ASTROBIOLOGYVolume 4 Number 1 2004copy Mary Ann Liebert Inc

Research Paper

Meteors Do Not Break Exogenous Organic Moleculesinto High Yields of Diatomics

PETER JENNISKENS1 EMILY L SCHALLER2 CHRISTOPHE O LAUX3

MICHAEL A WILSON4 GREG SCHMIDT4 and RICK L RAIRDEN5

ABSTRACT

Meteoroids that dominate the Earthrsquos extraterrestrial mass influx (50ndash300 mm size range) mayhave contributed a unique blend of exogenous organic molecules at the time of the origin oflife Such meteoroids are so large that most of their mass is ablated in the Earthrsquos atmosphereIn the process organic molecules are decomposed and chemically altered to molecules dif-ferently from those delivered to the Earthrsquos surface by smaller (50 mm) micrometeorites andlarger (10 cm) meteorites The question addressed here is whether the organic matter in thesemeteoroids is fully decomposed into atoms or diatomic compounds during ablation If notthen the ablation products made available for prebiotic organic chemistry and perhaps earlybiology might have retained some memory of their astrophysical nature To test this hy-pothesis we searched for CN emission in meteor spectra in an airborne experiment duringthe 2001 Leonid meteor storm We found that the meteorrsquos light-emitting air plasma whichincluded products of meteor ablation contained less than 1 CN molecule for every 30 mete-oric iron atoms This contrasts sharply with the nitrogeniron ratio of 112 in the solid mat-ter of comet 1PHalley Unless the nitrogen content or the abundance of complex organic mat-ter in the Leonid parent body comet 55PTempel-Tuttle differs from that in comet 1PHalleyit appears that very little of that organic nitrogen decomposes into CN molecules during me-teor ablation in the rarefied flow conditions that characterize the atmospheric entry of mete-oroids 50 mmndash10 cm in size We propose that the organics of such meteoroids survive instead as larger compounds Key Words Prebiotic chemistrymdashOrigin of lifemdashMeteorsmdashExogenous organics Astrobiology 4 67ndash79

67

1Center for the Study of Life in the Universe SETI Institute Mountain View California2Dartmouth College Hanover New Hampshire3High Temperature Gasdynamics Laboratory Mechanical Engineering Department Stanford University Stanford

California4NASA Ames Research Center Moffett Field California5Lockheed Martin Space Sciences Laboratory Palo Alto California

INTRODUCTION

REFRACTORY ORGANIC CARBON in extraterrestrialmaterials is a likely source of prebiotic com-

pounds for the origin of life (Oroacute 1961 Chybaand Sagan 1992 1997 Delsemme 1992) Organiccarbon is abundant in the cometary grains probedby the GIOTTO and VEGA spacecraft (Kissel andKrueger 1987 Jessberger and Kissel 1991) mak-ing up some 23 by weight of the comet massfraction and some 66 by weight of dust grainsonce the volatile compounds have evaporated(Greenberg 2000) Primitive asteroids contain or-ganic matter as well up to 12 by weight basedon the carbon content of interplanetary dust par-ticles (IDPs) that are 10 mm in size (Keller et al1995 Flynn et al 2000) Flynn et al (2000) foundfrom C-XANES and Fourier transform infraredspectroscopy that a substantial fraction of thiscarbon is organic Most carbon-rich meteoritescontain less than 3 by weight carbon but thesemeteorites may not be representative of the mostprimitive C-type asteroids

The organic carbon has high molecular massand survives exposure to the vacuum of spaceeven when approaching the Sun as close as Earthrsquosorbit [Tmax 250ndash300 K (Levasseur-Regourd et al2000)] In order for this organic matter to havebeen delivered to an early Earth it had to havesurvived impact with Earthrsquos atmosphere Of allinfalling matter it is believed that less than 8by weight of carbon survives intact in microme-teorites with initial sizes 5ndash50 mm but only inthose that derive from prograde asteroidal orbitsthat approach Earth from behind (Anders 1989)The remainder and the bulk of the mass influxin the form of meteoroids which includes all 50mmndash10 cm-sized meteoroids (Love and Brownlee1993) is ablated in the upper atmosphere duringa phase called a ldquometeorrdquo In particular we willuse the term meteor here to refer to the physicalcondition of rarefied flow that pertains to the in-teraction with Earthrsquos atmosphere of all mete-oroids small enough (less than about 10 cm) notto form a shock wave In that process the organicmolecules are chemically changed in a uniquemanner determined by the process of meteoroidablation and subsequent chemistry in the airplasma in the meteoroidrsquos wake (Jenniskens et al2000a)

This paper investigates the possibility that or-ganic matter carried to Earth via meteors is notnecessarily broken down into atoms or diatomic

molecules and may survive in the form of largermolecules On the primitive Earth this survivingmeteoritic material might have made an impor-tant contribution to the prebiotic organic matterThis idea comes at a time when other investiga-tors are also proposing the survivability of ex-traterrestrial organic matter at relatively hightemperatures Brownlee et al (2002) recently re-ported that refractory carbon in IDPs and mi-crometeorites survive relatively severe heating(1200 K) when exposed to the beam of an elec-tron microscope In these experiments the organicmatter was found to separate from the silicatemelt as a solid carbon-rich material with embed-ded metal It has not been established whetherthis amorphous carbonaceous material producedon heating is elemental or organic although thelatter is more likely In addition Glavin and Bada(1999) have demonstrated that several aminoacids and nucleobases from the Murchison mete-orite survive heating to 1100degC for several sec-onds at reduced pressures

Indeed micrometeorites are found to containsignificant amounts of organic matter that havesurvived the heating process Russell et al (2002)even found the C-H stretch vibration band of or-ganic matter in the debris left in the wake of abright Leonid fireball Based on measurements oforganic matter in micrometeorites found in oceansediments and Antarctic ice Maurette and col-leagues (Maurette 1998 Maurette et al 2000Toppani et al 2001) make the case that solidproducts from meteor ablation raining down tothe surface may have contributed organic matterto the early Earth

In contrast our work pursues the idea that at-mospheric chemistry is essential in creating use-ful prebiotic compounds Jenniskens et al (2000a)provided the first evidence that organic mole-cules may survive the meteor phase with the dis-covery that most light emission and thereforemost high temperature chemistry occurs in acomparatively mild temperature (4300 K) airplasma in almost local thermodynamic equilib-rium just behind the meteoroid (Boyd 2000 Jen-niskens et al 2000a) The product of that chem-istry depends on the time available for reactionsin the high temperature meteor plasma and thebalance between oxygen and nitrogen insertionand polymerization versus carbon extraction andmolecular breakup Reactions with the ablated or-ganic compounds that are too infrequent to occurwith high probability in the allotted time are not

JENNISKENS ET AL68

expected to reach equilibrium After being chem-ically altered the products will continue to reactwhile settling to the Earth surface Before ad-dressing the latter it is necessary to understandthe products that derive from the meteor phaseitself

Meteors represent elusive physical conditionsof rarefied high Mach number (up to 270) flowand non-equilibrium chemistry which makes itdifficult to simulate the conditions in the labora-tory without a better understanding of the phys-ical processes and fate of organic matter in realcases The intense 1998ndash2002 Leonid storms haveoffered an opportunity to obtain data on theseprocesses using modern instruments that wouldotherwise have only a small chance of detectinga meteor If the physical conditions can be de-scribed quantitatively then this will set bound-aries to theoretical models that can provide in-sight into this elusive chemistry

A key constraint for such models is the need topredict the production of the most abundant di-atomic molecules created during decompositionof the complex organic matter When complex or-ganic matter is heated to high temperatures orsputtered by fast ions or energetic photons smallmolecular fragments and atoms are lost from thecomplex organic matter (such as H O H2 O2OH H2O CO CO2 CN and CH4) in a processcalled carbonization (Jenniskens and Russell1993 Fristrom 1995 p 318) Further heating (upto 700 K) leads to additional loss of H and H2 ina process of polymerization which is accompa-nied by the growth of aromatic ring structuresProlonged heating in a confined environmentleads to graphitization which is the stacking ofthose rings into layers to form a crystal (egKoidl et al 1989) Further temperature increaseleads to decomposition by loss of H2 CN CHC2 HCN NH3 and C2H In this latter stage ofdecomposition in the absence of rapid radiativecooling after formation hydrocarbon radicalshigher than C2 rapidly fission into lower-molec-ular-weight products leaving only the thermallystable single and double bonded carbon radicalsCarbon atoms are also present but at low abun-dance due to thermodynamic factors (Fristrom1995 Rairden et al 2000) Upon cooling theselower-molecular -weight products can recon-dense into amorphous carbon to form soot

We propose that the best chance of probing thefate of organic matter in the rarefied high Machnumber flow of small meteoroids is to look for

the production of the CN radical This radical ismost easily detected because of a strong B R Xtransition of low excitation energy with a bandhead at 388 nm CN is a product of reactions in-volving nitrogen embedded in the complex or-ganic matter or interactions between the meteoriccarbon atoms and atmospheric N2 In general ni-trogen is closely associated with the organic mat-ter in IDPs in particular with polyaromatic hy-drocarbons and it is enriched in 15N (Keller et al1995)

Early work by Ceplecha (1971) identified CNemissions in the 212 magnitude flare of a come-tary meteoroid but the identification is doubtfulbecause of the many iron lines in the near-UVOur first efforts at detecting the 388 nm bandhead of CN during the 1999 Leonid Multi-In-strument Aircraft Campaign (MAC) (Jenniskensand Butow 1999 Jenniskens et al 2000b) focusedon slit-less spectroscopy of meteors in first orderin the same spectral region as observed by Ce-plecha using a UV-sensitive intensified chargecoupled device (CCD) camera (Abe et al 2000Rairden et al 2000) The 1999 Leonid storm wasseen under good conditions and the high meteorrates resulted in several bright spectra Our mostprecisely measured spectrum set an upper limitof 1 CN molecule per 3 Fe atoms only about afactor of 2 less than expected if all of the nitrogenin complex organic matter would have been re-leased in the form of CN radicals (Rairden et al2000)

In order to improve upon this result we neededto resolve the underlaying iron emission lines inthe meteor spectrum Here we report results froma high-resolution unintensified slit-less CCDspectrograph that obtained spectra with sufficientspectral resolution to separate the iron linesFrom the new data we conclude that CN pro-duction in meteors is not only low but it isstrongly inhibited

METHODS

The two-stage thermoelectrically cooled slit-less CCD spectrograph for Meteor Astronomyand Astrobiology (ldquoASTROrdquo) was specially de-veloped for the detection of small molecules inthe optical spectra of relatively faint meteors Thedesign provides for a high spectral resolutionwith reasonable detection rates for meteors thatare faint enough to have rarefied flow conditions

METEORS AND EXOGENOUS ORGANIC MOLECULES 69

typical for even smaller (but too faint) 150 mmmeteors The instrument layout is shown in Fig1 An objective grating disperses the light of a me-teor seen at some angle to the viewing directionand a lens projects the spectrum on a cooled CCDdetector A typical result is shown in Fig 2

The detector a Pixelvision camera with a two-stage thermoelectrically cooled 1024 3 1024 pixelback-illuminated SI003AB CCD with 24 3 24 mmpixel size (245 3 245 mm image region) wasused in an unintensified mode to keep as high aspectral resolution as possible The disadvantageof such an unintensified CCD detector is the rel-atively long readout time which is 095 s for 1 34 binning Exposure times needed to be kept to aminimum (01ndash08 s) so as not to have too manyzero-order star images and star spectra overlapthe meteor spectrum Minimum exposure times

were also needed to prevent smearing of thosestar images from Earthrsquos rotation and aircraft mo-tion Though the star background serves to cali-brate the meteor intensity once meteor intensitycalibration was accomplished the star imageswere removed from the spectra by subtraction ofa no-meteor image that was taken in the sameviewing direction and with the same instrumentsetting only seconds after a meteor was detectedNoise residuals of the star images were removedby replacing the residue pixels with low-intensitypixels from neighboring rows A 01 s exposureresulted in a maximum star magnitude that couldbe captured of 1106 In general the limitingmagnitude for meteors was about 157 becauseof their higher angular velocity and meteorsbrighter than magnitude 14 show the most in-tense first-order spectral lines Our best spectra

JENNISKENS ET AL70

FIG 1 Instrument setup on ldquoFISTArdquo and a schematic of the instrumental layout The instrument consists of acooled CCD detector optical system and grating The field of view of the instrument is only 5deg narrow but stretchesin the dispersion direction of the grating for the full range of unvigneted viewing The camera is typically positionedat an angle of 20deg to the aircraftrsquos optical window

probed 22 magnitude meteors at altitudes of90ndash120 km which corresponds to about 1 cm sizemeteoroids in the case of Leonid entry speeds of72 kms These are about 5 magnitudes fainter (a factor of 100 in mass less) than those studiedfrom traditional photographic spectroscopy tech-niques This is small enough to have the requiredrarefied flow field that characterizes meteor con-ditions for the bulk of meteoroid mass influx (Jen-niskens et al 2000a)

The AF-S Nikkor f28300D IF-ED 300ndashmmtelephoto lens collection optics determined thefield of view (5deg) and the short wavelength cut-off of sensitivity which extended down to about360 nm The grating dispersion was chosen in or-der to be able to capture any section of the zerofirst and full second order spectrum (Fig 1) Sec-ond order refers to the integer wavelength dif-ference of 2 for diffractions from two neighbor-ing grating grooves Based on this we chose an11 3 11-cm plane transmission grating 35-54-20-660 with aluminum coating on 12-mm BK7 sub-strate manufactured by Richardson Grating Lab-oratory (Rochester NY) We did not use an orderseparation filter a holder for which is providedby the telephoto lens Without such an order sep-aration filter the near-UV part of the third orderspectrum is also captured which overlaps thesecond order spectrum above 540 nm (3 3 360nm 5 2 3 540 nm) Grating and detector were po-sitioned such that the dispersion occurred alongthe rows of the CCD We measured a dispersionof 540 lmm and blaze wavelength of 698 nm(34deg) This configuration provided a full secondorder spectrum out to about 925 nm The factorygrating specifications suggested a good sensitiv-ity out to 11 mm but we found that the gratingefficiency fell off rapidly above 850 nm Figure 3shows the response curves for the camera cali-

brated for various orders which were measuredby using an approximate point source (an opticalfiber manufactured by Ocean Optics) illuminatedwith a mercury-argon lamp (wavelength calibra-tion) and a tungsten lamp (absolute intensity cal-ibration) The measured efficiencies have beenscaled to the product of the factory responsecurve for the transmission grating the transmis-sion of the camera lens and the CCD quantumefficiency (dashed line Fig 3)

The response curve shown in Fig 3 does notinclude the effects of (a) vigneting (or flatfield-ing) or (b) the geometric correction (a simple co-sine function with angle of incidence) for the ori-entation of the grating relative to the direction ofthe meteor By measuring the response of the in-strument to a homogeneously illuminated flat

METEORS AND EXOGENOUS ORGANIC MOLECULES 71

FIG 2 Second order spectrum of the 090558 UT November 18 2001 Leonid meteor Lines other than thosecaused by iron are marked Brackets indicate emission lines in the overlapping third order

FIG 3 Measured spectral response curves for first(1st) second (2nd) and third (3rd) order The second andthird orders are scaled down for comparison The dashedline shows the theoretical response for the combined fac-tory given grating efficiency CCD quantum efficiencyand camera optics transparency

surface we found that vigneting varied as a func-tion of row and pixel position This allowed us todetermine the flatfielding intensity calibrationcorrection factor ( f ) for a given position of thespectrum on the CCD Two examples are shownin Fig 4 each representing the mean over a bandof CCD rows

The position of the meteor on the sky and thechoice of focus determined what part of the spec-trum was recorded Together with its encasingthe camera left an effective hunting ground of astrip of sky 5deg wide and up to 886deg from the for-ward direction (Fig 1) The wavelength scale wasclose to linear over the full range which spannedabout 13600 pixels

l (nm) 5 15737 1 013382 3 pixel (1)

A small 2 nm residual described by the follow-ing second order polynomial remained

l (observed 2 calculated) (nm) 5 106382 00030227 pix 1 18322e-7 3 pix 3 pix (2)

For each individual spectrum the scale was alsonearly linear and the pixel position of higher-or-der lines was simply a multiple of the pixel po-sition at first order

The highest possible resolution was obtainedonly when the meteor moved nearly perpendic-ular to the dispersion direction A 4-pixel binning

was applied perpendicular to the dispersion di-rection to shorten the readout time but this alsodecreased the resolution if the meteor did notmove perpendicular to the dispersion directionUnfortunately the meteor orientation varies as afunction of the angle on the sky and the leftrightposition relative to the viewing direction in thefield of view Under optimal conditions the fullwidth at half-maximum (FWHM) spectral reso-lution in third order was as low as 18 Aring whichoutperforms photographic spectral analysis tech-niques by a factor of 3

RESULTS

The instrument was deployed during the in-tense 2001 Leonid meteor storm over the conti-nental United States onboard the USAF418thFLTS-operated NKC-135 ldquoFISTArdquo research air-craft during the 2001 Leonid MAC mission at thetime of the intense November 18th 1040 univer-sal time (UT) Leonid storm peak (Jenniskens andRussell 2003)

Figure 2 shows our best result a Leonid witha brief terminal flare which appeared in a west-northwestern direction while flying over LittleRock AR at 090558 UT November 18 (9297W13480N 37000 ft) No meteor trajectory infor-mation is available from a stereoscopic perspec-tive because this was a single-plane mission butthe typical altitude at which such meteors end is93 6 5 km (Jenniskens et al 1998 Betlem et al2000) While most of the spectrum consists of sec-ond order lines of sodium third order lines ofionized calcium are readily identified and verystrong The region shortward of the sodium linesis dominated by intrinsically bright third orderiron lines which were detected despite the factorof 50 lower instrumental sensitivity for third or-der lines (Fig 3)

The spectrum shown in Fig 3 is slightly out offocus at one side of the CCD because of a smallmisalignment of the optics and CCD plane As aresult the spectral resolution increases from aFWHM 5 13 Aring at 370 nm to about 24 Aring at 410nm This does not affect the analysis of the dataother than to require line profile integration priorto comparing relative intensities The sodiumemissions detected above this main portion of thespectrum on the CCD frame are due to the per-sistent afterglow or recombination line emissionfrom sodium atoms released at a time prior to the

JENNISKENS ET AL72

FIG 4 Examples of vigneting in the camera for twodifferent positions of the spectrum on the CCD The firstexample applies to a band near the upper edge of the CCDframe The second example is that valid for the positionof the spectrum in Fig 2

start of the exposure No correction was made forthis (small) contribution to the main spectrum

After summing all relevant rows we foundthat emission lines were on top of a continuumemission identified as the First Positive Band ofN2 observed in second order (Fig 5) No knownsecond or third order molecular band from at-mospheric emissions other than that of CN wasexpected in the region around the CN band headAfter correction for second order instrument re-sponse (Fig 3) we fit the N2 band contour to thewhole spectrum The N2 band contour definedthe background level A small correction for in-strument response to the remaining iron lines wasmade by subtracting the N2 band contour andcorrecting the resulting spectrum in a manner ap-propriate for the third order response (Fig 3)

The CN molecular band

Once instrument and background correctionswere made we focused our attention on the thirdorder spectrum around 388 nm (Fig 6) The newthird order spectrum resolved the weaker ironlines considerably as compared with our earlierfirst order spectra (Rairden et al 2000) The ex-pected Fe line emission spectrum shown as adashed line in Fig 6 was calculated for excita-tion temperatures of 4300 and 4500 K using allneutral iron lines in the NIST Atomic SpectraDatabase Version 20 (NIST Standard ReferenceDatabase 78) A comparison of the observed andcalculated emission lines established their iden-

tity and showed that weak iron lines did not con-tribute to the background continuum It shouldbe noted that second order iron lines are presentelsewhere in the spectrum

As indicated in Fig 6 two iron lines bracket theposition of the CN molecular band Because ofthe absence of any additional faint line emissionsin this region an upper limit on the CN abun-dance could be established Though the CN bandmight have in fact been detected as a small bumpit is at the noise level of the spectrum

The relative abundance of CN molecules versus iron atoms was calculated using theNEQAIR2 model of heated air in thermodynamicequilibrium (Laux 1993 Park et al 1997) The CNband profile for a resolution of 18 Aring FWHM isshown in Fig 6 The band head at 38844 nm isunresolved with its width reflecting the instru-ment profile With ICN equal to the total inte-grated emission intensity of CN in the range375ndash393 nm and IFe equal to the total integratedemission intensity of all third order Fe lines in therange 3813ndash3867 nm we calculated for T 5 4500that IFenFe 5 324 3 10216 Wsr and ICNnCN 5558 3 10217 Wsr where nFe is the total concen-tration of Fe atoms in cm23 and nCN is the totalconcentration of CN moleculescm3 For T 54300 K IFenFe 5 213 3 10216 Wsr and ICNnCN 5 393 3 10217 Wsr To calculate these val-ues we assumed that the emitting levels were in

METEORS AND EXOGENOUS ORGANIC MOLECULES 73

FIG 5 Extracted spectrum of Fig 2 after correction forinstrument response and vigneting The dashed lineshows the contribution from the first positive band of N2in second order

FIG 6 Detail of Fig 5 showing the position of the CNband and the iron lines in third order A model spec-trum at 4300 K of iron atom line emission is shown by adashed line The continuum background is the band con-tour of N2 in second order The thick line is a theoreticalprofile of the CN band for air plasma at 4300 K and in-strumental broadening FWHM 5 18 Aring

Boltzmann equilibrium with the ground state ofFe and CN at the given temperature These areconcentrations per total Fe atoms and total CNmolecules (that is not per molecule of CN in theexcited electronic B state)

From these calculations we found that the ra-tio of the total number of atoms of CN and Fe inthe meteor plasma was nCNnFe 0017 for an ex-citation temperature of 4500 K (or 0019 for anexcitation temperature of 4300 K) The largest un-certainty was caused by the position of the back-ground level However given the good fit of theN2 band contour to the overall spectrum short-ward of 590 nm (Fig 5) a conservative error ofless than 50 places the upper limit at nCNnFe 003

An even more precise upper limit was derivedfrom a Geminid spectrum obtained during the2001 Geminid shower with the same instrumentdeployed from a ground site at Fremont Peak Ob-servatory California (Fig 7) For that spectrumwe calculated for an excitation temperature of4500 K that nCNnFe 0011 (or 0012 at 4300 K)However Geminid meteoroids frequently ap-proach the Sun to within 014 AU which heatsthe meteoroids every 16 years to at least T 5 510K (the temperature of small zodiacal cloud mete-oroids) and possibly as high as 1800 K if thelarger meteoroids are more like a black body(Levasseur-Regourd et al 2000) This is hotenough to change their morphology into morecompact objects and can result in the loss of someor all of the organic matter

The more intense Geminid spectrum confirmsthat no weak second or third order atomic linesof any kind are present at the position of the CNband Hence we conclude that the meteor plasmaresponsible for most of the Leonid meteorrsquos emis-sion contains 1 CN molecule per 30 Fe atomsThis result improves by more than a factor of 10the upper limit of 1 CN per 3 Fe atoms found ear-lier by Rairden et al (2000) These numbers aresignificantly less than expected If all of the ni-trogen in the complex organic matter of the cometdust was decomposed in the form of CN mole-cules this would give nCNnFe 5 079 6 002based on the observed nitrogen abundance ofcomplex organic matter in small dust grains ofcomet 1PHalley (Delsemme 1991) The ob-served abundance is a factor of 26 lower

CN can also be formed by reactions in the me-teor air plasma involving CO2 and N2 in the am-bient atmosphere Though CN formation via suchaerothermochemistry is not well understood inthe case of thermodynamic equilibrium therewould not be enough CN made from this mech-anism to be detectable The calculated CN abun-dance for a 4300 K air plasma in local thermo-dynamic equilibrium at 95 km altitude equals10 3 105 molecules of CNcm3 The 22 magni-tude Leonid meteor studied here has a mass ofabout 044 g of which 0033 g is iron (Delsemme1991) That iron is ablated and distributed over apath length of about 17 km (Betlem et al 2000)in a cylindrical volume of at least a 2-m radius(Boyd 2000) which indicates a mean density ofnFe 5 2 3 109 atomscm3 and an expected nCNnFe of about 5 3 1025 a factor of 6 less than ourdetection limit

DISCUSSION

There are several reasons why the abundanceof CN in the meteor plasma can be a factor of 26lower than expected from the decomposition ofmeteoric organic matter during ablation (1) theabundance of organic matter in the dust of55PTempel-Tuttle is not known (2) the abun-dance of nitrogen in that organic matter is un-known (3) the fate of organic matter in the in-terplanetary medium due to space weathering isnot known (4) the organic matter could havebeen lost at higher altitude (earlier in flight) thanwhere the measurement was made (5) theprocess of ablation could favor the release of ni-

JENNISKENS ET AL74

FIG 7 Detail of the 084127 UT Geminid meteor spec-trum observed on December 13 2001 with symbols andlines the same as in Fig 6

trogen in atomic form or in the form of some othersmall molecules or (6) the nitrogen could survivein the form of larger organic compounds that arenot detected The latter possibility could providea new pathway towards prebiotic compoundsWe will now consider the significance and im-plication of these arguments one by one

Since the abundance of organic matter in thedust of 55PTemplel-Tuttle is not known it ispossible that there was little or no organic mat-ter in the Leonid particles to begin with Conse-quently the absence of nitrogen-bearing organicmatter in the Leonid meteoroids may have re-sulted in our inability to detect CN The onlyother measurement that could have detected ni-trogen-bearing organic matter in the dust (as op-posed to molecules in the gas phase) is in situmass spectroscopy The Leonid parent comet hasnot been visited In the coma of comet 1PHal-ley nitrogen-bearing organic matter was detectedin particles (the CHON particles) analyzed by themass spectrometers on the GIOTTO and VEGAspacecraft Because comet Halley is not the par-ent of the Leonid stream it is recommended thatfuture work focus on the meteoroids of the Ori-onid and eta-Aquarid streams that do originatefrom comet Halley Though bright meteors fromthese showers are much more difficult to observemuch the same result would be expected TheLeonid parent comet is remarkable only becauseit passes so close to Earthrsquos orbit Most impor-tantly since 55PTempel-Tuttle is in a Halley-type orbit and it should originate from the Oortcloud just as 1PHalley there is no reason to ex-pect that it is less rich in organic matter

For the same reason there is no reason to ex-pect that the nitrogen content or composition ofthe organic matter would differ from that of1PHalley Halley dust is as rich in nitrogen thanprimitive meteorites Though the nitrogen con-tent in the organic matter of comet Halley hasbeen questioned because of calibration problemswith the detectors there is no reason to expectthere to be more nitrogen in the less primitive me-teorites The heating and irradiation that alter theorganics in meteorites tend to evolve the organicmatter towards materials less rich in nitrogen andoxygen (Jenniskens et al 1993)

It is unknown how the organic content changeswith dust grain size and exposure to the plane-tary medium hence the fate of the organic mat-ter in the interplanetary medium is not knownHowever we know when the observed Leonids

were ejected from the parent comet in 1767 The090558 UT Leonid meteoroid never came closerto the Sun than 0976 AU and remained in the in-terplanetary medium only for 225 years (sevenorbits) most of that time far from the Sun

Might the organic matter have been lost in theEarthrsquos atmosphere early in the trajectory athigher altitudes when the meteoroid first warmedup to above the evaporation and decompositiontemperature of T 700 K This scenario was pro-posed by Elford et al (1997) who observed thatradar meteors are initially detected at heights upto 140 km They concluded that a more volatilecomponent than stony minerals must be ablatingat those heights The specific case of fast Leonidswas discussed by Steel (1998) who calculated thealtitude of evaporation for a range of small or-ganic compounds with relatively low sublimationtemperatures finding ablation heights in excessof 140 km for some compounds However thelight curve of bright Leonid fireballs above 136km was found to be very flat and starting at about200 km (Spurny et al 2000ab) which is incon-sistent with the light curves calculated by Elfordet al (1997) and Steel (1998) At 200 km the grainsare not warm enough to evaporate the organiccompounds that survived the vacuum of space

Indeed the Leonid studied here shows nosign of differential ablation of minerals in order oftheir volatility The elements Na Fe Mg and Caare lost in the same proportion (Fig 2) exceptfor the last wisp of matter after the catastrophicfragmentation during the terminal flare (a topicof future work) Sodium is at least partiallyfound in more volatile minerals than those con-taining magnesium and iron Though somesmaller Leonid meteoroids are known to losesodium early in their trajectory this is onlyfound in fragile cometary shower meteoroidssmall enough to expose the minerals efficientlyto the impinging air flow (BorovicIuml ka et al 1999)Sodium may be contained in the larger Leonidsat lower altitudes because of the limited timeavailable for inward diffusion of heat and out-ward diffusion of molecules Although the or-ganic and mineral components are separate en-tities they are intimately mixed down to verysmall size scales in comet Halley dust and thediffusion of those volatile compounds out of thegrains (if at all possible) will take time Such con-siderations are not included in the model of Steel(1998) Hence the organic matter is not lost earlyin flight

METEORS AND EXOGENOUS ORGANIC MOLECULES 75

The low CN abundance in the air plasma of the090558 UT Leonid meteoroid cannot be under-stood by merely allowing some fraction of nitro-gen to escape as N NH N2 NH3 NH2 NCONO HNO or HCN etc The configuration of Nin the complex organic matter of meteoroids isnot known However from meteorite and mi-crometeorite studies (where much organic matterhas been altered by aqueous alteration processes)we know that a significant fraction of the nitro-gen is contained in the organic framework (seereferences in Keller et al 1995 Greenberg 2000)Aqueous alteration tends to move nitrogen fromthe organic framework to functional groups suchas -NH In cometary matter no aqueous alter-ation has occurred So if N is in the frameworkin meteorite organics it should also exist in theorganic framework of cometary matter Coal-bound nitrogen also resides principally in hete-rocyclic ring moieties mostly five-memberedpyrrolic rings and six-membered pyridinic rings(Smith et al 1994 Zhang 2001) In pyrolysis stud-ies of tar such nitrogen is observed to escape pre-dominantly in the form of HCN and NH3 (Chenand Niksa 1992 Nelson et al 1992 Ledesma etal 1998) Ammonia is a product of amino com-pounds At the low pressure and high tempera-ture conditions in meteor plasma CN radicalsrather than HCN are stable and ammonia willbreak apart Small hydrocarbon radicals higherthan CN are so excited by their formation thatthey rapidly fission into lower products leavingonly the thermally stable single and double car-bon radicals unless prevented by rapid radiativecooling Carbon atoms are present too but atlower abundance because of thermodynamic con-siderations The carbon atoms react with N2 toform CN At 4400 K CN and N2 are only par-tially dissociated while O2 and CO (and CO2) arefully dissociated leading to high abundances ofCN if the plasma is in local thermodynamic equi-librium (Jenniskens et al 2000a) Only if oxygenplays a role in extracting the nitrogen after abla-tion then NOx compounds are expected to dom-inate Hence CN should be a significant productof the ablation of organic matter unless the or-ganic matter is released as larger compounds

