The Ordovician chondrite from Brunflo, central Sweden, II. Secondary minerals

The Ordovician chondrite from Brunflo, central Sweden, II. Secondary minerals

Lithos, 27 (1991) 167-185 Elsevier Science Publishers B.V., Amsterdam 167 The Ordovician chondrite from Brunflo, central Sweden, II. Secondary miner...

2MB Sizes 0 Downloads 0 Views

Lithos, 27 (1991) 167-185 Elsevier Science Publishers B.V., Amsterdam


The Ordovician chondrite from Brunflo, central Sweden, II. Secondary minerals J a n O l o v N y s t r 6 m a a n d F r a n s E. W i c k m a n b aSwedish Museum of Natural History, S-10405 Stockholm, Sweden bDepartment of Geology and Geochemistry, Stockholm University, S-10691 Stockholm, Sweden (Received March 28, 1991; revised and accepted August 30, 1991 )


ABSTRACT Nystr6m, J.O. and Wickman, F.E., 1991. The Ordovician chondrite from Brunflo, central Sweden, II. Secondary minerals. Lithos, 27:167-185. Brunflo is an altered H4-H 5 stony meteorite that fell in calcareous mud on the continental shelf of the Iapetus Ocean about 460-470 Ma B.P. Except for relicts of chromite all the original minerals are replaced by calcite, barite, an illite-type mineral and some minor to rare phases: carbonate-fluorapatite, quartz, titanium dioxide, chalcopyrite, three different members of the cobaltite-gersdorffite series (including a phase with a composition close to CoNi (ASS)2 ), rammelsbergite, safflorite, nickeline, maucherite, orcelite, native nickel, "anomalous bornite", galena, pyrite, sphalerite and two Cu-phases (Cu 1.66S and Cu4FeS4 ). The secondary mineralogy varies rapidly at a ram-scale, indicating non-equilibrium conditions during the alteration. By the time the mud was cemented to limestone most of the original minerals might have been transformed to secondary phases. During subsequent stages of diagenesis the temperature remained within the 0-50°C range for millions of years, probably explaining the formation of the phases of unusual compositions. A temperature rise to 200-300 ° C chiefly due to regional heating during the Caledonian orogeny in Silurian-Devonian time might have modified the secondary mineralogy.


Samples and methods

Brunflo, an altered stony meteorite discovered by Thorslund and W i c k m a n ( 1 9 8 1 ) in a slab o f Ordovician limestone from central Sweden, is the first reported fossil chondrite. The present minerals in Brunflo differ greatly from those found in newly fallen chondrites. Except for relicts o f p r i m a r y chromite and scarce c h r o m e spinel (Thorslund et al., 1984 ), they are all secondary, ranging from oxygen-bearing minerals to a native element ( N i ) and a n u m b e r o f arsenides, sulfar-sedines and sulfides, a few rare or up to now unreported from studies o f natural and synthetic systems. Here, we describe these minerals and discuss their formation. A preliminary report was published by W i c k m a n and N y s t r 6 m (1986).

A general description o f Brunflo, tentatively classified as a H 4 - H 5 chondrite, is given in Thorslund et al. (1984). The present study is based on 17 polished thin sections prepared from five drill cores with a diameter o f 7.5 m m in Brunflo. One o f the cores (no. 4; 17 m m long) goes through its outer part and penetrates 3 m m into the limestone. The others were shorter ( 4 - 1 2 m m ), since m u c h o f the meteorite was r e m o v e d m a n y decades ago by chiseling from the back o f the slab and,these chips subsequently have been lost (Thorslund and Wickman, 1981 ). The compositions o f the secondary minerals were determined with electron microprobe, using A R L S E M Q instruments. Most o f the arsenides, sulfarsenides and sulfides were analyzed at the Geologi-

0024-4937/91/$03.50 © 1991 Elsevier Science Publishers B.V. All rights reserved.


cal Survey of Sweden ( S G U ) , Uppsala, by C. Alinder. The acceleration voltage was 20 kV, the sample current 16 nA on brass and the counting time 20 s on peak and background. Natural sulfides and pure metals (Co, Ni, Bi) were used as standards and the results were corrected with the Magic IV program. Some As- and S-bearing minerals were also analyzed in a reconnaissance study at the Metal Research Institute (IMF), Stockholm, employing synthetic standards, and Stanton troilite for Fe and S; the procedure is described in Thorslund et al. (1984). Apatite, silicates, barite and calcite were analyzed in Uppsala ( S G U ) and at the Nordic Volcanological Institute (NVI), Iceland, the conditions being similar to those given above except for a lower accelerating voltage (15 kV) and sample current ( 10 nA). The analyses given in the tables were made in Uppsala, if nothing else is stated. In the tables each analysis referring to Brunflo has a label consisting of three parts: an initial number for the drill core, a letter for the polished section and a second number indicating the analyzed point in the section. Total iron is reported either as metal or as FeO. The following abbreviations are used: bd = below the detection limit and n d = not determined. The small grain size of the arsenides, sulfarsenides and many sulfides is reflected in the often low totals obtained; the phases in their immediate surroundings (calcite, barite, and less often layer silicate and apatite) account for the missing elements in the analyses. Analyses with low (sometimes high) totals are given for small grains in single-mineral environments, since we assume that the reader is more interested in the minerals of the unique meteorite than in analyses in the range 98101 wt.%. The precise crystallographic nature of some minerals is uncertain. No single crystal X-ray work has been performed due to the small grain size or quantity (or both) of the material.

Secondary minerals The chondrite structure of Brunflo is surprisingly well-preserved locally, in spite of the almost total replacement of its original minerals. Pseudomorphs of the various chondrule types observed in Brunflo are illustrated in Thorslund et al. (1984). The secondary mineralogy (Table 1 ) is highly variable over small distances and it appears to be unrelated to the


TABLE 1 Secondary minerals in Brunflo, Calcite, barite and layer silicate a r e the predominant minerals, embedding the minor constituents which range from ubiquitous (titanium dioxide) to sporadic (see text) Minor minerals

Native nickel Cobaltite-gersdorffite Gersdorffite Rammelsbergite/ Fe-gersdorffite Nickeline*2/safflorite Orcelite Chalcopyrite 'Bornite' 'Chalcocite' Pyrite Galena Sphalerite Titanium dioxide Quartz .3




+ (+ )



+ + +


Layer silicate*~



cg cp, ni cp, ni + cg cp

(+ )

The following abbreviations are used: cg=cobaltite-gersdorffite, cp = chalcopyrite, ni = nickeline (or maucherite ). The exact nature of the "chalcocite" and "bornite" is not known. *~lncludes apatite which sometimes is mixed with the layer silicate (the latter occurs in two generations; the younger is vein-forming and not referred to here). *2Or maucherite. *3Forms a discontinuous fringe between the layer silicate and barite.

primary structure. Nearby chondrules (the term is used for Brunflo solely in its structural sense) might, for example, be composed of calcite, layer silicate, barite or apatite (Fig. 1 ). Calcite is the predominant constituent of Brunflo, except in its structurally less preserved central part, where barite is common (Fig. 2). Small titanium dioxide grains and chromite relicts are ubiquitous. Layer silicate (two generations), apatite, cobaltite-gersdorffite and chalcopyrite are present in widely varying amounts. Several other Co-Ni arsenides, quartz, galena, secondary copper minerals, sphalerite, native nickel and pyrite occur as very subordinate or sporadic constituents. Besides the rare pyrite grains, none of the other iron minerals common in terrestrial rocks (magnetite, hematite and pyrrhotite) have been observed in Brunflo.

Native nickel A 0.04 X 0.01 mm large grain of native nickel with smooth outline has been found in one thin section and an additional one, too small for accurate anal-



Fig. I. Distribution of secondary minerals in Brunflo (a part of section 2A). The outline of former chondrules is indicated with a thick line. The secondary minerals are in many places difficult to identify due to presence of more than one phase, especially the domains with layer silicate and apatite. The legend refers therefore to the dominant mineral present. The proportion of secondary minerals depicted here is not typical; calcite is usually much more abundant and layer silicate, apatite and cobaltitegersdorfl]te more subordinate. ysis, has been identified with microprobe in another section. Both grains are e m b e d d e d in barite. They are white in reflected light, and optically homogeneous. The nickel is remarkably low in iron (Table 2). Influence from the surrounding barite probably accounts for the sulfur in the analyses. Small a m o u n t s o f iron, nickel and cobalt occurring as native metals or alloyed with each other are

well-known from serpentinites all over the world; particularly awaruite (Ni,Fe) is widespread. The first description o f native nickel appears to be by R a m d o h r ( 1 9 6 7 ) and refers to a sample from New Caledonia. Since then, nickel has been reported from several localities. It is generally agreed that nickel and related metals are secondary phases f o r m e d under reducing conditions during the low-tempera-



ture serpentinization process (Ramdohr, 1967; Kanehira et al., 1975; Frost, 1985 ).

