The Distribution of Platinum and Palladium Metals in Iron Meteorites and in the Metal Phase of Ordinary Chondrites

Home , Ataxite, Bear Creek (meteorite), Iron meteorite, Kamacite, Pultusk (meteorite), Toluca (meteorite)

3OURNALOF GEOPHYSICALRESEARCtt VOL. 70, NO. 2 3ANUARY15, 1965

WALTER •ICI-IIPORI3K AND •-•ARRISON BROWN

Division of Geological Sciences California Institute o] Technology, Pasadena

Abstract. The concentrations of Ru, Rh, Pd, Ir, and Pt have been determined spectrographically in twenty-four iron meteorites and in the metal phase of five ordinary chondrites. It was found that most iron meteorites fall into three distinct groups with regard to the Ru and Rh concentrationsand into three groupswith regard to the Ir and Pt concentrations,each Ir-Pt group correspondingto one of the Ru-Rh groups. Correlations are observed between these Ir-Pt and Ru-Rh groups on the one hand, and the Ga-Ge groups found by previous workers on the other, but the relationshipsare by no means clear-cut. Compared with Ru, Ir, and Pt, Pd was found to vary over a relatively small range. The metal phasesof all chondrites studied are chemically identical with the iron meteorites of the largest Ru-Rh and the largest Ir-Pt group,namely the Ru-Rh group which containsabout 6 ppm Ru and 1.5 ppm Rh, and the Ir-Pt group which containsabout 2 ppm Ir and 7 ppm Pt. The following atomic abundances(St: 106) of Ru, Rh, Pd, Ir, and Pt have been derived from our data: Ru Rh Pd Ir Pt Basedon metal phaseof high-iron-groupchondrites 1.66 0.27 1.05 0.40 1.22 Basedon iron meteorites,assuming they constitute10% of mean meteoritic matter 1.44 0.23 0.52 0.31 0.89

These valuesgenerally agreewithin a factor of lessthan 2 with the abundancescalculated by recent workers.

INTRODUCTION concentrations of Pt and Pd metals in iron meteorites selectedby previous investigators. Pt and Pd metals are concentrated almost en- Brown and Goldberg [1949] and Goldberg tirely in the metal phase2 of meteoritic matter. et al. [1951] have determined Pd in forty-five In view of this, it is desirablethat accurate de- iron meteoritesby neutron activation, and Hara terminations of their distribution in iron me- and Sandell [1960] have determined Ru in teorites and in the metal phase of chondritesbe seventeeniron meteoritesand in the metal phase obtained. Abundances of these elements have of two pallasiresby absorptionspectrophotom- been determined spectroscopicallyin iron me- etry. Herr ei ah [1958] have also determined teorites by Noddack and Noddack [1930] and Ru in the Carbo iron meteorite by neutron ac- alsoby Goldschmidtand Peters [1932], but the tivation. The ranges of concentration obtained estimatesare only approximate [Suessand Urey, by all these workers are shownin Table 2 and 1956]. Goldschmidt [1938] averaged these comparedwith the rangesof concentrationob- analysesand Suess and Urey [1956] reported served in this study. interpolated values. Table I lists the average The averagesand the precisionof the data of Goldberg et al. for Pd are respectively3.7 ppm x Contribution 1275, Division of Geological Sci- and approximately 10%, and those of the data ences,Seeley W. Mudd Laboratory, California In- of Itara and Sandell for Ru are respectively6.6 stitute of Technology, Pasadena. ppm and about 5%. •' We thank Dr. Goles for pointing out to us that As we shall see, the ranges of concentration the metal phase as used in this paper is not a observedin this study of Ir, Pt (see also Nichi- phase in the strict sensebut rather is a mixture of true phases,kamacite and taenite, which were not poruk and Brown [1962]), and Rh in iron me- separated in this investigation. teorites are just as large as if not larger than

459 460 NICHIPORUK AND BROWN TABLE 1. Abundances of Pt an] Pd Metals in HC1 gas into distilled water in a flask. HN03 Iron Meteorites Selectedby Previous Investigators was distilled under vacuum in a quartz still. The residuewas filtered, dried, and weighed; nitrate Goldschmidt Suess and was destroyed by evaporation with HC1. The Metals [19381 Brown [1949] Urey [1956] solution was made 8.1 N in HC1, then trans- fered into a l-liter separatory funnel. This solu- Ru, ppm 10 10.6 10 tion contained all Pt and Pd metals except Os. Rh, ppm 5 4.1 1.44 The Fe (as HFeCl•) was extractedtwice with Pd, ppm 9 3.7* 5.7 Os, ppm 8 7.6 12.2 600-mlportions of redistilledisopropyl ether Ir, ppm 4 3.0 10.1 [Dodsonei al., 1936;Nachirieb and Fryxell, Pt, ppm 20 19 20.2 1948],and the aqueousHC1 layer containing the metals was drawn off into a 200-ml Berzelius *The average of determinations by Brown and beaker. After the addition of 40 mg of Cu as a Goldberg[1949]. carrier and 0.5 mg of Mo as an internal standard, the aqueouslayer was evaporatedto about the rangesof concentrationsof Pd and Ru. By 20 ml and then saturated with C12gas to oxidize comparison,the rangesof concentrationof Pt all Mo to the +6 state. The solutions were and Pd metalsin the metallic-ironphase of ordi- evaporatedto drynessand the solidswere dis- nary, high-iron-groupchondrites were found to solved in 20 ml of 1.5N HC1 which contained be very small. A1C13'6H•Oas a coagulant for Ir sulfide. The