We conclude that the most likely explanationfor the lack of CN molecules in the meteor plasmais that meteoric organic matter is lost in the formof a large range of relatively complex moleculesor perhaps as an amorphous carbon solid withembedded metal as proposed by Brownlee et al

(2002) Until now no evidence has been found ofsuch meteoric soot particles in the Earth envi-ronment and it remains unknown how well theheating conditions in the electron microscopeused in the experiment by Brownlee et al (2002)mimic the actual physical conditions during me-teoroid ablation

If the organic matter survives as large mole-cules this implies that bonds are prevented frombreaking by rapid radiative cooling or becauseheat is carried away by collisional fragments thattransfer much of the kinetic energy to the ambi-ent air The latter mechanism is the more likelyone and comes about because of conservation ofmomentum in a collision The result is that thelower mass fragment will carry most of the ki-netic energy and it will carry proportionallymore kinetic energy as the mass of the mainresidue increases This is an indirect form of thecollisional cooling that facilitates laser desorp-tion a technique widely used in the laboratory tostudy high-molecular-mass polyaromatic hydro-carbons in the gas phase and for detecting com-plex organic compounds in meteorites

Once released from the surface a single colli-sion will cause the molecule to slow down andlag behind relative to the meteoroid The mole-cule then is slowed to thermal speeds in another10 collisions and remains in the warm wake ofthe meteor for 50 collisions Only about 3 ofair collisions with atoms are inelastic (Oumlpik1958) For collisions with molecules that per-centage is higher but only a fraction of collisionswill result in chemical reactions Typical reactionsthat occur in the burning of propene for exam-ple are (eg Marinov et al 1996)

C3H6 1 O R C2H5 1 HCO (3)

C3H6 1 O R CH2CO 1 CH3 1 H (4)

C3H6 1 O R CH3CHCO 1 H 1 H (5)

Hydrogen can react in an early Earth-like atmos-phere with CO2 to form OH radicals which con-tinue to attack the organic matter via reactionssuch as

C3H6 1 OH R C3H6OH (6)

Radical attack can also lead to the growth of car-bon chains by polymerization reactions in therare case that organic compounds meet Hencethe result of chemistry in the meteor plasma is

JENNISKENS ET AL76

likely a product of relatively high molecular massthat is enriched in oxygen (and nitrogen) Suchcompounds are much more interesting for prebi-otic chemistry than the very-high-molecular-mass functional group-poor compounds foundin (micro-) meteorites

CONCLUSIONS

We conclude that complex organic matter inmeteoroids does not fully decompose into di-atomic constituents or atoms during the most in-tense phase of meteor ablation and that meteoricorganics are most likely lost in the form of largemolecular fragments cooled by radiative coolingand the loss of hot molecular fragments Theproduct of chemical interactions in the air plasmain the meteoroidrsquos wake is most likely a large va-riety of molecules and radicals that are chemicallyaltered functional group-enriched molecularfragments and condensation products

This result marks an alternative route to pre-biotic molecules from organic matter contributedto the Earth by particles at the peak of the massndashfrequency distribution of infalling matter At thetime of the origin of life there would have beentypically as much extraterrestrial matter impact-ing Earth in the form of dust as in the form oflarge asteroids and comets (Ceplecha 1992) ifthey arrived on elliptic orbits to have time to de-posit debris Moreover the organic content ofthose meteoroids should have been similar to thatof the impacting parent bodies with the excep-tion of the volatile compounds lost during dustejection This makes the product of meteor abla-tion and subsequent chemistry potentially thedominant source of complex organic moleculeson the early Earth at the time of the origin of life

ACKNOWLEDGMENTS

We thank George Flynn and an anonymous ref-eree for discussions that helped improve this pa-per The cooled CCD slit-less spectrograph wasdeveloped with support of the NASA Ames Re-search Centerrsquos Directorrsquos Discretionary Fundduring preparation of the Leonid MAC missionsWe thank Mike Koop (California Meteor Society)and Gary Palmer for technical support at variousstages in this project The 2001 Leonid MAC mis-sion was sponsored by NASArsquos Astrobiology

ASTID and Planetary Astronomy programs The2001 mission was executed by the USAF 418thFlight Test Squadron of Edwards AFB in Califor-nia with logistic support from the SETI Institute

ABBREVIATIONS

CCD charge coupled device FWHM full widthat half-maximum IDP interplanetary dust parti-cle MAC Multi-Instrument Aircraft CampaignUT universal time

REFERENCES

Abe S Yano H Ebizuka N and Watanabe J-I (2000)First results of high-definition TV spectroscopic obser-vations of the 1999 Leonid meteor shower Earth MoonPlanets 82-83 369ndash377

Anders E (1989) Pre-biotic organic matter from cometsand asteroids Nature 342 255ndash257

Betlem H Jenniskens P Spurny P Docters vanLeeuwen G Miskotte K Ter Kuile CR Zerubin Pand Angelos C (2000) Precise trajectories and orbits ofmeteoroids from the 1999 Leonid meteor storm EarthMoon Planets 82-83 277ndash284

BorovicIuml ka J Stork R and Bocek J (1999) First resultsfrom video spectroscopy of 1998 Leonid meteors Me-teoritics Planet Sci 34 987ndash994

Boyd ID (2000) Computation of atmospheric entry flowabout a Leonid meteoroid Earth Moon Planets 82-8393ndash108

Brownlee DE Joswiak DJ Kress ME Taylor S andBradley J (2002) Survival of carbon in moderately tostrongly heated IDPs and micrometeorites [abstract1786] In 33rd Lunar and Planetary Science Conference LPIContribution No 1109 Lunar and Planetary InstituteHouston

Ceplecha Z (1971) Spectral data on terminal flare andwake of double-station meteor No 38421 (OndrejovApril 21 1963) Bull Astron Inst Czech 22 219ndash304

Ceplecha Z (1992) Earth influx of interplanetary bodiesAstron Astrophys 263 361ndash366

Chen JC and Niksa S (1992) Coal devolatilization dur-ing rapid transient heating 1 Primary devolatilizationEnergy Fuels 6 254ndash264

Chyba CF and Sagan C (1992) Endogenous productionexogenous delivery and impact shock synthesis of or-ganic molecules An inventory for the origins of lifeNature 355 125ndash132

Chyba CF and Sagan C (1997) Comets as a source ofprebiotic organic molecules for the early Earth InComets and the Origin and Evolution of Life edited by PJThomas CF Chyba and CP McKay Springer VerlagNew York pp 147ndash173

Delsemme AH (1991) Nature and history of the organiccompounds in comets an astrophysical view In Astro-

METEORS AND EXOGENOUS ORGANIC MOLECULES 77

INTRODUCTION

REFRACTORY ORGANIC CARBON in extraterrestrialmaterials is a likely source of prebiotic com-

pounds for the origin of life (Oroacute 1961 Chybaand Sagan 1992 1997 Delsemme 1992) Organiccarbon is abundant in the cometary grains probedby the GIOTTO and VEGA spacecraft (Kissel andKrueger 1987 Jessberger and Kissel 1991) mak-ing up some 23 by weight of the comet massfraction and some 66 by weight of dust grainsonce the volatile compounds have evaporated(Greenberg 2000) Primitive asteroids contain or-ganic matter as well up to 12 by weight basedon the carbon content of interplanetary dust par-ticles (IDPs) that are 10 mm in size (Keller et al1995 Flynn et al 2000) Flynn et al (2000) foundfrom C-XANES and Fourier transform infraredspectroscopy that a substantial fraction of thiscarbon is organic Most carbon-rich meteoritescontain less than 3 by weight carbon but thesemeteorites may not be representative of the mostprimitive C-type asteroids

The organic carbon has high molecular massand survives exposure to the vacuum of spaceeven when approaching the Sun as close as Earthrsquosorbit [Tmax 250ndash300 K (Levasseur-Regourd et al2000)] In order for this organic matter to havebeen delivered to an early Earth it had to havesurvived impact with Earthrsquos atmosphere Of allinfalling matter it is believed that less than 8by weight of carbon survives intact in microme-teorites with initial sizes 5ndash50 mm but only inthose that derive from prograde asteroidal orbitsthat approach Earth from behind (Anders 1989)The remainder and the bulk of the mass influxin the form of meteoroids which includes all 50mmndash10 cm-sized meteoroids (Love and Brownlee1993) is ablated in the upper atmosphere duringa phase called a ldquometeorrdquo In particular we willuse the term meteor here to refer to the physicalcondition of rarefied flow that pertains to the in-teraction with Earthrsquos atmosphere of all mete-oroids small enough (less than about 10 cm) notto form a shock wave In that process the organicmolecules are chemically changed in a uniquemanner determined by the process of meteoroidablation and subsequent chemistry in the airplasma in the meteoroidrsquos wake (Jenniskens et al2000a)

This paper investigates the possibility that or-ganic matter carried to Earth via meteors is notnecessarily broken down into atoms or diatomic

molecules and may survive in the form of largermolecules On the primitive Earth this survivingmeteoritic material might have made an impor-tant contribution to the prebiotic organic matterThis idea comes at a time when other investiga-tors are also proposing the survivability of ex-traterrestrial organic matter at relatively hightemperatures Brownlee et al (2002) recently re-ported that refractory carbon in IDPs and mi-crometeorites survive relatively severe heating(1200 K) when exposed to the beam of an elec-tron microscope In these experiments the organicmatter was found to separate from the silicatemelt as a solid carbon-rich material with embed-ded metal It has not been established whetherthis amorphous carbonaceous material producedon heating is elemental or organic although thelatter is more likely In addition Glavin and Bada(1999) have demonstrated that several aminoacids and nucleobases from the Murchison mete-orite survive heating to 1100degC for several sec-onds at reduced pressures

Indeed micrometeorites are found to containsignificant amounts of organic matter that havesurvived the heating process Russell et al (2002)even found the C-H stretch vibration band of or-ganic matter in the debris left in the wake of abright Leonid fireball Based on measurements oforganic matter in micrometeorites found in oceansediments and Antarctic ice Maurette and col-leagues (Maurette 1998 Maurette et al 2000Toppani et al 2001) make the case that solidproducts from meteor ablation raining down tothe surface may have contributed organic matterto the early Earth

In contrast our work pursues the idea that at-mospheric chemistry is essential in creating use-ful prebiotic compounds Jenniskens et al (2000a)provided the first evidence that organic mole-cules may survive the meteor phase with the dis-covery that most light emission and thereforemost high temperature chemistry occurs in acomparatively mild temperature (4300 K) airplasma in almost local thermodynamic equilib-rium just behind the meteoroid (Boyd 2000 Jen-niskens et al 2000a) The product of that chem-istry depends on the time available for reactionsin the high temperature meteor plasma and thebalance between oxygen and nitrogen insertionand polymerization versus carbon extraction andmolecular breakup Reactions with the ablated or-ganic compounds that are too infrequent to occurwith high probability in the allotted time are not

JENNISKENS ET AL68

expected to reach equilibrium After being chem-ically altered the products will continue to reactwhile settling to the Earth surface Before ad-dressing the latter it is necessary to understandthe products that derive from the meteor phaseitself

Meteors represent elusive physical conditionsof rarefied high Mach number (up to 270) flowand non-equilibrium chemistry which makes itdifficult to simulate the conditions in the labora-tory without a better understanding of the phys-ical processes and fate of organic matter in realcases The intense 1998ndash2002 Leonid storms haveoffered an opportunity to obtain data on theseprocesses using modern instruments that wouldotherwise have only a small chance of detectinga meteor If the physical conditions can be de-scribed quantitatively then this will set bound-aries to theoretical models that can provide in-sight into this elusive chemistry

A key constraint for such models is the need topredict the production of the most abundant di-atomic molecules created during decompositionof the complex organic matter When complex or-ganic matter is heated to high temperatures orsputtered by fast ions or energetic photons smallmolecular fragments and atoms are lost from thecomplex organic matter (such as H O H2 O2OH H2O CO CO2 CN and CH4) in a processcalled carbonization (Jenniskens and Russell1993 Fristrom 1995 p 318) Further heating (upto 700 K) leads to additional loss of H and H2 ina process of polymerization which is accompa-nied by the growth of aromatic ring structuresProlonged heating in a confined environmentleads to graphitization which is the stacking ofthose rings into layers to form a crystal (egKoidl et al 1989) Further temperature increaseleads to decomposition by loss of H2 CN CHC2 HCN NH3 and C2H In this latter stage ofdecomposition in the absence of rapid radiativecooling after formation hydrocarbon radicalshigher than C2 rapidly fission into lower-molec-ular-weight products leaving only the thermallystable single and double bonded carbon radicalsCarbon atoms are also present but at low abun-dance due to thermodynamic factors (Fristrom1995 Rairden et al 2000) Upon cooling theselower-molecular -weight products can recon-dense into amorphous carbon to form soot

We propose that the best chance of probing thefate of organic matter in the rarefied high Machnumber flow of small meteoroids is to look for

the production of the CN radical This radical ismost easily detected because of a strong B R Xtransition of low excitation energy with a bandhead at 388 nm CN is a product of reactions in-volving nitrogen embedded in the complex or-ganic matter or interactions between the meteoriccarbon atoms and atmospheric N2 In general ni-trogen is closely associated with the organic mat-ter in IDPs in particular with polyaromatic hy-drocarbons and it is enriched in 15N (Keller et al1995)

Early work by Ceplecha (1971) identified CNemissions in the 212 magnitude flare of a come-tary meteoroid but the identification is doubtfulbecause of the many iron lines in the near-UVOur first efforts at detecting the 388 nm bandhead of CN during the 1999 Leonid Multi-In-strument Aircraft Campaign (MAC) (Jenniskensand Butow 1999 Jenniskens et al 2000b) focusedon slit-less spectroscopy of meteors in first orderin the same spectral region as observed by Ce-plecha using a UV-sensitive intensified chargecoupled device (CCD) camera (Abe et al 2000Rairden et al 2000) The 1999 Leonid storm wasseen under good conditions and the high meteorrates resulted in several bright spectra Our mostprecisely measured spectrum set an upper limitof 1 CN molecule per 3 Fe atoms only about afactor of 2 less than expected if all of the nitrogenin complex organic matter would have been re-leased in the form of CN radicals (Rairden et al2000)

In order to improve upon this result we neededto resolve the underlaying iron emission lines inthe meteor spectrum Here we report results froma high-resolution unintensified slit-less CCDspectrograph that obtained spectra with sufficientspectral resolution to separate the iron linesFrom the new data we conclude that CN pro-duction in meteors is not only low but it isstrongly inhibited

METHODS

The two-stage thermoelectrically cooled slit-less CCD spectrograph for Meteor Astronomyand Astrobiology (ldquoASTROrdquo) was specially de-veloped for the detection of small molecules inthe optical spectra of relatively faint meteors Thedesign provides for a high spectral resolutionwith reasonable detection rates for meteors thatare faint enough to have rarefied flow conditions

METEORS AND EXOGENOUS ORGANIC MOLECULES 69

typical for even smaller (but too faint) 150 mmmeteors The instrument layout is shown in Fig1 An objective grating disperses the light of a me-teor seen at some angle to the viewing directionand a lens projects the spectrum on a cooled CCDdetector A typical result is shown in Fig 2

The detector a Pixelvision camera with a two-stage thermoelectrically cooled 1024 3 1024 pixelback-illuminated SI003AB CCD with 24 3 24 mmpixel size (245 3 245 mm image region) wasused in an unintensified mode to keep as high aspectral resolution as possible The disadvantageof such an unintensified CCD detector is the rel-atively long readout time which is 095 s for 1 34 binning Exposure times needed to be kept to aminimum (01ndash08 s) so as not to have too manyzero-order star images and star spectra overlapthe meteor spectrum Minimum exposure times

were also needed to prevent smearing of thosestar images from Earthrsquos rotation and aircraft mo-tion Though the star background serves to cali-brate the meteor intensity once meteor intensitycalibration was accomplished the star imageswere removed from the spectra by subtraction ofa no-meteor image that was taken in the sameviewing direction and with the same instrumentsetting only seconds after a meteor was detectedNoise residuals of the star images were removedby replacing the residue pixels with low-intensitypixels from neighboring rows A 01 s exposureresulted in a maximum star magnitude that couldbe captured of 1106 In general the limitingmagnitude for meteors was about 157 becauseof their higher angular velocity and meteorsbrighter than magnitude 14 show the most in-tense first-order spectral lines Our best spectra

JENNISKENS ET AL70

FIG 1 Instrument setup on ldquoFISTArdquo and a schematic of the instrumental layout The instrument consists of acooled CCD detector optical system and grating The field of view of the instrument is only 5deg narrow but stretchesin the dispersion direction of the grating for the full range of unvigneted viewing The camera is typically positionedat an angle of 20deg to the aircraftrsquos optical window

probed 22 magnitude meteors at altitudes of90ndash120 km which corresponds to about 1 cm sizemeteoroids in the case of Leonid entry speeds of72 kms These are about 5 magnitudes fainter (a factor of 100 in mass less) than those studiedfrom traditional photographic spectroscopy tech-niques This is small enough to have the requiredrarefied flow field that characterizes meteor con-ditions for the bulk of meteoroid mass influx (Jen-niskens et al 2000a)

The AF-S Nikkor f28300D IF-ED 300ndashmmtelephoto lens collection optics determined thefield of view (5deg) and the short wavelength cut-off of sensitivity which extended down to about360 nm The grating dispersion was chosen in or-der to be able to capture any section of the zerofirst and full second order spectrum (Fig 1) Sec-ond order refers to the integer wavelength dif-ference of 2 for diffractions from two neighbor-ing grating grooves Based on this we chose an11 3 11-cm plane transmission grating 35-54-20-660 with aluminum coating on 12-mm BK7 sub-strate manufactured by Richardson Grating Lab-oratory (Rochester NY) We did not use an orderseparation filter a holder for which is providedby the telephoto lens Without such an order sep-aration filter the near-UV part of the third orderspectrum is also captured which overlaps thesecond order spectrum above 540 nm (3 3 360nm 5 2 3 540 nm) Grating and detector were po-sitioned such that the dispersion occurred alongthe rows of the CCD We measured a dispersionof 540 lmm and blaze wavelength of 698 nm(34deg) This configuration provided a full secondorder spectrum out to about 925 nm The factorygrating specifications suggested a good sensitiv-ity out to 11 mm but we found that the gratingefficiency fell off rapidly above 850 nm Figure 3shows the response curves for the camera cali-

brated for various orders which were measuredby using an approximate point source (an opticalfiber manufactured by Ocean Optics) illuminatedwith a mercury-argon lamp (wavelength calibra-tion) and a tungsten lamp (absolute intensity cal-ibration) The measured efficiencies have beenscaled to the product of the factory responsecurve for the transmission grating the transmis-sion of the camera lens and the CCD quantumefficiency (dashed line Fig 3)

The response curve shown in Fig 3 does notinclude the effects of (a) vigneting (or flatfield-ing) or (b) the geometric correction (a simple co-sine function with angle of incidence) for the ori-entation of the grating relative to the direction ofthe meteor By measuring the response of the in-strument to a homogeneously illuminated flat

METEORS AND EXOGENOUS ORGANIC MOLECULES 71

FIG 2 Second order spectrum of the 090558 UT November 18 2001 Leonid meteor Lines other than thosecaused by iron are marked Brackets indicate emission lines in the overlapping third order

FIG 3 Measured spectral response curves for first(1st) second (2nd) and third (3rd) order The second andthird orders are scaled down for comparison The dashedline shows the theoretical response for the combined fac-tory given grating efficiency CCD quantum efficiencyand camera optics transparency

surface we found that vigneting varied as a func-tion of row and pixel position This allowed us todetermine the flatfielding intensity calibrationcorrection factor ( f ) for a given position of thespectrum on the CCD Two examples are shownin Fig 4 each representing the mean over a bandof CCD rows

The position of the meteor on the sky and thechoice of focus determined what part of the spec-trum was recorded Together with its encasingthe camera left an effective hunting ground of astrip of sky 5deg wide and up to 886deg from the for-ward direction (Fig 1) The wavelength scale wasclose to linear over the full range which spannedabout 13600 pixels

l (nm) 5 15737 1 013382 3 pixel (1)

A small 2 nm residual described by the follow-ing second order polynomial remained

l (observed 2 calculated) (nm) 5 106382 00030227 pix 1 18322e-7 3 pix 3 pix (2)

For each individual spectrum the scale was alsonearly linear and the pixel position of higher-or-der lines was simply a multiple of the pixel po-sition at first order

The highest possible resolution was obtainedonly when the meteor moved nearly perpendic-ular to the dispersion direction A 4-pixel binning

was applied perpendicular to the dispersion di-rection to shorten the readout time but this alsodecreased the resolution if the meteor did notmove perpendicular to the dispersion directionUnfortunately the meteor orientation varies as afunction of the angle on the sky and the leftrightposition relative to the viewing direction in thefield of view Under optimal conditions the fullwidth at half-maximum (FWHM) spectral reso-lution in third order was as low as 18 Aring whichoutperforms photographic spectral analysis tech-niques by a factor of 3

RESULTS

The instrument was deployed during the in-tense 2001 Leonid meteor storm over the conti-nental United States onboard the USAF418thFLTS-operated NKC-135 ldquoFISTArdquo research air-craft during the 2001 Leonid MAC mission at thetime of the intense November 18th 1040 univer-sal time (UT) Leonid storm peak (Jenniskens andRussell 2003)

Figure 2 shows our best result a Leonid witha brief terminal flare which appeared in a west-northwestern direction while flying over LittleRock AR at 090558 UT November 18 (9297W13480N 37000 ft) No meteor trajectory infor-mation is available from a stereoscopic perspec-tive because this was a single-plane mission butthe typical altitude at which such meteors end is93 6 5 km (Jenniskens et al 1998 Betlem et al2000) While most of the spectrum consists of sec-ond order lines of sodium third order lines ofionized calcium are readily identified and verystrong The region shortward of the sodium linesis dominated by intrinsically bright third orderiron lines which were detected despite the factorof 50 lower instrumental sensitivity for third or-der lines (Fig 3)

The spectrum shown in Fig 3 is slightly out offocus at one side of the CCD because of a smallmisalignment of the optics and CCD plane As aresult the spectral resolution increases from aFWHM 5 13 Aring at 370 nm to about 24 Aring at 410nm This does not affect the analysis of the dataother than to require line profile integration priorto comparing relative intensities The sodiumemissions detected above this main portion of thespectrum on the CCD frame are due to the per-sistent afterglow or recombination line emissionfrom sodium atoms released at a time prior to the

JENNISKENS ET AL72

FIG 4 Examples of vigneting in the camera for twodifferent positions of the spectrum on the CCD The firstexample applies to a band near the upper edge of the CCDframe The second example is that valid for the positionof the spectrum in Fig 2

start of the exposure No correction was made forthis (small) contribution to the main spectrum

After summing all relevant rows we foundthat emission lines were on top of a continuumemission identified as the First Positive Band ofN2 observed in second order (Fig 5) No knownsecond or third order molecular band from at-mospheric emissions other than that of CN wasexpected in the region around the CN band headAfter correction for second order instrument re-sponse (Fig 3) we fit the N2 band contour to thewhole spectrum The N2 band contour definedthe background level A small correction for in-strument response to the remaining iron lines wasmade by subtracting the N2 band contour andcorrecting the resulting spectrum in a manner ap-propriate for the third order response (Fig 3)

The CN molecular band

Once instrument and background correctionswere made we focused our attention on the thirdorder spectrum around 388 nm (Fig 6) The newthird order spectrum resolved the weaker ironlines considerably as compared with our earlierfirst order spectra (Rairden et al 2000) The ex-pected Fe line emission spectrum shown as adashed line in Fig 6 was calculated for excita-tion temperatures of 4300 and 4500 K using allneutral iron lines in the NIST Atomic SpectraDatabase Version 20 (NIST Standard ReferenceDatabase 78) A comparison of the observed andcalculated emission lines established their iden-

tity and showed that weak iron lines did not con-tribute to the background continuum It shouldbe noted that second order iron lines are presentelsewhere in the spectrum

As indicated in Fig 6 two iron lines bracket theposition of the CN molecular band Because ofthe absence of any additional faint line emissionsin this region an upper limit on the CN abun-dance could be established Though the CN bandmight have in fact been detected as a small bumpit is at the noise level of the spectrum

The relative abundance of CN molecules versus iron atoms was calculated using theNEQAIR2 model of heated air in thermodynamicequilibrium (Laux 1993 Park et al 1997) The CNband profile for a resolution of 18 Aring FWHM isshown in Fig 6 The band head at 38844 nm isunresolved with its width reflecting the instru-ment profile With ICN equal to the total inte-grated emission intensity of CN in the range375ndash393 nm and IFe equal to the total integratedemission intensity of all third order Fe lines in therange 3813ndash3867 nm we calculated for T 5 4500that IFenFe 5 324 3 10216 Wsr and ICNnCN 5558 3 10217 Wsr where nFe is the total concen-tration of Fe atoms in cm23 and nCN is the totalconcentration of CN moleculescm3 For T 54300 K IFenFe 5 213 3 10216 Wsr and ICNnCN 5 393 3 10217 Wsr To calculate these val-ues we assumed that the emitting levels were in

METEORS AND EXOGENOUS ORGANIC MOLECULES 73

FIG 5 Extracted spectrum of Fig 2 after correction forinstrument response and vigneting The dashed lineshows the contribution from the first positive band of N2in second order

FIG 6 Detail of Fig 5 showing the position of the CNband and the iron lines in third order A model spec-trum at 4300 K of iron atom line emission is shown by adashed line The continuum background is the band con-tour of N2 in second order The thick line is a theoreticalprofile of the CN band for air plasma at 4300 K and in-strumental broadening FWHM 5 18 Aring

Boltzmann equilibrium with the ground state ofFe and CN at the given temperature These areconcentrations per total Fe atoms and total CNmolecules (that is not per molecule of CN in theexcited electronic B state)

From these calculations we found that the ra-tio of the total number of atoms of CN and Fe inthe meteor plasma was nCNnFe 0017 for an ex-citation temperature of 4500 K (or 0019 for anexcitation temperature of 4300 K) The largest un-certainty was caused by the position of the back-ground level However given the good fit of theN2 band contour to the overall spectrum short-ward of 590 nm (Fig 5) a conservative error ofless than 50 places the upper limit at nCNnFe 003

An even more precise upper limit was derivedfrom a Geminid spectrum obtained during the2001 Geminid shower with the same instrumentdeployed from a ground site at Fremont Peak Ob-servatory California (Fig 7) For that spectrumwe calculated for an excitation temperature of4500 K that nCNnFe 0011 (or 0012 at 4300 K)However Geminid meteoroids frequently ap-proach the Sun to within 014 AU which heatsthe meteoroids every 16 years to at least T 5 510K (the temperature of small zodiacal cloud mete-oroids) and possibly as high as 1800 K if thelarger meteoroids are more like a black body(Levasseur-Regourd et al 2000) This is hotenough to change their morphology into morecompact objects and can result in the loss of someor all of the organic matter

The more intense Geminid spectrum confirmsthat no weak second or third order atomic linesof any kind are present at the position of the CNband Hence we conclude that the meteor plasmaresponsible for most of the Leonid meteorrsquos emis-sion contains 1 CN molecule per 30 Fe atomsThis result improves by more than a factor of 10the upper limit of 1 CN per 3 Fe atoms found ear-lier by Rairden et al (2000) These numbers aresignificantly less than expected If all of the ni-trogen in the complex organic matter of the cometdust was decomposed in the form of CN mole-cules this would give nCNnFe 5 079 6 002based on the observed nitrogen abundance ofcomplex organic matter in small dust grains ofcomet 1PHalley (Delsemme 1991) The ob-served abundance is a factor of 26 lower

CN can also be formed by reactions in the me-teor air plasma involving CO2 and N2 in the am-bient atmosphere Though CN formation via suchaerothermochemistry is not well understood inthe case of thermodynamic equilibrium therewould not be enough CN made from this mech-anism to be detectable The calculated CN abun-dance for a 4300 K air plasma in local thermo-dynamic equilibrium at 95 km altitude equals10 3 105 molecules of CNcm3 The 22 magni-tude Leonid meteor studied here has a mass ofabout 044 g of which 0033 g is iron (Delsemme1991) That iron is ablated and distributed over apath length of about 17 km (Betlem et al 2000)in a cylindrical volume of at least a 2-m radius(Boyd 2000) which indicates a mean density ofnFe 5 2 3 109 atomscm3 and an expected nCNnFe of about 5 3 1025 a factor of 6 less than ourdetection limit

DISCUSSION

There are several reasons why the abundanceof CN in the meteor plasma can be a factor of 26lower than expected from the decomposition ofmeteoric organic matter during ablation (1) theabundance of organic matter in the dust of55PTempel-Tuttle is not known (2) the abun-dance of nitrogen in that organic matter is un-known (3) the fate of organic matter in the in-terplanetary medium due to space weathering isnot known (4) the organic matter could havebeen lost at higher altitude (earlier in flight) thanwhere the measurement was made (5) theprocess of ablation could favor the release of ni-

JENNISKENS ET AL74

FIG 7 Detail of the 084127 UT Geminid meteor spec-trum observed on December 13 2001 with symbols andlines the same as in Fig 6

trogen in atomic form or in the form of some othersmall molecules or (6) the nitrogen could survivein the form of larger organic compounds that arenot detected The latter possibility could providea new pathway towards prebiotic compoundsWe will now consider the significance and im-plication of these arguments one by one

Since the abundance of organic matter in thedust of 55PTemplel-Tuttle is not known it ispossible that there was little or no organic mat-ter in the Leonid particles to begin with Conse-quently the absence of nitrogen-bearing organicmatter in the Leonid meteoroids may have re-sulted in our inability to detect CN The onlyother measurement that could have detected ni-trogen-bearing organic matter in the dust (as op-posed to molecules in the gas phase) is in situmass spectroscopy The Leonid parent comet hasnot been visited In the coma of comet 1PHal-ley nitrogen-bearing organic matter was detectedin particles (the CHON particles) analyzed by themass spectrometers on the GIOTTO and VEGAspacecraft Because comet Halley is not the par-ent of the Leonid stream it is recommended thatfuture work focus on the meteoroids of the Ori-onid and eta-Aquarid streams that do originatefrom comet Halley Though bright meteors fromthese showers are much more difficult to observemuch the same result would be expected TheLeonid parent comet is remarkable only becauseit passes so close to Earthrsquos orbit Most impor-tantly since 55PTempel-Tuttle is in a Halley-type orbit and it should originate from the Oortcloud just as 1PHalley there is no reason to ex-pect that it is less rich in organic matter

