples led toa finding that the lunar highland anor-thosites have an age of about 4.5 Ga,a testimony to the Moon’s lack of platetectonics and weathering. However, evenfor the Moon, its age of first formation issubject to considerable uncertainty (11).The true age of the Solar System, whichmust be carefully defined in this contextas the age of formation of the first solidbodies and not the age of the start of cloudcollapse, comes from CAIs. Their com-positional similarity to the predicted hightemperature condensates from a hot solargas and their unique isotopic propertiessuggested the possibility of extreme age.It was shown by Gray et al. (12) thatCAIs possess the lowest initial 87Sr/86Srof any known Solar System materials,proving their extreme antiquity. It waslater shown (13) that the Pb-Pb age forseveral Allende CAIs ranged in value from4.55 to 4.57 Ga. Most recently, therehave been greatly improved precisions ofPb-Pb measurements (14, 15) such thatthe precision is now better than 1 my.

Discovery of Presolar Grains inMeteoritesAs late as 1970, all that was known aboutinterstellar grains was their reddeningeffect on distant starlight. However, asnoted above, peculiar isotopic signaturesin bulk meteorites led people to suspectthat actual grains might be preserved incertain types of chondritic meteorites. Along and difficult study began with the goal

Fig. 4. Artist’s rendition of the stages in the formation of Earth’s Moon. (A) Giant impact hitting Earth.(B) An ejected cloud of molten and gaseous matter goes into orbit around Earth. (C) The ring of orbitingmaterial accretes into the proto-moon. (D) Early crystallizing plagioclase feldspar on the molten Moonfloats to the surface and coalesces into “rockbergs,” which will become the Lunar highlands. (E) Basalticvolcanism on the solidified Moon fills in giant impact basins, forming the great dark Maria. SmithsonianInstitution images. Don Davis, artist. (Reproduced with permission.) The image in the middle is themodern Moon, with its light highlands and dark basins. (Photo reproduced with permission from NASA).

19132 | www.pnas.org/cgi/doi/10.1073/pnas.1111493108 MacPherson and Thiemens

Dow

nloa

ded

by g

uest

on

June

22,

202

0

of isolating these grains. Finally thepainstaking diligence paid off with theisolation of presolar diamonds fromAllende by Lewis et al. (16). There fol-lowed quickly the isolation of presolarsilicon carbide, graphite (Fig. 5), and var-ious oxide grains. These grains containisotopic abundance patterns that are di-rectly traceable to different kinds of nu-cleosynthesis inside of stars. In otherwords, not only are we seeing the rawmatter from which our Solar Systemformed and the nature of the grains thatpopulate interstellar space, we can directlytest theories of stellar nucleosythesis.

Meteoritic Oxygen Isotopes and theOrigin of the Solar SystemThe observation (8) of meteoritic oxygenisotopic anomalies, suggested as derivingfrom admixture of alien nucleosyntheticcomponents based on their isotopic dis-tribution in three isotope spaces (Fig. 6),has been uniquely fundamental in cosmo-chemistry, because oxygen is the main el-ement in stony meteorites and planets; asshown in Fig. 6, it exists at a whole-rocklevel compared with minor admixed pre-solar components discussed in the worksby Zinner et al. (17) and Davis (18).Subsequent work by Thiemens and Hei-denreich (19) revealed that the diagonalline of Fig. 6, thought to be exclusivelyattributed to a nuclear process, could bechemically or photochemically produced.

Others have advocated self-shielding byCO in either the solar nebula near theSun (20) or interstellar molecular clouds

(21, 22). This question remains an out-standing issue in cosmochemistry, and thesource of these anomalies is undefined. Areview by Thiemens (23, 24) discussessome of the major issues, and an entiremonograph (25) has been dedicated tooxygen in the Solar System. As discussedin the work by Burnett et al. (26), meas-urements of the oxygen isotopic composi-tion of the solar wind have revealed thatmore, rather than fewer, issues remain tobe addressed in the unraveling of themeteoritic oxygen isotopic record. A re-cent paper by Krot et al. (20) discussed thepossible linkage between the oxygen iso-topic character of calcium aluminum in-clusions and processes by which oxygenisotopes were partitioned between dustand gas in the early Solar System.