ANALYTICAL METHODS Ru, Rh, Pd, Ir, and Pt sulfideswere precipitated with Cu and Mo sulfidesby bubbling It, S gas Pt and Pd metals were isolated from iron me- into warm solutions in 200-ml beakers for 3 teorites and the metal phase of chondritesby hours and allowing the precipitates and super- precipitationas sulfidesand were determined nates to stand for 48 hours. The sulfides were spectrographically. collectedin sintered glasscrucibles, and the pre- Samplesranging in weight from 8 to 20 g cipitation was repeated. The two precipitates were cut from larger portionsof iron meteorites, were thoroughly mixed in dilute ItC1 and ethyl care being taken to avoid discernibleinclusions alcoholand dried at 100øC; their weights ranged of troilite, schreibersite,and cohenire.Samples from 120 to 130 mg for all samplesanalyzed. of Ituizopa, Indian Valley, and Tazewell were All compoundsand metals used in this work in the form of fine sawingsand were residues were 'specpure'grade, obtainedfrom Johnson, from other work. All sampleswere washedin Matthey and Co., Ltd. A series of standards 6 N HC1, quadruplydistilled water, and ethyl was prepared containingfrom 15 to 4000 ppm alcohol to remove surface impurities. of each metal and 1.0 mg of Mo as an internal Specimensof chondritesweighing 70 to 90 g standard. The metals were present in a matrix were crushed in a Plattner mortar to about 100 of 120 to 130 mg of CuS. The standard solutions mesh and separatedinto 'magnetic'and 'non- composedof 15 g of Fe, 1.4 g of Ni, and 0.08 g magnetic' fractions with a hand magnet. The of Co were converted into Cu-Mo sulfides in the 'magnetic'fractions used in this work were com- same manner as the meteorite solutions,thus posedof metallic grains with inclusionsof fine silicate and troilite particles which were em- beddedin the grainsduring the crushingprocess. TABLE 2. Ranges of Concentration of Pd and The 'magnetic'fractions ranged in weight from Ru in Iron Meteorites As Found by Previous Workers and in This Study 7 to 13 g. All chemistry was performed in pyrex glass- Investigators Ru, ppm Pd, ppm ware which before use had been thoroughly washed and kept for at least 2 hours in hot, Goldbertet al. [1951] 1.44-9.88 concentrated, reagent grade ItNO•. The sam- Herr et al. [1958] 52.4 -v 1.5 ples were brought into solutionin 400-ml beak- Hara and Sandell [1960] 0.8-14.8 ers in a mixture of hot, concentrated ItC1 and This work 0.5-23.7 1.45-6.47 HNO•. The HC1 was obtainedby bubblingtank DISTRIBUTION OF Pt AND Pd IN METEORITES 461 ensuringclose physical and chemicalresemblenee Plate calibration. Thirteen selected iron lines. of both samples and standards. Inasmuch as Each plate was calibrated, and standards,in- iron meteorites contain about 10 ppm Mo cluding a blank, were exposedon each plate. [Kuroda and Sandell, 1954], an equivalent Densitometer. Jarrel-Ash model 2100. amount of the element (0.17 mg) was added to Sensitivity. Ru 30 ppm; Rh 10 ppm; Pd 10 each standard solution before its extraction with ppm; Ir (Ir 2849.725A) 30 ppm; Pt 15 ppm; ether. in a matrix of copper and molybdenumsulfides To establish whether significantly variable correctedfor quartz dilution and assuming100% yields of the investigated metals might have precipitationyields of the metals. beenobtained, minute quantitiesof 0sTM (•-par- ticle emitter 0.18 Mev, 15 days half-life) tracer PRECISION AND ACCURACY were added to three HC1 solutions of Fe, each The concentration of Ir in the Bear Creek, of which contained Ni and Co, and the meas- ured traces of all Pt and Pd metals. Within a Sandia Mountains, and Sikhote-Alinmeteorites was somewhatbelow the sensitivitylimits of the counting accuracy of 5%, the yields of 0sTM, element in the enriched samplesunder the con- correctedfor absorptionand scatteringeffects in ditions described.All sampleswere exposedbe- a matrix of Cu and Mo sulfides,were found to tween two and four times, and the arithmetic be constant. mean is reported. As ions of all Pt and Pd metals are expected An indication of the precisionof the method to behave similarly, it appears reasonablycer- may be obtained from the data on six iron me- tain that the yields of Ru, Rh, Pd, Ir, and Pt, teorites shown in Table 3. The duplicate sam- like the yields of Os, which were not determined pies were of approximately equal size and were in this study, did not vary appreciably under taken from either adjoining or separated loca- the conditionsof the enrichmentprocedures de- tions of each of the six meteorites listed. scribed. The differences between two samples could The standardsand sampleswere exposedusing arise in part from possible variations in chem- the following equipment and methods: ical yields, from spectrographieerrors (e.g., due to self-absorption in the analytical lines at Spectrograph. Jarrell-Ash 3.4-m grating in- higher concentrationsof the metals), from the strument with 15-thousand-lines-per-inchgrat- inhomogeneity of the samples, and from con- ing; Wadsworth mount, dispersion5.2 A/ram in first order.

Electrodes. High-purity x/5•-in.graphite rods TABLE 3. Concentrations of Five Pd and Pt as anode. Pointed x/s-in. rods as cathode. Metals in Duplicate Samplesof Six Iron Meteorites Excitation. 10-amp dc arc (13-amp short circuit) from a Jarrell-Ash Uni-source. Sample Nanhe of Ru, Rh, Pd, Ir, Pt, as anode. Analytical gap 4 mm magnified 5X Meteorite ppm ppm ppm ppm ppm and focused on the slit. Central 2 mm used with a slit width of 10 •. Samplesand standards Arispe I 15.1 1.7 7.3 ground with pure quartz in proportions of I to Arispe II 12.3 1.7 3.1 7.9 17.2 0.5, and 25-mg portions burned to completion Canyon DiabloI 5.9 1.5 1.6 6.6 after an initial pre-arcing at 2.5 to 3.0 amp; Canyon Diablo II 6.9 1.5 3.6 1.9 8.1 total time of burning 4 min. Total energymethod Henbury I 6.5 1.4 4.4 2.7 6.4 with internal standardization. Henbury II 5.6 1.1 3.4 1.9 4.9 Wavelengths. The analytical lines used were Ru 3436.737 A, Rh 3396.85 A, Pd 3421.24 A, Ir Mount TabbyI 17.6 2.2 2.8 10.5 25.2 2543.971 A, Ir 2849.725 A, and Pt 2659.454 A. Mount TabbyII 15.6 2.3 3.3 7.6 21.9 The internal standard line was Mo 2816.154 A Odessa I 6.0 1.3 1.9 8.9 (singly ionized molybdenum). Odessa II 6.0 1.4 3.7 1.8 8.1 Plates. Eastman Kodak III-0 developed for Toluca I 4.9 1.0 1.9 7.5 4 rain at 20øC in DK-50 developer, 30 sec in Toluca II 5.7 1.2 4.0 2.2 5.6 short stop, 10 rain in acid fix, and 20 rain wash. 462 NICItIPORUK AND BROWN TABLE 4. Ru and Pd Contentsof Iron MeteoritesDetermined by DifferentMethods of Analysis*

