Blackwell Publishing AsiaMelbourne, AustraliaIAR Island Arc1038-48712006 Blackwell Publishing Asia Pty LtdMarch 2006151143155Thematic Article Tectonic control of bioalterationH. Furnes et al.
Island Arc (2006) 15, 143–155
Thematic Article Tectonic control of bioalteration in modern and ancient oceanic crust as evidenced by carbon isotopes HARALD FURNES,1,* YILDIRIM DILEK,2 KARLIS MUEHLENBACHS3 AND NEIL R. BANERJEE4 1
Department of Earth Science, University of Bergen, Allegt. 41, 5007 Bergen, Norway (email:
[email protected]), 2Department of Geology, Miami University, Oxford, OH 45056, USA, 3Department of Earth and Atmospheric Sciences, University of Alberta, Edmonton, AB T6G OE2, Canada and 4Department of Earth Sciences, University of Western Ontario, London, Ontario, N6A 5B7, Canada
Abstract We review the carbon-isotope data for finely disseminated carbonates from bioaltered, glassy pillow rims of basaltic lava flows from in situ slow- and intermediatespreading oceanic crust of the central Atlantic Ocean (CAO) and the Costa Rica Rift (CRR). The δ13C values of the bioaltered glassy samples from the CAO show a large range, between −17 and +3‰ (Vienna Peedee belemnite standard), whereas those from the CRR define a much narrower range, between −17‰ and −7‰. This variation can be interpreted as the product of different microbial metabolisms during microbial alteration of the glass. In the present study, the generally low δ13C values (less than −7‰) are attributed to carbonate precipitated from microbially produced CO2 during oxidation of organic matter. Positive δ13C values >0‰ likely result from lithotrophic utilization of CO2 by methanogenic Archaea that produce CH4 from H2 and CO2. High production of H2 at the slow-spreading CAO crust may be a consequence of fault-bounded, high-level serpentinized peridotites near or on the sea floor, in contrast to the CRR crust, which exhibits a layer-cake pseudostratigraphy with much less faulting and supposedly less H2 production. A comparison of the δ13C data from glassy pillow margins in two ophiolites interpreted to have formed at different spreading rates supports this interpretation. The Jurassic Mirdita ophiolite complex in Albania shows a structural architecture similar to that of the slow-spreading CAO crust, with a similar range in δ13C values of biogenic carbonates. The Late Ordvician Solund–Stavfjord ophiolite complex in western Norway exhibits structural and geochemical evidence for evolution at an intermediate-spreading mid-ocean ridge and displays δ13C signatures in biogenic carbonates similar to those of the CRR. Based on the results of this comparative study, it is tentatively concluded that the spreading rate-dependent tectonic evolution of oceanic lithosphere has a significant control on the evolution of microbial life and hence on the δ13C biosignatures preserved in disseminated biogenic carbonates in glassy, bioaltered lavas. Key words: bioalteration, carbon isotopes, Costa Rica Rift, lavas, microbial life, mid-ocean ridges, Mirdita ophiolite, oceanic crust, ophiolites, Solund–Stavfjord ophiolite, spreading rates, upper oceanic crust.
INTRODUCTION Until recently, only sediments have been considered to be suitable habitats for microbial activity. *Correspondence. Received 18 November 2005; accepted for publication 03 December 2005. © 2006 The Authors Journal compilation © 2006 Blackwell Publishing Asia Pty Ltd
However, it is now well known that microbial life has also colonized the Earth’s crust to considerable depths, where carbon and energy sources are available and where physical conditions can support life (e.g. Lovley & Chapelle 1995; Pedersen 1997; Pedersen et al. 1997; Amend & Teske 2005; Schippers et al. 2005). During the past decade, it has also been demonstrated that the upper volcadoi:10.1111/j.1440-1738.2006.00516.x
