Geochemistry of Carbonaceous Impactites from the Gardnos Impact Structure, Norway

Home , Suevite

Geochimica et Cosmochimica Acta, Vol. 67, No. 20, pp. 3889–3903, 2003 Copyright © 2003 Elsevier Ltd Pergamon Printed in the USA. All rights reserved 0016-7037/03 $30.00 ϩ .00 doi:10.1016/S0016-7037(03)00213-8 Geochemistry of carbonaceous impactites from the Gardnos impact structure, Norway

1, 2 1 1 1 3 3 I. GILMOUR, *B.M.FRENCH, I. A. FRANCHI, J. I. ABBOTT, R. M. HOUGH, J. NEWTON, and C. KOEBERL 1Planetary and Space Sciences Research Institute, The Open University, Milton Keynes, MK7 6AA, UK 2Smithsonian Institution, Washington, DC 20560, USA 3Institute of Geochemistry, University of Vienna, Althanstrasse 14, A-1090 Vienna, Austria

(Received September 17, 2001; revised 14 October 2002; accepted in revised form October 14, 2002)

Abstract—The Gardnos impact structure in southern Norway is one of only two known impact structures (among ϳ175) whose impactites contain signiﬁcant amounts (typically 0.2–1.0 wt.%) of carbon, or 5 to 10 times the amount present in the target rocks; Sudbury, Canada is the other. This study extends a previous investigation of the geochemistry and petrology of Gardnos impactites (French et al., 1997) with additional sampling and a detailed investigation of the nature and possible origin of the carbonaceous material present. Two principal carbon components have been identiﬁed in Gardnos impactites: (1) impact-produced diamonds, 0.5 to 1 ␮m in size, with a cubic crystal structure, predominantly hexagonal morphologies with platey layers and an estimated concentration of Ͻ 0.19 ppm in Gardnos suevites and (2) graphitic carbon ranging from poorly ordered to moderately crystalline. Geochemical data suggests that there are no suitable target rocks that could provide a single source for the carbon in Gardnos impactites. However, Raman spectroscopy, stable isotope analysis and transmission electron microscopy of the impact diamonds and graphitic carbon suggests that there were at least two episodes of C emplacement in Gardnos impactites: an impact-related incorporation and shock transformation of graphitic material from target rocks followed by later mobilization of C, possibly during postimpact cooling or later regional metamorphism. Copyright © 2003 Elsevier Ltd

1. INTRODUCTION present overlying the basement rocks at the impact site, was suggested as a possible carbon source by French et al. (1997). The Gardnos impact structure (Dons and Naterstad, 1992; The original study (French et al., 1997) was based on a suite French et al., 1997) is located in southern Norway (60°40ЈN, of samples collected in 1994. During the summer of 1996, a 9°00ЈE), ϳ125 km northwest of Oslo, Norway. The structure, larger sample suite was collected at Gardnos to expand the originally ϳ5 km in diameter, has been deeply eroded. It lacks available suite of known basement (target) rocks and impactites any distinctive circular form and is now represented by out- and to investigate in more detail the source of the carbon in the crops of impact-produced breccias, both lithic and melt-bear- impactites. Gardnos samples were collected both from outcrop ing. Its impact origin has been convincingly established from and from a tunnel dump along the Dokkelvi River, which had three lines of evidence (French et al., 1997): (1) the presence of provided fresh samples of a wide range of basement rocks and melt-bearing breccias containing granitic rock fragments with impactites (French et al., 1997). In addition, samples of two definite Planar Deformation Features (PDFs) in quartz and black shale units, which might originally have been present feldspar; (2) the demonstration, using chemical mixing models, among the Gardnos target rocks, were collected from the clos- that the breccia compositions could be produced by mixing the est exposures to Gardnos. The Cambrian Alum Shale (Anders- exposed target rocks; (3) an extraterrestrial signature from the son et al., 1985; Thickpenny and Leggett, 1987; Bharati et al., projectile in the breccias, established using osmium-isotope 1995) was sampled from near Bjørgo, ϳ50 km ENE of Gard- analyses. The age of the structure is poorly constrained between nos, and the Proterozoic Biri Shale (Bjørlykke et al., 1976; a metamorphic age for the basement target rocks (ϳ900 Ma) Tucker, 1983; Vidal and Nystuen, 1990) from the west shore of and the Caledonian orogeny at ϳ385 Ma (Grier et al., 1999). Lake Mjøsa, ϳ100 km E of Gardnos. A major unresolved problem about the Gardnos structure is This paper presents results of geochemical studies on the the unusually high content of carbon in its impact-produced new group of samples, describes detailed studies on the carbo- rocks. Gardnos is one of only two known impact structures naceous material in the Gardnos impactites and related rocks, (among ϳ175) whose impactites contain significant amounts and evaluates the role of black shales in general, and the Biri (typically 0.2–1.0 wt.%) of carbon, or 5 to 10 times the amount and Alum Shales in particular, as sources for the carbonaceous present in the target rocks (French et al., 1997); Sudbury, material in the Gardnos impactites. Canada is the other (French, 1968; Bunch et al., 1999; Hey- mann et al., 1999). Values of ␦13C for the Gardnos carbon range from Ϫ28 to Ϫ32‰ (French et al., 1997), strongly 2. SAMPLE DESCRIPTIONS suggesting an origin from biogenically derived carbon. The carbonaceous Proterozoic Biri Shale, which could have been Based on hand-specimen and petrographic examinations, no new lithologic types not described earlier (French et al., 1997) were identified at Gardnos among either the target rocks or the * Author to whom correspondence should be addressed various types of impactites. The new samples (see Appendix) ([email protected]). can be included in the earlier nomenclature, and the earlier 3889 3890 I. Gilmour et al. petrographic descriptions can be applied to the current suite of 3. EXPERIMENTAL Gardnos samples (French et al., 1997, pp. 877–883). 3.1. XRF and INAA Three basement (target) lithologies have been distinguished (French et al., 1997): (1) a variable suite of granitic gneisses Major element oxides and some trace elements (Rb, Sr, Y, Zr, Nb, Ͼ Co, Ni, Cu, Zn, V, Cr and Ba) concentrations were determined on that form the majority ( 75–80 area %) of the outcrop area; (2) powdered samples by X-ray fluorescence (XRF) spectrometry at the amphibolite, generally present as crosscutting dikes in the Department of Geology, University of Witwatersrand, Johannesburg, granitic gneisses; (3) a coarse-grained metamorphic ortho- South Africa. Details on procedures, precision and accuracy are de- quartzite, which is white and massive outside the structure and scribed by (Reimold et al., 1994). Other trace elements (Sc, Cr, Co, Ni, becomes black and highly fractured within it. In addition, a Zn, As, Se, Br, Rb, Sr, Zr, Sb, Cs, Ba, Hf, Ta, W, Ir, Au, Th and U) and the REEs were determined using Instrumental neutron activation anal- wide range of impactites are exposed within the structure. The ysis (INAA). These analyses were carried out at the Institute of Geo- term impactite here designates all rocks, both coherent and chemistry, University of Vienna, Austria. Details on analytical proce- fragmental, produced by the action of shock waves (Stoeffler dures (including standard data) are given in (Koeberl, 1993). and Grieve, 1994). Impactites at Gardnos are: (1) shocked quartzite, which is black and highly fractured within the struc- 3.2. Isolation and Identification of Carbon Components ture (French et al., 1997, pp. 877–879); (2) lithic breccias: (a) Powdered samples (ϳ4 g) were partially demineralised using pro- the well-known subcrater “Gardnos Breccia” (Broch, 1945), cedures described previously (Gilmour et al., 1992; Hough et al., 1997) which consists of fragments of white granitic gneiss in a with the exception that microwave assisted dissolution was used for the pulverized black matrix; (b) a “black-matrix breccia” (French initial removal of silicates. The acid-dissolved residues were sub- sampled and treated with chromic acid (6 mol/L at 70°C) to remove et al., 1997, p. 881), which is similar to the “Gardnos Breccia,” amorphous and organic carbon. Aliquots of this residue were then but contains fewer fragments and a higher percentage of gen- treated with fuming perchloric acid at 110°C to remove crystalline erally darker matrix; (3) melt-bearing breccias: (a) suevite, in graphite. The residues were examined using a transmission electron which fragments of melt and basement rock occur in the clastic microscope (TEM), a JEOL 2000FX operated at 200 kV, and equipped matrix; (b) melt-matrix breccias (or impact-melt breccias) in with an EDS system for chemical characterization. The TEM was also to obtain selected-area electron diffraction (SAED) patterns to confirm which crystallized melt forms a matrix to rock and mineral the identity of carbon components. clasts (French et al., 1997). Samples of two carbon-bearing black shale lithologies now 3.3. Carbon Abundance and Isotopic Measurements exposed in the region around Gardnos, the Alum Shale (Cam- brian) and the Biri Shale (Proterozoic), were also sampled to Whole-rock carbon abundance and isotopic composition were determined on powdered samples treated with dilute HCl to remove any examine their suitability as sources for the carbon, on the carbonate components. Triplicate analyses were performed using an assumption that one or both of these units might have been elemental analyser coupled to a stable isotope mass spectrometer present at the site of the Gardnos structure at the time of the (Micromass Optima operated in continuous flow mode with a Carlo impact event (see French et al., 1997, pp. 898–899). Erba elemental analyser). Acid residue samples were also analysed using stepped combustion, incremental heating in oxygen, on a high The Alum Shale was collected from an outcrop near Bjørgo, sensitivity static vacuum mass spectrometer to determine the carbon ϳ50 km ENE of Gardnos, where it occurs as a crumpled and content and isotopic composition of components within the residue. deformed unit in parautochthonous rocks exposed beneath an (Prosser et al., 1990). overthrust sheet of Proterozoic quartzite (J. Naterstad, personal communication). In thin section, the shale displays a pro- 3.4. Raman Spectroscopy nounced contorted banding in which black, opaque, and organ- For all measurements a Jobin-Yvon Labram HR laser Raman mi- ic-rich layers alternate with bands and lenses of microcrystal- croprobe system was used. Excitation was with a 40-mW 514-nm line quartz on a submm scale. Ar-ion laser, power at the sample was 2 to 2.5 mW. Individual carbo- Biri Shale samples were collected from the west side of Lake naceous grains were supported on glass slides. Spectra were obtained using a 50 X objective and a 600 grids/mm grating from 600 to 2300 Mjøsa, ϳ100 km E of Gardnos and ϳ2 km S of the town of Ϫ Ϫ cm 1 with a spectral resolution of 2 cm 1. The detection time was Biri. At this site, the Proterozoic section that includes the Biri varied form 1 to 50 s, but most of the samples were analyzed with a Shale is interpreted as an allochthonous block, transported from detection time of 30 s. Peak areas, widths, positions and ratios were the NW during the Caledonian orogeny. Estimates of the post- determined using Jobin-Yvon Labspec software. depositional transport of the Biri shale during this event vary from 20 to 30 km (Bjørlykke et al., 1976) to Ͼ 150 km (Tucker, 4. RESULTS AND DISCUSSION 1983; Vidal and Nystuen, 1990), and its location with respect to 4.1. Geochemistry Gardnos at the time of the impact is unknown. At the sampling site, the Biri Shale is a dark grey to black, fissile, and highly We analysed 20 samples of rocks from the Gardnos impact contorted unit at least 10 m thick, which is interbedded with structure, as well as 4 samples of regionally occurring black more abundant carbonate rocks and intraformational carbonate shales, for their major and trace element composition. The breccias (Tucker, 1983). In thin section, the samples consist of results are given in Table 1 for basement rocks at Gardnos, and sub-mm organic-poor layers of fine detrital grains, dominantly in Table 2 for a variety of impactites, crater-fill sediments, and quartz with typical grain sizes of Ͻ 50 to 100 ␮m, interspersed the black shales. These data complement the existing database with finer-grained layers of dark, organic-rich material. The of the composition of 20 other samples from Gardnos, which appearance of the Biri Shale in thin section suggests a signif- were published in French et al. (1997). icantly lower content of organic material than that in the Alum The basement rock types at Gardnos include granitic gneis- Shale, an interpretation that is supported by the analyses of the ses, amphibolite, and quartzite. In the present work we present two units. data for five granitic gneisses and four amphibolites, bringing Carbonaceous impactites from the Gardnos Crater 3891

Table 1. Major and trace element contents of the two main basement rock types at Gardnos.