Cobalt-nickel arsenides

2 cm

Patches and veins of barite




to complete recrystallization

Partial recrystallization

'Atom-for-atom' replacement Transition zone

Fig. 2. Simplified division of the whole Brunflo stone into areas of different type (see text for further explanation).

Eight different C o - N i phases (three sulfarsenides and five arsenides) collectively referred to as arsenides have been found in Brunflo. The by far most abundant of them is an intermediate member of the cobaltite-gersdorffite series occurring as spongy aggregates. Rammelsbergite rimmed by Ferich gersdorffite, and nickeline and maucherite associated with safflorite are present as tiny grains; a single grain of orcelite has also been observed. The names rammelsbergite and saffiorite are used in a broad sense, because it has not been possible to distinguish these minerals from pararammelsbergite and clinosaffiorite, respectively, due to the small grain size. A third compositional type of gersdorfrite forms patches in chalcopyrite grains. The overall proportions of Co and Ni in the Brunflo arsenides are approximately equal. This is rather astonishing if it is taken into account how dominant nickel must have been in the original meteorite. The spongy aggregates of cobaltite-gersdorffite are common in structurally well-preserved parts of Brunflo now replaced by calcite. They occur preferentially in the matrix and marginal parts ofchondrules (the arsenide is seen as white irregular patches in Thorslund et al., 1984, Figs. 4 and 6-13 ), with a clear tendency to form rims around the chon-

TABLE 2 Microprobe analyses of native nickel in Brunflo 2X-2


Wt.% Fe Co

Ni Cu S As

Total At. prop. Fe Co

Ni Cu S As

0.16 0.19 99.14 0.21 0.25 0.07

0.15 0.1 l 99.67 0.26 0.24 0.08



0.17 0.19 98.94 0.19 0.46 0.05

0.15 0.11 98.99 0.24 0.44 0.07

Fig. 3. Rims of cobaltite-gersdorffite aggregates (white) around former chondrules in Brunflo, replaced by calcite (medium gray) and layer silicate (dark gray). Reflected light.


Fig. 4. SEM photograph of a cobaltite-gersdorffite aggregate embedded in calcite (black) in Brunflo. The medium gray areas within the aggregate consist of chalcopyrite (the white spot is a charge effect ) and the light gray grains at both sides of the aggregate are barite.

drules (Fig. 3). The largest aggregates attain 0.5-1 mm in cross-section; their usual size range is 0.050.2 ram. They consist of innumerable white particles in nebulous to dense configurations (Figs. 45). The nebulous aggregates resemble clusters of empty shells because the arsenide is concentrated in the outer part of individual bodies. Many arsenide particles are annular with diameters around 2/~m and they resemble the atoll-textured cobaltite-gersdorffite described by Vokes and Strand (1982). The dense aggregates have a grate-like structure in thin


section due to coalescence of the arsenide particles into parallel sheet-like domains (Fig. 6 ). Large aggregates are usually composite, as revealed by differences in the orientation of the domains (Fig. 5 ). Many of the sheet-like arsenide domains have a core made up of or rich in chalcopyrite (Fig. 6 ). Galena is locally present in the aggregates. Representative analyses of spongy aggregates are presented in Table 3. The cobaltite-gersdorffite is quite uniform in composition with a Ni:Co ratio around 1 : 1 (Fig. 7). The scatter in the diagram becomes even smaller if the points for thin section 1A are excluded, because this section was analyzed under non-optimized conditions in a reconnaissance study. The remaining 20 analyses yield a range of 45-52 tool % NiAsS (average 48 mol %). The 1:1 proportion of the metals is probably not accidental. We speculate that the cobaltite-gersdorffite aggregates represent an intermediate compound, CoNi (AsS) 2. X-ray diffractograms of polished sections with abundant arsenide demonstrate that the cobaltite-gersdorffite is a single phase, not a mixture of Co- and Ni-rich members. However, the degree of ordering in the structure cannot be determined due to the nature of the material and the minute grain size. No intermediate compound seems to have been described from natural or synthetic assemblages. Phases rich in both Co and Ni are reported (Klemm

Fig. 5. SEM photograph of a cobaltite-gersdorffite aggregate. The domains in the central-left part are nebulous. A part of a chromite grain is seen in the lower left corner.



Fig. 6. Well-developed grate-like structure in a cobaltite-gersdorffite aggregate in Brunflo (SEM photograph ). Many of the parallel arsenide d o m a i n s have cores of chalcopyrite ( m e d i u m gray). Barite (white) partially fills a fracture in the center o f the photograph. TABLE 3 Representative microprobe analyses of spongy cobaltite-gersdorffite aggregates in Brunfio 4A-7

4A- 10

4A- 16

Wt.% Fe Co Ni Cu S As

0.73 16.90 16.71 0.50 19.15 42.28

0.81 18.00 15.17 0.22 19.16 42.28

0.78 16.29 17.36 0.70 19.12 44.19





At. prop. Ee Co Ni S As

0.01 0.50 0.50 1.01 0.99

0.02 0.53 0.45 1.02 0.98

0.01 0.47 0.51 0.99 1.01





Atomic proportions (corrected for influence from chalcopyrite) based on S+As=2: M= Fe+ Co+Ni. The contents ofSb, Bi, Pb and Zn are below the detection limits.

and Weiser, 1965; Rosner, 1970; Petruk et al., 1971; Vokes and Strand, 1982), but they contain considerable iron and they do not show a restricted compositional range suggestive of an intermediate compound. The gersdorffite patches in chalcopyrite constitute a second cobaltite-gersdorffite series mineral in

close association with chalcopyrite in Brunflo, but the two arsenides differ texturally and chemically. The chalcopyrite-gersdorffite association occurs as relatively large grains with sharp and irregular mutual contacts (Fig. 8). The gersdorffite patches, which are found only occasionally in the chalcopyrire (or as individual grains near the chalcopyrite) have sizes up to 0.03 ram. The spongy arsenidechalcopyrite aggregates, on the other hand, are made up of intimately associated particles. Chemically, the gersdorffite patches are richer in Ni and As than the spongy arsenide (Tables 3, 4). The rammelsbergite-gersdorffite and nickeline (-maucherite)-safflorite grains occur sparsely in Brunflo, single or in swarms. Their morphology varies with the host mineral. Grains embedded in calcite are often polygonal to euhedral or skeletal (Fig. 9 ), in contrast to the arsenides in barite which tend to be anhedral or have a corroded appearance. The typical size range of the arsenide grains is from a few to 30/Lm. Barite-hosted grains are generally larger and less concentrated in swarms than those in calcite. The rammelsbergite is invariably rimmed by Fe-gersdorffite, whereas only some nickeline grains are partly bordered by saffiorite; others are mosaics (Fig. 9) of two or possibly three phases, because slight color variations indicate that both nickeline and maucherite may occur together with the saffiorite. Some seemingly corroded nickeline




~ ~ Section I • Section 4A 'spongy" aggregates x Section 16A • Rimaroundrommelsbergite in.chalcopyrite ~~ O



• CoAsS




v• 80

\ NiAsS

Fig. 7. Compositional variation for the m e m b e r s of the cobaltite-gersdorffite-arsenopyrite system analyzed in Brunflo.

TABLE 4 Microprobe analyses of gersdorffite occurring as patches in chalcopyrire and rims around rammelsbergite, in Brunflo Patches in chalcopyrite

Rims around rammelsbergite





Wt.% Fe Co Ni Cu S As

1.20 9.86 22.28 0.47 18.24 45.27

0.76 5.81 25.22 0.06 16.79 45.86

8.45 0.28 24.08 0.14 16.15 48.61

4.29 1.32 26.19 0.17 12.86 55.38






At. prop. Fe Co Ni S As

0.02 0.29 0.66 0.96 1.04

0.02 0.17 0.76 0.92 1.08

0.26 0.01 0.71 0.87 1.13

0.13 0.04 0.79 0.70 1.30






Fig. 8. Sketch of chalcopyrite with patches of gersdorffite (stippled or indicated with g) and 'bornite' (black) in Brunflo.


c:~' 0.01 mm

Atomic proportions (excluding Cu) based on S+As=2; M= Fe + Co + Ni; the proportions of the first analysis are corrected for influence from surrounding chalcopyrite. The contents of Sb, Bi and Pb are below the detection limits. *~Same grain as 3B-7 in Table 5. *2Same grain as 3B-9 in Table 5.

