Geochemistry of hydrothermally altered oceanic crust: DSDP/ODP Hole 504B – Implications for seawater-crust exchange budgets and Sr- and Pb-isotopic evolution of the mantle



[1] This paper presents petrographic, chemical, and isotopic (Sr, S) analyses of whole rock samples from a 1.8 km section of upper ocean crust (DSDP/ODP Hole 504B). The samples were selected to cover all lithologies (pillows, flows, breccias, dikes) and alteration/mineralization styles. The chemical and petrographic data were used to calculate weighted averages for upper crustal composition, based on which seawater-ocean crust exchange fluxes were calculated. These results confirm earlier estimates that identify the upper crust as a significant sink for K and Mg and a source of Ca and Si to the oceans. Changes in trace element geochemistry implies that the upper ocean crust in 504B is a sink for CO2, Rb, Cs, and U, although the flux rates are an order of magnitude smaller than suggests by previous estimates for DSDP Sites 417 and 418 in 118 Ma Atlantic crust. Fluxes of these components are similar, within a factor of four, to flux rates estimated for the Juan de Fuca Ridge flank, which may relate to similarities in the thermal and hydrogeological evolution at both sites that is controlled by rapid termination of fluid circulation and conductive reheating of the upper crust. The contrast between the fluxes of trace elements derived for those settings and the open-ocean sites 417/418 likely reflects prolonged fluid-rock interaction at the latter location. If the Mg uptake and Sr exchange reconstructed from 504B core is representative, ridge flank hydrothermal alteration cannot account for the imbalance in the Mg and Sr budgets of the oceans. Up to 10% of the crustal Pb resides in the mineralized parts of the transition zone between the volcanic section and the sheeted dike complex. Combined, the Pb mobilized in the deepest parts of the hydrothermal systems (probably not penetrated in 504B) and hosted in metalliferous sediments and mineralized stockwork may account for the Pb surplus of the continental crust and the evolution of Ce/Pb of the mantle. Hydrothermal alteration results in net increases of Rb/Sr and U/Pb, in particular in the uppermost 600 m of crust, but the increases are not large enough to make altered upper ocean crust a plausible precursor for the HIMU mantle component. Moreover, the fractionation between Th and Pb, if any, is insufficient to account for the development of highly radiogenic 208Pb/204Pb in a HIMU mantle source. Potential HIMU precursors can be derived from altered ocean crust after 1–2 Ga, if on the order of 80–90% Pb, 40–55% Rb, 40% Sr, and 35–40%U are removed during partial dehydration in subduction zones.

1. Introduction

[2] Circulation of seawater and water-rock interaction within the ocean crust have important consequences for the transport of energy and mass on Earth. Chemical exchange between seawater and the crust influence ocean chemistry and composition of ocean crust that is ultimately returned into the mantle. Subduction and partial dehydration/melting of altered ocean crust is one of the principal fractionation mechanisms on Earth and profoundly influences the chemical and isotopic evolution of the continental crust and the mantle [e.g., Hofmann and White, 1982]. Physical properties of the lithosphere, such as porosity, permeability, magnetization, and sonic velocity, are also strongly affected by hydrothermal processes. Drill holes provide an opportunity to sample vertical sections through the oceanic crust and examine how hydrothermal alteration processes change the chemical and physical characteristics of the oceanic crust during aging.

[3] Here we present a comprehensive set of chemical (major elements, trace elements, volatiles, and Sr- and S-isotopes) and petrographic data for the deepest drill hole within ocean crust. These data are used to deduce chemical fluxes between oceans and crust and to quantify chemical changes during crustal aging that will affect geochemical cycling in subduction zones. Many aspects of the geochemistry of hydrothermal alteration in Hole 504B were reviewed by Alt et al. [1996a]. We focus on trace elements (Cs, Rb, U, Th, Pb) for which geochemical budgets were previously unavailable (Pb) or restricted to the volcanic section at DSDP Sites 417/418 in 118 Ma North Atlantic crust [Hart and Staudigel, 1982; Staudigel et al., 1995; Staudigel et al., 1996]. A companion paper focuses on Re-Os and PGE systematics of 504B and 417/418 (B. Peucker-Ehrenbrink, W. Bach, S.R. Hart, J.S. Blusztajn, and T. Abbruzzese, Rhenium-osmium isotope systematics and platinum group element concentrations in oceanic crust from DSDP/ODP Sites 504 and 417/418, manuscript submitted to Geochemistry, Geophysics, Geosystems, 2002, hereinafter referred to as Peucker-Ehrenbrink et al., submitted manuscript, 2002).

2. Hole 504B

[4] Hole 504B is located in ∼6.6 Ma intermediate-spread crust, 200 km south of the Costa Rica Rift Zone in the equatorial East Pacific (Figure 1). It is the deepest drill hole in the oceanic crust (2111 meters below seafloor; mbsf) and serves as a reference for the structure and petrology of upper ocean crust. Drilling and research history of Hole 504B comprises 20 years of intense technical and scientific effort during which indispensable information on the petrology, geochemistry, and physics of the upper oceanic crust has been collected. Figure 2 presents the principal lithostratigraphy of basaltic basement in Hole 504B. The drilled section comprises 274.5 m of pelagic sediments (mainly siliceous ooze), 571.5 m volcanic zone (VZ) of pillowed and massive basalt flows as well as interpillow breccia, a 209-m thick transition zone (TZ) of abundant dikes as well as pillows and flows, and finally >1045 m of sheeted dike complex (SDC). While ∼5% of the recovered rocks from the VZ and the SDC show evidence of brecciation, the percentage of dislocation breccias is unusually high in the TZ that may hence represent a long-lived fault system [Agar, 1991]. A stockwork-like metal sulfide mineralization zone is developed within the transition zone between 910 and 928 mbsf [Honnorez et al., 1985].

Figure 1.

Location map of Hole 504B, 200 km south of the Costa Rica Rift (CRR) in the eastern equatorial Pacific. EFZ and PFZ are the Ecuador and Panama Fracture Zones bounding the CRR.

Figure 2.

Overview of the lithostratigraphy of Hole 504B. Scalloped pattern represents pillow basalt, stitched pattern represents massive flows, and vertical stripes represent dikes. Sample coding is as follows: black, volcanic zone; orange, transition zone; green, sheeted dike complex; circles, pillow basalt; triangles, breccias; squares, flows; diamonds, diabase. Permeabilities are from Fisher [1998] and references therein.

[5] Rocks recovered from Hole 504B reveal a complex multistage alteration history (see reviews in [Alt et al., 1996a; Alt et al., 1996c]). Alteration is most intense in brecciated lithologies and along flow boundaries and fractures but is minor in massive rock. Overall, there is a transition to higher alteration temperatures and lower water-to-rock ratios with increasing depth. Low-temperature oxidative alteration is restricted to the uppermost 200–300 m of basement where permeability is greatest (Figure 2). The lowermost part of the VZ is characterized by non-oxidative alteration under elevated, but still low (<150°C) temperatures, reflecting circulation of chemically evolved seawater. The steep transition to higher alteration temperatures (>250°C), indicated by the appearance of chlorite, albite, actinolite, and titanite, is developed within the upper part of the TZ. The greenschist facies mineral assemblage prevails down to ∼1600 mbsf, below which hornblende and Ca-rich secondary plagioclase become more abundant, while chlorite and albite abundance decreases (Figure 2). This change in secondary mineralogy may relate to the increase in sonic velocities that places the seismic layer 2/3 transition well above the SDC/gabbro interface [Detrick et al., 1994; Swift et al., 1998]. The appearance of hornblende and secondary augite and anorthite suggests peak alteration temperatures of >500°C in the lowermost SDC [Vanko and Laverne, 1998].

2.1. Volcanic Zone (274.5 to 846.0 mbsf)

[6] Previous studies have shown that the upper extrusive section consists of chemically different types of basalts comprising pillowed and massive flows of unusually depleted basalts (D-MORB), with minor typical mid-ocean ridge basalts (N-MORB) and enriched basalts (E-MORB) [Etoubleau et al., 1983; Tual et al., 1985; Pedersen and Furnes, 2001]. The majority of the lavas are D-MORB with very limited chemical variability, ranging from slightly to moderately evolved compositions (Mg# = 0.59–0.69; [Natland et al., 1983]). Differences in Nd and Pb isotope composition between D-MORB and N-/E-MORB suggest mantle source heterogeneities [Pedersen and Furnes, 2001]; however, 143Nd/144Nd of D-MORB is ∼0.51312, similar to Pacific N-MORB (0.51313; [White et al., 1987]) implying that the depletion of the D-MORB source is not ancient. The highly depleted trace element characteristic of most rocks from Hole 504B (LaN/SmN ∼ 0.4, Zr/Nb > 100) can be accounted for by dynamic melting [Johnson et al., 1995] or multistage batch melting [Bach et al., 1996b].

[7] Alteration of basalts from the volcanic zone in Hole 504B and in rocks from the nearby 270-m deep ODP Hole 896A was described in detail [Honnorez et al., 1983; Alt et al., 1996c; Teagle et al., 1996]. Alteration is non-pervasive and focused along clay and carbonate veins. It is most intense in brecciated rocks. Alteration commonly produced red and dark gray halos along veins that are distinct in alteration mineralogy from the basalt more distal to veins.

