Embed Size (px)
Citation preview
Available online at www.sciencedirect.com
www.elsevier.com/locate/gca
ScienceDirect
Geochimica et Cosmochimica Acta 221 (2018) 109–126
Strontium and barium isotopes in presolar silicon carbidegrains measured with CHILI—two types of X grains
Thomas Stephan a,b,⇑, Reto Trappitsch a,b,1, Andrew M. Davis a,b,c,Michael J. Pellin a,b,c,d, Detlef Rost a,b,2, Michael R. Savina b,d,1, Manavi Jadhav a,b,
Christopher H. Kelly a,b, Frank Gyngard e, Peter Hoppe f, Nicolas Dauphas a,b,c
aDepartment of the Geophysical Sciences, The University of Chicago, Chicago, IL 60637, USAbChicago Center for Cosmochemistry, Chicago, IL, USA
cThe Enrico Fermi Institute, The University of Chicago, Chicago, IL 60637, USAdMaterials Science Division, Argonne National Laboratory, Argonne, IL 60439, USA
eLaboratory for Space Sciences and Department of Physics, Washington University, St. Louis, MO 63130, USAfMax Planck Institute for Chemistry, 55128 Mainz, Germany
Received 1 October 2016; accepted in revised form 1 May 2017; available online 10 May 2017
Abstract
We used CHILI, the Chicago Instrument for Laser Ionization, a new resonance ionization mass spectrometer developedfor isotopic analysis of small samples, to analyze strontium, zirconium, and barium isotopes in 22 presolar silicon carbidegrains. Twenty of the grains showed detectable strontium and barium, but none of the grains had enough zirconium to bedetected with CHILI. Nine grains were excluded from further consideration since they showed very little signals (<1000counts) for strontium as well as for barium. Among the 11 remaining grains, we found three X grains. The discovery of threesupernova grains among only 22 grains was fortuitous, because only �1% of presolar silicon carbide grains are type X, butwas confirmed by silicon isotopic measurements of grain residues with NanoSIMS. While one of the X grains showed stron-tium and barium isotope patterns expected for supernova grains, the two other supernova grains have 87Sr/86Sr < 0.5, valuesnever observed in any natural sample before. From their silicon isotope ratios, the latter two grains can be classified as X2grains, while the former grain belongs to the more common X1 group. The differences of these grains in strontium and bariumisotopic composition constrain their individual formation conditions in Type II supernovae.� 2017 Elsevier Ltd. All rights reserved.
Keywords: Presolar grains; Silicon carbide; Supernovae; Nucleosynthesis; Resonance ionization mass spectrometry (RIMS); Strontium is-otopes; Barium isotopes
http://dx.doi.org/10.1016/j.gca.2017.05.001
0016-7037/� 2017 Elsevier Ltd. All rights reserved.
⇑ Corresponding author at: Department of the GeophysicalSciences, The University of Chicago, Chicago, IL 60637, USA.
E-mail address: [email protected] (T. Stephan).1 Present address: Nuclear and Chemical Sciences Division,
Lawrence Livermore National Laboratory, Livermore, CA94550, USA.2 Present address: Department of Physics, University of Auck-
land, Auckland 1010, New Zealand.
1. INTRODUCTION
Primitive meteorites and interplanetary dust particlescontain small quantities of isotopically anomalous refrac-tory dust grains that are older than our Solar System andcommonly called ‘‘presolar grains” (Davis, 2011; Zinner,2014). They condensed in the winds of evolved stars andin the ejecta of stellar explosions, i.e., they represent a sam-
110 T. Stephan et al. /Geochimica et Cosmochimica Acta 221 (2018) 109–126
ple of stardust that can be analyzed in the laboratory.Presolar minerals identified to date include, e.g., diamond,silicon carbide (SiC), graphite, silicon nitride (Si3N4), oxi-des, and various types of silicates. Silicon carbide is the bestcharacterized presolar mineral. Based on the isotopic com-positions of carbon, nitrogen, and silicon, it is divided intodistinct populations, namely, mainstream (the majority ofSiC grains) and the minor subtypes AB, C, X, Y, Z, andnova grains. Mainstream, Y, and Z grains are from low-mass asymptotic giant branch (AGB) stars with about-solar (mainstream grains) and subsolar (Y and Z grains)metallicities. C and X grains are believed to come fromType II supernovae. The origin of AB grains is still a matterof debate; among the proposed stellar sources are J-typecarbon stars (Lambert et al., 1986), born-again AGB stars(Asplund et al., 1999), and Type II supernovae (Liu et al.,2016).
CHILI, the Chicago Instrument for Laser Ionization is anew resonance ionization mass spectrometry (RIMS)instrument developed for elemental and isotopic analysisof small samples, such as dust returned to Earth by space-craft and presolar grains from meteorites, at high spatialresolution and high sensitivity (Stephan et al., 2016).CHILI is especially suited for the analysis of trace elementisotopic compositions in presolar grains. We used CHILI tomeasure the isotopic compositions of strontium, zirconium,and barium in 22 presolar SiC grains. These elements areparticularly important for understanding s-process nucle-osynthesis in AGB stars (Lugaro et al., 2003) because oftheir sensitivity to branching between neutron captureand b-decay. These elements are also particularly importantsince they have isotopes with magic neutron numbers,which makes them sensitive to the total neutron exposure.Strontium, zirconium, and barium have been measuredbefore in presolar SiC with RIMS (Nicolussi et al., 1997,1998; Savina et al., 2003a; Barzyk et al., 2007; Liu et al.,2014, 2015) using the CHARISMA instrument (Ma et al.,1995; Savina et al., 2003b) at Argonne NationalLaboratory.
Among the 22 SiC grains analyzed in the present study,three grains turned out to be X grains, as was confirmed bysubsequent silicon isotopic analyses performed with theCAMECA NanoSIMS 50 at Washington University in St.Louis. Only a few X grains had been analyzed by RIMSfor strontium, zirconium, and barium isotopes before,and those measurements suffered from relatively largeuncertainties (Pellin et al., 1999, 2000, 2006; Davis et al.,2002). X grains are typically attributed to an origin in TypeII supernovae (Amari et al., 1992; Davis, 2011; Zinner,2014). According to a study by Hoppe et al. (1994), inwhich 720 SiC grains ranging in size from 1 to 10 mm,including 181 KJG grains, were analyzed, only �1% ofpresolar SiC grains from the KJG size fraction from theMurchison meteorite, which was analyzed in the presentstudy, should be X grains. Smaller size fractions containup to 2% X grains (Hoppe et al., 2010). Assuming an Xgrain abundance of 1–2% among presolar SiC grains, theprobability for finding three X grains out of 22 grains ana-lyzed was calculated, using the binomial distribution, to be(1.3 � 8.4) � 10�3.
The new data obtained with CHILI and discussed hereprovide new insights into the nucleosynthetic processes pre-ceding grain formation. We compare our results to modelsof supernova nucleosynthesis with a full nuclear network(Rauscher et al., 2002). Such models have been successfullyused before to explain X grain properties, at least qualita-tively, by involving extensive multizone mixing (Yoshidaand Hashimoto, 2004; Yoshida, 2007), but have failed toquantitatively explain all observed X grain isotope systemsso far (Lin et al., 2010; Hoppe et al., 2010, 2012; Davis,2011; Zinner, 2014). Pignatari et al. (2013, 2015) recentlypresented new supernova models, in which a carbon- andsilicon-rich zone at the bottom of the helium-burning shellcould be the site for SiC grain formation without involvingselective, large-scale mixing. In these one-dimensional mod-els, the maximum temperature at the bottom of the helium-burning shell has a large impact on the isotopic signaturesof such grains. Multidimensional models (Nomoto et al.,2013; Muller, 2016) taking into account the asymmetry ofcore-collapse supernovae (Grefenstette et al., 2014) shouldprovide a more realistic picture in order to explain the widerange of isotopic compositions observed in supernovagrains (Hoppe et al., 2018). The data presented here addnew constraints to existing and to future models of super-nova nucleosynthesis.
2. EXPERIMENTAL
2.1. Samples
Presolar SiC grains that had been extracted from theMurchison CM2 meteorite more than 20 years ago wereanalyzed in this study. The grains are from the KJG sam-ple, which refers to the 1.5–3 mm size fraction of the KJSiC separate (Amari et al., 1994). In contrast to recent workon Murchison SiC grains (Levine et al., 2009; Liu et al.,2014, 2015; Trappitsch et al., 2018), the samples in thisstudy were not additionally treated with concentrated acidsto remove contamination from parent-body or terrestrialmaterial. The grains were mounted on a high-purity goldfoil by depositing them from a suspension and pressingthem into the gold with a sapphire window. Prior to RIMSanalysis, energy-dispersive X-ray spectroscopy (EDS)images of the mount were acquired by scanning electronmicroscopy (SEM) to locate the SiC grains. From theseimages, 22 silicon-rich grains were randomly selected foranalysis with CHILI.
2.2. RIMS measurements
A detailed description of the CHILI instrument and theanalytical procedures has been given by Stephan et al.(2016). Here, we only provide a short summary of the ana-lytical procedures, especially those specific to these mea-surements. The SiC grains were located using CHILI’sintegrated scanning electron microscope. The samples wereablated by a 351 nm wavelength laser beam from afrequency-tripled Nd:YLF desorption laser, focused to�1 mm with a Schwarzschild-type optical microscope, andrastered over an area of about 10 � 10 mm2. The beam
Tab
le1
Silicon,strontium,an
dbarium
isotopedataa
,bof11
presolarSiC
grainsan
alyzed
withCHIL
Ian
dNan
oSIM
S.