For the same reason there is no reason to ex-pect that the nitrogen content or composition ofthe organic matter would differ from that of1PHalley Halley dust is as rich in nitrogen thanprimitive meteorites Though the nitrogen con-tent in the organic matter of comet Halley hasbeen questioned because of calibration problemswith the detectors there is no reason to expectthere to be more nitrogen in the less primitive me-teorites The heating and irradiation that alter theorganics in meteorites tend to evolve the organicmatter towards materials less rich in nitrogen andoxygen (Jenniskens et al 1993)

It is unknown how the organic content changeswith dust grain size and exposure to the plane-tary medium hence the fate of the organic mat-ter in the interplanetary medium is not knownHowever we know when the observed Leonids

were ejected from the parent comet in 1767 The090558 UT Leonid meteoroid never came closerto the Sun than 0976 AU and remained in the in-terplanetary medium only for 225 years (sevenorbits) most of that time far from the Sun

Might the organic matter have been lost in theEarthrsquos atmosphere early in the trajectory athigher altitudes when the meteoroid first warmedup to above the evaporation and decompositiontemperature of T 700 K This scenario was pro-posed by Elford et al (1997) who observed thatradar meteors are initially detected at heights upto 140 km They concluded that a more volatilecomponent than stony minerals must be ablatingat those heights The specific case of fast Leonidswas discussed by Steel (1998) who calculated thealtitude of evaporation for a range of small or-ganic compounds with relatively low sublimationtemperatures finding ablation heights in excessof 140 km for some compounds However thelight curve of bright Leonid fireballs above 136km was found to be very flat and starting at about200 km (Spurny et al 2000ab) which is incon-sistent with the light curves calculated by Elfordet al (1997) and Steel (1998) At 200 km the grainsare not warm enough to evaporate the organiccompounds that survived the vacuum of space

Indeed the Leonid studied here shows nosign of differential ablation of minerals in order oftheir volatility The elements Na Fe Mg and Caare lost in the same proportion (Fig 2) exceptfor the last wisp of matter after the catastrophicfragmentation during the terminal flare (a topicof future work) Sodium is at least partiallyfound in more volatile minerals than those con-taining magnesium and iron Though somesmaller Leonid meteoroids are known to losesodium early in their trajectory this is onlyfound in fragile cometary shower meteoroidssmall enough to expose the minerals efficientlyto the impinging air flow (BorovicIuml ka et al 1999)Sodium may be contained in the larger Leonidsat lower altitudes because of the limited timeavailable for inward diffusion of heat and out-ward diffusion of molecules Although the or-ganic and mineral components are separate en-tities they are intimately mixed down to verysmall size scales in comet Halley dust and thediffusion of those volatile compounds out of thegrains (if at all possible) will take time Such con-siderations are not included in the model of Steel(1998) Hence the organic matter is not lost earlyin flight

METEORS AND EXOGENOUS ORGANIC MOLECULES 75

The low CN abundance in the air plasma of the090558 UT Leonid meteoroid cannot be under-stood by merely allowing some fraction of nitro-gen to escape as N NH N2 NH3 NH2 NCONO HNO or HCN etc The configuration of Nin the complex organic matter of meteoroids isnot known However from meteorite and mi-crometeorite studies (where much organic matterhas been altered by aqueous alteration processes)we know that a significant fraction of the nitro-gen is contained in the organic framework (seereferences in Keller et al 1995 Greenberg 2000)Aqueous alteration tends to move nitrogen fromthe organic framework to functional groups suchas -NH In cometary matter no aqueous alter-ation has occurred So if N is in the frameworkin meteorite organics it should also exist in theorganic framework of cometary matter Coal-bound nitrogen also resides principally in hete-rocyclic ring moieties mostly five-memberedpyrrolic rings and six-membered pyridinic rings(Smith et al 1994 Zhang 2001) In pyrolysis stud-ies of tar such nitrogen is observed to escape pre-dominantly in the form of HCN and NH3 (Chenand Niksa 1992 Nelson et al 1992 Ledesma etal 1998) Ammonia is a product of amino com-pounds At the low pressure and high tempera-ture conditions in meteor plasma CN radicalsrather than HCN are stable and ammonia willbreak apart Small hydrocarbon radicals higherthan CN are so excited by their formation thatthey rapidly fission into lower products leavingonly the thermally stable single and double car-bon radicals unless prevented by rapid radiativecooling Carbon atoms are present too but atlower abundance because of thermodynamic con-siderations The carbon atoms react with N2 toform CN At 4400 K CN and N2 are only par-tially dissociated while O2 and CO (and CO2) arefully dissociated leading to high abundances ofCN if the plasma is in local thermodynamic equi-librium (Jenniskens et al 2000a) Only if oxygenplays a role in extracting the nitrogen after abla-tion then NOx compounds are expected to dom-inate Hence CN should be a significant productof the ablation of organic matter unless the or-ganic matter is released as larger compounds

We conclude that the most likely explanationfor the lack of CN molecules in the meteor plasmais that meteoric organic matter is lost in the formof a large range of relatively complex moleculesor perhaps as an amorphous carbon solid withembedded metal as proposed by Brownlee et al

(2002) Until now no evidence has been found ofsuch meteoric soot particles in the Earth envi-ronment and it remains unknown how well theheating conditions in the electron microscopeused in the experiment by Brownlee et al (2002)mimic the actual physical conditions during me-teoroid ablation

If the organic matter survives as large mole-cules this implies that bonds are prevented frombreaking by rapid radiative cooling or becauseheat is carried away by collisional fragments thattransfer much of the kinetic energy to the ambi-ent air The latter mechanism is the more likelyone and comes about because of conservation ofmomentum in a collision The result is that thelower mass fragment will carry most of the ki-netic energy and it will carry proportionallymore kinetic energy as the mass of the mainresidue increases This is an indirect form of thecollisional cooling that facilitates laser desorp-tion a technique widely used in the laboratory tostudy high-molecular-mass polyaromatic hydro-carbons in the gas phase and for detecting com-plex organic compounds in meteorites

Once released from the surface a single colli-sion will cause the molecule to slow down andlag behind relative to the meteoroid The mole-cule then is slowed to thermal speeds in another10 collisions and remains in the warm wake ofthe meteor for 50 collisions Only about 3 ofair collisions with atoms are inelastic (Oumlpik1958) For collisions with molecules that per-centage is higher but only a fraction of collisionswill result in chemical reactions Typical reactionsthat occur in the burning of propene for exam-ple are (eg Marinov et al 1996)

C3H6 1 O R C2H5 1 HCO (3)

C3H6 1 O R CH2CO 1 CH3 1 H (4)

C3H6 1 O R CH3CHCO 1 H 1 H (5)

Hydrogen can react in an early Earth-like atmos-phere with CO2 to form OH radicals which con-tinue to attack the organic matter via reactionssuch as

C3H6 1 OH R C3H6OH (6)

Radical attack can also lead to the growth of car-bon chains by polymerization reactions in therare case that organic compounds meet Hencethe result of chemistry in the meteor plasma is

JENNISKENS ET AL76

likely a product of relatively high molecular massthat is enriched in oxygen (and nitrogen) Suchcompounds are much more interesting for prebi-otic chemistry than the very-high-molecular-mass functional group-poor compounds foundin (micro-) meteorites

CONCLUSIONS

We conclude that complex organic matter inmeteoroids does not fully decompose into di-atomic constituents or atoms during the most in-tense phase of meteor ablation and that meteoricorganics are most likely lost in the form of largemolecular fragments cooled by radiative coolingand the loss of hot molecular fragments Theproduct of chemical interactions in the air plasmain the meteoroidrsquos wake is most likely a large va-riety of molecules and radicals that are chemicallyaltered functional group-enriched molecularfragments and condensation products

This result marks an alternative route to pre-biotic molecules from organic matter contributedto the Earth by particles at the peak of the massndashfrequency distribution of infalling matter At thetime of the origin of life there would have beentypically as much extraterrestrial matter impact-ing Earth in the form of dust as in the form oflarge asteroids and comets (Ceplecha 1992) ifthey arrived on elliptic orbits to have time to de-posit debris Moreover the organic content ofthose meteoroids should have been similar to thatof the impacting parent bodies with the excep-tion of the volatile compounds lost during dustejection This makes the product of meteor abla-tion and subsequent chemistry potentially thedominant source of complex organic moleculeson the early Earth at the time of the origin of life

ACKNOWLEDGMENTS

We thank George Flynn and an anonymous ref-eree for discussions that helped improve this pa-per The cooled CCD slit-less spectrograph wasdeveloped with support of the NASA Ames Re-search Centerrsquos Directorrsquos Discretionary Fundduring preparation of the Leonid MAC missionsWe thank Mike Koop (California Meteor Society)and Gary Palmer for technical support at variousstages in this project The 2001 Leonid MAC mis-sion was sponsored by NASArsquos Astrobiology

ASTID and Planetary Astronomy programs The2001 mission was executed by the USAF 418thFlight Test Squadron of Edwards AFB in Califor-nia with logistic support from the SETI Institute

ABBREVIATIONS

CCD charge coupled device FWHM full widthat half-maximum IDP interplanetary dust parti-cle MAC Multi-Instrument Aircraft CampaignUT universal time

REFERENCES

Abe S Yano H Ebizuka N and Watanabe J-I (2000)First results of high-definition TV spectroscopic obser-vations of the 1999 Leonid meteor shower Earth MoonPlanets 82-83 369ndash377

Anders E (1989) Pre-biotic organic matter from cometsand asteroids Nature 342 255ndash257

Betlem H Jenniskens P Spurny P Docters vanLeeuwen G Miskotte K Ter Kuile CR Zerubin Pand Angelos C (2000) Precise trajectories and orbits ofmeteoroids from the 1999 Leonid meteor storm EarthMoon Planets 82-83 277ndash284

BorovicIuml ka J Stork R and Bocek J (1999) First resultsfrom video spectroscopy of 1998 Leonid meteors Me-teoritics Planet Sci 34 987ndash994

Boyd ID (2000) Computation of atmospheric entry flowabout a Leonid meteoroid Earth Moon Planets 82-8393ndash108

Brownlee DE Joswiak DJ Kress ME Taylor S andBradley J (2002) Survival of carbon in moderately tostrongly heated IDPs and micrometeorites [abstract1786] In 33rd Lunar and Planetary Science Conference LPIContribution No 1109 Lunar and Planetary InstituteHouston

Ceplecha Z (1971) Spectral data on terminal flare andwake of double-station meteor No 38421 (OndrejovApril 21 1963) Bull Astron Inst Czech 22 219ndash304

Ceplecha Z (1992) Earth influx of interplanetary bodiesAstron Astrophys 263 361ndash366

Chen JC and Niksa S (1992) Coal devolatilization dur-ing rapid transient heating 1 Primary devolatilizationEnergy Fuels 6 254ndash264

Chyba CF and Sagan C (1992) Endogenous productionexogenous delivery and impact shock synthesis of or-ganic molecules An inventory for the origins of lifeNature 355 125ndash132

Chyba CF and Sagan C (1997) Comets as a source ofprebiotic organic molecules for the early Earth InComets and the Origin and Evolution of Life edited by PJThomas CF Chyba and CP McKay Springer VerlagNew York pp 147ndash173

Delsemme AH (1991) Nature and history of the organiccompounds in comets an astrophysical view In Astro-

METEORS AND EXOGENOUS ORGANIC MOLECULES 77

expected to reach equilibrium After being chem-ically altered the products will continue to reactwhile settling to the Earth surface Before ad-dressing the latter it is necessary to understandthe products that derive from the meteor phaseitself

Meteors represent elusive physical conditionsof rarefied high Mach number (up to 270) flowand non-equilibrium chemistry which makes itdifficult to simulate the conditions in the labora-tory without a better understanding of the phys-ical processes and fate of organic matter in realcases The intense 1998ndash2002 Leonid storms haveoffered an opportunity to obtain data on theseprocesses using modern instruments that wouldotherwise have only a small chance of detectinga meteor If the physical conditions can be de-scribed quantitatively then this will set bound-aries to theoretical models that can provide in-sight into this elusive chemistry

A key constraint for such models is the need topredict the production of the most abundant di-atomic molecules created during decompositionof the complex organic matter When complex or-ganic matter is heated to high temperatures orsputtered by fast ions or energetic photons smallmolecular fragments and atoms are lost from thecomplex organic matter (such as H O H2 O2OH H2O CO CO2 CN and CH4) in a processcalled carbonization (Jenniskens and Russell1993 Fristrom 1995 p 318) Further heating (upto 700 K) leads to additional loss of H and H2 ina process of polymerization which is accompa-nied by the growth of aromatic ring structuresProlonged heating in a confined environmentleads to graphitization which is the stacking ofthose rings into layers to form a crystal (egKoidl et al 1989) Further temperature increaseleads to decomposition by loss of H2 CN CHC2 HCN NH3 and C2H In this latter stage ofdecomposition in the absence of rapid radiativecooling after formation hydrocarbon radicalshigher than C2 rapidly fission into lower-molec-ular-weight products leaving only the thermallystable single and double bonded carbon radicalsCarbon atoms are also present but at low abun-dance due to thermodynamic factors (Fristrom1995 Rairden et al 2000) Upon cooling theselower-molecular -weight products can recon-dense into amorphous carbon to form soot

We propose that the best chance of probing thefate of organic matter in the rarefied high Machnumber flow of small meteoroids is to look for

the production of the CN radical This radical ismost easily detected because of a strong B R Xtransition of low excitation energy with a bandhead at 388 nm CN is a product of reactions in-volving nitrogen embedded in the complex or-ganic matter or interactions between the meteoriccarbon atoms and atmospheric N2 In general ni-trogen is closely associated with the organic mat-ter in IDPs in particular with polyaromatic hy-drocarbons and it is enriched in 15N (Keller et al1995)

Early work by Ceplecha (1971) identified CNemissions in the 212 magnitude flare of a come-tary meteoroid but the identification is doubtfulbecause of the many iron lines in the near-UVOur first efforts at detecting the 388 nm bandhead of CN during the 1999 Leonid Multi-In-strument Aircraft Campaign (MAC) (Jenniskensand Butow 1999 Jenniskens et al 2000b) focusedon slit-less spectroscopy of meteors in first orderin the same spectral region as observed by Ce-plecha using a UV-sensitive intensified chargecoupled device (CCD) camera (Abe et al 2000Rairden et al 2000) The 1999 Leonid storm wasseen under good conditions and the high meteorrates resulted in several bright spectra Our mostprecisely measured spectrum set an upper limitof 1 CN molecule per 3 Fe atoms only about afactor of 2 less than expected if all of the nitrogenin complex organic matter would have been re-leased in the form of CN radicals (Rairden et al2000)

In order to improve upon this result we neededto resolve the underlaying iron emission lines inthe meteor spectrum Here we report results froma high-resolution unintensified slit-less CCDspectrograph that obtained spectra with sufficientspectral resolution to separate the iron linesFrom the new data we conclude that CN pro-duction in meteors is not only low but it isstrongly inhibited

METHODS

The two-stage thermoelectrically cooled slit-less CCD spectrograph for Meteor Astronomyand Astrobiology (ldquoASTROrdquo) was specially de-veloped for the detection of small molecules inthe optical spectra of relatively faint meteors Thedesign provides for a high spectral resolutionwith reasonable detection rates for meteors thatare faint enough to have rarefied flow conditions

METEORS AND EXOGENOUS ORGANIC MOLECULES 69

typical for even smaller (but too faint) 150 mmmeteors The instrument layout is shown in Fig1 An objective grating disperses the light of a me-teor seen at some angle to the viewing directionand a lens projects the spectrum on a cooled CCDdetector A typical result is shown in Fig 2

The detector a Pixelvision camera with a two-stage thermoelectrically cooled 1024 3 1024 pixelback-illuminated SI003AB CCD with 24 3 24 mmpixel size (245 3 245 mm image region) wasused in an unintensified mode to keep as high aspectral resolution as possible The disadvantageof such an unintensified CCD detector is the rel-atively long readout time which is 095 s for 1 34 binning Exposure times needed to be kept to aminimum (01ndash08 s) so as not to have too manyzero-order star images and star spectra overlapthe meteor spectrum Minimum exposure times

were also needed to prevent smearing of thosestar images from Earthrsquos rotation and aircraft mo-tion Though the star background serves to cali-brate the meteor intensity once meteor intensitycalibration was accomplished the star imageswere removed from the spectra by subtraction ofa no-meteor image that was taken in the sameviewing direction and with the same instrumentsetting only seconds after a meteor was detectedNoise residuals of the star images were removedby replacing the residue pixels with low-intensitypixels from neighboring rows A 01 s exposureresulted in a maximum star magnitude that couldbe captured of 1106 In general the limitingmagnitude for meteors was about 157 becauseof their higher angular velocity and meteorsbrighter than magnitude 14 show the most in-tense first-order spectral lines Our best spectra

JENNISKENS ET AL70

FIG 1 Instrument setup on ldquoFISTArdquo and a schematic of the instrumental layout The instrument consists of acooled CCD detector optical system and grating The field of view of the instrument is only 5deg narrow but stretchesin the dispersion direction of the grating for the full range of unvigneted viewing The camera is typically positionedat an angle of 20deg to the aircraftrsquos optical window

probed 22 magnitude meteors at altitudes of90ndash120 km which corresponds to about 1 cm sizemeteoroids in the case of Leonid entry speeds of72 kms These are about 5 magnitudes fainter (a factor of 100 in mass less) than those studiedfrom traditional photographic spectroscopy tech-niques This is small enough to have the requiredrarefied flow field that characterizes meteor con-ditions for the bulk of meteoroid mass influx (Jen-niskens et al 2000a)

The AF-S Nikkor f28300D IF-ED 300ndashmmtelephoto lens collection optics determined thefield of view (5deg) and the short wavelength cut-off of sensitivity which extended down to about360 nm The grating dispersion was chosen in or-der to be able to capture any section of the zerofirst and full second order spectrum (Fig 1) Sec-ond order refers to the integer wavelength dif-ference of 2 for diffractions from two neighbor-ing grating grooves Based on this we chose an11 3 11-cm plane transmission grating 35-54-20-660 with aluminum coating on 12-mm BK7 sub-strate manufactured by Richardson Grating Lab-oratory (Rochester NY) We did not use an orderseparation filter a holder for which is providedby the telephoto lens Without such an order sep-aration filter the near-UV part of the third orderspectrum is also captured which overlaps thesecond order spectrum above 540 nm (3 3 360nm 5 2 3 540 nm) Grating and detector were po-sitioned such that the dispersion occurred alongthe rows of the CCD We measured a dispersionof 540 lmm and blaze wavelength of 698 nm(34deg) This configuration provided a full secondorder spectrum out to about 925 nm The factorygrating specifications suggested a good sensitiv-ity out to 11 mm but we found that the gratingefficiency fell off rapidly above 850 nm Figure 3shows the response curves for the camera cali-

brated for various orders which were measuredby using an approximate point source (an opticalfiber manufactured by Ocean Optics) illuminatedwith a mercury-argon lamp (wavelength calibra-tion) and a tungsten lamp (absolute intensity cal-ibration) The measured efficiencies have beenscaled to the product of the factory responsecurve for the transmission grating the transmis-sion of the camera lens and the CCD quantumefficiency (dashed line Fig 3)

The response curve shown in Fig 3 does notinclude the effects of (a) vigneting (or flatfield-ing) or (b) the geometric correction (a simple co-sine function with angle of incidence) for the ori-entation of the grating relative to the direction ofthe meteor By measuring the response of the in-strument to a homogeneously illuminated flat

METEORS AND EXOGENOUS ORGANIC MOLECULES 71

FIG 2 Second order spectrum of the 090558 UT November 18 2001 Leonid meteor Lines other than thosecaused by iron are marked Brackets indicate emission lines in the overlapping third order

FIG 3 Measured spectral response curves for first(1st) second (2nd) and third (3rd) order The second andthird orders are scaled down for comparison The dashedline shows the theoretical response for the combined fac-tory given grating efficiency CCD quantum efficiencyand camera optics transparency

surface we found that vigneting varied as a func-tion of row and pixel position This allowed us todetermine the flatfielding intensity calibrationcorrection factor ( f ) for a given position of thespectrum on the CCD Two examples are shownin Fig 4 each representing the mean over a bandof CCD rows

The position of the meteor on the sky and thechoice of focus determined what part of the spec-trum was recorded Together with its encasingthe camera left an effective hunting ground of astrip of sky 5deg wide and up to 886deg from the for-ward direction (Fig 1) The wavelength scale wasclose to linear over the full range which spannedabout 13600 pixels

l (nm) 5 15737 1 013382 3 pixel (1)

A small 2 nm residual described by the follow-ing second order polynomial remained

l (observed 2 calculated) (nm) 5 106382 00030227 pix 1 18322e-7 3 pix 3 pix (2)

For each individual spectrum the scale was alsonearly linear and the pixel position of higher-or-der lines was simply a multiple of the pixel po-sition at first order

The highest possible resolution was obtainedonly when the meteor moved nearly perpendic-ular to the dispersion direction A 4-pixel binning

was applied perpendicular to the dispersion di-rection to shorten the readout time but this alsodecreased the resolution if the meteor did notmove perpendicular to the dispersion directionUnfortunately the meteor orientation varies as afunction of the angle on the sky and the leftrightposition relative to the viewing direction in thefield of view Under optimal conditions the fullwidth at half-maximum (FWHM) spectral reso-lution in third order was as low as 18 Aring whichoutperforms photographic spectral analysis tech-niques by a factor of 3

RESULTS

The instrument was deployed during the in-tense 2001 Leonid meteor storm over the conti-nental United States onboard the USAF418thFLTS-operated NKC-135 ldquoFISTArdquo research air-craft during the 2001 Leonid MAC mission at thetime of the intense November 18th 1040 univer-sal time (UT) Leonid storm peak (Jenniskens andRussell 2003)

Figure 2 shows our best result a Leonid witha brief terminal flare which appeared in a west-northwestern direction while flying over LittleRock AR at 090558 UT November 18 (9297W13480N 37000 ft) No meteor trajectory infor-mation is available from a stereoscopic perspec-tive because this was a single-plane mission butthe typical altitude at which such meteors end is93 6 5 km (Jenniskens et al 1998 Betlem et al2000) While most of the spectrum consists of sec-ond order lines of sodium third order lines ofionized calcium are readily identified and verystrong The region shortward of the sodium linesis dominated by intrinsically bright third orderiron lines which were detected despite the factorof 50 lower instrumental sensitivity for third or-der lines (Fig 3)

The spectrum shown in Fig 3 is slightly out offocus at one side of the CCD because of a smallmisalignment of the optics and CCD plane As aresult the spectral resolution increases from aFWHM 5 13 Aring at 370 nm to about 24 Aring at 410nm This does not affect the analysis of the dataother than to require line profile integration priorto comparing relative intensities The sodiumemissions detected above this main portion of thespectrum on the CCD frame are due to the per-sistent afterglow or recombination line emissionfrom sodium atoms released at a time prior to the

JENNISKENS ET AL72

FIG 4 Examples of vigneting in the camera for twodifferent positions of the spectrum on the CCD The firstexample applies to a band near the upper edge of the CCDframe The second example is that valid for the positionof the spectrum in Fig 2

start of the exposure No correction was made forthis (small) contribution to the main spectrum

After summing all relevant rows we foundthat emission lines were on top of a continuumemission identified as the First Positive Band ofN2 observed in second order (Fig 5) No knownsecond or third order molecular band from at-mospheric emissions other than that of CN wasexpected in the region around the CN band headAfter correction for second order instrument re-sponse (Fig 3) we fit the N2 band contour to thewhole spectrum The N2 band contour definedthe background level A small correction for in-strument response to the remaining iron lines wasmade by subtracting the N2 band contour andcorrecting the resulting spectrum in a manner ap-propriate for the third order response (Fig 3)

The CN molecular band

Once instrument and background correctionswere made we focused our attention on the thirdorder spectrum around 388 nm (Fig 6) The newthird order spectrum resolved the weaker ironlines considerably as compared with our earlierfirst order spectra (Rairden et al 2000) The ex-pected Fe line emission spectrum shown as adashed line in Fig 6 was calculated for excita-tion temperatures of 4300 and 4500 K using allneutral iron lines in the NIST Atomic SpectraDatabase Version 20 (NIST Standard ReferenceDatabase 78) A comparison of the observed andcalculated emission lines established their iden-

tity and showed that weak iron lines did not con-tribute to the background continuum It shouldbe noted that second order iron lines are presentelsewhere in the spectrum

As indicated in Fig 6 two iron lines bracket theposition of the CN molecular band Because ofthe absence of any additional faint line emissionsin this region an upper limit on the CN abun-dance could be established Though the CN bandmight have in fact been detected as a small bumpit is at the noise level of the spectrum

The relative abundance of CN molecules versus iron atoms was calculated using theNEQAIR2 model of heated air in thermodynamicequilibrium (Laux 1993 Park et al 1997) The CNband profile for a resolution of 18 Aring FWHM isshown in Fig 6 The band head at 38844 nm isunresolved with its width reflecting the instru-ment profile With ICN equal to the total inte-grated emission intensity of CN in the range375ndash393 nm and IFe equal to the total integratedemission intensity of all third order Fe lines in therange 3813ndash3867 nm we calculated for T 5 4500that IFenFe 5 324 3 10216 Wsr and ICNnCN 5558 3 10217 Wsr where nFe is the total concen-tration of Fe atoms in cm23 and nCN is the totalconcentration of CN moleculescm3 For T 54300 K IFenFe 5 213 3 10216 Wsr and ICNnCN 5 393 3 10217 Wsr To calculate these val-ues we assumed that the emitting levels were in

METEORS AND EXOGENOUS ORGANIC MOLECULES 73

FIG 5 Extracted spectrum of Fig 2 after correction forinstrument response and vigneting The dashed lineshows the contribution from the first positive band of N2in second order

FIG 6 Detail of Fig 5 showing the position of the CNband and the iron lines in third order A model spec-trum at 4300 K of iron atom line emission is shown by adashed line The continuum background is the band con-tour of N2 in second order The thick line is a theoreticalprofile of the CN band for air plasma at 4300 K and in-strumental broadening FWHM 5 18 Aring

Boltzmann equilibrium with the ground state ofFe and CN at the given temperature These areconcentrations per total Fe atoms and total CNmolecules (that is not per molecule of CN in theexcited electronic B state)

From these calculations we found that the ra-tio of the total number of atoms of CN and Fe inthe meteor plasma was nCNnFe 0017 for an ex-citation temperature of 4500 K (or 0019 for anexcitation temperature of 4300 K) The largest un-certainty was caused by the position of the back-ground level However given the good fit of theN2 band contour to the overall spectrum short-ward of 590 nm (Fig 5) a conservative error ofless than 50 places the upper limit at nCNnFe 003

An even more precise upper limit was derivedfrom a Geminid spectrum obtained during the2001 Geminid shower with the same instrumentdeployed from a ground site at Fremont Peak Ob-servatory California (Fig 7) For that spectrumwe calculated for an excitation temperature of4500 K that nCNnFe 0011 (or 0012 at 4300 K)However Geminid meteoroids frequently ap-proach the Sun to within 014 AU which heatsthe meteoroids every 16 years to at least T 5 510K (the temperature of small zodiacal cloud mete-oroids) and possibly as high as 1800 K if thelarger meteoroids are more like a black body(Levasseur-Regourd et al 2000) This is hotenough to change their morphology into morecompact objects and can result in the loss of someor all of the organic matter

The more intense Geminid spectrum confirmsthat no weak second or third order atomic linesof any kind are present at the position of the CNband Hence we conclude that the meteor plasmaresponsible for most of the Leonid meteorrsquos emis-sion contains 1 CN molecule per 30 Fe atomsThis result improves by more than a factor of 10the upper limit of 1 CN per 3 Fe atoms found ear-lier by Rairden et al (2000) These numbers aresignificantly less than expected If all of the ni-trogen in the complex organic matter of the cometdust was decomposed in the form of CN mole-cules this would give nCNnFe 5 079 6 002based on the observed nitrogen abundance ofcomplex organic matter in small dust grains ofcomet 1PHalley (Delsemme 1991) The ob-served abundance is a factor of 26 lower

CN can also be formed by reactions in the me-teor air plasma involving CO2 and N2 in the am-bient atmosphere Though CN formation via suchaerothermochemistry is not well understood inthe case of thermodynamic equilibrium therewould not be enough CN made from this mech-anism to be detectable The calculated CN abun-dance for a 4300 K air plasma in local thermo-dynamic equilibrium at 95 km altitude equals10 3 105 molecules of CNcm3 The 22 magni-tude Leonid meteor studied here has a mass ofabout 044 g of which 0033 g is iron (Delsemme1991) That iron is ablated and distributed over apath length of about 17 km (Betlem et al 2000)in a cylindrical volume of at least a 2-m radius(Boyd 2000) which indicates a mean density ofnFe 5 2 3 109 atomscm3 and an expected nCNnFe of about 5 3 1025 a factor of 6 less than ourdetection limit

DISCUSSION

There are several reasons why the abundanceof CN in the meteor plasma can be a factor of 26lower than expected from the decomposition ofmeteoric organic matter during ablation (1) theabundance of organic matter in the dust of55PTempel-Tuttle is not known (2) the abun-dance of nitrogen in that organic matter is un-known (3) the fate of organic matter in the in-terplanetary medium due to space weathering isnot known (4) the organic matter could havebeen lost at higher altitude (earlier in flight) thanwhere the measurement was made (5) theprocess of ablation could favor the release of ni-

JENNISKENS ET AL74

FIG 7 Detail of the 084127 UT Geminid meteor spec-trum observed on December 13 2001 with symbols andlines the same as in Fig 6

trogen in atomic form or in the form of some othersmall molecules or (6) the nitrogen could survivein the form of larger organic compounds that arenot detected The latter possibility could providea new pathway towards prebiotic compoundsWe will now consider the significance and im-plication of these arguments one by one

Since the abundance of organic matter in thedust of 55PTemplel-Tuttle is not known it ispossible that there was little or no organic mat-ter in the Leonid particles to begin with Conse-quently the absence of nitrogen-bearing organicmatter in the Leonid meteoroids may have re-sulted in our inability to detect CN The onlyother measurement that could have detected ni-trogen-bearing organic matter in the dust (as op-posed to molecules in the gas phase) is in situmass spectroscopy The Leonid parent comet hasnot been visited In the coma of comet 1PHal-ley nitrogen-bearing organic matter was detectedin particles (the CHON particles) analyzed by themass spectrometers on the GIOTTO and VEGAspacecraft Because comet Halley is not the par-ent of the Leonid stream it is recommended thatfuture work focus on the meteoroids of the Ori-onid and eta-Aquarid streams that do originatefrom comet Halley Though bright meteors fromthese showers are much more difficult to observemuch the same result would be expected TheLeonid parent comet is remarkable only becauseit passes so close to Earthrsquos orbit Most impor-tantly since 55PTempel-Tuttle is in a Halley-type orbit and it should originate from the Oortcloud just as 1PHalley there is no reason to ex-pect that it is less rich in organic matter