Granitic gneiss Amphibolite

NGF-96-114 NGF-96-115 NGF-96-116 NGF-96-119 NGF-96-134 NGF-96-107 NGF-96-112 NGF-96-117 NGF-96-118

SiO2 69.22 71.89 71.45 77.49 74.82 47.18 43.39 53.36 53.80 TiO2 0.67 0.57 0.55 0.40 0.39 1.75 2.54 2.26 2.20 Al2O3 14.56 13.05 13.83 10.07 12.13 15.77 16.14 13.67 12.36 Fe2O3 4.27 2.87 3.00 3.50 2.28 14.59 16.41 13.54 13.48 MnO 0.09 0.07 0.06 0.05 0.03 0.24 0.25 0.19 0.22 MgO 0.69 0.53 0.60 0.28 0.24 7.56 7.91 4.67 4.72 CaO 1.84 1.54 1.20 1.00 0.53 6.89 6.02 5.21 7.58

Na2O 3.69 3.65 3.03 2.80 3.40 1.45 1.75 3.02 2.76 K2O 4.38 5.20 5.96 3.80 5.96 3.41 2.69 2.04 1.61 P2O5 0.06 0.09 0.09 0.05 0.02 0.20 0.40 0.47 0.51 LOI 0.35 0.41 0.81 0.75 0.53 1.36 3.03 1.97 1.21 Total 99.82 99.87 100.58 100.19 100.33 100.40 100.53 100.40 100.45 Sc 8.96 7.33 6.85 5.01 2.59 34.1 38.1 34.6 35.5 V 22 24 24 15 15 271 275 249 266 Cr 1.8 6.2 4.8 2.1 3.5 120 142 76.7 101 Co 2.48 3.07 2.94 1.64 0.69 42.7 42.7 31.2 32.8 Ni 10 10 8 12 10 95 164 50 55 Cu 2 2 Ͻ22 213145795 Zn 90 56 53 57 74 143 189 144 110 Ga 31 15 9 22 4 40 50 120 14 As 0.39 0.44 0.42 0.46 0.52 0.53 0.53 0.69 0.29 Se 0.15 0.14 0.12 0.09 0.11 0.18 0.12 0.06 0.07 Br 0.39 0.17 0.22 0.93 0.06 0.51 0.56 0.82 1.52 Rb 219 184 219 158 222 190 130 84.9 67.1 Sr 170 160 145 115 55 125 270 170 145 Y 704643679130425846 Zr 490 350 290 417 380 110 210 292 213 Nb 24 18 16 15 22 6 10 12 11 Ag 0.03 0.11 0.03 0.03 0.09 0.02 0.07 0.04 0.13 Sb 0.07 0.09 0.33 0.39 0.07 0.69 0.82 0.13 0.27 Cs 2.61 5.01 6.57 10.1 7.91 9.05 10.4 6.49 1.96 Ba 1100 990 940 870 650 910 502 430 262 La 80.7 56.1 42.8 54.3 58.6 7.44 14.5 26.1 19.3 Ce 174 115 91.8 120 133 18.3 36.5 59.4 46.5 Nd 87.5 51.9 41.9 64.3 73.1 13.5 24.1 35.1 28.9 Sm 17.6 9.91 7.71 14.1 15.1 4.51 7.25 9.39 7.17 Eu 2.39 1.17 1.01 1.98 1.34 1.29 2.04 2.14 2.09 Gd 17.8 9.52 8.04 14.1 17.1 4.24 8.7 9.55 7.65 Tb 2.58 1.38 1.21 2.17 2.57 0.78 1.45 1.59 1.35 Tm 1.28 0.82 0.74 1.28 1.42 0.45 0.64 0.85 0.65 Yb 8.63 5.51 4.78 8.67 10.1 3.19 3.88 5.49 4.73 Lu 1.13 0.77 0.67 1.12 1.32 0.46 0.53 0.72 0.63 Hf 16.8 11.2 9.31 15.3 13.5 2.78 5.42 7.57 5.49 Ta 0.96 1.04 0.97 0.79 1.28 0.074 0.33 0.37 0.33 W 1.42 0.87 0.54 0.81 0.67 0.26 1.67 1.44 0.31 Ir (ppb) Ͻ0.3 Ͻ1 Ͻ1 Ͻ0.7 Ͻ1 Ͻ1 Ͻ1 Ͻ1 Ͻ1 Au (ppb) 0.4 Ͻ1 Ͻ1 0.5 0.5 0.2 0.2 0.6 Ͻ1 Hg 0.03 0.02 0.03 0.03 0.03 0.02 Ͻ0.02 0.04 0.12 Th 24.2 18.7 16.6 15.6 20.6 0.22 0.22 4.02 1.99 U 5.09 6.01 3.35 4.77 3.76 0.31 0.31 0.79 0.75 K/U 7171 7210 14,826 6639 13,209 91,667 72,312 21,519 17,889 Zr/Hf 29.2 31.3 31.1 27.3 28.1 39.6 38.7 38.6 38.8 La/Th 3.33 3.00 2.58 3.48 2.84 33.8 65.9 6.49 9.70 Hf/Ta 17.5 10.8 9.60 19.4 10.5 37.6 16.4 20.5 16.6 Th/U 4.75 3.11 4.96 3.27 5.48 0.71 0.71 5.09 2.65

LaN/YbN 6.32 6.88 6.05 4.23 3.92 1.58 2.53 3.21 2.76 Eu/Eu* 0.42 0.38 0.40 0.44 0.26 0.92 0.80 0.71 0.88

ϭ Major element data in wt.%, trace element data in ppm, except as noted. All Fe as Fe2O3.N chondrite-normalized. gneiss data to a total of nine samples and amphibolite data to matrix breccias. One sample of crater-ﬁll sediment was anal- six samples. No further samples of unshocked quartzite were ysed, one Alum shale and three Biri shale samples. Average analysed. Of the impactites, we analysed one more sample of compositions for those rock types for which more than two shocked (black) quartzite, two Gardnos Breccia samples, two samples were analysed (granitic gneiss, amphibolite, suevite, Black-matrix Breccia samples, three suevites, and two melt- and Biri shale) are presented in Table 3. 82I imu tal. et Gilmour I. 3892 Table 2. Major and trace element contents of impactites from the Gardnos structure, crater-ﬁll sediments, and regional carbon-rich shales.

Gardnos breccia Black-matrix breccia Gardnos suevite Melt-matrix breccia Black shales Black Crater-ﬁll quartzite sediment Alum shale Biri shale Biri shale Biri shale NGF-96-143 NGF-96-120 NGF-96-121 NGF-96-126 NGF-96-176 NGF-96-133 NGF-96-137 NGF-96-178 NGF-96-179 NGF-96-175 NGF-96-160 NGF-96-164 NGF-96-169 NGF-96-170 NGF-96-171