Grain
Typ
ed2
9Si[‰
]d3
0Si[‰
]d8
4Sr[‰
]d8
7Sr[‰
]d8
8Sr[‰
]d1
30Ba[‰
]d1
32Ba[‰
]d1
34Ba[‰
]d1
35Ba[‰
]d1
37Ba[‰
]d1
38Ba[‰
]
#2
Mc
+82
±84
+90±11
0�4
20±24
0�5
0±200
+70
±13
0–
–+25
0±600
+10±31
0�7
0±26
0�7
0±210
#4
M+10
0±66
+110±
86�4
60±27
0+60±10
0�3
0±66
––
�40±
200
�581
±78
�337
±88
�260
±73
#5
M+18
±61
+18±78
––
––
–+60
±50
0�6
00±18
0�4
80±15
0�3
40±13
0#11
M+11
±71
�56±
85+60±41
0+38±82
+15
±54
––
+10
±23
0�1
50±11
0�7
4±98
�99±
76#13
M+80
±68
+80±87
�780
±13
0�1
8±65
�108
±40
�110
±73
0�5
40±64
0+12
0±140
�532
±56
�330
±59
�281
±47
#18
M+11
1±74
+79±92
�750
±10
0+1±
50�1
27±30
+210±
920
+260±
930
�140
±14
0�5
49±64
�294
±72
�287
±55
#22
M+9±
63+62±85
–0±
340
+50
±16
0–
–�6
0±19
0+20±14
0+40
±13
0�9
±98
#23
M+12
0±69
+134±
90–
+190±
300
+10
0±190
––
––
––
#14
X1
�220
±45
�389
±46
�518
±73
�234
±29
+70
2±40
+260±
770
+560±
790
�300
±11
0�5
89±53
+31
0±100
+1670
±18
0#6
X2
�415
±35
�398
±47
�300
±12
00�3
80±21
0�4
74±93
––
+31
0±420
�650
±13
0�2
10±16
0�1
00±14
0#15
X2
�460
±34
�434
±46
�670
±20
0�4
24±55
�440
±31
––
+64
0±440
�600
±12
0�3
30±15
0�3
50±11
0
aIsotoperatiosarereported
asd-va
lues,givingthedeviationfrom
terrestrialisotoperatios(M
eija
etal.,20
16),norm
alized
to28Si,
86Sr,an
d136Ba,
respectively,in
parts
per
thousand(‰
).bUncertainties
are2r
resultingfrom
countingstatistics
and,forsilicon,from
reproducibilityofthestan
dards.
cM:mainstream
.
T. Stephan et al. /Geochimica et Cosmochimica Acta 221 (2018) 109–126 111
intensity was controlled in order to keep desorption rates ata level such that significant dead time effects of the detectorwere avoided. CHILI has six Ti:sapphire lasers, which arepumped by three 40 W Nd:YLF lasers with 527 nm wave-length, allowing typically to resonantly ionize two or threeelements simultaneously. In the present study, light fromthe six Ti:sapphire lasers, tuned individually to specificwavelengths, was first combined with a three-prism beamcombiner, then sent through the cloud of neutrals generatedby the desorption laser, and bounced back through thecloud onto itself. The Ti:sapphire lasers were tunedfor resonance ionization of strontium (k1 = 460.862 nm,k2 = 405.214 nm; Liu et al., 2015), zirconium (k1 =296.172 nm, k2 = 442.533 nm; Barzyk et al., 2007), andbarium (k1 = 307.247 nm, k2 = 883.472 nm; Savina et al.,2003b). Photoions were then separated according to theirmass-to-charge ratio in a reflectron-type time-of-flight massspectrometer and detected with a microchannel plate detec-tor. Data were corrected for dead time effects using Poissonstatistics as described by Stephan et al. (1994).
2.3. RIMS standards
Isotopic standards used in this study are NIST SRM855a with 180 ppm Sr, NIST SRM 1264a with 690 ppmZr, zirconium metal, and terrestrial BaTiO3, all of whichwere assumed to be of average terrestrial isotopic composi-tion (Meija et al., 2016). This is a reasonable assumption atprecisions of a few ‰ or more.
Instrumental isotopic fractionation was determined tobe smaller than the statistical error of typically a few tensof ‰ for all isotopes but 91Zr (Stephan et al., 2016). TheRIMS sensitivity for 91Zr was enhanced by �200‰ dueto an odd-even effect (Fairbank et al., 1989; Wunderlichet al., 1992), which had been observed for zirconium before(Nicolussi et al., 1997).
2.4. NanoSIMS
Subsequent to RIMS analysis with CHILI, residues ofall 22 grains were analyzed with the CAMECA NanoSIMS50 (Stadermann et al., 1999; Hoppe et al., 2013) at Wash-ington University in St. Louis in order to identify differentSiC grain types (Davis, 2011; Zinner, 2014). A Cs+ primaryion beam was used, and the negative secondary ions 12C�,13C�, 28Si�, 29Si�, and 30Si� were measured simultaneously.Fine-grained synthetic SiC was used as a standard. Sincemost of the grain material had been already consumed dur-ing analysis with CHILI, limited material was left forNanoSIMS analysis. Therefore, analytical errors of the lat-ter measurements, which include both counting statisticsand reproducibility of the standards, were larger thanusual.
3. RESULTS
Zirconium was not detected in any of the grains ana-lyzed in this study. Two of the grains also showed nodetectable strontium or barium. Nine grains, althoughshowing detectable strontium and barium, only yielded
-800 -600 -400 -200 0 +200-800
-600
-400
-200
0
+200mainstreamX1X2mainstream (PGD)X (PGD)
solar
sola
r
X2
X1
29Si
/28Si
[‰]
30Si/ 28Si [‰]
X0
δ
δ
Fig. 1. Silicon isotope data of 11 SiC grains analyzed in this study. Error bars are 2r resulting from counting statistics and reproducibility ofthe standards. Eight of the grains plot in the typical area of mainstream grains, while three grains show isotope ratios typical for X grains.Literature data for mainstream and X grains are taken from the Washington University presolar grain data base (PGD; Hynes andGyngard, 2009, and references therein), excluding data points with uncertainties (2r) > 40‰. Following the classification by Lin et al. (2010),regions for X0, X1, and X2 subtypes are shown, with the majority of X grains belonging to X1, lying along a line through solar values with aslope of �2/3.
112 T. Stephan et al. /Geochimica et Cosmochimica Acta 221 (2018) 109–126
<1000 ion counts for each of those two elements. Theirresults were regarded as being insignificant and thereforeexcluded from further consideration. Table 1 shows iso-topic data for the 11 presolar SiC grains for which at least1000 strontium or barium ions were detected with CHILI.As discussed by Stephan et al. (2016), we would haveexpected to find zirconium in 30–50% of the grains, sincetrace element analyses by SIMS and synchrotron X-ray flu-orescence suggest that �50% of SiC grains are enriched inZr/Si by more than a factor of two compared to CI mete-orites (Amari et al., 1995; Kashiv et al., 2010). The nonde-tection of zirconium in our grains was seen as an indicationthat the resonance ionization scheme used for zirconiumwas less effective than expected. Recent technical improve-ments of the mass spectrometer, a slightly modified ioniza-tion scheme for zirconium, and the replacement of somemirrors, which had insufficient reflectance at lower wave-lengths, important for the first step (296.172 nm) in the zir-conium scheme, finally led to the first detection ofzirconium in presolar silicon carbide grains with CHILI(Stephan et al., 2017).
3.1. Carbon
Carbon isotopic analysis with the NanoSIMS wasimpeded by the fact that the grains had been almost entirelyconsumed during RIMS analysis. While all 22 grains had
sufficient silicon in the residue associated with the laser des-orption pit to perform an isotopic analysis, carbon wasfound all over the sample mount. NanoSIMS analysesshowed 12C/13C ratios close to the terrestrial value—a clearindication for contamination—for all but one grain residue(grain residue #22 showed a 12C/13C ratio of 45.4 ± 1.4; allerrors in this work are given as 2r). Furthermore, subse-quent SEM-EDS analyses yielded C/Si elemental ratios of2–30, while an untouched SiC grain on the same samplemount showed the expected ratio close to unity. We there-fore base our grain classification solely on silicon data,which appear to be uncompromised.
3.2. Silicon
From silicon isotopic ratios, eight of the 11 grains wereclassified as mainstream grains; three grains showed strongenrichments in 28Si and were classified as X grains (Table 1).The remaining 11 grains that had no or little detectablestrontium and barium are also, from their silicon isotopes,consistent with a classification as mainstream grains, butwill not be discussed any further here. Based on silicon iso-topes alone, some ambiguity remains with regard to thegrains classified as mainstream. In addition, the relativelylarge uncertainties in the silicon isotopic measurements ofgrain residues after RIMS analysis precluded more accurateclassification. While we expect that most grains are indeed
T. Stephan et al. /Geochimica et Cosmochimica Acta 221 (2018) 109–126 113
mainstream grains, some of them could also be Y or Zgrains, which are attributed to AGB stars as well, or ABgrains, for which the proposed stellar sources include J-type carbon stars, born-again AGB stars, and core collapsesupernovae. Following Lin et al. (2010), we attribute one ofthe X grains to the subtype X1, while the other two Xgrains can be classified as X2. Silicon isotope ratios areshown in Fig. 1. Shown for comparison are isotopic datafrom the Presolar Grain Database (PGD; Hynes andGyngard, 2009; http://presolar.wustl.edu/Laboratory_for_Space_Sciences/Presolar_Grain_Database.html and refer-ences therein), which, as of April 2017, contains silicon iso-topic data for 13,331 mainstream and 588 X grains. InFig. 1, we only show data for 10,022 mainstream and 415X grains, excluding all grains with uncertainties (2r)> 40‰ in d-value for at least one of the silicon isotopes.It should be noted here that X grains are overrepresentedin the PGD, since it includes data of studies specificallyaddressing X grains and therefore not reporting data for
-500
0
+500
+1000
δ(87
Sr/86
Sr) [
‰]
-1000 -500
-500
0
+500
δ(88
Sr/86
Sr) [
‰]
δ(84Sr/Fig. 2. Strontium isotope data for eight presolar SiC grains (this work) cobars are 2r resulting from counting statistics.
mainstream grains that had been found together with theX grains. A narrow band of X grains in the silicon three-isotope plot (Fig. 1) defines the X1 gains. The 3–5% ofgrains above this band are classified as X0 grains. Depend-ing on the exact definition, 24–28% of the grains plot belowthe X1 band in Fig. 1 and are classified as X2 grains.
The two X2 grains in this study are very close to eachother in all measured isotope systems (Table 1). However,from their different location on the sample mount (morethan 1 mm apart) and a �1.5 � difference in strontium-to-barium signal ratio, we conclude that the two grains arenot fragments of a larger grain that broke apart duringseparation.
3.3. Strontium
Strontium isotope data for eight grains are comparedwith literature data on mainstream grains (Liu et al.,2015) in Fig. 2. For two mainstream grains (#22 and
mainstreamX1X2Liu et al., 2015
0 +500 +100086Sr) [‰]mpared to mainstream grain literature data (Liu et al., 2015). Error
-1000
-500
0
+500
+1000 mainstreamX1X2Liu et al.,
2014, 2015
δ(13
4 Ba/13
6 Ba) [
‰]
-1000
-500
0
+500
δ(13
7 Ba/13
6 Ba) [
‰]
-1000 -500 0 +500 +1000-1000
-500
0
+500
+1000
+1500
δ(13
8 Ba/13
6 Ba) [
‰]
δ(135Ba/136Ba) [‰]Fig. 3. Barium isotope data for 10 presolar SiC grains (this work)compared to mainstream grain literature data (Liu et al., 2014,2015). Error bars are 2r resulting from counting statistics.