[8] Borehole packer tests reveal that the uppermost 200–300 m of basement in Hole 504B are characterized by high permeabilities (∼10–15m2) that facilitate the circulation of large quantities of seawater [Becker et al., 1983; Fisher, 1998; Fisher and Becker, 2000] (Figure 2). Consequently, time-integrated fluid flow and intensity of low-temperature alteration is highest in the uppermost 300 m of basement. The basalts react with oxygenated deep-seawater to form secondary minerals, including Fe-oxyhydroxides, smectite, and celadonite [Alt, 1995; Honnorez, 1981]. These phases replace glass, olivine, sulfides, and to lesser extents plagioclase and clinopyroxene, and they fill fractures and void space in the crust. Zoned oxidation halos along clay/carbonate/oxyhydroxide filled veins are commonly developed [Honnorez et al., 1983; Alt et al., 1996c; Teagle et al., 1996] and indicate (1) diffusion-limited transport in the fairly impermeable rock away from the main fluid pathway, and (2) multiple episodes of seawater-rock interaction. The redox conditions in the pore fluid away from the main fractures change rapidly resulting in precipitation of a variety of Fe-bearing minerals that range from ferric oxides and hydroxides to celadonite and Mg-rich saponite to Fe-rich saponite and pyrite [Andrews, 1979]. Honnorez et al. [1983] found red alteration halos down to 310 m sub-basement. Below this depth, oxidative alteration is not developed suggesting the rocks were interacting with more evolved fluids that had largely lost their oxidizing capacity. The relative timing of different alteration stages in the uppermost crust has been summarized by Alt et al. [1996c]. The oldest secondary formations are celadonite and Fe-oxyhydroxides in dark gray to black halos, followed by Fe-oxyhydroxides and saponite in red halos. During this early oxidative alteration, formation temperatures were less than 25–50°C, and the crust was open to free circulation of seawater under oxidizing alteration conditions. The whole rock chemical changes involved increases in K, Rb, Cs and minor increases in H2O, Fe3+/ΣFe, U, P, Li, δ18O and 87Sr/86Sr, local losses of S, and minor losses of Ca and Mg [Teagle et al., 1996]. While the black, celadonite-bearing halos make up only ∼2% of the upper volcanic section in Hole 504B, red alteration halos are about 10 times more abundant [Alt et al., 1996c]. A later period of saponite and pyrite formation in the rocks reflects restricted circulation of evolved seawater, less oxidizing conditions, and temperatures >40°C. Increased Mg, H2O, δ18O, slight increases in alkali elements, and local gains of S and Tl (up to 0.4 ppm) were the prevalent chemical changes during this alteration stage. Carbonates and zeolites are the youngest secondary phases and formed at temperatures ∼60°C [Teagle et al., 1996] similar to the present-day temperature in the shallow basement in the area of Site 504 [Becker et al., 1983]. Zeolites are particularly abundant in a zone from 528.5 to 583 mbsf that may represent a zone of focused fluid up-flow [Alt et al., 1996a].

2.2. Transition Zone (846 to 1055 mbsf)

[9] The transition zone (TZ) is defined as an interval where both flows and dikes are abundant, and where physical properties of the crust change significantly (sonic velocity increases, porosity and permeability decrease). Pillows, flows, and dikes in the transition zone are intensely brecciated and altered [Alt et al., 1986; Honnorez et al., 1985]. The fraction of dislocation versus interpillow breccia is unknown and a relation between the intense alteration and stockwork mineralization and deformation is uncertain. Down to ∼900 mbsf the section is dominated by lavas that can be distinguished from rocks of the overlying volcanic zone by the development of titanite, laumontite, and anhydrite. Below 900 mbsf, chlorite, albite, and actinolite are abundant, and pyrite is common, disseminated in the rock and replacing silicates. Vein fill and breccia cements are chlorite-rich with varying proportions of quartz, actinolite, epidote, laumontite, heulandite, calcite, and pyrite. Metal sulfide mineralization is most abundant in the stockwork zone from 910 to 928 mbsf, but highly sulfidized breccias and sulfide impregnated massive rocks are developed locally throughout the lower section of the TZ and into the uppermost SDC. Honnorez et al. [1985] observed the following sequence of hydrothermal alteration in a detailed investigation of the stockwork zone, but similar paragenic sequences can be observed in other intervals of the TZ as well. Early periods of quartz veining and silicification in the stockwork zone, were followed by massive chloritization of the host rock and deposition of chlorite-rich veins. During a later stage, these fissures were reopened and quartz ± chlorite, epidote, sulfides were deposited. Multiple generations of sulfides (pyrite, sphalerite, chalcopyrite, and trace galena) indicate fluctuating temperatures and fluid compositions. Zeolites, calcite, and minor anhydrite form during the latest stage of hydrothermal alteration filling void space and replacing quartz (laumontite).

2.3. Sheeted Dike Complex (1055 to >2111 mbsf)

[10] The upper SDC (down to 1500 mbsf) is variably altered to greenschist-facies mineral assemblages. Alteration is most intense along chlorite ± actinolite veins and in cm-sized, light-gray alteration patches that may represent areas if increased primary porosity [Alt et al., 1996a]. The alteration mineralogy is similar to that in the lower TZ, but zeolites and calcite are less abundant, talc-magnetite alteration of olivine is common, and prehnite is more abundant filling veins and replacing plagioclase. Generally, fracturing and alteration intensity are decreased relative to the TZ. The sequence of vein mineral precipiation is (1) chlorite ± actinolite ± titanite, (2) quartz ± epidote ± sulfide, and (3) zeolite, prehnite, or anhydrite [Alt et al., 1996a].

[11] The lowermost SDC in Hole 504B appears to be representative of the root zone of hydrothermal vent fluids as indicated by depletions in Cu, Zn, and S depletions [Alt et al., 1996a]. Such root zones have been found in ophiolites just above the SDC/gabbro transition [Harper et al., 1988; Schiffman and Smith, 1988].

[12] Alt et al. [1996a] proposed five alteration stages in the lower sheeted dike section beginning with circulation of high temperature fluids (500°–600°C) along permeable fluid conduits and formation of secondary clinopyroxene and calcic plagioclase in veins, halos, and patches of strong alteration. In the second stage, pervasive circulation of up to 400°C hydrothermal fluids resulted in the formation of actinolitic hornblende, titanite and oligoclase, followed by amphibole, chlorite, and albite at lower temperatures. This episode was followed by upwelling of reacted fluids (Mg-depleted) at temperatures of about 320°C and precipitation of epidote and quartz in crosscutting veins and infill in pore space. During the fourth stage anhydrite formation occurred locally as a result of recharge of seawater into the hot crust. Finally, highly evolved hydrothermal fluids at lower temperatures (<250°C) circulated through the off-axis, causing the formation of laumontite and prehnite.

3. Samples and Analytical Methods

[13] More than 70 samples were taken from representative intervals of Hole 504B, with focus on the VZ, TZ, and upper SDC for which ICP-MS trace element data were previously not available. 56 thin-sections were made for microscopic examination of primary and secondary mineralogy and visual estimation of the degree of alteration. Seventy-one bulk rock samples were sand-blasted (quartz) to remove potential contaminants from the surface. Sample were then carefully cleaned in distilled water, air-dried, disaggregated in a ceramic jaw-crusher, and powdered in a new agate shatterbox for chemical and isotopic analyses.

[14] Major and trace elements were determined by X-ray fluorescence and ICP-MS at Washington State University. Relative external precisions of these techniques are <1–5% for the major elements, and <5–10% for trace elements. Precise U, Th, and Pb concentrations were measured by isotope dilution at WHOI, using a Finnigan Element ICP-MS. Carbon and S concentrations (CHS elemental analyzer) and S isotopic composition were measured at Indiana University, Bloomington, using a method described in Studley et al. [2002]. At GFZ-Potsdam, ferrous iron was determined by manganometric titration, and CO2 and H2O were measured by a LECO™ analyzer RC-412. CO2 data represent total carbon (carbonate, C, CH4 and other reduced carbon species). Sr isotope analyses were carried out at WHOI with a VG 354 thermal ionization mass spectrometer. Samples were dissolved with HF/HClO4 in Teflon beakers. Strontium was separated on quartz columns with a 5 ml resin bed of AG50W-X8 200–400 mesh. 87Sr/86Sr ratios are reported relative to NBS 987 = 0.71024. External precision (2σ) of Sr isotope analyses is 30 ppm.

4. Results and Discussion

[15] Results of petrographic and geochemical studies are listed in Tables 1 and 2. A detailed review of major element, isotopic, and mineralogical variation in Hole 504B is given by Alt et al. [1996b]. The whole rock chemistry database for Hole 504B that existed to this date is biased toward relatively fresh and unbrecciated samples from the uppermost 1200 m of basement. We focused our sampling effort specifically on including breccias and highly altered/mineralized rocks. This allows us to examine the geochemistry of highly altered and mineralized rocks from the VZ and TZ and work those into estimates of bulk upper crust composition and chemical fluxes. Average chemical compositions of VZ, TZ, SDC and total penetrated crust (super composite, SC) were calculated by weighting each sample according the lithologic abundance of rock types (Tables 3 and 4). For many major and trace elements these weighted averages deviate from those reported by Peucker-Ehrenbrink et al. (submitted manuscript, 2002) for samples that were analyzed for Re-Os. We use the averages that are based on the full set of analyses for the following discussion.

