The chemical composition of the silicate inclusions in the weekeroo station iron meteorite

The chemical composition of the silicate inclusions in the weekeroo station iron meteorite

EARTH AND PLANETARY SCIENCE LETTERS 8 (1970) 261-266 . NORTH-HOLLAND PUBLISHING COMPANY THE CHEMICAL COMPOSITION OF THE SILICATE INCLUSIONS IN THE WE...

563KB Sizes 7 Downloads 51 Views

EARTH AND PLANETARY SCIENCE LETTERS 8 (1970) 261-266 . NORTH-HOLLAND PUBLISHING COMPANY

THE CHEMICAL COMPOSITION OF THE SILICATE INCLUSIONS IN THE WEEKEROO STATION IRON METEORITE Eward OLSEN

Field Museum of Natural History, Chicago, Illinois 60605, USA

and Eugene JAROSEWICH

United States National Museum, Washington D. C 20506, USA

Received 2 February 1970

A chemical analysis of tlle silicate portion of the iron meteorite, Weekeroo Station, is compared with those of other meteorite types . The composition is not similar to that of a% chondrite type, however, it is similar to the residue of a bulk chondritic composition from which 5046 olivine ( s Fo) has been subtracted, with only the minor elements Cr and P showing significant anomalies. Within some silicate nodules there is textural evidence for silicate liquid immiscibility,

1 . Introduction Iron meteorites with silicate inclusions have created considerable interest recently [ 1 -13] . Their relation to other irons and to the large group 3f chondrites is not clear. Several different typ 5 ara readily distinguishable [ 1,41 , however, the rarity (A silicate material has made it impossible to obtain sufficient amounts for good quality bulk wet chemical analyses without destroying large portions of such meteorites . Thus, comparisons of the compositions of silicate portions of these irons with those of other types of meteorites have most often been estimated by making visual identifications of the various minerals, with estimated proportions of each, to arrive at some idea of the bulk coi . position [1,13,271 . Weekeroo Station, "woodbine, and Campo dei Ciélu (El Taco), three meteorites of this group, however, do have sufficient silicate inclusions to permit collection of quantities for analysis, El Taco has been described elsewhere [ I I ] and its bulk compcsition analysis will be discussed in detail elsewhere [141 . Woodbine has been described and discussed by Mason [ l ] . The

mineralogy, textures, and aspects of the isotope chemistry and isotopic dating ofWeekeroo Station have been discussed in other papers [2,4,5,8,9] and these data will not be repeated here. 2. Sampling for analysis A slab of Weekeroo Station, 48 X 17 cm, 6445 g, from the Field Museum collections (cat. no. Me 2118) was sampled . Of over one hundred silicate inclusions twelve were selected at random~and mined out by using hardened steel chisels and picks, small high speed drills, and various dental tools. Petrographic examination [2] indicated that each nodule would have to be completely removed in order to eliminate a serious sampling bias becaustr the pyFoxene crystal's are usually concentrated in the interiors of the nodules and coarse chromite grains are usually clustered near the interface with enclosing metal . Thus, each nodule was chiseled clear to the walls of metal, which were also scaled . The resulting sample consisted of granules and powder which was cleaned with a hand magnet to remove any

E.OLSEN and EJAROSEWICH

262

Table 1 Analysis of silicate inclusions of the Weekeroo Station iron meteorite (E.Jarosewich, analyst). Si02 Ti02 A1203 Cr203 Fe 2 03 FeO

54 .54 0.29 5.22 2.76 0.93* 10.67

MnO Mg0 CaO Na20 K20 P205

0.59 16.24 5.16 1 .80 0.20 0.10

1120+ H20Ni C S Sum

ND 0.12 0.05 0.91;** less than 0.1 99 .65

Observed phases : Orthopyroxene, Ca-pyroxene, plagioclase, K-feldspar, tridymite, chromite, rutile, ilmenite, whitlockite, glass, graphite . Mode : Orthopyroxene K-feldspar Whitlockite