114 T. Stephan et al. /Geochimica et Cosmochimica Acta 221 (2018) 109–126
#23), strontium data show large statistical errors, and 84Srwas not unambiguously detected (Table 1). Therefore, thesegrains are not shown in Fig. 2. All mainstream grains ana-lyzed in the present study show a range of strontium isotoperatios similar to the previous studies and are consistent withformation in low-mass AGB stars (Liu et al., 2015). How-ever, there seems to be a tendency to plot closer to normalthan the Liu et al. (2015) data; residual contamination fromparent-body or terrestrial material could be responsible forsuch a trend. Grains that plot closer to normal also showlarger error bars due to the fact that less strontium wasfound in these grains, which also makes them more suscep-tible to contamination. Two mainstream grains, #13 and#18, show relatively small error bars and d84Sr/86Sr below�700‰. In accordance with Liu et al. (2015), we assumethese grains to be free of contamination. Besides the strongdepletion in 84Sr, mainstream grains show rather normal87Sr and a small depletion in 88Sr normalized to 86Sr andterrestrial abundances.
Both types of X grains show distinct properties in stron-tium isotopes (Fig. 2). The X1 grain has a substantial excessin 88Sr and is depleted in 84Sr and in 87Sr relative to 86Sr andterrestrial abundances. The two X2 grains show even stron-ger depletions in 87Sr than the X1 grain and, in contrast tothe X1 grain, are also depleted in 88Sr. The 87Sr/86Sr ratiosof 0.44 ± 0.15 for grain #6 and 0.41 ± 0.04 for grain #15,respectively, are, to our knowledge, the lowest 87Sr/86Srratios ever reported for any natural material. For compar-ison, the 87Sr/86Sr ratio of the Solar System when it formed4.56 Ga ago was �0.699 (Papanastassiou and Wasserburg,1969). When compared to data given in Table 1 or Fig. 2,the early Solar System value is equivalent to d87Sr = �16‰,since we normalized all data in this work to modern terres-trial abundances according to Meija et al. (2016). Theincrease of 87Sr by 16‰ in the Earth over the last 4.56 Gais due to the decay of 87Rb with a half-life of 49 Ga.
3.4. Barium
Barium isotope data for 10 grains are compared with lit-erature data on mainstream grains (Liu et al., 2014, 2015) inFig. 3. As for strontium, all mainstream grains show a sim-ilar range in barium isotopes as observed in previous stud-ies and are consistent with formation in low-mass AGBstars. Again, some grains show isotope ratios closer to nor-mal than expected, which can be attributed to residual con-tamination with parent-body or terrestrial material. Clearevidence for barium contamination of SiC grains has beenobserved before (Marhas et al., 2007). Liu et al. (2014)regarded d135Ba data above �400‰ to be the result of con-tamination. In the present study, four mainstream grains,#4, #5, #13, and #18, show d135Ba below �500‰, whichwe interpret as being free of substantial barium contamina-tion. Besides the strong depletion in 135Ba, mainstreamgrains are also depleted in 137Ba and in 138Ba, but showrather normal 134Ba when normalized to 136Ba and terres-trial abundances. For most mainstream grains, 130Ba and132Ba were not detected, and for the two grains that showed
T. Stephan et al. /Geochimica et Cosmochimica Acta 221 (2018) 109–126 115
detectable 130Ba and 132Ba, statistical errors are too large(Table 1) to draw any conclusions. Liu et al. (2014, 2015)found these isotopes to be heavily depleted (relative toSolar System abundances) in mainstream grains.
The X1 grain shows major excesses in 137Ba and, espe-cially, in 138Ba as well as depletion in 134Ba compared tothe mainstream grains. For 130Ba and 132Ba, there mightbe a statistically weak indication (Table 1) that they areenriched in the X1 grain. The X2 grains show barium iso-topic ratios very close to the mainstream grains with theexception of 134Ba, which seems to be slightly enriched.Barium-130 and 132Ba could not be detected in the X2grains.
4. DISCUSSION
Since the SiC grains investigated in this study were notadditionally treated with concentrated acids, unlike thoseanalyzed by Liu et al. (2014, 2015), it is no surprise thatsome of the samples in the present study have isotope ratiosshifted towards isotopically normal strontium and barium,probably due to contamination with parent-body or terres-trial material. For the following discussion, we rigorouslyexcluded mainstream grains (all except #13 and #18) thatshow indications of contaminations either in strontium orbarium. To lower statistical uncertainties, isotope ratios
28 29 30-1000
-500
0
+500
+1000
+1500
+2000
84 86 87 88
mainstream (#13 & #18) X1 (#14) X2 (#6 & #15)
δ(x Si
/28Si
) or δ
(x Sr/86
Sr) o
r δ(x Ba
/136 Ba
) [‰
]
Si isotopes Sr isotopesFig. 4. Silicon, strontium, and barium isotopes normalized to 28Si, 86Sr, atypes. To lower statistical uncertainties, the mass spectra from two mainwere the spectra of the two X2 grains. Error bars are 2r resulting frostandards.
for mainstream grains were calculated from the added sig-nal intensities of the mass spectra for the two least-contaminated mainstream grains. We will, however, in thefollowing focus mostly on X grains, the differences betweenX1 and X2, and how they differ from mainstream grains. Amore elaborate discussion on mainstream grains based on amuch larger data base was given by Liu et al. (2014, 2015).
Since all of the X grains also show very low 84Sr/86Srand 135Ba/136Ba ratios, we conclude that their compositionsare not dominated by major contamination either.Although this might not be a necessary requirement for Xgrains, since we don’t know their 84Sr/86Sr and 135Ba/136Baratios beforehand, the huge deviations from normal84Sr/86Sr and 135Ba/136Ba can be considered as a sufficientcriterion. In any case, as always with presolar grains, dilu-tion with isotopically normal material has to be considered.We added the mass spectra from grains #6 and #15 andrefer to them as X2 grain data. This may seem arbitrarybut is justified since all measured isotope data for thesegrains overlap within 2r of each other (Table 1). Resultsfor mainstream, X1, and X2 grains are summarized inFig. 4.
For a proper discussion of the results, we need to take acloser look at the chart of nuclides, of which two partialsections, krypton to zirconium and xenon to cerium,respectively, are shown in Fig. 5. Branching factors
130 132 134 135 136 137 138
Ba isotopesnd 136Ba, respectively, showed different behavior for different grainstream grains that showed no hint of contamination were added asm counting statistics and, for silicon, from reproducibility of the
116 T. Stephan et al. /Geochimica et Cosmochimica Acta 221 (2018) 109–126
fn = kn/(kn + kb) of some relevant nuclides as a function ofneutron density for various temperatures are shown inFig. 6. Especially, decay rates for 134Cs and 135Cs increasedrastically with increasing temperature. In high-temperature
130Ba0.106 %755 mb
131Ba11.52 d
132Ba0.101 %397 mb
133Ba10.551 a
134Ba2.417 %176.0 mb
129Cs32.06 h
130Cs29.21 m
131Cs9.689 d
132Cs6.480 d
133Cs100 %502 mb
56
55
128Xe2.234 %262.5 mb
129Xe27.463 %617 mb
130Xe4.378 %132.0 mb
131Xe21.802 %340 mb
132Xe26.355 %63.8 mb
54
74 75 76 77 78
p
84Sr0.5580 %321 mb
85Sr64.853 d
86Sr9.8678 %60.0 mb
87Sr6.8961 %93.8 mb
88Sr82.6781%6.31 mb
85Rb70.844 %240 mb
82Kr11.655 %
90 mb85gKr10.776 a
73 mb
83Kr11.546 %249 mb
84Kr56.903 %33.1 mb
86Kr17.208 %4.76 mb
83Rb86.2 d
84Rb32.82 d 49.23 Ga
16.9 mb
87Rb29.156 %
85mKr4.480 h
38
37
36
46 47 48 49 50
86Rb18.642 d292 mb
58
57131La
59 m
132Ce3.51 h
1570 mb
132La4.8 h
133Ce97 m
2600 mb
131La3.912 h
134Ce3.16 d967 mb
134La6.45 m
135Ce17.7 h
1320 mb
136Ce0.186 %328 mb
135La19.5 h
40
39
86Zr16.5 h
87Zr1.68 h
88Zr83.4 d
89Zr78.41 h
85Y2.68 h
86Y14.74 h
87Y79.8 h
88Y106.626 d
89Y100 %
20.5 mb
90Zr51.452 %19.3 mb
Fig. 5. Sections of the chart of the nuclides showing s-process branchespanel) isotopes. For each stable or long-lived nuclide, the isotopic abundato Lodders et al. (2009). The half-lives at room temperature are givenindicating strong temperature dependence (more than a factor of two vaaccording to Takahashi and Yokoi (1987). Neutron capture cross sectkT = 30 keV from the Karlsruhe Astrophysical Database of Nucleosynthewith numbers in italics based on theoretical calculations only.
stellar environments, faster decaying, thermally populatedexcited states increase the decay rate of nuclides like, e.g.,134Cs (Takahashi and Yokoi, 1987). For nuclides with smalldecay energies (Qb < 300 keV) like, e.g., 135Cs, bound-state
138La0.091 %
135Ba6.592 %455 mb
136Ba7.853 %61.1 mb
137Ba11.232 %76.2 mb
138Ba71.699 %4.13 mb
139Ba83.25 m
135Cs2.3 Ma163 mb
136Cs13.16 d
137Cs30.08 a
138Cs33.41 m
133Xe5.2475 d127 mb
134Xe9.661 %21.3 mb
135Xe9.14 h
136Xe7.868 %0.98 mb
137Xe3.818 m
79 80 81 82 83
134Cs2.0652 a724 mb
n
89Sr50.53 d15.2 mb
88Rb17.773 m
87Kr76.3 m
51
139La99.909 %32.4 mb
140Ce88.450 %11.73 mb
138Ce0.250 %179 mb
137Ce9.0 h
973 mb
139Ce137.641 d
214 mb
141Ce32.508 d
76 mb
136La9.87 m
137La60 ka
140La40.285 h
90Y64.00 h104 mb
91Zr11.223 %
63 mb
102 Ga419 mb
s-process only nuclide
r-process only nuclide
s- and r-process nuclide
p-process nuclide
β –decaying nuclide
β + and ε decaying nuclide
magic neutron number
main s-process path
alternative paths
minor reactions
leading to the various strontium (lower panel) and barium (uppernce 4.56 Ga ago, when the Solar System formed, is given accordingfor unstable nuclides (Audi et al., 2012), with numbers in italicsriation in half-life at temperatures between 5 � 107 and 5 � 108 K)ions are given as Maxwellian-averaged cross sections (MACS) atsis in Stars (KADoNiS v1.0; Dillmann et al., 2014) for each nuclide,
0.0
0.2
0.4
0.6
0.8
1.0
f n 0.5×108 K 1.0×108 K 2.0×108 K 3.0×108 K 4.0×108 K 5.0×108 K
85Kr 86Rb
0.0
0.2
0.4
0.6
0.8
1.0
87Rb
f n
133Xe
105 106 107 108 109 1010 10110.0
0.2
0.4
0.6
0.8
1.0134Cs
f n
Nn [cm-3]105 106 107 108 109 1010 1011 1012
135Cs
Nn [cm-3]Fig. 6. Branching factors fn = kn/(kn + kb) as a function of neutron density Nn and temperature for various nuclides. We used thetemperature-dependent b-decay rates reported by Takahashi and Yokoi (1987) and adapted neutron capture cross sections from KADoNiSv1.0 (Dillmann et al., 2014).
