Nature © Macmillan Publishers Ltd 1998

8

NATURE | VOL 396 | 19 NOVEMBER 1998 | www.nature.com 237

letters to nature

Ameteorite from theCretaceous/TertiaryboundaryFrank T. Kyte

Institute of Geophysics and Planetary Physics, University of California,

Los Angeles, California 90095-1567, USA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Cretaceous/Tertiary boundary sediments are now widely recog-nized to contain the record of a large asteroid or comet impactevent1, probably at the site of the Chicxulub crater on the Yucatanpeninsula2. After nearly two decades of intensive research, how-ever, much remains unknown about the speci®c nature of theprojectile and of the impact event itself. Here we describe a 2.5-mm fossil meteorite found in sediments retrieved from theCretaceous/Tertiary boundary in the North Paci®c Ocean thatwe infer may be a piece of the projectile responsible for theChicxulub crater. Geochemical and petrographic analyses of thismeteorite indicate that it probably came from a typical metal- andsulphide-rich carbonacous chondrite rather than the porousaggregate type of interplanetary dust considered typical of come-tary materials3. The fact that meteorite survival should beenhanced by impacts at low (asteroidal) velocities4 also impliesthat this meteorite had an asteroidal rather than a cometaryorigin.

Cretaceous/Tertiary (K/T) boundary sediments in DSDP Hole576 (328 21.49 N, 1648 16.59 E) are dark-brown, oxidized, pelagicclays, typical of deep-ocean sediments in the North Paci®c. Thesilicate fraction of this sediment is ®ne-grained (,3 mm) wind-blown dust derived from North America and Asia5. 65 Myr ago, thissite was located in the central portion of the palaeo-North Paci®cbasin (Fig. 1), 9,000 km west of the Chicxulub impact structure2,6,and thousands of kilometres from the nearest continent. The K/Tboundary here has a broad iridium anomaly7, shocked quartzgrains8 as large as 190 mm and grains of magnesioferrite spinel9

(presumably derived from ,250-mm spherules). One sedimentsample (Hole 576 core 8, section 1, 50±52 cm) from the base ofthe iridium anomaly contained an anomalous ,4-mm inclusion oflight-brown clay. This inclusion was separated, air-dried, and splitopen to reveal a 2.5-mm lithic clast (Fig. 2) which we have found is a

fossil meteorite. The clast has a brecciated texture with a ®ne-grained matrix enclosing rounded to angular inclusions of variouscolours, and its surface is peppered with numerous small (,50 mm)opaque grains.

Six chips from separate portions of the meteorite were mountedin polished sections (a total of 2.4 mm2) to examine its interior. Inpolished section, the meteorite is distinguished from the surround-ing clays by its high concentrations of iron oxides that ranged from,1 mm to 250 mm in size. Distinct textural domains can be dis-tinguished (Fig. 3a, b). In some areas, clays and oxides are inter-grown on a micrometre scale. Sometimes there are sharp bordersaround regions of nearly pure clay, which often contain largeinclusions of iron oxides. The clay-rich inclusion shown in Fig. 3bis interpreted to be a pseudomorphic replacement of pre-existingolivine, and the patches of iron oxides within it are probablypseudomorphs after Fe±Ni metal. For comparison, an olivinegrain with metal inclusions from the CM chondrite LEW85001 isshown in Fig. 3c.