(or maucherite) grains contain small patches of Cu sulfides (chalcopyrite, "bornite" and "chalcocite"). For simplicity, we refer only to nickeline in the following for the pinkish white to pink phases associated with safflorite and Cu sulfides. Chemical analyses of the arsenide grains are given in Tables 4-6 and metal ratios are plotted in Figs. 7 and 10. All the Brunflo arsenides with the exception of the rim-forming gersdorffite are low in Fe. The

Fig. 9. Sketch of nickeline-safflorite grains in Brunflo [ nickeline (and maucherite ) = white; safflorite = black ].







~ eAs2

l oo.o


Rammelsbergite J



Rammelsbergite, Bou A z z e r








Fig. 10. C o m p o s i t i o n a l v a r i a t i o n for t h e m e m b e r s o f t h e s a f f i o r i t e - r a m m e l s b e r g i t e (-1611ingite) s y s t e m a n a l y z e d in Brunflo. TABLE5


Representative microprobe analyses of rammelsbergite and saffiorite in Brunflo

Representative microprobe analyses of nickeline and maucherite in Brunflo


Saffiorite Bou Azzer.3

3B-5 .4


3B-7 *~

3B-9 .2


Wt.% Fe Co Ni S As Sb

0.20 8.0t 20.01 0.90 71.75 0.32

0.18 4.98 23.39 0.67 73.11 0.21

0.3 12.5 17.5 0.6 69.2 nd

1.04 21.86 5.87 0.56 71.76 0.15

0.00 23.86 4.69 0.21 72.29 0.32







At. prop. Fe Co Ni S -ks Sb

0.01 0.28 0.69 0.06 1.94 0.01

0.01 0.17 0.80 0.04 1.96 0.00

0.01 0.45 0.63 0.04 1.96 -

0.04 0.76 0.20 0.04 1.96 0.00

0.00 0.83 0.16 0.01 1.98 0.01










3B-2 *~

3B-I 1

Wt.% Fe Co Ni Cu S As

0.21 3.76 37.65 0.23 0.59 54.73

0.06 2.33 39.99 0.13 0.57 55.23

0.02 4.27 47.35 0.11 0.96 47.17

0.00 9.39 41.11 0.17 1.29 47.06






At. prop. Fe Co Ni Cu S As

0.03 0.51 5.14 0.03 0.15 5.85

0.01 0.31 5.41 0.02 0.14 5.86

0.00 0.88 9.78 0.02 0.36 7.64

0.00 1.91 8.38 0.03 0.48 7.52






Atomic proportions based on S + A s + S b = 2 ; M = F e + C o + N i . The contenls of Bi, Pb and Cu are below the detection limits. All the Brunflo grains are embedded in barite. *~Same grain as 3B-8 in Table 4. *2Same grain as 3B-10 in Table 4. *3Nickel-rich ore in ophiolite, Vinogradova et al. ( 1972 ). *4Same grain as 3B-2 in Table 6.

Atomic proportions based on S + As = 6 for nickeline and 8 for maucherite; M = F e + C o + N i + C u . The contents ofSb, Bi and Pb are below the detection limits. All the grains except 3A-3 (calcite-hosted) are embedded in barite. *~Same grain as 3B-5 in Table 5.

rims corresponds to a replacement of Co and As by Fe and S. Because of the low Fe contents the CoAs2NiAs~-FeAs2 system can be regarded as pseudobinary (Fig. 10). Roseboom (1963) reported that at 800°C there is complete miscibility between synthetic rammelsbergite and saffiorite and that a miscibility gap appeared at lower temperatures. There is an extensive substitution in the rammelsbergite (14-35 mol % CoAs2) and saffiorite ( 15-42 mol % NiAs2) of Brunflo. None of the ana-

lyzed grains approaches end-member composition and no point falls in the intermediate range occupied by the spongy arsenide aggregates in Fig. 7. Sulfur replaces As to a small extent in the two phases (2-4.5 at.% for rammelsbergite and 0.5-2 at.% for saffiorite). The only chemically similar rammelsbergite we have found in the literature is from the Bou Azzer district, Morocco (Table 5; Vinogradova et al., 1972 ). These authors suggested that the solid solution field of rammelsbergite extends fur-


ORDOV1CIAN CHONDRITE FROM BRUNFLO, SWEDEN TABLE 7 Representative microprobe analyses of chalcopyrite and pyrite in Brunflo Chalcopyrite IX-4

Pyrite 4C-2

Wt.% Cu Fe Co Ni S As

34.50 30.43 0.02 0.02 35.11 0.12

0.03 45.32 0.10 0.10 51.24 1.30




0.99 0.99 0.00 0.00 2.00 0.00

0.00 1.00 0.00 0.00 1.98 0.02

At. prop. Cu Fe Co Ni S As

,Atomic proportions based on S + As = 2.

TABLE 8 Microprobe analyses of"anomalous bornite" A





Wt.% Cu Fe S

59.98 10.40 26.19

64.64 11.22 28.15

61.13 12,06 27.80

60.8 11.4 26.2

61.5 11.5 26.5







5.00 0.99 4.32

5.00 0.99 4.31

At. prop. Cu Fe S

5.00 h 12 4.51

5.00 1.07 4.27

5.00 1.06 4.27

Atomic proportions based on Cu = 5. A=Brunflo (analysis IE-3, IMF, Stockholm); Co, Ni and As are below the detection limits. B= Brunflo (analysis 4B-42). C = Mina Esmeralda, Copiap6, Chile (C, = surface and C2 = 10 m level; Sillitoe and Clark, 1969). D = Copper mineralization in thin fissure, Sommerkahl, Germany (von Gehlen, 1964).

ther towards the saffiorite end member than previously thought, a suggestion supported by our data. The minerals of the saffiorite-rammelsbergite system are the only Sb-bearing phases in Brunflo (Table 5), but the Sb content is low. The chemical composition of nickeline usually ranges in the interval msAs6 to M 6 A s 6 , where M = Fe + Co + Ni + Cu. In Brunflo M varies from 5.45 to 5.84, most analyses being closer to the latter

value, with Co contents of about 2-4 wt.% (Table 6). The data are insufficient to show if there is any systematic difference in composition between the nickeline associated with saffiorite and with Cu sulfides. The maucherite formula can be written Mt tAss, often with a deficiency of M. The Brunflo maucherites ( M = 10.32-10.68) contain a variable and often quite high amount of Co (4-9 wt.%). Similar values for saffiorite-free and saffiorite-bearing grains suggest that the Co content is a characteristic feature of the Brunflo maucherites. Petruk et al. ( 1971 ) reported a maucherite from Cobalt (Ontario) with the formula Ni9.72Coo.81As7.97Sb0.o3 , i.e. similar to analysis 3B-2 in Table 6 with respect to Co and Ni. Orcelite(Ms_xAs2), first described by Caill6re et al. (1959) from the Ti6baghi harzburgite massif in New Caledonia, has been regarded as very rare, but several similar occurrences in serpentinized peridotite elsewhere have now been discovered. The mineral is considered to form through interaction between Ni-bearing serpentinizing solutions and maucherite, which in turn may be derived from reaction with nickeline (Oen et al., 1980; Lorand and Pinet, 1984). Orcelite is rare in Brunflo. Only one grain, with the composition Ni3.87Coo.33Feo.06ASl.91So.o9 calculated from an analysis with a very low total, has been found. Its M value (4.26; M = Fe + Co + Ni ) is at the low end of the known range, given as 4.2-5.0 by Fleischer ( 1960 ) and as 4.4-5.5 by Lorand and Pinet (1984). The grain (lost after repolishing) was polygonal and 4/~m in diameter; it was embedded in calcite.