54 .1% 1 .2% 0.217,

Plagioclase (An 20) Silica Ilmenite

22 .0% 3.5% 0.4%

Ca-pyroxene Chromite Graphite

15 .6% 3.1% 0.1%**

* Due to terrestrial oxidation. ** Some of the carbon may be due to contamination introduced by oils used in the uriginal cutting of the meteorite. metal particles. Very little material was picked up by the magnet and this consisted of flakes and curls of metal which were cut out of the walls in the scaling process. ;Slightly over two grams were collected. This sample was overall brown in color indicating some partial terrestrial oxidation (Weekeroo Station is a t,nd). 3 . Analytical procedures The major elements : Si, Mg, Fe, Ca, Al and Ti, were determined by the classical method of Hildebrand et al . [241 and Peck [251 . Silica was separated from hydrochloric acid solution by double dehydration and the residual silica retained in the R20 ., group was rec_wered by dehydration frorn the sulfuric acid solution of this group . One aliquot of the solution containing the R203 group was used for the determination of iron . From a second aliquot iron was removed by mercury-cathode, one portion of this ironfree solution being used to determine titanium colorimetrically (using Tiron), and the remaining portion used to determine aluminum by precipitation with 8hydroxyquinoline . Calcium was precipitated with oxalic acid and magnesium with dibasic ammonium phosphate. Nickel, sulfur, manganese, chromium and

phosphorus were determined, from a separate sample which was dissolved in Br water and HN03 and then treated witli HF and HN03 io remove silica . Nickel was determined colorimetrically witli dimethylglyoxime and sulfur witli BaCl 2 . Manganese was determined colmimetrically as perrnarrganate, chv .mium witli dipherylcarbazide, and phosphorus with molybdovanadate . Sodium and potassium were determined by flame photometer after solution of the sample it HF and H2SO4 . The overall precision of the analysis is the same as that discussed by Hildebrand et al . [241 and Peck 1251 . 4. Discussion of analysis The bu :A composition of the silicate fraction of Weekeroo Station bears only a small resemblance to that of tlrc ordinary chondrites, which have been recomputed to a metal-free and sulfide-free basis for comparison (table 2) . Similarly enstatite chondrites and carbonaceous chondrites show little resemblance to it . The fact that potassic alkali feldspars are known from Weekeroo Station [21 might suggest that both K20 and Na 2 0 would be highly enriched relative to the ordinary chondrites . While Na,0 shows some enri-,.Iirnent 1 .6 to 2 times relative to he ordinary chondrite,,, i~'-20 is quite normal .

CHEMICAL COMPOSITION OF SILICATE INCLUSIONS

26 3

Table 2 Meteorite analyses recalculated to a metal-free and sulfide-free basis . H

b X C

Si02 TiO2 A1 2 03 Cr203 rcO* MnO Mg0 CaO Na20 K20 P2os C

b r á. E ti O v B . .., u

51 .41 0.16 2.65 0.31 5.05 0.40 36 .77 1.92 0.91 0.08 0.28 -

W O O 3 N

55 .20 0.13 3.42 0.19 3 .16 0.19 33 .50 2.19 1 .87 0.13 --

O O as aJ y

O

.+ ei ;1

54 .54 0.29 5 .22 2.76 11 .51 0.59 16 .24 5.16 1.80 0.20 0.10 0.98

10 11 3 O = N N 4 ...r

49 .96 0.11 8.75 0.42 15 .69 0.78 16.17 6.56 0.95 0.28 0.07 -

O v

c.

W Ca .

O .~

r.

M _

U Ó N O N N h ,O .L: ci 39 .23 0.13 3.01 0.48 17 .54

0.32 26 .45 2.57 1 .29 0.12 0.46 6.41

60 .41

ç,

á. s

~+ J

. o

C,

1

O O O ti'n

J4>

y 1. 1. r_ 3 r_ _ 0 Q r7 0 ; :7 0 O N 0 . S... N . 1; C-4 O - .C N ~" U - 00 U ~- . 0% U - -

+

"

C O

.+ ~

1.