Table 1. Petrographic Description and Visually Estimated Degree of Alteration
LegHoleCoreTypeSectionInterval Top(cm) BottomPieceDepth (mbsf)Rock typeOlivine alterationPlagioclase alterationCpx alterationVoid and vein fillCommentsExtent of alteration
69504B3R1125131256279.3Fine-grained, sparsely plagioclase-olivine-phyric basaltComplete to brown clay and hematiteIncipient to pale clay along fracturesFreshReddish-brown clay 15
69504B5R2109117384292.1Microcrystalline to fine-grained, moderately plagioclase-olivine-phyric basaltCompletely replaced by pale-green clay/talc and later brown clayPartly replaced by yellowish-green clayFresh  20
69504B7R33235478310.8Microcrystalline, moderately olivine-plagioclase-phyric basaltComplete to pale clay and talc, brown clay and hematitePatchy to clay, up to 10%Fresh Sulfide globules in glassy mesostasis20
69504B13R44751806366.5Fine-grained, sparsely plagioclase-olivine-phyric basaltComplete to either brown clay and hematite or clay and pyriteIncipient to brown clayFresh Pyrite replacing oxide and olivine15
69504B16R40121003388.5Microcrystalline to fine-grained, moderately plagioclase-olivine-phyric basaltComplete to brown clay and hematiteIncipient to brown clay along cracksFresh Glassy mesostasis appears clay-altered30
69504B19R120251115403.2Microcrystalline to fine-grained, sparsely plagioclase-phyric basaltSparse microphenocrysts are altered to brown clay and hematiteWeakly altered to clayFresh  20
69504B23R133401315439.3Microcrystalline, highly plagioclase-olivine-phyric basaltComplete to pale-green clayWeak to clay, more altered near veinsFreshGreen clayGlassy mesostasis appears highly altered to brown clay30
69504B28R190991496475.9Fine-grained, sparsely plagioclase-olivine-phyric basaltHighly altered to light-brown clayFreshFresh Sulfide globules in glassy mesostasis8
69504B28R21031081514477.5Spherulitic, highly plagioclase-clinopyroxene-olivine-phyric basaltComplete to clayPatchy to clay, up to 10%Fresh Glassy mesostasis appears highly altered to brown clay35
69504B28R358641529478.6Spherulitic to microcrystalline, highly plagioclase-olivine-phyric basaltComplete to light brown clay and talcWeakly altered to clayFreshFew clay veinletsGlassy mesostasis appears highly altered to brown clay30
70504B32R11923139507.7Basalt breccia (clasts of spherulitic, moderately plagioclase-phyric basalt)Complete to brown clayLocally high to green clay   45
70504B33R2132138217 red519.2Fine-grained, moderately plagioclase-phyric basaltComplete to brown clayFreshFresh1 mm wide clay vein with brown selvages 10
70504B36R42634329547.9Fine-grained, moderately plagioclase-olivine-phyric basaltComplete to pale clayFreshFresh Trace secondary sulfides10
70504B37R3512390555.6Microcrystalline, sparsely plagioclase-olivine phyric basaltComplete to yellowish-reddish-brown clayPartly to green clay and pyriteFreshPale-green clay and zeolite with selvages of brown clayOxidative alteration clearly pre-dates non-oxidative alteration25
70504B39R1122126472571.7Basalt brecciaComplete to green and brown clay   Poor thin section, matrix plucked 
70504B39R27584490572.8Fine-grained, highly plagioclase-olivine-phyric basaltComplete to brown and dark-green clayIncipient to green clayFresh  12
70504B40R3130138558 red583.8Fine-grained, moderately plagioclase-olivine-phyric basaltComplete to reddish-brown clay and hematiteIncipient to yellowish clay  Mesostasis partly altered to brown clay and hematite20
70504B40R3130138558 gray583.8Fine-grained, moderately plagioclase-olivine-phyric basaltComplete to pale clay, pyrite, and calciteIncipient to green clay  Pyrite after hematite and magnetite20
70504B44R23040721612.8Microcrystalline to variolitic, moderately plagioclase-phyric basaltComplete to brown clayIncipient to green clay  Pyrite common in mesostasis10
70504B49R22530938657.8Basalt breccia (clasts of microcrystalline to variolitic, moderately plagioclase-phyric basalt) Weak to green clay Veinlets of brown and green clayClast-supported jig-saw breccia; breccia cement dominantly green clay25
70504B51R18187981670.3Fine-grained, moderately plagioclase-olivine-phyric basaltComplete to green clay, talc , and magnetiteFreshFresh Pyrite and rare chalcopyrite in mesostasis10
70504B56R126291102705.8Basalt breccia (clasts of microcrystalline, sparsely plagioclase-phyric basalt)Complete to brown and green clayFresh  Pillow-rind breccia; clast/cement: 80/20; cement is green and, less abundant, brown clay30
70504B60R176851264742.3Fine-grained, moderately plagioclase-olivine-phyric basaltComplete to green and brown clayIncipient to green clay  Pyrite and rare chalcopyrite in mesostasis15
70504B64R114201404733.1Microcrystalline, moderately plagioclase-olivine-phyric basalt grading into brecciaComplete to green clay and trace calciteWeak to green clay Vein network (pale-green clay)Pyrite after skeletal magnetite in altered mesostasis, no Py in vein but common in vein halo25
70504B66R21301361522793.8Fine-grained, aphyric basaltComplete to green clayIncipient to green clayFreshGreen clay veinlet with abundant pyriteRare brown clay after olivine, in mesostasis and in a single veinlet12
70504B70R123301550827.2Basalt breccia (clasts of microcrystalline, sparsely plagioclase-phyric basalt)Complete to brown clayIncipient to green clay Vein network (green clay, minor brown clay, trace carbonate, zeolite, pyrite)Jig-saw fit of angular fragments35
83504B72R163706A844.1Microcrystalline, sparsely plagioclase-phyric basaltComplete to green and brown clayIncipient to green clay Veins of green clay 10
83504B73R174805E853.2Fine-grained, moderately olivine-plagioclase-phyric basaltComplete to clay and opaline silica with talc-hematite-pyrite-quartz domainsPartly to green clay and carbonateIncipient to clay  20
83504B75R247543872.4Microcrystalline, moderately plagioclase-phyric basaltComplete to green clayIncipient to clayIncipient to clay  15
83504B79R251593A906.5Fine-grained, sparsely plagioclase-phyric basaltComplete to green clay and chloritePartly to chlorite and quartzIncipient to chlorite Pyrite, hematite, and magnetite are enriched in irregular bands35
83504B80R162705910.6Microcrystalline, moderately plagioclase-phyric basalt with strongly developed quartz-chlorite vein networkComplete to chlorite and pyritePartly to chlorite, quartz, and pyriteIncipient to chloriteQuartz-chlorite vein network with trace titanite, pyritePyrite is only sulfide and is concnetrated in the interiors of basalt kernels, away from the veins50
83504B80R210161911.4Microcrystalline, moderately plagioclase-phyric basalt Incipient to chlorite Network of chlorite-quartz-pyrite-sphalerite veins 35
83504B80R347534A913.3Basalt breccia (clasts of microcrystalline, moderately plagioclase-phyric basalt) Variably to chlorite and quartz, heavy near vein Cement is chlorite with tiny crystallites of titanite and epidote; late calcite-quartz-chalcopyrite veinletsPyrite is rare, disseminated in clasts50
83504B82R226344930.3Basalt breccia (clasts of microcrystalline, sparsely plagioclase-clinopyroxene-phyric basalt) Incipient to albite and quartz-chloriteFreshNetwork of chlorite veins with minor titanite and actinolite in center and quartz selvages; some late-stage zeolite veinsPyrite is abundant in vein halos50
83504B82R287906E930.8Microcrystalline, highly plagioclase-olivine-clinopyroxene-phyric basaltComplete to chlorite, quartz, and magnetiteVariably to chlorite-quartz and quartz-epidoteIncipient to actinoliteVeins of chlorite with actinolite and titanite (rare epidote) in center and quartz selvagesPyrite in patches and clusters away from veins50
83504B82R39132A931.5Basalt breccia (clasts of microcrystalline, sparsely plagioclase-olivine-phyric basalt)Complete to chlorite-quartzWeak to albite, quartz, chlorite, pyrite Quartz-chlorite veins with rare titanite and pyritePyrite concentrated in centers of clasts, replacing mesostasis and plagioclase50
83504B83R122262B937.7Fine-grained, aphyric basaltComplete to chlorite and actinoliteWeak to quartz and chloriteWeak to actinoliteNetwork of quartz-chlorite veins with little pyrite 40
83504B83R182905A938.3Fine-grained, aphyric basaltComplete to chlorite and actinoliteWeak to quartz, chlorite, and rare epidoteWeak to actinoliteMonomineralic quartz vein, cross-cut by a quartz-laumontite vein, and a chlorite veinGroundmass is notably chloritized40
83504B84R131361D946.8Microcrystalline, highly plagioclase-olivine-clinopyroxene-phyric basaltComplete to chlorite, quartz, and titaniteWeak to quartz, chlorite, titanite, and actinoliteWeak to actinoliteQuartz-chlorite veins with minor chalcopyritePyrite is concentrated in the interiors of kernels, away from the veins50
83504B84R114114611947.9Microcrystalline, highly plagioclase-olivine-clinopyroxene-phyric basaltComplete to chlorite and quartzWeak to quartz and chloriteAppears freshCyclic quartz-chlorite veinPyrite-rich stringers40
83504B85R242452957.2Microcrystalline, highly plagioclase-olivine-clinopyroxene-phyric basaltHigh to quartz and chloriteHigh to quartz and chlorite, also albitizedHigh to quartz and chloriteQuartz-chlorite veins, later calcite-quartz veins, and zeolite veins 60
83504B85R289994B957.7DiabaseComplete to chlorite, actinolite, and titaniteWeak to chlorite and actinoliteWeak to actinolite Sulfides in groundmass30
83504B87R117232B967.2Basalt breccia (clasts of microcrystalline, moderately plagioclase-phyric basalt) High to quarz, chlorite, titanite and actinolite, albitizedHigh to quarz, chlorite, titanite and actinoliteNetwork of chlorite, quartz-chlorite, and quartz-epidote veinsChalcopyrite is abundant in chlorite vein90
83504B89R123282C985.7Microcrystalline, highly plagioclase-olivine-clinopyroxene-phyric basalt Near veins: complete to quartz, chlorite, laumontite (after quartz), epidote, and chalcopyrite; heavily albitized away from veinHigh to chlorite, actinolite, and titaniteChlorite-quartz-actinolite-sulfide veinsLaumontite palimpsests after quartz80
83504B90R228364996.3Basalt breccia (clasts of microcrystalline, aphyric basalt) Moderate to albite Quartz-chlorite-actinolite-epidote cementGroundmass is chloritized and has trace of pyrite50
83504B91R345516B1007.0Fine-grained, aphyric basalt Moderate to laumontite, chlorite, actinoliteWeak to chlorite-actinoliteQuartz-chlorite-actinolite-titanite veins; quartz is partly replaced by laumontiteTrace pyrite in veins and groundmass60
83504B92R111311813B1013.6Basalt breccia (clasts of glass and microcrystalline, plagioclase-phyric basalt) Moderate to chlorite and quartz Quartz-actinolite-chlorite matrix; narrow quartz-pyrite veinletsGroundmass is chloritized and has trace of pyrite70
83504B94R220292B1032.2DiabaseComplete to chlorite, actinolite, and pyriteWeakly albitized, minor actinoliteWeak to actinoliteChlorite veinsPyrite and minor chalcopyrite are disseminated in rock, preferentially after olivine35
83504B98R1576381062.6DiabaseComplete to chlorite, actinolite, and chalcopyriteModerate to chlorite, albite, actinoliteWeak to actinoliteChlorite veins 45
83504B99R144526C1071.9DiabaseComplete to chlorite or talc-quartz-mtModerate to chlorite, albite, quartzModerate to chlorite and actinolite  50
83504B101R110511110B1090.6Fine-grained, plagioclase-clinopyroxene-olivine-phyric basaltComplete to chlorite and pyriteModerate to albite, chlorite, actinolite, pyriteWeak to actinoliteChlorite veins 40
83504B107R1104111141144.5DiabaseComplete to chlorite, actinolite, and magnetiteWeak to albite and actinoliteIncipient to actinolite  10
83504B122R1768561214.3Diabase brecciaComplete to chloriteHigh to chlorite, albite, and quartzHigh to chlorite and actinolite Matrix is all chloritized80
83504B130R269824A1281.2DiabaseComplete to talc, actinolite, quartz, magnetite, and pyriteIncipient to actinoliteIncipient to actinoliteChlorite veinlet 10
83504B133R2344241305.8DiabaseComplete to chlorite, actinolite, and magnetiteModerate to actinolite, albite, chlorite, and quartzHigh to actinolite, chlorite, titanite, and magnetite  75
83504B134R1136143191314.4DiabaseComplete to chlorite, actinolite, and magnetiteModerate to actinolite, albite, and chloriteHigh to actinolite, chlorite, titanite, and magnetite  70
83504B141R1505771346.0Chilled margin breccia   Cement of prehnite, chlorite, and anhydriteCompletely chloritized very fine-grained rock fragments100
Table 2. Results of Chemical and Isotopic Analyses of Samples From Hole 504B
TopBottom(mbsf)zonetypeSiO2TiO2Al2O3Fe2O3FeOMnOMgOCaONa2OK2OP2O5H2OCO2CO2SFeO, totalNi ppmCr ppmSc ppmV ppmSr ppmZr ppmY ppmGa ppmCu ppmZn ppmLa ppmCe ppmPr ppmNd ppmSm ppmEu ppmGd ppmTb ppmDy ppmHo ppmEr ppmTm ppmYb ppmLu ppmBa ppmTh ppmNb ppmY ppmHf ppmTa ppmU ppmPb ppmRb ppmCs ppmSr ppmSc ppmU (ppm)Th (ppm)Pb (ppm)87Sr/86Srδ34Sa
  • a

    Numbers in parenthesis are values for pyrite separates.