Electron-microprobe analyses of the oxides (Table 1) typicallytotal ,90% when Fe is measured as FeO, indicating that theprincipal oxide is haematite (Fe2O3). Nickel oxide is detected inmost of the larger haematite grains. Although NiO is typically ,1%,several grains were found with 2±7% NiO (Table 1). Some haema-tite contains ®nely dispersed, submicrometre, Ni±Fe metal and Ni±Fe sulphides (Fig. 3d). Mostly these are too small to analyse in themicroprobe, but qualitatively Ni/Fe ratios are always high. Only onemetal grain was large enough to analyse quantitatively (Fig. 3d) andyielded a composition of Ni87Fe13. Microprobe analyses of the claysare dif®cult because of the hydrous, porous nature of these samples.Even with broad-beam (,10-mm) analyses, totals typically are, 85%. The relatively pure clays in oxide-free zones are character-ized by high MgO and FeO concentrations, and are compositionallydistinct from the surrounding clays of the light-brown rim (Table1). Typically MgO concentrations range from 6% to 10%, butconcentrations as high as 15% have been measured. Based oncation ratios, these clays are probably interlayered glauconite andsmectite. The region with large patches of haematite on the left sideof Fig. 3a has some intergrown clays with 24% MgO (Table 1) thatare identi®ed as saponite. The clays of the light-brown rim aretexturally and compositionally similar to the dark-brown claystypical of the bulk sediments. Both contain phosphate grains upto 50 mm in size that are probably ®sh scale and bone debris. The

577 576577 576

Figure 1 Palaeoreconstruction map. Shown are the locations of DSDP sites 576

and 577 (®lled circles) and the Chicxulub impact structure (star) at the time of the

K/T boundary impact (65Myr ago).

Figure 2 Photograph of separated meteorite and surrounding clays. A fossil

meteorite (A) was found encased in light-brown clays (B). Portions of the

meteorite still line the interior cavity of the light-brown clays (C), and the white spot

in the cavity is a portion of the black and white inclusion at the centre of the

meteorite surface. We note that some light-brown clays remain on the surface of

the meteorite (D). Typical sediments from the K/T boundary at DSDP Site 576 are

dark brown. Small pieces of dark-brown clay (E) can be seen on the bottom of the

piece of light brown rim. Largest dimension of the meteorite is 2.5mm.

Nature © Macmillan Publishers Ltd 1998

8

main difference between these two samples is the relative absence ofMnO in the light-brown rim.

Neutron activation analyses (Table 2) also show the composi-tional difference of the fossil meteorite from the surroundingsediments. Compared to the dark-brown sediments of the K/Tboundary, this clast is highly enriched in the siderophile elements Ir,Au and Cr and the commonly chalcophile elements As and Sb, whileit is depleted in lithophile elements such as Cs, La, Sm and Th.Although the concentrations of lithophiles are signi®cantly higherthan in chondritic materials, this may result from contaminationfrom the rim, which was not entirely removed from this sample(Fig. 2). The light-brown clays that rim the clast are also highlyenriched in Ir, Au and As relative to the brown clays, but to a muchlesser degree. However, the lithophile contents of the light-brownand dark-brown clays are similar to each other, an observationconsistent with the petrographic and microprobe data.

For comparison, the best documented occurrence of a fossilmeteorite is that of Brun¯o, a 10-cm chondritic clast recoveredfrom an Ordovician limestone10,11. Brun¯o is remarkable in thatdetailed textures of meteoritic chondrules have been preserved,despite nearly complete chemical exchange with surrounding car-bonate sediments. The primary ma®c silicates, metal and sulphidesof Brun¯o were altered to calcite, barite, illite and a host of tracemineral phases. The fossil meteorite from DSDP Hole 576 alsoappears to have retained primary textures, but there is no reason tobelieve that any primary mineralogy has been preserved. The tracesof Ni-rich metal and sulphide are probably not primary phases, but

residues from alteration of pre-existing Fe±Ni metal and sulphides.A similar process has been described in cosmic spherules fromPaci®c sediments, where oxidation of the metallic core in an ironsphere12 formed a metallic residue with 89% Ni. Although saponiteis a common mineral in some carbonaceous chondrites13, havingformed by aqueous alteration of ultrama®c phases, its presence inthis rock could also be the result of secondary alteration followingburial.

