4. CHEMISTRY OF THE LOWER SHEETED DIKE COMPLEX, HOLE …

Alt, J.C., Kinoshita, H., Stokking, L.B., and Michael, P.J. (Eds.), 1996Proceedings of the Ocean Drilling Program, Scientific Results, Vol. 148

4. CHEMISTRY OF THE LOWER SHEETED DIKE COMPLEX, HOLE 504B (LEG 148):INFLUENCE OF MAGMATIC DIFFERENTIATION AND HYDROTHERMAL ALTERATION1

Wolfgang Bach,2,3 Jörg Erzinger,2 Jeffrey C. Alt,4 and Damon A.H. Teagle4

ABSTRACT

We present major- and trace-element data on the lower sheeted dike complex (SDC) drilled during ODP Leg 148 in Hole504B. Similar to previously studied rocks from the upper sections of Hole 504B, the rocks recovered during Leg 148 are mildlyto moderately evolved mid-ocean ridge basalts (MORB) ranging in their MgO concentrations from 7.8 to 10.0 wt%. The rocksexhibit the same sort of depletion as the lavas of Hole 504B, characterized by very low incompatible element abundances (e.g.,TiO2 = 0.54-1.03 wt%, Zr = 29-53 ppm, and Ce = 2.2-5.5 ppm). Chondrite-normalized La/Sm and Nb/Zr ratios are markedlylower than average N-MORB, suggesting a highly depleted mantle source. Fractional crystallization of plagioclase, olivine, ±clinopyroxene can account for most of the chemical variability, whereby the degree of fractionation is less than 50%. Thechemical effects of alteration are more difficult to distinguish because they can sometimes have the same consequences as mag-matic differentiation. Formation of chlorite, amphibole, albite-oligoclase, and titanite replacing groundmass and primary min-erals are typical for all samples. The results of this type of alteration are slight increases of MgO and H2O and decreases of CaOand SiO2. TiO2, MnO, Zr, and the REEs appear to slightly decrease with increasing extents of alteration. Halos and patches ofintensive alteration are common in the lower SDC. The alteration halos show perplexing chemical trends with no clear relation-ship to the degree of alteration. The alteration patches are significantly enriched in H2O and highly depleted in Cu (Zn) and S.This make the patches in the lower SDC possible sources for base-metal deposits in the oceanic crust. The patches are also sig-nificantly depleted in REE, Zr, Nb, Ti, and exhibit positive Eu anomalies. Theoretical considerations and mass-balance calcula-tions indicate that this behavior is probably not only the consequence of intensive alteration but requires primary chemicaldifferences between the patches and the neighboring diabases. It remains unclear how far such differences could be related tomagmatic differentiation processes within the dikes. The puzzling question of the nature of these alteration patches deservesfurther attention.

INTRODUCTION

Ocean Drilling Program (ODP) Leg 148 was the seventh return toHole 504B and deepened the drill hole to a final depth of 2111 m be-low seafloor (mbsf) Hole 504B is the deepest drill hole in the oceaniccrust and serves as a reference for the structure and petrology of oce-anic crustal Layers 2. Hole 504B drilling and research history com-prises 15 years of intense technical and scientific effort during whichindispensable information on the petrology, chemistry and physics ofthe upper oceanic crust has been collected. Previous studies haveshown that the upper extrusive section consists of chemically differ-ent types of basalts comprising highly depleted basalts, typical mid-ocean ridge basalts (N-MORB), and enriched basalts (E-MORB)(e.g., Tual et al., 1985), thus displaying a time record of the temporalvariability of the mantle source. Studies of secondary minerals result-ed in the reconstruction of a complex multi-stage alteration history inthe volcanic section and the sheeted dikes including ore-forming pro-cesses in a transition zone (e.g., Alt et al., 1986).

Other points of great interest are the physical properties of theoceanic crust. The provocative finding of Legs 140 and 148 has beenthat seismic wave velocities steadily increase downward and reachvalues thought to be typical for oceanic Layer 3 (Dick, Erzinger,Stokking, et al., 1992; Alt, Kinoshita, Stokking, et al., 1993). How-ever, the rocks of the last section drilled during Leg 148 are still dia-

1 Alt, J.C., Kinoshita, H., Stokking, L.B., and Michael, P.J. (Eds.), 1996. Proc. ODP,Sci. Results, 148: College Station, TX (Ocean Drilling Program).