Copper sulfides The ternary system C u - F e - S has been studied repeatedly ever since the first systematic investigation by Merwin and Lombard (1937). It is exceedingly complex at low temperatures with many polymorphs and non-stoichiometric members. Examination of the thin sections with reflected light suggests the presence of four copper sulfides: chalcopyrite, a bornite-like mineral and two different Cu-S phases. Chalcopyrite is the only widespread copper sulfide in Brunflo. As mentioned above it is intimately associated with the spongy cobaltite-gersdorffite aggregates (Figs. 4 - 6 ) , and it occurs as relatively


coarse grains with occasional patches of other Cu sulfides and gersdorffite (Fig. 8). Only chalcopyrire associated with gersdorfilte has been possible to analyze quantitatively. These grains are of very irregular shape (Fig. 8) and up to 0.2×0.4 mm in size, dimensions around 0.02-0.05 mm being more common. Most of them are concentrated in up to 1 mm long elliptical clusters embedded in calcite (one cluster is indicated in Fig. l ). Analyses, given in Table 7, show that the chalcopyrite has a stoichiometric composition with very low contents of Co and Ni. A reddish brown bornite-like mineral ("anomalous bornite") occurs locally as 5-15/tm large bodies intergrown with chalcopyrite. It also occurs as small one-phase grains in the vicinity of the chalcopyrite grains, which they resemble morphologically. A few lamellae of chalcopyrite penetrate the "bornite" along cleavage planes (Fig. 8). Two microprobe analyses of the "bornite" are given in Table 8. They are similar in spite of being determined in different drill cores with different instruments. Their Cu:Fe ratio is quite normal 5: 1, but the S values (4.31 and 4.32 at. prop.) are much higher than for bornite proper (4.00). The x-bornite of Yund and Kullerud (1966) has S=4.05, which is far from the Brunflo values. Sillitoe and Clark ( 1969 ) described "anomalous bornite" as an intermediate oxidation product of bornite in the supergene alteration zone of several Chilean ores and von Gehlen ( 1964 ) has reported a similar phase. Three of their analyses are included in Table 8, demonstrating that our phase chemically falls within the range o f " a n o m a l o u s bornite". Another reddish brown sulfide but with less Cu than the "anomalous bornite" was observed as a small grain in barite. Its composition is close to Cu4FeS4 (Cu3.97Feo.95S 4 ). To the best of our knowledge no mineral or synthetic phase with this composition has been described. The most common of the two Cu-S minerals in Brunflo occurs as tiny pm-sized patches in grains of chalcopyrite and nickeline. It has a grayish blue color and looks like chalcocite, but analysis of a relatively large patch (10 #m) in nickeline gives CUl.665 (Cu5S3) which does not correspond to any reported Cu-S mineral. However, Koch and Mclntyre (1976) obtained a synthetic phase with similar composition (Cu 1.65-1.68S ) during anodic oxidation of chalcocite. X-ray diffraction showed that struc-


turally this phase resembled the metastable 5a-type digenite of Morimoto and Kullerud ( 1963 ). A dark blue copper sulfide looking like covellite is associated with the CUl.665 mineral at two places in Brunflo.

Other sulfides A single almost euhedral grain of pyrite has been observed in Brunflo. It is a 0.02 mm large section through a pyritohedron situated in barite. Repeated analyses (Table 7) show that it lacks Co and Ni, and is As-bearing with 1.1-1.3 wt.% As. There is no optical evidence that the As values are due to arsenide impurities. Galena of stoichiometric composition is present as scattered 5-20 #m large grains with shapes like ink spots in some cobaltite-gersdorffite aggregates. The textural relationship suggests that the galena is replacing the arsenide. A few chalcopyrite grains contain small amounts of sphalerite (or wurtzite) in their outer part, too small for a determination of the Fe content.

Titanium dioxide A titanium dioxide mineral of stoichiometric composition occurs as numerous grains of uniform size (0.02-0.04 m m ) . It is preferentially associated with the layer silicate (Fig. 11 ), but many grains are also embedded in calcite; very few are found in barite. The shape of the grains varies from euhedral

Fig. 11. Euhedral grains of titanium dioxide (white) associated with shreds of layer silicate (dark gray ) in Brunflo (light gray = barite; medium gray = calcite). Reflected light.



(square to rectangular cross-sections; Fig. 11 ) to irregular, though bounded by planar growth surfaces in detail. We have not been able to determine whether the phase is rutile, anatase or brookite, but it is probably one of the two latter polymorphs because they are more typical of low-temperature environments (Lindsley, 1976 ).

Calcite The parts of Brunflo with well-preserved chondrite structure illustrated by Thorslund et al. ( 1984 ) are composed of calcite ( + l a y e r silicate). These areas are labeled 'atom-for-atom' replacement in Fig. 2, because the original chondritic structure is preserved. The calcite here and in other parts of Brunflo is almost pure. A few analyzed points selected at random show only a limited solid solution ofMg, Fe and Mn (0.1-0.6 wt.% each).

Barite White fissure fillings of barite is a conspicuous feature of Brunflo (Fig. 2; Thorslund et al., 1984, Fig. 2), but Fig. 1 illustrates that it also occurs as a patchy replacement. Individual grains are large, usually many millimeters across. The barite is very rich in fluid inclusions containing mainly aqueous solutions and subordinate vapor; there is no indication of carbon dioxide or other gases (C. Broman, Stockholm, pers. commun.). Chemically, the barite is nearly pure BaSO4 with a small amount of S r (0.9-2.6 wt.% SrO). This range is typical of marine barites (Church, 1979). It is the only Sr-bearing mineral in Brunflo; the apatite contains very little of this element.

Apatite One of the chondrules observed in Brunflo consists essentially of apatite (Fig. 12 ), but otherwise the apatite is associated with layer silicate and calcite and unevenly distributed in matrix as well as chondrules (Fig. 1 ). About 25 analyses of apatite were made in the area depicted in Fig. 1, the majority in the apatite chondrule, but the matrix-replacing apatite is also represented. No systematic corn-

Fig. 12. Two chondrules in Brunflo: the upper one (possibly a pseudomorph after a radial pyroxene chondrule) has a pigmented inner part and is composed of apatite with titanium dioxide as small dark grains, and the lower one consists of calcite and layer silicate. The nature of the surrounding phases is seen in Fig. I (central upper part). positional difference related to site could be detected. The analyses show that it is a Cr- and Febearing fluorapatite (Table 9) with less than 0.05 wt.% C1 and 0.1 wt.% MnO. A deficiency in P ( < 6 at. prop.; Table 9) combined with sometimes high F contents ( > 2 at. prop. ) indicates that the apatite may be carbonatebearing. McClellan (1980) and Nathan (1984) discuss the geochemistry and mineralogy of carbonate-fluorapatite (francolite) on basis of the extensive literature. The most favored idea, supported by experimental evidence, is a substitution of PO43- by the group (CO3, F) 3-. Substitution by (CO3, OH) 3- might also take place, although the evidence for it is less strong. Several analyses of Brunflo apatite have Ca values exceeding 10 atoms, suggesting the presence of small amounts of calcite. The Cr content of the apatite displays a considerable variation, with most values in the range 0.05-


178 TABLE 9

Representative microprobe analyses of apatite in Brunflo 2A-24

l E-2

Wt.% CaO FeO Cr203 P205 F SiO2

55.94 0.32 nd 41.16 3.56 0.38

54.80 0.07 0.12 41.06 4.17 0.50

S O~F Sum

101.40 1.51 99.89

100.80*l 1.76 99.04

10.19 0.05 5.92 1.91

10.01 0.01 0.02 5.93 2.25



At. prop. Ca Fe 2+ Cr 3+ P F M

The analyses were made in chondrules (2A-15 and 24 in the apatite chondrule illustrated in Fig. 12). Atomic proportions based on O + F = 2 6 ; M = C a + C r + F e . The contents of CI, Sr, Mn, Na, K, AI and Ti are below the detection limits and the SiO2 values at least partly reflect contamination. *lSum includes 0.08 wt.% NiO.

cate in tiny veinlets; some calcite grains contain numerous small bodies of silica. The phase is quartz, because the (non-fibrous) crystals are too coarse to be chalcedony, and the birefringence is too high for tridymite and cristobalite. Occurrence of quartz rather than tridymite and cristobalite is consistent with the nature of the layer silicate (see below), and heating of the limestone hosting Brunflo to 200300 ° C during burial (cf. Thorslund et al., 1984).