T. Stephan et al. /Geochimica et Cosmochimica Acta 221 (2018) 109–126 117
118 T. Stephan et al. /Geochimica et Cosmochimica Acta 221 (2018) 109–126
b-decay can drastically increase the decay rates in the ion-ized stellar plasma (Takahashi and Yokoi, 1987; Kappeleret al., 1989).
The following Sections 4.1 and 4.2 discuss nucleosynthe-sis of strontium and barium in our grains, neutron capturepathways, branchings, and their dependence on tempera-ture and neutron density, as well as element fractionationduring condensation. Section 4.3, eventually, uses theRauscher et al. (2002) Type II supernova models to discusspossible mixing scenarios that would generate the observedisotope ratios for the three X grains analyzed.
4.1. Strontium
The branch point at 85Kr (Fig. 5) plays a major role indetermining strontium isotope ratios during nucleosynthe-sis. Krypton-85 has a half-life of �11 a and decays (b�)to 85Rb. However, this is only true for the ground state85gKr. The isomer 85mKr can either decay directly to 85Rb(b�; 78.6%) or to the ground state (c; 21.4%). The half-life for this metastable state is 4.480 h. The ground statepreferentially captures a neutron before it decays whenthe neutron density exceeds 0.9–5 � 108 cm�3 (dependingon temperature, see Fig. 6), whereas neutron capture forthe metastable state is negligible (Raut et al., 2013). At tem-peratures >3 � 108 K, possible thermalization of the isomercould reduce the effective half-life of 85gKr (Takahashi andYokoi, 1987; Raut et al., 2013). At typical s-process condi-tions in AGB stars, however, thermalization of 85Kr doesnot occur (Ward, 1977; Cosner et al., 1980), and the pro-duction ratio 85mKr/85gKr from 84Kr(n,c)85Kr is close tounity (Beer, 1991). This explains why 85Kr behaves differ-ently from all other nuclides shown in Fig. 6. Even at highneutron densities, the branching factor fn only reaches amaximum of 54%, since 84Kr(n,c)85Kr produces 58.6%85mKr (KADoNiS v1.0; Dillmann et al., 2014), of which78.6% b-decays to 85Rb. Hence, 46% of 85Kr is not avail-able for another neutron capture.
Nevertheless, at high neutron densities, up to 54% of thetotal 85Kr captures a neutron, generating 86Kr, eventuallyleading to an enrichment of 88Sr (Fig. 5). However, at least46% of the 85Kr decays to 85Rb, which could lead, via 86Rb(�19 d half-life), to 86Sr. Again, at high neutron densities,87Rb is produced, eventually leading to an enrichment of88Sr (Fig. 5). The decay of 87Rb has been ignored in theseconsiderations. With its half-life of �49 Ga, it can be trea-ted as stable here, although at temperatures >5 � 108 K, thehalf-life drops below 105 a (Takahashi and Yokoi, 1987).But even at such temperatures, decay of 87Rb would stillplay a very minor role at neutron densities >107 cm�3 (cf.Fig. 6). Furthermore, the refractory element strontiumand the volatile element rubidium should be highly frac-tionated during SiC grain condensation (Lugaro et al.,2003; Liu et al., 2015), leaving the grain relatively depletedin rubidium. Therefore, any contribution of decaying rubid-ium to strontium isotopes after grain formation isnegligible.
The s-process in low-mass AGB stars, the sources of themainstream grains, occurs in thermal pulses in the heliumintershell, the region between the helium-burning and
hydrogen-burning shells, when neutrons are released by13C(a,n)16O and 22Ne(a,n)25Mg. The first reaction occursat a temperature of 9 � 107 K and produces a neutron den-sity of about 107 cm�3 over a relatively long time of about104 a; the second reaction takes place at a temperature of3 � 108 K, lasts only a few years, but with a high peak neu-tron density of about 1010 cm�3 (Busso et al., 1999; Lugaroet al., 2003; Straniero et al., 2006; Cristallo et al., 2009). Thes-process is following the main path (Fig. 5) with branchingfactors fn for the two neutron sources of about 5% and 53%,respectively, at 85Kr and 0.2% and 63%, respectively, at86Rb (Fig. 6). On the main path, after the decay of 86Rbto 86Sr, some 86Sr is consumed by producing 87Sr and even-tually 88Sr via neutron captures. Strontium-88 with itsmagic neutron number and very low neutron capture crosssection is considered a bottleneck for the s-process. As afinal result, the activation of the 13C neutron source leadsto strong enrichments of 86Sr, 87Sr, and 88Sr, while theabundances of 86Sr and 87Sr are slightly reduced when the22Ne neutron source is activated (Lugaro et al., 2003).
In order to explain the strontium isotopes in X grains,the so-called weak s-process (e.g., The et al., 2007;Pignatari et al., 2010), which takes place in massive starsprior to supernova explosion has to be considered. Here,neutrons are provided by the 22Ne(a,n)25Mg reaction atthe end of core helium burning and in the subsequent con-vective carbon-burning shell. Mean neutron densities ofabout 105–106 cm�3 and peak densities of up to 3 � 107
cm�3 at the end of the core helium burning, as suggestedby The et al. (2007) and Pignatari et al. (2010), would notlead to significant branching at 85Kr or 86Rb (Fig. 6) at typ-ical temperatures of >2.8 � 108 K (Frischknecht et al.,2016). During shell carbon burning, however, temperaturesof about 109 K and neutron densities >1010 cm�3 were sug-gested at the bottom of the convective shell with a final neu-tron burst of up to 5 � 1012 cm�3 at temperatures of1.4 � 109 K (The et al., 2007; Pignatari et al., 2010). There-fore, neutron capture would become dominant at 85Kr and86Rb during shell carbon burning.
Such conditions could therefore lead to 88Sr enrichmentas observed for the X1 grain by following the alternativepath shown in Fig. 5. The lower 87Sr/86Sr ratio comparedto the mainstream grains (Fig. 4), however, can only beexplained if nucleosynthesis also followed in part the mains-process path producing some 86Sr (Fig. 5). While thepeculiarities of the 85Kr branch point, as discussed above,would lead to 46% 85Rb even at high neutron densities,the branch point at 86Rb would bypass 86Sr entirely. How-ever, neutron capture nucleosynthesis is localized at thebottom of the convection zone, which has characteristicconvective turnover timescales of hours (Pignatari et al.,2010). If the material was removed from neutron exposureat timescales comparable to or shorter than the �19 d half-life of 86Rb, significant decay of 86Rb would occur after thatremoval, and 86Sr would be produced without producingany substantial 87Sr. This could explain the low 87Sr/86Srratios observed. Another scenario to explain 88Sr enrich-ment would invoke a supernova neutron burst (T = 109
K, Nn = 1017 cm�3, duration 1 s) as suggested by Meyeret al. (2000) in order to explain molybdenum isotopic
T. Stephan et al. /Geochimica et Cosmochimica Acta 221 (2018) 109–126 119
signatures in X grains reported by Pellin et al. (1999, 2000)following a model by Howard et al. (1992). Although thiswould naturally explain 88Sr enrichment, it would also pre-dict most 87Sr and even more 86Sr being destroyed by theneutron burst (Howard et al., 1992), contrary to our obser-vation for the X1 grain in the present study. Some 88Sr inthe X1 grain could also have formed in an r-process, butthis would also not explain the low 87Sr/86Sr ratio sinceboth isotopes would be bypassed. The p-process nuclide84Sr is depleted. Therefore, r- and p-process as well as asupernova neutron burst can be excluded from playing asignificant role in strontium production in the X1 grainanalyzed here.
Nucleosynthesis for the X2 grains must have followed apath similar to the main s-process path (Fig. 5). The neu-tron density in the region where the X2 grains formed musthave been low enough that the 85gKr(n,c)86Kr reactionplayed no substantial role in order to explain the low88Sr/86Sr ratio (Fig. 4). A possible scenario to explain thestrontium isotopes in X2 grains is core helium burning dur-ing the weak s-process as described above. As for the X1grain, the low 87Sr/86Sr ratio compared to the mainstreamgrains (Fig. 4) puts some time constraints on the durationof the nucleosynthetic processes for the X2 grains. Signifi-cant decay of 87Rb must have occurred after convectionremoved material from the helium-burning core leadingto enriched 86Sr without producing any substantial 87Sr.
All neutron capture pathways in Fig. 5 bypass 84Sr,which is actually destroyed in the s-process, explainingthe depletion in the isotope compared to the other stron-tium isotopes in all types of presolar grains shown in Fig. 4.
4.2. Barium
Barium production is heavily influenced by a branchpoint at 134Cs (Fig. 5). The decay rate of 134Cs shows astrong temperature-dependence (Takahashi and Yokoi,1987). While at temperatures up to 5 � 107 K the half-lifeis �2 a, it drops to �21 h at 5 � 108 K (Takahashi andYokoi, 1987). However, since the neutron capture cross sec-tion for 134Cs is rather high (724 mb; Fig. 5), high neutrondensities could still lead to significant 135Cs production(alternative path in Fig. 5). Cesium-135 has a decay ratethat is also strongly temperature-dependent. While at tem-peratures up to 5 � 107 K, the half-life is �2 Ma, it drops to�4 a at 5 � 108 K (Takahashi and Yokoi, 1987). However,this is still long enough on the s-process timescale thatalmost all of the 135Cs produced from 134Cs will capture aneutron (Fig. 6) and turn into 136Cs, which has a half-lifeof �13 d. Cesium-136 either decays to 136Ba or capturesanother neutron to produce 137Cs. Cesium-137 decays to137Ba with a half-life of 30 a or, via another neutron cap-ture, transforms to 138Cs, ultimately producing 138Ba.Barium-137 and 136Ba are also partially converted via neu-tron capture to the magic neutron number nuclide 138Ba,which is another bottleneck in the s-process due to its verysmall neutron capture cross section.