144 H. Furnes et al.
nic part of the in situ oceanic crust is a habitat for microbial life (e.g. Thorseth et al. 1992, 1995, 2001, 2003; Furnes et al. 1996, 1999, 2001a,b; Fisk et al. 1998, 2003; Torsvik et al. 1998; Furnes & Staudigel 1999; Banerjee & Muehlenbachs 2003; Staudigel & Furnes 2004; Staudigel et al. 2004). The microbial colonization of oceanic volcanic rocks is best observed along fractures in the glassy selvages of pillow lavas, leaving behind biosignals that reveal their former presence. The biosignals are marked by biogenic alteration textures at the interface between fresh and altered glass. Furnes and Staudigel (1999) have shown that various bioalteration textures exist in the glassy rims of pillow lavas occurring at a depth of 500 m in the extrusive sequence drilled in Deep Sea Drilling Program– Ocean Drilling Project (DSDP–ODP) Hole 504B at the Costa Rica Rift (CRR). The biogenically altered material that contains these alteration textures is characterized by elevated element concentrations of C, N, P, K and S, and a wide range of δ13C values. The carbon of the putative life may come from either dissolved organic matter in pore waters or from reduction of dissolved CO2. The δ13C values found in the bioaltered, chilled pillow margins to a large extent fall outside the range that can be easily explained in terms of abiological processes. The observed differences in the structural architecture of the oceanic crust may control the diversity of microbial life through the availability and production of different energy sources. At slow-spreading ridges, pillow lavas locally rest directly on crustal ultramafic rocks and/or upper mantle peridotites, which have been exhumed during periods of amagmatic extension (Cannat et al. 1995; Lagabrielle et al. 1998). The H2 that is produced during alteration of the ultramafic rocks may serve as an energy source for microbial activity. This results in higher abiotic production of H2 in the slow-spreading crust, where extensive serpentinization occurs as a result of the penetration of seawater into the lower crust and upper mantle along deep-seated extensional normal faults (Mevel et al. 1991; Cannat et al. 1995; Lagabrielle et al. 1998). The availability of abiotic H2 in slow-spreading environments may hence control the occurrence and diversity of microbial life colonizing volcanic rocks in the upper oceanic crust. The absence of exhumed serpentinized mantle at intermediate- to fastspreading ridges, in contrast, results in less H2 and lower activity of methanogenic microorganisms in this environment. © 2006 The Authors Journal compilation © 2006 Blackwell Publishing Asia Pty Ltd
In the present paper, new δ13C data are reported and the published values for pillow lavas from modern oceanic crust drilled in the slow-spreading Mid-Atlantic Ridge (MAR) and the intermediatespreading CRR are reviewed. The δ13C values of volcanic rocks from two ophiolite complexes are then examined, the evolution of which is well constrained based on extensive systematic structural field and geochemical studies. This comparative examination of the potential links between bioalteration products and the structural architecture of oceanic crust in modern and ancient sea floor spreading environments suggests that the preserved C isotopes are correlated to the spreading rate.
BIOALTERATION OF UPPER OCEANIC CRUST AND C-ISOTOPIC SIGNATURE Recent studies of the volcanic sequence of in situ and fossil oceanic crust have shown textural evidence for bioalteration within the glassy margins of pillow lavas (Fig. 1), whereas the crystalline interior parts of these lava flows lack any biogenic textures (Thorseth et al. 1995; Furnes et al. 1996, 2001b). Microbes living on and within volcanic glass commonly leave behind metabolic biosignatures recorded in the pillow margins. In contrast, the crystalline interior of pillow lavas has isotope values bracketed between typical magmatic basalt values (inherited from mantle CO2) and inorganically precipitated carbonate (in veins and amygdales). These contrasting C-isotopic values from pillow rims and interiors are extremely important in investigating the activity of microbial
Fig. 1 Photomicrograph showing bioalteration textures (GT, granular texture; TT, tubular textures) in a basaltic glass (FG, fresh glass) sample from a pillow lava rim. ODP sample 148-896A-9R-1, 17–21 cm.