Layer silicate In the previous papers about Brunflo (Thorslund and Wickman, 1981; Thorslund et al., 1984) it was stated that the layer silicate is phengite. However, chemical analysis (Table 10, column A) and X-ray diffractograms of polished thin sections rich in the silicate suggest that it is a strongly disordered illitetype mineral. Beds rich in layer silicate in outcrops and quarries of the Brunflo limestone show no sign of disintegration, indicating that the silicate is nonTABLE 10

0.4 wt.% C r 2 0 3 ; there are also several values around 0.8 wt.% and even higher. This suggests that Cr is incorporated in the apatite structure. Most of the Cr-rich analyses were made in the apatite chondrule (Fig. 12) where no chromite grains are seen. The apatite also contains iron (.¢=0.21 wt.%; S D = 0 . 1 5 ) . The Cr:Fe ratio varies, in agreement with the indication that Cr (and Fe) is substituted in the apatite rather than occurring as tiny included chromite grains. The S i O 2 content is probably due to contamination, because (a) SiO2 shows a large variation within small areas, (b) quartz is a widespread constituent of Brunflo and (c) apatite in sedimentary rocks (phosphorites) does not contain noticeable amounts of SiO2 in the crystal structure.

Quartz Small amounts of silica are widespread in Brunrio, but this phase would easily be overlooked were it not for its relatively high hardness which has given the polished surface a considerable relief. Silica forms a discontinuous fringe between the layer silicate and barite and it is associated with layer sili-

Microprobe analyses of layer silicate in Brunflo A main silicate

B veinlet silicate

Wt.% SiO2 TiO2 A1203 Cr203" V203 *~ FeO t°t MgO CaO Na20 K20

51.5 0.08 21.0 1.06 _+0.21 1.15_+0.11 5.51 3.22 0.25 0.03 7.66

50.5 0.03 19.4 0.74 _+0.14 0.78+0.07 6.61 3.69 0.24 0.10 8.63




At. prop. Si AI Cr V Ti Fe 2+ Mg Ca Na K Fe: Mg *~

6.88 3.74 0.13 0.14 0.01 0.70 0.73 0.04 0.01 1.48 0.98 _+0.06

6.88 3.53 0.10 0.12 0.00 0.85 0.85 0.04 0.03 1.70 1.12 _+0,06

Atomic proportions based on O = 22 ( a n h y d r o u s cell ). The content of Mn is below the detection limit. *1 Average value.



main silicate of Brunflo, their ratio are the same (Table 10).

The formation of the secondary minerals 0.9





Mg ratio

Fig. 13. Variation of the Fe: Mg ratio in Brunflo layer silicate. swelling. The layer silicate occurs alone or together with calcite as a replacement after chondrules and matrix (Fig. 1 ) and in small amounts as veinlets crossing the other secondary minerals. The main silicate is olive to brownish green and the veinlet silicate is emerald green. The two generations of layer silicate and especially the younger one vary in composition, but they are rather similar chemically (Table 10) in spite of the color difference. The veinlet silicate contains less A1, Cr and V, more K, Fe and Mg and it has a higher Fe: Mg ratio (Fig. 13 ). The Fe: Mg ratio of the main silicate is quite uniform, considering that its contents of Fe and Mg vary (4.3-6.4 wt.% FeO and 2.53.7 wt.% MgO, excluding extreme values). It has long been known that total Fe and Mg occur in fairly constant and characteristic amounts in the two groups of ordinary chondrites, H and L. For example, Michaelis et al. (1969) reported an average Fe:Mg ratio of 0.85 (highest value 0.91 ) for nine H-falls and 0.63 (highest value 0.74) for 16 Lfalls. The Brunflo ratio (0.98) is much closer to the H-ratio than to the value for L-chondrites, which is consistent with the classification of Brunflo as an H-group chondrite (see Thorslund et al., 1984). The elements V and Cr are of particular interest because both of them occur in about one weight percent in the layer silicate (Table 10). The Cr might be derived from the alteration of chromite, but analysis in profiles towards relicts of this mineral shows no obvious trends with regard to Cr. The high and likewise rather constant values of V suggest an external source for this element. In ordinary chondrites the amounts of V and Cr differ much. The V value is usually in the range 50-90 ppm whereas Cr is around 2000-4000 ppm (Greenland and Lovering, 1965). Although less V and Cr became incorporated in the veinlet silicate than in the

Local geologic history around Brunflo after its fall Brunflo fell on the continental shelf of the Iapetus Ocean in Llanvirnian time at about 460-470 Ma B.P. and came to rest in calcareous mud (Thorslund et al., 1984). The water column must have strongly reduced the velocity of the stone, which therefore only could penetrate the bottom sediments to a small depth. The calcareous mud was of hemipelagic type deposited at a slow rate in a very tranquil environment (Jaanusson, 1982) and hardgrounds formed periodically (Lindstrrm, 1979). The mud was part of a thin sedimentary cover (preserved as Cambrian sandstones and alum shales and Ordovician limestones and shales) above a Precambrian peneplain composed of granite and olivine diabase dikes (SGU, 1984). Metasomatic reactions between the meteorite and surrounding water-mud mixture started immediately after the fall. By the time the sediment was cemented most of the original minerals were in all likelihood transformed to secondary phases, because they were highly unstable in the new environment. In addition, the high porosity of chondrites (Christophe Michel-Lrvy, 1978 ) favored fast reactions; most of the reactions were metastable. The rapid variation of secondary mineralogy at a mmscale indicates non-equilibrium conditions. Studies of very similar Ordovician epicontinental limestones in other parts of Sweden suggest that the mud probably was lithified about 1000-100,000 years after its deposition (M. Lindstrrm, Stockholm, pers. commun., 1989). The formation of secondary phases continued during subsequent stages of diagenesis. The next known event in the region is the formation of an impact structure (the Lockne structure) according to Wickman ( 1988 ). Its wall was located about 6 km south of the finding-place of Brunflo. Larsson ( 1973 ) described a locality where the limestone that hosted Brunflo is overlain by anomalous clastic rocks (cf. Lindstr~m et al., 1983) now interpreted as ejecta from the shock crater. The impact



took place in Caradoc, 20-40 Ma after the fall of Brunflo. During the Caledonian orogeny in Silurian-Devonian time there was a general pressure from west to east. Some Swedish geologists consider a movement of a few kilometers possible for the sedimentary cover in the area where Brunflo was found (near the eastern border with the Precambrian basement), while others think that the sedimentary rocks here are practically autochthonous. Paleothermal analysis of conodonts (L6fgren, 1978; BergstriSm, 1980) and vitrinite reflectance studies (Kisch, 1980) suggested temperatures in the range 200300°C for sedimentary rocks from the finding-place of Brunflo. Bergstr6m (1980) attributed the elevated temperatures to burial of the rocks beneath a perhaps 3-7 km thick pile of nappes. The interval following the Caledonian orogeny to the present is not distinguished by any event of relevance for this investigation so far we know. The area where Brunflo was found was probably covered by Old Red Sandstones, eroded away after uplift.

The original chemical and mineralogical composition of Brunflo Thorslund et al. (1984) classified Brunflo as a H4 or H5 chondrite based on two independent lines of evidence: frequency of different chondrule types and chromite chemistry. Average chemical compositions and ranges (highest and lowest values) for these meteorite groups are given in Table 11. It can be assumed that Brunflo had a composition within the listed ranges. Average abundances of some minor and trace elements in H-group falls have been given by Greenland and Lovering ( 1965; values in ppm): Cr=2455, B a = 7 , C u = 109, F = 130, Sr= 11 and V = 69. Hamaguchi et al. (1965) reported an As content of about 2 ppm in a H5 fall. According to the criteria of Van Schmus and Wood (1967) H4 and H5 chondrites consist of homogeneous grains of olivine and pyroxene, predominantly microcrystalline aggregates of feldspar and nickel-iron and troilite; chromite is a common accessory. The amounts of silicate (largely olivine and pyroxene), metal and troilite in H-chondrites are seen in Table 11. The original minerals of Brunflo must have occurred in similar proportions.