Chemical alteration has certainly severely perturbed the com-position of the meteorite reported here. Remarkably, Ir concentra-tions remain within the range typical of chondritic meteorites, andFe and Cr concentrations remain within a factor of two ofthese values (Table 2). As meteorites are not compositionallyhomogenous on a millimetre scale, the abundances of these threeelements could be considered a match with chondritic abundances.The low concentrations of Ni and Co are readily explained by lossduring diagenetic alteration. For example, weathered samples ofMaralinga, an Australian CK chondrite, have suffered 70% loss of Niand 40% loss of Co relative to unweathered CK chondrites, whileconcentrations of Cr, Fe and Ir were relatively unaffected14. There isno simple explanation for the 1,000 times enrichment of Au overchondritic values. Apparently Au is also mobilized during altera-tion, as a 60% loss of this element was observed in Maralinga. Theonly explanation that we can propose for the extreme Au enrich-ment is that during early diagenesis of the initial K/T boundaryejecta deposit, Au was mobilized and the reducing microenviron-ment of the meteorite fragment acted as a sink for Au diffusing

letters to nature

238 NATURE | VOL 396 | 19 NOVEMBER 1998 | www.nature.com

Figure 3 Backscatter electron images of polished sections from the fossil

meteorite. Relict textures of mineral phases that havebeen replaced by haematite

(white grains) and clays (dark-grey matrix) can be seen. a, The region with large

hematite grains on the left of the image is a portionof the black andwhite inclusion

on the surface of the meteorite in Fig. 2. Haematite in this inclusion contains 1±5%

NiO and traces of micrometre-sized nickel sulphides. Dark clays in the centre of

this inclusion are saponite. b, Expanded view of a region from the top of a, which

shows a clay-rich zone with sharp boundaries and spheroidal inclusions of

haematite. This is probably a pseudomorph after euhedral olivine in a ®ne-

grained matrix. Equant bright grey grains that stand out in relief at the left-centre of

the image are NaCl crystals that crystallized from interstitial salts after the section

was made. c, Euhedral olivine with metal inclusions in ®ne-grained matrix from

the CM chondrite LEW85001. d, Backscatter (top) and X-ray images for Fe and Ni

(bottom) from a haematite grain with metal inclusions. The large Ni-rich spot has a

composition of Ni87Fe13.

Table 1 Selected electron microprobe analyses

Lithicoxide

Lithicoxide

Lithicclay

Lithicclay

Lithicclay

Light-brownrim

Dark-brownclay

.............................................................................................................................................................................

SiO2 0.06 0.00 37.69 41.65 53.85 50.55 48.03TiO2 0.26 0.06 0.35 0.21 0.07 0.45 0.67Al2O3 0.03 0.00 12.67 11.39 3.30 16.81 15.05Cr2O3 1.52 0.23 0.42 0.35 0.36 0.00 0.06FeO 86.98 82.47 20.92 12.98 4.77 7.14 6.81MnO 0.07 0.00 0.06 0.03 0.03 0.27 2.82MgO 0.21 0.57 7.80 15.57 24.06 3.73 3.62NiO 0.60 5.50 0.04 0.23 0.12 0.12 0.01CaO 0.05 0.28 0.16 0.31 0.06 1.14 1.77Na2O 0.00 0.07 1.43 2.64 0.78 1.42 1.62K2O 0.00 0.00 3.11 2.98 0.76 4.30 4.09Cl 0.00 0.04 0.53 1.53 0.41 0.29 0.66Total 89.78 89.22 85.18 89.87 88.57 86.22 85.21.............................................................................................................................................................................

All values in per cent.

Table 2 Trace-element concentrations in bulk samples

Lithicclast

Light-brownclays

Dark-brownclays

CI* CV*

.............................................................................................................................................................................