Barium and cesium condensation behaves similar to thestrontium-rubidium pair during SiC grain formation.Refractory barium should condense into grains, in
agreement with high barium abundances in presolar SiCgrains (Amari et al., 1995), whereas volatile cesium shouldnot condense. Consequently, any contribution to bariumfrom decaying cesium after grain formation should be neg-ligible (Lugaro et al., 2003; Liu et al., 2014).
The measured barium isotope ratios (Fig. 4) for thethree grain types (mainstream, X1, and X2) can beexplained analogously to the considerations for strontiumisotopes: The s-process in low-mass AGB stars, where themainstream grains formed, in the helium intershell is fol-lowing the main s-process path (Fig. 5) with branching fac-tors fn for the
13C and 22Ne neutron sources of 0.02% and11%, respectively, at 133Xe, 6% and 46%, respectively, at134Cs, and about 100% for both neutron sources at 135Cs(Fig. 6). The branching point at 134Cs is extremely sensitiveto the actual temperature and neutron density. After thedecay of 134Cs, some 134Ba is consumed to produce 135Ba,136Ba, 137Ba, and, eventually, 138Ba via several neutron cap-tures. As a final result, all barium isotopes, except 130Ba and132Ba, will be enriched, with the highest enrichments in134Ba and 136Ba, and the lowest in 135Ba (Lugaro et al.,2003).
High neutron densities during shell carbon burning inmassive stars prior to the supernova explosion lead to the137Ba and 138Ba enrichments and the 134Ba depletionobserved in the X1 grain. Both the r- and p-processes couldhave played a minor role as well. Indeed, both p-processnuclides 130Ba and 132Ba might be slightly enriched in theX1 grain, but the large statistical errors do not allow anyfirm conclusions to be drawn here. A supernova neutronburst as suggested by Howard et al. (1992) and Meyeret al. (2000) would explain the 138Ba enrichment but wouldalso predict an almost total destruction of 134Ba (Howardet al., 1992), which is not observed for our X1 grain. Thes- and r-process nuclide 138Ba is strongly enriched com-pared to the barium isotopes that are only formed by thes-process (Fig. 5). Barium-137, which is also an s- and r-process nuclide, is less enriched than 138Ba. This can beexplained by the �30 a half-life of 137Cs, leading to a some-what delayed production of 137Ba after the grain had con-densed. This might also be the reason why there is noenrichment of 135Ba, another s- and r-process nuclide, inthe X1 grain. The long half-life of 135Cs (2.3 Ma) meansthat 135Ba would be produced long after the grain has con-densed without cesium in it.
As for strontium, barium production for the X2 grainsmust have followed the main s-process path (Fig. 5), prob-ably during core helium burning. An elevated temperatureincreased the decay rate of 134Cs according to Takahashiand Yokoi (1987). This and the relatively low neutron den-sity ensured that 134Cs(n,c)135Cs played no substantial role.Consequently, 134Ba is enriched in X2 grains. The fact thatX2 grains have an elevated 134Ba/135Ba ratio compared tothe mainstream grains (Fig. 4) again puts some time con-straints on the duration of the nucleosynthetic processes.As with 87Rb, significant decay of 134Cs (�2 a half-life, orshorter at higher temperatures) must have occurred afterconvection removed material from the helium-burning core,thus stopping the s-process path at 134Ba and not leading to135Ba.
120 T. Stephan et al. /Geochimica et Cosmochimica Acta 221 (2018) 109–126
4.3. Supernova mixing models
To evaluate if present supernova models can explain ourobservations, we used models from Rauscher et al. (2002)for Type II supernovae for stars with 15, 19, 20, 21, and25 M☉ initial mass and included calculations taking intoaccount element fractionation during grain condensation.The models from Rauscher et al. (2002) give the post-supernova zonal composition 25,000 s after the corebounce. Core bounce with the formation of a shock wavedescribes the onset of a Type II supernova explosion, when,after the collapse of the inner core, nuclear densities (�1014
g/cm3) are reached causing the implosion to halt and infall-ing material to bounce back (Janka et al., 2007). Twenty-five thousand seconds after the core bounce, a significantnumber of radioactive nuclides, which subsequentlydecayed to their stable products, were still present. How-ever, if grain condensation occurred prior to the decay ofa given nuclide, fractionation between parent and daughterelement would have taken place. Since rubidium andcesium are much more volatile than the refractory elementsstrontium and barium, we assumed that only the latter twowere incorporated into SiC during grain condensation.Thus, radioactive rubidium and cesium atoms as well asnoble gas (krypton and xenon) atoms that ultimatelydecayed into strontium or barium did not contribute tothe strontium and barium inventories of the grains today.On the other hand, radioactive atoms from refractory ele-ments yttrium, zirconium, lanthanum, and cerium that ulti-mately decayed to strontium or barium were probablyincorporated into the condensing grains. However, theircontribution is minor due to the low abundance of the rel-evant nuclides.
The results of such calculations, assuming grain conden-sation 1 a after the Rauscher et al. (2002) model, are shownin Fig. 7. For carbon and oxygen, the elemental mass frac-tions in the gas for the various supernova ejecta zones in anobject with 15 M☉ are given at the time of grain condensa-tion. For silicon, strontium, and barium, the mass fractionsof the various isotopes 4.567 Ga later are shown, assumingthat all elements that are not expected to condense into SiCgrains were removed 1 a after the initial model. The timechosen, 4.567 Ga, is the age of the Solar System and shouldbe regarded as the minimum age of the presolar grains. Atthis time, all relevant radioactive nuclides that had con-densed into the grain have decayed long ago.
In order to explain isotopic signatures of X grains aswell as a carbon-to-oxygen ratio above one, the conditionfor SiC condensation under equilibrium conditions(Larimer and Bartholomay, 1979; Lodders and Fegley,1997), deep and inhomogeneous mixing is necessary(Zinner, 2014). According to Zinner (2014), material fromthe inner Ni, Si/S, and O/Si zones, which experienced sili-con burning and an a-rich freeze-out from nuclear statisti-cal equilibrium (Ni zone), oxygen burning (Si/S zone), andneon and partial oxygen burning (O/Si zone), had to mixwith material from the outer He/C and He/N zones, whichexperienced incomplete helium burning (He/C zone) andhydrogen burning (He/N zone). We therefore tried to finda best match for the observed X1 and X2 isotopic
signatures by mixing different zones from the model andvarying the time of condensation.
For the observed X1 signature, it is not possible to comeup with a reasonable mixing model that would explain thesilicon, strontium, and barium isotopes and also yield C/O > 1 by mixing entire zones. A best match with theobserved isotope ratios is obtained by mixing material fromSi/S, O/Si, and O/Ne zones with the He/N zone (X1amodel, Table 2). However, such a complex mixture, involv-ing that many zones in arbitrary proportions not onlyseems very unlikely but also would yield a C/O ratio of0.016, which is impossible to reconcile with SiC condensa-tion. A higher C/O ratio of 0.32 is obtained by addingmaterial from the outer part of the O/C zone (X1b model,Table 2). Dividing the O/C zone into two distinct regions(O/CI and O/CII) seems reasonable with respect to the highvariability observed for strontium and barium isotopes inthat zone (Fig. 7). However, the proposed mix is even morecomplex, and the C/O ratio remains <1. Finally, a C/Oratio of 2.7 is obtained by including material from the verybottom of the He/C zone (X1c model, Table 2). The mixingwould be as complex as in the X1b model, and for bothX1b and X1c models, it seems unreasonable to invokematerial from almost all zones in arbitrary proportionsbut specifically excluding inner parts of the O/C zone andouter parts of the He/C zone. Although the X1c modelseems to deliver the best overall match with the observa-tional data and also yields C/O > 1, it does not reproducethe definitive feature of X1 grains, a d29Si/d30Si ratio of2/3 (Fig. 1).
Carbon isotope data would be helpful to distinguishbetween the three X1 models, but as mentioned before, suchdata are not available for the grain in this study. However,the X1b and X1c models return 12C/13C ratios higher thansolar, which is more typical for X grains (Zinner, 2014) thanthe low ratio suggested by the X1a model (Table 2),although the high 12C/13C ratio of model X1c is at theupper end of observed 12C/13C ratios in X grains.
If we, for the time being, ignore that grain formationunder the conditions discussed above seems impossible, itis still interesting to take a look at the influence of the con-jectured condensation time on isotope ratios. The isotopeabundances that are most susceptible to the actual conden-sation time are those for 86Sr, 134Ba, and 137Ba, which areaffected by the decay of 86Rb, 134Cs, and 137Cs, respectively(cf. Fig. 5). These nuclides are present in significant quanti-ties right after the core bounce, especially in the O/Ne zoneaccording to the model by Rauscher et al. (2002), and thiszone contributed significantly to the strontium and bariuminventory of the X1 grain (Table 2). Dust condensation inthe ejecta of SN1987A was observed about 1.4 a(�500 days; Wooden, 1997) after the explosion, which isin the range of the half-lives of these nuclides. A time scalefor grain condensation in supernova ejecta of about 1 a wasalso inferred by evidence for incorporation of live 49V (330d half-life) in X grains (Hoppe and Besmehn, 2002). Fig. 8shows the influence of the condensation time on the variousisotope ratios. Especially d(137Ba/136Ba) is strongly depen-dent on the condensation time, and a perfect match withthe X1 grain data was achieved for the X1a, X1b, and
Ni Si/S O/Si O/Ne O/C He/C He/N H
10-8
10-6
10-4
10-2
100 IIIII
Mas
s Fr
actio
n
C O
I
10-8
10-6
10-4
10-2
100
28Si 29Si 30SiM
ass
Frac
tion
10-17
10-15
10-13
10-11
10-9
10-7
10-5
84Sr 86Sr 87Sr 88Sr
Mas
s Fr
actio
n
2.0 2.5 3.0 3.5 4.0 4.5 5.010-19
10-17
10-15
10-13
10-11
10-9
10-7
130Ba 132Ba 134Ba 135Ba 136Ba 137Ba 138Ba
Mas
s Fr
actio
n
Interior Mass [M☉]Fig. 7. Model predictions for a 15M☉ Type II supernova according to Rauscher et al. (2002). The various zones are labeled according to theirmost abundant element(s) following the nomenclature of Meyer et al. (1995). The O/C and He/C zones are each subdivided into two separateregions (I and II) as indicated by the vertical dashed lines. The carbon and oxygen abundances are given at the assumed time of graincondensation, one year after the core bounce. After this time, only refractory elements are allowed to contribute to the isotopic abundancesshown via radioactive decay.