SiO2 95.44 67.93 72.76 70.14 69.63 61.52 66.32 69.13 66.40 69.57 64.59 57.20 58.68 59.61 56.15 TiO2 0.21 0.78 0.38 0.22 0.74 0.79 0.42 0.95 0.89 0.74 1.67 0.92 0.89 0.89 0.89 Al2O3 2.19 13.92 12.23 14.56 11.30 13.90 12.95 10.88 13.62 13.62 16.57 14.99 15.96 15.67 17.17 Fe2O3 0.31 3.97 3.19 1.75 5.65 6.96 6.57 4.93 4.58 2.66 4.76 5.17 9.37 7.54 9.83 MnO 0.02 0.06 0.05 0.03 0.16 0.13 0.14 0.10 0.07 0.06 0.04 0.03 0.24 0.33 0.10 MgO 0.33 1.27 0.45 0.56 1.79 2.9 2.22 2.41 1.48 1.16 1.29 1.51 2.92 2.74 2.67 CaO 0.01 1.69 0.67 0.16 1.03 1.2 0.97 1.57 1.19 1.11 0.02 0.57 0.64 1.65 0.47 Na2O 0.06 3.60 3.11 4.68 1.64 1.34 1.57 2.42 1.77 1.95 0.55 0.17 1.88 2.14 0.01 K2O 0.63 5.51 5.81 6.41 5.94 7.18 7.04 4.47 8.18 8.28 5.99 5.24 3.14 3.20 3.77 P2O5 0.01 0.13 0.06 0.03 0.17 0.18 0.08 0.19 0.26 0.17 0.02 0.13 0.22 0.20 0.16 LOI 1.24 1.59 1.05 1.66 1.73 2.90 1.87 2.98 1.49 1.22 3.78 14.67 5.38 5.53 7.90 Total 100.45 100.45 99.76 100.20 99.78 99.00 100.15 100.03 100.50 100.54 99.28 100.60 99.32 99.5 99.12 Sc 0.76 11.2 4.29 1.15 9.42 11.5 4.67 13.4 12.8 10.2 17.8 15.1 18.2 16.2 18.2 V 17 41 17 14 35 40 36 72 49 33 90 757 80 74 15 Cr 1.1 14.5 2.3 2.1 32.1 35.6 21.4 59.5 23.3 100 55.3 88.2 54.1 49.2 59.7 Co 0.42 5.52 3.05 2.59 8.24 11.6 12.9 9.92 6.88 6.11 12.7 2.32 18.2 11.5 14.7 Ni 5 10 14 5 60 82 145 80 35 120 46 50 45 35 60 Cu Ͻ222Ͻ2 3 2 162 215722372820 Zn 6.8 87 76 19 78 81 99 45 40 64 36 315 485 300 156 Ga2 6 61710121836202517457 3510 As 0.04 0.34 0.42 0.46 1.34 0.87 0.28 3.09 0.61 0.68 7.94 58.9 25.1 17.4 49.1 Se 0.12 0.22 0.43 0.28 0.11 0.11 0.25 0.09 0.11 0.11 0.41 0.22 0.36 0.28 0.68 Br 0.45 0.18 0.28 0.23 0.25 0.08 0.09 0.35 0.41 0.18 0.04 0.54 0.38 0.88 0.19 Rb 24.3 265 213 185 199 235 254 160 257 265 290 190 133 127 160 Sr 30 215 105 110 100 110 85 115 76 90 40 45 110 190 72 Y 3 60239 4653213549574932655555 Zr 110 405 180 50 320 328 195 257 350 365 670 186 270 248 312 Nb 8 23 10 8 14 16 10 13 15 16 27 19 23 22 20 Ag 0.01 0.07 0.03 0.02 0.06 0.08 0.04 0.12 0.12 0.19 0.07 0.08 0.05 0.12 0.06 Sb 0.031 0.16 0.14 0.09 0.22 0.079 0.12 0.26 0.12 0.041 1.42 8.29 0.34 0.45 0.67 Cs 0.58 10.7 11.5 3.37 7.11 3.68 9.89 3.83 5.61 2.07 25.7 6.17 6.17 6.15 8.92 Ba 150 970 1050 1370 1990 3460 2400 1690 2200 2410 1460 1050 790 1010 925 La 0.45 41.4 60.1 7.77 48.7 28.5 6.04 24.8 25.1 16.3 56.1 32.1 94.2 78.7 73.3 Ce 1.13 100 118 13.1 136 64.5 20.5 53.2 75.1 48.8 118 54.8 199 154 128 Nd 0.83 51.6 53.5 8.52 56.1 35.5 11.7 29.4 45.1 32.9 60.9 26.4 95.3 74.8 68.7 Sm 0.27 11.9 8.03 2.35 12.1 7.79 3.14 7.07 9.46 9.09 10.2 6.94 17.2 13.4 10.1 Eu 0.046 1.51 1.08 0.35 2.01 1.41 0.53 1.36 1.51 1.11 1.45 1.06 2.61 2.11 1.56 Gd 0.4 11.5 5.65 2.11 8.92 8.71 2.82 6.72 9.62 9.33 10.1 7.1 17.6 13.8 12.2 Tb 0.086 1.78 0.76 0.23 1.71 1.45 0.52 1.07 1.42 1.56 1.49 0.99 2.26 1.61 1.56 Tm 0.1 1.03 0.42 0.11 0.89 0.86 0.36 0.61 0.86 0.86 1.04 0.58 1.16 0.92 0.89 Yb 0.78 7.35 2.61 0.71 6.21 6.14 2.82 4.03 5.79 6.41 7.33 3.52 6.96 5.69 5.91 Lu 0.13 0.98 0.39 0.084 0.86 0.84 0.44 0.62 0.77 0.87 0.99 0.52 0.88 0.84 0.76 Hf 2.84 12.9 5.16 1.11 9.46 9.43 4.58 7.61 10.1 9.41 20.9 5.02 7.99 7.57 5.65 Ta 0.094 1.34 0.42 0.19 0.76 0.87 0.34 0.63 0.84 0.71 1.62 1.31 1.43 1.27 1.36 W 0.12 0.73 0.26 0.17 1.19 1.25 0.67 0.82 1.83 0.62 4.13 1.86 0.85 1.25 1.16 Ir (ppb) Ͻ0.1 Ͻ1 0.2 Ͻ0.5 0.5 2.1 0.9 0.4 Ͻ1 4.7 Ͻ0.5 Ͻ1 Ͻ1 Ͻ1 Ͻ1 Au (ppb) 0.1 0.4 0.2 0.1 0.2 0.4 0.2 0.3 0.4 0.8 0.4 0.5 0.5 0.7 0.3 Hg 0.01 0.06 0.08 0.06 0.03 0.21 0.04 0.08 0.06 0.02 0.08 0.03 0.15 0.12 0.21 Th 2.38 23.2 6.69 2.61 10.6 10.5 4.55 8.07 12.6 11.2 25.4 12.4 22.4 20.5 23.7 U 0.54 5.25 2.02 1.06 3.78 3.05 2.24 3.71 3.69 4.22 5.04 23.2 4.06 4.21 4.08 K/U 9722 8746 23,969 50,393 13,095 19,617 26,190 10,040 18,473 16,351 9904 1882 6445 6334 7700 Zr/Hf 38.7 31.4 34.9 45.0 33.8 34.8 42.6 33.8 34.7 38.8 32.1 37.1 33.8 32.8 55.2 La/Th 0.19 1.78 8.98 2.98 4.59 2.71 1.33 3.07 1.99 1.46 2.21 2.59 4.21 3.84 3.09 Hf/Ta 30.2 9.63 12.3 5.84 12.4 10.8 13.5 12.1 12.0 13.3 12.9 3.83 5.59 5.96 4.15 Th/U 4.41 4.42 3.31 2.46 2.80 3.44 2.03 2.18 3.41 2.65 5.04 0.53 5.52 4.87 5.81 LaN /YbN 0.39 3.81 15.6 7.40 5.30 3.14 1.45 4.16 2.93 1.72 5.17 6.16 9.15 9.35 8.38 Eu/Eu* 0.44 0.40 0.50 0.49 0.61 0.54 0.56 0.62 0.50 0.38 0.45 0.47 0.47 0.49 0.44 Carbonaceous impactites from the Gardnos Crater 3893

Table 3. Average compositions of basement rocks and suevites from the Gardnos structure, and regional Biri shales, based on data in Tables 1 and 2.

Granitic gneiss Amphibolites Suevite Biri shale

Ϯ Ϯ Ϯ Ϯ SiO2 72.97 3.22 49.43 5.04 65.66 3.85 58.15 1.79 Ϯ Ϯ Ϯ Ϯ TiO2 0.52 0.12 2.19 0.33 0.72 0.27 0.89 0.00 Ϯ Ϯ Ϯ Ϯ Al2O3 12.73 1.74 14.49 1.79 12.58 1.54 16.27 0.80 Ϯ Ϯ Ϯ Ϯ Fe2O3 3.18 0.75 14.51 1.37 6.15 1.08 8.91 1.21 MnO 0.06 Ϯ 0.02 0.23 Ϯ 0.03 0.12 Ϯ 0.02 0.22 Ϯ 0.12 MgO 0.47 Ϯ 0.20 6.22 Ϯ 1.76 2.51 Ϯ 0.35 2.78 Ϯ 0.13 CaO 1.22 Ϯ 0.50 6.43 Ϯ 1.03 1.25 Ϯ 0.30 0.92 Ϯ 0.64 Ϯ Ϯ Ϯ Ϯ Na2O 3.31 0.39 2.25 0.76 1.78 0.57 1.34 1.16 Ϯ Ϯ Ϯ Ϯ K2O 5.06 0.96 2.44 0.79 6.23 1.53 3.37 0.35 Ϯ Ϯ Ϯ Ϯ P2O5 0.06 0.03 0.40 0.14 0.15 0.06 0.19 0.03 LOI 0.57 Ϯ 0.20 1.89 Ϯ 0.83 2.58 Ϯ 0.62 6.27 Ϯ 1.41 Total 100.16 100.45 99.73 99.31 Sc 6.15 Ϯ 2.44 35.6 Ϯ 1.78 9.86 Ϯ 4.59 17.5 Ϯ 1.15 V20Ϯ 5 265 Ϯ 11 49 Ϯ 20 56 Ϯ 36 Cr 3.68 Ϯ 1.85 110 Ϯ 27.8 38.8 Ϯ 19.3 54.3 Ϯ 5.25 Co 2.16 Ϯ 1.00 37.4 Ϯ 6.21 11.5 Ϯ 1.49 14.8 Ϯ 3.35 Ni 10 Ϯ 191Ϯ 53 102 Ϯ 37 47 Ϯ 13 Cu 2 45 Ϯ 39 7 Ϯ 828Ϯ 9 Zn 66 Ϯ 16 147 Ϯ 32 75 Ϯ 27 314 Ϯ 165 Ga 16 Ϯ 11 56 Ϯ 45 22 Ϯ 12 17 Ϯ 15 As 0.45 Ϯ 0.05 0.51 Ϯ 0.16 1.41 Ϯ 1.48 30.5 Ϯ 16.5 Se 0.12 Ϯ 0.02 0.11 Ϯ 0.06 0.15 Ϯ 0.09 0.44 Ϯ 0.21 Br 0.35 Ϯ 0.34 0.85 Ϯ 0.47 0.17 Ϯ 0.15 0.48 Ϯ 0.36 Rb 200 Ϯ 28.4 118 Ϯ 54.8 216 Ϯ 49.7 140 Ϯ 17.6 Sr 129 Ϯ 46 178 Ϯ 64 103 Ϯ 16 124 Ϯ 60 Y63Ϯ 20 44 Ϯ 12 36 Ϯ 16 58 Ϯ 6 Zr 385 Ϯ 75 206 Ϯ 75 260 Ϯ 67 277 Ϯ 33 Nb 19 Ϯ 410Ϯ 313Ϯ 322Ϯ 2 Ag 0.06 Ϯ 0.04 0.07 Ϯ 0.05 0.08 Ϯ 0.04 0.08 Ϯ 0.04 Sb 0.19 Ϯ 0.16 0.48 Ϯ 0.33 0.15 Ϯ 0.09 0.49 Ϯ 0.17 Cs 6.44 Ϯ 2.84 6.98 Ϯ 3.72 5.80 Ϯ 3.54 7.08 Ϯ 1.59 Ba 910 Ϯ 168 526 Ϯ 275 2517 Ϯ 891 908 Ϯ 111 La 58.5 Ϯ 13.8 16.8 Ϯ 7.87 19.8 Ϯ 12.0 82.1 Ϯ 10.8 Ce 127 Ϯ 30.3 40.2 Ϯ 17.3 46.1 Ϯ 22.9 160 Ϯ 35.9 Nd 63.7 Ϯ 17.8 25.4 Ϯ 9.12 25.5 Ϯ 12.4 79.6 Ϯ 13.9 Sm 12.9 Ϯ 4.01 7.08 Ϯ 2.00 6.00 Ϯ 2.50 13.6 Ϯ 3.55 Eu 1.58 Ϯ 0.58 1.89 Ϯ 0.40 1.10 Ϯ 0.49 2.09 Ϯ 0.53 Gd 13.3 Ϯ 4.40 7.54 Ϯ 2.33 6.08 Ϯ 3.00 14.5 Ϯ 2.77 Tb 1.98 Ϯ 0.65 1.29 Ϯ 0.36 1.01 Ϯ 0.47 1.81 Ϯ 0.39 Tm 1.11 Ϯ 0.31 0.65 Ϯ 0.16 0.61 Ϯ 0.25 0.99 Ϯ 0.15 Yb 7.54 Ϯ 2.28 4.32 Ϯ 1.00 4.33 Ϯ 1.68 6.19 Ϯ 0.68 Lu 1.00 Ϯ 0.27 0.59 Ϯ 0.11 0.63 Ϯ 0.20 0.83 Ϯ 0.06 Hf 13.2 Ϯ 3.02 5.32 Ϯ 1.96 7.21 Ϯ 2.45 7.07 Ϯ 1.25 Ta 1.01 Ϯ 0.18 0.28 Ϯ 0.14 0.61 Ϯ 0.27 1.35 Ϯ 0.08 W 0.86 Ϯ 0.34 0.92 Ϯ 0.74 0.91 Ϯ 0.30 1.09 Ϯ 0.21 Ir (ppb) 1.1 Ϯ 0.9 Au (ppb) 0.5 Ϯ 0.1 0.3 Ϯ 0.2 0.3 Ϯ 0.1 0.5 Ϯ 0.2 Hg 0.03 Ϯ 0.00 0.06 Ϯ 0.05 0.11 Ϯ 0.09 0.16 Ϯ 0.05 Th 19.1 Ϯ 3.42 1.61 Ϯ 1.81 7.71 Ϯ 2.99 22.2 Ϯ 1.61 U 4.60 Ϯ 1.06 0.54 Ϯ 0.27 3.00 Ϯ 0.74 4.12 Ϯ 0.08 K/U 9811 Ϯ 3889 50,847 Ϯ 36,848 18,616 Ϯ 8121 6826 Ϯ 759 Zr/Hf 29.4 Ϯ 1.78 38.9 Ϯ 0.44 37.0 Ϯ 4.82 40.6 Ϯ 12.7 La/Th 3.05 Ϯ 0.37 29.0 Ϯ 27.5 2.37 Ϯ 0.92 3.71 Ϯ 0.57 Hf/Ta 13.6 Ϯ 4.52 22.8 Ϯ 10.0 12.1 Ϯ 1.32 5.23 Ϯ 0.95 Th/U 4.31 Ϯ 1.06 2.29 Ϯ 2.08 2.55 Ϯ 0.78 5.40 Ϯ 0.48 Ϯ Ϯ Ϯ Ϯ LaN/YbN 5.48 1.32 2.52 0.69 2.91 1.37 8.96 0.51 Eu/Eu* 0.38 Ϯ 0.07 0.83 Ϯ 0.10 0.57 Ϯ 0.04 0.46 Ϯ 0.02