TABLE 11 Chemical compositions of H4 and H5 meteorites H4






Fe Ni Co FeS SiO2 TiO2 A1203 Cr203 FeO MnO MgO CaO Na20 KzO P205

16.99 1.67 0.09 5.18 36.93 0.11 2.36 0.52 9.18 0.30 23.44 1.69 0.92 0.10 0.21

14.74-18.39 1.49- 1.81 0.07- 0.13 4.02- 6.11 36.30-38.38 0.07- 0.13 1.98- 2.93 0.43- 0.67 8.52-11.04 0.21- 0.32 22.84-24.18 1.22- 2.00 0.82- 1.01 0.08- 0.14 0.05- 0.32

17.26 1.70 0.09 5.65 36.33 0.12 2.24 0.49 9.37 0.29 23.37 1.69 0.90 0.10 0.28

14.70-26.19 1.38- 2.24 0.06- 0.13 2.87- 9.65 30.08-39.00 0.10- 0.15 1.98- 2.78 0.39- 0.54 5.25-11.04 0.22- 0.32 19.77-25.00 1.20- 1.84 0.69- 1.10 0.07- 0.13 0.22- 0.36



Metal Troilite Silicate Total Fe

18.75 5.18 75.76 27.42

99.88 16.34-20.88 4.02- 6.11 73.01-79.64 26.10-28.77

19.05 5.65 75.18 28.13

16.45-28.56 2.87- 9.65 61.79-79.50 25.36-36.40

The values, based on the compositions of ten H4 and nine H5 meteorites determined after 1960 by distinguished analysts, are taken from the following sources: H4: Clarke et al., 1975 (Kiffa); Easton and EIliott, 1977 (Ochansk); Hutchison et al., 1981 (Menow, Monroe, Quenggouk ); Mason and Wiik, 1965 (Weston); Mason and Wiik, 1966 (Bath); Noonan et al., 1972 (Forest Vale); Shima et al., 1974 (Marilia); and an average between analyses in Clarke el al., 1975, and Hutchison et al., 1973 (Kabo). H5: Clarke et al., 1971 (Lost City, Piibram, Ucera ); Easton and Elliott, 1977 (Richardton); Jarosewich, 1966 (Ehole); Mason and Wiik, 1965 (Forest City, Geidam); Mason and Wiik, 1966 (Rose City); and an average between analyses in Easton and Elliott, 1977, and Jarosewich and Mason, 1969 (Allegan).

Compositional changes during the alteration Point counter analysis ( 10248 points counted in seven thin sections) gives the following approximate proportions of the main constituents of Brunflo: calcite 67%, barite 17%, layer silicate 10%, cobaltite-gersdorffite 3%, quartz 2%, chromite 0.4%, chalcopyrite 0.3% and titanium dioxide 0.2%. Apatite ( < 3%) is included in the layer silicate and calcite values. The barite content was obtained by point counting the entire cross section of Brunflo (Fig. 2 ), because the distribution of barite is too uneven at thin section scale. The value for cobaltite-gersdorfrite, including all other arsenides which together amount to less than 2% of the cobaltite-gersdorfrite, is probably too high due to the spongy nature of this mineral.



its total sum (in layer silicate, apatite and corroded chromite relicts) is too low for a H4 or H5 meteorite. Some elements (A1, Ti and Mn) were immobile during the alteration of Brunflo. The titanium dioxide grains most likely represent Ti originally present in the pyroxene. Besides titanium dioxide, cobaltite-gersdorffite is the only secondary mineral in Brunflo that shows influence of a precursor. The grate-like structure of the spongy aggregates (Fig. 6 ) gives them a superficial similarity to altered pyrrhotite with development of marcasite/pyrite along the basal cleavage. It is therefore possible that the cobaltite-gersdorffite aggregates are pseudomorphs after troilite, with pyrrhotite as an intermediate alteration product. This is consistent with their local occurrence as discontinuous rims around chondrules (Figs. 12, 13 in Thorslund et al., 1984).

Bi2 m

10 000









Ve 001


Fig. 14. Estimated gains and losses during the alteration of Brunflo ( 1=original composition ). The CO2 and H20 additions are not shown. An estimate of gains and losses of major and some minor elements during the alteration of Brunflo is illustrated in Fig. 14. It is based on the original chemical composition inferred for Brunflo (represented by the averages of the H 4 - H 5 values in Table 11 ) and its present mineralogical composition and mineral chemistry described in preceeding sections. The uncertainty involved in the calculation means that the gain/loss values must be regarded as very approximative. The alteration involves drastic chemical changes. Two of the three major constituents (Fe and Mg) are almost completely lost and only a small part of the third (Si) remains. Their place are largely taken by C, Ca and Ba. Considerable amounts of As, S, K, P and Co have been added and the contents of Sr, V, Cu and F have been significantly increased. Most of the original Ni is lost, whereas Co is added, changing the Ni:Co ratio from about 10:1 to 1 : 1. There also appears to be a partial loss of Cr, because

Alteration process The present composition of Brunflo is the result of complex reactions showing surprisingly many similarities with fossilization processes in sediments. The most important of these processes are replacement, recrystallization and cementation (Rolfe and Brett, 1969 ). Very slow gradual replacement of the original constituents by inorganic matter through substitution, described in an oversimplified way as 'atom-for-atom' replacement, is responsible for the delicate preservation of microstructures in many fossils. This process can also be inferred from Brunflo (cf. Fig. 2 ). The best preserved chondrules occur where the original minerals are replaced by calcite alone. Where olivine and pyroxene were replaced by layer silicate the resulting volume increase led to a partial or total obliteration of primary structures. Only the outline of former chondrules can be discerned locally. Destruction of structures was particularly strong in the central part of Brunflo, coinciding with the occurrence of numerous barite-filled cracks (Fig. 2). Increases in grain size during recrystallization which produce cracking and other features likewise destroy microstructures in fossils. Brunflo was in a highly reduced state when it became embedded in the calcareous mud. Its iron occurred as native metal and ferrous Fe in silicates and troilite and there was no sulfate. The only ferric phase was magnetite in the fusion skin. The biofacies of the limestone indicates that the mud was in


an oxidized state. The original constituents of Brunflo were unstable to sea water; the order of increasing stability was nickel-iron, troilite, olivine, other silicates and chromite. The original Ni content of the nickel-iron is unknown, since nothing is left for a study. In general, the stability increases with higher Ni content. Metallic iron cannot exist at equilibrium in the presence of water at any pH (Garrels and Christ, 1965, Figs. 7.3 to 7.5) and it seems safe to assume that also nickel-iron is unstable and therefore was dissolved with formation of gaseous hydrogen, leaving voids in the meteorite. Troilite, absent in Brunflo, must also have been dissolved. The H2S produced from troilite was partly oxidized to sulfate by 02 dissolved in the sea water. Barium ions in the sea water precipitated sulfate as barite. The silicates are more resistant to alteration and this may be one of the reasons that Brunflo is so well-preserved. Analogy with submarine weathering of basalts (Velde, 1985 ) suggests that smectites form after olivine and less readily after pyroxene. Weathering of ultrabasic rocks under subaerial conditions also produces smectite (Velde, 1985). The present layer silicates in Brunflo, two generations of illite-type minerals, are later products of low-temperature metamorphism. However, most of the silicates were replaced by calcite, which is to be expected in an environment consisting of lithifying calcareous mud, since a high CO2 activity stabilizes carbonates at the expense of Ca-A1 silicates during diagenesis and low-grade metamorphism (Liou et al., 1987). The similarity between Brunflo and serpentinized ultrabasic rocks with regard to accessory minerals is remarkable, although both magnetite (and serpentine ) are absent in Brunfio. The minerals described by Ramdohr (1967) and others from these rocks comprise native metals (including native nickel), sulfides low in sulfur as well as arsenides. The Brunflo sulfides and arsenides are probably the result of slow reactions (time scales of 104-106 years) at very low temperature (Wickman and Nystr6m, 1986). Such conditions are especially indicated by the "anomalous bornite" and perhaps the spongy aggregates of cobaltite-gersdorffite. Temperatures in the range 0 - 5 0 ° C prevailed for a long time in the limestone hosting Brunflo, due to the slow rate of sedimentation in the tectonically stable area (Jaanusson, 1982). The conditions