Sc (mgg-1) 49 38 31 5.8 11.4Cr (mgg-1) 6.54 0.11 0.05 2.7 3.6Fe (mgg-1) 413 54 54 182 237Co (mgg-1) 76 51 182 508 659Ni (mgg-1)) 1,370 216 389 10,700 13,400Zn (mgg-1) 565 176 203 312 116As (mgg-1) 312 44 17 1.8 1.5Rb (mgg-1) ,60 112 89 2.2 NASb (mgg-1) 21 3.9 2.7 0.15 0.08Cs (mgg-1) 1.3 7.0 6.2 0.18 NALa (mgg-1) 11 165 159 0.24 0.47Sm (mgg-1) ,0.5 0.7 0.7 0.02 NAIr (ngg-1) 690 38 2.6 460 771Au (ngg-1) 213,000 3,500 ,10 144 143Th (mgg-1) 0.8 8.7 8.6 0.03 NA.............................................................................................................................................................................

NA, not available.*Concentrations in CI (ref. 29) and CV (ref. 31) chondrites provided for comparison.

Nature © Macmillan Publishers Ltd 1998

8

through sediment pore waters. This admittedly far-fetched schemerequires all of the Au deposited in ejecta over at least 30 cm2 of theocean ¯oor to have diffused into this 2.5-mm particle. There isevidence for chemical diffusion at least on a millimetre-scale. Theclays surrounding the meteorite are light brown because they lackthe hydrogenous manganese oxides that are ubiquitous in pelagicclays and which give them their dark-brown colour. During diage-netic alteration of metal and sulphides within the meteorite, a zoneof reducing conditions in the surrounding sediments must havedeveloped, reducing Mn to the soluble Mn2+ (ref. 15) which couldthen diffuse outwards. Glauconite, which we infer to be a principalclay mineral in the altered meteorite, is also known to form onlyunder reducing conditions16.

The Ir anomaly in DSDP Hole 576 was smeared across at least30 cm by bioturbation7. In such slowly accumulating clays5 thisinterval could represent as much as 0.5 Myr of sedimentation, so it isimpossible to state with certainty that this object accreted at thesame moment as the K/T projectile. Interplanetary dust from theasteroid belt is an unlikely source because it should be stronglybiased against particles this large which are destroyed by collisionsbefore solar wind induced drag can decelerate them into Earth-crossing orbits17. In any case, over 0.5 Myr, only ,2 3 1016 g ofinterplanetary dust should accrete to the Earth, an amount which is,4% of the mass of the 10-km chondritic asteroid proposed for theK/T boundary1. Dust produced by a signi®cant comet shower18 is apotential source, but there is no independent physical evidence thatsuch an event occurred 65 Myr ago. The K/T projectile itself iscapable of producing unmelted meteorite fragments and should beconsidered the most plausible source. Numerical simulations of theChicxulub impact show that vertical impact of a 10-km projectile atasteroidal velocities could result in as much as 10% of the projectileexperiencing shock pressures below the melting point4. At moretypical angles approaching 458, even more material could survive.This has also been demonstrated in laboratory experiments ofoblique impacts19, which have been discussed as the possible causeof asymmetries in the Chicxulub crater20. Also, a late Pliocenehypervelocity impact is known to have produced millimetre- tocentimetre-sized unmelted meteorites21 so the potential for meteor-ite survival is an established fact.

The fossil meteorite described in this study is, to our knowledge,the ®rst K/T boundary sample with suf®cient information fromtextural and chemical data to make inferences concerning itsorigin. Shuryatz et al.22 reported micrometre-sized Ir nuggets inimpact melt rocks from Chicxulub, and speculated that theymight be derived directly from the projectile. Robin et al.23 reported,250-mm spheroidal debris with chondritic Ir concentrations andshapes reminiscent of partially melted interplanetary dust. Thesewere in K/Tsediments from DSDP Site 577, only 500 km west of Site576. They made no speci®c interpretation of the source materials,other than that they were chondritic. The fossil meteorite fromDSDP Hole 576 appears to be from (1) a chondritic meteorite with(2) signi®cant amounts of metal and sulphide (4±8%), (3) largeinclusions (.200 mm) of ma®c minerals that also contained metal,and (4) 30±60% ®ne-grained matrix. The known meteorite groupsthat best ®t these criteria could be the CV, CO and CR carbonaceouschondrites24±26. Type CM carbonaceous chondrites are possible, butthey typically contain only 1±3% opaque minerals27. Unequili-brated ordinary chondrites are a less likely source because theyhave only 15% ®ne-grained matrix28. Carbonaceous chondritesconstitute only a few per cent of observed meteorite falls29, butthey may be an important component of the asteroid belt, as most ofthe Antarctic micrometeorites have characteristics typical of CMand CR chondrites30. M