Ru, ppm Pd, ppm

Photometric, NeutronActivation, Nameof Hara andSandell Spectrographic, Goldberget al. Spectrographic, Meteorite Class [1960] This Work [1951] This Work

Altonab Of 3.6 4.7 4.45 4.2 Arispe Og 10.0 13.7 2 69 3.1 Bear Creek Om 5 57 5.4 Bristol Of 4 36 3.9 Canyon Diablo Og 5.9 6.4 3 98 3.6 Cape of Good Hope D 7 O7 4.1 Coahuila 1 44 2.0 Costilia Peak Om 12.4 13.4 2 02 2.5 Coya Norte It 14.8 15.2 Edmonton Of 0.8 0.5 5 97 65 Goose Lake Om 4.6 6.4 3 34 39 Henbury Om 3.6 6.0 2 02 39 Indian Valley H 1 54 15 Odessa Og 4 15 37 Sandia Mountains It 7.5 7.2 2 24 21 Spearman Om 4.8 4.6 3 67 29 Tazewell Off 7.73 57 Toluca Om 4.72 40

* Most of the specimensfor which comparisonsare made in this table are from the originalcollection usedby Goldberget al. [1951].

tamination of the sampleswith troilite or schrei- sesof the Henbury meteorite. As all determina- bersite. tions were made on the meteoritespecimen used The average deviation from the mean is esti- by Goldberget al. [1951], the observeddis- mated to be 8% for Ru, 10% for Rh, 9% for crepancywould appear to indicate that appre- Pd, 17% for Ir, and 12% for Pt. ciablefluctuation exist in the chemicalcomposi- To obtain an indication regarding the ac- tion of the meteorite. It is possible,however, curacy of the spectrographicmethod a compari- that samplesof this meteoritemight have been son was made of the Pt and Pd metal results in accidentallymixed with another. this work with the resultsobtained by Goldberg The comparisonpresented in Table 5 shows et al. [1951], by Hara and Sandell [1960], and that whereasthe threesets of the determinedRu, by Yavnel' [1950]. Further, the Ir resultson the Rh, Pd, and Pt valuesof Ya.vnel' [1950] are metal phaseof several chondriteswere compared generallyof poor precision,his selectedvalues, with thoseobtained by Rushbrookand Ehmann exceptthat for Pd, agreeclosely with our values. [1962]. As for Ir, it is clear from Table 6 that the From the comparisonspresented in Table 4 spectrographicmethod yields values that are in it can be seen that spectrographicRu deter- good agreementwith thoseobtained by neutron minationsare generallyabout 20 to 30% higher activation.The fact that Yavnel' [1950] did not than those determinedby absorptionspectro- detectIr in the Sikhote-Aliniron meteoritespec- photometry. Pd determinedspectrographically trographically (footnote, Table 5) providesan actually showstwo trends. At concentrationsbe- independentcheck, even though only qualita- low about 3 ppm spectrographicPd determina- tive, on those of our spectrographicIr values tions tend to be higher than Pd determinedby which are lower than 0.5 ppm (Table 7). neutron activation; abovethe 3-ppm level they RESULTS tend to be lower.The moreserious discrepancy, lying outsidethe limits of reproducibility(Table The concentrationsof Ru, Rh, Pd, Ir, and Pt 3) of the spectrographiemethod, is in the analy- in twenty-four iron meteorites are given in DISTRIBUTION OF Pt AND Pd IN METEORITES 463 TABLE 5. Ru, Rh, Pd, and Pt Contents of the among the different groups in the following Sikhote-Alin Meteorite manner: Of the nine meteoritesbelonging to Ru-Rh Ru, Rh, Pd, Pt, group I, six belongto Ir-Pt group I; one belongs ppm ppm ppm ppm to Ir-Pt group III; and two belong to Ir-Pt group I on the basisof their Pt contentsand to Yavnel' (1950)* Ir-Pt group II on the basisof their Ir contents. Sample 1 0.4 0.2 2.8 2.4 Of the twelve meteorites which belong to Sample2 5.7 0.9 6.9 4.6 Ru-Rh groupII, as many asnine belongto Ir-Pt Sample3 3.4 0.8 2.0 2.7 group II; the remainingthree belongto Ir-Pt Accepted group III. contents 5.7 0.9 6.9 4.6 The only two meteorites which are in Ru-Rh This work 5.0 4-0.1 1.6 4-0.1 3.9 4-0.3 group III do not belong to any of the Ir-Pt (esti- groups. mated) It is of interest that the Ru/Ilh ratio varies only from 5.7 to 10.3 for the meteorites of * High-temperature extraction of the metals into Ru-Rh group I, from 3.1 to 6.7 for those of a lead globulefollowed by spectrographicdetermina- tion. The highest of the three sets of values selected group II, and from 1.6 to 3.2 for thoseof group because the low values are believed to be the result III; the Pt/Ir ratio varies from 1.2 to 3.7 for of probable high-temperaturevolatilization lossesof the meteoritesof Ir-Pt group I, from 1.8 to 4.7 the metals. Analytical error is estimated to be about for those of group II, and from 6.0 to 33.8 for 20%. Ir and Os have not been detected. Sample sizesin the range of 25 g. those of group III. Structures and nickel contents. When the Table 7. The correspondingconcentrations in structuresand the known Ni contents (Table 7) the metal phasesof five ordinary chondrites,to- of the meteoritesbelonging to each Ru-Rh and gether with the observed proportions of these Ir-Pt group are examined,it is found that the phases,are given in Table 8. meteoritesof Ru-Rh group I and Ir-Pt group I Ru-Rh and It-Pt groups. The Ru contents have Ni contentslying between 5.5 and 16.5% of the iron meteorites are plotted against the and embrace hexahedrites,coarse and medium Rh contentsin Figure 1, and the Ir contentsof the meteorites are plotted against the Pt con- TABLE 6. Ir Content of the Metal Phase of Ordinary Chondrites by Different Methods of tents in Figure 2. Inspectionof the figuresshows Analysis that the Ru and Ir levels are concentrated in three well-separatedand mutually correspond- Ir, ppm, Ir, ppm, ing regions which for conveniencewe called Neutron Spectro- Ru-Rh groups I, II, and III and Ir-Pt groups Name of Activation,* graphic, I, II, and III. The ranges of concentrationas- Mete- Rushbrook and This orite Ehmann [1962] Name Work sociatedwith each of the groupsare as follows:

Elenovka 3.0 Ru-Rh Groups Ir-Pt Groups Alamogordo 3.0 4- 0.3 Forest 2.7 Gilgoin Station 2.6 4- 0.3 Ru, Rh, Ir, Pt, City ppm ppm ppm ppm Okhansk 3.0 Okhansk 2.9 4- 0.4

Group I 11.9-21.5 1.7-2.5 7.6-15.3 11.8-29.3 Plainview 3.0 Plainview 3.2 4-0.6 4.6- 7.2 0.7-1.6 1.4- 3.2 4.4- 8.5 Group II Pultusk 2.4 Gladstone 2.1 4-0.2 Group III (1.0 (0.5 (0.5 2.4- 9.8 Average 2.85 Average 2.8 4- 0.4 Only one of the meteorites studied (Bear value value Creek) was found to have a Ru content and two (Edmonton and Tazewell) to have a Pt content * Complete samples of the chondrites were ana- lyzed for Ir and the results normalized to the known lying outside these rather narrow limits. The metal phase content, assuming 100% siderophile remaining twenty-one meteorites are divided character of the element. 464 NICHIPORUK AND BROWN

TABLE 7. Ru, Rh, Pd, Ir, and Pt Contents of Iron Meteorites

N•me of Ni, l Ru, Rh, Pd, Ir, Pt, Meteorite Type* % ppm ppm ppm ppm ppm

Altonah Of 8 56 4.7 -V 0.1 0.7 q- 0.1 4.2 -V 0.1 1.6 q- 0.2 5.1 q- 0.6 Arispe Og 6 77 13.7 4- 1.5 1.7 -V 0.1 3.1 q- 0.2 7.6 q- 0.3 17.2 -V 0.1 Bear Creek Om 10 14 2.5 -V 0.4 1.5 q- 0.3 5.4 q- 0.6 <0.4 2.4 -V 0.1 Bendego Og 68 11.9 4- 0.7 2.1 4- 0.4 2.7 4- 0.9 0.3 4- 0.1 9.8 4- 1.0 Bristol Off 815 5.1 -V 1.2 0.8 q- 0.1 3.9 4- 0.3 1.7 q- 0.4 5.0 -V 0.6 Canyon Diablo Og 718 6.4 4- 0.4 1.5 + 0.1 3.6 + 0.2 1.8 + 0.2 7.4 4- 0.7 Cape of Good Hope D• 16 48 13.3 4- 0.2 1.7 4- 0.3 4.1 4- 0.1 8.5 4- 2.1 11.8 4- 0.2 Coahuila H 5 65 23.7 4- 2.5 2.3 4- 0.1 2.0 + 0.1 14.8 -V 0.1 24.2 -V 0.1 Costilia Peak Om 7 59 13.4 4- 0.2 1.7 -V 0.1 2.5 4- 0.2 15.3 4- 2.3 17.7 4- 3.3 Coya Norte H 55 15.2 -V 2.0 2.5 + 0.2 1.8 4- 0.3 3.3 4- 0.5 21.9 4- 5.6 Edmonton Off-Of 12 66 0.5 -V 0.1 0.3 q- 0.0 6.5 4- 0.8 0.6 q- 0.1 0.5 4- 0.1 Goose Lake Om 8 46 6.4 -V 0.5 1.2 + 0.1 3.9 q- 0.5 2.2 q- 0.3 6.8 -V 0.7 Henbury Om 7 66 6.0 4- 0.1 1.2 + 0.2 3.9 4- 0.5 2.3 4- 0.4 5.6 4- 0.8 Huizopa Of 7 81 5.7 + 0.5 0.9 + 0.1 4.8 + 0.2 2.5 + 0.1 4.4 4- 0.3 Indian Valley H; D•gr-H 5 64 21.0 -V 1.2 2.1 + 0.1 1.5 4-0.1 7.9 4- 1.7 29.3 4-2.6 Moonbi Om-Of 7 99 6.3 q- 0.1 1.3 -V 0.1 3.5 q- 0.1 1.4 q- 0.1 9.2 q- 0.7 Mount Tabby (?) Og 6 82 16.6 -V 1.0 2.2 4-0.1 3.0 4-0.2 9.1 4- 1.4 23.5 4- 1.6 Odessa Og-Ogg 7 40 6.0 q- 0.1 1.4 4- 0.1 3.7 q- 0.2 1.8 q- 0.1 8.5 q- 0.4 Rio Loa H 5 70 21.5 q- 2.4 2.5 -V 0.6 2.2 4- 0.5 3.4 4- 0.6 26.9 q- 4.6 Sandia Mountains Hgr 5 94 7.2 -V 1.0 1.5 4- 0.3 2.1 4- 0.4 <0.4 9.0 4- 1.0 Sikhote-Alin Hgr-Ogg 5 68 5.0 q- 0.1 1.6 q- 0.1 <0.3 3.9 -V 0.3 Spearman Om 8 39 4.6 q- 0.1 1.2 4- 0.1 2.9 q- 0.2 0.4 q- 0.1 6.7 -V 0.2 Tazewell Off 16 69 0.5 q- 0.1 0.14 4- 0.01 5.7 q- 0.8 0.3 q- 0.1 0.5 -V 0.1 Toluca Om 8 31 5.3 4- 0.4 1.1 4- 0.1 4.0 4- 0.2 2.0 4- 0.1 6.5 4- 0.9