changed abruptly with the impact event forming the Lockne structure 20-40 Ma after the fall of Brunflo. The impact was accompanied by a temperature rise in the vicinity of the structure, but the temperature increase cannot be distinguished from the regional heating to 200-300°C attributed to burial below nappes (Bergstr6m, 1980). Local movements of the Cambro-Ordovician sedimentary cover could also have contributed to the rise in temperature. The original meteorite behaved as an open system in contact with sea water. This explains the severe loss of major elements and gain by some other elements (Fig. 14). The water-rock exchange was largely completed when the calcareous mud became cemented, because the porosity of the resulting limestone was very low. For example, the considerable amount of barite in Brunflo ( 17 vol.°/0) cannot have been added to the stone after the lithification. There are no veins or veinlets crossing Brunflo that could serve as channelways for later solutions and the concentric zones of bleached rock surrounding it (Thorslund et al., 1984) are undisturbed. Apatite is a widespread minor constituent in some beds of the epicontinental limestones deposited in Sweden during the Early to Middle Ordovician (LindstfiSm and Vortisch, 1983). Typically, it occurs associated with clay minerals as a dissemination and forms massive phosphate aggregates. Its presence in Brunflo is therefore quite normal for this environment. Impregnations by calcium phosphate are characteristic ofhardgrounds (Bathurst, 1976). The secondary mineralogy of Brunflo presents a number of difficult problems. This is not surprising because there is nothing to compare with. Brunflo is the first recovered stony meteorite that has been altered in sea water. Buddhue ( 1957 ) has discussed the secondary minerals of meteorites, but all his data are derived from continental finds. Moreover, Brunflo fell a long time ago and has a complicated terrestrial history. Our knowledge of the stone is essentially two-dimensional, based on the study of one surface through the meteorite ( = the limestone slab) and a few short drill cores, and thus fragmentary. In spite of all our efforts, we have no real chemical view of the stone. A major problem is the source of some of the elements enriched in Brunflo, such as Ba, As, Co, V and Cu. Their abundances in sea water are low, ranging between 20 × 10 - 3 ppm ( Ba ) and 0.4 × 10- 3 ppm (Co) (Mason, 1982). However, they might



have been released to the water through leaching of rocks below the thin sequence of limestone. The alum shales and olivine diabase dikes occurring here are possible sources. Precipitation by redox reactions (e.g. reduction of AsO43- and VO 3- ) and perhaps adsorption by colloid clay minerals could concentrate these elements in Brunflo. The high oxidation state of sulfur in barite is also a problem. Most of the other S- and As-bearing phases in Brunflo (including the native nickel), as well as the absence of Fe-oxides and the surrounding zones of bleached limestone demonstrate that reducing conditions prevailed here, at least during the later stages of diagenesis.

Conclusions Brunflo is an altered stony meteorite found in a limestone slab from the R6dbrottet quarry near Brunflo village, J~imtland County, almost at the geographical center of Sweden. It fell into calcareous mud on the continental shelf of the Iapetus Ocean during the Middle Ordovician (Llanvirnian) at about 460-470 Ma B.P. All the original minerals except chromite and rare grains of chrome spinel are replaced by secondary phases, but the chondritic structure is surprisingly well-preserved. A statistical treatment of the different chondrite types and the composition of the chromite relicts indicate independently that Brunflo is a H4-H5 chondrite (Thorslund et al., 1984). The following conclusions can be stated from a study of the secondary mineralogy of Brunflo. (1) The inferred original minerals, largely olivine and pyroxene, nickel-iron and troilite, are replaced by calcite (two thirds of the present stone), barite, layer silicate and small amounts of apatite, quartz, titanium dioxide, eight different Co-Ni arsenides/sulfarsenides and four Cu-bearing phases; a few grains of native nickel, galena, pyrite and sphalerite have also been found (Table 1 ). (2) The calcite is quite pure and the barite is slightly Sr-bearing (typical of the marine depositional environment). The layer silicate seems to be a strongly disordered illite-type mineral; it occurs in two generations. Chemically, the apatite is a carbonate-fluorapatite. (3) The most common Co-Ni phase, whose composition is close to CoNi (AsS)2, might be a new

mineral. It forms spongy aggregates closely associated with chalcopyrite. There are also two other members of the cobaltite-gersdorffite system in Brunflo: gersdorffite patches in a second variety of chalcopyrite and a Fe-rich gersdorffite rimming grains of rammelsbergite. Other Co-Ni minerals are tiny mosaics of nickeline (and/or maucherite ) and safflorite, and rare orcelite. (4) All the Cu minerals, besides chalcopyrite, are rare or previously unobserved. They comprise "anomalous bornite" (known as a rare oxidation product after bornite), CUl.66S (only known as a synthetic phase) and Cu4FeS4. (5) Besides a few pyrite grains of anomalous composition (=As-bearing), none of the other common iron minerals in normal rocks (magnetite, hematite, pyrrhotite) occur in Brunflo. (6) The secondary mineralogy varies rapidly at a mm-scale, indicating non-equilibrium formation conditions. The alteration process was accompanied by chemical changes. The major constituents (Mg, Fe, Si) are almost completely lost and their place is taken by C, Ca and Ba. Much As, S, K, P, Co, Sr, V, Cu and F have also been added; the Ni: Co ratio has changed from about 10: 1 to 1 : 1. (7) Most of the exchange reactions took place in the calcareous mud-sea water mixture before the mud was cemented to limestone, which is considered to have taken place about 1000-100,000 years after the deposition of the sediment. During subsequent stages of diagenesis the temperature remained within the 0-50°C range for many millions of years, the long time span at ambient temperature probably being a controlling factor for the formation of the CoNi(AsS) 2 phase, the "anomalous bornite" and other phases of unusual composition. Rising temperature due to an impact event 20-40 Ma after the fall of Brunflo and regional heating to 200-300°C during the Caledonian orogeny in Silurian-Devonian time might have modified the secondary mineralogy.

Acknowledgements We thank B. Levi, B. Lindqvist and M. Lindstrrm for constructive criticism. I. Arnstr/Sm and S. Jevall drew the fgures, D. Holtstam helped with the SEM photographs, and C. Broman studied the f u i d inclusions in barite. We are grateful to G.E. Sigval-

184 d a s o n for p e r m i s s i o n to use t h e m i c r o p r o b e at the N o r d i c V o l c a n o l o g i c a l I n s t i t u t e a n d to K. G r 6 n v o i d for assistance. W e are also i n d e b t e d to C. A1i n d e r for his a n a l y t i c a l skill a n d to C. W e s t m a n for advice. Economic support from the Wallenberg F o u n d a t i o n is a c k n o w l e d g e d .

References Bathurst, R.G.C., 1976. Carbonate Sediments and Their Diagenesis. Elsevier, Amsterdam, 660 pp. Bergstr6m, S.M., 1980. Conodonts as paleotemperature tools in Ordovician rocks of the Caledonides and adjacent areas in Scandinavia and the British Isles. Geol. F/Sren. Stockholm F~3rh., 102: 377-392. Buddhue, J.D., 1957. The oxidation and weathering of meteorites. Univ. New Mexico Publ. in Meteoritics, 3: 1-161. Caillrre, S., Avias, J. and Falgueirettes, J., 1959. Drcouverte en Nouvelle-Calrdonie d'une minrralisation arsenicale sous forme d'un nouvel arsrniure de nickel Ni2As. C.R.A. Sci., Paris, 249: 1771-1773. Christophe Michel-Lrvy, M., 1978. Estimation de la porosit6 de quelques chondrites par analyse d'images de leurs sections polies. Meteoritics, 13: 305-309. Church, T.M., 1979. Marine barite. In: R.G. Burns (Editor), Marine Minerals. Mineral. Soc. Am. Rev. Mineral., 6:175209. Clarke, R.S., Jarosewich, E. and Nelen, J., 1971. The Lost City, Oklahoma, meteorite: an introduction to its laboratory investigation and comparisons with the P~'ibram and Ucera. J. Geophys. Res., 76: 4135-4143. Clarke, R.S., Jarosewich, E. and Noonan, A.F., 1975. Preliminary data on eight observed-fall chondritic meteorites. Smithson. Contrib. Earth Sci., 1975( 14): 63-70. Easton, A.J. and Elliott, C.J., 1977. Analyses of some meteorites from the British Museum (Natural History) collection. Meteoritics, 12: 409-416. Fleischer, M., 1960. New mineral names. Orcelite. Am. Mineral., 45: 753-754. Frost, B.R., 1985. On the stability of sulfides, oxides, and nalive metals in serpentinile. J. Petrol., 26:31-63. Garrels, R.M. and Christ, C.L., 1965. Solutions, Minerals, and Equilibria. Harper & Row, New York, 450 pp. Gehlen, K., von, 1964. Anomaler Bornit und seine Umbildung zu ldait und "Chalkopyrit" in deszendenten Kupfererzen von Sommerkahl (Spessart). Fortsehr. Mineral., 41 : 163. Greenland, L. and Lovering, J.F., 1965. Minor and trace element abundances in chondritic meteorites. Geochim. Cosmochim. Acta, 29:821-858. Hamaguchi, H., Onuma, N., Furukawa, M., Hirao, Y. and Bando, S., 1965. Tin, antimony, and arsenic in carbonaceous chondrites. In: K.H. Wedepohl (Editor), Handbook of Geochemistry, v. II-3. Springer, Berlin, p. 33-C-2. Hutchison, R., Ajakaiye, D.E., Elliott, C.J. and Fry, F.A., 1973. The Kabo, Nigeria, meteorite fall. Mineral. Mag., 39: 340345. Hutchison, R., Bevan, A.W.R., Easton, A.J. and Agrell, S.O.,