Received 2 June; accepted 17 September 1998.

1. Alvarez, L. W., Alvarez, W., Asaro, F. & Michel, H. V. Extraterrestrial cause for the Cretaceous-Tertiary

extinction. Science 208, 1095±1108 (1980).2. Hildebrand, A. R. et al. Chicxulub crater: A possible Cretaceous/Tertiary boundary impact crater on

the YucataÂn Peninsula, Mexico. Geology 19, 867±871 (1991).

3. Brownlee, D. E. Cosmic dust: collection and research. Annu. Rev. Earth Planet. Sci. 13, 147±173(1985).

4. Pierazzo, E., Kring, D. A. & Melosh, H. J. Hydrocode simulation of the Chicxulub impact event and the

production of climatically active gases. J. Geophys. Res. E (submitted).

5. Rea, D. K., Leinen, M. & Janecek, T. R. Geologic approach to the long-term history of atmospheric

circulation. Science 227, 721±725 (1985).6. Kyte, F. T., Bostwick, J. A. & Zhou, L. The Cretaceous-Tertiary boundary on the Paci®c plate:

composition and distribution of impact debris. Geol. Soc. Am. Spec. Pap 307, 389±401 (1996).

7. Kyte, F. T., Bostwick, J. A. & Zhou, L. Identi®cation of the Cretaceous-Tertiary boundary at ODP Site

886, ODP Site 803, and DSDP Site 576. Proc. ODP Sci. Res. 145, 427±434 (1995).

8. Bostwick, J. A. & Kyte, F. T. The size and abundance of shocked quartz in Cretaceous-Tertiaryboundary sediments from the Paci®c basin. Geol. Soc. Am. Spec. Pap. 307, 403±415 (1996).

9. Kyte, F. T. & Bostwick, J. A. Magnesioferrite spinel in Cretaceous-Tertiary boundary sediments of the

Paci®c basin: Hot, early condensates of the Chicxulub impact? Earth Planet. Sci. Lett. 132, 113±127

(1995).

10. Thorslund, P. & Wickman, F. E. Middle Ordovician chondrite in fossiliferous limestone from Brun¯o,central Sweden. Nature 289, 285±286 (1981).

11. Thorslund, P., Wickman, F. E. & Nystrom, J. A. The Ordovician chondrite from Brun¯o, central

Sweden, I. General description and primary minerals. Lithos 17, 87±100 (1984).

12. Blanchard, M. B., Brownlee, D. E., Bunch, T. E., Hodge, P. W. & Kyte, F. T. Meteroid ablation spheresfrom deep-sea sediments. Earth Planet. Sci. Lett. 46, 178±190 (1980).

13. Zolensky, M., Barrett, R. & Browning, L. Mineralogy and composition of matrix and chondrule rims

in carbonaceous chondrites. Geochim. Cosmochim. Acta 57, 3123±3148 (1993).

14. Kellemeyn, G. W., Rubin, A. E. & Wasson, J. T. The compositional classi®cation of chondrites: V. The

Karronda (CK) group of carbonaceous chondrites. Geochim. Cosmochim. Acta 55, 893±904 (1991).15. Lynne, E. C. & Bonatti, D. Mobility of manganese in diagenesis of deep-sea sediments. Mar. Geol. 3,

457±474 (1965).