70504B33R2132138217 red527.3VZMR48.370.8815.66  0.168.9113.281.940.110.078  0.736<0.00599.737.772.570.71 154456282488550231984621.614.870.875.012.180.873.100.624.140.912.590.382.330.372.00.050.7223.421.410. 
70504B33R2132138217 gray527.3VZMG47.660.8715.26  0.179.9712.681.840.230.076  0.8550.00299.687.313.610.74 160460372628549251675621.514.650.834.892.120.863.090.614.100.902.590.372.320.376.10.050.6823.571.360.050.040.382.90.0678442.90.04500.00070.06580.00220.25870.00520.70373816 
70504B40R3130138558 red582.9VZMR48.030.7017.293.893.920.228.1513.691.970.160.0481.740.6520.619<0.005100.457.602.450.690.461145438352317035211471500.942.940.573.461.640.682.530.513.530.792.300.332.100.338.50.030.3120. 
70504B40R3130138558 gray582.9VZMG48.320.7316.853.244.900.168.1213.331.910.020.0501.560.1810.1860.08399.457.961.820.680.366137421312456538211478601.033.210.603.681.740.712.620.543.670.832.380.352.230.350.90.030.3421.
70504B44R23040721619.3VZB48.500.8815.81  0.148.3411.952.050.050.044 99400432816143271689701.033.670.734.512.100.853.260.664.621.022.940.442.770.440.70.030.2926.711.320.         
70504B47R26267830643.1VZB48.990.9014.61  0.168.7011.802.130.100.053 87291412925846251487661.053.580.714.262.060.823.140.634.370.962.750.402.540.400.80.030.3025.051.330.030.170.341.10.0065949.4         
70504B56R126291102706.9VZB48.251.2814.15  0.159.3510.032.460.070.142    99.379.454.040.67 792474831810596341883713.429.501.528.         
83504B72R163706A845.0VZPG47.350.9216.22  0.158.7012.561.940.020.054    99.428.133.370.69 126372432666949251480641.123.970.774.632.130.883.220.624.290.952.720.402.490.392.00.040.2925.321.420.030.070.400.20.0036751.4         
83504B75R247543882.9TZMG47.250.8816.03  0.168.5812.702.010.010.051    100.048.713.660.67 122361362467549251992561.113.940.774.672.100.873.070.604.110.922.630.382.400.371.00.040.2424.651.380.020.100.360.10.0017349.3         
83504B80R2991038916.0TZ/SWPG                             0.361.440.301.910.890.371.400.281.960.431.        (4.0)
83504B81R110232920.8TZ/SWB61.990.539.82  0.368.271.320.470.010.031  0.1820.77399.0311.095.130.61 682002619152816135769270.662.240.412.561.170.421.810.362.490.541.620.231.460.  4.4
83504B81R19910111927.1TZ/SWPG49.200.8715.401.896.970.238.7413.101.790.020.0601.770.0800.1900.189100.318.67 0.680.196146       811091.103.610.694.221.900.743.080.624.380.962.850.412.550.400.50.050.2924.741.320.        2.4
83504B82R287906E936.1TZPG49.660.8815.13  0.229.1212.312.010.010.062    99.988.621.960.69 101363352745644251461781.163.800.734.352.000.813.100.634.250.942.770.402.580.411.00.030.2824.921.300.<0.0015242.9         
83504B83R122262B939.3TZPG54.300.5212.23  0.348.248.131.820.020.031 1043163421633271713199190.622.050.402.551.240.451.990.402.790.611.820.261.600.<0.0013140.4         
83504B84R131361D947.9TZPG48.890.8714.92  0.279.0312.101.870.010.058    98.848.272.550.70 119390402755245241461961.083.620.704.211.960.773.090.624.400.962.740.402.550.410.80.030.3326.421.330.<0.0015351.4      0.70289814 
83504B84R234406954.6TZPG                              3.196.530.904.531.453.252.070.372.460.561.600.231.420.        (4.0)
83504B85R242452961.0TZPG53.200.8012.51  0.307.759.553.260.020.059    98.687.593.630.68 762613526911342241410841.073.520.674.071.860.812.950.594.090.902.620.372.350.383.30.040.3824.551.         
83504B90R228364997.8TZB59.840.7310.43  0.256.825.071.700.010.049    98.959.694.370.60 592494121727392111151121.353.760.633.691.620.842.480.503.490.762.240.322.060.333.20.090.3520.901.         
83504B90R312012418C1002.2TZB60.040.5211.31  0.246.454.270.670.010.033    99.0510.235.280.57 672062419426281918361291.463.690.613.481.471.212.210.442.980.651.910.271.700.         
83504B92R111311813B1015.4TZB53.570.5612.48    98.7511.126.340.63 812612822133291920441680.972.710.493.001.351.192.170.422.960.651.910.271.670.<0.0013230.3         
83504B98R1576381069.1SDCD52.241.0314.20  0.217.8311.451.850.010.071    99.438.162.380.67 912924430448522715247711.174.050.784.832.230.773.490.714.991.103.230.472.900.461.00.040.4028.971.570.<0.0014646.0         
83504B123R2586571234.4SDCD49.070.7516.31  0.159.3913.421.63<0.010.048    99.497.621.100.72 14441736232633721158261                                   
83504B130R269824A1285.3SDCD           1.370.0710.1010.088               1.053.410.633.791.720.752.760.553.810.862.440.352.240.340.20.130.3020.641.<0.0016242.00.01410.00030.02700.00120.14800.00340.702600140.5
83504B133R2344241315.1SDCD50.050.8614.58  0.208.3612.    99.028.551.850.67 88329432795340251468660.923.130.623.931.920.803.090.624.410.982.850.412.560.400.70.090.2924.<0.0015146.0         
83504B141R1505771350.3SDCB34.020.6616.262.8511.060.106.2514.320.860.010.0446.600.1170.1201.69497.3814.926.370.470.17272218322533740203251710.922.790.533.211.460.672.220.443.120.702.060.291.820.28 0.050.2817.
111504B145R1202231378.8SDCD49.600.8015.901.786.330.168.9612.901.770.020.0601.60 0.0300.01399.928.311.770.690.193139 46     28 0.983.280.633.921.840.762.860.563.910.872.500.362.310.350.50.040.2822.681.<0.0015744.7         
111504B163R21311514.6SDCD49.000.9016.402.376.270.157.9812.302.350.020.0601.480.030 0.01299.328.90 0.650.240108 47     28 1.073.710.714.372.020.793.190.624.380.982.860.412.590.400.30.030.2925.481.380.         
140504B193R15860141680.8SDCD50.200.5114.501.386.750.137.9210.901.48<0.010.0405.570.0900.1280.11699.598.28 0.670.15011139347 3926171336640.632.110.402.441.170.781.850.372.650.601.740.251.580.<0.0013940.30.01440.00030.01210.00090.07710.00170.7029341411.5
140504B214R1737681819.9SDCD48.800.4113.90  0.128.9912.401.72<0.010.0403.850.0700.1490.01398.958.64 0.69 16337640 502316121420.551.810.362.081.010.631.630.322.280.511.500.221.330.<0.0014735.40.00360.00020.01320.00100.36850.00720.703066143.0
148504B241R1117118202024.8SDCD48.250.6717.251.765.840.138.5212.971.900.010.0401.62  0.00198.947.78 0.700.20413538344 593216138410.702.350.462.811.330.592.230.443.070.692.060.301.900.300.40.040.2418.640.900.020.010.300.3<0.0016044.30.00380.00020.01830.00140.21270.00480.70280216 
148504B249R14822071.6SDCD49.571.0314.78  0.198.3712.972.11<0.010.0701.34  0.02197.757.34 0.71 8626751 6653271577531.384.530.845.142.320.923.530.694.851.083.060.452.780.451.40.050.5328.831.550.040.010.840.3<0.0016750.3         
Table 3. Recipe for 504B Composite
  1. a

    Lithological abundances are from Alt et al. [1996a, 1996c].

From274.5 846 1055
To846 1055 2100
Thickness571.5 209 1045
Pillows76% 50% 0%
Flows/Dikes18% 40% 96%
Breccias6% 10% 4%
Red Halos23%SW8.6%  
Gray Rock77%TZ-SW91.4%  
Total Meters
 Total (m)Total (%) Key 
VZ/PR99.95.5 Lithology zone 
VZ/PG334.418.3 VZ = volcanic zone 
VZ/MR23.71.3 TZ = transition zone 
VZ/MG79.24.3 SDC = sheeted dike zone 
SW/PG15.00.8 Lithology type 
SW/B3.00.2 PR = red halo, pillow 
TZ/PG95.55.2 PG = gray pillow 
TZ/MG76.44.2 MR = red halo, massive flow 
TZ/B19.11.0 MG = gray, massive flow/dike 
SDC/D1003.255.0 B = breccia 
SDC/B41.82.3 D = dike 
Table 4. Weighted Averages for Volcanic Zone (VZ), Transition Zone (TZ), Sheeted Dike Complex (SDC) and Total Penetrated Crust (Super Composite, SC)
  All SamplesAll SamplesAll SamplesAll SamplesPrecursora
  • a

    Precursor composition is from Emmermann [1985] and this work (marked “*”). See text and footnotes of Table 5 for detail.