ϭ Major element data in wt.%, trace element data in ppm, except as noted. All Fe as Fe2O3.N chondrite-normalized.

A comparison between the new data and those published by rock type there are no important systematic differences to the French et al. (1997) shows a good agreement between the two earlier results. Results obtained from mixing calculations re- datasets. For most elements, the variation between the two sets garding the percentages of basement rocks required to repro- of data are within the range of calculated standard deviations duce the various impactites (Tables 6 and 7 in French et al., (Table 3; compare Table 3 in French et al., 1997) and within the 1997) are considered valid for the purposes of the present work range of compositions shown by the various samples of each as well. 3894 I. Gilmour et al.

Fig. 1. Chondrite-normalized rare earth element (REE) patterns for the major basement and impactite rocks at Gardnos. Normalization factors from Taylor and McLennan (1985).

A comparison of the chondrite-normalized rare earth element (REE) data between the different rock types is shown in Figure 1. Similar to earlier observations (French et al., 1997), the amphibolites have a much less pronounced negative Eu anom- aly (Eu/Eu* ϭ 0.83) than the gneisses (Eu/Eu* ϭ 0.38). With the exception of low light REE (LREE) contents for the amphibolites, the suevites have the lowest overall REE abundances, which might be at least partly related to a higher volatile content. Both melt-matrix breccia and black-matrix breccia have LREE patterns that seem to indicate slight positive Ce anomalies, which might indicate postformational alteration processes. The Biri shale has relatively high overall REE con- Fig. 2. Calculated major element (a) and trace element (b) suevite tents, with a chondrite-normalized pattern that is similar to that composition, based on 18.8% Biri shale, 56.1% granitic gneiss, and of the gneiss, Gardnos Breccia, and suevite. The REE contents 25.1% amphibolite, compared to measured composition (Table 3). “Old” average refers to data from French et al. (1997), “new” to present of the Biri shale samples are about twice that of the Alum shale data. For C content, the highest value of the three samples measured (Table 3). The REE patterns observed in the present work are (NGF-96-169) was used. almost indistinguishable from those published previously (Figs. 13a–13c in French et al., 1997). Based on the results of the mixing calculations reported by and Ba. With exception of Ba, these elements are higher in a French et al. (1997), it is possible to compare the compositions calculated mixture (because of the high abundances in the of the calculated mixtures of the various impactites with mea- shales) than in the suevites actually measured. The reason for sured compositions. The best fit for melt-matrix breccias was the high Ba contents of the suevites has been discussed by 84.9% granitic gneiss, 12.1% amphibolite, and 2.9% black French et al. (1997). Using the Alum shale in place of the Biri shale, whereas for the black-matrix breccia the values are shale in the mixing calculations results in higher C abundances 80.4% granitic gneiss, 7.7% quartzite, and 11.9% black shale. than those actually measured in the suevites (Fig. 3a), while The agreement between calculated major element compositions still showing deviations in the Ca, K, and Na contents. The high for melt-matrix and black-matrix breccias with the actually Ca in the calculated mixture content is probably the result of observed compositions is very good, with the largest deviation the high CaO content of the amphibolites. Compared to the being no more than 5 rel.%. Thus, we feel that the HMX measured suevites, the mixture containing Alum shale has (harmonic least-squares) mixing calculations (French et al., much higher V, As, and Sb, and lower Ba contents (Fig. 3b). 1997) represent the source rocks for these breccias very well. There could be several reasons for these differences; for exam- Slightly larger deviations are observed for suevite. ple, the volatile elements As and Sb could be lost during suevite Using a mixture that includes 18.8% Biri shale, 56.1% gra- formation. However, this does not explain the significant dif- nitic gneiss, and 25.1% amphibolite, Figure 2a shows the ference in V content (which, on the other hand, is not evident calculated abundances stay within a factor of 2 of the measured when using Biri shale). Thus, it is likely that a shale with a abundances, with the most significant deviations for Ca, Na, chemical composition intermediate between Biri and Alum and K. Also, the calculated C content (based on a maximum C shale was incorporated in the suevites. Given the natural vari- abundance of 1.66 wt.% for the Biri shale) is much lower than ation in composition exhibited by the analysed shale samples, the content observed in suevites. For trace element contents the existence of such a component among the surficial target (Fig. 2b), the most significant deviations are for As, Se, Br, Sb, rocks of Gardnos is likely. Carbonaceous impactites from the Gardnos Crater 3895

Table 4. Carbon content, carbon isotopic composition and occurrence of graphite and diamond in Gardnos impactites, possible target rocks and crater-ﬁll sediments.

␦13 [C] CPDB Sample (wt.%) (‰ Ϯ 1␴)GrD

Shocked quartzite (black) NG-94-16 0.17 –29.1 Ϯ 0.6 NG-94-17a 0.39 –29.5 Ϯ 0.4 NGF96-143 0.53 –29.2 Ϯ 0.1 Gardnos breccia NG-94-4a 0.70 –29.1 Ϯ 0.4 NGF-120 0.07 –32.3 Ϯ 3.6 ͱ n.d. NGF-121 0.55 –29.0 Ϯ 0.4 Black matrix breccia NG-94-31 1.04 –27.7 Ϯ 0.4 NG-94-6a 0.38 –29.6 Ϯ 0.4 NGF96-176 1.01 –29.5 Ϯ 0.1 Suevite NG-94-29a 1.11 –30.1 Ϯ 0.7 NG-94-30a 1.01 –30.1 Ϯ 0.1 NGF96-133 0.71 –29.9 Ϯ 0.2 ͱͱ NGF96-137 0.23 –28.5 Ϯ 1.4 ͱͱ NGF96-178 1.15 –30.0 Ϯ 0.1 ͱͱ Melt-matrix breccia NG-94-27 0.20 –28.8 Ϯ 0.3 NG-94-28a 0.34 –31.1 Ϯ 0.4 NG-94-99 0.13 –29.2 Ϯ 0.2 NGF96-175 0.19 –30.0 Ϯ 1.8 NGF96-179 0.13 –28.8 Ϯ 0.6 Crater-ﬁll sediments NG-94-21 0.16 –29.3 Ϯ 0.1 NG-94-22 2.13 –29.9 Ϯ 0.4 NGF96-160 0.51 –32.7 Ϯ 0.8 Biri shale (off-structure) Fig. 3. Calculated major element (a) and trace element (b) suevite NGF96-169 0.37 –28.0 Ϯ 0.6 ͱ n.d. composition, based on 18.8% Alum shale, 56.1% granitic gneiss, and NGF96-170 0.30 –27.6 Ϯ 0.6 25.1% amphibolite, compared to measured composition (Table 3). NGF96-171 1.66 –30.0 Ϯ 0.6 “Old” average refers to data from French et al. (1997), “new” to present Alum shale (off-structure) data. NGF96-164 7.64 –29.6 Ϯ 0.1 ͱ n.d.

a Also measured by French et al. (1997). presence of diamond in the three suevite samples studied but Gr ϭ graphite, D ϭ diamond, n.d. ϭ not detected, ͱ ϭ detected. not the lithic breccia or off-structure samples.