Tectonic control of bioalteration 145
life in the uppermost in situ oceanic crust and ancient ophiolites (Furnes et al. 2002, 2004, 2005; Banerjee et al. 2006). Biotic and/or abiotic processes can fractionate carbon isotopes (Oremland 1988). Marine carbonates and biogenic organic matter, representing the largest carbon reservoirs on Earth, show isotopic compositions with mean δ13C values of approximately 0‰ and −25‰, respectively (Hoefs 1997). The δ13C value of mantle CO2, extracted from fresh basalts, ranges from approximately −5 to −7 (Alt et al. 1996; Hoefs 1997). The most efficient metabolism by microorganisms is to obtain carbon and energy from the oxidation of organic matter. During respiration, organic material is oxidized to CO2, which may subsequently bind with Ca, Mg and Fe ions within the rock matrix, resulting in the precipitation of carbonate minerals. As a consequence of this process, biogenic fractionation leads to the generation of 12C-enriched CO2, and hence the resulting carbonate mineral becomes depleted in δ13C. In contrast, positive δ13C signatures may result from lithotrophic utilization and reduction of CO2 in which methanogenic Archaea produce methane (CH4) from H2 and CO2. These lithoautotrophic ecosystems require abiologically produced H2 to provide substantial energy for various electron acceptors (Amend & Shock 2001). During this process, an extreme isotopic fractionation enriches the methane in 12C (Kelts & McKenzie 1982). The remaining CO2 (to form carbonate) becomes enriched in 13C (because of the loss of the 12Cenriched methane), which leads to 13C-enriched carbonate. The occurrence of diagenetic dolomite with δ13C as high as +14 has been reported from sediments recovered from DSDP–ODP Hole 479 (Gulf of California), suggesting a biogenic CO2 reservoir related to active methanogenesis (Kelts & McKenzie 1982). Similarly, pore water measured in anoxic sediments from a number of DSDP holes also show δ13C values in the range of +1 to +14 (Hoefs 1997). Figure 2 shows a sketch that explains the opposite δ13C paths generated by different microbial metabolism, as well as the δ13C range of fresh basalts and diagenetic marine carbonates.
PRODUCTION OF ABIOTIC HYDROGEN Abiotic H2 of crustal origin can be produced in different ways, that is: (i) by serpentinization and oxidation of Fe2+-bearing minerals; (ii) by forma-
δ
Fig. 2 Diagram depicting the opposite δ13C paths generated by different microbial metabolism.
tion of FeS2 from FeS; (iii) by fracture-induced reduction of water; and (iv) by radiolysis of water (see review by Lin et al. 2005). In the case of the rocks in question, it is the first alternative that is the most relevant, and it is briefly discussed below. During hydration and transformation of olivine to serpentine, Fe2+ in olivine oxidizes to Fe3+, which then forms magnetite, along with concomitant reduction of water to molecular hydrogen (H2) (e.g. Berndt et al. 1996; Holm & Charlou 2001). This alteration of olivine-rich rocks (mantle and lower oceanic crust) takes place at pH conditions in the range 5–11 (Holm & Charlou 2001; Stevens & McKinley 2001), as documented from ocean floor serpentinites (Haggerty 1991), the Oman ophiolite (Neal & Stanger 1983) and deep aquifers in the Columbia River basalt (Stevens & McKinley 1995; Stevens & McKinley 2001). The relevant reactions can be expressed as follows: Olivine (5Mg2SiO4 + Fe2SiO4) + water (9H2O) ⇒ serpentine (3Mg3Si2O5(OH)4) + brucite (Mg(OH)2) + iron (II) hydroxide 2Fe(OH)2, and further 3Fe(OH)2 ⇒ magnetite Fe3O4 + H2 + 2H2O. Studies by Kelley et al. (1993, 2005) and Kelley (1996) infer that significant CH4 and H2 production may result from serpentinization of upper mantle peridotites as a result of seawater penetrating along deep-seated faults. The production of H2 and CH4 as a result of serpentinization of ultramafic rocks is presently occurring in the Zambales ophiolite (Abrajano et al. 1990). In addition, H2 may also be produced to a lesser extent by the oxidation of ferromagnesian silicates (e.g. pyroxenes) during the alteration of fresh basalt, as shown by Stevens and McKinley (1995) from the Columbia River basalts. The oceanic crust at magma-poor slowspreading systems is more pervasively deformed because of extensive faulting in comparison to oce© 2006 The Authors Journal compilation © 2006 Blackwell Publishing Asia Pty Ltd
146 H. Furnes et al.
anic crust evolving at fast-spreading ridges, where a high heat budget and magmatically compensated extension preclude widespread, deep-seated faulting (Mevel et al. 1991; Alt & Teagle 2000; Kelley & Früh-Green 2000; Manning et al. 2000). Thus, it is inferred that the H2 production is greater at slowspreading ridges where tectonic extension dominates than at intermediate- to fast-spreading ridges, where magmatic output is high at a steadystate rate. The structure of the oceanic crust in different tectonic settings can therefore result in variable degrees of serpentinization and H2 production, which in turn have an important effect on the style of microbial colonization.