2GeoForschungsZentrum Potsdam, Projektbereich 4.2, Telegrafenberg A50, D-14473Potsdam, Federal Republic of Germany. Bach: [email protected]; Erzinger:[email protected]

'Present address: Universitat Potsdam, Institut fiir Geowissenschaften, Postfach601553, D-14415 Potsdam, Federal Republic of Germany.

department of Geological Sciences, 2534 C.C. Little Building, The University ofMichigan, Ann Arbor, MI 48109-1063, U.S.A. Alt: [email protected]; Teagle:[email protected]

bases of the sheeted dike complex. This finding has stimulated a newdiscussion on the nature of the seismic Layer 2/3 boundary and hasgiven rise to the question of whether the seismic Layer 2/3 boundaryan alteration front rather than the dike/gabbro transition (Detrick etal., 1994).

The current depth of Hole 504B appears to be representative ofthe lowermost part of the sheeted dike complex. Evidence for thispresumption arises from the observation that base-metal depletionsthought to be typical of root zones of hydrothermal vent fluids havebeen detected in the lower part of the hole (Alt et al., 1995; Zulegeret al., 1995). Such root zones have been found to occur in ophiolitesjust above the magma reservoir (Richardson et al., 1987; Harper etal., 1988; Schiffman et al., 1990; Nehlig et al., 1994).

In this paper we report major and trace-element data for 30 rocksfrom the lowermost 110.6 m of the sheeted dike complex. The studyis a continuation of the work by Zuleger et al. (1995) and follows thesame sampling and analyzing strategy, making our data directly com-parable to the Zuleger et al. (1995) data set. Most of the samples werealso studied for their stable isotope composition and secondary min-eral chemistry and the results are given by Alt et al. (this volume). Weplaced emphasis on the variability of chemical compositions in dia-bases altered to various extents. Our results indicate increasing mo-bility of nominally immobile elements with increasing degree ofalteration, resulting in incompatible element depletions of highly al-tered rocks. In contrast, changes in major-element compositions areless obvious. Possible primary and secondary mechanisms causingthese depletions are discussed below.

GEOLOGICAL SETTING AND PREVIOUS STUDIES

Hole 504B is located in 5.9 Ma crust, about 200 km south of theCosta Rica Rift Zone in the Eastern Pacific ocean (1O13.611'N,83°43.818'W; see Fig. 1). The Costa Rica Rift Zone is an east-west-

W. BACH, J. ERZINGER, J.C. ALT, D.A.H. TEAGLE

10° N

- 5C

- 0c

90° W - 5°S

Figure 1. Location of DSDP Sites 501 and 505, DSDP/ODP Site 504, and ODP SiteHobartetal., 1985).

south of the Costa Rica Rift in the eastern equatorial Pacific (after

trending intermediate spreading center with a southward spreadingrate of 36 mm/yr (Hey et al., 1977).

Several previous studies have shown that the basaltic rocks fromHole 504B are olivine- and plagioclase- (± clinopyroxene-) bearingtholeiites, representing moderately evolved MORB (Autio andRhodes, 1983; Etoubleau et al., 1983; Hubberten et al., 1983; Marshet al., 1983; Emmermann, 1985; Kempton et al., 1985; Tual et al.,1985; Autio et al., 1989; Sparks et al., 1995; Zuleger et al., 1995). Interms of their trace-element compositions most rocks (>98%) arehighly depleted in incompatible elements (Na, Ti, Zr, Y, REE, Ta,

Nb, etc.) when compared to average N-MORBs. In addition to thisdominant type of depleted composition are four other types of basalts(Kempton et al., 1985) that display either higher degrees of magmaticdifferentiation (Kempton et al.'s type M') or variations in their trace-element characteristics (Kempton et al.'s types M, T, and E). Using amore common terminology, one can say that less than 2% of therocks are not highly depleted and range in composition from typicalN-MORB to T-MORB and E-MORB.The numerous investigators ofthe igneous petrology and geochemistry of Hole 504B rocks agreeoverall in attributing these differences to mantle heterogeneities and/

40

CHEMISTRY OF THE LOWER SHEETED DIKE COMPLEX, HOLE 504B

Drilling history Lithostratigraphy Alteration

200

400

600

800

1000

αα>O 1200

1400

1600

1800

2000

2111

Leg 69

Leg 70

Leg 83

Leg 111

Leg 137

Leg 140

Leg 148

Sediment

Volcanic sectionof mainly pillowand subordinate

sheet lavas

Transition

Sheeted dikes,dipping 60°to 68°

N in the upper• part and 79°-84°

N in thelower part

• Chilled marginsand aphyric rocks

• less apparent orlacking in the

• lowermost part

• Low temperatureseawateralteration

Subo×ic toanoxic

alteration60°-110°C

Mineralization-zone

Greenschistmineral

assemblage

Temperature ofpervasive back-ground alterationincreasing down-

ward

' Patches and veinhalos of intense

' fluid-rock inter-action, moreabundant in

the lower part

Figure 2. Summary of the site drilling history, and the lithostratigraphy andalteration of Hole 504B rocks.

or to multi-stage melting processes, with the depleted basalts repre-senting a mantle that was depleted as a result of earlier melting pro-cesses.