* According to Prior [1953] and Lovering et al. [1957]. • Accordingto Goldberget al. [1951], Loveringet al. [1957], and Yavnel' [1954]. octahedrites,according to the classificationpro- two groupsappear to lie in the neighborhoodof posedby Lovering et al. [1957], and also a Ni- 5.5%. rich ataxite. The former group contains about The meteorites of Ru-Rh group II and Ir-P• 70% and the latter about 30% of the hexahe- group II, each of which containsabout one-half drites studied. The lower Ni boundaries of the of the meteorites studied, range in Ni content

TABLE 8. Ru, Rh, Pd, Ir, and Pt Contentsof the Metal Phaseof Chondrites

Metal Name of Phase, Ru, P•h, Pd, Ir, Pt, Meteorite Class % ppm ppm ppm ppm ppm

Alamogordo Crystalline spherical 13.5 6.14-0.8 0.94-0.1 4.24- 0.6 3.04- 0.3 7.94- 0.4 chondrite

Gilgoin Station Crystalline bronzite 19.6 5.4 4- 0.2 1.04-0.2 3.7 4- 0.3 2.64-0.3 8.0 4- 1.6 chondrite

Gladstone Black-veined crystalline 15.4 5.94-0.3 1.04-0.1 3.94- 0.2 2.1 4- 0.2 8.5 -4-- 1.4 spherical chondrite Okhansk Polymict brecciated 20.8 6.2 4- 0.6 1.1 4-0.1 3.74-0.1 2.94-0.4 9.14-1.7 spherical bronzite chondrite

Plainview Polymict brecciated 15.0 6.54-0.4 1.04-0.1 4.7 4- 0.3 3.24-0.6 9.04-0.9 veined intermediate chondrite DISTRIBUTION OF Pt AND Pd IN METEORITES 465

4O belongto both Ru-Rh group III and Ir-Pt group 50 III, the relationships,if any, between these 2O I "%1Group I , %,, Rht7-2.5ppm groupsand Ga-Ge groups,in particular Ga-Ge •.j Rut2- 25 pprn group III, are not completelyclear.

r ...... ] Groupff A more comprehensivepicture of the gen- , Rh 0.7-1.7 ppm eral relationships between the groups, struc- •____.4_•Ru 4.5-7. Sppm E tural classes,and Ni contentsof the meteorites is shown in Figure 3, where Ni concentrationis Explanation plotted againstkamacite bandwidth in order of [] Normal hexahedrites .-= 1.o ß Granular hexahedrdes decreasingcoarseness. The Ru-Rh, Ir-Pt, and - N•ckel-poor ataxites Ga-Ge groupsare found to be only nonuniquely A Coarse octahedrites *-0.5 !-(•...... (• GroupfTf ...... Rh (0.5 pprn ß Medium octahedrites associatedwith the well-defined structural fields, Ru (1.0pprn o Fine octahedrites as shown in Figure 3. Meteorites lying within ß Nickel- rich ataxites ß Chondrites, metal phase any of these fields will with high probability have Ru, Rh, Ir, Pt, Ga, and Ge contentscorre- 0.10.1 O.' 5 1.•0 t•0 spondingto any of the three Ru-Rh, Ir-Pt, and Ga-Ge groups,thus makingit appearthat there Rhodium, ppm is little or no immediately discerniblerelation- Fig. 1. Relation between Ru and Rh contents ship betweenstructure and Ga-Ge, Ru-Rh, and of meteorites. Ir-Pt groups. Palladium. The behavior of Pd was found to from 5.5 to 9.0% and consistchiefly of medium be quite different from that of Ru, Rh, Ir, and and fine octahedrites. Pt. While the Ru, Rh, Ir, and Pt concentrations The two meteoritesof Ru-Rh group III have generally increasewith increasingconcentration Ni contents of 12.7 and 16.7% and are fine of one another, the Pd concentration, which octahedrites; the five meteorites of Ir-Pt group varies only over the relatively narrow range of III range in Ni content between 5.7 and 10.1% 1.45 to 6.47 ppm, decreaseswith increasingcon- and encompasshexahedrites and medium and centration of each of these metals. Although the coarse octahedrites. Ru, Rh, Ir, and Pt concentrationsgenerally de- CORRELATIONS creasewith increasingNi concentrationand de- creasingkamacite bandwidth, the Pd concentra- Ga-Ge groups. It will be recalled that the tion increases. ranges of concentration in the four known Goldberg et al. [1951] pointed out that Pd Ga-Ge groups are as follows'

Ga, ppm Ge, ppm 20 I I Explanation GroupI I Group I 80-100 300-420 n Normalhexahedrites Ir 7-16 ppm I , lO ß Granularhexahedrdes Pt11-50 ppm Group II 40-65 130-230 a N•ckel-poorataxites Group III 8-24 15-80 A Coarseoctahedrites ß Medium octahedr•tes ...... --,-, Group IV 1-3 < 1-1 o F•neoctahedr•tes r e•'],*'.... •OA_e ©N•ckel-rich atax•tes I! ,• • AA!©I GroupIr 1.5-5.5ppmTT Of these four groups, Ga-Ge groups II, III, ©Chondr•tes, metalphase L_•_____• Pt4.5-9.5ppm and IV, which comprise twenty out of the twenty-four meteorites examined in this study, are represented in the correlations shown in [ß ß ß i GroupfiT Aj Ir (O.Spprn Table 9. It is interesting to note that the me- Pt 2.5-10 ppr teorites of Ru-Rh group I and Ir-Pt group I without exception belong to Ga-Ge group II, 0t' whereasthe meteorites of Ru-Rh group II and 0t ti.0 to• 20• 5 • 40' and Ir-Pt group II can belongto either Ga-Ge Plotinum, ppm group II or group IV. Since there are no me- Fig. 2. Relation between Ir and Pt contents of teorites among those listed in the table which meteorites. 466 NICYIIPORUK AND BROWN

TABLE 9. Correlationsbetween Ru-Rh, Ir-Pt, and Ga-Ge Groups [Loveringet al., 1957]