4.3. Transmission Electron Microscopy 4.2. Carbonaceous Components Graphite crystals were identified in all of the samples studied We determined the carbon content and isotopic composition from the 3.35-Å d-spacing in SAED patterns. They were pre- for 28 samples of rocks from the Gardnos impact structure and dominantly hexagonal, platey structures ranging in size from regionally occurring black shales (Table 4). It is apparent that ϳ400 nm to ϳ5 ␮m. SAED patterns varied from well-defined the black colour of many of the impactites: shocked quartzite, spot patterns indicating a high degree of crystallinity to more Gardnos breccia, and black matrix breccia, reflects their ele- diffuse ring patterns that are indicative of polycrystalline or less vated C contents. Six of these samples were subjected to acid crystalline graphite (Buseck and Huang, 1985). demineralisation followed by optical examination, electron mi- Diamond was identified in the three suevite samples studied, croscopy, further stable isotope analysis and Raman spectros- SAED gave d-spacings of 2.06, 1.26, and 1.08 Å characteristic copy to characterize the carbon phases present in both Gardnos of cubic diamond (Fig. 4). No evidence of the hexagonal impactites (NGF-96-120, NGF-96-133, NGF-96-137 and NGF- polymorph of diamond (lonsdaleite) was detected in any of the 96-178) and off-structure possible target rocks (NGF-96-164 carbon residues analysed. However, since TEM is not capable and NGF-96-169). Initial optical examination with a petro- of analysing large sample volumes and the platey nature of the graphic microscope of the acid residues showed them to be diamond crystals results in most samples only being studied in predominantly black and fine-grained (5–50 ␮m) with occa- a single orientation, the presence of lonsdaleite cannot be sional small zircon crystals (10–50 ␮m). The results of the entirely excluded. Diamonds were not detected in the lithic electron microscopic examination of the acid residues are sum- breccia or off-structure samples. The diamond crystals were marized in Table 4. Graphite and amorphous C were identified predominantly hexagonal with platey layers that resulted in in all six acid residues following HF/HCl dissolution, further double reflections in some SAED patterns. The layers occa- acid treatment with chromic and perchloric acids revealed the sionally showed rotation around the axis of the basal plane (Fig. 3896 I. Gilmour et al.

Fig. 5. Raman spectra of carbonaceous material from Gardnos im- ␮ Fig. 4. Bright field TEM image of a 2- m-sized diamond grain from pactites and an off-strucuture Black shale (Biri shale), showing the suevite NG-96-137. The crystalline nature of the grain is clearly ap- order (O) and disorder (D) peaks for graphite. parent from the SAED pattern as is the identification of diamond d111 ϭ 2.06 Å. Its morphology is layered and there are hints of a pseudohex- agonal form. shows a shoulder developed on the high wavenumber side of the 1582 cmϪ1 peak indicating disorder in the graphite. Car- 4). Stacking faults were well developed in the diamond crystals bonaceous material from the Biri shale shows much broader Ϫ1 Ϫ1 and are similar to those observed in other impact diamonds 1582 cm peaks (shifted to high wavenumbers ϳ1597 cm ) Ϫ1 from Lappaja¨rvi and Popigai (Koeberl et al., 1997; Vishnevsky together with a broad 1360 cm peak. et al., 1999). Stacking faults with a spacing of Ͻ 10 nm can be Raman spectroscopy identified a range of ordering for the seen in both the surface layer and several underlying layers carbon present in Gardnos impactites from poorly ordered C in resulting in a cross-hatching pattern at ϳ60° (Fig. 4). the Biri shale to crystalline graphite in the isolated residues from suevite NGF-96-137. The range in ordering of the C is Ϫ1 4.4. Raman Spectroscopic Characterization of Carbon shown in Figure 6 which plots the position of the 1582 cm O-peak against its width. The presence of graphite, identified as First order spectra were analyzed from 600 to 2300 cmϪ1 for a major constituent of Gardnos impactites, by TEM is con- carbonaceous residues from five samples together with spectra firmed by the Raman study. Graphitic material in the Gardnos from a suite of metamorphic graphites that had experienced Breccia and Suevites is more ordered than in the Biri shale with metamorphic grades higher than amphibolite facies. Between some samples approaching the crystallinity of well-ordered three and seven spectra were measured for carbonaceous ma- crystalline graphite. Raman spectra of C in impactites from the terial from each sample. The spectra from the Gardnos samples Sudbury impact structure (Heymann et al., 1999) show a more all contained the first-order single band at ϳ1582 cmϪ1 (O- restricted range of ordering than the Gardnos impactites. peak) that is characteristic of well-crystallized graphite (Fig. 5). Ϫ1 Disorder in graphite appears as a broadening of the 1582 cm 4.5. Carbon Abundance and Isotope Compositions band together with a shift toward higher wavenumbers as a result of the development of an additional band near 1360 cmϪ1 Table 4 gives the C abundance and isotopic compositions for (D-peak) (Wopenka and Pasteris, 1993). With the exception of 28 samples representing 8 lithologies from the Gardnos crater a chemically isolated graphite sample from the suevite NGF- and possible C-rich target rocks. These data include a re- 96-137 carbon from all of the Gardnos impactite samples analysis of 6 samples studied by French et al. (1997) that were studied shows two large peaks, at ϳ1582 and 1360 cmϪ1 (Fig. measured to enable a statistical comparison to be made between 5). Carbonaceous material from suevite NGF-96-178 also the carbon isotopic compositions of the different lithologies. Carbonaceous impactites from the Gardnos Crater 3897

lithic breccias (black matrix breccia and Gardnos breccia) have mean ␦13C values of Ϫ28.9 and Ϫ30.1‰, and the melt bearing breccias (suevite and melt-matrix breccia) have mean ␦13C values of Ϫ29.7 and Ϫ29.6‰ respectively. The crater-fill sediments and Biri shales have mean ␦13C values of Ϫ30.7 and Ϫ28.5‰ while the Alum shale has a ␦13C value of Ϫ29.6‰. However, there are no significant differences in the carbon isotope compositions between the suevite and melt-matrix ϭ ϭ breccia (t7 0.2, p 0.8) or between the suevite and the black ϭ ϭ matrix breccia (t3 1.1, p 0.34) which may imply a common source for the carbon in these rocks. Similarly, the carbon isotopic compositions are not significantly different between ϭ ϭ ␦13 suevite and the Biri shale (t3 1.4, p 0.24) and the C value obtained for the Alum shale is indistinguishable from the mean carbon isotopic composition of the suevite and other impactites. Fig. 6. Variation in Raman spectra O-peak (1582 cmϪ1) width at half The acid-resistant residues from two of the suevite samples height versus O-peak position for carbonaceous material from Gardnos in which diamonds were observed (NGF-96-137 and NGF-96- impactites, off-structure Biri shale and a suite of crystalline graphite 178) were further investigated (Table 5) using stepped com- standards. bustion (Swart et al., 1983). The two samples gave yields of 18.5 and 16.3 wt.% C confirming that these residues contain Carbon contents range from a minimum of 0.07 wt.% in a significant quantities of non-carbonaceous material, most prob- sample of Gardnos Breccia to a maximum of 7.6 wt.% for a ably zircon and rutile which are difficult to remove completely sample of the Late Cambrian Alum shale. Gardnos impactites by acid dissolution. The carbon contents of NGF-96-137 and have relatively high C contents: a mean of 0.35 wt.% for the NGF-96-178 correspond to 3.1 and 2.7 ppm respectively of the shocked quartzite, 0.81 and 0.44 wt.% for the two lithic breccia bulk suevites, providing an upper limit to the diamond content lithologies (black matrix breccia and Gardnos breccia), and in these samples. 0.84 and 0.20 wt.% for the melt-bearing suevite and melt- Figure 7a shows the carbon release profile for the diamond- matrix breccias. The crater-fill sediments contain an average of containing acid residue from NGF-96-137. This residue was 0.93 wt.% C while the off-structure Biri shales have a mean C treated with chromic acid, which removes a significant propor- content of 0.78 wt.%. The single sample of Alum shale mea- tion of the carbon including organic C, amorphous C and some sured in this study has 7.64 wt.% C; values reported in the poorly crystalline graphite. The majority (85 wt.%) of the literature are typically 10 to 12 wt.% C (Bharati et al., 1995). carbon combusts at temperatures below 500°C. The carbon A5‰ range was observed in ␦13C values for individual released below 500°C corresponds to the combustion temper- samples from the 8 lithologies studied. For the impactites, the ature of organic carbon (Gilmour and Pillinger, 1985) and is shocked quartzites have a mean ␦13C value of Ϫ29.4‰, the most likely organic contamination acquired by the sample

Table 5. Stepped combustion carbon yields and isotope compositions for acid-resistant residues from Gardnos suevites.

␦13 ␦13 [C] (ppm) CPDB (‰) [C] (ppm) CPDB (‰)

T (°C) NGF96-137 NGF96-178

200 5035 –29.9 6771 –29.4 300 14,126 –29.6 42,159 –28.6 350 54,261 –28.0 400 64,895 –30.5 13,153 –29.9 450 61,608 –30.8 17,747 –31.0 500 9650 –30.8 13,619 –31.3 550 3319 –26.6 4438 –29.3 600 2583 –25.2 1664 –26.6 650 3222 –24.1 4961 –32.3 700 5565 –24.0 2674 –27.7 750 7003 –24.0 376 –31.2 800 4557 –24.2 442 –22.5 900 1986 –23.4 416 –22.4 1200 1408 –25.9 525 –22.8 ¥[C] (ppm) 184,956 163,206 ¥[C] (ppm) (whole-rock) 3.1 2.7 ¥␦13C –29.6 –29.1 ¥␦13C(Ͻ500°C)a

a Weighted mean ␦13C for steps Ͻ 500°C. 3898 I. Gilmour et al.