For the same reason there is no reason to ex-pect that the nitrogen content or composition ofthe organic matter would differ from that of1PHalley Halley dust is as rich in nitrogen thanprimitive meteorites Though the nitrogen con-tent in the organic matter of comet Halley hasbeen questioned because of calibration problemswith the detectors there is no reason to expectthere to be more nitrogen in the less primitive me-teorites The heating and irradiation that alter theorganics in meteorites tend to evolve the organicmatter towards materials less rich in nitrogen andoxygen (Jenniskens et al 1993)

It is unknown how the organic content changeswith dust grain size and exposure to the plane-tary medium hence the fate of the organic mat-ter in the interplanetary medium is not knownHowever we know when the observed Leonids

were ejected from the parent comet in 1767 The090558 UT Leonid meteoroid never came closerto the Sun than 0976 AU and remained in the in-terplanetary medium only for 225 years (sevenorbits) most of that time far from the Sun

Might the organic matter have been lost in theEarthrsquos atmosphere early in the trajectory athigher altitudes when the meteoroid first warmedup to above the evaporation and decompositiontemperature of T 700 K This scenario was pro-posed by Elford et al (1997) who observed thatradar meteors are initially detected at heights upto 140 km They concluded that a more volatilecomponent than stony minerals must be ablatingat those heights The specific case of fast Leonidswas discussed by Steel (1998) who calculated thealtitude of evaporation for a range of small or-ganic compounds with relatively low sublimationtemperatures finding ablation heights in excessof 140 km for some compounds However thelight curve of bright Leonid fireballs above 136km was found to be very flat and starting at about200 km (Spurny et al 2000ab) which is incon-sistent with the light curves calculated by Elfordet al (1997) and Steel (1998) At 200 km the grainsare not warm enough to evaporate the organiccompounds that survived the vacuum of space

Indeed the Leonid studied here shows nosign of differential ablation of minerals in order oftheir volatility The elements Na Fe Mg and Caare lost in the same proportion (Fig 2) exceptfor the last wisp of matter after the catastrophicfragmentation during the terminal flare (a topicof future work) Sodium is at least partiallyfound in more volatile minerals than those con-taining magnesium and iron Though somesmaller Leonid meteoroids are known to losesodium early in their trajectory this is onlyfound in fragile cometary shower meteoroidssmall enough to expose the minerals efficientlyto the impinging air flow (BorovicIuml ka et al 1999)Sodium may be contained in the larger Leonidsat lower altitudes because of the limited timeavailable for inward diffusion of heat and out-ward diffusion of molecules Although the or-ganic and mineral components are separate en-tities they are intimately mixed down to verysmall size scales in comet Halley dust and thediffusion of those volatile compounds out of thegrains (if at all possible) will take time Such con-siderations are not included in the model of Steel(1998) Hence the organic matter is not lost earlyin flight

METEORS AND EXOGENOUS ORGANIC MOLECULES 75

The low CN abundance in the air plasma of the090558 UT Leonid meteoroid cannot be under-stood by merely allowing some fraction of nitro-gen to escape as N NH N2 NH3 NH2 NCONO HNO or HCN etc The configuration of Nin the complex organic matter of meteoroids isnot known However from meteorite and mi-crometeorite studies (where much organic matterhas been altered by aqueous alteration processes)we know that a significant fraction of the nitro-gen is contained in the organic framework (seereferences in Keller et al 1995 Greenberg 2000)Aqueous alteration tends to move nitrogen fromthe organic framework to functional groups suchas -NH In cometary matter no aqueous alter-ation has occurred So if N is in the frameworkin meteorite organics it should also exist in theorganic framework of cometary matter Coal-bound nitrogen also resides principally in hete-rocyclic ring moieties mostly five-memberedpyrrolic rings and six-membered pyridinic rings(Smith et al 1994 Zhang 2001) In pyrolysis stud-ies of tar such nitrogen is observed to escape pre-dominantly in the form of HCN and NH3 (Chenand Niksa 1992 Nelson et al 1992 Ledesma etal 1998) Ammonia is a product of amino com-pounds At the low pressure and high tempera-ture conditions in meteor plasma CN radicalsrather than HCN are stable and ammonia willbreak apart Small hydrocarbon radicals higherthan CN are so excited by their formation thatthey rapidly fission into lower products leavingonly the thermally stable single and double car-bon radicals unless prevented by rapid radiativecooling Carbon atoms are present too but atlower abundance because of thermodynamic con-siderations The carbon atoms react with N2 toform CN At 4400 K CN and N2 are only par-tially dissociated while O2 and CO (and CO2) arefully dissociated leading to high abundances ofCN if the plasma is in local thermodynamic equi-librium (Jenniskens et al 2000a) Only if oxygenplays a role in extracting the nitrogen after abla-tion then NOx compounds are expected to dom-inate Hence CN should be a significant productof the ablation of organic matter unless the or-ganic matter is released as larger compounds

We conclude that the most likely explanationfor the lack of CN molecules in the meteor plasmais that meteoric organic matter is lost in the formof a large range of relatively complex moleculesor perhaps as an amorphous carbon solid withembedded metal as proposed by Brownlee et al

(2002) Until now no evidence has been found ofsuch meteoric soot particles in the Earth envi-ronment and it remains unknown how well theheating conditions in the electron microscopeused in the experiment by Brownlee et al (2002)mimic the actual physical conditions during me-teoroid ablation

If the organic matter survives as large mole-cules this implies that bonds are prevented frombreaking by rapid radiative cooling or becauseheat is carried away by collisional fragments thattransfer much of the kinetic energy to the ambi-ent air The latter mechanism is the more likelyone and comes about because of conservation ofmomentum in a collision The result is that thelower mass fragment will carry most of the ki-netic energy and it will carry proportionallymore kinetic energy as the mass of the mainresidue increases This is an indirect form of thecollisional cooling that facilitates laser desorp-tion a technique widely used in the laboratory tostudy high-molecular-mass polyaromatic hydro-carbons in the gas phase and for detecting com-plex organic compounds in meteorites

Once released from the surface a single colli-sion will cause the molecule to slow down andlag behind relative to the meteoroid The mole-cule then is slowed to thermal speeds in another10 collisions and remains in the warm wake ofthe meteor for 50 collisions Only about 3 ofair collisions with atoms are inelastic (Oumlpik1958) For collisions with molecules that per-centage is higher but only a fraction of collisionswill result in chemical reactions Typical reactionsthat occur in the burning of propene for exam-ple are (eg Marinov et al 1996)

C3H6 1 O R C2H5 1 HCO (3)

C3H6 1 O R CH2CO 1 CH3 1 H (4)

C3H6 1 O R CH3CHCO 1 H 1 H (5)

Hydrogen can react in an early Earth-like atmos-phere with CO2 to form OH radicals which con-tinue to attack the organic matter via reactionssuch as

C3H6 1 OH R C3H6OH (6)

Radical attack can also lead to the growth of car-bon chains by polymerization reactions in therare case that organic compounds meet Hencethe result of chemistry in the meteor plasma is

JENNISKENS ET AL76

likely a product of relatively high molecular massthat is enriched in oxygen (and nitrogen) Suchcompounds are much more interesting for prebi-otic chemistry than the very-high-molecular-mass functional group-poor compounds foundin (micro-) meteorites

CONCLUSIONS

We conclude that complex organic matter inmeteoroids does not fully decompose into di-atomic constituents or atoms during the most in-tense phase of meteor ablation and that meteoricorganics are most likely lost in the form of largemolecular fragments cooled by radiative coolingand the loss of hot molecular fragments Theproduct of chemical interactions in the air plasmain the meteoroidrsquos wake is most likely a large va-riety of molecules and radicals that are chemicallyaltered functional group-enriched molecularfragments and condensation products

This result marks an alternative route to pre-biotic molecules from organic matter contributedto the Earth by particles at the peak of the massndashfrequency distribution of infalling matter At thetime of the origin of life there would have beentypically as much extraterrestrial matter impact-ing Earth in the form of dust as in the form oflarge asteroids and comets (Ceplecha 1992) ifthey arrived on elliptic orbits to have time to de-posit debris Moreover the organic content ofthose meteoroids should have been similar to thatof the impacting parent bodies with the excep-tion of the volatile compounds lost during dustejection This makes the product of meteor abla-tion and subsequent chemistry potentially thedominant source of complex organic moleculeson the early Earth at the time of the origin of life

ACKNOWLEDGMENTS

We thank George Flynn and an anonymous ref-eree for discussions that helped improve this pa-per The cooled CCD slit-less spectrograph wasdeveloped with support of the NASA Ames Re-search Centerrsquos Directorrsquos Discretionary Fundduring preparation of the Leonid MAC missionsWe thank Mike Koop (California Meteor Society)and Gary Palmer for technical support at variousstages in this project The 2001 Leonid MAC mis-sion was sponsored by NASArsquos Astrobiology

ASTID and Planetary Astronomy programs The2001 mission was executed by the USAF 418thFlight Test Squadron of Edwards AFB in Califor-nia with logistic support from the SETI Institute

ABBREVIATIONS

CCD charge coupled device FWHM full widthat half-maximum IDP interplanetary dust parti-cle MAC Multi-Instrument Aircraft CampaignUT universal time

REFERENCES

Abe S Yano H Ebizuka N and Watanabe J-I (2000)First results of high-definition TV spectroscopic obser-vations of the 1999 Leonid meteor shower Earth MoonPlanets 82-83 369ndash377

Anders E (1989) Pre-biotic organic matter from cometsand asteroids Nature 342 255ndash257

Betlem H Jenniskens P Spurny P Docters vanLeeuwen G Miskotte K Ter Kuile CR Zerubin Pand Angelos C (2000) Precise trajectories and orbits ofmeteoroids from the 1999 Leonid meteor storm EarthMoon Planets 82-83 277ndash284

BorovicIuml ka J Stork R and Bocek J (1999) First resultsfrom video spectroscopy of 1998 Leonid meteors Me-teoritics Planet Sci 34 987ndash994

Boyd ID (2000) Computation of atmospheric entry flowabout a Leonid meteoroid Earth Moon Planets 82-8393ndash108

Brownlee DE Joswiak DJ Kress ME Taylor S andBradley J (2002) Survival of carbon in moderately tostrongly heated IDPs and micrometeorites [abstract1786] In 33rd Lunar and Planetary Science Conference LPIContribution No 1109 Lunar and Planetary InstituteHouston

Ceplecha Z (1971) Spectral data on terminal flare andwake of double-station meteor No 38421 (OndrejovApril 21 1963) Bull Astron Inst Czech 22 219ndash304

Ceplecha Z (1992) Earth influx of interplanetary bodiesAstron Astrophys 263 361ndash366

Chen JC and Niksa S (1992) Coal devolatilization dur-ing rapid transient heating 1 Primary devolatilizationEnergy Fuels 6 254ndash264

Chyba CF and Sagan C (1992) Endogenous productionexogenous delivery and impact shock synthesis of or-ganic molecules An inventory for the origins of lifeNature 355 125ndash132

Chyba CF and Sagan C (1997) Comets as a source ofprebiotic organic molecules for the early Earth InComets and the Origin and Evolution of Life edited by PJThomas CF Chyba and CP McKay Springer VerlagNew York pp 147ndash173

Delsemme AH (1991) Nature and history of the organiccompounds in comets an astrophysical view In Astro-

METEORS AND EXOGENOUS ORGANIC MOLECULES 77

typical for even smaller (but too faint) 150 mmmeteors The instrument layout is shown in Fig1 An objective grating disperses the light of a me-teor seen at some angle to the viewing directionand a lens projects the spectrum on a cooled CCDdetector A typical result is shown in Fig 2

The detector a Pixelvision camera with a two-stage thermoelectrically cooled 1024 3 1024 pixelback-illuminated SI003AB CCD with 24 3 24 mmpixel size (245 3 245 mm image region) wasused in an unintensified mode to keep as high aspectral resolution as possible The disadvantageof such an unintensified CCD detector is the rel-atively long readout time which is 095 s for 1 34 binning Exposure times needed to be kept to aminimum (01ndash08 s) so as not to have too manyzero-order star images and star spectra overlapthe meteor spectrum Minimum exposure times

were also needed to prevent smearing of thosestar images from Earthrsquos rotation and aircraft mo-tion Though the star background serves to cali-brate the meteor intensity once meteor intensitycalibration was accomplished the star imageswere removed from the spectra by subtraction ofa no-meteor image that was taken in the sameviewing direction and with the same instrumentsetting only seconds after a meteor was detectedNoise residuals of the star images were removedby replacing the residue pixels with low-intensitypixels from neighboring rows A 01 s exposureresulted in a maximum star magnitude that couldbe captured of 1106 In general the limitingmagnitude for meteors was about 157 becauseof their higher angular velocity and meteorsbrighter than magnitude 14 show the most in-tense first-order spectral lines Our best spectra

JENNISKENS ET AL70

FIG 1 Instrument setup on ldquoFISTArdquo and a schematic of the instrumental layout The instrument consists of acooled CCD detector optical system and grating The field of view of the instrument is only 5deg narrow but stretchesin the dispersion direction of the grating for the full range of unvigneted viewing The camera is typically positionedat an angle of 20deg to the aircraftrsquos optical window

probed 22 magnitude meteors at altitudes of90ndash120 km which corresponds to about 1 cm sizemeteoroids in the case of Leonid entry speeds of72 kms These are about 5 magnitudes fainter (a factor of 100 in mass less) than those studiedfrom traditional photographic spectroscopy tech-niques This is small enough to have the requiredrarefied flow field that characterizes meteor con-ditions for the bulk of meteoroid mass influx (Jen-niskens et al 2000a)

The AF-S Nikkor f28300D IF-ED 300ndashmmtelephoto lens collection optics determined thefield of view (5deg) and the short wavelength cut-off of sensitivity which extended down to about360 nm The grating dispersion was chosen in or-der to be able to capture any section of the zerofirst and full second order spectrum (Fig 1) Sec-ond order refers to the integer wavelength dif-ference of 2 for diffractions from two neighbor-ing grating grooves Based on this we chose an11 3 11-cm plane transmission grating 35-54-20-660 with aluminum coating on 12-mm BK7 sub-strate manufactured by Richardson Grating Lab-oratory (Rochester NY) We did not use an orderseparation filter a holder for which is providedby the telephoto lens Without such an order sep-aration filter the near-UV part of the third orderspectrum is also captured which overlaps thesecond order spectrum above 540 nm (3 3 360nm 5 2 3 540 nm) Grating and detector were po-sitioned such that the dispersion occurred alongthe rows of the CCD We measured a dispersionof 540 lmm and blaze wavelength of 698 nm(34deg) This configuration provided a full secondorder spectrum out to about 925 nm The factorygrating specifications suggested a good sensitiv-ity out to 11 mm but we found that the gratingefficiency fell off rapidly above 850 nm Figure 3shows the response curves for the camera cali-

brated for various orders which were measuredby using an approximate point source (an opticalfiber manufactured by Ocean Optics) illuminatedwith a mercury-argon lamp (wavelength calibra-tion) and a tungsten lamp (absolute intensity cal-ibration) The measured efficiencies have beenscaled to the product of the factory responsecurve for the transmission grating the transmis-sion of the camera lens and the CCD quantumefficiency (dashed line Fig 3)

The response curve shown in Fig 3 does notinclude the effects of (a) vigneting (or flatfield-ing) or (b) the geometric correction (a simple co-sine function with angle of incidence) for the ori-entation of the grating relative to the direction ofthe meteor By measuring the response of the in-strument to a homogeneously illuminated flat

METEORS AND EXOGENOUS ORGANIC MOLECULES 71

FIG 2 Second order spectrum of the 090558 UT November 18 2001 Leonid meteor Lines other than thosecaused by iron are marked Brackets indicate emission lines in the overlapping third order

FIG 3 Measured spectral response curves for first(1st) second (2nd) and third (3rd) order The second andthird orders are scaled down for comparison The dashedline shows the theoretical response for the combined fac-tory given grating efficiency CCD quantum efficiencyand camera optics transparency

surface we found that vigneting varied as a func-tion of row and pixel position This allowed us todetermine the flatfielding intensity calibrationcorrection factor ( f ) for a given position of thespectrum on the CCD Two examples are shownin Fig 4 each representing the mean over a bandof CCD rows

The position of the meteor on the sky and thechoice of focus determined what part of the spec-trum was recorded Together with its encasingthe camera left an effective hunting ground of astrip of sky 5deg wide and up to 886deg from the for-ward direction (Fig 1) The wavelength scale wasclose to linear over the full range which spannedabout 13600 pixels

l (nm) 5 15737 1 013382 3 pixel (1)

A small 2 nm residual described by the follow-ing second order polynomial remained

l (observed 2 calculated) (nm) 5 106382 00030227 pix 1 18322e-7 3 pix 3 pix (2)

For each individual spectrum the scale was alsonearly linear and the pixel position of higher-or-der lines was simply a multiple of the pixel po-sition at first order

The highest possible resolution was obtainedonly when the meteor moved nearly perpendic-ular to the dispersion direction A 4-pixel binning

was applied perpendicular to the dispersion di-rection to shorten the readout time but this alsodecreased the resolution if the meteor did notmove perpendicular to the dispersion directionUnfortunately the meteor orientation varies as afunction of the angle on the sky and the leftrightposition relative to the viewing direction in thefield of view Under optimal conditions the fullwidth at half-maximum (FWHM) spectral reso-lution in third order was as low as 18 Aring whichoutperforms photographic spectral analysis tech-niques by a factor of 3

RESULTS

The instrument was deployed during the in-tense 2001 Leonid meteor storm over the conti-nental United States onboard the USAF418thFLTS-operated NKC-135 ldquoFISTArdquo research air-craft during the 2001 Leonid MAC mission at thetime of the intense November 18th 1040 univer-sal time (UT) Leonid storm peak (Jenniskens andRussell 2003)

Figure 2 shows our best result a Leonid witha brief terminal flare which appeared in a west-northwestern direction while flying over LittleRock AR at 090558 UT November 18 (9297W13480N 37000 ft) No meteor trajectory infor-mation is available from a stereoscopic perspec-tive because this was a single-plane mission butthe typical altitude at which such meteors end is93 6 5 km (Jenniskens et al 1998 Betlem et al2000) While most of the spectrum consists of sec-ond order lines of sodium third order lines ofionized calcium are readily identified and verystrong The region shortward of the sodium linesis dominated by intrinsically bright third orderiron lines which were detected despite the factorof 50 lower instrumental sensitivity for third or-der lines (Fig 3)

The spectrum shown in Fig 3 is slightly out offocus at one side of the CCD because of a smallmisalignment of the optics and CCD plane As aresult the spectral resolution increases from aFWHM 5 13 Aring at 370 nm to about 24 Aring at 410nm This does not affect the analysis of the dataother than to require line profile integration priorto comparing relative intensities The sodiumemissions detected above this main portion of thespectrum on the CCD frame are due to the per-sistent afterglow or recombination line emissionfrom sodium atoms released at a time prior to the

JENNISKENS ET AL72

FIG 4 Examples of vigneting in the camera for twodifferent positions of the spectrum on the CCD The firstexample applies to a band near the upper edge of the CCDframe The second example is that valid for the positionof the spectrum in Fig 2

start of the exposure No correction was made forthis (small) contribution to the main spectrum

After summing all relevant rows we foundthat emission lines were on top of a continuumemission identified as the First Positive Band ofN2 observed in second order (Fig 5) No knownsecond or third order molecular band from at-mospheric emissions other than that of CN wasexpected in the region around the CN band headAfter correction for second order instrument re-sponse (Fig 3) we fit the N2 band contour to thewhole spectrum The N2 band contour definedthe background level A small correction for in-strument response to the remaining iron lines wasmade by subtracting the N2 band contour andcorrecting the resulting spectrum in a manner ap-propriate for the third order response (Fig 3)

The CN molecular band

Once instrument and background correctionswere made we focused our attention on the thirdorder spectrum around 388 nm (Fig 6) The newthird order spectrum resolved the weaker ironlines considerably as compared with our earlierfirst order spectra (Rairden et al 2000) The ex-pected Fe line emission spectrum shown as adashed line in Fig 6 was calculated for excita-tion temperatures of 4300 and 4500 K using allneutral iron lines in the NIST Atomic SpectraDatabase Version 20 (NIST Standard ReferenceDatabase 78) A comparison of the observed andcalculated emission lines established their iden-

tity and showed that weak iron lines did not con-tribute to the background continuum It shouldbe noted that second order iron lines are presentelsewhere in the spectrum

As indicated in Fig 6 two iron lines bracket theposition of the CN molecular band Because ofthe absence of any additional faint line emissionsin this region an upper limit on the CN abun-dance could be established Though the CN bandmight have in fact been detected as a small bumpit is at the noise level of the spectrum

The relative abundance of CN molecules versus iron atoms was calculated using theNEQAIR2 model of heated air in thermodynamicequilibrium (Laux 1993 Park et al 1997) The CNband profile for a resolution of 18 Aring FWHM isshown in Fig 6 The band head at 38844 nm isunresolved with its width reflecting the instru-ment profile With ICN equal to the total inte-grated emission intensity of CN in the range375ndash393 nm and IFe equal to the total integratedemission intensity of all third order Fe lines in therange 3813ndash3867 nm we calculated for T 5 4500that IFenFe 5 324 3 10216 Wsr and ICNnCN 5558 3 10217 Wsr where nFe is the total concen-tration of Fe atoms in cm23 and nCN is the totalconcentration of CN moleculescm3 For T 54300 K IFenFe 5 213 3 10216 Wsr and ICNnCN 5 393 3 10217 Wsr To calculate these val-ues we assumed that the emitting levels were in

METEORS AND EXOGENOUS ORGANIC MOLECULES 73

FIG 5 Extracted spectrum of Fig 2 after correction forinstrument response and vigneting The dashed lineshows the contribution from the first positive band of N2in second order

FIG 6 Detail of Fig 5 showing the position of the CNband and the iron lines in third order A model spec-trum at 4300 K of iron atom line emission is shown by adashed line The continuum background is the band con-tour of N2 in second order The thick line is a theoreticalprofile of the CN band for air plasma at 4300 K and in-strumental broadening FWHM 5 18 Aring

Boltzmann equilibrium with the ground state ofFe and CN at the given temperature These areconcentrations per total Fe atoms and total CNmolecules (that is not per molecule of CN in theexcited electronic B state)

From these calculations we found that the ra-tio of the total number of atoms of CN and Fe inthe meteor plasma was nCNnFe 0017 for an ex-citation temperature of 4500 K (or 0019 for anexcitation temperature of 4300 K) The largest un-certainty was caused by the position of the back-ground level However given the good fit of theN2 band contour to the overall spectrum short-ward of 590 nm (Fig 5) a conservative error ofless than 50 places the upper limit at nCNnFe 003

An even more precise upper limit was derivedfrom a Geminid spectrum obtained during the2001 Geminid shower with the same instrumentdeployed from a ground site at Fremont Peak Ob-servatory California (Fig 7) For that spectrumwe calculated for an excitation temperature of4500 K that nCNnFe 0011 (or 0012 at 4300 K)However Geminid meteoroids frequently ap-proach the Sun to within 014 AU which heatsthe meteoroids every 16 years to at least T 5 510K (the temperature of small zodiacal cloud mete-oroids) and possibly as high as 1800 K if thelarger meteoroids are more like a black body(Levasseur-Regourd et al 2000) This is hotenough to change their morphology into morecompact objects and can result in the loss of someor all of the organic matter

The more intense Geminid spectrum confirmsthat no weak second or third order atomic linesof any kind are present at the position of the CNband Hence we conclude that the meteor plasmaresponsible for most of the Leonid meteorrsquos emis-sion contains 1 CN molecule per 30 Fe atomsThis result improves by more than a factor of 10the upper limit of 1 CN per 3 Fe atoms found ear-lier by Rairden et al (2000) These numbers aresignificantly less than expected If all of the ni-trogen in the complex organic matter of the cometdust was decomposed in the form of CN mole-cules this would give nCNnFe 5 079 6 002based on the observed nitrogen abundance ofcomplex organic matter in small dust grains ofcomet 1PHalley (Delsemme 1991) The ob-served abundance is a factor of 26 lower

CN can also be formed by reactions in the me-teor air plasma involving CO2 and N2 in the am-bient atmosphere Though CN formation via suchaerothermochemistry is not well understood inthe case of thermodynamic equilibrium therewould not be enough CN made from this mech-anism to be detectable The calculated CN abun-dance for a 4300 K air plasma in local thermo-dynamic equilibrium at 95 km altitude equals10 3 105 molecules of CNcm3 The 22 magni-tude Leonid meteor studied here has a mass ofabout 044 g of which 0033 g is iron (Delsemme1991) That iron is ablated and distributed over apath length of about 17 km (Betlem et al 2000)in a cylindrical volume of at least a 2-m radius(Boyd 2000) which indicates a mean density ofnFe 5 2 3 109 atomscm3 and an expected nCNnFe of about 5 3 1025 a factor of 6 less than ourdetection limit

DISCUSSION

There are several reasons why the abundanceof CN in the meteor plasma can be a factor of 26lower than expected from the decomposition ofmeteoric organic matter during ablation (1) theabundance of organic matter in the dust of55PTempel-Tuttle is not known (2) the abun-dance of nitrogen in that organic matter is un-known (3) the fate of organic matter in the in-terplanetary medium due to space weathering isnot known (4) the organic matter could havebeen lost at higher altitude (earlier in flight) thanwhere the measurement was made (5) theprocess of ablation could favor the release of ni-

JENNISKENS ET AL74

FIG 7 Detail of the 084127 UT Geminid meteor spec-trum observed on December 13 2001 with symbols andlines the same as in Fig 6

trogen in atomic form or in the form of some othersmall molecules or (6) the nitrogen could survivein the form of larger organic compounds that arenot detected The latter possibility could providea new pathway towards prebiotic compoundsWe will now consider the significance and im-plication of these arguments one by one

Since the abundance of organic matter in thedust of 55PTemplel-Tuttle is not known it ispossible that there was little or no organic mat-ter in the Leonid particles to begin with Conse-quently the absence of nitrogen-bearing organicmatter in the Leonid meteoroids may have re-sulted in our inability to detect CN The onlyother measurement that could have detected ni-trogen-bearing organic matter in the dust (as op-posed to molecules in the gas phase) is in situmass spectroscopy The Leonid parent comet hasnot been visited In the coma of comet 1PHal-ley nitrogen-bearing organic matter was detectedin particles (the CHON particles) analyzed by themass spectrometers on the GIOTTO and VEGAspacecraft Because comet Halley is not the par-ent of the Leonid stream it is recommended thatfuture work focus on the meteoroids of the Ori-onid and eta-Aquarid streams that do originatefrom comet Halley Though bright meteors fromthese showers are much more difficult to observemuch the same result would be expected TheLeonid parent comet is remarkable only becauseit passes so close to Earthrsquos orbit Most impor-tantly since 55PTempel-Tuttle is in a Halley-type orbit and it should originate from the Oortcloud just as 1PHalley there is no reason to ex-pect that it is less rich in organic matter

For the same reason there is no reason to ex-pect that the nitrogen content or composition ofthe organic matter would differ from that of1PHalley Halley dust is as rich in nitrogen thanprimitive meteorites Though the nitrogen con-tent in the organic matter of comet Halley hasbeen questioned because of calibration problemswith the detectors there is no reason to expectthere to be more nitrogen in the less primitive me-teorites The heating and irradiation that alter theorganics in meteorites tend to evolve the organicmatter towards materials less rich in nitrogen andoxygen (Jenniskens et al 1993)

It is unknown how the organic content changeswith dust grain size and exposure to the plane-tary medium hence the fate of the organic mat-ter in the interplanetary medium is not knownHowever we know when the observed Leonids

were ejected from the parent comet in 1767 The090558 UT Leonid meteoroid never came closerto the Sun than 0976 AU and remained in the in-terplanetary medium only for 225 years (sevenorbits) most of that time far from the Sun

Might the organic matter have been lost in theEarthrsquos atmosphere early in the trajectory athigher altitudes when the meteoroid first warmedup to above the evaporation and decompositiontemperature of T 700 K This scenario was pro-posed by Elford et al (1997) who observed thatradar meteors are initially detected at heights upto 140 km They concluded that a more volatilecomponent than stony minerals must be ablatingat those heights The specific case of fast Leonidswas discussed by Steel (1998) who calculated thealtitude of evaporation for a range of small or-ganic compounds with relatively low sublimationtemperatures finding ablation heights in excessof 140 km for some compounds However thelight curve of bright Leonid fireballs above 136km was found to be very flat and starting at about200 km (Spurny et al 2000ab) which is incon-sistent with the light curves calculated by Elfordet al (1997) and Steel (1998) At 200 km the grainsare not warm enough to evaporate the organiccompounds that survived the vacuum of space

Indeed the Leonid studied here shows nosign of differential ablation of minerals in order oftheir volatility The elements Na Fe Mg and Caare lost in the same proportion (Fig 2) exceptfor the last wisp of matter after the catastrophicfragmentation during the terminal flare (a topicof future work) Sodium is at least partiallyfound in more volatile minerals than those con-taining magnesium and iron Though somesmaller Leonid meteoroids are known to losesodium early in their trajectory this is onlyfound in fragile cometary shower meteoroidssmall enough to expose the minerals efficientlyto the impinging air flow (BorovicIuml ka et al 1999)Sodium may be contained in the larger Leonidsat lower altitudes because of the limited timeavailable for inward diffusion of heat and out-ward diffusion of molecules Although the or-ganic and mineral components are separate en-tities they are intimately mixed down to verysmall size scales in comet Halley dust and thediffusion of those volatile compounds out of thegrains (if at all possible) will take time Such con-siderations are not included in the model of Steel(1998) Hence the organic matter is not lost earlyin flight