cn 00

O O O t N + y y C O O .- ; t.~p rO

48.64

47 .29

44 .65

47.27

48 .60

0 .20

0.17

0 .18

0.21

0.14

0 .14

3 .29

3 .24

2 .74

2.48

2.61

2 .61

0 .59

0.48

0 .52

0 .34

1 .38

1 .38

0.00

11 .82

15 .56

19 .84

12 .94

5 .76

0.39

0 .33

0 .32

0 .36

0 .30

0 .30

30.20

31 .27

29 .62

28 .47

30 .94

36 .74

2 .41

2 .42

2 .25

2 .20

2 .58

2 .58

1 .57

1 .13

1 .04

0 .98

0 .90

0 .90

0.22

0 .19

0.17

0 .24

0 .10

0 .10

0 .31

0.31

0 .25

0 .05

0.05

0 .49

0.49

0 .72

-

-

-

* Total ferrous and ferric calculated as ferrous.

A large number of analyses of many meteorite types were compared with this analysis . Campo del Cielo and Woodbine (table 2) diir in major aspects from Weekeroo Station . Because they belong to a different class of irons with silicate inclusions they have a significantly different mineralogy . Weekeroo Station, foe example, is olivine-free and has free silica (as tridymite) as opposed to Campo del Cielo and Woodbine . Both of these are moderately close to H-group cliondrites and enstatite cliondrites . In this respect they are close to the H-group chondrite, Rose City, which exhibits the greatest degree of reduction and has a composition somewhat intermediate between the H-group and the enstatite chondrites . The bulk analysis of Weekeroo Station compares moderately well in some respects with the polymictorthonvroxnne-nivrnnite-ni ±çP achondritez t .p ._ .- . .  t ,,, r . o - .. . . r -,toincl _ ; or howardites . For example, a plot of FeO/FeO+MgO versus CaO (28 M %, and 5.1-6 st . `/~-, respectively) of the kind used by Mason [ 151 falls close to the cluster of points representing the howardites (also see Duke 1161 ) . The mineralogy is, however, not close to that of the howardites, which have basic plagioclase (90 An) rather than the very acid alkali plagioclase found in Weekeroo Station (I I An). The approximately 40%

lower A1203 content of Weekeroo Station coupled with the only 20% lower CaO content reflects the greater abundance of CaO in pyroxene in Weekeroo Station relative to CaO in plagioclase in the howardites . The same plot for the silicate portions of Campo del Cielo and Woodbine do not fall anywhere near the point for Weekeroo Station. The superficial resemblance to the bulk composition of the howardites has no significance and a genetic discussion based upon the system anorthite-forsteritesilica, as discussed by Mason for the acliondrites 115, fig. 431 cannot be applied here . If, however, the orthopyroxene, Ca-pyroxene, and tridymite are computed from the analysis and plotted in the system forsterite-diopside-silica [ 171 it falls near the boundary between silica and py roxenes. Such a plot ignores the iron content of the pyroxene and the presence of the feldspars, which comprise major phases in the ruck . It is noteworthy, nevertheless, that this composition does not plot in some absolutely arbitrary position but near the boundary where a residual liquid might plot after the fractional crystallization of olivine from an initial bulk composition which is a good deal less siliceous. This leads one to consider the possibility that