Fig. 7. Stepped combustion profile of carbon from Gardnos suevites acid-resistant residues. (a) HF/HCl and chromic acid treated residue. (b) HF/HCl and chromic acid treated residue followed by removal of graphitic carbon with perchloric acid. The histograms give carbon yield information, and the line graph with filled circles corresponds to the isotopic composition at each temperature step. during laboratory handling. The main component comprising diamond with nanometre-sized impact-produced diamond crys- ϳ70 wt.% of the total C in the residue combusts between 400 tals combusting at ϳ500°C while larger crystals combust at and 500°C and has a mean ␦13C value of Ϫ30.8‰. However, higher temperatures, 800 to 850°C in the case of millimetre- a distinct carbon phase comprising ϳ15 wt.% of the total C is sized diamonds (Ash et al., 1990; Gilmour et al., 1992). It present that combusts at temperatures of between 650 and therefore seems unlikely that any of the C combusting below 800°C and is different in 13C abundance by possibly 6‰ (Ϫ24 500°C is diamond, rather, these components are predominantly vs. Ϫ30.8‰). organic contamination together with acid-resistant amorphous The release profile for the second acid residue sample, NGF- C in NGF-96-137. However, the component combusting be- 96-178, is more complex (Fig. 7b). This residue was further tween 600 and 700°C in NGF-96-178 has a combustion tem- treated with fuming perchloric acid at 110°Cfor4dtoremove perature consistent with the 400 nm to 1.5 ␮m size range more resistant carbonaceous components such as graphite. A observed for the diamonds by TEM. If this component, with a sizeable proportion (65 wt.%) of the carbon combusts below mean ␦13CofϪ29.4‰ is indeed diamond, then summing the C 400°C, this most likely reflects a greater relative proportion of released between 600 and 700°C gives an upper limit for the organic contamination in this sample. Between 400 and 500°C diamond concentration of 0.19 ppm for the suevite NGF-96- some 25 wt.% of the carbon is released and the isotopic 178. composition reaches a minimum of Ϫ31.3‰ (cf. Ϫ30.8‰ for the same release temperature in NGF-96-137). The shift in 4.6. Origin and Preservation of Impact Diamonds ␦13C values between 400 and 500°C is the result of the con- tinued combustion of organic carbon alongside the more 13C- Graphite is ubiquitous in the impactites and potential target depleted component that combusts over this temperature range. rock sources of carbon examined at Gardnos. The transforma- Above 500°C, ␦13C values become more 13C-enriched until the tion of graphite to diamond requires several crystallographic release of a distinct C component between 600 and 700°C(ϳ5 changes: an increase in the interatomic distance within the wt.% C) with more negative ␦13C values that apparently co- individual carbon planes of the graphite of 0.12 Å and a combusts alongside the more 13C-enriched component. decrease in the interplanar spacing of 1.86 Å. However, the Comparison between the perchloric and non-perchloric acid hexagonal ring system of graphite can withstand a considerable treated residues indicates that the majority of the 13C-enriched amount of static compression before the reconstructive trans- component combusting above 650°C in NGF-96-137 is re- formation to cubic diamond or the hexagonal polymorph, moved by perchloric acid, suggesting that this C phase is most lonsdaleite occurs. At pressures Ͼϳ80 GPa, the excess energy likely graphitic in nature. This component is isotopically dis- stored in bond compression is catastrophically released thereby tinct from the C measured in Gardnos impactites (Table 5), distorting the graphite planes to from diamond (Fahy et al., however, it is present in extremely low abundances (ϳ0.4 ppm) 1986). The response of graphite to dynamic high pressures is and therefore represents a minor C component in the suevites markedly different; laboratory experiments have shown that the that is only identifiable as a result of the extensive acid-etching. shock-induced transformation of graphite to diamond can occur Identifying the diamond component is more difficult; grain size at pressures as low as 20 GPa (Erskine and Nellis, 1991) is the primary influence on the combustion temperature of suggesting that a more energy efficient transformation mecha- Carbonaceous impactites from the Gardnos Crater 3899

Table 6. Mean Raman peak parameters of carbonaceous residues from Gardnos impactites and Biri shale.

D peak position D peak width O peak position O peak width Sample (cm–1) (cm–1) (cm–1) (cm–1)

NG-96-169 (n ϭ 7) 1349 Ϯ 4 194 Ϯ 23 1596 Ϯ 6 60.2 Ϯ 9.0 NG-96-120 (n ϭ 4) 1347 Ϯ 152Ϯ 2 1590 Ϯ 4 45.5 Ϯ 2.0 NG-96-133 (n ϭ 5) 1343 Ϯ 553Ϯ 3 1579 Ϯ 3 37.3 Ϯ 5.3 NG-96-178 (n ϭ 3) 1352 Ϯ 156Ϯ 5 1585 Ϯ 1 47.1 Ϯ 3.8 NG-96-137a (n ϭ 2) 1355 Ϯ 137Ϯ 2 1585 Ϯ 1 22.4 Ϯ 1.9

a Graphite-rich residue after treatment with chromic acid.

nism is involved. Several mechanisms have been proposed observed at the Ries crater or diamond-silicon carbide inter- including ultrafast annealing of a glassy carbon phase (Pujols growths (Hough et al., 1995). and Boisard, 1970), martensitic shear transformation of graph- The diamond concentration determined from the stepped ite (Erskine and Nellis, 1991), partial dislocations brought combustion experiments (0.19 ppm) is considerably lower than about by the dissociation of twin boundaries to produce low the concentrations of impact diamonds observed at the Chicxu- energy “diamond-like” stacking faults (Freise and Kellt, 1961; lub crater, 3.6 to 18 ppm (Gilmour et al., 1992; Hough et al., Pujols and Boisard, 1970; Solovev, 1976), and a mechanism 1997), or at the Ries crater, 3 to 10 ppm (Hough et al., 1995; involving a liquid or quasi-liquid interface layer between Abbott, 2000). The preservation of diamond formed by the graphite and shock-produced diamond (Kleiman et al., 1984; shock transformation of graphite requires rapid quenching to Ͻ Heimann and Kleiman, 1988). The strong morphologic simi- 1000 K to inhibit re-graphitisation. The small size of diamonds larity of millimetre-sized impact diamonds to graphite at craters in Gardnos suevites may reflect the size distribution of graphite such as Popigai and Ries has favoured the martensitic shear within the target rocks, graphite may also have been less transformation of graphite as the predominant formation mech- abundant than at other craters (aside from the off-structure Biri anism for these diamonds (Koeberl et al., 1997; Masaitis, and Alum shales, target rocks at Gardnos were apparently low 1998). However, small non-platey diamonds found at the Ries in carbon before the impact), or they may represent the relict crater epitaxially intergrown with silicon carbide have led to cores from re-graphitisation of the diamond. While some areas the suggestion that vapour phase processes may also be in- of the suevite will not have experienced very high tempera- volved in the production of some impact diamonds (Hough et tures, regions in close proximity to the melt matrix breccia will al., 1995). Vapour phase growth of nanometre-sized diamond have experienced higher temperatures that could cause re- crystallites formed by the explosive detonation of graphite has graphitisation of the diamond. A second process that may have also been proposed as a formation mechanism in laboratory resulted in the loss of diamonds in Gardnos suevites is regional experiments (Yamada and Sawaoka, 1994). The observation of metamorphism. The rocks of the Gardnos structure have been the linear carbon polytype, carbyne, in Ries suevites (El Goresy subjected to low greenschist facies metamorphism Ϫ and Donnay, 1968) also implies that carbon in impact rocks (350–400°C, 4 7 kbar) and there is evidence that the abundance of nanometre-sized diamonds in meteorites is controlled may undergo high temperature or liquid phase transformations by metamorphic processes (Huss, 1990). since carbynes are formed at temperatures in excess of 2500 K and may play a role in the shock-induced transformation of graphite to diamond (Wang et al., 1993; Heimann, 1994). 4.7. Origin of Carbon in Gardnos Impactites ␦13 The C value of the putative diamond component identi- Our studies do not support the earlier suggestion (French et fied in the stepped combustion of the perchloric acid residue al., 1997, pp. 898–899) that the Biri Shale could be the source from suevite NGF-96-178 is indistinguishable from the other for the carbon in the Gardnos impactites since carbon contents carbon phases at Gardnos. Despite their small size, the layered in the Biri Shale appear to be too low. Our own measurements plate-like morphology of Gardnos impact-diamonds is charac- (Ͻ2 wt.%; average 0.8 wt.%, see Table 4), are also consistent teristic of a probable graphite precursor. Indeed, there is labo- with other determinations, e.g., 0.5 to 3 wt.%, average 1.5 wt.% ratory evidence that a smaller crystallite size and lower crys- (Tucker, 1983) and Ͻ3 wt.% (Vidal and Nystuen, 1990). Other, tallinity elevates the initial energy states of graphite making it more indirect, arguments against the Biri Shale as the source of relatively easier to transcend the activation-energy barrier to carbon include: (1) the general absence of thick (Ͼ25 m) and diamond (Hirai et al., 1995). However, graphite does not ap- uniform black shale layers in the sequence (Bjørlykke et al., pear to be a prerequisite for the shock-induced formation of 1976; Tucker, 1983; Vidal and Nystuen, 1990), making it diamond in laboratory experiments. A range of carbon phases, unlikely that a random impact would exactly sample an ade- including amorphous carbon blacks, soot and fullerenes, have quate thickness of high-carbon material; (2) the apparent tec- been used to form diamond by shock (Yoo and Nellis, 1991; tonic transport of the Biri Shale from an unknown depositional Donnet et al., 1996). Most of these experiments yield diamond site and the corresponding uncertainty about whether it was in the form of nanometre-sized polycrystalline powders with ever present in the Gardnos area. few larger scale morphologic features. TEM examination of By contrast, the Alum Shale sample analyzed contains sig- diamond-containing residues did not identify any morphologic nificantly more C (7.74 wt.%), a value which is also consistent characteristics similar to skeletal polycrystalline aggregates with other measurements (e.g., 12–22 wt.%, Bharati et al., 3900 I. Gilmour et al.