METEORS AND EXOGENOUS ORGANIC MOLECULES 75

The low CN abundance in the air plasma of the090558 UT Leonid meteoroid cannot be under-stood by merely allowing some fraction of nitro-gen to escape as N NH N2 NH3 NH2 NCONO HNO or HCN etc The configuration of Nin the complex organic matter of meteoroids isnot known However from meteorite and mi-crometeorite studies (where much organic matterhas been altered by aqueous alteration processes)we know that a significant fraction of the nitro-gen is contained in the organic framework (seereferences in Keller et al 1995 Greenberg 2000)Aqueous alteration tends to move nitrogen fromthe organic framework to functional groups suchas -NH In cometary matter no aqueous alter-ation has occurred So if N is in the frameworkin meteorite organics it should also exist in theorganic framework of cometary matter Coal-bound nitrogen also resides principally in hete-rocyclic ring moieties mostly five-memberedpyrrolic rings and six-membered pyridinic rings(Smith et al 1994 Zhang 2001) In pyrolysis stud-ies of tar such nitrogen is observed to escape pre-dominantly in the form of HCN and NH3 (Chenand Niksa 1992 Nelson et al 1992 Ledesma etal 1998) Ammonia is a product of amino com-pounds At the low pressure and high tempera-ture conditions in meteor plasma CN radicalsrather than HCN are stable and ammonia willbreak apart Small hydrocarbon radicals higherthan CN are so excited by their formation thatthey rapidly fission into lower products leavingonly the thermally stable single and double car-bon radicals unless prevented by rapid radiativecooling Carbon atoms are present too but atlower abundance because of thermodynamic con-siderations The carbon atoms react with N2 toform CN At 4400 K CN and N2 are only par-tially dissociated while O2 and CO (and CO2) arefully dissociated leading to high abundances ofCN if the plasma is in local thermodynamic equi-librium (Jenniskens et al 2000a) Only if oxygenplays a role in extracting the nitrogen after abla-tion then NOx compounds are expected to dom-inate Hence CN should be a significant productof the ablation of organic matter unless the or-ganic matter is released as larger compounds

We conclude that the most likely explanationfor the lack of CN molecules in the meteor plasmais that meteoric organic matter is lost in the formof a large range of relatively complex moleculesor perhaps as an amorphous carbon solid withembedded metal as proposed by Brownlee et al

(2002) Until now no evidence has been found ofsuch meteoric soot particles in the Earth envi-ronment and it remains unknown how well theheating conditions in the electron microscopeused in the experiment by Brownlee et al (2002)mimic the actual physical conditions during me-teoroid ablation

If the organic matter survives as large mole-cules this implies that bonds are prevented frombreaking by rapid radiative cooling or becauseheat is carried away by collisional fragments thattransfer much of the kinetic energy to the ambi-ent air The latter mechanism is the more likelyone and comes about because of conservation ofmomentum in a collision The result is that thelower mass fragment will carry most of the ki-netic energy and it will carry proportionallymore kinetic energy as the mass of the mainresidue increases This is an indirect form of thecollisional cooling that facilitates laser desorp-tion a technique widely used in the laboratory tostudy high-molecular-mass polyaromatic hydro-carbons in the gas phase and for detecting com-plex organic compounds in meteorites

Once released from the surface a single colli-sion will cause the molecule to slow down andlag behind relative to the meteoroid The mole-cule then is slowed to thermal speeds in another10 collisions and remains in the warm wake ofthe meteor for 50 collisions Only about 3 ofair collisions with atoms are inelastic (Oumlpik1958) For collisions with molecules that per-centage is higher but only a fraction of collisionswill result in chemical reactions Typical reactionsthat occur in the burning of propene for exam-ple are (eg Marinov et al 1996)

C3H6 1 O R C2H5 1 HCO (3)

C3H6 1 O R CH2CO 1 CH3 1 H (4)

C3H6 1 O R CH3CHCO 1 H 1 H (5)

Hydrogen can react in an early Earth-like atmos-phere with CO2 to form OH radicals which con-tinue to attack the organic matter via reactionssuch as

C3H6 1 OH R C3H6OH (6)

Radical attack can also lead to the growth of car-bon chains by polymerization reactions in therare case that organic compounds meet Hencethe result of chemistry in the meteor plasma is

JENNISKENS ET AL76

likely a product of relatively high molecular massthat is enriched in oxygen (and nitrogen) Suchcompounds are much more interesting for prebi-otic chemistry than the very-high-molecular-mass functional group-poor compounds foundin (micro-) meteorites

CONCLUSIONS

We conclude that complex organic matter inmeteoroids does not fully decompose into di-atomic constituents or atoms during the most in-tense phase of meteor ablation and that meteoricorganics are most likely lost in the form of largemolecular fragments cooled by radiative coolingand the loss of hot molecular fragments Theproduct of chemical interactions in the air plasmain the meteoroidrsquos wake is most likely a large va-riety of molecules and radicals that are chemicallyaltered functional group-enriched molecularfragments and condensation products

This result marks an alternative route to pre-biotic molecules from organic matter contributedto the Earth by particles at the peak of the massndashfrequency distribution of infalling matter At thetime of the origin of life there would have beentypically as much extraterrestrial matter impact-ing Earth in the form of dust as in the form oflarge asteroids and comets (Ceplecha 1992) ifthey arrived on elliptic orbits to have time to de-posit debris Moreover the organic content ofthose meteoroids should have been similar to thatof the impacting parent bodies with the excep-tion of the volatile compounds lost during dustejection This makes the product of meteor abla-tion and subsequent chemistry potentially thedominant source of complex organic moleculeson the early Earth at the time of the origin of life

ACKNOWLEDGMENTS

We thank George Flynn and an anonymous ref-eree for discussions that helped improve this pa-per The cooled CCD slit-less spectrograph wasdeveloped with support of the NASA Ames Re-search Centerrsquos Directorrsquos Discretionary Fundduring preparation of the Leonid MAC missionsWe thank Mike Koop (California Meteor Society)and Gary Palmer for technical support at variousstages in this project The 2001 Leonid MAC mis-sion was sponsored by NASArsquos Astrobiology

ASTID and Planetary Astronomy programs The2001 mission was executed by the USAF 418thFlight Test Squadron of Edwards AFB in Califor-nia with logistic support from the SETI Institute

ABBREVIATIONS

CCD charge coupled device FWHM full widthat half-maximum IDP interplanetary dust parti-cle MAC Multi-Instrument Aircraft CampaignUT universal time

REFERENCES

Abe S Yano H Ebizuka N and Watanabe J-I (2000)First results of high-definition TV spectroscopic obser-vations of the 1999 Leonid meteor shower Earth MoonPlanets 82-83 369ndash377

Anders E (1989) Pre-biotic organic matter from cometsand asteroids Nature 342 255ndash257

Betlem H Jenniskens P Spurny P Docters vanLeeuwen G Miskotte K Ter Kuile CR Zerubin Pand Angelos C (2000) Precise trajectories and orbits ofmeteoroids from the 1999 Leonid meteor storm EarthMoon Planets 82-83 277ndash284

BorovicIuml ka J Stork R and Bocek J (1999) First resultsfrom video spectroscopy of 1998 Leonid meteors Me-teoritics Planet Sci 34 987ndash994

Boyd ID (2000) Computation of atmospheric entry flowabout a Leonid meteoroid Earth Moon Planets 82-8393ndash108

Brownlee DE Joswiak DJ Kress ME Taylor S andBradley J (2002) Survival of carbon in moderately tostrongly heated IDPs and micrometeorites [abstract1786] In 33rd Lunar and Planetary Science Conference LPIContribution No 1109 Lunar and Planetary InstituteHouston

Ceplecha Z (1971) Spectral data on terminal flare andwake of double-station meteor No 38421 (OndrejovApril 21 1963) Bull Astron Inst Czech 22 219ndash304

Ceplecha Z (1992) Earth influx of interplanetary bodiesAstron Astrophys 263 361ndash366

Chen JC and Niksa S (1992) Coal devolatilization dur-ing rapid transient heating 1 Primary devolatilizationEnergy Fuels 6 254ndash264

Chyba CF and Sagan C (1992) Endogenous productionexogenous delivery and impact shock synthesis of or-ganic molecules An inventory for the origins of lifeNature 355 125ndash132

Chyba CF and Sagan C (1997) Comets as a source ofprebiotic organic molecules for the early Earth InComets and the Origin and Evolution of Life edited by PJThomas CF Chyba and CP McKay Springer VerlagNew York pp 147ndash173

Delsemme AH (1991) Nature and history of the organiccompounds in comets an astrophysical view In Astro-

METEORS AND EXOGENOUS ORGANIC MOLECULES 77

probed 22 magnitude meteors at altitudes of90ndash120 km which corresponds to about 1 cm sizemeteoroids in the case of Leonid entry speeds of72 kms These are about 5 magnitudes fainter (a factor of 100 in mass less) than those studiedfrom traditional photographic spectroscopy tech-niques This is small enough to have the requiredrarefied flow field that characterizes meteor con-ditions for the bulk of meteoroid mass influx (Jen-niskens et al 2000a)

The AF-S Nikkor f28300D IF-ED 300ndashmmtelephoto lens collection optics determined thefield of view (5deg) and the short wavelength cut-off of sensitivity which extended down to about360 nm The grating dispersion was chosen in or-der to be able to capture any section of the zerofirst and full second order spectrum (Fig 1) Sec-ond order refers to the integer wavelength dif-ference of 2 for diffractions from two neighbor-ing grating grooves Based on this we chose an11 3 11-cm plane transmission grating 35-54-20-660 with aluminum coating on 12-mm BK7 sub-strate manufactured by Richardson Grating Lab-oratory (Rochester NY) We did not use an orderseparation filter a holder for which is providedby the telephoto lens Without such an order sep-aration filter the near-UV part of the third orderspectrum is also captured which overlaps thesecond order spectrum above 540 nm (3 3 360nm 5 2 3 540 nm) Grating and detector were po-sitioned such that the dispersion occurred alongthe rows of the CCD We measured a dispersionof 540 lmm and blaze wavelength of 698 nm(34deg) This configuration provided a full secondorder spectrum out to about 925 nm The factorygrating specifications suggested a good sensitiv-ity out to 11 mm but we found that the gratingefficiency fell off rapidly above 850 nm Figure 3shows the response curves for the camera cali-

brated for various orders which were measuredby using an approximate point source (an opticalfiber manufactured by Ocean Optics) illuminatedwith a mercury-argon lamp (wavelength calibra-tion) and a tungsten lamp (absolute intensity cal-ibration) The measured efficiencies have beenscaled to the product of the factory responsecurve for the transmission grating the transmis-sion of the camera lens and the CCD quantumefficiency (dashed line Fig 3)

The response curve shown in Fig 3 does notinclude the effects of (a) vigneting (or flatfield-ing) or (b) the geometric correction (a simple co-sine function with angle of incidence) for the ori-entation of the grating relative to the direction ofthe meteor By measuring the response of the in-strument to a homogeneously illuminated flat

METEORS AND EXOGENOUS ORGANIC MOLECULES 71

FIG 2 Second order spectrum of the 090558 UT November 18 2001 Leonid meteor Lines other than thosecaused by iron are marked Brackets indicate emission lines in the overlapping third order

FIG 3 Measured spectral response curves for first(1st) second (2nd) and third (3rd) order The second andthird orders are scaled down for comparison The dashedline shows the theoretical response for the combined fac-tory given grating efficiency CCD quantum efficiencyand camera optics transparency

surface we found that vigneting varied as a func-tion of row and pixel position This allowed us todetermine the flatfielding intensity calibrationcorrection factor ( f ) for a given position of thespectrum on the CCD Two examples are shownin Fig 4 each representing the mean over a bandof CCD rows

The position of the meteor on the sky and thechoice of focus determined what part of the spec-trum was recorded Together with its encasingthe camera left an effective hunting ground of astrip of sky 5deg wide and up to 886deg from the for-ward direction (Fig 1) The wavelength scale wasclose to linear over the full range which spannedabout 13600 pixels

l (nm) 5 15737 1 013382 3 pixel (1)

A small 2 nm residual described by the follow-ing second order polynomial remained

l (observed 2 calculated) (nm) 5 106382 00030227 pix 1 18322e-7 3 pix 3 pix (2)

For each individual spectrum the scale was alsonearly linear and the pixel position of higher-or-der lines was simply a multiple of the pixel po-sition at first order

The highest possible resolution was obtainedonly when the meteor moved nearly perpendic-ular to the dispersion direction A 4-pixel binning

was applied perpendicular to the dispersion di-rection to shorten the readout time but this alsodecreased the resolution if the meteor did notmove perpendicular to the dispersion directionUnfortunately the meteor orientation varies as afunction of the angle on the sky and the leftrightposition relative to the viewing direction in thefield of view Under optimal conditions the fullwidth at half-maximum (FWHM) spectral reso-lution in third order was as low as 18 Aring whichoutperforms photographic spectral analysis tech-niques by a factor of 3

RESULTS

The instrument was deployed during the in-tense 2001 Leonid meteor storm over the conti-nental United States onboard the USAF418thFLTS-operated NKC-135 ldquoFISTArdquo research air-craft during the 2001 Leonid MAC mission at thetime of the intense November 18th 1040 univer-sal time (UT) Leonid storm peak (Jenniskens andRussell 2003)

Figure 2 shows our best result a Leonid witha brief terminal flare which appeared in a west-northwestern direction while flying over LittleRock AR at 090558 UT November 18 (9297W13480N 37000 ft) No meteor trajectory infor-mation is available from a stereoscopic perspec-tive because this was a single-plane mission butthe typical altitude at which such meteors end is93 6 5 km (Jenniskens et al 1998 Betlem et al2000) While most of the spectrum consists of sec-ond order lines of sodium third order lines ofionized calcium are readily identified and verystrong The region shortward of the sodium linesis dominated by intrinsically bright third orderiron lines which were detected despite the factorof 50 lower instrumental sensitivity for third or-der lines (Fig 3)

The spectrum shown in Fig 3 is slightly out offocus at one side of the CCD because of a smallmisalignment of the optics and CCD plane As aresult the spectral resolution increases from aFWHM 5 13 Aring at 370 nm to about 24 Aring at 410nm This does not affect the analysis of the dataother than to require line profile integration priorto comparing relative intensities The sodiumemissions detected above this main portion of thespectrum on the CCD frame are due to the per-sistent afterglow or recombination line emissionfrom sodium atoms released at a time prior to the

JENNISKENS ET AL72

FIG 4 Examples of vigneting in the camera for twodifferent positions of the spectrum on the CCD The firstexample applies to a band near the upper edge of the CCDframe The second example is that valid for the positionof the spectrum in Fig 2

start of the exposure No correction was made forthis (small) contribution to the main spectrum

After summing all relevant rows we foundthat emission lines were on top of a continuumemission identified as the First Positive Band ofN2 observed in second order (Fig 5) No knownsecond or third order molecular band from at-mospheric emissions other than that of CN wasexpected in the region around the CN band headAfter correction for second order instrument re-sponse (Fig 3) we fit the N2 band contour to thewhole spectrum The N2 band contour definedthe background level A small correction for in-strument response to the remaining iron lines wasmade by subtracting the N2 band contour andcorrecting the resulting spectrum in a manner ap-propriate for the third order response (Fig 3)

The CN molecular band

Once instrument and background correctionswere made we focused our attention on the thirdorder spectrum around 388 nm (Fig 6) The newthird order spectrum resolved the weaker ironlines considerably as compared with our earlierfirst order spectra (Rairden et al 2000) The ex-pected Fe line emission spectrum shown as adashed line in Fig 6 was calculated for excita-tion temperatures of 4300 and 4500 K using allneutral iron lines in the NIST Atomic SpectraDatabase Version 20 (NIST Standard ReferenceDatabase 78) A comparison of the observed andcalculated emission lines established their iden-

tity and showed that weak iron lines did not con-tribute to the background continuum It shouldbe noted that second order iron lines are presentelsewhere in the spectrum

As indicated in Fig 6 two iron lines bracket theposition of the CN molecular band Because ofthe absence of any additional faint line emissionsin this region an upper limit on the CN abun-dance could be established Though the CN bandmight have in fact been detected as a small bumpit is at the noise level of the spectrum

The relative abundance of CN molecules versus iron atoms was calculated using theNEQAIR2 model of heated air in thermodynamicequilibrium (Laux 1993 Park et al 1997) The CNband profile for a resolution of 18 Aring FWHM isshown in Fig 6 The band head at 38844 nm isunresolved with its width reflecting the instru-ment profile With ICN equal to the total inte-grated emission intensity of CN in the range375ndash393 nm and IFe equal to the total integratedemission intensity of all third order Fe lines in therange 3813ndash3867 nm we calculated for T 5 4500that IFenFe 5 324 3 10216 Wsr and ICNnCN 5558 3 10217 Wsr where nFe is the total concen-tration of Fe atoms in cm23 and nCN is the totalconcentration of CN moleculescm3 For T 54300 K IFenFe 5 213 3 10216 Wsr and ICNnCN 5 393 3 10217 Wsr To calculate these val-ues we assumed that the emitting levels were in

METEORS AND EXOGENOUS ORGANIC MOLECULES 73

FIG 5 Extracted spectrum of Fig 2 after correction forinstrument response and vigneting The dashed lineshows the contribution from the first positive band of N2in second order

FIG 6 Detail of Fig 5 showing the position of the CNband and the iron lines in third order A model spec-trum at 4300 K of iron atom line emission is shown by adashed line The continuum background is the band con-tour of N2 in second order The thick line is a theoreticalprofile of the CN band for air plasma at 4300 K and in-strumental broadening FWHM 5 18 Aring

Boltzmann equilibrium with the ground state ofFe and CN at the given temperature These areconcentrations per total Fe atoms and total CNmolecules (that is not per molecule of CN in theexcited electronic B state)

From these calculations we found that the ra-tio of the total number of atoms of CN and Fe inthe meteor plasma was nCNnFe 0017 for an ex-citation temperature of 4500 K (or 0019 for anexcitation temperature of 4300 K) The largest un-certainty was caused by the position of the back-ground level However given the good fit of theN2 band contour to the overall spectrum short-ward of 590 nm (Fig 5) a conservative error ofless than 50 places the upper limit at nCNnFe 003

An even more precise upper limit was derivedfrom a Geminid spectrum obtained during the2001 Geminid shower with the same instrumentdeployed from a ground site at Fremont Peak Ob-servatory California (Fig 7) For that spectrumwe calculated for an excitation temperature of4500 K that nCNnFe 0011 (or 0012 at 4300 K)However Geminid meteoroids frequently ap-proach the Sun to within 014 AU which heatsthe meteoroids every 16 years to at least T 5 510K (the temperature of small zodiacal cloud mete-oroids) and possibly as high as 1800 K if thelarger meteoroids are more like a black body(Levasseur-Regourd et al 2000) This is hotenough to change their morphology into morecompact objects and can result in the loss of someor all of the organic matter

The more intense Geminid spectrum confirmsthat no weak second or third order atomic linesof any kind are present at the position of the CNband Hence we conclude that the meteor plasmaresponsible for most of the Leonid meteorrsquos emis-sion contains 1 CN molecule per 30 Fe atomsThis result improves by more than a factor of 10the upper limit of 1 CN per 3 Fe atoms found ear-lier by Rairden et al (2000) These numbers aresignificantly less than expected If all of the ni-trogen in the complex organic matter of the cometdust was decomposed in the form of CN mole-cules this would give nCNnFe 5 079 6 002based on the observed nitrogen abundance ofcomplex organic matter in small dust grains ofcomet 1PHalley (Delsemme 1991) The ob-served abundance is a factor of 26 lower

CN can also be formed by reactions in the me-teor air plasma involving CO2 and N2 in the am-bient atmosphere Though CN formation via suchaerothermochemistry is not well understood inthe case of thermodynamic equilibrium therewould not be enough CN made from this mech-anism to be detectable The calculated CN abun-dance for a 4300 K air plasma in local thermo-dynamic equilibrium at 95 km altitude equals10 3 105 molecules of CNcm3 The 22 magni-tude Leonid meteor studied here has a mass ofabout 044 g of which 0033 g is iron (Delsemme1991) That iron is ablated and distributed over apath length of about 17 km (Betlem et al 2000)in a cylindrical volume of at least a 2-m radius(Boyd 2000) which indicates a mean density ofnFe 5 2 3 109 atomscm3 and an expected nCNnFe of about 5 3 1025 a factor of 6 less than ourdetection limit

DISCUSSION

There are several reasons why the abundanceof CN in the meteor plasma can be a factor of 26lower than expected from the decomposition ofmeteoric organic matter during ablation (1) theabundance of organic matter in the dust of55PTempel-Tuttle is not known (2) the abun-dance of nitrogen in that organic matter is un-known (3) the fate of organic matter in the in-terplanetary medium due to space weathering isnot known (4) the organic matter could havebeen lost at higher altitude (earlier in flight) thanwhere the measurement was made (5) theprocess of ablation could favor the release of ni-

JENNISKENS ET AL74

FIG 7 Detail of the 084127 UT Geminid meteor spec-trum observed on December 13 2001 with symbols andlines the same as in Fig 6

trogen in atomic form or in the form of some othersmall molecules or (6) the nitrogen could survivein the form of larger organic compounds that arenot detected The latter possibility could providea new pathway towards prebiotic compoundsWe will now consider the significance and im-plication of these arguments one by one

Since the abundance of organic matter in thedust of 55PTemplel-Tuttle is not known it ispossible that there was little or no organic mat-ter in the Leonid particles to begin with Conse-quently the absence of nitrogen-bearing organicmatter in the Leonid meteoroids may have re-sulted in our inability to detect CN The onlyother measurement that could have detected ni-trogen-bearing organic matter in the dust (as op-posed to molecules in the gas phase) is in situmass spectroscopy The Leonid parent comet hasnot been visited In the coma of comet 1PHal-ley nitrogen-bearing organic matter was detectedin particles (the CHON particles) analyzed by themass spectrometers on the GIOTTO and VEGAspacecraft Because comet Halley is not the par-ent of the Leonid stream it is recommended thatfuture work focus on the meteoroids of the Ori-onid and eta-Aquarid streams that do originatefrom comet Halley Though bright meteors fromthese showers are much more difficult to observemuch the same result would be expected TheLeonid parent comet is remarkable only becauseit passes so close to Earthrsquos orbit Most impor-tantly since 55PTempel-Tuttle is in a Halley-type orbit and it should originate from the Oortcloud just as 1PHalley there is no reason to ex-pect that it is less rich in organic matter

For the same reason there is no reason to ex-pect that the nitrogen content or composition ofthe organic matter would differ from that of1PHalley Halley dust is as rich in nitrogen thanprimitive meteorites Though the nitrogen con-tent in the organic matter of comet Halley hasbeen questioned because of calibration problemswith the detectors there is no reason to expectthere to be more nitrogen in the less primitive me-teorites The heating and irradiation that alter theorganics in meteorites tend to evolve the organicmatter towards materials less rich in nitrogen andoxygen (Jenniskens et al 1993)

It is unknown how the organic content changeswith dust grain size and exposure to the plane-tary medium hence the fate of the organic mat-ter in the interplanetary medium is not knownHowever we know when the observed Leonids

were ejected from the parent comet in 1767 The090558 UT Leonid meteoroid never came closerto the Sun than 0976 AU and remained in the in-terplanetary medium only for 225 years (sevenorbits) most of that time far from the Sun

Might the organic matter have been lost in theEarthrsquos atmosphere early in the trajectory athigher altitudes when the meteoroid first warmedup to above the evaporation and decompositiontemperature of T 700 K This scenario was pro-posed by Elford et al (1997) who observed thatradar meteors are initially detected at heights upto 140 km They concluded that a more volatilecomponent than stony minerals must be ablatingat those heights The specific case of fast Leonidswas discussed by Steel (1998) who calculated thealtitude of evaporation for a range of small or-ganic compounds with relatively low sublimationtemperatures finding ablation heights in excessof 140 km for some compounds However thelight curve of bright Leonid fireballs above 136km was found to be very flat and starting at about200 km (Spurny et al 2000ab) which is incon-sistent with the light curves calculated by Elfordet al (1997) and Steel (1998) At 200 km the grainsare not warm enough to evaporate the organiccompounds that survived the vacuum of space

Indeed the Leonid studied here shows nosign of differential ablation of minerals in order oftheir volatility The elements Na Fe Mg and Caare lost in the same proportion (Fig 2) exceptfor the last wisp of matter after the catastrophicfragmentation during the terminal flare (a topicof future work) Sodium is at least partiallyfound in more volatile minerals than those con-taining magnesium and iron Though somesmaller Leonid meteoroids are known to losesodium early in their trajectory this is onlyfound in fragile cometary shower meteoroidssmall enough to expose the minerals efficientlyto the impinging air flow (BorovicIuml ka et al 1999)Sodium may be contained in the larger Leonidsat lower altitudes because of the limited timeavailable for inward diffusion of heat and out-ward diffusion of molecules Although the or-ganic and mineral components are separate en-tities they are intimately mixed down to verysmall size scales in comet Halley dust and thediffusion of those volatile compounds out of thegrains (if at all possible) will take time Such con-siderations are not included in the model of Steel(1998) Hence the organic matter is not lost earlyin flight

METEORS AND EXOGENOUS ORGANIC MOLECULES 75

The low CN abundance in the air plasma of the090558 UT Leonid meteoroid cannot be under-stood by merely allowing some fraction of nitro-gen to escape as N NH N2 NH3 NH2 NCONO HNO or HCN etc The configuration of Nin the complex organic matter of meteoroids isnot known However from meteorite and mi-crometeorite studies (where much organic matterhas been altered by aqueous alteration processes)we know that a significant fraction of the nitro-gen is contained in the organic framework (seereferences in Keller et al 1995 Greenberg 2000)Aqueous alteration tends to move nitrogen fromthe organic framework to functional groups suchas -NH In cometary matter no aqueous alter-ation has occurred So if N is in the frameworkin meteorite organics it should also exist in theorganic framework of cometary matter Coal-bound nitrogen also resides principally in hete-rocyclic ring moieties mostly five-memberedpyrrolic rings and six-membered pyridinic rings(Smith et al 1994 Zhang 2001) In pyrolysis stud-ies of tar such nitrogen is observed to escape pre-dominantly in the form of HCN and NH3 (Chenand Niksa 1992 Nelson et al 1992 Ledesma etal 1998) Ammonia is a product of amino com-pounds At the low pressure and high tempera-ture conditions in meteor plasma CN radicalsrather than HCN are stable and ammonia willbreak apart Small hydrocarbon radicals higherthan CN are so excited by their formation thatthey rapidly fission into lower products leavingonly the thermally stable single and double car-bon radicals unless prevented by rapid radiativecooling Carbon atoms are present too but atlower abundance because of thermodynamic con-siderations The carbon atoms react with N2 toform CN At 4400 K CN and N2 are only par-tially dissociated while O2 and CO (and CO2) arefully dissociated leading to high abundances ofCN if the plasma is in local thermodynamic equi-librium (Jenniskens et al 2000a) Only if oxygenplays a role in extracting the nitrogen after abla-tion then NOx compounds are expected to dom-inate Hence CN should be a significant productof the ablation of organic matter unless the or-ganic matter is released as larger compounds

We conclude that the most likely explanationfor the lack of CN molecules in the meteor plasmais that meteoric organic matter is lost in the formof a large range of relatively complex moleculesor perhaps as an amorphous carbon solid withembedded metal as proposed by Brownlee et al

(2002) Until now no evidence has been found ofsuch meteoric soot particles in the Earth envi-ronment and it remains unknown how well theheating conditions in the electron microscopeused in the experiment by Brownlee et al (2002)mimic the actual physical conditions during me-teoroid ablation

If the organic matter survives as large mole-cules this implies that bonds are prevented frombreaking by rapid radiative cooling or becauseheat is carried away by collisional fragments thattransfer much of the kinetic energy to the ambi-ent air The latter mechanism is the more likelyone and comes about because of conservation ofmomentum in a collision The result is that thelower mass fragment will carry most of the ki-netic energy and it will carry proportionallymore kinetic energy as the mass of the mainresidue increases This is an indirect form of thecollisional cooling that facilitates laser desorp-tion a technique widely used in the laboratory tostudy high-molecular-mass polyaromatic hydro-carbons in the gas phase and for detecting com-plex organic compounds in meteorites

Once released from the surface a single colli-sion will cause the molecule to slow down andlag behind relative to the meteoroid The mole-cule then is slowed to thermal speeds in another10 collisions and remains in the warm wake ofthe meteor for 50 collisions Only about 3 ofair collisions with atoms are inelastic (Oumlpik1958) For collisions with molecules that per-centage is higher but only a fraction of collisionswill result in chemical reactions Typical reactionsthat occur in the burning of propene for exam-ple are (eg Marinov et al 1996)

C3H6 1 O R C2H5 1 HCO (3)

C3H6 1 O R CH2CO 1 CH3 1 H (4)

C3H6 1 O R CH3CHCO 1 H 1 H (5)

Hydrogen can react in an early Earth-like atmos-phere with CO2 to form OH radicals which con-tinue to attack the organic matter via reactionssuch as

C3H6 1 OH R C3H6OH (6)

Radical attack can also lead to the growth of car-bon chains by polymerization reactions in therare case that organic compounds meet Hencethe result of chemistry in the meteor plasma is

JENNISKENS ET AL76

likely a product of relatively high molecular massthat is enriched in oxygen (and nitrogen) Suchcompounds are much more interesting for prebi-otic chemistry than the very-high-molecular-mass functional group-poor compounds foundin (micro-) meteorites

CONCLUSIONS

We conclude that complex organic matter inmeteoroids does not fully decompose into di-atomic constituents or atoms during the most in-tense phase of meteor ablation and that meteoricorganics are most likely lost in the form of largemolecular fragments cooled by radiative coolingand the loss of hot molecular fragments Theproduct of chemical interactions in the air plasmain the meteoroidrsquos wake is most likely a large va-riety of molecules and radicals that are chemicallyaltered functional group-enriched molecularfragments and condensation products

This result marks an alternative route to pre-biotic molecules from organic matter contributedto the Earth by particles at the peak of the massndashfrequency distribution of infalling matter At thetime of the origin of life there would have beentypically as much extraterrestrial matter impact-ing Earth in the form of dust as in the form oflarge asteroids and comets (Ceplecha 1992) ifthey arrived on elliptic orbits to have time to de-posit debris Moreover the organic content ofthose meteoroids should have been similar to thatof the impacting parent bodies with the excep-tion of the volatile compounds lost during dustejection This makes the product of meteor abla-tion and subsequent chemistry potentially thedominant source of complex organic moleculeson the early Earth at the time of the origin of life

ACKNOWLEDGMENTS

We thank George Flynn and an anonymous ref-eree for discussions that helped improve this pa-per The cooled CCD slit-less spectrograph wasdeveloped with support of the NASA Ames Re-search Centerrsquos Directorrsquos Discretionary Fundduring preparation of the Leonid MAC missionsWe thank Mike Koop (California Meteor Society)and Gary Palmer for technical support at variousstages in this project The 2001 Leonid MAC mis-sion was sponsored by NASArsquos Astrobiology

ASTID and Planetary Astronomy programs The2001 mission was executed by the USAF 418thFlight Test Squadron of Edwards AFB in Califor-nia with logistic support from the SETI Institute

ABBREVIATIONS

CCD charge coupled device FWHM full widthat half-maximum IDP interplanetary dust parti-cle MAC Multi-Instrument Aircraft CampaignUT universal time