16. Deer, W. A., Howie, R. A. & Zussman, J. An Introduction to the Rock-forming Minerals 2nd edn (Wiley,

New York, 1992).

17. Love, S. G. & Brownlee, D. E. A direct measurement of the terrestrial mass accretion rate of cosmicdust. Science 262, 550±553 (1993).

18. Hut, P. et al. Comet showers as a cause of mass extinctions. Nature 329, 118±126 (1987).

19. Schultze, P. H. & Gault, D. E. Prolonged global catastrophes from oblique impacts. Geol. Soc. Am. Spec.

Pap. 247, 239±262 (1990).

20. Schultz, P. H. & D'Hondt, S. D. Cretaceous-Tertiary (Chicxulub) impact angle and its consequences.Geology 24, 963±967 (1996).

21. Gersonde, R. et al. Geological record and reconstruction of the late Pliocene impact of the Eltanin

asteroid in the Southern Ocean. Nature 390, 357±363 (1997).

22. Schuraytz, B. L. et al. Iridium metal in Chicxulub impact melt: forensic chemistry on the K-Tsmoking

gun. Science 271, 1573±1576 (1996).23. Robin, E., Froget, L., JeÂhanno, C. & Rocchia, R. Evidence for a K/T impact event in the Paci®c Ocean.

Nature 363, 615±617 (1993).

24. McSween, H. Y. Jr Petrographic variations among carbonaceous chondrites of the Vigarano type.

Geochim. Cosmochim. Acta 41, 1777±1790 (1977).

25. McSween, H. Y. Jr Carbonaceous chondrites of the Ormans type: a metamorphic sequence. Geochim.Cosmochim. Acta. 41, 477±491 (1977).

26. Weisberg, M. K., Prinz, P., Clayton, R. N. & Mayeda, T. K. The CR (Renazzo-type) carbonaceous

chondrite group and its implications. Geochim. Cosmochim. Acta 43, 1761±1770 (1979).

27. McSween, H. Y. Jr Alteration in CM carbonaceous chondrites inferred from modal and chemical

variations in matrix. Geochim. Cosmochim. Acta 57, 1567±1586 (1993).28. Grossman, J. N., Rubin, A. E., Nagahara, H. & King, E. A. in Meteorites and the Early Solar System (eds

Kerridge, J. F. & Matthews, M. S.) 619±659 (Univ. Arizona Press, Tucson, 1994).

29. Wasson, J. T. Meteorites: their Record of Early Solar System History (Freeman, New York, 1985).

30. Engrand, C. & Maurette, M. Carbonaceous micrometeorites from Antarctica. Meteoritics 33, 565±580

(1998).31. Kellemeyn, G. W. & Wasson, J. T. The compositional classi®cation of chondrites-I. The carbonaceous

chondrite groups. Geochim. Cosmochim. Acta 45, 1217±1230 (1981).

Acknowledgements. This manuscript bene®ted signi®cantly from discussions with A. E. Rubin andcomments from H. McSween, E. Pierazzo and M. Grady. This work was supported by the Geology andPaleontology Program of the National Science Foundation. Curation of DSDP cores is supported by theNSF.

Correspondence and requests for materials should be addressed to the author (e-mail: kyte@igpp.ucla.edu).

letters to nature

NATURE | VOL 396 | 19 NOVEMBER 1998 | www.nature.com 239

Coherent quantumcontrol oftwo-photon transitionsbya femtosecond laser pulseDoron Meshulach & Yaron Silberberg

Department of Physics of Complex Systems, Weizmann Institute of Science,

Rehovot 76100, Israel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Coherent quantum control1±3 has attracted interest as a means toin¯uence the outcome of a quantum-mechanical interaction. Inprinciple, the quantum system can be steered towards a desiredstate by its interaction with light. For example, in photoinducedtransitions between atomic energy levels, quantum interferenceeffects can lead to enhancement or cancellation of the totaltransition probability. The interference depends on the spectral