XRFFeO, tot.8.338.518.658.539.90

[16] A major uncertainty in calculating an upper crustal composite is the low recovery rate in Hole 504B (30% in VZ, 25% in TZ, and 14% in SDC [Alt et al., 1996a]). A common concern is that soft lithologies such as breccias and highly clay-altered rocks are under-represented in the recovered core material. However, a recent study [Haggas et al., 2002] has shown that formation microscanner logging data and core description data provide similar proportions of breccia (∼10%) for Hole 896A, despite the fact that core recovery rate was only 27%. This report supersedes an earlier communication [Brewer et al., 1995] based on which large sampling biases were inferred. In any case, the low recovery rates make it impossible to estimate the statistical error associated with our composite calculation and the flux estimate. These uncertainties are likely large, on the order of ± 50%.

[17] In the following sections the terms “enrichment” and “depletion” are relative to fresh crust (see estimate in Table 4) and do not refer to primary differences related to magmatic differentiation.

4.1. Major Elements

[18] The downhole variations of major element concentrations are shown in Figure 3. Silica contents of the majority of rocks is ∼50 wt.%, but silica is enriched in silicified rocks from the TZ and depleted in highly chloritized rocks from the TZ and SDC. TiO2 is slightly enriched in volcanic units with E-MORB (core 19, 130 m sub-basement) and N-MORB (core 56, 430 m sub-basement) chemistries. Titanium (and other high-field strength elements, HFSE) is depleted in highly altered rocks from the TZ and the SDC. This depletion cannot solely be attributed to dilution of HFSE in the rock by hydrothermal infill of void space, but reflects mobility of these elements during hydrothermal alteration [Bach et al., 1996a]. Aluminum concentrations vary little, with the exception of pronounced Al depletions in the lowermost TZ. Iron is locally enriched in pyrite-rich and Fe-chlorite-rich intervals of the TZ and SDC. Manganese is enriched in the TZ and increases upward within the TZ. Alt et al. [1996a] have shown that this trend in whole rock composition mirrors the upward increase of Mn concentrations of chlorite and chlorite-smectite. Magnesium is locally enriched in saponite-rich lithologies. Similar to SiO2 and Al2O3, MgO contents are variable in the TZ and tend to be depleted in the silicified rocks (chiefly in the lower TZ) and enriched in the chlorite-rich upper TZ. Calcium is depleted in some intervals, in particular in the TZ, indicating local leaching of Ca from the rock. Leaching of Ca is less pronounced in the lower SDC, where secondary Ca phases (calcic plagioclase and amphibole) are stable. Sodium is variably enriched and depleted with the strongest variability in the TZ, reflecting patchy silicification and albitization. Potassium is enriched in most rocks from the uppermost 300 m of basement where celadonite (a K-bearing mica) is a common alteration phase. Like TiO2, phosphorous is enriched in the E- and N-MORB lithologies. A mild enrichment of P2O5 in the oxidatively altered uppermost 300 m of crust is likely related to the sorption of phosphorous onto Fe-oxyhydroxides [e.g., Wheat et al., 1996].

Figure 3.

Downhole variation of major element composition of rocks from this study (solid circles) and previous studies (open squares; see Alt et al. [1996a] for data sources). The pair of horizontal lines represents the location of the transition zone.

4.2. Extent of Alteration, Volatile Contents, Fe Oxidation, and S, Sr Isotopes

[19] Visual estimates of the percentage of secondary minerals in thin-section suggest that the extent of alteration is usually <50% in the VZ while it is highly variable (10–100%) in the TZ and SDC (Table 1). In the upper crust, where chlorite and smectite are the main alteration phases, water contents are strongly correlated with the extent of alteration (Figure 4). In the lower SDC, amphibole and secondary plagioclase are the most abundant secondary phases, and even in highly altered rocks, H2O contents are usually <2 wt.%. (This indicates that water concentration is not a good indicator of alteration if samples cover a range of metamorphic grades.) CO2 is most enriched in rocks from the lower VZ where saponite, carbonate, zeolites, and pyrite are common secondary phases.

Figure 4.

Downhole variation of the extent of alteration and geochemical indicators of alteration and mineralization. Data from Alt et al. [1996a], Teagle et al. [1998a], and references therein. The pair of horizontal lines represents the location of the transition zone. Solid vertical lines mark the estimated primary values (see text), while dashed vertical lines represent present-day seawater compositions.

[20] The extent of ferrous iron oxidation (Fe3+/ΣFe) is notably higher in the VZ than in TZ and SDC (Figure 4). In basaltic liquids, Fe3+/ΣFe is ∼0.1, but during cooling and crystallization Fe3+/ΣFe increases by 0.04 to 0.10 [Christie et al., 1986; Bach and Erzinger, 1995] leading to pre-alteration values of Fe3+/ΣFe between 0.14 and 0.20. Fe3+/ΣFe is elevated above these pre-alteration values throughout the entire VZ, indicating significant oxidation of the uppermost 600 m of basement by interaction with seawater. The upper 200 m of basement is marginally more oxidized (Fe3+/ΣFe = 0.40 ± 0.06) than the underlying 400 m of volcanics (Fe3+/ΣFe = 0.33 ± 0.07), but it important to note that oxidation of ferrous iron is not restricted to the zone of visible oxidative alteration.

[21] The S geochemistry of 504B rocks has been studied in great detail [e.g., Alt et al., 1989; Alt, 1995]. Primary sulfur contents can be estimated on the basis of major element composition and oxidation state of a basaltic melt [e.g., Mathez, 1976; Wallace and Carmichael, 1992]. Using an empirical relation between iron and sulfur contents of sulfide-bearing basalt glass [Mathez, 1979], it can be estimated that the primary S concentration of 504B lavas is ∼900 ppm. Oxidation of primary sulfides (pentlandite, pyrrhotite, and chalcopyrite) has led to almost pervasive S depletion in the uppermost 200–300 m of basement (Figures 4 and 5). Rocks that do not appear to have suffered S loss are more common in the lower VZ. The abundance of secondary pyrite in this interval suggests that sulfur is redistributed with little, if any, net sulfur loss [Andrews, 1979]. Sulfur is enriched in the mineralized stockwork zone and in certain intervals in the TZ and SDC. Sulfur isotopes indicate that many of basalts from the VZ have retained their primary sulfur isotopic signature (Figure 5). Rocks from the TZ have δ34S values between around 3–4‰, similar to typical hydrothermal vent fluids from mid-ocean ridges [Shanks et al., 1995]. These values suggest that sulfide in solution was dominated by sulfur leached from the basaltic basement with a minor contribution of reduced seawater sulfate [Alt, 1995; Shanks et al., 1995]. We measured δ34S of three pyrite separates that have values between 3.4 and 4.0‰, indicating that sulfides control the S budgets of rocks from the TZ. The patchy occurrence of rocks with elevated S concentration and δ34S indicates the addition of seawater sulfate in the form of anhydrite in certain intervals (Figure 5). In the case of a sample from core 141 (1350 mbsf) this could be confirmed by optical identification of anhydrite in a prehnite-rich breccia cement. Alt [1995] concluded that despite local sulfate and sulfide enrichments, the bulk crust does not appear to act as a significant sink for sulfur. Our super composite (Table 4) has a slightly higher S content that Alt's [1995] (0.14 versus 0.10 wt.% S) suggesting that the crust might act as a small sink for sulfur. However, our sample set may be biased toward S-rich samples by focusing on mineralized samples and including more sulfate-rich samples from the SDC than Alt [1995] (Figure 5). We do hence not feel that our data should supersede the results of the Alt [1995] study, which was specifically geared toward establishing a sulfur budget for altered crust.

Figure 5.

Plot of δ34S versus S concentrations of whole rock samples from Hole 504B. Arrows represent schematic trends of primary sulfide dissolution (leaching), addition of seawater sulfate in form of anhydrite (δ34S = 21‰), addition of hydrothermal sulfides (δ34S = 3–4‰), and addition of isotopically light sulfides of possible microbial origin. Note that unaltered crust has a δ34S of 0‰ and a sulfur concentration of ∼900 ppm. Symbols as in Figure 2.

[22] Isotopically light sulfur was reported for sulfide in some bulk rock samples and pyrite separates [e.g., Alt et al., 1989]. Andrews [1979] explained S isotope fractionation in altered basalt by the formation of sulfite and thiosulfate as intermediate oxidation products that disproportionate to 34S-enriched sulfate and 32S-enriched bisulfide. Contributions to 32S-enrichment from microbial sulfate reduction, however, cannot be ruled out.

[23] Sr isotope geochemistry of rocks and minerals from Hole 504B has been studied in considerable detail [Alt et al., 1996a, Teagle et al., 1998a; Teagle et al., 1998b, and references therein]. Downhole trends in Sr isotopic composition reflect exchange of crustal Sr (87Sr/86Sr ∼ 0.7025; Shimizu et al., 1989; Barrett and Friedrichsen, 1982) and seawater (87Sr/86Sr ∼ 0.709). This exchange of Sr is kinetically limited [Bickle and Teagle, 1992] and hence depends on the fluid mass flux as well as the temperature-dependent rate of Sr exchange between solid and fluid. 87Sr/86Sr decreases within the VZ as a consequence of a decreasing fluid flux down section. Within the transition zone, alteration is more pervasive, and rocks and hydrothermal precipitates (representing the Sr isotopic composition of the fluid) are isotopically equilibrated (Figure 4). This is a consequence of the combination of high alteration temperatures (>200°C) at still fairly high fluid flux rates (time-integrated water-to-rock mass ratios of ∼50 [Alt et al., 1986]). In the SDC, alteration is non-pervasive and highly variable with a general trend of decreasing 87Sr/86Sr and diminishing discrepancy between the 87Sr/86Sr of hydrothermal precipitates and host rock, indicating that isotopic equilibrium was approached.

4.3. Rare Earth Elements

[24] Rare earth elements (REE) concentrations of rocks from the VZ (Figure 6a) are uniformly light REE depleted with the exception of samples 19R-1, 20–25 cm (E-MORB) and 56R-1, 26–29 cm (N-MORB). The moderate degrees of low-temperature alteration in lavas from 504B did not result in any notable mobility of REE. A similar conclusion was reached by Teagle et al. [1996] for lavas from nearby Hole 896A (1 km SE of Hole 504B). Only where water-to-rock ratios of low-temperature alteration are extremely high, such as at the seafloor, alteration-related changes in REE contents of seafloor basalt, usually in the form of light REE enrichments, have been noticed [Ludden and Thompson, 1978]. Some samples from the transition zone exhibit modification of REE concentrations by hydrothermal alteration. Several samples from the TZ have large positive Eu anomalies, sometimes associated with light REE enrichment, while other samples have developed negative Eu anomalies (Figure 6b). It appears that Eu anomalies are limited to strongly altered samples (Figure 7a) and that the development of negative Eu anomalies is related to the breakdown of primary plagioclase, the main host of Eu in fresh basalt (Figure 7b). Positive Eu anomalies likely reflect net addition of Eu to the rock in the form of vein minerals (e.g., actinolite, epidote) that have precipitated from hydrothermal fluids with large positive Eu anomalies. Rocks from the SDC are variably REE depleted. Two of the diabase samples are chlorite and actinolite-rich alteration patches that show REE depletion and positive Eu anomalies typical of this rock type [Bach et al., 1996a; Bach and Irber, 1998].