REFERENCES

Abe S Yano H Ebizuka N and Watanabe J-I (2000)First results of high-definition TV spectroscopic obser-vations of the 1999 Leonid meteor shower Earth MoonPlanets 82-83 369ndash377

Anders E (1989) Pre-biotic organic matter from cometsand asteroids Nature 342 255ndash257

Betlem H Jenniskens P Spurny P Docters vanLeeuwen G Miskotte K Ter Kuile CR Zerubin Pand Angelos C (2000) Precise trajectories and orbits ofmeteoroids from the 1999 Leonid meteor storm EarthMoon Planets 82-83 277ndash284

BorovicIuml ka J Stork R and Bocek J (1999) First resultsfrom video spectroscopy of 1998 Leonid meteors Me-teoritics Planet Sci 34 987ndash994

Boyd ID (2000) Computation of atmospheric entry flowabout a Leonid meteoroid Earth Moon Planets 82-8393ndash108

Brownlee DE Joswiak DJ Kress ME Taylor S andBradley J (2002) Survival of carbon in moderately tostrongly heated IDPs and micrometeorites [abstract1786] In 33rd Lunar and Planetary Science Conference LPIContribution No 1109 Lunar and Planetary InstituteHouston

Ceplecha Z (1971) Spectral data on terminal flare andwake of double-station meteor No 38421 (OndrejovApril 21 1963) Bull Astron Inst Czech 22 219ndash304

Ceplecha Z (1992) Earth influx of interplanetary bodiesAstron Astrophys 263 361ndash366

Chen JC and Niksa S (1992) Coal devolatilization dur-ing rapid transient heating 1 Primary devolatilizationEnergy Fuels 6 254ndash264

Chyba CF and Sagan C (1992) Endogenous productionexogenous delivery and impact shock synthesis of or-ganic molecules An inventory for the origins of lifeNature 355 125ndash132

Chyba CF and Sagan C (1997) Comets as a source ofprebiotic organic molecules for the early Earth InComets and the Origin and Evolution of Life edited by PJThomas CF Chyba and CP McKay Springer VerlagNew York pp 147ndash173

Delsemme AH (1991) Nature and history of the organiccompounds in comets an astrophysical view In Astro-

METEORS AND EXOGENOUS ORGANIC MOLECULES 77

surface we found that vigneting varied as a func-tion of row and pixel position This allowed us todetermine the flatfielding intensity calibrationcorrection factor ( f ) for a given position of thespectrum on the CCD Two examples are shownin Fig 4 each representing the mean over a bandof CCD rows

The position of the meteor on the sky and thechoice of focus determined what part of the spec-trum was recorded Together with its encasingthe camera left an effective hunting ground of astrip of sky 5deg wide and up to 886deg from the for-ward direction (Fig 1) The wavelength scale wasclose to linear over the full range which spannedabout 13600 pixels

l (nm) 5 15737 1 013382 3 pixel (1)

A small 2 nm residual described by the follow-ing second order polynomial remained

l (observed 2 calculated) (nm) 5 106382 00030227 pix 1 18322e-7 3 pix 3 pix (2)

For each individual spectrum the scale was alsonearly linear and the pixel position of higher-or-der lines was simply a multiple of the pixel po-sition at first order

The highest possible resolution was obtainedonly when the meteor moved nearly perpendic-ular to the dispersion direction A 4-pixel binning

was applied perpendicular to the dispersion di-rection to shorten the readout time but this alsodecreased the resolution if the meteor did notmove perpendicular to the dispersion directionUnfortunately the meteor orientation varies as afunction of the angle on the sky and the leftrightposition relative to the viewing direction in thefield of view Under optimal conditions the fullwidth at half-maximum (FWHM) spectral reso-lution in third order was as low as 18 Aring whichoutperforms photographic spectral analysis tech-niques by a factor of 3

RESULTS

The instrument was deployed during the in-tense 2001 Leonid meteor storm over the conti-nental United States onboard the USAF418thFLTS-operated NKC-135 ldquoFISTArdquo research air-craft during the 2001 Leonid MAC mission at thetime of the intense November 18th 1040 univer-sal time (UT) Leonid storm peak (Jenniskens andRussell 2003)

Figure 2 shows our best result a Leonid witha brief terminal flare which appeared in a west-northwestern direction while flying over LittleRock AR at 090558 UT November 18 (9297W13480N 37000 ft) No meteor trajectory infor-mation is available from a stereoscopic perspec-tive because this was a single-plane mission butthe typical altitude at which such meteors end is93 6 5 km (Jenniskens et al 1998 Betlem et al2000) While most of the spectrum consists of sec-ond order lines of sodium third order lines ofionized calcium are readily identified and verystrong The region shortward of the sodium linesis dominated by intrinsically bright third orderiron lines which were detected despite the factorof 50 lower instrumental sensitivity for third or-der lines (Fig 3)

The spectrum shown in Fig 3 is slightly out offocus at one side of the CCD because of a smallmisalignment of the optics and CCD plane As aresult the spectral resolution increases from aFWHM 5 13 Aring at 370 nm to about 24 Aring at 410nm This does not affect the analysis of the dataother than to require line profile integration priorto comparing relative intensities The sodiumemissions detected above this main portion of thespectrum on the CCD frame are due to the per-sistent afterglow or recombination line emissionfrom sodium atoms released at a time prior to the

JENNISKENS ET AL72

FIG 4 Examples of vigneting in the camera for twodifferent positions of the spectrum on the CCD The firstexample applies to a band near the upper edge of the CCDframe The second example is that valid for the positionof the spectrum in Fig 2

start of the exposure No correction was made forthis (small) contribution to the main spectrum

After summing all relevant rows we foundthat emission lines were on top of a continuumemission identified as the First Positive Band ofN2 observed in second order (Fig 5) No knownsecond or third order molecular band from at-mospheric emissions other than that of CN wasexpected in the region around the CN band headAfter correction for second order instrument re-sponse (Fig 3) we fit the N2 band contour to thewhole spectrum The N2 band contour definedthe background level A small correction for in-strument response to the remaining iron lines wasmade by subtracting the N2 band contour andcorrecting the resulting spectrum in a manner ap-propriate for the third order response (Fig 3)

The CN molecular band

Once instrument and background correctionswere made we focused our attention on the thirdorder spectrum around 388 nm (Fig 6) The newthird order spectrum resolved the weaker ironlines considerably as compared with our earlierfirst order spectra (Rairden et al 2000) The ex-pected Fe line emission spectrum shown as adashed line in Fig 6 was calculated for excita-tion temperatures of 4300 and 4500 K using allneutral iron lines in the NIST Atomic SpectraDatabase Version 20 (NIST Standard ReferenceDatabase 78) A comparison of the observed andcalculated emission lines established their iden-

tity and showed that weak iron lines did not con-tribute to the background continuum It shouldbe noted that second order iron lines are presentelsewhere in the spectrum

As indicated in Fig 6 two iron lines bracket theposition of the CN molecular band Because ofthe absence of any additional faint line emissionsin this region an upper limit on the CN abun-dance could be established Though the CN bandmight have in fact been detected as a small bumpit is at the noise level of the spectrum

The relative abundance of CN molecules versus iron atoms was calculated using theNEQAIR2 model of heated air in thermodynamicequilibrium (Laux 1993 Park et al 1997) The CNband profile for a resolution of 18 Aring FWHM isshown in Fig 6 The band head at 38844 nm isunresolved with its width reflecting the instru-ment profile With ICN equal to the total inte-grated emission intensity of CN in the range375ndash393 nm and IFe equal to the total integratedemission intensity of all third order Fe lines in therange 3813ndash3867 nm we calculated for T 5 4500that IFenFe 5 324 3 10216 Wsr and ICNnCN 5558 3 10217 Wsr where nFe is the total concen-tration of Fe atoms in cm23 and nCN is the totalconcentration of CN moleculescm3 For T 54300 K IFenFe 5 213 3 10216 Wsr and ICNnCN 5 393 3 10217 Wsr To calculate these val-ues we assumed that the emitting levels were in

METEORS AND EXOGENOUS ORGANIC MOLECULES 73

FIG 5 Extracted spectrum of Fig 2 after correction forinstrument response and vigneting The dashed lineshows the contribution from the first positive band of N2in second order

FIG 6 Detail of Fig 5 showing the position of the CNband and the iron lines in third order A model spec-trum at 4300 K of iron atom line emission is shown by adashed line The continuum background is the band con-tour of N2 in second order The thick line is a theoreticalprofile of the CN band for air plasma at 4300 K and in-strumental broadening FWHM 5 18 Aring

Boltzmann equilibrium with the ground state ofFe and CN at the given temperature These areconcentrations per total Fe atoms and total CNmolecules (that is not per molecule of CN in theexcited electronic B state)

From these calculations we found that the ra-tio of the total number of atoms of CN and Fe inthe meteor plasma was nCNnFe 0017 for an ex-citation temperature of 4500 K (or 0019 for anexcitation temperature of 4300 K) The largest un-certainty was caused by the position of the back-ground level However given the good fit of theN2 band contour to the overall spectrum short-ward of 590 nm (Fig 5) a conservative error ofless than 50 places the upper limit at nCNnFe 003

An even more precise upper limit was derivedfrom a Geminid spectrum obtained during the2001 Geminid shower with the same instrumentdeployed from a ground site at Fremont Peak Ob-servatory California (Fig 7) For that spectrumwe calculated for an excitation temperature of4500 K that nCNnFe 0011 (or 0012 at 4300 K)However Geminid meteoroids frequently ap-proach the Sun to within 014 AU which heatsthe meteoroids every 16 years to at least T 5 510K (the temperature of small zodiacal cloud mete-oroids) and possibly as high as 1800 K if thelarger meteoroids are more like a black body(Levasseur-Regourd et al 2000) This is hotenough to change their morphology into morecompact objects and can result in the loss of someor all of the organic matter

The more intense Geminid spectrum confirmsthat no weak second or third order atomic linesof any kind are present at the position of the CNband Hence we conclude that the meteor plasmaresponsible for most of the Leonid meteorrsquos emis-sion contains 1 CN molecule per 30 Fe atomsThis result improves by more than a factor of 10the upper limit of 1 CN per 3 Fe atoms found ear-lier by Rairden et al (2000) These numbers aresignificantly less than expected If all of the ni-trogen in the complex organic matter of the cometdust was decomposed in the form of CN mole-cules this would give nCNnFe 5 079 6 002based on the observed nitrogen abundance ofcomplex organic matter in small dust grains ofcomet 1PHalley (Delsemme 1991) The ob-served abundance is a factor of 26 lower

CN can also be formed by reactions in the me-teor air plasma involving CO2 and N2 in the am-bient atmosphere Though CN formation via suchaerothermochemistry is not well understood inthe case of thermodynamic equilibrium therewould not be enough CN made from this mech-anism to be detectable The calculated CN abun-dance for a 4300 K air plasma in local thermo-dynamic equilibrium at 95 km altitude equals10 3 105 molecules of CNcm3 The 22 magni-tude Leonid meteor studied here has a mass ofabout 044 g of which 0033 g is iron (Delsemme1991) That iron is ablated and distributed over apath length of about 17 km (Betlem et al 2000)in a cylindrical volume of at least a 2-m radius(Boyd 2000) which indicates a mean density ofnFe 5 2 3 109 atomscm3 and an expected nCNnFe of about 5 3 1025 a factor of 6 less than ourdetection limit

DISCUSSION

There are several reasons why the abundanceof CN in the meteor plasma can be a factor of 26lower than expected from the decomposition ofmeteoric organic matter during ablation (1) theabundance of organic matter in the dust of55PTempel-Tuttle is not known (2) the abun-dance of nitrogen in that organic matter is un-known (3) the fate of organic matter in the in-terplanetary medium due to space weathering isnot known (4) the organic matter could havebeen lost at higher altitude (earlier in flight) thanwhere the measurement was made (5) theprocess of ablation could favor the release of ni-

JENNISKENS ET AL74

FIG 7 Detail of the 084127 UT Geminid meteor spec-trum observed on December 13 2001 with symbols andlines the same as in Fig 6

trogen in atomic form or in the form of some othersmall molecules or (6) the nitrogen could survivein the form of larger organic compounds that arenot detected The latter possibility could providea new pathway towards prebiotic compoundsWe will now consider the significance and im-plication of these arguments one by one

Since the abundance of organic matter in thedust of 55PTemplel-Tuttle is not known it ispossible that there was little or no organic mat-ter in the Leonid particles to begin with Conse-quently the absence of nitrogen-bearing organicmatter in the Leonid meteoroids may have re-sulted in our inability to detect CN The onlyother measurement that could have detected ni-trogen-bearing organic matter in the dust (as op-posed to molecules in the gas phase) is in situmass spectroscopy The Leonid parent comet hasnot been visited In the coma of comet 1PHal-ley nitrogen-bearing organic matter was detectedin particles (the CHON particles) analyzed by themass spectrometers on the GIOTTO and VEGAspacecraft Because comet Halley is not the par-ent of the Leonid stream it is recommended thatfuture work focus on the meteoroids of the Ori-onid and eta-Aquarid streams that do originatefrom comet Halley Though bright meteors fromthese showers are much more difficult to observemuch the same result would be expected TheLeonid parent comet is remarkable only becauseit passes so close to Earthrsquos orbit Most impor-tantly since 55PTempel-Tuttle is in a Halley-type orbit and it should originate from the Oortcloud just as 1PHalley there is no reason to ex-pect that it is less rich in organic matter

For the same reason there is no reason to ex-pect that the nitrogen content or composition ofthe organic matter would differ from that of1PHalley Halley dust is as rich in nitrogen thanprimitive meteorites Though the nitrogen con-tent in the organic matter of comet Halley hasbeen questioned because of calibration problemswith the detectors there is no reason to expectthere to be more nitrogen in the less primitive me-teorites The heating and irradiation that alter theorganics in meteorites tend to evolve the organicmatter towards materials less rich in nitrogen andoxygen (Jenniskens et al 1993)

It is unknown how the organic content changeswith dust grain size and exposure to the plane-tary medium hence the fate of the organic mat-ter in the interplanetary medium is not knownHowever we know when the observed Leonids

were ejected from the parent comet in 1767 The090558 UT Leonid meteoroid never came closerto the Sun than 0976 AU and remained in the in-terplanetary medium only for 225 years (sevenorbits) most of that time far from the Sun

Might the organic matter have been lost in theEarthrsquos atmosphere early in the trajectory athigher altitudes when the meteoroid first warmedup to above the evaporation and decompositiontemperature of T 700 K This scenario was pro-posed by Elford et al (1997) who observed thatradar meteors are initially detected at heights upto 140 km They concluded that a more volatilecomponent than stony minerals must be ablatingat those heights The specific case of fast Leonidswas discussed by Steel (1998) who calculated thealtitude of evaporation for a range of small or-ganic compounds with relatively low sublimationtemperatures finding ablation heights in excessof 140 km for some compounds However thelight curve of bright Leonid fireballs above 136km was found to be very flat and starting at about200 km (Spurny et al 2000ab) which is incon-sistent with the light curves calculated by Elfordet al (1997) and Steel (1998) At 200 km the grainsare not warm enough to evaporate the organiccompounds that survived the vacuum of space

Indeed the Leonid studied here shows nosign of differential ablation of minerals in order oftheir volatility The elements Na Fe Mg and Caare lost in the same proportion (Fig 2) exceptfor the last wisp of matter after the catastrophicfragmentation during the terminal flare (a topicof future work) Sodium is at least partiallyfound in more volatile minerals than those con-taining magnesium and iron Though somesmaller Leonid meteoroids are known to losesodium early in their trajectory this is onlyfound in fragile cometary shower meteoroidssmall enough to expose the minerals efficientlyto the impinging air flow (BorovicIuml ka et al 1999)Sodium may be contained in the larger Leonidsat lower altitudes because of the limited timeavailable for inward diffusion of heat and out-ward diffusion of molecules Although the or-ganic and mineral components are separate en-tities they are intimately mixed down to verysmall size scales in comet Halley dust and thediffusion of those volatile compounds out of thegrains (if at all possible) will take time Such con-siderations are not included in the model of Steel(1998) Hence the organic matter is not lost earlyin flight

METEORS AND EXOGENOUS ORGANIC MOLECULES 75

The low CN abundance in the air plasma of the090558 UT Leonid meteoroid cannot be under-stood by merely allowing some fraction of nitro-gen to escape as N NH N2 NH3 NH2 NCONO HNO or HCN etc The configuration of Nin the complex organic matter of meteoroids isnot known However from meteorite and mi-crometeorite studies (where much organic matterhas been altered by aqueous alteration processes)we know that a significant fraction of the nitro-gen is contained in the organic framework (seereferences in Keller et al 1995 Greenberg 2000)Aqueous alteration tends to move nitrogen fromthe organic framework to functional groups suchas -NH In cometary matter no aqueous alter-ation has occurred So if N is in the frameworkin meteorite organics it should also exist in theorganic framework of cometary matter Coal-bound nitrogen also resides principally in hete-rocyclic ring moieties mostly five-memberedpyrrolic rings and six-membered pyridinic rings(Smith et al 1994 Zhang 2001) In pyrolysis stud-ies of tar such nitrogen is observed to escape pre-dominantly in the form of HCN and NH3 (Chenand Niksa 1992 Nelson et al 1992 Ledesma etal 1998) Ammonia is a product of amino com-pounds At the low pressure and high tempera-ture conditions in meteor plasma CN radicalsrather than HCN are stable and ammonia willbreak apart Small hydrocarbon radicals higherthan CN are so excited by their formation thatthey rapidly fission into lower products leavingonly the thermally stable single and double car-bon radicals unless prevented by rapid radiativecooling Carbon atoms are present too but atlower abundance because of thermodynamic con-siderations The carbon atoms react with N2 toform CN At 4400 K CN and N2 are only par-tially dissociated while O2 and CO (and CO2) arefully dissociated leading to high abundances ofCN if the plasma is in local thermodynamic equi-librium (Jenniskens et al 2000a) Only if oxygenplays a role in extracting the nitrogen after abla-tion then NOx compounds are expected to dom-inate Hence CN should be a significant productof the ablation of organic matter unless the or-ganic matter is released as larger compounds

We conclude that the most likely explanationfor the lack of CN molecules in the meteor plasmais that meteoric organic matter is lost in the formof a large range of relatively complex moleculesor perhaps as an amorphous carbon solid withembedded metal as proposed by Brownlee et al

(2002) Until now no evidence has been found ofsuch meteoric soot particles in the Earth envi-ronment and it remains unknown how well theheating conditions in the electron microscopeused in the experiment by Brownlee et al (2002)mimic the actual physical conditions during me-teoroid ablation

If the organic matter survives as large mole-cules this implies that bonds are prevented frombreaking by rapid radiative cooling or becauseheat is carried away by collisional fragments thattransfer much of the kinetic energy to the ambi-ent air The latter mechanism is the more likelyone and comes about because of conservation ofmomentum in a collision The result is that thelower mass fragment will carry most of the ki-netic energy and it will carry proportionallymore kinetic energy as the mass of the mainresidue increases This is an indirect form of thecollisional cooling that facilitates laser desorp-tion a technique widely used in the laboratory tostudy high-molecular-mass polyaromatic hydro-carbons in the gas phase and for detecting com-plex organic compounds in meteorites

Once released from the surface a single colli-sion will cause the molecule to slow down andlag behind relative to the meteoroid The mole-cule then is slowed to thermal speeds in another10 collisions and remains in the warm wake ofthe meteor for 50 collisions Only about 3 ofair collisions with atoms are inelastic (Oumlpik1958) For collisions with molecules that per-centage is higher but only a fraction of collisionswill result in chemical reactions Typical reactionsthat occur in the burning of propene for exam-ple are (eg Marinov et al 1996)

C3H6 1 O R C2H5 1 HCO (3)

C3H6 1 O R CH2CO 1 CH3 1 H (4)

C3H6 1 O R CH3CHCO 1 H 1 H (5)

Hydrogen can react in an early Earth-like atmos-phere with CO2 to form OH radicals which con-tinue to attack the organic matter via reactionssuch as

C3H6 1 OH R C3H6OH (6)

Radical attack can also lead to the growth of car-bon chains by polymerization reactions in therare case that organic compounds meet Hencethe result of chemistry in the meteor plasma is

JENNISKENS ET AL76

likely a product of relatively high molecular massthat is enriched in oxygen (and nitrogen) Suchcompounds are much more interesting for prebi-otic chemistry than the very-high-molecular-mass functional group-poor compounds foundin (micro-) meteorites

CONCLUSIONS

We conclude that complex organic matter inmeteoroids does not fully decompose into di-atomic constituents or atoms during the most in-tense phase of meteor ablation and that meteoricorganics are most likely lost in the form of largemolecular fragments cooled by radiative coolingand the loss of hot molecular fragments Theproduct of chemical interactions in the air plasmain the meteoroidrsquos wake is most likely a large va-riety of molecules and radicals that are chemicallyaltered functional group-enriched molecularfragments and condensation products

This result marks an alternative route to pre-biotic molecules from organic matter contributedto the Earth by particles at the peak of the massndashfrequency distribution of infalling matter At thetime of the origin of life there would have beentypically as much extraterrestrial matter impact-ing Earth in the form of dust as in the form oflarge asteroids and comets (Ceplecha 1992) ifthey arrived on elliptic orbits to have time to de-posit debris Moreover the organic content ofthose meteoroids should have been similar to thatof the impacting parent bodies with the excep-tion of the volatile compounds lost during dustejection This makes the product of meteor abla-tion and subsequent chemistry potentially thedominant source of complex organic moleculeson the early Earth at the time of the origin of life

ACKNOWLEDGMENTS

We thank George Flynn and an anonymous ref-eree for discussions that helped improve this pa-per The cooled CCD slit-less spectrograph wasdeveloped with support of the NASA Ames Re-search Centerrsquos Directorrsquos Discretionary Fundduring preparation of the Leonid MAC missionsWe thank Mike Koop (California Meteor Society)and Gary Palmer for technical support at variousstages in this project The 2001 Leonid MAC mis-sion was sponsored by NASArsquos Astrobiology

ASTID and Planetary Astronomy programs The2001 mission was executed by the USAF 418thFlight Test Squadron of Edwards AFB in Califor-nia with logistic support from the SETI Institute

ABBREVIATIONS

CCD charge coupled device FWHM full widthat half-maximum IDP interplanetary dust parti-cle MAC Multi-Instrument Aircraft CampaignUT universal time

REFERENCES

Abe S Yano H Ebizuka N and Watanabe J-I (2000)First results of high-definition TV spectroscopic obser-vations of the 1999 Leonid meteor shower Earth MoonPlanets 82-83 369ndash377

Anders E (1989) Pre-biotic organic matter from cometsand asteroids Nature 342 255ndash257

Betlem H Jenniskens P Spurny P Docters vanLeeuwen G Miskotte K Ter Kuile CR Zerubin Pand Angelos C (2000) Precise trajectories and orbits ofmeteoroids from the 1999 Leonid meteor storm EarthMoon Planets 82-83 277ndash284

BorovicIuml ka J Stork R and Bocek J (1999) First resultsfrom video spectroscopy of 1998 Leonid meteors Me-teoritics Planet Sci 34 987ndash994

Boyd ID (2000) Computation of atmospheric entry flowabout a Leonid meteoroid Earth Moon Planets 82-8393ndash108

Brownlee DE Joswiak DJ Kress ME Taylor S andBradley J (2002) Survival of carbon in moderately tostrongly heated IDPs and micrometeorites [abstract1786] In 33rd Lunar and Planetary Science Conference LPIContribution No 1109 Lunar and Planetary InstituteHouston

Ceplecha Z (1971) Spectral data on terminal flare andwake of double-station meteor No 38421 (OndrejovApril 21 1963) Bull Astron Inst Czech 22 219ndash304

Ceplecha Z (1992) Earth influx of interplanetary bodiesAstron Astrophys 263 361ndash366

Chen JC and Niksa S (1992) Coal devolatilization dur-ing rapid transient heating 1 Primary devolatilizationEnergy Fuels 6 254ndash264

Chyba CF and Sagan C (1992) Endogenous productionexogenous delivery and impact shock synthesis of or-ganic molecules An inventory for the origins of lifeNature 355 125ndash132

Chyba CF and Sagan C (1997) Comets as a source ofprebiotic organic molecules for the early Earth InComets and the Origin and Evolution of Life edited by PJThomas CF Chyba and CP McKay Springer VerlagNew York pp 147ndash173

Delsemme AH (1991) Nature and history of the organiccompounds in comets an astrophysical view In Astro-

METEORS AND EXOGENOUS ORGANIC MOLECULES 77

start of the exposure No correction was made forthis (small) contribution to the main spectrum

After summing all relevant rows we foundthat emission lines were on top of a continuumemission identified as the First Positive Band ofN2 observed in second order (Fig 5) No knownsecond or third order molecular band from at-mospheric emissions other than that of CN wasexpected in the region around the CN band headAfter correction for second order instrument re-sponse (Fig 3) we fit the N2 band contour to thewhole spectrum The N2 band contour definedthe background level A small correction for in-strument response to the remaining iron lines wasmade by subtracting the N2 band contour andcorrecting the resulting spectrum in a manner ap-propriate for the third order response (Fig 3)

The CN molecular band

Once instrument and background correctionswere made we focused our attention on the thirdorder spectrum around 388 nm (Fig 6) The newthird order spectrum resolved the weaker ironlines considerably as compared with our earlierfirst order spectra (Rairden et al 2000) The ex-pected Fe line emission spectrum shown as adashed line in Fig 6 was calculated for excita-tion temperatures of 4300 and 4500 K using allneutral iron lines in the NIST Atomic SpectraDatabase Version 20 (NIST Standard ReferenceDatabase 78) A comparison of the observed andcalculated emission lines established their iden-

tity and showed that weak iron lines did not con-tribute to the background continuum It shouldbe noted that second order iron lines are presentelsewhere in the spectrum

As indicated in Fig 6 two iron lines bracket theposition of the CN molecular band Because ofthe absence of any additional faint line emissionsin this region an upper limit on the CN abun-dance could be established Though the CN bandmight have in fact been detected as a small bumpit is at the noise level of the spectrum

The relative abundance of CN molecules versus iron atoms was calculated using theNEQAIR2 model of heated air in thermodynamicequilibrium (Laux 1993 Park et al 1997) The CNband profile for a resolution of 18 Aring FWHM isshown in Fig 6 The band head at 38844 nm isunresolved with its width reflecting the instru-ment profile With ICN equal to the total inte-grated emission intensity of CN in the range375ndash393 nm and IFe equal to the total integratedemission intensity of all third order Fe lines in therange 3813ndash3867 nm we calculated for T 5 4500that IFenFe 5 324 3 10216 Wsr and ICNnCN 5558 3 10217 Wsr where nFe is the total concen-tration of Fe atoms in cm23 and nCN is the totalconcentration of CN moleculescm3 For T 54300 K IFenFe 5 213 3 10216 Wsr and ICNnCN 5 393 3 10217 Wsr To calculate these val-ues we assumed that the emitting levels were in

METEORS AND EXOGENOUS ORGANIC MOLECULES 73

FIG 5 Extracted spectrum of Fig 2 after correction forinstrument response and vigneting The dashed lineshows the contribution from the first positive band of N2in second order

FIG 6 Detail of Fig 5 showing the position of the CNband and the iron lines in third order A model spec-trum at 4300 K of iron atom line emission is shown by adashed line The continuum background is the band con-tour of N2 in second order The thick line is a theoreticalprofile of the CN band for air plasma at 4300 K and in-strumental broadening FWHM 5 18 Aring

Boltzmann equilibrium with the ground state ofFe and CN at the given temperature These areconcentrations per total Fe atoms and total CNmolecules (that is not per molecule of CN in theexcited electronic B state)

From these calculations we found that the ra-tio of the total number of atoms of CN and Fe inthe meteor plasma was nCNnFe 0017 for an ex-citation temperature of 4500 K (or 0019 for anexcitation temperature of 4300 K) The largest un-certainty was caused by the position of the back-ground level However given the good fit of theN2 band contour to the overall spectrum short-ward of 590 nm (Fig 5) a conservative error ofless than 50 places the upper limit at nCNnFe 003

An even more precise upper limit was derivedfrom a Geminid spectrum obtained during the2001 Geminid shower with the same instrumentdeployed from a ground site at Fremont Peak Ob-servatory California (Fig 7) For that spectrumwe calculated for an excitation temperature of4500 K that nCNnFe 0011 (or 0012 at 4300 K)However Geminid meteoroids frequently ap-proach the Sun to within 014 AU which heatsthe meteoroids every 16 years to at least T 5 510K (the temperature of small zodiacal cloud mete-oroids) and possibly as high as 1800 K if thelarger meteoroids are more like a black body(Levasseur-Regourd et al 2000) This is hotenough to change their morphology into morecompact objects and can result in the loss of someor all of the organic matter

The more intense Geminid spectrum confirmsthat no weak second or third order atomic linesof any kind are present at the position of the CNband Hence we conclude that the meteor plasmaresponsible for most of the Leonid meteorrsquos emis-sion contains 1 CN molecule per 30 Fe atomsThis result improves by more than a factor of 10the upper limit of 1 CN per 3 Fe atoms found ear-lier by Rairden et al (2000) These numbers aresignificantly less than expected If all of the ni-trogen in the complex organic matter of the cometdust was decomposed in the form of CN mole-cules this would give nCNnFe 5 079 6 002based on the observed nitrogen abundance ofcomplex organic matter in small dust grains ofcomet 1PHalley (Delsemme 1991) The ob-served abundance is a factor of 26 lower

CN can also be formed by reactions in the me-teor air plasma involving CO2 and N2 in the am-bient atmosphere Though CN formation via suchaerothermochemistry is not well understood inthe case of thermodynamic equilibrium therewould not be enough CN made from this mech-anism to be detectable The calculated CN abun-dance for a 4300 K air plasma in local thermo-dynamic equilibrium at 95 km altitude equals10 3 105 molecules of CNcm3 The 22 magni-tude Leonid meteor studied here has a mass ofabout 044 g of which 0033 g is iron (Delsemme1991) That iron is ablated and distributed over apath length of about 17 km (Betlem et al 2000)in a cylindrical volume of at least a 2-m radius(Boyd 2000) which indicates a mean density ofnFe 5 2 3 109 atomscm3 and an expected nCNnFe of about 5 3 1025 a factor of 6 less than ourdetection limit

DISCUSSION

There are several reasons why the abundanceof CN in the meteor plasma can be a factor of 26lower than expected from the decomposition ofmeteoric organic matter during ablation (1) theabundance of organic matter in the dust of55PTempel-Tuttle is not known (2) the abun-dance of nitrogen in that organic matter is un-known (3) the fate of organic matter in the in-terplanetary medium due to space weathering isnot known (4) the organic matter could havebeen lost at higher altitude (earlier in flight) thanwhere the measurement was made (5) theprocess of ablation could favor the release of ni-