J.O, NYSTROMAND F.E.WICKMAN 1981. Mineral chemistry and genetic relations among Hgroup chondrites. Proc. R. Soc. London, A374:159-178. Jaanusson, V., 1982. Introduction to the Ordovician of Sweden. Palaeontol. Contrib. Univ. Oslo, 279: 1-9. Jarosewich, E., 1966. Chemical analyses of ten stony meteorites. Geochim. Cosmochim. Acta, 30:1261 - 1265. Jarosewich, E. and Mason, B., 1969. Chemical analyses with notes on one mesosiderite and seven chondrites. Geochim. Cosmochim. Acta, 33:411-416. Kanehira, K., Banno, S. and Yui, S., 1975. Awaruite, heazlewoodite, and native copper in serpentinized peridotite from the Mineoka district, southern Boso Peninsula. J. Jpn. Assoc. Mineral. Petrol. Econ. Geol., 70: 388-394. Kisch, H.J., 1980. Incipient metamorphism of Cambro-Silurian clastic rocks from J~imtland Supergroup, central Scandinavian Caledonides, western Sweden: illite crystallinity and vitrinite reflectance. J. Geol. Soc. London, 137: 271-288. Klemm, D.D. and Weiser, T., 1965. Hochtemperierte Glanzkobalt-Gersdorffit-Mischkristalle von Outokumpu/ Finnland. N. Jahrb. Mineral. Monatsh., pp. 236-241. Koch, D.F.A. and Mclntyre, R.J., 1976. The application of reflectance spectroscopy to a study of the anodic oxidation of cuprous sulphide. J. Electroanal. Chem., 71: 285296. Larsson, K., 1973. The Lower Viruan in the autochthonous Ordovician sequence of J~imtland. Sver. Geol. Unders., C683: 1-82. Lindsley, D.H., 1976. Experimental studies of oxide minerals. In: D. Rumble (Editor), Oxide Minerals. Mineral. Soc. Am. Rev. Mineral., 3: L61-L88. Lindstrrm, M., 1979. Diagenesis of Lower Ordovician hardgrounds in Sweden. Geol. Pal~eontol., 13: 9-30. Lindstrrm, M., Simon, S., Paul, B. and Kessler, K., 1983. The Ordovician and its mass movements in the Lockne Area near the Caledonian margin, central Sweden. Geol. Pal~eontol., 17: 17-27. LindstrSm, M. and Vortisch, W., 1983. Indications of upwelling in the Lower Ordovician of Scandinavia. In: J. Thiede and E. Suess (Editors), Coastal Upwelling. Plenum, New York, pp. 535-551. Liou, J.G., Maruyama, S. and Cho, M., 1987. Very low-grade metamorphism of volcanic and volcaniclastic rocks-mineral assemblages and mineral facies. In: M. Frey (Editor), Low Temperature Metamorphism. Blackie, Glasgow, pp. 59-113. Lrfgren, A., 1978. Arenigian and Llanvirnian conodonds from J~imtland, northern Sweden. Fossils Strata, 13: 1-129. Lorand, J.P. and Pinet, M., 1984. L'orcelite des prridotites de Beni Bousera (Maroc), Ronda (Espagne), Table Mountain et Blow-Me-Down Mountain (Terre-Neuve) et du Pinde septentrional (Grbce). Can. Mineral., 22: 553560. Mason, B., 1982. Principles of Geochemistry. 4th ed. J. Wiley, New York, 344 pp. Mason, B. and Wiik, H.B., 1965. The composition of the Forest City, Tennasilm, Weston, and Geidam meteorites. Amer. Museum Novitates, 1965 (2220): 1-20. Mason, B. and Wiik, H.B., 1966. The composition of the Bath,


Frankfort, Kakangari, Rose City, and Tadjera meteorites. Am. Mus. Novit., 1966(2272): 1-24. McClellan, G.H., 1980. Mineralogy of carbonate fluoraparites. J. Geol. Soc. London, 137: 657-681. Merwin, H.E. and Lombard, R.H., 1937. The system, CuFe-S. Econ. Geol., 32: 203-284. Michaelis, H., von, Ahrens, L.H. and Willis, J.P., 1969. The composition of stony meteorites. II. The analytical data and an assessment of their quality. Earth planet. Sci. Lett., 5: 387-394. Morimoto, N. and Kullerud, G., 1963. Polymorphism in digenite. Am. Mineral., 48:110-123. Nathan, Y., 1984. The mineralogy and geochemistry ofphosphorites. In: J.O. Nriagu and P.B. Moore (Editors), Phosphate Minerals. Springer, Berlin, pp. 275-291. Noonan, A.F., Fredriksson, K.A. and Nelen, J., 1972. Forest Vale meteorite. Smithson. Contrib. Earth Sci., 1972(9): 57-64. Oen, I.S., Kieft, C., Burke, E.A.J. and Westerhof, A.B., 1980. Orcelite and associated minerals in the N i - F e - A s - S system in chromitites and orthopyroxenites of Nebral, M~ilaga, Spain. Bull. Minrral., 103: 198-208. Petruk, W., Harris, D.C. and Stewart, J.M., 1971. Characteristics of the arsenides, sulpharsenides, and antimonides. Can. Mineral., 11: 150-186. Ramdohr, P., 1967. A widespread mineral association, connected with serpentinization with notes on some new or insufficiently defined minerals. Neues Jahrb. Mineral. Abh., 107: 241-265. Rolfe, W.D.I. and Brett, D.W., 1969. Fossilization processes. In: G. Eglinton and M.T.J. Murphy (Editors), Organic Geochemistry - Methods and Results. Springer, New York, pp. 213-244. Roseboom, E.H., 1963. C o - N i - F e diarsenides: compositions and cell dimensions. Am. Mineral., 48:271-299. Rosner, B., 1970. Mikrosonden-Untersuchungen an natiirlischen Gersdorffiten. Neues Jahrb. Min. Monatsh., pp. 483-498.

185 SGU, 1984. Bedrock geology of J~imtland County, 1:200,000. Sver. Geol. Unders. Ca53 (map in 4 sheets). Shima, M., Jochum, K.P., Sigholfi, G.P. and Hintenberger, H., 1974. The chemical composition of major elements and heavy trace metals in chondrites Parambfl and Marilia. Meteoritics, 9: 199-207. Sillitoe, R.H. and Clark, A.H., 1969. Copper and copper-iron sulfides as the initial products of supergene oxidation, Copiap6 mining district, northern Chile. Am. Mineral., 54: 1684-1710. Thorslund, P. and Wickman, F.E., 1981. Middle Ordovician chondrite in fossiliferous limestone from Brunflo, central Sweden. Nature, 289: 285-286. Thorslund, P., Wickman, F.E. and Nystr/Sm, J.O., 1984. The Ordovician chondrite from Brunflo, central Sweden, I. General description and primary minerals. Lithos, 17: 87100. Van Schmus, W.R. and Wood, J.A., 1967. A chemical-petrologic classification for the chondritic meteorites. Geochim. Cosmochim. Acta, 31: 747-765. Velde, B., 1985. Clay Minerals. A Physico-Chemical Explanation of Their Occurrence. Elsevier, Amsterdam, 428 pp. Vinogradova, R.A., Yeremin, N.I. and Krutov, G.A., 1972. Cobalt-rich rammelsbergite from the Bou Azzer district, Morocco. Dokl. Akad. Nauk. SSSR, 207:118-120. Vokes, F.M. and Strand, G.S., 1982. Atoll texture in minerals of the cobaltite-gersdorffite series from the Raipas mine, Finnmark, Norway. Soc. Geol. Appl. Miner. Deposits, Spec. Publ., 2: 118-130. Wickman, F.E., 1988. Possible impact structures in Sweden. In: A. Bodrn and K.G. Eriksson (Editors), Exploration of Deep Continental Crust, v. 1. Deep Drilling in Crystalline Bedrock. Springer, Berlin, pp. 298-327. Wickman, F.E. and Nystrtim, J.O., 1986. Brunflo, the first fossil chondrite. Meteoritics, 21: 538. Yund, R.A. and Kullerud, G., 1966. Thermal stability of assemblages in the Cu-Fe-S system. J. Petrol.. 7: 454-488.