ConclusionsThe papers in this special issue of PNASreview some of these discoveries and oth-ers. A contribution by Davis (18) discussesthe current state of knowledge of presolargrains and their implications for stellarnucleosynthesis. MacPherson and Boss(27) describe how studies of carbonaceouschondrites have led to an understanding ofthe processes by which our Solar Systemformed and how we recognize those sameprocesses occurring now in newly formingstars within our own galactic neighbor-hood. The work by Burnett et al. (26)summarizes the findings from NASA’s

Fig. 5. Transmission electron microscope image of a presolar graphite grain with an embedded titaniumcarbide grain. This grain formed in the atmosphere of an evolved red giant star. [Photo reproduced with per-mission from Thomas Bernatowicz (© 1994 American Geosciences Institute and used with their permission).]

Fig. 6. A three-isotope plot of oxygen for extraterrestrial materials. All nonred symbols representdifferent classes of meteorites, separable by their oxygen isotopic composition in three-isotope space.The red circles reflect the high-temperature CAIs from carbonaceous chondritic meteorites. The line witha slope of ∼1.0 was originally thought to represent the addition of pure 16O to the early Solar System,but now is thought to reflect a chemical or photochemical process that changed isotope ratios in a mass-independent manner.

MacPherson and Thiemens PNAS | November 29, 2011 | vol. 108 | no. 48 | 19133

Dow

nloa

ded

by g

uest

on

June

22,

202

0

Genesis mission, which actually collectedsamples of the Sun in the form of im-planted solar wind ions and brought thecollector materials back for laboratoryanalysis. Because the Sun is 99+% of theSolar System, knowing the Sun’s compo-sition with great accuracy establishes thebulk composition of the Solar System andthe starting material from which all of theplanets, moons, and other bodies firstformed. The work by McCoy et al. (28)reviews the current understanding of theplanet Mars, not only through analysis ofMartian meteorites in terrestrial labora-tories but also through a series of in-creasingly sensitive and precise roboticspacecraft that has landed on Mars.Righter and O’Brien (29) explain what wenow understand about how the terrestrial(rocky) planets formed, including Earth.Some meteorites and comets containabundant organic matter of nonbiologicorigin. How that matter forms and evolvesis complex, and the work by Cody et al.(30) reviews what we know about the

processes that occur on small bodies thatproduce complex organic molecules. Fi-nally, in acknowledgment of the fact thatadvances in cosmochemistry are highlytechnology-dependent, two papers arededicated to the analytical methodsthemselves. The work by Zinner et al. (17)describes the huge advances in laboratoryanalytical instrumentation that have en-abled major discoveries, advances not onlyin precision and sensitivity but also spatialresolution, which enables analysis of in-dividual submicrometer grains. The workby McSween (31) documents just how farspacecraft instrumentation has advancedfrom the early observational days of thelunar landscape. Ideally, one would like toanalyze every sample in an Earth-basedlaboratory under perfectly controlledconditions with the most advanced instru-ments unconstrained by limitations ofpower, weight, or size, but sample returnmissions to other worlds are hugely ex-pensive. Thus, robotic missions are an es-sential component of cosmochemistry.

However, spacecraft instruments are seri-ously hobbled by all of the constraints justlisted. Spacecraft instruments will neverattain the measurement ability of labora-tory instruments, but they are gettingamazingly better. The current precisionand spatial resolution of laboratory meas-urements will certainly not be attainedby spacecraft measurements for decadesand for some kinds of study (e.g., trans-mission EM), possibly ever. However,McSween (31) shows that they now aregood enough to actually answer many (notall) important scientific questions, andthus, there is a major contribution fromspacecraft and laboratory analysis.This series of articles aims to highlight

the excitement and accomplishments of themodern field of cosmochemistry. However,readers are reminded that each one ofthese topics could be expanded to fill manybooks, and therefore, there should be noexpectation of in-depth treatment in thesenecessarily brief reviews.

1. Marvin UB (1992) The meteorite of Eisensheim: 1492–1992. Meteoritics 27:28–72.

2. Clarke RS, Jr., et al. (1970) The Allende, Mexico,meteorite shower. Smithson Contrib Earth Sci 5:1–30.

3. Reynolds JH, Turner G (1964) Rare gases in thechondrite Renazzo. J Geophys Res 69:3263–3281.

4. Black DC, Pepin RO (1969) Trapped neon in meteoritesII. Earth Planet Sci Lett 6:395–405.

5. McSween HY, Stolper EM (1979) Basaltic meteorites. SciAm 242:54–63.

6. Bogard DD, Johnson P (1983) Martian gases in anantarctic meteorite? Science 221:651–654.

7. McKay DS, et al. (1996) Search for past life on Mars:Possible relic biogenic activity in martian meteoriteALH84001. Science 273:924–930.