T. Stephan et al. /Geochimica et Cosmochimica Acta 221 (2018) 109–126 121
X1c models by assuming condensation times after the corebounce of 22 a, 17 a, and 12 a, respectively. If condensationof SiC would not fractionate cesium and barium or if con-densation occurred much later, after 137Cs has completelydecayed, 137Ba and 138Ba should be equally enriched rela-tive to 136Ba. A perfect match for the observed 138Ba/137Baratio would be obtained by assuming condensation times of3.6 a, 10.1 a, and 9.6 a in the X1a, X1b, and X1c models,
respectively. The X1c model also provides a perfect matchfor the 87Sr/86Sr and 134Ba/136Ba ratios. We note that thesecondensation time scales are larger than those observed insupernova ejecta and inferred from incorporation of 49Vin X grains.
The condensation times calculated from these models,however, should be taken with a grain of salt. Fig. 8 ismeant to demonstrate a general concept, namely how
Table 2Comparison between X grain data and Type II supernova model calculations according to Rauscher et al. (2002).
X1 grain X1a model X1b model X1c model X2 grainsa X2 model
Pre-supernova stellar mass [M☉] – 15 15 15 – 15Condensation time [a]b – 22 17 12 – 0.19Ni zone [%]c – 0 0 0 – 0Si/S zone [%] – 3.13 2.42 1.32 – 0.15O/Si zone [%] – 0.31 0.23 0.19 – 0O/Ne zone [%] – 28.09 19.20 8.43 – 0O/C zone [%] (I/II)d – 0 0/20.66 0 – 0He/C zone [%] (I/II)e – 0 0 58.15/0 – 70.74He/N zone [%] – 68.47 57.49 31.91 – 29.11H zone [%] – 0 0 0 – 0C/O [atoms/atoms]f – 0.016 0.32 2.7 – 8.112C/13C [atoms/atoms]g – 50 1492 7878 – 1729d(29Si/28Si) [‰]h �220 ± 45 �236 �265 �303 �431 ± 33 �420d(30Si/28Si) [‰] �389 ± 46 �376 �353 �317 �411 ± 45 �407d(84Sr/86Sr) [‰] �518 ± 73 �681 �639 �599 �570 ± 340 �649d(87Sr/86Sr) [‰] �234 ± 29 �355 �332 �188 �414 ± 68 �368d(88Sr/86Sr) [‰] +702 ± 40 +750 +1063 +1191 �449 ± 34 �353d(130Ba/136Ba) [‰] +260 ± 770 +524 +491 +529 – �703d(132Ba/136Ba) [‰] +560 ± 790 +344 +328 +319 – �583d(134Ba/136Ba) [‰] �300 ± 110 �153 �138 �306 +450 ± 310 +552d(135Ba/136Ba) [‰] �589 ± 53 �311 �279 �413 �629 ± 94 �521d(137Ba/136Ba) [‰] +310 ± 100 +309 +315 +308 �260 ± 110 �294d(138Ba/136Ba) [‰] +1670 ± 180 +935 +1368 +1567 �204 ± 91 �220
a Data for X2 grains are calculated by adding ion counts for SiC grains #6 and #15.b Condensation time is the time after the supernova core bounce when the SiC grain is assumed to have condensed.c Model values in % for various zones describe the mass fraction of the grain that is supposed to come from that zone.d The O/C zone was subdivided into regions I and II in the X1b model, which included material from the O/CII region (outer part of the
zone).e The He/C zone was subdivided into regions I and II in the X1c model, which included material from the He/CI region (inner part of the
zone).f The C/O ratio of the gas at time of condensation.g Model calculations for 12C/13C and for the d-values are predictions for a SiC grain observed today taking into account element
fractionation at the time of grain condensation.h Uncertainties for measured d-values are 2r resulting from counting statistics and, for silicon isotopes, from reproducibility of the
standards.
122 T. Stephan et al. /Geochimica et Cosmochimica Acta 221 (2018) 109–126
condensation times for some isotopes can have a largeimpact on their relative abundances, rather than providinga definite answer on the actual condensation time of our X1grain. It should be noted that these isotope ratios are notalways monotonic functions of condensation time. Both134Ba/136Ba and 137Ba/136Ba first decrease with increasingcondensation time due to the decay of 136Cs (13 d half-life) producing 136Ba. After most 136Cs has decayed, theratios increase due to decaying 134Cs (2 a half-life) and137Cs (30 a half-life).
Besides the 15M☉ supernova model, we also tested if 19,20, 21, and 25M☉ supernova models (Rauscher et al., 2002)would deliver a better match with our X1 grain data. How-ever, these models seem to be even less suited than the 15M☉ supernova model to reproduce the observational data.The pre-explosion zonal composition models also providedby Rauscher et al. (2002) did not match our X1 grain dataat all.
For the X2 grains, finding a mixing model that wouldreproduce the observed isotopic signature is much morestraightforward. An almost perfect match is achieved by
assuming a mixture of a small amount from the Si/S zonewith material from the He/C and He/N zones (Table 2).This is more plausible than the multizone mixing in the var-ious X1 models. In addition, the X2 model suggests a C/Oration of 8.1 at time of condensation for the grains, whichseems to be a favorable condition for SiC graincondensation.
Confining the time of condensation for the X2 grains,however, is less stringent. The 87Sr/86Sr, 134Ba/136Ba, and137Ba/136Ba ratios depend much less on the condensationtime in the X2 model than in the X1 models, and any time>0.036 a (>13 d) after the core bounce would be consistentwith the measured data (Fig. 8). The best match with theobservational data was achieved by assuming condensation0.19 a after the core bounce. However, since, according tothe model by Rauscher et al. (2002), only very little 86Rb,134Cs, and 137Cs is present right after the core bounce inthe He/C and He/N zones, the source of strontium and bar-ium in the X2 grains (Table 2), no significant conclusionscan be drawn from this data about condensation times ofthe X2 grains.
-500
0
+500
+1000
+1500
δ-va
lue
[‰]
Model X1a δ(87Sr/86Sr) δ(134Ba/136Ba) δ(137Ba/136Ba)
Model X1b
0.001 0.01 0.1 1 10 100-500
0
+500
+1000
Model X1c
δ-va
lue
[‰]
Condensation time [a]0.001 0.01 0.1 1 10 100 1000
Model X2
Condensation time [a]Fig. 8. Dependence of d(87Sr/86Sr), d(134Ba/136Ba), and d(137Ba/136Ba) values on the assumed condensation times for the three discussed X1models and the X2 model. The vertical line in each plot shows the condensation time assumed for the respective model to get the best fit(Table 2).
T. Stephan et al. /Geochimica et Cosmochimica Acta 221 (2018) 109–126 123
Besides the 15 M☉ supernova model, the 19, 20, 21, and25 M☉ supernova models (Rauscher et al., 2002) matchedour X2 grain data almost equally well, but the best matchwas still achieved with the 15 M☉ supernova model. Thepre-explosion zonal composition models (Rauscher et al.,2002) matched our X2 grain data also quite well, which isno surprise since the supernova explosion has little effecton strontium and barium isotopes in the He/C and He/Ne zones, where, according to our model, >99.8% of thematerial of the X2 grains comes from.
Multizone mixing models had some success in qualita-tively reproducing isotope signatures of X grains before(Yoshida and Hashimoto, 2004; Yoshida, 2007), but oftenfailed in quantitatively explaining all isotope ratios involved(Lin et al., 2010; Hoppe et al., 2010, 2012; Davis, 2011;Zinner, 2014), especially for X1 grains (Lin et al., 2010).One of the major difficulties in reproducing X grain signa-tures is the C/O > 1 requirement. It has been suggested thatcondensation of carbonaceous phases in Type II supernova
ejecta could be possible even at C/O < 1 (Clayton et al.,1999; Deneault et al., 2003, 2006). However, Ebel andGrossman (2001) showed that SiC should not condenseunder such conditions, and Lin et al. (2010) argued thatthe isotopic ratios in X grains would not agree with thatscenario. When comparing the X grain data for the ele-ments analyzed in this study with the models, C/O > 1was easy to achieve for the X2 grains only.
New supernova models by Pignatari et al. (2013, 2015)suggest a carbon- and silicon-rich C/Si zone at the bottomof the helium-burning shell as a site for SiC grain formationwithout involving selective, large-scale mixing. However,since strontium and barium data from these models have,to our knowledge, not been published yet, we were not ableto evaluate if they would fit better our grain data. It isworth mentioning that even by using the models byRauscher et al. (2002), the best fit for our X1 grain (X1cmodel) involves major contributions from the bottom ofthe He/C zone (named He/CI region in Fig. 7 and Table 2),
124 T. Stephan et al. /Geochimica et Cosmochimica Acta 221 (2018) 109–126
which coincides with the C/Si zone discussed by Pignatariet al. (2013, 2015).
Furthermore, our discussion in Sections 4.1 and 4.2 sug-gests that X1 grains should contain material from zonesthat have been influenced by shell carbon burning, whileX2 grains should not. Our mixing models invoking materialfrom the O/Ne zone, which has seen carbon burning for theX1 grains, and no such material for the X2 grains at leastqualitatively confirms these predictions.
5. CONCLUSIONS AND OUTLOOK
We found three X grains among 22 randomly selectedpresolar SiC grains. After Lin et al. (2002) first introducedthe X1 and X2 subtypes and later added X0 (Lin et al.,2010), our observations now show what might be the mostpronounced differences between X1 and X2 in strontiumand barium isotopes. While the X1 grain is stronglyenriched in 88Sr and 138Ba, the two X2 grains are enrichedin 86Sr and 134Ba.
The strontium and barium isotopic signatures suggestthat these elements in mainstream grains have been formedby the s-process in low-mass AGB stars, occurring in ther-mal pulses in the helium intershell. Strontium and bariumisotopes in X grains can be explained by the weak s-process, which takes place in massive stars prior to super-nova explosion. The X2 signature can be explained by neu-trons provided by the 22Ne(a,n)25Mg reaction at the end ofcore helium burning, whereas the X1 signature requiressubsequent carbon burning in a convective shell.