1995). The Alum Shale could, therefore, supply the observed to the spectra obtained in this study for the Biri shale, i.e., C carbon content of the impactites by the addition of only 1 to 15 with a lower degree of ordering than the C present in the vol.%, an amount that is much more plausible geologically. suevites and breccias but consistent with the temperatures However, the breccias do not show the expected higher As reached during the Caledonian orogeny. This suggests that the contents that should be produced by the addition of even small degree of ordering of C in Gardnos impactites may be related amounts of Alum Shale (As ϭ 58.9 ppm). In addition, incor- to postimpact temperatures with C in suevites having the great- poration of the Cambrian Alum Shale into the breccia requires est degree of ordering, breccias less ordered, and shocked a post-Cambrian age for the Gardnos impact, while some quartzites the least ordered. Andersen and Burke (1996) also indirect geological arguments (J. Naterstad, personal commu- concluded that hydrocarbon fluid inclusions present in the nication) are more consistent with a Proterozoic age. Radio- quartzite were trapped at temperatures of ϳ320°C and pres- metric dating of the Gardnos impact has so far been hampered sures of ϳ2.5 kbar, trapping conditions consistent with condi- by the Caledonian metamorphic effects at ϳ400 Ma (Grier et tions during the Caledonian orogeny but not with the higher al., 1999), and successful dating of the actual impact event temperatures and lower pressures associated with postimpact would help decide between these two alternatives. cooling. Using TEM, stepped combustion analysis and Raman spec- These observations suggest that there are at least two epi- troscopy we have identified several forms of carbon in impac- sodes of C emplacement in Gardnos impactites. An initial tites from the Gardnos crater: micron-sized diamonds, crystal- impact-related incorporation and shock transformation of gra- line graphite and more poorly ordered carbon. Hydrocarbon phitic material from target rocks followed by secondary em- fluid inclusions have also been identified in shocked quartzites placement as a result of mobilization of C as a possible con- from the Gardnos crater that have been interpreted as evidence sequence postimpact hydrothermal activity or during the later for the mobilization of C during metamorphism as a result of regional metamorphism that affected the structure. Stable iso- the Caledonian Orogeny (Andersen and Burke, 1996). The tope analysis by stepped combustion indicates that while the occurrence of impact-produced diamond is evidence that car- majority of graphite is isotopically indistinguishable from the bon in the target rocks was directly incorporated into the whole-rock C isotope compositions of impactites and possible impactites during the impact process. The morphologic simi- C sources, a minor 13C-enriched component was observed larities of these diamonds to a graphitic precursor is suggestive suggesting that more than one source of C was directly incor- that at least some of the graphite now present in the impactites porated during the impact event. Possible scenarios for the later was also incorporated during the impact event. mobilization of C include: (1) mobilization during postimpact The apparent geochemical and geological difficulties of in- cooling when organic-rich material from possibly either the corporating sufficient C directly during the impact event from Biri, Alum or other shales could have migrated into the heavily either the Biri or Alum shales suggests that some of the C may fractured and deformed impactites; (2) mobilization of C dur- have been incorporated into the impactites either during ing the later regional metamorphic event that evidently led to postimpact cooling or at a later stage. More than one episode of the trapping of hydrocarbons in fluid inclusions. C-emplacement may also explain the variations in graphite The model we propose for the incorporation of C within ordering in Gardnos impactites measured by Raman spectros- Gardnos impactites is different from that proposed by Hey- copy (Fig. 6; Table 6), which range from well-crystallized graphite to more poorly-ordered C. The Raman spectra of mann et al. (1999) to explain the high C contents of impactites graphite formed in-situ from the metamorphism of organic at the Sudbury impact structure. The apparent lack of a C-rich matter can be used to estimate the rock’s peak metamorphic target rock in the vicinity of the Sudbury structure led Hey- temperature (Wopenka and Pasteris, 1993). Graphitic material mann et al. (1999) to suggest that the high carbon contents of produced by the progressive metamorphism of organic matter the breccias of the Black Member of the Onaping Formation at shows an increase in crystallinity with metamorphic grade. In Sudbury are the result of biogenically produced organic matter contrast, graphite deposited due to nucleation and crystal and necessitating the slow accumulation of the breccias. Such growth from fluids generally does not show the same degree of a conclusion is at odds with the interpretation of the breccias as ordering as metamorphic graphite subjected to the same tem- suevitic fall-back impact deposits (see Dressler et al., 1996, for peratures (Pasteris and Chou, 1998). Metamorphic tempera- a discussion of the origin of Onaping Formation breccias). tures at Gardnos reached 350 to 400°C, consistent with the However, as at Gardnos there is evidence, in the form of degree of ordering observed in Raman spectra of carbonaceous impact-produced diamonds (Masaitis et al., 1999), that some of material from the Biri shale (Wopenka and Pasteris, 1993). The the C in the Black Member of the Onaping Formation at higher degree of ordering in much of the graphitic material in Sudbury must have been incorporated from C in the target Gardnos suevites and Breccias suggests that this C was either rocks. derived from more crystalline graphite incorporated into these The occurrence of C in unequivocal impact breccias and rocks during the impact event or that it has experienced tem- suevites at Gardnos precludes an origin via the syn-depositional peratures higher than those due to the regional metamorphism biogenic activity model of Heymann et al. (1999). Furthermore, of the Caledonian orogeny. The more crystalline nature of the mechanism we propose for the postimpact mobilization and graphite in the suevites and breccias also apparently precludes incorporation of C at Gardnos would, with suitable C-rich rocks a fluid-deposited origin during regional metamorphism. in the vicinity, be applicable to the Sudbury structure where Andersen and Burke (1996) measured Raman spectra for car- extensive hydrothermal activity has occurred subsequent to the bonaceous material in Gardnos quartzite that were very similar impact event. Carbonaceous impactites from the Gardnos Crater 3901

5. CONCLUSIONS REFERENCES

1. Comparison of the new geochemical data with those of Abbott J. I (2000) Carbon in Giant Impact Craters. Ph.D. dissertation, French et al. (1997) shows a good agreement between the The Open University. Andersen T. and Burke E. A. J. (1996) Methane inclusions in shocked datasets. Mixing model calculations indicate that incorpo- quartz from the Gardnos impact breccia, South Norway. Eur. J. Min. ration of ϳ20% of either the Proterozoic Biri shale or the 8, 927–936. Paleozoic Alum shale gives a reasonable fit for the majority Andersson A., Dahlman B., Gee D. G., and Snall S. (1985) The of elements in Gardnos suevites. However, for the Biri shale Scandinavian Alum Shales. Sveriges Geologiska Undersoekning. there are significant discrepancies in the amount of C that Ash R. D., Arden J. W., Grady M. M., Wright I. P., and Pillinger C. T. (1990) Recondite interstellar carbon components in the Allende could be incorporated (too low) as well as deviations for Ca, Meteorite revealed by preparative precombustion. Geochim. Cosmo- Na and K. For the Alum shale, mixing calculations give chim. Acta 54, 455–468. sufficient C in the suevites but there are significant devia- Bharati S., Patience R. L., Larter S. R., Standen G., and Poplett I. J. F. tions in some trace elements: V, As, and Sb. These results (1995) Elucidation of the Alum shale kerogen structure using a multidisciplinary approach. Org. Geochem. 23, 1043–1058. do not support the earlier suggestion (French et al., 1997) Bjørlykke K., Elvsborg A., and Hoey T. (1976) Late Precambrian that the Biri shale could be the source for the carbon in the sedimentation in the central sparagmite basin of South Norway. Gardnos impactites. Norsk Geol. Tidsskrift. 56, 233–290. 2. Carbon isotopic measurements of C from both impactites Broch O. A. (1945) Gardnosbreksjen i Hallingdal. Norsk Geol. and possible C sources show no significant differences in Tidsskrift. 25,16–25. Bunch T. E., Becker L., Des Marais D., Tharpe A., Schultz P. H., ␦13 C values. Wolbach W., Glavin D. P., Brinton K. L., and Bada J. L. (1999) 3. Impact-produced diamonds have been identified in Gardnos Carbonaceous matter in the rocks of the Sudbury Basin, Ontario, suevites by transmission electron microscopy. The diamond Canada. In International Conference on Large Meteorite Impacts crystals have predominantly hexagonal morphologies with and Planetary Evolution, Vol. 339 (eds. B. O. Dressler and V. L. Sharpton), pp. 331–343. Geological Society of America, Boulder, platey layers, however, selected area diffraction patterns CO. indicate that the diamonds have a cubic crystal structure. Buseck P. R. and Huang B. (1985) Conversion of carbonaceous mate- Diamonds were typically 0.5 to 1 ␮m in size and contained rial to graphite during metamorphism. Geochim. Cosmochim. Acta stacking faults with spacings of Ͻ 10 nm. Stepped combus- 49, 2003–2016. tion analysis of two acid-resistant residues identified a pos- Donnet J. B., Lemoigne C., Wang T. K., Kurdyumov A. V., Borimchuk N. I., and Ostrovskaya N. F. (1996) Characterization of diamond sible diamond component with a concentration in the suevite obtained by shock compression of carbon black. Bull. Soc. Chim. sample studied of Ͻ 0.19 ppm. France 133, 295–300. 4. The concentration of diamond at Gardnos is much lower Dons J. A. and Naterstad J. (1992) The Gardnos impact structure, than the concentrations of impact diamonds observed at Norway. Meteoritics 27, 215. Dressler B. O., Weiser T., and Brockmeyer P. (1996) Recrystallized other impact craters. The preservation of diamond formed impact glasses of the Onaping Formation and the Sudbury igneous by shock-transformation requires rapid quenching to tem- complex, Sudbury Structure, Ontario, Canada. Geochim. Cosmo- peratures of Ͻ 1000 K to inhibit regraphitisation. If suevites chim. Acta 60, 2019–2036. at Gardnos were not quenched rapidly then the diamond El Goresy A. and Donnay G. (1968) A new allotropic form of carbon concentration would be lower. However, the Gardnos struc- from the Ries crater. Science 161, 363–364. Erskine D. J. and Nellis W. J. (1991) Shock-induced martensitic ture has also been affected by metamorphic processes during phase-transformation of oriented graphite to diamond. Nature 349, the Caledonian orogeny which may also have had a control- 317–319. ling influence on the concentration of diamond. Fahy S., Louie S. G., and Cohen M. L. (1986) Pseudopotential total- 5. Raman spectroscopy identified a range of ordering for the energy study of the transitional from rhomohedral graphite to diamond. Phys. Rev. B: Condens. Matter 34, 1191–1199. carbon present in Gardnos impactites from poorly ordered C Freise E. J. and Kellt A. (1961) Twinning in graphite. Proc. R. Soc. in a sample of the Biri shale to crystalline graphite in the London, Ser. A 264, 269–276. isolated residues from suevites. The maximum ordering of C French B. M. (1968) Sudbury structure, Ontario; some petrographic observed is higher that would be expected had the C only evidence for an origin by meteorite impact. In Shock Metamorphism been subjected to the greenschist facies metamorphism of of Natural Minerals (eds. B. M. French and N. M. Short), pp. 383–412. Mono Book Co., Baltimore, MD. the Caledonian orogeny. French B. M., Koeberl C., Gilmour I., Shirey S. B., Dons J. A., and 6. The variations observed in the ordering of C in Gardnos Naterstad J. (1997) The Gardnos impact structure, Norway: Petrol- impactites suggest that there were at least two episodes of C ogy and geochemistry of target rocks and impactites. Geochim. emplacement in Gardnos impactites. An initial impact-re- Cosmochim. Acta 61, 873–904. Gilmour I. and Pillinger C. T. (1985) Stable carbon isotopic analysis of lated incorporation and shock transformation of graphitic sedimentary organic matter by stepped combustion. Org. Geochem. material from target rocks followed by later mobilization of 8, 421–426. C. Gilmour I., Russell S. S., Arden J. W., Lee M. R., Franchi I. A., and Pillinger C. T. (1992) Terrestrial carbon and nitrogen isotopic-ratios from Cretaceous-Tertiary boundary nanodiamonds. Science 258, 1624–1626. Acknowledgments—This work was supported by the U.K. PPARC and Grier J. A., Swindle T. D., Kring D. A., and Melosh H. J. (1999) the Austrian FWF. We are grateful to Jill Pasteris, Dave Kring and an Argon-40/argon-39 analyses of samples from the Gardnos impact anonymous reviewer for comments that significantly improved the structure, Norway. Meteoritics Planet. Sci. 34, 803–807. manuscript. B. M. F. thanks the Barringer Crater Company for support Heimann R. B. (1994) Linear finite carbon chains (carbynes)—Their for field studies and sample collection. role during dynamic transformation of graphite to diamond, and their geometric and electronic-structure. Diamond Relat. Mater. 3, 1151– Associate editor: B. R. Frost 1157. 3902 I. Gilmour et al.