Name of Ru-Rh Ir-Pt Ga-Ge* Meteorite Type Group Group Group

Altonab Of II II IV Arispe Og I I II Bear Creek Om Anomalous III III Bendego Og I III II Bristol Off II II IV Canyon Diablo Og II II Anomalous Cape of Good Hope D, I I Not determined Coahuila H I I II Costilla Peak Om I I Not determined Coya Norte H I I-II II Edmonton Off-Of III Anomalous III Goose Lake Om II II II Henbury Om II II II Huizopa Of II II IV Indian Valley . D2gr-H I I II Moonbi Om-Of II II Anomalous Mount Tabby Og I I II Odessa Og II II Not determined Rio Loa H I I-II II Sandia Mountains Hgr II III II Sikhote-Alin Hgr-Ogg II III Not determined Spearman Om II III III Tazewell Off III Anomalous IV Toluca Om II II II

* None of the eight meteoritesbelonging to Ga-Ge groupI has beenexamined for Ru, Rh, Ir, and Pt in this study. concentrationincreases with increasingNi con- same trends. The behavior of Pd and that of centration but decreaseswith increasing Ga Ru, Rh, Ir, and Pt in relation to one another concentration.They showedthat for a given Ni and also in relation to Ni is illustrated diagra- content, the meteoritesbelonging to Ga classII matically in Figures 4 through 6. have lower Pd contents than the meteorites be- In general,meteorites of low Pd content (1.5 longing to Ga classesI and III. More recently to 3.5 ppm) tend to fall into Ru-Rh group I, Hara and Sandell [1960], in a paper on the meteoritic abundanceof Ru, have concludedinde- ! ! Ru-Rh I pendently on the basis of their own Ru deter- Ru-Rh .[[,TIT ß Ir-Pt I' minations on thirteen iron meteorites and the Ir-Pt Tr, anomalous Ga- Ge Trr, Pd values of Goldberget al. for these sameme- Of- teorites that Pd concentration generally decreaseswith increasingRu concentration. Om ...... ['a• ..... •,• Ru-Ir-PtRh I, I,Tr,rr,anomalous TIT In general, whereas hexahedritesand coarse Ga- Ge]I, TIT,anomalous octahedriteshave the highest Ru, Rh, Ir, and .....[•,,•] Ir-PtRu-Rh I,]I,I,II ]]I Explanation Pt contents and fine octahedrites have the low- Ga-GeTr, caNormal hexahedrites anomalous ß Granular hexahedrites est, the same hexahedriteshave the lowest Pd H- r-• •, Ru-Rh I, 1T a Nmkel-poorataxites content and the fine octahedriteshave the high- i iIr-PtGa- GeI,I-TT, 11 rrI ß• MediumCoarse octahedritesoctahedrites ' ' e Fine octahedrites est. A significant exception to these trends is i i t.a_.J ß Nickel-rich ataxites Cape of Good Hope. This is a meteorite of very D•- I I I I I fine structure; yet the Ru, Rh, Ir, and Pt contents appear much too high, consideringthe Nickel, per cent trends observed. The Pd content of this me- Fig. 3. Relation between structural classes,Ni teorite appears much too low consideringthese content and Ru-Rh, Ir-Pt, and Ga-Ge groups. DISTRIBUTION OF Pt AND Pd IN METEORITES 467

] ' E•plana•ion n Normal hexahedrites •o - 20 ß Granular hexahedrites 2o - Nickel-poor ataxites A Coarse octahedrites ß Medium octahedrites o Fine octahedrites E ß Nickel-rich ataxites

1.0 _• m..•\\ / Granularex,,o.o,,o.hexahedrites ,.o[ 0.5 a Nickel-poorataxites 0.5 • Coarse octahedrites & Medium octahedrites o Fine octahedrites ß Nickel-rich ataxites ß Chondrites, metal phase 0.1 1.0 I . • I 5.0 100 20.0 5.0••o la Normalt hexahedrites 'io1/.0•0.0 Nickel, 20.0 per cent Fig. 6. Variation of Ru and Rh contents of Fig. 4. Relations between Pd and Pt and between meteorites with Ni content. Ru and Pt contents of meteorites.

Ir-Pt group I, and Ga-Ge group II; meteorites Similarity o.f Canyon Diablo and Odessa. with intermediatePd contents(3.5 to 5.0 ppm) Goldberg et al. observed that three pairs of tend to fall into Ru-Rh group II, Ir-Pt groupII, meteorites,namely CanyonDiablo-Odessa, Hen-' and Ga-Ge groupII or IV; and meteoriteswith bury-Costilla Peak, and Altonah-Bristol, have high Pd contents (greater than 5.0 ppm), of compositionsand structures which are quite which only three are representedhere, tend to close to each other. They pointed out that of fall into Ga-Ge group III or IV and into these pairs only Canyon Diablo-Odessais made Ru-Rh group III or Ir-Pt group III. up of meteorites which lie close to each other geographically.The Odessacraters lie about 600 miles east-southeastof Meteor Crater, Arizona. •0 It is apparent also from this work that Canyon 20 Diablo and Odessaare very similar in composition, as shown in Table 10, where these two 10 meteoritesare listed alongwith two other pairs of very similar meteorites.

TABLE 10. Pairs of Meteorites of Similar Compositions - Name of Ni, Ru, Rh, Pd, Ir, Pt,

Normal hexahedrites Meteorite Class % ppm ppm ppm ppm ppm Granular hexahedrites N•ckel-poor ataxites Canyon Coarse octahedrites Medium octahedrites Diablo Og 7.18 6.4 1.5 3.6 1.8 7.4 Fine octahedrites Odessa Og 7.24 6.0 1.4 3.7 1.8 8.5 N•ckel-rich ataxites Goose Lake Om 8.46 6.4 1.2 3.9 2.2 6.8 5o ' • o o I 5o 4oo, o.o Henbury Om 7.66 6.0 1.2 3.9 2.3 5.6 Nickel, per cent Altonab Of 8.56 4.7 0.7 4.2 1.6 5.1 Fig. 5. Variation of Pd and Pt contentsof Bristol Of 8.15 5.1 0.8 3.9 1.7 5.0 meteorites with Ni content. 468 NICHIPORUK AND BROWN TABLE 11. Dissimilar Pt and Pd Metal Contents In this work, hexahedriteswere frequently of Hexahedrites found to be dissimilarin trace elementcontents, as can be seenfrom inspectionof Table 11. The Name of Ni, Ru, Rh, Pd, Ir, Pt, greatest of these dissimilaritiesappears to be re- Meteorite % ppm ppm ppm ppm ppm lated to the physical structures of the hexa- bedrites. The normal hexahedrites (Coahuila, Coahuila 5.56 23.7 2.3 2.0 14.8 24.2 Coya Norte, Rio Loa), which are made up of CoyaNorte 5.5 15.2 2.5 1.8 3.3 21.9 a uniform kamacite phase,have high concentra- Indian tions of Ru, Ir, and Pt; whereasthe granular Valley 5.64 21.0 2.1 1.5 7.9 29.3 hexahedrites(Sandia Mountains, Sikhote-Alin),