Figure 6.

Rare earth element concentrations of rocks from the VZ, TZ, and SDC normalized against C1-chondrite [Sun and McDonough, 1989]. Average global N-MORB composition (thick, black line) from Sun and McDonough [1989] is shown for comparison.

Figure 7.

Plots of Eu/Eu* versus extent of (a) alteration and (b) CaO contents. Eu/Eu* = EuN/[(SmN + GdN)/2]. Trend 1 represents the development of a small negative Eu anomaly as alteration and plagioclase breakdown proceeds. Trend 2 illustrates the effect of hydrothermal mineral precipitation within void space. Symbols as in Figure 2.

4.4. Base Metals, Rb, Sr, U, Th, and Pb

[25] Zinc and lead are strongly enriched in the TZ, in particular in the mineralized stockwork zone between 636 and 654 m sub-basement (Figure 8). Copper is variably enriched and depleted, and the weighted average (Table 4) of 109 ppm Cu in the TZ indicates a net Cu enrichment of 35 ppm over Cu in the VZ (74 ppm). The corresponding enrichment of Zn and Pb in the TZ is 114 ppm and 0.76 ppm, respectively. In the lowermost 500 m of Hole 504B, rocks that are depleted in Cu and Zn are common [Zuleger et al., 1995; Bach et al., 1996a]. In this study, the lower SDC was probably under-sampled for reliable estimates of the base metal depletion. However, unweighted average values for the lower SDC (69 ppm Cu and 53 ppm Zn, 110 analyses) suggest that the net depletion in base metals is fairly small (5 ppm Cu, 13 ppm Zn; compared to weighted averages for VZ in Table 4). To account for the observed enrichment of Cu and Zn in 209 m of TZ, it is required to leach these quantities of Cu and Zn out of ∼1.5–2 km of crust. Depletions on the order of 15 ppm Cu and 48 ppm Zn were required to leach the Cu and Zn enriched in the TZ out of a 500 m thick reaction zone. These simple calculations suggest that there is an imbalance of about a factor of 3–4 between the observed base metal depletions in the lower SDC and the enrichments in the TZ. This discrepancy can be accounted for if the leaching zone is more extensive than usually proposed (kilometers rather than hundreds of meters) or, more likely, if it is assumed that transport was not strictly one-dimensional and that the average lateral extent of leaching zones is greater that that of stockwork zones.

Figure 8.

Downhole variation of selected trace elements. Note the strong enrichments of Zn and Pb in the transition zone (marked by the gray horizontal lines). Zn and Cu are somewhat depleted in the lower SDC, while no depletion can be deduced from the small number of Pb analyses available for the lower SDC. Also not that U is enriched throughout the VZ and TZ, while Rb is only enriched in the uppermost 300 m of basement.

[26] Similar to K, Rb is strongly enriched in the uppermost 300 m of the VZ. Strontium is depleted in the TZ, while U is enriched throughout the VZ and TZ (Figure 8). Consequently, Rb/Sr and U/Pb are elevated in most of the upper 800 m of crust at Site 504. The implications for Sr and Pb isotope geochemistry are discussed in detail in section 4.6. The enrichment of U in rocks of the lower VZ and the TZ is surprising, as previous reports of U uptake by the crust are restricted to low-temperature (≪100°C) altered rocks [e.g., Aumento, 1971; Bloch, 1980; Hart and Staudigel, 1982]. It was uncertain if the virtually complete U removal observed in vent fluids [Michard et al., 1983] is solely related to low-temperature uptake of U in the recharge zone. Unless affected by a cryptic low-temperature overprint, our data suggest that U uptake continues under conditions of the lower greenschist facies. U data for ridge flank fluids venting at Baby Bare in 3.5 Ma crust [Mottl et al., 1998], however, indicate that temperatures higher than 60°C are not essential for complete U removal from circulating seawater. Fe-oxyhydroxides and carbonates were proposed to host U in low-temperature altered rocks [Aumento, 1971; Staudigel et al., 1996; Teagle et al., 1996; Bach et al., 2001]. Both phases may play a role in U uptake as indicated by loose correlations of CO2 and Fe3+/ΣFe with U (Figure 9). The fact that U is enriched throughout the uppermost 800 m of basement where there are marked changes in secondary mineralogy (Figure 2) suggests the presence of more than one host phase for U. We speculate that Fe-oxyhydroxides are the most important U host in the uppermost VZ, while carbonates may have a stronger control on U distribution in the lower VZ and TZ.

Figure 9.

U concentration versus extent of oxidation (Fe3+/ΣFe)(A) and versus CO2 concentration (B). Data for 504B are compared to 417/418 data [Staudigel et al., 1996] that are much more enriched in U. Note that weak correlations between U and both Fe3+ /ΣFe and CO2 appear when the data are plotted in a log-log fashion (insets). Symbols as in Figure 2.

4.5. Implications for Chemical Basalt-Seawater Exchange

[27] We used weighted averages (Table 4) for the VZ, SDC, and total upper crust (SC) to estimate the directions and magnitudes of chemical exchange between seawater and crust. Assuming a crustal production rate of 3.45 km2/yr [Parsons, 1981] and an invariable density of the ocean crust (2800 kg/m3), we calculated corresponding chemical fluxes for axial and off-axial systems and total flux (Table 5). The underlying assumption is that chemical changes in the VZ took place dominantly in an off-axis setting, while the exchange in the SDC reflects high-temperature processes in an axial hydrothermal system. The latter is only correct if the whole rock chemistry of SDC rocks was not modified during later stages of hydrothermal alteration that took place off-axis. While the thermal history of 504B crust is poorly understood, it is generally assumed that late-stage alteration and anhydrite precipitation took place predominantly in the down-welling limb of an axial hydrothermal system [Teagle et al., 1998a]. Circulation of seawater extends well into the gabbro sequence [e.g., Gregory and Taylor, 1981; Hart et al., 1999], and therefore penetration in Hole 504B did possibly not go deep enough to sample the entire section of crust that interacted with heated seawater (see section 4.4). The assumption of constant crustal density introduces a large uncertainty, particularly for the major elements. Likely, crustal density increases with age, because the reduction in void space overwhelms the small decrease in density associated with alteration [Alt and Emmermann, 1985; Thompson, 1983]. Calculated major elements hence represent maximum values. The biggest uncertainty in the calculation of trace element fluxes is that the composition of the unaltered crust is not known as trace element analyses of fresh glass are not available. We scale the highly incompatible trace elements U, Cs, Rb, and K to Nb, which is depleted relative to N-MORB by a factor of 3 to 5 (weighted average of 0.72 ppm versus 2.3 ppm [Sun and McDonough, 1989] or 3.5 ppm [Hofmann, 1988] in N-MORB).

Table 5. Comparison of Ridge Flank Fluxes Derived for Hole 504B With Other Locations and the Dissolved River Flux
 VZSDCSCFreshOff-axisOn-axis?Total fluxTotal fluxTotal fluxOff-axisOff-axisOn-axisDRFi 
CA (ppm)aCA (ppm)aCA (ppm)aCi (ppm)b600 m VZc1000 m SDCc1800 m SCc504Bd504Be417/8 VZfJdFR flankgVentsh
  • a

    Note that positive values indicate flux into the ocean, while negative numbers indicate flux into crust.

  • a

    Weighted averages for volcanic zone (VZ), super composite (SC) and sheeted dike complex (SDC) are from Table 3.

  • b

    Intial concentrations for Si, Ca, and Mg are average values for 51 glass samples from 504B [Emmermann, 1985]. Trace elements (except Sr) were estimated to be 3 times depleted relative to N-MORB from Sun and McDonough [1989] (see text). CO2 was assumed to be 400 mg/kg in unaltered basalt [Fine and Stolper, 1986].

  • c

    Flux (mol/yr) = [−(CA−Ci)(g/kg) * thickness (m) * crustal production rate (3.45 × 106 m2/yr) * density (2800 kg/m3)] / molecular weight (g/mol).

  • d

    Data from: Alt et al. [1986] for 600 m VZ and 1000 m SDC and TZ; CO2 flux from Alt and Teagle [1999].

  • e

    Data from: Laverne et al. [2001].

  • f

    Data from: Staudigel et al. [1996] (Si, Ca, Mg, C, Sr) and Hart and Staudigel [1982] (K, Rb, Cs, U).

  • g

    Data (except Sr) from Wheat and Mottl [2000].

  • h

    Axial fluxes and dissolved river flux (except Sr) from Mottl et al. [1994].

  • i

    All fluxes are for mantle Sr with 87Sr/86Sr = 0.7025. Calculation of 504B Sr fluxes is based ppm Sr leached (numbers in parenthesis) derived from isotope mass balance, assuming 87Sr/86Sr of the fresh rock is 0.7025 and that of seawater is 0.709. JdF flank flux of 2.5 × 109 mol Sr/yr [Wheat and Mottl, 2000] with an isotopic composition of 0.7074 [Elderfield et al., 1999; Butterfield et al., 2001] was recalculated to a mantle Sr mantle Sr flux with 87Sr/86Sr = 0.7025. Axial Sr flux is from Palmer and Edmond [1989] recalculated to a mantle Sr flux with 87Sr/86Sr = 0.7025.

  • j

    Calculated from data in Pedersen and Furnes [2001] assuming 98% D-MORB and 2% N-/E-MORB.