Supernova models by Rauscher et al. (2002) can in partexplain the observed isotope patterns, but, especially for thecase of X1, require implausible mixing of many zoneswithin the supernova, excluding some of the zones inbetween the extremes. The silicon isotope pattern of theX1 grain could not be reproduced in models that had C/O > 1. However, the general trend seems to be that in orderto explain the X1 signature, more material, compared to X2grains, from the inner (Si/S, O/Si, and O/Ne) zones mustmix in a complex way with material from zones (O/C,He/C, and He/N) further out. The X2 grains, on the otherhand, are well explained by a mixing model, which onlyrequires a tiny amount of material from the Si/S zone tomix with material from He/C and He/N zones, which ismuch more plausible than the complex mixing required tomatch the X1 data.
Strontium and barium isotopes can in principle deliverinformation on condensation times of SiC grains after asupernova explosion. While our data suggest that the X1grain condensed �10 a after the supernova, the X2 grainsmight have condensed within less than a year after theexplosion. However, we note that the condensation timescale of �10 a inferred for X1 grains is high compared towhat was reported in other studies. Clearly, for a bettertime constraint, more measurements and more precisesupernova models are required.
It should be emphasized that the discussion in this workis solely based on data from three supernova grains. This isby far not enough to draw final conclusions about the nat-ure of X1 and X2 grains in general. Obviously, more data is
required. Modern RIMS instruments such as CHILI canmeasure two or three elements simultaneously (Stephanet al., 2016), and many elements in subsequent analysis ses-sions. Elements like strontium and barium, but probablyalso other elements suited for RIMS analysis such as zirco-nium and molybdenum, are valuable in that they providestrong constraints on neutron exposure conditions and con-densation timing of SiC grains in Type II supernovae. Inparallel, more precise supernova models like, e.g., thoseby Pignatari et al. (2013, 2015) are also needed to keepup with analytical improvements.
ACKNOWLEDGEMENTS
We are grateful to guest editor Kevin McKeegan and twoanonymous reviewers for their helpful comments. Constructionand development of CHILI was supported by the NASA Sample
Return Laboratory Instruments and Data Analysis Program andthe Laboratory Analysis of Returned Samples Program throughgrants NNX07AL94G, NNX11AC21G, and NNX15AF78G, bythe University of Chicago, and by Argonne National Laboratory.RT was supported by NASA Headquarters under the NASA Earth
and Space Science Fellowship Program through grantNNX12AL85H. AMD acknowledges support by the NASA Cos-
mochemistry Program through grant NNX09AG39G. LLNL-JRNL-729492.
REFERENCES
Amari S., Hoppe P., Zinner E. and Lewis R. S. (1992) InterstellarSiC with unusual isotopic compositions: grains from a super-nova? Astrophys. J. 394, L43–L46.
Amari S., Lewis R. S. and Anders E. (1994) Interstellar grains inmeteorites: I. Isolation of SiC, graphite, and diamond; sizedistributions of SiC and graphite. Geochim. Cosmochim. Acta
58, 459–470.Amari S., Hoppe P., Zinner E. and Lewis R. S. (1995) Trace-
element concentrations in single circumstellar silicon carbidegrains from the Murchison meteorite. Meteoritics 30, 679–693.
Asplund M., Lambert D. L., Kipper T., Pollacco D. and ShetroneM. D. (1999) The rapid evolution of the born-again giantSakurai’s object. Astron. Astrophys. 343, 507–518.
Audi G., Kondev F. G., Wang M., Pfeiffer B., Sun X., Blachot J.and MacCormick M. (2012) The NUBASE2012 evaluation ofnuclear properties. Chinese Phys. C 36, 1157–1286.
Barzyk J. G., Savina M. R., Davis A. M., Gallino R., Gyngard F.,Amari S., Zinner E., Pellin M. J., Lewis R. S. and Clayton R.N. (2007) Constraining the 13C neutron source in AGB starsthrough isotopic analysis of trace elements in presolar SiC.Meteorit. Planet. Sci. 42, 1103–1119.
Beer H. (1991) Capture cross section measurements of krypton andxenon isotopes and the fundamental parameters of the s-process. Astrophys. J. 375, 823–832.
Busso M., Gallino R. and Wasserburg G. J. (1999) Nucleosynthesisin asymptotic giant branch stars: relevance for galactic enrich-ment and solar system formation. Annu. Rev. Astron. Astro-phys. 37, 239–309.
Clayton D. D., Liu W. and Dalgarno A. (1999) Condensation ofcarbon in radioactive supernova gas. Science 283, 1290–1292.
Cosner K., Iben, Jr., I. and Truran J. W. (1980) The effects ofunthermalized isomeric states and of a time-varying neutronflux on s-process branching ratios. Astrophys. J. 238, L91–L96.
Cristallo S., Straniero O., Gallino R., Piersanti L., Domınguez I.and Lederer M. T. (2009) Evolution, nucleosynthesis, and yields
T. Stephan et al. /Geochimica et Cosmochimica Acta 221 (2018) 109–126 125
of low-mass asymptotic giant branch stars at different metal-licities. Astrophys. J. 696, 797–820.
Davis A. M. (2011) Stardust in meteorites. Proc. Natl. Acad. Sci.
108, 19142–19146.Davis A. M., Pellin M. J., Tripa C. E., Savina M. R., Lewis R. S.,
Clayton R. N. and Amari S. (2002) Multielement analyses ofsingle presolar SiC grains from supernovae. Geochim. Cos-
mochim. Acta 66, A171 (abstr.).Deneault E. A. N., Clayton D. D. and Heger A. (2003) Supernova
reverse shocks: SiC growth and isotopic composition. Astro-phys. J. 594, 312–325.
Deneault E. A. N., Clayton D. D. and Meyer B. S. (2006) Growthof carbon grains in supernova ejecta. Astrophys. J. 638, 234–240.
Dillmann I., Szucs T., Plag R., Fulop Z., Kappeler F., Mengoni A.and Rauscher T. (2014) The Karlsruhe astrophysical databaseof nucleosynthesis in stars project – status and prospects. Nucl.
Data Sheets 120, 171–174.Ebel D. S. and Grossman L. (2001) Condensation from supernova
gas made of free atoms. Geochim. Cosmochim. Acta 65, 469–477.
Fairbank, Jr., W. M., Spaar M. T., Parks J. E. and Hutchinson J.M. R. (1989) Anomalous odd- to even-mass isotope ratios inresonance ionization with broad-band lasers. Phys. Rev. A 40,2195–2198.
Frischknecht U., Hirschi R., Pignatari M., Maeder A., Meynet G.,Chiappini C., Thielemann F.-K., Rauscher T., Georgy C. andEkstrom S. (2016) S-process production in rotating massivestars at solar and low metallicities. Mon. Not. R. Astron. Soc.
456, 1803–1825.Grefenstette B. W., Harrison F. A., Boggs S. E., Reynolds S. P.,
Fryer C. L., Madsen K. K., Wik D. R., Zoglauer A., Ellinger C.I., Alexander D. M., An H., Barret D., Christensen F. E., CraigW. W., Forster K., Giommi P., Hailey C. J., Hornstrup A.,Kaspi V. M., Kitaguchi T., Koglin J. E., Mao P. H., MiyasakaH., Mori K., Perri M., Pivovaroff M. J., Puccetti S., Rana V.,Stern D., Westergaard N. J. and Zhang W. W. (2014)Asymmetries in core-collapse supernovae from maps ofradioactive 44Ti in Cassiopeia A. Nature 506, 339–342.
Hoppe P. and Besmehn A. (2002) Evidence for extinct vanadium-49 in presolar silicon carbide grains from supernovae. Astro-phys. J. 576, L69–L72.
Hoppe P., Amari S., Zinner E., Ireland T. and Lewis R. S. (1994)Carbon, nitrogen, magnesium, silicon, and titanium isotopiccompositions of single interstellar silicon carbide grains fromthe Murchison carbonaceous chondrite. Astrophys. J. 430, 870–890.
Hoppe P., Leitner J., Groner E., Marhas K. K., Meyer B. S. andAmari S. (2010) NanoSIMS studies of small presolar SiCgrains: new insights into supernova nucleosynthesis, chemistry,and dust formation. Astrophys. J. 719, 1370–1384.
Hoppe P., Fujiya W. and Zinner E. (2012) Sulfur moleculechemistry in supernova ejecta recorded by silicon carbidestardust. Astrophys. J. Lett. 745, L26 (5pp).
Hoppe P., Cohen S. and Meibom A. (2013) NanoSIMS:technical aspects and applications in cosmochemistry andbiological geochemistry. Geostand. Geoanal. Res. 37, 111–154.
Hoppe P., Pignatari M., Kodolanyi J., Groner E. and Amari S.(2018) NanoSIMS isotope studies of rare types of presolarsilicon carbide grains from the Murchison meteorite: implica-tions for supernova models and the role of 14C. Geochim.
Cosmochim. Acta 221, 182–199.Howard W. M., Meyer B. S. and Clayton D. D. (1992) Heavy-
element abundances from a neutron burst that produces Xe-H.Meteoritics 27, 404–412.
Hynes K. M. and Gyngard F. (2009) The presolar grain database:http://presolar.wustl.edu/~pgd. Lunar Planet. Sci. 40, #1198(abstr.).
Janka H.-Th., Langanke K., Marek A., Martınez-Pinedo G. andMuller B. (2007) Theory of core-collapse supernovae. Phys.
Rept. 442, 38–74.Kappeler F., Beer H. and Wisshak K. (1989) S-process nucleosyn-
thesis—nuclear physics and the classical model. Rep. Prog.
Phys. 52, 945–1013.Kashiv Y., Davis A. M., Gallino R., Cai Z., Lai B., Sutton S. R.
and Clayton R. N. (2010) Extinct 93Zr in single presolar SiCgrains from low mass asymptotic giant branch stars andcondensation from Zr-depleted gas. Astrophys. J. 713, 212–219.
Lambert D. L., Gustafsson B., Eriksson K. and Hinkle K. H.(1986) The chemical composition of carbon stars. I. Carbon,nitrogen, and oxygen in 30 cool carbon stars in the galacticdisk. Astrophys. J. Suppl. S. 62, 373–425.
Larimer J. W. and Bartholomay M. (1979) The role of carbon andoxygen in cosmic gases: some applications to the chemistry andmineralogy of enstatite chondrites. Geochim. Cosmochim. Acta
43, 1455–1466.Levine J., Savina M. R., Stephan T., Dauphas N., Davis A. M.,
Knight K. B. and Pellin M. J. (2009) Resonance ionization massspectrometry for precise measurements of isotope ratios. Int. J.Mass Spectrom. 288, 36–43.