8. Clayton RN, Grossman L, Mayeda TK (1973) A componentof primitive nuclear composition in carbonaceous mete-orites. Science 182:485–488.

9. Lee T, Papanastassiou DA, Wasserburg GJ (1976)Demonstration of 26Mg excess in Allende and evidencefor 26Al. Geophys Res Lett 3:109–112.

10. Patterson CC (1956) Age of meteorites and the Earth.Geochim Cosmochim Acta 10:230–237.

11. Halliday AN (2003) The origin and earliest history of theEarth. Meteorites, Comet and Planets, ed Davis AM(Elsevier, London). Treatise on Geochemistry, eds HollandHD, Turekian KK (Elsevier, London), Vol. 1, pp 509–557.

12. Gray CM, Papanastassiou DA, Wasserburg GJ (1973)The identification of early condensates from the solarnebula. Icarus 20:213–239.

13. Chen JH, Tilton GR (1976) Isotopic lead investigationson the Allende carbonaceous chondrite. GeochimCosmochim Acta 40:635–643.

14. Amelin Y, Krot AN, Hutcheon ID, Ulyanov AA(2002) Lead isotopic ages of chondrules and cal-cium-aluminum-rich inclusions. Science 297:1678–1683.

15. Bouvier A, Wadhwa M (2010) The age of the solarsystem redefined by the oldest Pb-Pb age of ameteoritic inclusion. Nat Geosci 3:637–641.

16. Lewis RS, Tang M, Wacker JF, Anders E, Steel E (1987)Interstellar diamonds in meteorites. Nature 326:160–162.

17. Zinner EK, Moynier F, Stroud RM (2011) Laboratorytechnology and cosmochemistry. Proc Natl Acad SciUSA 108:19135–19141.

18. Davis AM (2011) Stardust in meteorites. Proc NatlAcad Sci USA 108:19142–19146.

19. Thiemens MH, Heidenreich JE, 3rd (1983) The mass-independent fractionation of oxygen: A novel isotopeeffect and its possible cosmochemical implications.Science 219:1073–1075.

20. Krot AN, et al. (2010) Oxygen isotopic composition ofthe sun and mean oxygen isotopic composion of theprotosolar silicate dust: Evidence for refractoryinclusions. Astrophys J 713:1159–1166.

21. Lyons JR, Young ED (2005) Self shielding as the originof oxygen isotopic anomalies in the early solar system.Nature 435:317.

22. Yurimoto H, Kuramoto K (2004) Molecular cloud originfor the oxygen isotope heterogeneity in the solarsystem. Science 305:1763–1766.

23. Thiemens MH (2006) History and applications of mass-independent isotope effects. Annu Rev Earth Planet Sci34:217–262.

24. Clayton RN (2002) Self-shielding in the solar nebula.Nature 415:860–861.

25. MacPherson GJ, Mittlefeldht D, Jones JJ, eds. (2008)Oxygen in the Solar System. Reviews in Mineralogyand Geochemistry Vol. 68 (Mineralogical Society ofAmerica), 597 pp.

26. Burnett DS, Genesis Science Team (2011) Solar com-position from the Genesis Discovery Mission. Proc NatlAcad Sci USA 108:19147–19151.

27. MacPherson GJ, Boss A (2011) Cosmochemical evidence forastrophysical processes during the formation of our solarsystem. Proc Natl Acad Sci USA 108:19152–19158.

28. McCoy TJ, Corrigan CM, Herd CDK (2011) Combiningmeteorites and missions to explore Mars. Proc NatlAcad Sci USA 108:19159–19164.

29. Righter K, O’Brien DP (2011) Terrestrial planetformation. Proc Natl Acad Sci USA 108:19165–19170.

30. Cody GD, et al. (2011) Establishing a molecular relation-ship between chondritic and cometary organic solids.Proc Natl Acad Sci USA 108:19171–19176.

31. McSween HY, Jr., McNutt RL, Jr., Prettyman TH (2011)Spacecraft instrument technology and cosmochemistry.Proc Natl Acad Sci USA 108:19177–19182.

19134 | www.pnas.org/cgi/doi/10.1073/pnas.1111493108 MacPherson and Thiemens

Dow

nloa

ded

by g

uest

on

June

22,

202

0