JENNISKENS ET AL74

FIG 7 Detail of the 084127 UT Geminid meteor spec-trum observed on December 13 2001 with symbols andlines the same as in Fig 6

trogen in atomic form or in the form of some othersmall molecules or (6) the nitrogen could survivein the form of larger organic compounds that arenot detected The latter possibility could providea new pathway towards prebiotic compoundsWe will now consider the significance and im-plication of these arguments one by one

Since the abundance of organic matter in thedust of 55PTemplel-Tuttle is not known it ispossible that there was little or no organic mat-ter in the Leonid particles to begin with Conse-quently the absence of nitrogen-bearing organicmatter in the Leonid meteoroids may have re-sulted in our inability to detect CN The onlyother measurement that could have detected ni-trogen-bearing organic matter in the dust (as op-posed to molecules in the gas phase) is in situmass spectroscopy The Leonid parent comet hasnot been visited In the coma of comet 1PHal-ley nitrogen-bearing organic matter was detectedin particles (the CHON particles) analyzed by themass spectrometers on the GIOTTO and VEGAspacecraft Because comet Halley is not the par-ent of the Leonid stream it is recommended thatfuture work focus on the meteoroids of the Ori-onid and eta-Aquarid streams that do originatefrom comet Halley Though bright meteors fromthese showers are much more difficult to observemuch the same result would be expected TheLeonid parent comet is remarkable only becauseit passes so close to Earthrsquos orbit Most impor-tantly since 55PTempel-Tuttle is in a Halley-type orbit and it should originate from the Oortcloud just as 1PHalley there is no reason to ex-pect that it is less rich in organic matter

For the same reason there is no reason to ex-pect that the nitrogen content or composition ofthe organic matter would differ from that of1PHalley Halley dust is as rich in nitrogen thanprimitive meteorites Though the nitrogen con-tent in the organic matter of comet Halley hasbeen questioned because of calibration problemswith the detectors there is no reason to expectthere to be more nitrogen in the less primitive me-teorites The heating and irradiation that alter theorganics in meteorites tend to evolve the organicmatter towards materials less rich in nitrogen andoxygen (Jenniskens et al 1993)

It is unknown how the organic content changeswith dust grain size and exposure to the plane-tary medium hence the fate of the organic mat-ter in the interplanetary medium is not knownHowever we know when the observed Leonids

were ejected from the parent comet in 1767 The090558 UT Leonid meteoroid never came closerto the Sun than 0976 AU and remained in the in-terplanetary medium only for 225 years (sevenorbits) most of that time far from the Sun

Might the organic matter have been lost in theEarthrsquos atmosphere early in the trajectory athigher altitudes when the meteoroid first warmedup to above the evaporation and decompositiontemperature of T 700 K This scenario was pro-posed by Elford et al (1997) who observed thatradar meteors are initially detected at heights upto 140 km They concluded that a more volatilecomponent than stony minerals must be ablatingat those heights The specific case of fast Leonidswas discussed by Steel (1998) who calculated thealtitude of evaporation for a range of small or-ganic compounds with relatively low sublimationtemperatures finding ablation heights in excessof 140 km for some compounds However thelight curve of bright Leonid fireballs above 136km was found to be very flat and starting at about200 km (Spurny et al 2000ab) which is incon-sistent with the light curves calculated by Elfordet al (1997) and Steel (1998) At 200 km the grainsare not warm enough to evaporate the organiccompounds that survived the vacuum of space

Indeed the Leonid studied here shows nosign of differential ablation of minerals in order oftheir volatility The elements Na Fe Mg and Caare lost in the same proportion (Fig 2) exceptfor the last wisp of matter after the catastrophicfragmentation during the terminal flare (a topicof future work) Sodium is at least partiallyfound in more volatile minerals than those con-taining magnesium and iron Though somesmaller Leonid meteoroids are known to losesodium early in their trajectory this is onlyfound in fragile cometary shower meteoroidssmall enough to expose the minerals efficientlyto the impinging air flow (BorovicIuml ka et al 1999)Sodium may be contained in the larger Leonidsat lower altitudes because of the limited timeavailable for inward diffusion of heat and out-ward diffusion of molecules Although the or-ganic and mineral components are separate en-tities they are intimately mixed down to verysmall size scales in comet Halley dust and thediffusion of those volatile compounds out of thegrains (if at all possible) will take time Such con-siderations are not included in the model of Steel(1998) Hence the organic matter is not lost earlyin flight

METEORS AND EXOGENOUS ORGANIC MOLECULES 75

The low CN abundance in the air plasma of the090558 UT Leonid meteoroid cannot be under-stood by merely allowing some fraction of nitro-gen to escape as N NH N2 NH3 NH2 NCONO HNO or HCN etc The configuration of Nin the complex organic matter of meteoroids isnot known However from meteorite and mi-crometeorite studies (where much organic matterhas been altered by aqueous alteration processes)we know that a significant fraction of the nitro-gen is contained in the organic framework (seereferences in Keller et al 1995 Greenberg 2000)Aqueous alteration tends to move nitrogen fromthe organic framework to functional groups suchas -NH In cometary matter no aqueous alter-ation has occurred So if N is in the frameworkin meteorite organics it should also exist in theorganic framework of cometary matter Coal-bound nitrogen also resides principally in hete-rocyclic ring moieties mostly five-memberedpyrrolic rings and six-membered pyridinic rings(Smith et al 1994 Zhang 2001) In pyrolysis stud-ies of tar such nitrogen is observed to escape pre-dominantly in the form of HCN and NH3 (Chenand Niksa 1992 Nelson et al 1992 Ledesma etal 1998) Ammonia is a product of amino com-pounds At the low pressure and high tempera-ture conditions in meteor plasma CN radicalsrather than HCN are stable and ammonia willbreak apart Small hydrocarbon radicals higherthan CN are so excited by their formation thatthey rapidly fission into lower products leavingonly the thermally stable single and double car-bon radicals unless prevented by rapid radiativecooling Carbon atoms are present too but atlower abundance because of thermodynamic con-siderations The carbon atoms react with N2 toform CN At 4400 K CN and N2 are only par-tially dissociated while O2 and CO (and CO2) arefully dissociated leading to high abundances ofCN if the plasma is in local thermodynamic equi-librium (Jenniskens et al 2000a) Only if oxygenplays a role in extracting the nitrogen after abla-tion then NOx compounds are expected to dom-inate Hence CN should be a significant productof the ablation of organic matter unless the or-ganic matter is released as larger compounds

We conclude that the most likely explanationfor the lack of CN molecules in the meteor plasmais that meteoric organic matter is lost in the formof a large range of relatively complex moleculesor perhaps as an amorphous carbon solid withembedded metal as proposed by Brownlee et al

(2002) Until now no evidence has been found ofsuch meteoric soot particles in the Earth envi-ronment and it remains unknown how well theheating conditions in the electron microscopeused in the experiment by Brownlee et al (2002)mimic the actual physical conditions during me-teoroid ablation

If the organic matter survives as large mole-cules this implies that bonds are prevented frombreaking by rapid radiative cooling or becauseheat is carried away by collisional fragments thattransfer much of the kinetic energy to the ambi-ent air The latter mechanism is the more likelyone and comes about because of conservation ofmomentum in a collision The result is that thelower mass fragment will carry most of the ki-netic energy and it will carry proportionallymore kinetic energy as the mass of the mainresidue increases This is an indirect form of thecollisional cooling that facilitates laser desorp-tion a technique widely used in the laboratory tostudy high-molecular-mass polyaromatic hydro-carbons in the gas phase and for detecting com-plex organic compounds in meteorites

Once released from the surface a single colli-sion will cause the molecule to slow down andlag behind relative to the meteoroid The mole-cule then is slowed to thermal speeds in another10 collisions and remains in the warm wake ofthe meteor for 50 collisions Only about 3 ofair collisions with atoms are inelastic (Oumlpik1958) For collisions with molecules that per-centage is higher but only a fraction of collisionswill result in chemical reactions Typical reactionsthat occur in the burning of propene for exam-ple are (eg Marinov et al 1996)

C3H6 1 O R C2H5 1 HCO (3)

C3H6 1 O R CH2CO 1 CH3 1 H (4)

C3H6 1 O R CH3CHCO 1 H 1 H (5)

Hydrogen can react in an early Earth-like atmos-phere with CO2 to form OH radicals which con-tinue to attack the organic matter via reactionssuch as

C3H6 1 OH R C3H6OH (6)

Radical attack can also lead to the growth of car-bon chains by polymerization reactions in therare case that organic compounds meet Hencethe result of chemistry in the meteor plasma is

JENNISKENS ET AL76

likely a product of relatively high molecular massthat is enriched in oxygen (and nitrogen) Suchcompounds are much more interesting for prebi-otic chemistry than the very-high-molecular-mass functional group-poor compounds foundin (micro-) meteorites

CONCLUSIONS

We conclude that complex organic matter inmeteoroids does not fully decompose into di-atomic constituents or atoms during the most in-tense phase of meteor ablation and that meteoricorganics are most likely lost in the form of largemolecular fragments cooled by radiative coolingand the loss of hot molecular fragments Theproduct of chemical interactions in the air plasmain the meteoroidrsquos wake is most likely a large va-riety of molecules and radicals that are chemicallyaltered functional group-enriched molecularfragments and condensation products

This result marks an alternative route to pre-biotic molecules from organic matter contributedto the Earth by particles at the peak of the massndashfrequency distribution of infalling matter At thetime of the origin of life there would have beentypically as much extraterrestrial matter impact-ing Earth in the form of dust as in the form oflarge asteroids and comets (Ceplecha 1992) ifthey arrived on elliptic orbits to have time to de-posit debris Moreover the organic content ofthose meteoroids should have been similar to thatof the impacting parent bodies with the excep-tion of the volatile compounds lost during dustejection This makes the product of meteor abla-tion and subsequent chemistry potentially thedominant source of complex organic moleculeson the early Earth at the time of the origin of life

ACKNOWLEDGMENTS

We thank George Flynn and an anonymous ref-eree for discussions that helped improve this pa-per The cooled CCD slit-less spectrograph wasdeveloped with support of the NASA Ames Re-search Centerrsquos Directorrsquos Discretionary Fundduring preparation of the Leonid MAC missionsWe thank Mike Koop (California Meteor Society)and Gary Palmer for technical support at variousstages in this project The 2001 Leonid MAC mis-sion was sponsored by NASArsquos Astrobiology

ASTID and Planetary Astronomy programs The2001 mission was executed by the USAF 418thFlight Test Squadron of Edwards AFB in Califor-nia with logistic support from the SETI Institute

ABBREVIATIONS

CCD charge coupled device FWHM full widthat half-maximum IDP interplanetary dust parti-cle MAC Multi-Instrument Aircraft CampaignUT universal time

REFERENCES

Abe S Yano H Ebizuka N and Watanabe J-I (2000)First results of high-definition TV spectroscopic obser-vations of the 1999 Leonid meteor shower Earth MoonPlanets 82-83 369ndash377

Anders E (1989) Pre-biotic organic matter from cometsand asteroids Nature 342 255ndash257

Betlem H Jenniskens P Spurny P Docters vanLeeuwen G Miskotte K Ter Kuile CR Zerubin Pand Angelos C (2000) Precise trajectories and orbits ofmeteoroids from the 1999 Leonid meteor storm EarthMoon Planets 82-83 277ndash284

BorovicIuml ka J Stork R and Bocek J (1999) First resultsfrom video spectroscopy of 1998 Leonid meteors Me-teoritics Planet Sci 34 987ndash994

Boyd ID (2000) Computation of atmospheric entry flowabout a Leonid meteoroid Earth Moon Planets 82-8393ndash108

Brownlee DE Joswiak DJ Kress ME Taylor S andBradley J (2002) Survival of carbon in moderately tostrongly heated IDPs and micrometeorites [abstract1786] In 33rd Lunar and Planetary Science Conference LPIContribution No 1109 Lunar and Planetary InstituteHouston

Ceplecha Z (1971) Spectral data on terminal flare andwake of double-station meteor No 38421 (OndrejovApril 21 1963) Bull Astron Inst Czech 22 219ndash304

Ceplecha Z (1992) Earth influx of interplanetary bodiesAstron Astrophys 263 361ndash366

Chen JC and Niksa S (1992) Coal devolatilization dur-ing rapid transient heating 1 Primary devolatilizationEnergy Fuels 6 254ndash264

Chyba CF and Sagan C (1992) Endogenous productionexogenous delivery and impact shock synthesis of or-ganic molecules An inventory for the origins of lifeNature 355 125ndash132

Chyba CF and Sagan C (1997) Comets as a source ofprebiotic organic molecules for the early Earth InComets and the Origin and Evolution of Life edited by PJThomas CF Chyba and CP McKay Springer VerlagNew York pp 147ndash173

Delsemme AH (1991) Nature and history of the organiccompounds in comets an astrophysical view In Astro-

METEORS AND EXOGENOUS ORGANIC MOLECULES 77

Boltzmann equilibrium with the ground state ofFe and CN at the given temperature These areconcentrations per total Fe atoms and total CNmolecules (that is not per molecule of CN in theexcited electronic B state)

From these calculations we found that the ra-tio of the total number of atoms of CN and Fe inthe meteor plasma was nCNnFe 0017 for an ex-citation temperature of 4500 K (or 0019 for anexcitation temperature of 4300 K) The largest un-certainty was caused by the position of the back-ground level However given the good fit of theN2 band contour to the overall spectrum short-ward of 590 nm (Fig 5) a conservative error ofless than 50 places the upper limit at nCNnFe 003

An even more precise upper limit was derivedfrom a Geminid spectrum obtained during the2001 Geminid shower with the same instrumentdeployed from a ground site at Fremont Peak Ob-servatory California (Fig 7) For that spectrumwe calculated for an excitation temperature of4500 K that nCNnFe 0011 (or 0012 at 4300 K)However Geminid meteoroids frequently ap-proach the Sun to within 014 AU which heatsthe meteoroids every 16 years to at least T 5 510K (the temperature of small zodiacal cloud mete-oroids) and possibly as high as 1800 K if thelarger meteoroids are more like a black body(Levasseur-Regourd et al 2000) This is hotenough to change their morphology into morecompact objects and can result in the loss of someor all of the organic matter

The more intense Geminid spectrum confirmsthat no weak second or third order atomic linesof any kind are present at the position of the CNband Hence we conclude that the meteor plasmaresponsible for most of the Leonid meteorrsquos emis-sion contains 1 CN molecule per 30 Fe atomsThis result improves by more than a factor of 10the upper limit of 1 CN per 3 Fe atoms found ear-lier by Rairden et al (2000) These numbers aresignificantly less than expected If all of the ni-trogen in the complex organic matter of the cometdust was decomposed in the form of CN mole-cules this would give nCNnFe 5 079 6 002based on the observed nitrogen abundance ofcomplex organic matter in small dust grains ofcomet 1PHalley (Delsemme 1991) The ob-served abundance is a factor of 26 lower

CN can also be formed by reactions in the me-teor air plasma involving CO2 and N2 in the am-bient atmosphere Though CN formation via suchaerothermochemistry is not well understood inthe case of thermodynamic equilibrium therewould not be enough CN made from this mech-anism to be detectable The calculated CN abun-dance for a 4300 K air plasma in local thermo-dynamic equilibrium at 95 km altitude equals10 3 105 molecules of CNcm3 The 22 magni-tude Leonid meteor studied here has a mass ofabout 044 g of which 0033 g is iron (Delsemme1991) That iron is ablated and distributed over apath length of about 17 km (Betlem et al 2000)in a cylindrical volume of at least a 2-m radius(Boyd 2000) which indicates a mean density ofnFe 5 2 3 109 atomscm3 and an expected nCNnFe of about 5 3 1025 a factor of 6 less than ourdetection limit

DISCUSSION

There are several reasons why the abundanceof CN in the meteor plasma can be a factor of 26lower than expected from the decomposition ofmeteoric organic matter during ablation (1) theabundance of organic matter in the dust of55PTempel-Tuttle is not known (2) the abun-dance of nitrogen in that organic matter is un-known (3) the fate of organic matter in the in-terplanetary medium due to space weathering isnot known (4) the organic matter could havebeen lost at higher altitude (earlier in flight) thanwhere the measurement was made (5) theprocess of ablation could favor the release of ni-

JENNISKENS ET AL74

FIG 7 Detail of the 084127 UT Geminid meteor spec-trum observed on December 13 2001 with symbols andlines the same as in Fig 6

trogen in atomic form or in the form of some othersmall molecules or (6) the nitrogen could survivein the form of larger organic compounds that arenot detected The latter possibility could providea new pathway towards prebiotic compoundsWe will now consider the significance and im-plication of these arguments one by one

Since the abundance of organic matter in thedust of 55PTemplel-Tuttle is not known it ispossible that there was little or no organic mat-ter in the Leonid particles to begin with Conse-quently the absence of nitrogen-bearing organicmatter in the Leonid meteoroids may have re-sulted in our inability to detect CN The onlyother measurement that could have detected ni-trogen-bearing organic matter in the dust (as op-posed to molecules in the gas phase) is in situmass spectroscopy The Leonid parent comet hasnot been visited In the coma of comet 1PHal-ley nitrogen-bearing organic matter was detectedin particles (the CHON particles) analyzed by themass spectrometers on the GIOTTO and VEGAspacecraft Because comet Halley is not the par-ent of the Leonid stream it is recommended thatfuture work focus on the meteoroids of the Ori-onid and eta-Aquarid streams that do originatefrom comet Halley Though bright meteors fromthese showers are much more difficult to observemuch the same result would be expected TheLeonid parent comet is remarkable only becauseit passes so close to Earthrsquos orbit Most impor-tantly since 55PTempel-Tuttle is in a Halley-type orbit and it should originate from the Oortcloud just as 1PHalley there is no reason to ex-pect that it is less rich in organic matter

For the same reason there is no reason to ex-pect that the nitrogen content or composition ofthe organic matter would differ from that of1PHalley Halley dust is as rich in nitrogen thanprimitive meteorites Though the nitrogen con-tent in the organic matter of comet Halley hasbeen questioned because of calibration problemswith the detectors there is no reason to expectthere to be more nitrogen in the less primitive me-teorites The heating and irradiation that alter theorganics in meteorites tend to evolve the organicmatter towards materials less rich in nitrogen andoxygen (Jenniskens et al 1993)

It is unknown how the organic content changeswith dust grain size and exposure to the plane-tary medium hence the fate of the organic mat-ter in the interplanetary medium is not knownHowever we know when the observed Leonids

were ejected from the parent comet in 1767 The090558 UT Leonid meteoroid never came closerto the Sun than 0976 AU and remained in the in-terplanetary medium only for 225 years (sevenorbits) most of that time far from the Sun

Might the organic matter have been lost in theEarthrsquos atmosphere early in the trajectory athigher altitudes when the meteoroid first warmedup to above the evaporation and decompositiontemperature of T 700 K This scenario was pro-posed by Elford et al (1997) who observed thatradar meteors are initially detected at heights upto 140 km They concluded that a more volatilecomponent than stony minerals must be ablatingat those heights The specific case of fast Leonidswas discussed by Steel (1998) who calculated thealtitude of evaporation for a range of small or-ganic compounds with relatively low sublimationtemperatures finding ablation heights in excessof 140 km for some compounds However thelight curve of bright Leonid fireballs above 136km was found to be very flat and starting at about200 km (Spurny et al 2000ab) which is incon-sistent with the light curves calculated by Elfordet al (1997) and Steel (1998) At 200 km the grainsare not warm enough to evaporate the organiccompounds that survived the vacuum of space

Indeed the Leonid studied here shows nosign of differential ablation of minerals in order oftheir volatility The elements Na Fe Mg and Caare lost in the same proportion (Fig 2) exceptfor the last wisp of matter after the catastrophicfragmentation during the terminal flare (a topicof future work) Sodium is at least partiallyfound in more volatile minerals than those con-taining magnesium and iron Though somesmaller Leonid meteoroids are known to losesodium early in their trajectory this is onlyfound in fragile cometary shower meteoroidssmall enough to expose the minerals efficientlyto the impinging air flow (BorovicIuml ka et al 1999)Sodium may be contained in the larger Leonidsat lower altitudes because of the limited timeavailable for inward diffusion of heat and out-ward diffusion of molecules Although the or-ganic and mineral components are separate en-tities they are intimately mixed down to verysmall size scales in comet Halley dust and thediffusion of those volatile compounds out of thegrains (if at all possible) will take time Such con-siderations are not included in the model of Steel(1998) Hence the organic matter is not lost earlyin flight

METEORS AND EXOGENOUS ORGANIC MOLECULES 75

The low CN abundance in the air plasma of the090558 UT Leonid meteoroid cannot be under-stood by merely allowing some fraction of nitro-gen to escape as N NH N2 NH3 NH2 NCONO HNO or HCN etc The configuration of Nin the complex organic matter of meteoroids isnot known However from meteorite and mi-crometeorite studies (where much organic matterhas been altered by aqueous alteration processes)we know that a significant fraction of the nitro-gen is contained in the organic framework (seereferences in Keller et al 1995 Greenberg 2000)Aqueous alteration tends to move nitrogen fromthe organic framework to functional groups suchas -NH In cometary matter no aqueous alter-ation has occurred So if N is in the frameworkin meteorite organics it should also exist in theorganic framework of cometary matter Coal-bound nitrogen also resides principally in hete-rocyclic ring moieties mostly five-memberedpyrrolic rings and six-membered pyridinic rings(Smith et al 1994 Zhang 2001) In pyrolysis stud-ies of tar such nitrogen is observed to escape pre-dominantly in the form of HCN and NH3 (Chenand Niksa 1992 Nelson et al 1992 Ledesma etal 1998) Ammonia is a product of amino com-pounds At the low pressure and high tempera-ture conditions in meteor plasma CN radicalsrather than HCN are stable and ammonia willbreak apart Small hydrocarbon radicals higherthan CN are so excited by their formation thatthey rapidly fission into lower products leavingonly the thermally stable single and double car-bon radicals unless prevented by rapid radiativecooling Carbon atoms are present too but atlower abundance because of thermodynamic con-siderations The carbon atoms react with N2 toform CN At 4400 K CN and N2 are only par-tially dissociated while O2 and CO (and CO2) arefully dissociated leading to high abundances ofCN if the plasma is in local thermodynamic equi-librium (Jenniskens et al 2000a) Only if oxygenplays a role in extracting the nitrogen after abla-tion then NOx compounds are expected to dom-inate Hence CN should be a significant productof the ablation of organic matter unless the or-ganic matter is released as larger compounds

We conclude that the most likely explanationfor the lack of CN molecules in the meteor plasmais that meteoric organic matter is lost in the formof a large range of relatively complex moleculesor perhaps as an amorphous carbon solid withembedded metal as proposed by Brownlee et al

(2002) Until now no evidence has been found ofsuch meteoric soot particles in the Earth envi-ronment and it remains unknown how well theheating conditions in the electron microscopeused in the experiment by Brownlee et al (2002)mimic the actual physical conditions during me-teoroid ablation

If the organic matter survives as large mole-cules this implies that bonds are prevented frombreaking by rapid radiative cooling or becauseheat is carried away by collisional fragments thattransfer much of the kinetic energy to the ambi-ent air The latter mechanism is the more likelyone and comes about because of conservation ofmomentum in a collision The result is that thelower mass fragment will carry most of the ki-netic energy and it will carry proportionallymore kinetic energy as the mass of the mainresidue increases This is an indirect form of thecollisional cooling that facilitates laser desorp-tion a technique widely used in the laboratory tostudy high-molecular-mass polyaromatic hydro-carbons in the gas phase and for detecting com-plex organic compounds in meteorites

Once released from the surface a single colli-sion will cause the molecule to slow down andlag behind relative to the meteoroid The mole-cule then is slowed to thermal speeds in another10 collisions and remains in the warm wake ofthe meteor for 50 collisions Only about 3 ofair collisions with atoms are inelastic (Oumlpik1958) For collisions with molecules that per-centage is higher but only a fraction of collisionswill result in chemical reactions Typical reactionsthat occur in the burning of propene for exam-ple are (eg Marinov et al 1996)

C3H6 1 O R C2H5 1 HCO (3)

C3H6 1 O R CH2CO 1 CH3 1 H (4)

C3H6 1 O R CH3CHCO 1 H 1 H (5)

Hydrogen can react in an early Earth-like atmos-phere with CO2 to form OH radicals which con-tinue to attack the organic matter via reactionssuch as

C3H6 1 OH R C3H6OH (6)

Radical attack can also lead to the growth of car-bon chains by polymerization reactions in therare case that organic compounds meet Hencethe result of chemistry in the meteor plasma is

JENNISKENS ET AL76

likely a product of relatively high molecular massthat is enriched in oxygen (and nitrogen) Suchcompounds are much more interesting for prebi-otic chemistry than the very-high-molecular-mass functional group-poor compounds foundin (micro-) meteorites

CONCLUSIONS

We conclude that complex organic matter inmeteoroids does not fully decompose into di-atomic constituents or atoms during the most in-tense phase of meteor ablation and that meteoricorganics are most likely lost in the form of largemolecular fragments cooled by radiative coolingand the loss of hot molecular fragments Theproduct of chemical interactions in the air plasmain the meteoroidrsquos wake is most likely a large va-riety of molecules and radicals that are chemicallyaltered functional group-enriched molecularfragments and condensation products

This result marks an alternative route to pre-biotic molecules from organic matter contributedto the Earth by particles at the peak of the massndashfrequency distribution of infalling matter At thetime of the origin of life there would have beentypically as much extraterrestrial matter impact-ing Earth in the form of dust as in the form oflarge asteroids and comets (Ceplecha 1992) ifthey arrived on elliptic orbits to have time to de-posit debris Moreover the organic content ofthose meteoroids should have been similar to thatof the impacting parent bodies with the excep-tion of the volatile compounds lost during dustejection This makes the product of meteor abla-tion and subsequent chemistry potentially thedominant source of complex organic moleculeson the early Earth at the time of the origin of life

ACKNOWLEDGMENTS

We thank George Flynn and an anonymous ref-eree for discussions that helped improve this pa-per The cooled CCD slit-less spectrograph wasdeveloped with support of the NASA Ames Re-search Centerrsquos Directorrsquos Discretionary Fundduring preparation of the Leonid MAC missionsWe thank Mike Koop (California Meteor Society)and Gary Palmer for technical support at variousstages in this project The 2001 Leonid MAC mis-sion was sponsored by NASArsquos Astrobiology

ASTID and Planetary Astronomy programs The2001 mission was executed by the USAF 418thFlight Test Squadron of Edwards AFB in Califor-nia with logistic support from the SETI Institute

ABBREVIATIONS

CCD charge coupled device FWHM full widthat half-maximum IDP interplanetary dust parti-cle MAC Multi-Instrument Aircraft CampaignUT universal time

REFERENCES

Abe S Yano H Ebizuka N and Watanabe J-I (2000)First results of high-definition TV spectroscopic obser-vations of the 1999 Leonid meteor shower Earth MoonPlanets 82-83 369ndash377

Anders E (1989) Pre-biotic organic matter from cometsand asteroids Nature 342 255ndash257

Betlem H Jenniskens P Spurny P Docters vanLeeuwen G Miskotte K Ter Kuile CR Zerubin Pand Angelos C (2000) Precise trajectories and orbits ofmeteoroids from the 1999 Leonid meteor storm EarthMoon Planets 82-83 277ndash284

BorovicIuml ka J Stork R and Bocek J (1999) First resultsfrom video spectroscopy of 1998 Leonid meteors Me-teoritics Planet Sci 34 987ndash994

Boyd ID (2000) Computation of atmospheric entry flowabout a Leonid meteoroid Earth Moon Planets 82-8393ndash108

Brownlee DE Joswiak DJ Kress ME Taylor S andBradley J (2002) Survival of carbon in moderately tostrongly heated IDPs and micrometeorites [abstract1786] In 33rd Lunar and Planetary Science Conference LPIContribution No 1109 Lunar and Planetary InstituteHouston

Ceplecha Z (1971) Spectral data on terminal flare andwake of double-station meteor No 38421 (OndrejovApril 21 1963) Bull Astron Inst Czech 22 219ndash304

Ceplecha Z (1992) Earth influx of interplanetary bodiesAstron Astrophys 263 361ndash366

Chen JC and Niksa S (1992) Coal devolatilization dur-ing rapid transient heating 1 Primary devolatilizationEnergy Fuels 6 254ndash264

Chyba CF and Sagan C (1992) Endogenous productionexogenous delivery and impact shock synthesis of or-ganic molecules An inventory for the origins of lifeNature 355 125ndash132

Chyba CF and Sagan C (1997) Comets as a source ofprebiotic organic molecules for the early Earth InComets and the Origin and Evolution of Life edited by PJThomas CF Chyba and CP McKay Springer VerlagNew York pp 147ndash173

Delsemme AH (1991) Nature and history of the organiccompounds in comets an astrophysical view In Astro-

METEORS AND EXOGENOUS ORGANIC MOLECULES 77

trogen in atomic form or in the form of some othersmall molecules or (6) the nitrogen could survivein the form of larger organic compounds that arenot detected The latter possibility could providea new pathway towards prebiotic compoundsWe will now consider the significance and im-plication of these arguments one by one

Since the abundance of organic matter in thedust of 55PTemplel-Tuttle is not known it ispossible that there was little or no organic mat-ter in the Leonid particles to begin with Conse-quently the absence of nitrogen-bearing organicmatter in the Leonid meteoroids may have re-sulted in our inability to detect CN The onlyother measurement that could have detected ni-trogen-bearing organic matter in the dust (as op-posed to molecules in the gas phase) is in situmass spectroscopy The Leonid parent comet hasnot been visited In the coma of comet 1PHal-ley nitrogen-bearing organic matter was detectedin particles (the CHON particles) analyzed by themass spectrometers on the GIOTTO and VEGAspacecraft Because comet Halley is not the par-ent of the Leonid stream it is recommended thatfuture work focus on the meteoroids of the Ori-onid and eta-Aquarid streams that do originatefrom comet Halley Though bright meteors fromthese showers are much more difficult to observemuch the same result would be expected TheLeonid parent comet is remarkable only becauseit passes so close to Earthrsquos orbit Most impor-tantly since 55PTempel-Tuttle is in a Halley-type orbit and it should originate from the Oortcloud just as 1PHalley there is no reason to ex-pect that it is less rich in organic matter

For the same reason there is no reason to ex-pect that the nitrogen content or composition ofthe organic matter would differ from that of1PHalley Halley dust is as rich in nitrogen thanprimitive meteorites Though the nitrogen con-tent in the organic matter of comet Halley hasbeen questioned because of calibration problemswith the detectors there is no reason to expectthere to be more nitrogen in the less primitive me-teorites The heating and irradiation that alter theorganics in meteorites tend to evolve the organicmatter towards materials less rich in nitrogen andoxygen (Jenniskens et al 1993)