Lin Y., Amari S. and Pravdivtseva O. (2002) Presolar grains fromthe Qingzhen (EH3) meteorite. Astrophys. J. 575, 257–263.
Lin Y., Gyngard F. and Zinner E. (2010) Isotopic analysis ofsupernova SiC and Si3N4 grains from the Qingzhen (EH3)chondrite. Astrophys. J. 709, 1157–1173.
Liu N., Savina M. R., Davis A. M., Gallino R., Straniero O.,Gyngard F., Pellin M. J., Willingham D. G., Dauphas N. andPignatari M. (2014) Barium isotopic composition of main-stream silicon carbides from Murchison: constraints for s-process nucleosynthesis in asymptotic giant branch stars.Astrophys. J. 786, 66 (20pp).
Liu N., Savina M. R., Gallino R., Davis A. M., Bisterzo S.,Gyngard F., Kappeler F., Cristallo S., Dauphas N., Pellin M. J.and Dillmann I. (2015) Correlated strontium and bariumisotopic compositions of acid-cleaned single mainstream siliconcarbides from Murchison. Astrophys. J. 803, 12 (23pp).
Liu N., Nittler L. R., Alexander C. M. O’D., Wang J., PignatariM., Jose J. and Nguyen A. (2016) Stellar origins of extremely13C- and 15N-enriched presolar SiC grains: novae or super-novae? Astrophys. J. 820, 140 (14pp).
Lodders K. and Fegley, Jr., B. (1997) Condensation chemistry ofcarbon stars. In Astrophysical Implications of the Laboratory
Study of Presolar Materials, AIP Conference Proc. 402 (eds. T.J. Bernatowicz and E. K. Zinner). AIP, Woodbury, New York,pp. 391–423.
Lodders K., Palme H. and Gail H.-P. (2009) Abundances of theelements in the solar system. In Landolt-Bornstein, New Series
VI/4B (ed. J. E. Trumper). Springer Verlag, Berlin, Heidelberg,New York, pp. 34.1–34.59.
Lugaro M., Davis A. M., Gallino R., Pellin M. J., Straniero O. andKappeler F. (2003) Isotopic compositions of strontium, zirco-nium, molybdenum, and barium in single presolar SiC grainsand asymptotic giant branch stars. Astrophys. J. 593, 486–508.
Ma Z., Thompson R. N., Lykke K. R., Pellin M. J. and Davis A.M. (1995) New instrument for microbeam analysis incorporat-ing submicron imaging and resonance ionization mass spec-trometry. Rev. Sci. Instrum. 66, 3168–3176.
Marhas K. K., Hoppe P. and Ott U. (2007) NanoSIMS studies ofBa isotopic compositions in single presolar silicon carbidegrains from AGB stars and supernovae. Meteorit. Planet. Sci.
42, 1077–1101.
126 T. Stephan et al. /Geochimica et Cosmochimica Acta 221 (2018) 109–126
Meija J., Coplen T. B., Berglund M., Brand W. A., De Bievre P.,Groning M., Holden N. E., Irrgeher J., Loss R. D., Walczyk T.and Prohaska T. (2016) Isotopic composition of the elements2013 (IUPAC Technical Report). Pure Appl. Chem. 88, 293–306.
Meyer B. S., Weaver T. A. and Woosley S. E. (1995) Isotope sourcetable for a 25 M☉ supernova. Meteoritics 30, 325–334.
Meyer B. S., Clayton D. D. and The L.-S. (2000) Molybdenum andzirconium isotopes from a supernova neutron burst. Astrophys.J. 540, L49–L52.
Muller B. (2016) The status of multi-dimensional core-collapsesupernova models. Publ. Astron. Soc. Aust. 33, e048 (29pp).
Nicolussi G. K., Davis A. M., Pellin M. J., Lewis R. S., Clayton R.N. and Amari S. (1997) S-Process zirconium in presolar siliconcarbide grains. Science 277, 1281–1283.
Nicolussi G. K., Pellin M. J., Lewis R. S., Davis A. M., Clayton R.N. and Amari S. (1998) Strontium isotopic composition inindividual circumstellar silicon carbide grains: a record of s-process nucleosynthesis. Phys. Rev. Lett. 81, 3583–3586.
Nomoto K., Kobayashi C. and Tominaga N. (2013) Nucleosyn-thesis in stars and the chemical enrichment of galaxies. Annu.Rev. Astron. Astrophys. 51, 457–509.
Papanastassiou D. A. and Wasserburg G. J. (1969) Initialstrontium isotopic abundances and the resolution of small timedifferences in the formation of planetary objects. Earth Planet.
Sci. Lett. 5, 361–376.Pellin M. J., Davis A. M., Lewis R. S., Amari S. and Clayton R. N.
(1999) Molybdenum isotopic composition of single siliconcarbide grains from supernovae. Lunar Planet. Sci. 30, #1969(abstr.).
Pellin M. J., Calaway W. F., Davis A. M., Lewis R. S., Amari S.and Clayton R. N. (2000) Towards complete isotopic analysisof individual presolar silicon carbide grains: C, N, Si, Sr, Zr,Mo, and Ba in single grains of type X. Lunar Planet. Sci. 31,#1917 (abstr.).
Pellin M. J., Savina M. R., Calaway W. F., Tripa C. E., Barzyk J.G., Davis A. M., Gyngard F., Amari S., Zinner E., Lewis R. S.and Clayton R. N. (2006) Heavy metal isotopic anomalies insupernovae presolar grains. Lunar Planet. Sci. 37, #2041(abstr.).
Pignatari M., Gallino R., Heil M., Wiescher M., Kappeler F.,Herwig F. and Bisterzo S. (2010) The weak s-process in massivestars and its dependence on the neutron capture cross sections.Astrophys. J. 710, 1557–1577.
Pignatari M., Wiescher M., Timmes F. X., de Boer R. J.,Thielemann F.-K., Fryer C., Heger A., Herwig F. and HirschiR. (2013) Production of carbon-rich presolar grains frommassive stars. Astrophys. J. Lett. 767, L22 (6pp).
Pignatari M., Zinner E., Hoppe P., Jordan C. J., Gibson B. K.,Trappitsch R., Herwig F., Fryer C., Hirschi R. and Timmes F.X. (2015) Carbon-rich presolar grains from massive stars:subsolar 12C/13C and 14N/15N ratios and the mystery of 15N.Astrophys. J. Lett. 808, L43 (6pp).
Rauscher T., Heger A., Hoffman R. D. and Woosley S. E. (2002)Nucleosynthesis in massive stars with improved nuclear andstellar physics. Astrophys. J. 576, 323–348.
Raut R., Tonchev A. P., Rusev G., Tornow W., Iliadis C., LugaroM., Butain J., Goriely S., Kelley J. H., Schwengner R., Banu A.and Tsoneva N. (2013) Cross-section measurements of the 86Kr(c, n) reaction to probe the s-process branching at 85Kr. Phys.Rev. Lett. 111, 112501 (5pp).
Savina M. R., Davis A. M., Tripa C. E., Pellin M. J., Clayton R.N., Lewis R. S., Amari S., Gallino R. and Lugaro M. (2003a)Barium isotopes in individual presolar silicon carbide grainsfrom the Murchison meteorite. Geochim. Cosmochim. Acta 67,3201–3214.
Savina M. R., Pellin M. J., Tripa C. E., Veryovkin I. V., CalawayW. F. and Davis A. M. (2003b) Analyzing individual presolargrains with CHARISMA. Geochim. Cosmochim. Acta 67, 3215–3225.
Stadermann F. J., Walker R. M. and Zinner E. (1999) Sub-micronisotopic measurements with the CAMECA NanoSIMS. LunarPlanet. Sci. 30, #1407 (abstr.).
Stephan T., Zehnpfenning J. and Benninghoven A. (1994) Correc-tion of dead time effects in time-of-flight mass spectrometry. J.Vac. Sci. Technol. A 12, 405–410.
Stephan T., Trappitsch R., Davis A. M., Pellin M. J., Rost D.,Savina M. R., Yokochi R. and Liu N. (2016) CHILI – theChicago Instrument for Laser Ionization – a new tool forisotope measurements in cosmochemistry. Int. J. Mass Spec-
trom. 407, 1–15.Stephan T., Trappitsch R., Boehnke P., Davis A. M., Pellin M. J.
and Pardo O. S. (2017) Analysis of strontium, zirconium,molybdenum, and barium isotopes in presolar silicon carbidegrains with CHILI. Lunar Planet. Sci. 48, #2513 (abstr.).
Straniero O., Gallino R. and Cristallo S. (2006) S process in low-mass asymptotic giant branch stars. Nucl. Phys. A 777, 311–339.
Takahashi K. and Yokoi K. (1987) Beta-decay rates of highlyionized heavy atoms in stellar interiors. Atom. Data Nucl. Data
Tables 36, 375–409.The L.-S., El Eid M. F. and Meyer B. S. (2007) S-process
nucleosynthesis in advanced burning phases of massive stars.Astrophys. J. 655, 1058–1078.
Trappitsch R., Stephan T., Savina M. R., Davis A. M., Pellin M.J., Rost D., Gyngard F., Gallino R., Bisterzo S., Cristallo S.and Dauphas N. (2018) Simultaneous iron and nickelisotopic analyses of presolar silicon carbide grains. Geochim.
Cosmochim. Acta. 221, 87–108.Ward R. A. (1977) The importance of long-lived isomeric states in
s-process branching. Astrophys. J. 216, 540–547.Wooden D. H. (1997) Observational evidence for mixing and dust
condensation in core-collapse supernovae. In Astrophysical
Implications of the Laboratory Study of Presolar Materials, AIP
Conference Proc. 402 (eds. T. J. Bernatowicz and E. K. Zinner).AIP, Woodbury, New York, pp. 317–376.
Wunderlich R. K., Wasserburg G. J., Hutcheon I. D. and Blake G.A. (1992) Systematics of the odd-even effect in the resonanceionization of Os and Ti. Inst. Phys. Conf. Ser. 128, 127–130.
Yoshida T. and Hashimoto M. (2004) Numerical analyses ofisotopic ratios of presolar grains from supernovae. Astrophys. J.606, 592–604.
Yoshida T. (2007) Supernova mixtures reproducing isotopic ratiosof presolar grains. Astrophys. J. 666, 1048–1068.
Zinner E. K. (2014) Presolar grains. In Meteorites and Cosmo-
chemical Processes (ed. A. M. Davis), Vol. 1 Treatise on
Geochemistry 2nd Ed. (exec. eds. H. D. Holland and K. K.Turekian), Elsevier, Oxford, pp. 181–213.
Associate Editor: Anders Meibom