STRUCTURE OF THE MODERN OCEANIC CRUST The structure and architecture of modern oceanic crust is controlled mainly by the spreading rate, magma supply and thermal regime beneath and along the spreading axes (e.g. Macdonald 1982, 1983; Phipps Morgan et al. 1994; Dilek et al. 1998). Therefore, the oceanic crust developed at midocean ridges with different spreading rates (slow vs fast) displays a great variability in the mode and nature of magmatic, tectonic and hydrothermal activities (Alt & Teagle 2000; Humphris & Tivey 2000; Kelley & Früh-Green 2000; Manning et al. 2000; Robinson et al. 2000). The magma budget along the slow-spreading MAR is highly variable. The slow-spreading MAR oceanic crust displays significant changes in crustal thickness to the order of 1–4 km, both along and across the axis, and is disrupted tectonically near the segment boundaries (Detrick et al. 1995; Cannat 1996). During periods of tectonic extension, in the absence of a steady-state magma supply, lower crustal and upper mantle rocks become exposed on the sea floor (Cannat et al. 1995; Karson 1998; Lagabrielle et al. 1998). These exhumed upper mantle and lower crustal rocks are locally covered by pillow lava flows near and within the neovolcanic zone along the ridge axis, indicating renewed phases of volcanism following amagmatic periods of sea floor spreading (Macdonald 1983; Mevel et al. 1991; Cannat et al. 1995). In the present paper, the definitions of Dilek (2003) are adopted, where incomplete, disrupted oceanic crust has been defined as ‘Hess-type’ oceanic crust, in contrast to an idealized, complete pseudostratigraphy termed ‘Penrose-type’ oceanic crust, which evolved at magma-rich, fast-spreading centers. At slow-spreading ridges (e.g. ∼2.5 cm/year), the © 2006 The Authors Journal compilation © 2006 Blackwell Publishing Asia Pty Ltd
majority of volcanic products consist of pillow lavas, with much less common sheet flows. At intermediate- to fast-spreading ridges the magma budget is consistently high, and hence magmatic construction keeps pace with sea floor spreading without major tectonic disruption (e.g. Karson & Rona 1990; Sinton & Detrick 1992; Phipps Morgan et al. 1994). Faulting is less extensive and commonly limited to the upper oceanic crust, in which the extensional normal faults are generally planar in geometry with steep dip angles (Dilek et al. 1998). The modern oceanic crust developed at intermediate- to fast-spreading ridges commonly has a complete, Penrose-type pseudostratigraphy with gradational igneous contacts. Sheeted dykes are well developed with transitional contacts with the overlying extrusive sequence and the underlying plutonic complex (Dilek 1998). At intermediate to fast-spreading ridges (e.g. ∼6–12 cm/year) massive sheet flows are widespread, and the extrusive rocks are locally intensely jointed and fractured and may be crosscut in places by normal growth faults (e.g. Bicknell et al. 1987; Bonatti & Harrison 1988; Delaney et al. 1992; Alt 1995; Fouquet et al. 1996; Macdonald 1998). In summary, the modern oceanic crust appears to be more disrupted by crustal- to lithosphericscale normal faults at slow-spreading ridges (i.e. MAR, Southwest Indian Ocean Ridge) in comparison to the relatively less deformed oceanic crust evolved at intermediate- to fast-spreading ridges (i.e. CRR and East Pacific Rise). Serpentinized upper mantle rocks, which have been tectonically and diapirically transported to shallow crustal levels and/or exposed on the sea floor (e.g. Dick 1989; Karson 1990; Mevel et al. 1991; Cannat et al. 1995), are expected to be common features in the slow-spreading Atlantic Ocean. Figure 3 shows generalized structural sections of the modern oceanic crust observed along the slow-spreading MAR and the intermediate-spreading CRR.
STRUCTURAL AND CARBON ISOTOPE PROFILES IN OCEANIC CRUST The internal structure and stratigraphy of the investigated in situ and ancient oceanic crust are described below to show the fundamental differences in their crustal architecture. The oceanic crust generated at the MAR is a typical example of a heterogeneously deformed slow-spreading oceanic crust, whereas the oceanic crust evolved at
Tectonic control of bioalteration 147 IN SITU OCEANIC CRUST
the intermediate-spreading CRR is relatively uniform in its internal structure and is more analogous to the fast-spreading oceanic crust of the East Pacific Rise. In the present bioalteration study, volcanic rocks from in situ oceanic crust in these two tectonic settings are examined because of the wealth of data and observations available. For the ancient oceanic crust, two different ophiolites have been chosen, the Jurassic Mirdita ophiolite in Albania and the Late Ordovician Solund–Stavfjord ophiolite in western Norway, for which the tectonic environment of formation is well established in the literature.