It is unknown how the organic content changeswith dust grain size and exposure to the plane-tary medium hence the fate of the organic mat-ter in the interplanetary medium is not knownHowever we know when the observed Leonids

were ejected from the parent comet in 1767 The090558 UT Leonid meteoroid never came closerto the Sun than 0976 AU and remained in the in-terplanetary medium only for 225 years (sevenorbits) most of that time far from the Sun

Might the organic matter have been lost in theEarthrsquos atmosphere early in the trajectory athigher altitudes when the meteoroid first warmedup to above the evaporation and decompositiontemperature of T 700 K This scenario was pro-posed by Elford et al (1997) who observed thatradar meteors are initially detected at heights upto 140 km They concluded that a more volatilecomponent than stony minerals must be ablatingat those heights The specific case of fast Leonidswas discussed by Steel (1998) who calculated thealtitude of evaporation for a range of small or-ganic compounds with relatively low sublimationtemperatures finding ablation heights in excessof 140 km for some compounds However thelight curve of bright Leonid fireballs above 136km was found to be very flat and starting at about200 km (Spurny et al 2000ab) which is incon-sistent with the light curves calculated by Elfordet al (1997) and Steel (1998) At 200 km the grainsare not warm enough to evaporate the organiccompounds that survived the vacuum of space

Indeed the Leonid studied here shows nosign of differential ablation of minerals in order oftheir volatility The elements Na Fe Mg and Caare lost in the same proportion (Fig 2) exceptfor the last wisp of matter after the catastrophicfragmentation during the terminal flare (a topicof future work) Sodium is at least partiallyfound in more volatile minerals than those con-taining magnesium and iron Though somesmaller Leonid meteoroids are known to losesodium early in their trajectory this is onlyfound in fragile cometary shower meteoroidssmall enough to expose the minerals efficientlyto the impinging air flow (BorovicIuml ka et al 1999)Sodium may be contained in the larger Leonidsat lower altitudes because of the limited timeavailable for inward diffusion of heat and out-ward diffusion of molecules Although the or-ganic and mineral components are separate en-tities they are intimately mixed down to verysmall size scales in comet Halley dust and thediffusion of those volatile compounds out of thegrains (if at all possible) will take time Such con-siderations are not included in the model of Steel(1998) Hence the organic matter is not lost earlyin flight

METEORS AND EXOGENOUS ORGANIC MOLECULES 75

The low CN abundance in the air plasma of the090558 UT Leonid meteoroid cannot be under-stood by merely allowing some fraction of nitro-gen to escape as N NH N2 NH3 NH2 NCONO HNO or HCN etc The configuration of Nin the complex organic matter of meteoroids isnot known However from meteorite and mi-crometeorite studies (where much organic matterhas been altered by aqueous alteration processes)we know that a significant fraction of the nitro-gen is contained in the organic framework (seereferences in Keller et al 1995 Greenberg 2000)Aqueous alteration tends to move nitrogen fromthe organic framework to functional groups suchas -NH In cometary matter no aqueous alter-ation has occurred So if N is in the frameworkin meteorite organics it should also exist in theorganic framework of cometary matter Coal-bound nitrogen also resides principally in hete-rocyclic ring moieties mostly five-memberedpyrrolic rings and six-membered pyridinic rings(Smith et al 1994 Zhang 2001) In pyrolysis stud-ies of tar such nitrogen is observed to escape pre-dominantly in the form of HCN and NH3 (Chenand Niksa 1992 Nelson et al 1992 Ledesma etal 1998) Ammonia is a product of amino com-pounds At the low pressure and high tempera-ture conditions in meteor plasma CN radicalsrather than HCN are stable and ammonia willbreak apart Small hydrocarbon radicals higherthan CN are so excited by their formation thatthey rapidly fission into lower products leavingonly the thermally stable single and double car-bon radicals unless prevented by rapid radiativecooling Carbon atoms are present too but atlower abundance because of thermodynamic con-siderations The carbon atoms react with N2 toform CN At 4400 K CN and N2 are only par-tially dissociated while O2 and CO (and CO2) arefully dissociated leading to high abundances ofCN if the plasma is in local thermodynamic equi-librium (Jenniskens et al 2000a) Only if oxygenplays a role in extracting the nitrogen after abla-tion then NOx compounds are expected to dom-inate Hence CN should be a significant productof the ablation of organic matter unless the or-ganic matter is released as larger compounds

We conclude that the most likely explanationfor the lack of CN molecules in the meteor plasmais that meteoric organic matter is lost in the formof a large range of relatively complex moleculesor perhaps as an amorphous carbon solid withembedded metal as proposed by Brownlee et al

(2002) Until now no evidence has been found ofsuch meteoric soot particles in the Earth envi-ronment and it remains unknown how well theheating conditions in the electron microscopeused in the experiment by Brownlee et al (2002)mimic the actual physical conditions during me-teoroid ablation

If the organic matter survives as large mole-cules this implies that bonds are prevented frombreaking by rapid radiative cooling or becauseheat is carried away by collisional fragments thattransfer much of the kinetic energy to the ambi-ent air The latter mechanism is the more likelyone and comes about because of conservation ofmomentum in a collision The result is that thelower mass fragment will carry most of the ki-netic energy and it will carry proportionallymore kinetic energy as the mass of the mainresidue increases This is an indirect form of thecollisional cooling that facilitates laser desorp-tion a technique widely used in the laboratory tostudy high-molecular-mass polyaromatic hydro-carbons in the gas phase and for detecting com-plex organic compounds in meteorites

Once released from the surface a single colli-sion will cause the molecule to slow down andlag behind relative to the meteoroid The mole-cule then is slowed to thermal speeds in another10 collisions and remains in the warm wake ofthe meteor for 50 collisions Only about 3 ofair collisions with atoms are inelastic (Oumlpik1958) For collisions with molecules that per-centage is higher but only a fraction of collisionswill result in chemical reactions Typical reactionsthat occur in the burning of propene for exam-ple are (eg Marinov et al 1996)

C3H6 1 O R C2H5 1 HCO (3)

C3H6 1 O R CH2CO 1 CH3 1 H (4)

C3H6 1 O R CH3CHCO 1 H 1 H (5)

Hydrogen can react in an early Earth-like atmos-phere with CO2 to form OH radicals which con-tinue to attack the organic matter via reactionssuch as

C3H6 1 OH R C3H6OH (6)

Radical attack can also lead to the growth of car-bon chains by polymerization reactions in therare case that organic compounds meet Hencethe result of chemistry in the meteor plasma is

JENNISKENS ET AL76

likely a product of relatively high molecular massthat is enriched in oxygen (and nitrogen) Suchcompounds are much more interesting for prebi-otic chemistry than the very-high-molecular-mass functional group-poor compounds foundin (micro-) meteorites

CONCLUSIONS

We conclude that complex organic matter inmeteoroids does not fully decompose into di-atomic constituents or atoms during the most in-tense phase of meteor ablation and that meteoricorganics are most likely lost in the form of largemolecular fragments cooled by radiative coolingand the loss of hot molecular fragments Theproduct of chemical interactions in the air plasmain the meteoroidrsquos wake is most likely a large va-riety of molecules and radicals that are chemicallyaltered functional group-enriched molecularfragments and condensation products

This result marks an alternative route to pre-biotic molecules from organic matter contributedto the Earth by particles at the peak of the massndashfrequency distribution of infalling matter At thetime of the origin of life there would have beentypically as much extraterrestrial matter impact-ing Earth in the form of dust as in the form oflarge asteroids and comets (Ceplecha 1992) ifthey arrived on elliptic orbits to have time to de-posit debris Moreover the organic content ofthose meteoroids should have been similar to thatof the impacting parent bodies with the excep-tion of the volatile compounds lost during dustejection This makes the product of meteor abla-tion and subsequent chemistry potentially thedominant source of complex organic moleculeson the early Earth at the time of the origin of life

ACKNOWLEDGMENTS

We thank George Flynn and an anonymous ref-eree for discussions that helped improve this pa-per The cooled CCD slit-less spectrograph wasdeveloped with support of the NASA Ames Re-search Centerrsquos Directorrsquos Discretionary Fundduring preparation of the Leonid MAC missionsWe thank Mike Koop (California Meteor Society)and Gary Palmer for technical support at variousstages in this project The 2001 Leonid MAC mis-sion was sponsored by NASArsquos Astrobiology

ASTID and Planetary Astronomy programs The2001 mission was executed by the USAF 418thFlight Test Squadron of Edwards AFB in Califor-nia with logistic support from the SETI Institute

ABBREVIATIONS

CCD charge coupled device FWHM full widthat half-maximum IDP interplanetary dust parti-cle MAC Multi-Instrument Aircraft CampaignUT universal time

REFERENCES

Abe S Yano H Ebizuka N and Watanabe J-I (2000)First results of high-definition TV spectroscopic obser-vations of the 1999 Leonid meteor shower Earth MoonPlanets 82-83 369ndash377

Anders E (1989) Pre-biotic organic matter from cometsand asteroids Nature 342 255ndash257

Betlem H Jenniskens P Spurny P Docters vanLeeuwen G Miskotte K Ter Kuile CR Zerubin Pand Angelos C (2000) Precise trajectories and orbits ofmeteoroids from the 1999 Leonid meteor storm EarthMoon Planets 82-83 277ndash284

BorovicIuml ka J Stork R and Bocek J (1999) First resultsfrom video spectroscopy of 1998 Leonid meteors Me-teoritics Planet Sci 34 987ndash994

Boyd ID (2000) Computation of atmospheric entry flowabout a Leonid meteoroid Earth Moon Planets 82-8393ndash108

Brownlee DE Joswiak DJ Kress ME Taylor S andBradley J (2002) Survival of carbon in moderately tostrongly heated IDPs and micrometeorites [abstract1786] In 33rd Lunar and Planetary Science Conference LPIContribution No 1109 Lunar and Planetary InstituteHouston

Ceplecha Z (1971) Spectral data on terminal flare andwake of double-station meteor No 38421 (OndrejovApril 21 1963) Bull Astron Inst Czech 22 219ndash304

Ceplecha Z (1992) Earth influx of interplanetary bodiesAstron Astrophys 263 361ndash366

Chen JC and Niksa S (1992) Coal devolatilization dur-ing rapid transient heating 1 Primary devolatilizationEnergy Fuels 6 254ndash264

Chyba CF and Sagan C (1992) Endogenous productionexogenous delivery and impact shock synthesis of or-ganic molecules An inventory for the origins of lifeNature 355 125ndash132

Chyba CF and Sagan C (1997) Comets as a source ofprebiotic organic molecules for the early Earth InComets and the Origin and Evolution of Life edited by PJThomas CF Chyba and CP McKay Springer VerlagNew York pp 147ndash173

Delsemme AH (1991) Nature and history of the organiccompounds in comets an astrophysical view In Astro-

METEORS AND EXOGENOUS ORGANIC MOLECULES 77

The low CN abundance in the air plasma of the090558 UT Leonid meteoroid cannot be under-stood by merely allowing some fraction of nitro-gen to escape as N NH N2 NH3 NH2 NCONO HNO or HCN etc The configuration of Nin the complex organic matter of meteoroids isnot known However from meteorite and mi-crometeorite studies (where much organic matterhas been altered by aqueous alteration processes)we know that a significant fraction of the nitro-gen is contained in the organic framework (seereferences in Keller et al 1995 Greenberg 2000)Aqueous alteration tends to move nitrogen fromthe organic framework to functional groups suchas -NH In cometary matter no aqueous alter-ation has occurred So if N is in the frameworkin meteorite organics it should also exist in theorganic framework of cometary matter Coal-bound nitrogen also resides principally in hete-rocyclic ring moieties mostly five-memberedpyrrolic rings and six-membered pyridinic rings(Smith et al 1994 Zhang 2001) In pyrolysis stud-ies of tar such nitrogen is observed to escape pre-dominantly in the form of HCN and NH3 (Chenand Niksa 1992 Nelson et al 1992 Ledesma etal 1998) Ammonia is a product of amino com-pounds At the low pressure and high tempera-ture conditions in meteor plasma CN radicalsrather than HCN are stable and ammonia willbreak apart Small hydrocarbon radicals higherthan CN are so excited by their formation thatthey rapidly fission into lower products leavingonly the thermally stable single and double car-bon radicals unless prevented by rapid radiativecooling Carbon atoms are present too but atlower abundance because of thermodynamic con-siderations The carbon atoms react with N2 toform CN At 4400 K CN and N2 are only par-tially dissociated while O2 and CO (and CO2) arefully dissociated leading to high abundances ofCN if the plasma is in local thermodynamic equi-librium (Jenniskens et al 2000a) Only if oxygenplays a role in extracting the nitrogen after abla-tion then NOx compounds are expected to dom-inate Hence CN should be a significant productof the ablation of organic matter unless the or-ganic matter is released as larger compounds

We conclude that the most likely explanationfor the lack of CN molecules in the meteor plasmais that meteoric organic matter is lost in the formof a large range of relatively complex moleculesor perhaps as an amorphous carbon solid withembedded metal as proposed by Brownlee et al

(2002) Until now no evidence has been found ofsuch meteoric soot particles in the Earth envi-ronment and it remains unknown how well theheating conditions in the electron microscopeused in the experiment by Brownlee et al (2002)mimic the actual physical conditions during me-teoroid ablation

If the organic matter survives as large mole-cules this implies that bonds are prevented frombreaking by rapid radiative cooling or becauseheat is carried away by collisional fragments thattransfer much of the kinetic energy to the ambi-ent air The latter mechanism is the more likelyone and comes about because of conservation ofmomentum in a collision The result is that thelower mass fragment will carry most of the ki-netic energy and it will carry proportionallymore kinetic energy as the mass of the mainresidue increases This is an indirect form of thecollisional cooling that facilitates laser desorp-tion a technique widely used in the laboratory tostudy high-molecular-mass polyaromatic hydro-carbons in the gas phase and for detecting com-plex organic compounds in meteorites

Once released from the surface a single colli-sion will cause the molecule to slow down andlag behind relative to the meteoroid The mole-cule then is slowed to thermal speeds in another10 collisions and remains in the warm wake ofthe meteor for 50 collisions Only about 3 ofair collisions with atoms are inelastic (Oumlpik1958) For collisions with molecules that per-centage is higher but only a fraction of collisionswill result in chemical reactions Typical reactionsthat occur in the burning of propene for exam-ple are (eg Marinov et al 1996)

C3H6 1 O R C2H5 1 HCO (3)

C3H6 1 O R CH2CO 1 CH3 1 H (4)

C3H6 1 O R CH3CHCO 1 H 1 H (5)

Hydrogen can react in an early Earth-like atmos-phere with CO2 to form OH radicals which con-tinue to attack the organic matter via reactionssuch as

C3H6 1 OH R C3H6OH (6)

Radical attack can also lead to the growth of car-bon chains by polymerization reactions in therare case that organic compounds meet Hencethe result of chemistry in the meteor plasma is

JENNISKENS ET AL76

likely a product of relatively high molecular massthat is enriched in oxygen (and nitrogen) Suchcompounds are much more interesting for prebi-otic chemistry than the very-high-molecular-mass functional group-poor compounds foundin (micro-) meteorites

CONCLUSIONS

We conclude that complex organic matter inmeteoroids does not fully decompose into di-atomic constituents or atoms during the most in-tense phase of meteor ablation and that meteoricorganics are most likely lost in the form of largemolecular fragments cooled by radiative coolingand the loss of hot molecular fragments Theproduct of chemical interactions in the air plasmain the meteoroidrsquos wake is most likely a large va-riety of molecules and radicals that are chemicallyaltered functional group-enriched molecularfragments and condensation products

This result marks an alternative route to pre-biotic molecules from organic matter contributedto the Earth by particles at the peak of the massndashfrequency distribution of infalling matter At thetime of the origin of life there would have beentypically as much extraterrestrial matter impact-ing Earth in the form of dust as in the form oflarge asteroids and comets (Ceplecha 1992) ifthey arrived on elliptic orbits to have time to de-posit debris Moreover the organic content ofthose meteoroids should have been similar to thatof the impacting parent bodies with the excep-tion of the volatile compounds lost during dustejection This makes the product of meteor abla-tion and subsequent chemistry potentially thedominant source of complex organic moleculeson the early Earth at the time of the origin of life

ACKNOWLEDGMENTS

We thank George Flynn and an anonymous ref-eree for discussions that helped improve this pa-per The cooled CCD slit-less spectrograph wasdeveloped with support of the NASA Ames Re-search Centerrsquos Directorrsquos Discretionary Fundduring preparation of the Leonid MAC missionsWe thank Mike Koop (California Meteor Society)and Gary Palmer for technical support at variousstages in this project The 2001 Leonid MAC mis-sion was sponsored by NASArsquos Astrobiology

ASTID and Planetary Astronomy programs The2001 mission was executed by the USAF 418thFlight Test Squadron of Edwards AFB in Califor-nia with logistic support from the SETI Institute

ABBREVIATIONS

CCD charge coupled device FWHM full widthat half-maximum IDP interplanetary dust parti-cle MAC Multi-Instrument Aircraft CampaignUT universal time

REFERENCES

Abe S Yano H Ebizuka N and Watanabe J-I (2000)First results of high-definition TV spectroscopic obser-vations of the 1999 Leonid meteor shower Earth MoonPlanets 82-83 369ndash377

Anders E (1989) Pre-biotic organic matter from cometsand asteroids Nature 342 255ndash257

Betlem H Jenniskens P Spurny P Docters vanLeeuwen G Miskotte K Ter Kuile CR Zerubin Pand Angelos C (2000) Precise trajectories and orbits ofmeteoroids from the 1999 Leonid meteor storm EarthMoon Planets 82-83 277ndash284

BorovicIuml ka J Stork R and Bocek J (1999) First resultsfrom video spectroscopy of 1998 Leonid meteors Me-teoritics Planet Sci 34 987ndash994

Boyd ID (2000) Computation of atmospheric entry flowabout a Leonid meteoroid Earth Moon Planets 82-8393ndash108

Brownlee DE Joswiak DJ Kress ME Taylor S andBradley J (2002) Survival of carbon in moderately tostrongly heated IDPs and micrometeorites [abstract1786] In 33rd Lunar and Planetary Science Conference LPIContribution No 1109 Lunar and Planetary InstituteHouston

Ceplecha Z (1971) Spectral data on terminal flare andwake of double-station meteor No 38421 (OndrejovApril 21 1963) Bull Astron Inst Czech 22 219ndash304

Ceplecha Z (1992) Earth influx of interplanetary bodiesAstron Astrophys 263 361ndash366

Chen JC and Niksa S (1992) Coal devolatilization dur-ing rapid transient heating 1 Primary devolatilizationEnergy Fuels 6 254ndash264

Chyba CF and Sagan C (1992) Endogenous productionexogenous delivery and impact shock synthesis of or-ganic molecules An inventory for the origins of lifeNature 355 125ndash132

Chyba CF and Sagan C (1997) Comets as a source ofprebiotic organic molecules for the early Earth InComets and the Origin and Evolution of Life edited by PJThomas CF Chyba and CP McKay Springer VerlagNew York pp 147ndash173

Delsemme AH (1991) Nature and history of the organiccompounds in comets an astrophysical view In Astro-

METEORS AND EXOGENOUS ORGANIC MOLECULES 77

likely a product of relatively high molecular massthat is enriched in oxygen (and nitrogen) Suchcompounds are much more interesting for prebi-otic chemistry than the very-high-molecular-mass functional group-poor compounds foundin (micro-) meteorites

CONCLUSIONS

We conclude that complex organic matter inmeteoroids does not fully decompose into di-atomic constituents or atoms during the most in-tense phase of meteor ablation and that meteoricorganics are most likely lost in the form of largemolecular fragments cooled by radiative coolingand the loss of hot molecular fragments Theproduct of chemical interactions in the air plasmain the meteoroidrsquos wake is most likely a large va-riety of molecules and radicals that are chemicallyaltered functional group-enriched molecularfragments and condensation products

This result marks an alternative route to pre-biotic molecules from organic matter contributedto the Earth by particles at the peak of the massndashfrequency distribution of infalling matter At thetime of the origin of life there would have beentypically as much extraterrestrial matter impact-ing Earth in the form of dust as in the form oflarge asteroids and comets (Ceplecha 1992) ifthey arrived on elliptic orbits to have time to de-posit debris Moreover the organic content ofthose meteoroids should have been similar to thatof the impacting parent bodies with the excep-tion of the volatile compounds lost during dustejection This makes the product of meteor abla-tion and subsequent chemistry potentially thedominant source of complex organic moleculeson the early Earth at the time of the origin of life

ACKNOWLEDGMENTS

We thank George Flynn and an anonymous ref-eree for discussions that helped improve this pa-per The cooled CCD slit-less spectrograph wasdeveloped with support of the NASA Ames Re-search Centerrsquos Directorrsquos Discretionary Fundduring preparation of the Leonid MAC missionsWe thank Mike Koop (California Meteor Society)and Gary Palmer for technical support at variousstages in this project The 2001 Leonid MAC mis-sion was sponsored by NASArsquos Astrobiology

ASTID and Planetary Astronomy programs The2001 mission was executed by the USAF 418thFlight Test Squadron of Edwards AFB in Califor-nia with logistic support from the SETI Institute

ABBREVIATIONS

CCD charge coupled device FWHM full widthat half-maximum IDP interplanetary dust parti-cle MAC Multi-Instrument Aircraft CampaignUT universal time

REFERENCES

Abe S Yano H Ebizuka N and Watanabe J-I (2000)First results of high-definition TV spectroscopic obser-vations of the 1999 Leonid meteor shower Earth MoonPlanets 82-83 369ndash377

Anders E (1989) Pre-biotic organic matter from cometsand asteroids Nature 342 255ndash257

Betlem H Jenniskens P Spurny P Docters vanLeeuwen G Miskotte K Ter Kuile CR Zerubin Pand Angelos C (2000) Precise trajectories and orbits ofmeteoroids from the 1999 Leonid meteor storm EarthMoon Planets 82-83 277ndash284

BorovicIuml ka J Stork R and Bocek J (1999) First resultsfrom video spectroscopy of 1998 Leonid meteors Me-teoritics Planet Sci 34 987ndash994

Boyd ID (2000) Computation of atmospheric entry flowabout a Leonid meteoroid Earth Moon Planets 82-8393ndash108

Brownlee DE Joswiak DJ Kress ME Taylor S andBradley J (2002) Survival of carbon in moderately tostrongly heated IDPs and micrometeorites [abstract1786] In 33rd Lunar and Planetary Science Conference LPIContribution No 1109 Lunar and Planetary InstituteHouston

Ceplecha Z (1971) Spectral data on terminal flare andwake of double-station meteor No 38421 (OndrejovApril 21 1963) Bull Astron Inst Czech 22 219ndash304

Ceplecha Z (1992) Earth influx of interplanetary bodiesAstron Astrophys 263 361ndash366

Chen JC and Niksa S (1992) Coal devolatilization dur-ing rapid transient heating 1 Primary devolatilizationEnergy Fuels 6 254ndash264

Chyba CF and Sagan C (1992) Endogenous productionexogenous delivery and impact shock synthesis of or-ganic molecules An inventory for the origins of lifeNature 355 125ndash132

Chyba CF and Sagan C (1997) Comets as a source ofprebiotic organic molecules for the early Earth InComets and the Origin and Evolution of Life edited by PJThomas CF Chyba and CP McKay Springer VerlagNew York pp 147ndash173

Delsemme AH (1991) Nature and history of the organiccompounds in comets an astrophysical view In Astro-

METEORS AND EXOGENOUS ORGANIC MOLECULES 77

physics and Space Science Library Vol 167 Comets in thePost-Halley Era Vol 1 edited by RL Newburn MNeugebauer and JH Rahe Kluwer Dordrecht TheNetherlands pp 377ndash428

Delsemme AH (1992) Cometary origin of carbon nitro-gen and water on the Earth Origins Life 21 279ndash298

Elford WG Steel DI and Taylor AD (1997) Impli-cations for meteoroid chemistry from the height dis-tribution of radar meteors Adv Space Res 201501ndash1504

Flynn GJ Keller LP Jacobsen C Wirick S andMiller MA (2000) Organic carbon in interplanetarydust particles In ASP Conference Series Vol 213 A NewEra in Bioastronomy Astronomical Society of the PacificPress San Francisco pp 191ndash194

Fristrom RM (1995) Flame Structure and Processes Ox-ford University Press New York

Glavin DP and Bada JL (1999) The sublimation andsurvival of amino acids and nucleobases in the Murchi-son meteorite during a simulated atmospheric heatingevent [abstract 1085] In 30th Lunar and Planetary ScienceConference LPI Contribution No 964 Lunar and Plan-etary Institute Houston

Greenberg JM (2000) From comets to meteors EarthMoon Planets 82ndash83 313ndash324

Jenniskens P and Butow SJ (1999) The 1998 Leonidmulti-instrument aircraft campaignmdashan early reviewMeteoritics Planet Sci 34 933ndash943

Jenniskens P and Russell RW (2003) The 2001 LeonidMulti-Instrument Aircraft Campaignmdashan early reviewISAS Rep SP 15 3ndash15

Jenniskens P Baratta GA Kouchi A de Groot MSGreenberg JM and Strazzulla G (1993) Carbon dustformation on interstellar grains Astron Astrophys 273583ndash600

Jenniskens P de Lignie M Betlem H Borovicka JLaux CO Packan D and Kruumlger CH (1998) Prepar-ing for the 199899 Leonid Storms Earth Moon Planets80 311ndash341

Jenniskens P Wilson MA Packan D Laux COBoyd ID Popova OP and Fonda M (2000a) Mete-ors A delivery mechanism of organic matter to theearly Earth Earth Moon Planets 82-83 57ndash70

Jenniskens P Butow SJ and Fonda M (2000b) The1999 Leonid Multi-Instrument Aircraft Campaignmdashanearly review Earth Moon Planets 82-83 1ndash26

Jessberger EK and Kissel J (1991) Chemical propertiesof cometary dust and a note on carbon isotopes In As-trophysics and Space Science Library Vol 167 Comets inthe Post-Halley Era Vol 2 edited by RL Newburn MNeugebauer and JH Rahe Kluwer Dordrecht TheNetherlands pp 1075ndash1092

Keller LP Thomas KL Bradley JP and McKay DS(1995) Nitrogen in interplanetary dust particles Mete-oritics 30 526ndash527

Kissel J and Krueger FR (1987) The organic componentin dust from comet Halley as measured by the PUMAmass spectrometer onboard Vega 1 Nature 326 755ndash760

Koidl P Wild Ch Dischler B Wagner J and Ram-steiner M (1989) Plasma deposition properties andstructure of amorphous hydrogenated carbon films Ma-terials Science Forum 52-53 41ndash70

Laux CO (1993) Optical diagnostics and radiative emis-sion of high temperature air plasmas [PhD Disserta-tion] HTGL Report T288 Stanford University Stan-ford CA

Ledesma EB Li CZ Nelson PF and Mackie JC(1998) Release of HCN NH3 and HNCO from the ther-mal gas-phase cracking of coal pyrolysis tars EnergyFuels 12 536ndash541

Levasseur-Regourd AC Mann I Dumont R and Han-ner MS (2000) Optical and thermal properties of in-terplanetary dust In Interplanetary Dust edited by EGruen BAringS Gustafson S Dermott and H FechtigSpringer Verlag Berlin pp 57ndash94

Love SG and Brownlee DE (1993) A direct measure-ment of the terrestrial mass accretion rate of cosmicdust Science 262 550ndash553

Marinov NM Pitz WJ Westbrook CK Castaldi MJand Senkan SM (1996) Modeling of aromatic andpolycyclic aromatic hydrocarbon formation in pre-mixed methane and ethane flames Combust Sci Tech-nol 116-117 211ndash287

Maurette M (1998) Carbonaceous micrometeorites andthe origin of life Origins Life Evol Biosphere 28 385ndash412

Maurette M Duprat J Engrand C Gounelle M Ku-rat G Matrajt G and Toppani A (2000) Accretion ofneon organics CO2 nitrogen and water from large in-terplanetary dust particles on the early Earth PlanetSpace Sci 48 1117ndash1137

Nelson PF Buchley AN and Kelly MD (1992) Func-tional forms of nitrogen in coals and the release of coalnitrogen as NOx precursors (HCN and NH3) In 24th

Symposium on Combustion The Combustion InstitutePittsburgh pp 1259ndash1267

Oumlpik E (1958) Physics of Meteor Flight in the AtmosphereInterscience Publishers New York

Oroacute J (1961) Comets and the formation of biochemicalcompounds on the primitive Earth Nature 190 389ndash390

Park CS Newfield ME Fletcher DG Goumlkccedilen T andSharma SP (1997) Spectroscopic Emission MeasurementsWithin the Blunt-Body Shock Layer in an Arc-Jet FlowAIAA 97ndash0990 American Institute of Aeronautics andAstronautics Reston VA

Rairden RL Jenniskens P and Laux CO (2000) Searchfor organic matter in Leonid meteoroids Earth MoonPlanets 82ndash83 71ndash80

Russell RW Rossano GS Chatelain MA LynchDK Tessensohn TK Abendroth E Kim D and Jen-niskens P (2002) Mid-infrared spectroscopy of persis-tent Leonid trains Earth Moon Planets 82-83 429ndash438

Smith KL Smoot LD Fletcher TH and Pugmire RJ(1994) The Structure and Reaction Processes of CoalPlenum Press New York

Spurny P Betlem H vanrsquot Leven J and Jenniskens P(2000a) Atmospheric behavior and extreme beginningheights of the thirteen brightest photographic Leonid

JENNISKENS ET AL78

meteors from the ground-based expedition to ChinaMeteoritics Planet Sci 35 243ndash249

Spurny P Betlem H Jobse K Koten P and vanrsquotLeven J (2000b) New type of radiation of bright Leonidmeteors above 130 km Meteoritics Planet Sci 351109ndash1115

Steel D (1998) The Leonid Meteors Compositions andconsequences Astron Geophys 39 24ndash26

Toppani A Libourel G Engrand C and Maurette M(2001) Experimental simulation of atmospheric entry ofmicrometeorites Meteoritics Planet Sci 36 1377ndash1396

Zhang H (2001) Nitrogen evolution and soot formationduring secondary coal pyrolysis [PhD Thesis] De-

partment of Chemical Engineering Brigham YoungUniversity Salt Lake City UT

Address reprint requests toP Jenniskens

Center for the Study of Life in the UniverseSETI Institute

2035 Landings DriveMountain View CA 94043

E-mail pjenniskensmailarcnasagov

METEORS AND EXOGENOUS ORGANIC MOLECULES 79