SiO249210048410048800050660013993617538250003250−70018100–11006400× 109 mol/yr
CaO1257001174001190001278002171791272926003675−140047005–220012000× 109 mol/yr
MgO84400828008390082000−345−192−820−1800288170−5400−960–−26005400× 109 mol/yr
CO2192010001380400−200−132−387−150 −2700−16062–130032000× 109 mol/yr
K860100340200−9825−62−130−100−210−330130–11001900× 109 mol/yr
Rb1.390.210.590.19−81−2−81  −490−26160–1500370× 106 mol/yr
Cs0.0370.00040.0120.002−1.50.1−1.3  −6.5 1.8–9.84.8× 106 mol/yr
U0.0550.0080.0260.016−0.90.3−0.7  −9.7−1.4−0.23– −0.6445× 106 mol/yr
Sri69 (9)51 (22)58 (17)70j59524263374  411 13200 × 106 mol/yr

[28] The numbers calculated for Hole 504B are compared to previous flux estimates in Table 5. The only other drill sites for which rigorous budget calculations were made are DSDP Sites 417 and 418 in 118 Ma North Atlantic crust [Staudigel et al., 1995; Staudigel et al., 1996]. These drill holes, however, did not penetrate deeper than 600 m into the volcanic basement. Moreover, an estimate of the Pb budget for 417/418 was plagued by Pb contamination of the drill core [Hart and Staudigel, 1989]. Estimates of hydrothermal exchange were made for ophiolite sections [Lecuyer et al., 1990; Bednarz and Schmincke, 1989], but we did not include those estimates in our comparison, because there are pronounced differences in alteration extent and style between ophiolites and modern crust [Bickle and Teagle, 1992]. Wheat and Mottl [2000] used the chemical composition of fluid samples from the eastern Juan de Fuca Ridge flank and heat flow constraints to calculate off-axis fluxes, assuming that the dissolved river flux of Mg is entirely balanced by Mg uptake in off-axis hydrothermal systems.

[29] For Si, Ca, Mg, and K the total hydrothermal flux we calculated for Hole 504B is similar to earlier estimates by Alt et al. [1986] on the basis of unweighted whole rock data and vein counts that were available at the time to a depths of 1000 m sub-basement. Laverne et al. [2001] made detailed estimates of chemical change within the sheeted dike complex and combined these with flux calculations for the uppermost 600 m by Alt et al. [1986] to calculate integrated fluxes for Si, Ca, Mg, and K. Our results for these elements are similar to the fluxes estimated in these earlier studies, suggesting that our sample selection and weighting technique did not introduce large biases. The results of our study confirms the conclusions reached by Alt et al. [1986] that (1) hydrothermal input to the oceans of Si and Ca are significant when compared to the dissolved river flux, and (2) hydrothermal systems provide a significant sink for Mg and K. Our calculations also support the conclusions of earlier studies [Bloch, 1980; Hart and Staudigel, 1982; Staudigel and Hart, 1983; Thompson, 1983; Staudigel et al., 1989; Staudigel et al., 1995; Staudigel et al., 1996; Dunk et al., 2002] that the upper crust is a significant sink for CO2, Rb, Cs, and U. The fluxes calculated for these components on the basis of the 504B data are 4 to 13 times smaller than what was estimated for Site 417/418 [Hart and Staudigel, 1982; Staudigel et al., 1989; Staudigel et al., 1996]. A likely explanation for this discrepancy is a longer duration of seawater circulation at Site 417/418 resulting in higher time-integrated water-to-rock ratios and larger magnitudes of chemical exchange. This could have been facilitated by the rough basement topography and high permeability, in particular in the area of a basement high around Hole 417A, where the upper crust is highly brecciated, cemented, and altered. A Rb-Sr age for smectite-celadonite from Hole 417A of 108 Ma [Richardson et al., 1980] suggests circulation of oxygenated seawater continued at least 10 Myrs after formation of the crust (118 Ma). Although the basement age at Site 504 is only 6.6 Ma, non-oxidative alteration is already superimposed on oxidative alteration [Alt et al., 1996c] indicating that, indeed, the duration of oxidative alteration was relatively short at Site 504B. A consequence of the rapid sedimentation rates at Site 504 is conductive reheating of the basement following rapid termination of rigorous seawater circulation. CO2, K, Rb, and U flux estimates based on Hole 504B correspond within a factor of 4 to estimates by Wheat and Mottl [2000] for the Juan de Fuca Ridge flank (also an area of rapid sedimentation and conductive reheating), and these fluxes may be representative of ‘warm’ ridge flanks. Alt et al. [1996c] and Alt and Teagle [1999] find that there is about twice as much carbonate in Hole 896A (located on a basement and heat flow high) than in Hole 504B (basement trough), suggesting a significant small-scale variability in CO2 enrichment that probably reflects intensity of fluid flow. Alt and Teagle [1999] suggest that there is considerable uncertainty in the timing of CO2 uptake, which may continue in crust of considerable age (>100 Ma) or may be completed within a few tens of million years.

[30] Chemical exchange in off-axis hydrothermal systems was argued to provide a solution to the Mg budget of the ocean [Wheat and Mottl, 2000] and, most recently, also the Sr isotope budget of the ocean [Butterfield et al., 2001]. These conclusions were reached on the basis of geochemical analyses of fluid samples and extrapolating to global fluxes by ratioing chemical change to heat. The rock record provides a time-integrated result of processes such as Mg uptake and Sr exchange and hence allows us ground-truthing the extrapolations from fluid chemistry. The data presented in Table 5 suggest that the upper crust is a relatively small sink for Mg (504B), or maybe even a source of Mg to the oceans (417/418). The hypothesis that ridge flanks are the main sink for Mg in the oceans can only be reconciled with the riverine Mg input of 1.3 × 1014 g Mg/yr if 5.8 × 1015 g upper crust produced per year (assuming 600 m thickness, production rate of 3.45 km2/yr, and a density of 2800 kg/m3) shows an enrichment of 3.7 wt.% MgO. Such large Mg increases are rarely found in upper crustal rocks, and the average Mg uptake is an order of magnitude smaller (Table 5). The argument can be made that the core recovery in drill holes is too small and that highly clay-altered lithologies, clay-cemented breccias and Mg-rich clay veins are under-sampled. Formation of Mg-smectite (assuming 20 wt.% MgO and a density of 2500 kg/m3) in void spaces of the upper crust can only provide a sink for 1.3 × 1014 g Mg/yr, if roughly 20% of the upper 600 m of ocean crust consists of smectite. This is an order of magnitude higher than the observed smectite vein abundance in Hole 504B of 1.7% [Alt et al., 1996c]. Johnson [1979] and Staudigel and Hart [1983] estimated that smectite veins in DSDP Hole 418A (70% core recovery) comprise between 7 and 10% of the uppermost 583 m of crust. The Mg uptake associated with smectite vein formation, however, is balanced by large Mg loss of altered glass, in particular in Hole 417D, so that the super composite for Sites 417/418 does not indicate net Mg uptake by altered upper ocean crust [Staudigel et al., 1996].

[31] Similar to Mg, the rock record suggests that the flux of mantle Sr is at least one order of magnitude smaller that what would be required to balance the river input (Table 5). Teagle et al. [2001] have also suggested that the Sr isotopic shifts in altered crust are too small to account for the missing source of mantle Sr to the oceans. We propose that, unless a resolution to the different magnitudes of fluxes derived from fluid chemistry and rock record can be found, the global oceanic Mg and Sr isotope budgets should not be considered solved.

4.6. Isotopic Evolution of Altered Ocean Crust

[32] The data presented here allow us, for the first time, to establish the Rb-Sr and U-Th-Pb inventories of 1.8 km of altered ocean crust and to predict how such a reservoir would evolve isotopically with time. To compare the isotopic composition of modern oceanic basalt with those of ancient altered ocean crust, we (1) used model Rb-Sr and U-Th-Pb ratios of bulk silicate earth and depleted mantle reservoirs to estimate initial isotopic compositions of ocean crust 1 or 2 billion years ago, and (2) inserted Rb-Sr and U-Th-Pb ratios calculated for the 1.8 km 504B super composite and the 600 m 504B and 417/418 super composites and let these model reservoirs evolve to present-day. The principle behind this exercise is illustrated in Figure 10 and the calculation results are presented in Table 6. We acknowledge that quantitative understanding of Ce/Pb and Pb isotope evolution of the earth's mantle requires open system behavior [Galer and O'Nions, 1985; White, 1993; Chauvel et al., 1995; Elliott et al., 1999] that is not integrated in our simple models. We are particularly interested in testing if altered oceanic crust, in a closed system, may develop into a viable precursor of the HIMU mantle source characterized by radiogenic Pb and unradiogenic Sr isotopes [e.g., Zindler and Hart, 1986]. We use this in a qualitative way to examine the role of hydrothermal alteration and the associated modification of U-Th-Pb and Rb-Sr distribution in the upper ocean crust in generating Sr- and Pb-isotopic heterogeneity in the mantle.

Figure 10.

Examples of multistage Pb and Sr isotope evolution of different earth reservoirs. CC, continental crust; DMM, depleted MORB mantle; BSE, bulk silicate earth. It is assumed that BSE evolves from 4.55 to 2 Ga, at which stage continental crust, oceanic crust, and DMM are separated. We let upper oceanic crust of the model composition of the VZ in 504B evolve for 2 billion years and, in a second model, assume evolution of upper crust of the composition of the 417/418 super composite from 1 Ga to present-day. Model inputs are listed in Table 6.

Table 6. Pb and Sr Isotopic Evolution of Different Earth Reservoirs in Comparison With That of Upper Ocean Crust
  1. a

    BSE, bulk silicate earth; DMM, depleted MORB mantle; CC, continental crust; VZ, volcanic zone; SC, super composite; SC-TZ, super composite minus Pb hosted in transition zone, S1, hypothetical subducted crust (504B SC - 35%U - 80%Pb - 55% Rb - 40% Sr); S2, hypothetical subducted crust (504B SC - 40%U - 90%Pb - 40% Rb - 40% Sr), μ (238U/204Pb), ω (232Th/204Pb), κ (232Th/238U) values and 87Rb/86Sr are from Faure [1986] (BSE), White [1993] and this work (DMM), Rudnick and Fountain [1995] (CC), Hart and Staudigel [1989] (417/8), and this work (504B). t is time in billion years of closed system evolution since origin of the Earth (4.55 Ga, BSE) and arbitrary ages of differentiation of BSE into DMM, CC, and altered crust (2 Ga), and DMM into altered crust (1Ga). Upper ocean crust U, Th, Pb, Rb, and Sr weighted concentrations are from Table 4.

BSE     8.5034.004.000.0854.5518.0215.6838.060.70464
DMM     6.3015.752.500.0102.017.2215.5836.160.70249
CC     7.1027.693.900.5042.017.5215.6237.400.71670
504B VZ0.0550.0900.2301.396916.9636.342.140.0572.021.1016.0638.300.70384
504B VZ         1.019.0115.7137.210.70316
417/8 VZ0.3210.0630.6909.0117.532.996.520.200.2162.026.9316.7735.200.70843
417/8 VZ         1.021.7015.9035.690.70544
504B SC0.0260.0510.3100.59585.9511.751.980.0292.017.1015.5635.750.70303
504B SC-TZ0.0260.0510.2170.59588.5016.791.980.0292.018.0215.6836.270.70303
504B S10.01690.0510.0620.2663519.3358.763.040.0212.021.9616.1640.640.70282
504B S20.01560.0510.0310.3543532.0290.392.820.0291.022.1515.9441.320.70276

[33] We first examine the effect of the heterogeneous nature of alteration to the isotopic evolution of a hypothetical section of subducted crust. Figure 11 presents the downhole variation of measured Rb/Sr and U/Pb ratios and predicted 87Sr/86Sr and 206Pb/204Pb ratios after closed-system evolution for 2 billion years. Rb/Sr and U/Pb are variably elevated throughout the VZ. This enrichment results in significant radiogenic ingrowth of 87Sr and 206Pb in the uppermost 500 m of crust, while the lower section is predicted to evolve similar to the depleted mantle.