°
(a)
(b)
°
The slow-spreading MAR samples were collected from DSDP cores at the Reykjanes Ridge (Holes 407, 408, 409) and the north-central (Holes 410A, 411, 412A), central (Holes 396B and 648B) and western parts (Hole 418A) of the Atlantic Ocean. The rock samples range in age from the Quaternary (Hole 648B) to the Early Cretaceous (Hole 418A). The depth of drillcore penetration into the volcanic basement in the central Atlantic Ocean varies between approximately 33 and 555 m. Pillow lavas make up the dominant component of extrusive rocks drilled in these holes. The core recovery is in general best in the oldest rocks and ranges from 15% in the youngest (in Hole 648B) to 72% in the oldest (in Hole 418A). For further details (geographical, description etc.) of the samples, see Table 1 and Furnes et al. (2001a). The full spreading rate at the MAR is 20–40 mm/ year (Bellaiche et al. 1974; Ballard & van Andel 1977). The samples from the intermediate-spreading CRR are from DSDP/ODP Hole 504B and ODP Hole 896A (Alt et al. 1996). The east–westtrending CRR has an intermediate, full-spreading rate of 72 mm/year to the south and 60 mm/year to the north (Hey et al. 1977). The age of the crust drilled in Hole 504B is 5.9 my, and pillow lavas and massive sheet flows occur approximately in the proportions of 60% and 40%, respectively, as observed in the drillcores (see Alt et al. 1993 and references therein). The core recovery is rather
¢
¢
Fig. 3 Simplified cross-sections from (a) the slow-spreading MidAtlantic Ridge oceanic crust (modified from Karson & Winters 1992; Dilek et al. 1998) and (b) the intermediate-spreading Costa Rica Rift oceanic crust (from Dilek 1998).
Table 1 Sample information of the drillholes from which carbon-isotope data were collected DSDP/ODP Leg (Site)
Atlantic Ocean 106/109 (648B) 49 (411) 412A 49 (409) 49 (410A) 46 (396B) 49 (408) 49 (407) 51/52/53 (418A) Costa Rica Rift 69/70/83/111/137 /140/148 (504B) 148 (896A)
Age (Ma)
Percent core recovery
Water depth (m)
Maximum depth (mivb)
Sediment cover (m)
Reference
Quaternary 1 1.6 2.3 10 10 20 38 110
15 25 17 30 47 23 61 40 72
3326 1935 2609 832 2977 4459 1624 2472 5511
33.3 45.5 131 239 51 255 39 158 555
0 74 163 80 331 150.5 324 300.5 313
Detrick et al. (1988) Luydendyk et al. (1978) Luydendyk et al. (1978) Luydendyk et al. (1978) Luydendyk et al. (1978) Dmitriev et al. (1978) Luydendyk et al. (1978) Luydendyk et al. (1978) Robinson et al. (1979)
5.9
22
3463
1836.5
274.5
Alt et al. (1993)
5.9
27
3448
273.9
195.1
Alt et al. (1993)
DSDP, Deep Sea Drilling Program; mivb, meters into volcanic basement; ODP, Ocean Drilling Program.