RioLoa 5.70 21.5 2.5 2.2 3.4 26.9 which are made up of separatecrystals of kamacite, have low concentrationsof these elements, Sandia in particular markedly low concentrationsof Ir. Mountains 5.94 7.2 1.5 2.1 •0.4 9.0 These relationshipsbetween the structuresof the Sikhote-Alin 5.68 5.0 1.6 •0.3 3.9 hexahedrites and their trace constituents can be found from the graphsin Figures 5 and 6, where all hexahedrites lie on separate branches of the It is interestingto note that the date of fall of curves representing the behavior of Pt and Pd Canyon Diablo has been estimated landers, metals as a function of Ni content. 1963] at _•2700 years ago and that of Odessa Chondrites. The five chondrites examined in at _•1400 and even at _•2900 years ago. These this study contain13 to 21% metallicphase and estimates are based on the content of the cosmic- are classifiedas belongingto the high-irongroup ray-producedAr 3"(325 years half-life) and CF6 of Urey and Craig [1953]. With the exceptionof (308,000years half-life), as reportedby various Okhansk,which has been previously classified by investigators.The C• content of Odessa,as de- Wiik [1956], these classificationsare new and termined by Goel and Kohman [1962], gives a are taken from analysesof stony meteoritesby date of fall more than 11,000 years ago. From X-ray fluorescence[Nichiporuk et al., 1965]. the chemical similarities given above and these It is evidentthat the metal phasesof the high- limiting dates of fall, it appears that the final iron-groupchondrites, unlike the highly variable proof as to whether Canyon Diablo and Odessa metal phasesof all iron meteoritesstudied, are are really fragments of the same shower of irons nearly identical in Pt and Pd metal contents and will depend upon further measurementsof the are quite similar in these contentsto the large cosmic-ray-inducedactivities. group of coarse,medium, and fine octahedrites Dissimilarities among hexahedrites. Hender- belongingto our Ru-Rh group II and Ir-Pt son [1941] noted that hexahedrites found in group II. Furthermore,with respectto Ru, the widely separated areas of the world are fre- metal phasesof the high-iron-groupchondrites quently indistinguishablefrom one another on are decidedly similar to the granular hexa- the basisof their Ni contents.Goldberg et al. healritesbelonging to Ru-Rh group II. [1951] confirmed the earlier observation on the This rather clear connectionbetween high- hexahedrites and stated further that individual iron-group chondrites and the two mutually hexahedrites are frequently identical with re- interlinked groups of iron meteoritessuggests spect to their trace constituents. that perhapsother groupsof chondrites,namely

TABLE 12. Average Concentrations of Five Pt and Pd Metals

Ru, Rh, Pd, Ir, Pt, Group ppm ppm ppm ppm ppm

High-iron-group chondrites, metal phase, 5 samples 6.0 :t: 0.5 0.98 :t: 0.10 4.1 :t: 0.3 2.8 :t: 0.4 8.5 :t: 1.2 Iron meteorites, 24 samples 9.2 •-0.8 1.47 :t:0.15 3.5 :t: 0.6 3.7 :t:0.6 11.0 :t: 1.3 DISTRIBUTION OF Pt AND Pd IN METEORITES 469

TABLE 13. Abundances of Pt and Pd Metals, Si = 106

Investigators Ru Rh Pd Ir Pt

This work, based on metal phase of high- iron-group chondrites 1.66 0.27 1.05 0.40 1.22 This work, based on iron meteorites 1.44 0.23 0.521 0.31 0.89 Hara and Sandell [1960], chondrites 1.5 Bate and Huizenga [1963], low-iron-group chondrites 1.10 high-iron-group chondrites 1.63 Schindewolfand Wahlgren [1960], low-iron-group chondrites 0.23 high-iron-group chondrites 0.33 Hamaguch, et al. (1961), chondrites 1.26 0.27 0.80 Rushbrookand Ehmann [1962], ehondrites o.38 Suessand Urey [1956], interpolated iron meteorite values of Goldschmidt[1938] 1.49 0.214 0.675* 0.821 1.625 Cameron[1959], nucleosynthesisand Suess and Urey [1956] 0.87 0.15 0.675 0.494 1.28 Claytonand Fowler [1961], nucleosynthesis 0.83 0.13 0.601 0.39 0.80

* Based on analyses of 45 iron meteorites by Goldberget al. [1951]. low-iron-groupand enstatite chondrites,are also the contribution of the metal phaseequal to the chemically connected to particular groups of actually determined proportions by weight of iron meteorites. It should be noted in this con- that phase as listed in Table 8. The abundances text that in the previous work [Lovering et al., are relative to Si taken as 106,and the calcula- 1957] the stony-iron meteorites,specifically the tions are basedon an amount of Si of 17.2% by pallasites, were found to be very clearly linked weight [Urey, 1964] for the high-iron-group in their Ga and Ge contentsto a large group of chondritesand an averageamount betweenthat iron meteorites composed mainly of medium percentage and the 18.6% [Urey, 1964] in the octahedritesand belongingto Ga-Ge group III. low-iron-group chondrites for the iron meteor- Abundancesof Pt and Pd metals. It is con- ites after normalization to mean meteoritic mat- venient at this point to calculatefrom the above ter. The results are shown in Tables 12 and 13. data the atomic abundances of Pt and Pd metals and see how they compare with the abundances Acknowledgments. We thank Elizabeth Bing- calculatedby recentworkers. The averagevalues ham and Arthur Chodos for helping us with the we have usedare shownin Table 12. All average spectrographic determinations. We are indebted to the following personswho made available samples Pd values and the averagesof the metal phaseof of a number meteorites used in this investigation: chondritesare from quite uniformly distributed E. P. Henderson of the United States National individual values and are straightforward. The Museum, C. F. Frondel of Harvard University, averagesof the iron meteoritesare the summed- A. A. Yavnel' of the Committee on Meteorites of the Soviet Academy of Sciences,and J. F. Lover- up proportionately weighted averagesof each ing of the Australian National University. The Ru-Rh and each Ir-Pt group. Entries in Table sample of Mount Tabby was obtained through 13 are in atomic abundances,and the calcula- the generosity of the Geology Museum of the Uni- tions have been made assumingin the caseof the versity of Utah. iron meteorites a contribution to mean mete- This work was supported by the National Aero- nautics and Space Administration, grant NsG 56- oritic matter of 10% [Suess and Urey, 1956; 60, and the Atomic Energy Commission,contract Ehmann, 1961] and in the caseof the chondrites AT(11-1)-208. 470 NICHIPORUK AND BROWN