26 4

E.OLSEN and E .JAROSEWICH

olivine may have been fractionally removed from a chondritec composition and that Weekeroa Station might be a residue silicate fraction. The removed olivine fraction would, within a matrix. of metal appear as a pallasite . The olivine in the average pallasite is about 85Fo in composition 1281 . Most chondrites contain from 25-70% olivine . by adding an average amount, 50°ío, of 85Fo -:o the Weokeroo Station analysis we obtaw tile analysis given in table 2 (no . 10) . The result is extremely close to an average chondrite for every major element and most of the minor elements except Cr203, which is three times too high, and P205, which is six times too low . The latter difference could be a result of sampling error. The question arises, however, that if olivine were fractionally removed why the chromite was not also? Certainly in bodies such as the Stillwater and Bushveldit Complexes olivine and chromite separate together into the ultramafic portions [23j . Well crystallized chromite occurs within most silicate nodules in Weekeroo Statiun. A small part of the bulk Cr203 content, however, is attributable to the Ca-pyroxene, which usually contains almost 1% by weight (in separate microprobe analyses [4] ). Another, more serious objection to such a scheme, of course, is that of the implied complementary relation between Weekeroo Station-type irons and the pallasites. Very clearly the implied proportion, 1 : 1, among known meteorites is off by one order of magnitude, it being about 1 :10 . Prior [ 18] suggested a similar complementary relation between eucrites and pallasites. In this case the suggested proportion of eucrite to pallasite, 1 :5 [ 19] , is not observed among known meteorites either, it being closer to 1 :2. In both cases. eucrite-pallasite and Weekeroo Stationtype-pallasite, the so-called "sociological factor" relating tile bias towards high metal meteorites versus scones (especia!ly achondrites) among finds is such that the implied proportion should be favored, rather than tile reverse, which are tile cases here. On the other hand, the close similarity between column 10, table 2, and tile average chondrite sug&::sts that some straightforward relation to chondrites must exist . Because Weekeroo Station-type irons are so rare (three known in this category, with Elga possibly a fourth) it is possible that tile olivine-rich complementary fraction is also rare . An olivine-rich stone, such as a chassi.gnyite, of ,omewhat lower oxidation

state, would fit the requiremen s both in mineralogy and ratity . A relation to chondrites is also indicated if the FeO/FeO+MgO percentage (2811,ó) is plotted against the Ni content (6.9%) of the metal, in a plot illustrating one of Prior's rules (cf. fig. :!, Ringwood [ 19] ). The point plots just slightly outside the cluster of points representing the trend among the chondrites. By adding olivine (column 10, table 2) the point moves into the middle of the chondrite; cluster of points and onto the trend curve .. Mason [ 1 ] , in discussing the Woodbine meteorite (which is not a Weekeroo Station type) suggested that "there is some indication that Prior's rules, est<
CHEMICAL. COMPOSITION OF SILICATE INCLUSIONS

the same physical feature of the two glasses "each embedded in and containing globules of the other" . Immiscibility in certain iron-rich silicate systems has been investigated [201 and Holgate 1211 has discussed the question thoroughly. Direct terrestrial field evidence for a truly unmixed silicate system is lacking because most terrestrial systems are too alumina and alkali rich to have bulk compositions which lie anywhere near those required for immiscible liquids to form. The Weekeroo Station analysis is, however, low in alkalies and alumina relative to crustal igneous systems. In fact, a plot of the Weekeroo Station analysis on the trilinear diagram used by Holgate ( [21 J, fig. 1) falls outside the field for terrestrial rocks and on the side of his empirical boundary, A--A', towards the immiscilbe region . Because most nodules in Weekeroo Station do not show silicate-silicate immiscibility we conclude that only in some cases did small masses of silicate melt, of somewhat different bulk composition, become separated within the metal mass and these were sufficiently different that they formed small immscible systems . In the larger immiscible system, silicate-metal, it seems reasonable that the low silicate to metal ratio could create such local inhomogeneities with the silicate portion . With regard to the silicate-silicate immiscibility features in some of these nodules, the question of time arises . Bowen's experiments [261 on interdiffusion of plagioclase and diopside melts indicate extreme sluggishness. For effects over short distances, however, such as the case here, the slow rates may not be a serious limiting factor for unmixing to occur. It is hoped that a sufficient number of moderately large nodules were sampled so that the analysis (table 1) provides a good average of the silicate system, averaging out any local inhomogeneities and providing one of the three available silicate analyses of this increasingly significant meteorite group, the irons with silicate inclusions. Acknow iedgments We acv grateful to Dr. Robert F .Mueller (NASA, Greenbelt, Maryland) and Dr. Robert Newton (University of Chicago) for discussions concerning the Weekeroo Station meteorite ; and to Dr. Arden Albee (California Institute of Technology) for critiCdlly