© 2006 The Authors Journal compilation © 2006 Blackwell Publishing Asia Pty Ltd
148 H. Furnes et al.
low, approximately 22–27% (Table 1). The southern flank of the ridge is defined by a basement topography characterized by east–west-trending abyssal hills and intervening troughs (e.g. Alt et al. 1993). These steeply north-dipping asymmetric escarpments have been interpreted to represent tilted fault blocks, with normal-fault displacements in the order of 100 m or more (e.g. Alt et al. 1993; Dilek 1998). OBDUCTED ANCIENT OCEANIC CRUST OPHIOLITES
The Jurassic Mirdita ophiolite in Albania occurs within the Dinaride–Hellenide mountain belt of the Alpine-Himalayan orogenic system. The Mirdita ophiolite has been subdivided into western (WMO) and eastern (EMO) type complexes based on the rock types and their geochemical characteristics, as established in previous studies (e.g. Shallo 1992, 1995; Robertson & Shallo 2000; Bortolotti et al. 2002; Shallo & Dilek 2003; Dilek et al. 2005). Although the basaltic rocks in the WMO have suffered prehnite–pumpellyite grade metamorphism, samples were collected containing fresh volcanic glass. The approximately 6-km-thick WMO contains 650 m of extrusive rocks composed mainly of pillow basalts. A well-developed sheeted dyke complex is absent in the WMO, although dyke swarms are locally observed feeding into the overlying volcanic rocks. Pillow lavas commonly rest directly on gabbros and/or mantle lherzolites and harzburgites (Nicolas et al. 1999; Dilek et al. 2005). The general lack of a sheeted dyke complex, the predominance of pillow lavas, and the Hess-type oceanic crustal architecture (Dilek 2003) in the WMO (Fig. 4a) collectively suggest that this (a)
(b)
Fig. 4 Cross-sections from (a) the Jurassic Mirdita ophiolite complex (Albania) (modified from Shallo & Dilek 2003) and (b) the Late Ordovician Solund–Stavfjord ophiolite complex (modified from Muehlenbachs et al. 2003). © 2006 The Authors Journal compilation © 2006 Blackwell Publishing Asia Pty Ltd
ancient oceanic crust was likely generated at a slow-spreading system in the Pindos-Mirdita Basin (Nicolas et al. 1999; Dilek & Flower 2003; Shallo & Dilek 2003; Dilek et al. 2005). The approximately 10-km-thick EMO contains approximately 1.1 km of an extrusive sequence consisting of basalt, basaltic andesite, andesite and dacite/rhyolite. This volcanic sequence is underlain by a welldeveloped sheeted dyke complex, and the entire EMO displays a Penrose-type pseudostratigraphy reminiscent of the fast-spreading East Pacific Rise oceanic crust (Dilek et al. 2005). The Late Ordovician Solund–Stavfjord ophiolite complex (SSOC) in the Norwegian Caledonides has a well-preserved volcanic sequence of basaltic pillow lavas, sheet flows and volcanic breccias, a sheeted dyke complex, and high-level gabbros (Furnes et al. 1990). The ocean floor hydrothermal alteration history of this ophiolite is well preserved in spite of the superimposed regional greenschist facies metamorphism that affected the entire Caledonide Orogenic Belt (Fonneland et al. 2005). A geological cross-section of the SSOC is shown in Figure 4b. Structural and geochemical features of the sheeted dyke complex and dyke swarms intruding the plutonic complex indicate that the SSOC developed during at least two discrete sea floor spreading events (Dilek et al. 1997; Furnes et al. 1998) in a marginal basin, analogous to the modern Andaman Sea (Furnes et al. 1990, 2000). The high proportion of sheet flows, the internal structure of the sheeted dyke complex, and the high proportion of ferrobasalts collectively suggest that the SSOC was generated at an intermediate- to fast-spreading mid-ocean ridge similar to the CRR or the Juan de Fuca Ridge (Dilek et al. 1997; Dilek 1998; Furnes et al. 1998, 2003, 2006). An estimate of the spreading rate at which the SSOC developed, based on the proportion of pillow lavas to sheet flows and the their Zr/ Y-Zr relationships, would indicate an average fullspreading rate of approximately 100 mm/year (Furnes et al. 2006). CARBON-ISOTOPIC BIOSIGNATURES
A large number of pillow lava samples from the modern and fossil oceanic crust described above were analyzed. Subsamples were prepared from the outermost glassy rim and the crystalline interior of the pillows to examine their bulk-rock carbonate for carbon-isotope ratios to see whether these rocks show any indication of biological fractionation. Prior to crushing, the samples were
Tectonic control of bioalteration 149
carefully trimmed with a saw to remove any trace of surface contamination, and those containing open fractures and pore spaces were avoided. Stable isotope analyses of carbonates were performed by pouring 100% phosphoric acid on whole-rock powders under vacuum (McCrea 1950) and analyzing the exsolved CO2 on a Finnigan MAT 252 mass spectrometer at the University of Alberta. The errors on isotope analyses are better than ±0.1‰. The data are reported in the usual delta-notation with respect to the Vienna Peedee belemnite standard (Craig 1957, 1961).
chilled margin and for the crystalline interior parts of pillow lavas from the Jurassic Mirdita and the Late Ordovician Solund–Stavfjord ophiolites are shown in Figure 6. The majority (∼90%) of the samples from the crystalline pillow centers from both ophiolite complexes define δ13C values between −8 and 0. The δ13C values of the glassy pillow rims from the WMO define an even distribution between −22 and +1.5 (Fig. 6a), whereas those from the Solund–Stavfjord ophiolite mostly define a range between −16 and −2.5, with the exception of one sample (Fig. 6b).