t•EFERENCES rural technetium, Geochim. Cosmochim.Acta, 14, 158, 1958. Anders, E., Meteorite ages, in The Moon, Mete- Kuroda, P. K., and E. B. Sandell, Geochemistryof orites, and Comets,edited by G. P. Kuiper and molybdenum, Geochim.Cosmochim. Acta, 6, 35- B. Middlehurst, chapter 13, pp. 402-495, The 63, 1954. University of ChicagoPress, 1963. Lovering, J. F., W. Nichiporuk, A. Chodos, and Bate, G. L., and J. R. Ituizenga, Abundances of tI. Brown, The distribution of gallium, germa- ruthenium, osmium and uranium in some cosmic nium, cobalt, chromium,and copperin iron and and terrestrial sources, Geochim. Cosmochim. stony-iron meteorites in relation to nickel con- Acta, 27, 345-360, 1963. tent and structure,Geochim. Cosmochim. Acta, Brown, It., A table of relative abundancesof nu- 11, 263-278, 1957. clear species,Rev. Mod. Phys., 21, 625-634, 1949. •Nachtrieb, N. I-I., and R. E. Fryxell, The extrac- Brown, It., and E. Goldberg, The neutron pile as tion of ferric chloride by isopropyl ether, 2, J. a tool in quantitative analysis; the gallium and Am. Chem. Soc., 70, 3552-3557,1948. palladium content of iron meteorites, Science, Nichiporuk, W., and I-I. Brown, Platinum and 109(2832), 347-353, 1949. iridium abundances in meteorites, Phys. Rev. Cameron, A. G. W., A revised table of abundances Letters, 9, 245-246, 1962. of the elements,Astrophs. J., 129, 676-699, 1959. Nichiporuk, W., E. I-Ielin, A. Chodos,and Y•. Brown, Clayton, D. D., and W. A. Fowler, Abundancesof X-ray fluorescenceanalysis of stony meteorites, heavy nuclides, Ann. Phys. (N.Y.), 16, 51-68, to be published in Geochim. Cosmichim. Acta, 1961. 1965. I)odson, R. W., G. J. Forhey, and E. It. Swift, Noddack, I., and W. Noddack, Die ItS.ufigkeit der Extraction of ferric chloride from hydrochloric chemischenElemente, Naturwiss., 18, 757-764, acid solutionsby isoproplether, J. Am. Chem. 1930. Soc.,58, 2573-2577,1936. Prior, G. T., Catalogue of Meteorites, 432 pp., Ehmann, W. D., Recent improvement in our British Museum, London, 1953. knowledge of cosmic abundances,J. Chem. Rushbrook, P. R., and W. D. Ehmann, Iridium in Educ., 38, 53-57, 1961. stone meteorites by neutron activation, Geo- Goel, P.S., and T. P. Kohman, Cosmogeniccar- chim. Cosmochim.Acta, 26, 649-657, 1962. bon-14 in meteorites and terrestrial ages of Schindewolf,U., and M. Wahlgren, The rhodium, "finds" and craters,Science, 136, 875-876,1962. silver, and indium content of some chondritic Goldberg, E., A. Uchiyama, and It. Brown, The meteorites, Geochim. Cosmochim.Acta, 18, 36- distribution of nickel, cobalt, gallium, palladium 41, 1960. and gold in iron meteorites, Geochem. Cosmo- Suess,It. E., and It. C. Urey, Abundancesof the chim. Acta, 2, 1-25, 1951. elements, Rev. Mod. Phys., 28, 53-74, 1956. Goldschmidt,V. M., GeochemischeVerteiIungsge- Urey, It. C., A review of atomic abundances in setze der Elemente, $kri•ter Norske Videnskaps- chondrites and the origin of meteorites, Rev. Akad. Oslo, I, Mat.-Naturv. Kl., 148 pp., 1938. Geophys.,2, 1-34, 1964. Goldschmidt,V. M., and C1. Peters, Zur Geo- Urey, It. C., and It. Craig, The compositionof the chemicder Edelmetalle, Nachr. Ges. Wiss.,GSt- stone meteorites and the origin of the meteorites, tingen,Math.-physik. Kl., no. 4, 377-401,1932. Geochim.Cosmochim. Acta, J, 36-82, 1953. YIamaguchi,It., T. Nakai, and V. Endo (quoted Wiik, It. B., The chemical composition of some by It. Urey), in 'A review of atomic abundances stony meteorites, Geochim. Cosmochim.Acta, 9, in chondritesand the origin of meteorites,'Rev. 279-289, 1956. Geophys.,2, 1-34, 1964. Yavnel', A. A., Spectrographicanalysis of the Sik- YIara, T., and E. B. Sandell, Meteoritic abundance hote-Alin meteorite (in Russian), Meteoritika, of ruthenium, Geochim. Cosmochim.Acta, 21, 8, 134-148, 1950. 145-150, 1960. ¾avnel', A. A., Concerning the uniformity of the Henderson, E. P., Chilean hexahedrites and the chemical composition ot• the Sikhote-Alin iron composition of all hexahedrites,Am. Mineral- meteorite (in Russian), Meteoritika, II, 107- ogist, 26, 546-550, 1941. 116, 1954. YIerr, W., E. Merz, P. Eberhardt,J. Geiss,C. Land, and P. Signer, Searchfor decay productsof (Manuscript receivedJuly 20, 1964.)