26 5

reading the manuscript . One of us (E.O.) was supported by grants from the National Science Foundation, and this support is gratefully acknowledged . References B.Mason, The Woodbine meteorite, with notes on silicates in iron meteorites, Mineral . Mag. 36 (1967) 120. T.E.Bunch and E.Olsen, Potassium feldspar in Weekeroo Station, Kodaikanal, and Colomera iron meteorites, Science 160 (1968) 1223 . G.J .Wasserburg, H.G .Sanz and A .E .Bence, Potassium feldspar phenocrysts in the surface of Colomera, an iron meteorite, Science 161 (1968) 684 . [41 T.E .Bunch, K.Keil and E.Olsen, Mineralogy and petrology of silicate inclusions in iron meteorites, Contrib. Mineral. and Petrol . (1970) in press. G.J .Wasserburg, D .S .Burnett and C.Frondei, Strontiumrubidium age of an iron meteorite, Science 150 (1965) 1814 . [61 D .S .Burnett and G .J .Wasserburg, Evidence for the formation of an iron meteorite at 3.8 X 109 years, Earth. Planet . Sci. Letters 2 (1967) 137. 171 I .Olsen and R.F .Mueller, Silicates in some iron meteorites, Nature 201 (1964) 596. 181 D.Bogard, D.Bumett, P.Eberhardt and G.J .Wasc^Tburg, 40 Ar-a° K ages of silicate inclusions in iron meteorites, Earth Planet. Sci. Letters (1968) . D.S.Burnett and G.J .Wasserburg, 87 Rb-87 Sr ages of silicate inclusions in iron meteorites, Earth Planet . Sci. Letters 2 (1967) 397. [101 R.Marshall and K.Keil, Polymineralic inclusions in the Odessa iron meteorite, lcaru,, 4 (1965) 461 . 1111 F.R .Park, T.E .Bunch and T.B tassalski, A study of the silicate inclusions and óther phases in the Campo del Cielo meteorite, Geochim. Cosmochim. Acta 30 (196S) 399. 1121 E.Olsen, Amphibole: first occurrence in a meteorite, Science 156 (1967) 61 . [ 131 E.Olsen and L.Fuchs, Krinovite, NaMg2CrSi301o: a new meteorite mineral, Science 161 (1968) 786. 114 1 F.Wlot7ka, in preparation . [ 151 B.Mason, Meteorites (John Wiley, New York, 1962). 1161 M.B .Duke, Petrology of eucrites, howardites, and mesosiderites, Geochim. Cosmochim . Acta 31 (1967) 1637 . . , . vüy~üio anu .1 Jt . O4110i1 L . ..'-. .., ,~t..... t., on the system ~.w data 11/.~ l.nF.J .aïa New diopside-forscerite-silica, Carnegie Inst ., Geophysical Lab. Yearbook 62 (1963) 95 . [ :81 G.T .Prior, on the mesosiderite-grahamite group of meteorites, Mineral. Mag. 18 (1918) 151 . 1191 A.E .Ringwood, Chemical and genetic relationships among meteorites, Geochim . Cosmochim. Acta 24 (1961) 159. 1201 E.Roeddcr, Low temperature liquid immiscilibity in the system K 20-FeO-AI203-SiO2, Am . Mineral. 36 (1951) 282.

26 6

E.OLSEN and RJAROSEWICH

1211 N.Holgate, The role of liquid immiscibility in igneous petrogenesis, J. Geol . 62 (1954) 439. 1221 K.Ked, Meteorite composition, in : Handbook of geochemistry, Vol. I (Springer Verlag, Berlin, 1969) p . 78 . 1231 E .D .Jackson . Primary textrues and mineral associations in the ultramafic zzone of the Stillwater Complex, Montana, U .S. Geol . Surv. Prof . Pape r 355 (1961) . 1241 W.F .Hildebrand, G.Lundell, H.Bright and J .Hoffmann, applied inorganic analysis (2nd pd.) (John Wiley, New York).

1251 L.C .Peck. Systematic analysis of silicates, Bull . U .S. Gcol . Surv . 1170 (1964) . 1261 N.L.Bowen, Diffusion in silicate meus, J . Geol . 29 (1921) 295. 1271 T.Hodge-Smith, The Weekeroo Station meteorite: a siderite from South Australia, Rec. Austral. Mus. 18 (1930) 312. 1281 P.buseck and J.Goldstein, Olivine compositions and cooling rates of pallasite meteorites, Bull. Geol . Soc. Am . 80 (1969) 2141 .