In situ oceanic crust
The relationships between the δ13C(carbonate) and wt% carbonate for the glassy, bioaltered chilled margin and the complementary crystalline interior part of pillow lavas are shown in Figure 5. The crystalline interior parts of the pillow lavas from the central MAR and the CRR show δ13C values between −7 and 0, only rarely lower than −7. The bioaltered, glassy samples from the slow- and intermediate-spreading oceanic crust show, in contrast, marked differences between their δ13C values. The samples from the central MAR show a large range in δ13C values between approximately −17 and +3 (Fig. 5a), whereas those from the CRR define a much narrower range between approximately −17 and −7 (Fig. 5b). Ophiolites
The relationships between the δ13C(carbonate) and wt% carbonate for the originally glassy, bioaltered
DISCUSSION AND CONCLUSIONS The large range in δ13C values for the pillow lavas (Figs 5,6) could represent the sum of different sources, identified as: (i) the primary mantle source (−5 to −7‰); (ii) marine carbonates (near 0‰); and (iii) different metabolic byproducts of microorganisms that may drive the δ13C values in opposite directions, as shown in Figure 2. Because textural evidence of bioalteration was only found in the glassy rim of pillows, all these samples were isolated. Furthermore, in order to avoid samples with marine carbonate precipitation, as in vesicles or vein fillings, samples with more than 1 wt% carbonate were omitted. The result, shown in Figure 7, demonstrates different δ13C distributions for the slow-spreading oceanic crust of the MAR and the Mirdita ophiolite vs the intermediate-spreading oceanic crust of the CRR and the Solund–Stavfjord ophiolite. For the in situ MAR
(a) Mirdita Ophiolite Complex
(a) Atlantic Oceanic Crust 100 Crystalline Glassy
10
% carbonate
%carbonate
100
1 0,1 0,01 -20
-15
-10
-5
0
10 1 0,1 0,01 0,001 -25
5
δ13 C (carbonate)
Crystalline Glassy
-20
-10
-5
0
5
0
5
δ 13 C (carbonate) (b) Solund-Stavfjord Ophiolite Complex
(b) Costa Rica Rift Oceanic Crust 10
1
% carbonate
%carbonate
-15
1 0,1 Crystalline Glassy
0,01 0,001 -20
-15
-10
-5
0
5
δ13 C (carbonate)
Fig. 5 Relationships between δ13C and wt% carbonate for the glassy and crystalline interior of pillow lava samples from (a) the Mid-Atlantic Ridge oceanic crust and (b) the Costa Rica Rift oceanic crust. Data sources: Atlantic oceanic crust, J. French (pers. comm., 1998) and Furnes et al. (2001a); Costa Rica Rift, Furnes et al. (1999).
Crystalline Glassy 0,1 0,01 0,001 -25
-20
-15
-10
-5
δ 13 C (carbonate)
Fig. 6 Relationships between the δ13C and wt% carbonate for the glassy and crystalline interior of pillow lava samples from (a) the Mirdita ophiolite complex and (b) the Solund–Stavfjord ophiolite complex. Data sources: Mirdita ophiolite, N. R. Banerjee (unpubl. data, 2006); Solund– Stavfjord ophiolite, Furnes et al. (2002). © 2006 The Authors Journal compilation © 2006 Blackwell Publishing Asia Pty Ltd
150 H. Furnes et al. (a) % carbonate
1
0,1
0,01
0,001 -25
Glassy Mirdita Ophiolite Glassy Atlantic Oceanic Crust
-20
-15
-10
-5
0
5
δ13 C (carbonate)
(b)
% carbonate
1
0,1
0,01 Glassy - Solund-Stavfjord Ophiolite Glassy - Costa Rica Rift Oceanic Crust
0,001 -25
-20
-15
-10
-5
0
5
δ13 C (carbonate)
Fig. 7 Comparison between δ13C and