Figure 11.

Downhole variation of 87Rb/86Sr, 238U/204Pb, and measured Sr isotope compositions in comparison with the Sr and Pb isotope profile extrapolated to t = 2 billion years. 238U and 204Pb were approximated from measured U and Pb concentrations by assuming the fraction of 238U is 0.993 and that of 204Pb is 0.014. Initial 206Pb/204Pb is 16.22, which corresponds to a present-day ratio of 18.5 [Pedersen and Furnes, 2001] if a 238U/204Pb of 6.3 for the depleted mantle is assumed [White, 1993]. Initial 87Sr/86Sr is 0.7022, corresponding to a present-day 87Sr/86Sr of 0.7025 for the depleted mantle at a 87Rb/86Sr ratio for DMM of 0.01 (cf. Table 6). Gray vertical lines represent the approximate composition of depleted upper mantle. Modification of the intial Sr isotope composition by seawater alteration was not considered; this has a negligible effect on the calculation results as the Sr isotopic composition of 2 Ga seawater is likely fairly unradiogenic.

[34] We next explore how average altered upper crust will evolve isotopically by assigning it the weighted compositions of specific crustal sections in 504B and 417/418 (Table 6). Figure 12 plots the predicted isotopic composition of these model reservoirs in 3D Pb-Pb-Pb and Sr-Pb-Pb isotope ratio diagrams together with the global MORB and OIB data sets. We make the following observations: (1) the 504B super composite composition is not a suitable HIMU precursor but is predicted to develop similar to the depleted MORB mantle; (2) upper crustal (600 m) super composites for 504B and 417/418 may develop into HIMU sources as far as 206Pb/204Pb and 207Pb/204Pb are concerned, but 208Pb/204Pb does not evolved to values >38 and Sr isotopes evolve to compositions that are too radiogenic; (3) further modification of the Rb-Sr and U-Th-Pb ratios in subduction zones are required for subducted altered ocean crust meeting HIMU mantle source characteristics. These conclusions are similar to those of Hart and Staudigel [1989]. Differences in the mobilization of these elements during dehydration of altered crust have indeed been suggested based on empirical and experimental studies [Hawkesworth et al., 1991; Pearce and Peate, 1995; Keppler, 1996]. Our results suggest that the extent to which elements must be extracted from the dehydrating subducting ocean crust in order to make it a suitable precursor for the HIMU component in the mantle is: Pb > Rb ≥ Sr ≥ U > Th. In two example calculation we show that if 80–90% Pb, 40–55% Rb, 40% Sr, 35–40% U, and 0% Th are removed from the dehydrating slab with the chemical composition of the 504B super composite, such a reservoir will develop HIMU end-member-like isotopic compositions in 1 to 2 billion years (S1 and S2 in Figure 12).

Figure 12.

(a) Pb-Pb-Pb and (b) Sr-Pb-Pb plots of modern oceanic basalts in comparison with the present-day isotopic composition of the model components derived in Table 6. Abbreviation as in Figure 10, and SC = 504B super composite, SC-TZ = 504B super composite minus Pb hosted in the transition zone, S1 and S2 = modified 504B SC (see Table 6). Note that the upper ocean crust as represented by the 504B SC remains similar to DMM in Sr and Pb isotopic composition. The uppermost 600 m of ocean crust (417/8 VZ and 504B VZ) evolve to radiogenic 207Pb/204Pb and 206Pb/204Pb, but not to radiogenic 208Pb/204Pb. Only additional modification of U-Th-Pb and Rb-Sr ratios during subduction of the altered ocean crust will lead to components (S1 and S2) that make potential precursors of the HIMU component in the mantle (Table 6). See the GERM webpage ( for links to the data sources for oceanic basalts. Ocean island basalt data are for young basalts from Austral-Cook Islands, Azores, Galapagos Islands, Hawaiian Islands, Iceland, Kerguelen, Marquesas, Samoan Islands, Society Islands, St. Helena Chain, and Tristan da Cunha Group.

[35] Geochemical studies of Pb enrichment in hydrothermal sediments, continental crust and island arcs have concluded that ∼30% of the Pb in the ocean crust, or roughly 6 × 109 g Pb (0.3 × 0.3 ppm Pb × 7000 m × 3.45 km2/yr × 2800 kg/m3) must be removed during subduction to account for the Pb surplus of the continental crust and the Ce/Pb ratio of the depleted upper mantle [Miller et al., 1994; Peucker-Ehrenbrink et al., 1994; Chauvel et al., 1995; Mühe et al., 1997]. Peucker-Ehrenbrink et al. [1994] found that only 12(±8)% of the crustal Pb (or 2.6(±2.0) × 109 g Pb) may be transferred to the continental crust by the formation of metalliferous sediments at the seafloor. Recognizing that a large fraction of the ‘hydrothermal Pb’ is probably redeposited in subseafloor sulfide stockworks, Mühe et al. [1997] suggested that preferred mobilization of that sulfide-bound Pb by breakdown of sulfide phases in the subduction zone may account for the missing flux of hydrothermal Pb pointed out by Peucker-Ehrenbrink et al. [1994]. The Pb enrichment in the 209-m TZ in 504B is ∼0.8 ppm (Table 4). If all that Pb is mobilized during subduction, the corresponding Pb flux is 1.6 × 109 g/yr (0.8 ppm × 209 m × 3.45 km2/yr × 2800 kg/m3). This flux, while similar to the hydrothermal Pb transfer flux to the seafloor calculated by Peucker-Ehrenbrink et al. [1994], probably represents an upper bound because the development of a stockwork zone is likely discontinuous (see section 4.4). The combined hydrothermal Pb flux (∼25% of crustal Pb) may be sufficient to account for Pb enrichments in arc lavas, the Pb surplus of the continental crust, and Ce/Pb evolution of the mantle. Our data hence suggest that Pb stored in stockwork deposits within the upper ocean crust may indeed constitute a significant source of mobile Pb in subduction zones. However, a disconnect remains between this Pb flux estimate and the >80% Pb loss required to have subducted altered ocean crust develop into a reservoir with HIMU-like Pb isotope composition.

5. Conclusions

[36] The geochemical budgets of ocean crust – seawater exchange that were here established for 504B are similar, within a factor of 4, to budgets for Ca, Si, CO2, U, K, and Rb derived from studies of Juan de Fuca Ridge flank fluid chemistry [Wheat and Mottl, 2000]. Fluxes of CO2, U, K, Rb, and Cs are an order of magnitude smaller that estimates for 118 Ma basement at sites 417/418. According to the fluxes derived from Hole 504B, upper ocean crust is too small a sink for the alkalis to balance the riverine input. This discrepancy between 504B/JdFR on the one hand and 417/418 on the other hand may reflect differences in the sedimentation rates between the sites that lead to drastically different thermal and hydrogeological evolution between the Pacific and the Atlantic sites. The relatively small gain in Mg and subtle shifts in Sr isotopic composition inferred for the uppermost crust at 504B suggests that fluxes of Mg and exchange rates of Sr are insufficient to account for the large role ridge flank hydrothermal circulation was believed to play in the oceanic budgets of these elements [Mottl and Wheat, 1994; Butterfield et al., 2001]. Unless there is a very high fraction of unsampled, highly veined and altered material in 504B, the Mg and Sr isotope oceanic budgets cannot be considered solved. Ridge flank hydrothermal fluxes are still extremely poorly constrained. Deep drilling with high core recovery rates will be required to mine the full potential of using the rock record for the estimation of the seawater – ocean crust exchange rates. This will be one of the goals of the next phase of ocean drilling.

[37] Hydrothermal alteration of the upper ocean crust strongly affects Rb, Sr, U, and Pb geochemistry . However, with the Rb/Sr, U/Pb, and Th/Pb inventory of 1.8 km of upper crust in Hole 504B such a reservoir will develop very similar to the depleted upper mantle in Sr and Pb isotope space. Altered ocean crust will hence not develop strongly radiogenic Pb isotopic signatures often associated with recycled oceanic crust. Preferred mobilization of Pb redistributed within the crust and bound in sulfides in the mineralized stockwork zone may contribute significantly to the Pb surplus of the continental crust. Additional mobilization of Pb relative to U and Th in subduction zones is required to make subducted altered ocean crust a viable precursor for the HIMU mantle component [cf., Hart and Staudigel, 1989].

[38] Seawater–ocean crust hydrothermal fluxes and budgets of chemical alteration of ocean crust are still extremely poorly constrained. While 504B is the only drill hole that allows sampling of a large part of the upper crust, the low core recovery, the unusually depleted nature of the mantle source, and the remarkably high sedimentation rate in the area raise the question of how representative 504B is for ‘normal’ ocean crust. Again, deep drilling with high core recovery rates will be required to mine the full potential of using the rock record for the estimation of the seawater – ocean crust exchange rates and bulk chemical properties of altered ocean crust. We need a number of deep holes in crust of variable age and architecture to better constrain the timing and spatial variability of alteration. The role of ocean crust alteration in the regulation of ocean chemistry and the differentiation of the Earth will have to be revisited when more deep holes are established during the next phase of ocean drilling.


[39] We thank Ed Ripley (Indiana University at Bloomington) for conducting the S isotope ratio and carbon and sulfur concentration measurements. Ferrous iron, H2O, and CO2 concentrations were determined in Jörg Erzinger's lab at GFZ-Postdam. Mark Kurz provided access to the Thermal Ionization Mass Spectrometry facility at WHOI and Lary Ball helped with the ID-ICP-MS measurements at the WHOI ICP facility. We thank Jerry Bode at the DSDP West Coast Core Repository and Brian Schroeder (WHOI) for their assistance during sampling and sample preparation. Jeff Alt provided samples from the lower sheeted dike complex. Thanks to Damon Teagle and Olivier Rouxel for their insightful reviews. Margaret Sulanowska helped with the artwork. This work was supported by NSF grant OCE 9811209. WHOI contribution number 10,676.