Heimann R. B. and Kleiman J. (1988) Shock-induced growth of su- Vidal G. and Nystuen J. P. (1990) Micropaleontology, depositional perhard materials. In Crystals: Growth, Properties and Applications environment, and biostratigraphy of the upper Proterozoic Hedmark (ed. H. C. Freyhardt), pp. 1–75. Springer-Verlag, New York. Group, Southern Norway. Proterozoic Evol. Environ. 290-A, 170– Heymann D., Dressler B. O., Knell J., Thiemens M. H., Buseck P. R., 211. Dunbar R. B., and Mucciarone D. (1999) Origin of carbonaceous Vishnevsky S. A., Palchik N. A., and Raitala J. (1999) Diamonds in matter, fullerenes, and elemental sulfur in rocks of the Whitewater impactites of the Lappajarvi astrobleme (Finland). Geologiya I Group, Sudbury impact structure, Ontario, Canada. In International Geofizika 40, 1506–1510. Conference on Large Meteorite Impacts and Planetary Evolution, Wang C. Z., Ho K. M., and Chan C. T. (1993) Structure and dynamics Vol. 339 (eds. B. O. Dressler and V. L. Sharpton), pp. 345–360. of liquid carbon. Phys. Rev. B: Condens. Matter 47, 14835–14841. Geological Society of America, Boulder, CO. Wopenka B. and Pasteris J. D. (1993) Structural characterization of Hirai H., Kukino S., and Kondo K. (1995) Predominant parameters in kerogens to granulite-facies graphite; applicability of Raman micro- the shock-induced transition from graphite to diamond. J. Appl. probe spectroscopy. Am. Mineral. 78, 533–557. Phys. 78, 3052–3059. Yamada K. and Sawaoka A. B. (1994) Nucleation and growth of Hough R. M., Gilmour I., Pillinger C. T., Arden J. W., Gilkes K. W. R., diamond in detonation products. J. Am. Ceram. Soc. 77, 1104–1106. Yuan J., and Milledge H. J. (1995) Diamond and silicon-carbide in Yoo C. S. and Nellis W. J. (1991) Phase transformations in carbon impact melt rock from the Ries impact crater. Nature 378,41–44. fullerenes at high shock pressures. Science 254, 1489–1491. Hough R. M., Gilmour I., Pillinger C. T., Langenhorst F., and Mon- tanari A. (1997) Diamonds from the iridium-rich K-T boundary layer Appendix at Arroyo el Mimbral, Tamaulipas, Mexico. Geology 25, 1019– 1022. PETROLOGICAL AND MINERALOGICAL DESCRIPTIONS Huss G. R. (1990) Ubiquitous interstellar diamond and SiC in primitive OF SAMPLES STUDIED IN THE PRESENT WORK chondrites: Abundances reflect metamorphism. Nature 347, 159. Kleiman J., Heimann R. B., Hawken D., and Salansky N. M. (1984) For detailed descriptions of specimens of the specific lithologies, see Shock compression and flash heating of graphite/metal mixtures at French et al. (1997). temperatures up to 3200 K and pressures up to 25 GPa. J. Appl. Phys. 56, 1440–1454. A.1. Basement (Target Rocks, Undeformed) Koeberl C. (1993) Instrumental neutron activation analysis of geo- NGF-96-107 Amphibolite. Dark blackish-green, foliated, lineated, chemical and cosmochemical samples; a fast and reliable method for fine-grained Amph Plag-Bio rock with minor cataclastic veins. Loose small sample analysis. J. Radioanal. Nucl. Chem. 168,47–60. block, from ϳ0.5 km S of farm Sollegard, SE edge of structure. Koeberl C., Masaitis V. L., Shafranovsky G. I., Gilmour I., Langen- NGF-96-112 Amphibolite. Massive, dark blackish-green, medium- horst F., and Schrauder M. (1997) Diamonds from the Popigai grained Amph-Bio-Plag rock with highly altered Plag. Outcrop sample, impact structure, Russia. Geology 25, 967–970. Garnos [sic] road bridge over Dokkelvi River. Masaitis V. L. (1998) Popigai crater: Origin and distribution of dia- NGF-96-114 Granitic gneiss. White, massive, fine- to medium- mond-bearing impactites. Meteoritics Planet. Sci. 33, 349–359. grained, Qtz-Fsp rock, with minor Bio (Stilp?) without prominent Masaitis V. L., Shafranovsky G. I., Grieve R. A., Langenhorst F., gneissic banding. Loose block. Dokkelvi River, tunnel dump. Peredery W. V., Therriault A. M., Balmasov E. L., and Fedorova NGF-96-115 Granitic gneiss. White, streakily banded, fine-to medi- I. G. (1999) Impact diamonds in the suevitic breccias of the black um-grained, Qtz-Fsp rock with minor Bio (Stilp?), Sph, Amph (rare). member of the Onaping Formation, Sudbury Structure, Ontario, Loose block, Dokkelvi River, tunnel dump. Canada. Geol. Soc. Am. Spec. Paper 339, 317–322. NGF-96-116 Granitic gneiss. White, medium-grained Qtz-Fsp-Bio Pasteris J. D. and Chou I. M. (1998) Fluid-deposited graphitic inclu- (Stilp?) rock, with irregular streaky foliation. Loose block, Dokkelvi sions in quartz; comparison between KTB (German continental River, tunnel dump. deep-drilling) core samples and artificially reequilibrated natural NGF-96-117 Amphibolite. Massive, blackish-green, medium- inclusions. Geochim. Cosmochim. Acta 62, 109–122. grained, Amph-Plag-Qtz rock with irregular foliation. Loose block, Prosser S. J., Wright I. P., and Pillinger C. T. (1990) A preliminary Dokkelvi River tunnel dump. investigation into the isotopic measurement of carbon at the pico- NGF-96-118 Amphibolite. Massive, dark greenish-black, medium- mole level using static vacuum mass spectrometry. Chem. Geol. 83, grained Amph-Plag-Qtz rock with well-developed foliation. Loose 71–88. block, Dokkelvi River, tunnel dump. Pujols H. and Boisard F. (1970) Effects of an intense shock wave on NGF-96-119 Granitic gneiss. Massive, pinkish, medium- to coarse- graphite. Carbon 8, 781–782. grained Kfs-Qtz-Plag rock. Kfs-rich; contains coarser pegmatitic veins Reimold W. U., Koeberl C., and Bishop J. (1994) Roter Kamm impact and secondary epidote. Loose block, Dokkelvi River tunnel dump. crater, Namibia; geochemistry of basement rocks and breccias. NGF-96-134 Granitic gneiss. Massive, white, medium-grained Kfs- Geochim. Cosmochim. Acta 58, 2689–2710. Plag-Qtz rock with well developed foliations. Resembles clasts in Solovev V. A. (1976) Stacking faults: Plasticity and deformation ki- “Gardnos Breccia.” Outcrop sample, N side of Dokkelvi River, near W netics under high pressure. High Temp. High Pressure Res. 8, edge of structure. 706–707. Stoeffler D. and Grieve R. A. F. (1994) Classification and nomenclature of impact metamorphic rocks: A proposal to the IUGS subcommis- A.2. Basement, Deformed (Quartzite) sion on the systematics of metamorphic rocks. Lunar Planet. Sci. 25, NGF-96.143 Quartzite (def.) Dark gray to black, highly-fractured, 1347–1348. medium- to coarse-grained orthoquartzite. Qtz grains show highly Swart P. K., Grady M. M., and Pillinger C. T. (1983) A method for the undulose to mosaic extinction, multiple sets of parallel fractures. Minor identification and elimination of contamination during carbon isoto- associated Musc. Outcrop sample, Road cut above farm “Mexico,” S pic analyses of extraterrestrial samples. Meteoritics 18, 137–154. part of structure. Taylor S. R. and McLennan S. M. (1985) The Continental Crust: Its composition and evolution. Blackwell Scientific Publications, Ox- A.3. Lithic Breccias ford, 312 pp. NGF-96-120 Gardnos breccia. Typical Gardnos breccia, massive Thickpenny A. and Leggett J. K. (1987) Stratigraphic distribution and rock with whith granitic gneiss inclusions up to several cm in size in a palaeo-oceanographic significance of European early Palaeozoic or- pervasive black matrix. Loose block, Dokkelvi River tunnel dump. ganic-rich sediments. In Geological Society of London, Marine Stud- NGF-96-121 Gardnos breccia. Massive breccia with white granitic ies and Petroleum Geochemistry Groups, Special Meeting, Vol. 26 gneiss inclusions up to several cm in size in a black pervasive matrix (eds. J. Brooks and A. J. Fleet), pp. 231–247. Geological Society of that fills crosscutting fractures. Generally darker than normal Gardnos London. breccia; contains fewer large inclusions and a higher amount of unusu- Tucker M. E. (1983) Sedimentation of organic-rich limestones in the ally black matrix. Matrix occasionally shiny on fracture surfaces (or- late Precambrian of southern Norway. Precam. Res. 22, 295–315. ganic material?). Loose block, Dokkelvi River tunnel dump. Carbonaceous impactites from the Gardnos Crater 3903

NGF-96-126 Black-matrix breccia. Massive black/white breccia Outcrop sample, from ϳ1 m above basement (Gardnos breccia), Flat- containing scattered cm-size granite gneiss fragments in a dark black dalselvi River, W part of structure. matrix. Breccia area cut by 2-cm “dike” of aphanitic brownish-green NGF-96-175 Melt-matrix breccia. Massive dense green, clast-rich material. (Analyzed material came from dark breccia.) Loose block, breccia, with cm-size rock clasts and mm-size mineral clasts in Dokkelvi River tunnel dump. massive greenish matrix. Outcrop sample, from slightly about base- NGF-96-176 Black-matrix breccia. Typical massive black/white ment/suevite contact. Dokkelvi River bed, ϳ75 m W (upstream) breccia, with scattered cm-size granite gneiss fragments in a dark gray from footbridge. to black matrix. Near contact between basement and overlying suevite; NGF-96-178 Suevite. Massive, very coherent, dark black breccia possibly an included block in suevite(?) Outcrop sample, Dokkelvi with irregular glass clasts typically Ͻ 1 cm in size, together with River bed, ϳ75 m W (upstream) of footbridge. mm-size rock and mineral clasts, in a very dense black matrix. Loose block, NE bank of Dokkelvi River near footbridge. NGF-96-179 Melt-matrix breccia. Massive dense, green rock, with A.4. Melt-Bearing Breccias scattered clasts up to 2 to 4 cm in a dense, aphanitic matrix with a NGF-96-133 Suevite. Masssive, dark green breccia containing irreg- greasy luster. Outcrop sample (from J. Naterstad, Univ. of Oslo). ular cm-size black glassy inclusions. Local block, from near outcrop, Dokkelvi River bed, ϳ75 m W (upstream) of footbridge. Same general Dokkelvi River bed, at footbridge, slightly above basement/suevite area as samples 175 to 178. contact. Abbreviations: Qtz ϭ quartz; Fsp ϭ feldspar; Kfs ϭ K-feldspar; NGF-96-137 Suevite Massive, dark-green breccia with irregular Plag ϭ plagioclase; Amph ϭ amphibole; Bio ϭ biotite; Chl ϭ chlorite; cm-size glassy inclusions (ϳ10–15%), occasional maﬁc inclusions. Epid ϭ epidote; Musc ϭ muscovite; Stilp ϭ stilpnomelane.