Isotopic studies support the idea of a heterogenous mantle (e.g.,Barrett, 1983; Kusakabe et al., 1989; Shimizu et al., 1989). Anotherpeculiarity of the Hole 504B basalts is their high Ca/Na ratio and thecorresponding calcic plagioclase compositions (up to An94, Natlandet al., 1983). Liquidus temperatures calculated from glass composi-tions vary in a narrow range possibly indicating a nearly steady-statemagmatic system (Natland et al., 1983).

Figure 2 summarizes the drilling history, structural units, and al-teration types observed in Hole 504B. Alteration style changes in theextrusive series from oxidative to non-oxidative and this correlateswith changes in chemical compositions (e.g., a drastic decrease in theconcentrations of alkaline elements; Hubberten et al., 1983; Alt et al.,1986). Leg 83 drilled a transition zone of extrusive flows cross-cut bynumerous dikes (Alt and Emmermann, 1985). The upper part of thistransition zone, the so-called stockwork mineralization zone, is char-acterized by profound impregnations with sulfides resulting in in-creased Zn, Cu, Fe, and S concentrations (Alt and Emmermann,1985; Alt et al., 1986). The stockwork ores are thought to be precip-itated when hot upwelling metal-bearing hydrothermal fluids mixedwith cold downwelling seawater at a prominent permeability bound-ary between the more permeable extrusives and the less permeablesheeted dikes (Alt et al., 1986).

Studies of secondary mineral chemistry and paragenesis made adetailed reconstruction of the post-magmatic history of the uppercrust in Hole 504B possible. Alt et al. (1995) and Laverne et al.

(1995) proposed five alteration stages in the lower sheeted dike sec-tion beginning with circulation of high-temperature fluids (500°-600°C) along fluid conduits of high permeability and formation ofsecondary clinopyroxene and calcic plagioclase in veins, halos, andpatches of strong alteration. The second stage was pervasive circula-tion of up to 400°C hydrothermal fluids and deposition of actinolitichornblende, titanite and oligoclase, followed by amphibole, chlorite,and albite at lower temperatures. This episode was followed by up-welling of reacted fluids (Mg-depleted) at temperatures of about320°C and precipitation of epidote and quartz in crosscutting veinsand fill in pore space. During the fourth stage, anhydrite formationoccurred 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 crust and led to the forma-tion of laumontite and prehnite.

The sheeted dikes below the transition zone have been altered un-der greenschist facies conditions with temperatures increasing down-ward (Alt et al., 1989; Alt et al., 1995; Laverne et al., 1995). Aninteresting feature of the lower sheeted dike complex is the presenceof highly altered zones along amphibole veins (halos) and irregularlyformed patches of strong alteration. The latter show drastically lowerabundances of incompatible and nominally immobile elements (Ti,Zr,Y,REE,etc.)(Zulegeretal., 1995;Sparks, 1995). It is yet unclearto what extent these chemical depletions are of primary origin as op-posed to alteration processes.

SAMPLE PREPARATIONAND ANALYTICAL PROCEDURES

Rock powders for all chemical analyses and chips for all thin sec-tions were prepared from the same piece of rock. After culling rockchips for thin sections from the remaining material, areas with largephenocrysts were cut off. The samples were then rinsed with distilledwater, crushed in a steel mortar, and powdered in an agate mill.

Major-element oxides were analyzed by X-ray fluorescence(XRF) on samples prepared as fused disks of lithium metaborate(sample-to-flux ratio 1:4). A Philips PW 1400 computerized spec-trometer was used and a Philips "alphas program" calculated the con-centrations. Ferrous iron was determined by manganometric titration,and H2O

+ by a LECO analyzer RC-412. Sulfur analyses were per-formed using a LECO sulfur determinator CS 225.

Determinations of Cr, Ni, Cu, Zn, Ga, Sr, Y, and Zr were conduct-ed by XRF on pressed powder pellets. The rhodium Compton peak ofthe X-ray tube was used for matrix corrections. International refer-ence rocks were used for calibration of the XRF method

The rare-earth elements (REE) Li, Rb, Sc, Nb, Th, and U were de-termined by mass spectrometry with an inductively coupled plasmaas ionization source (ICP-MS, VG Plasmaquad PQ 2+). We used asample preparation and a modified ICP-MS technique originally de-scribed by Garbe-Schönberg (1993). A detailed description of ourmodified method is given by Zuleger et al. (this volume). Cr, Ni, Cu,Zn, Ga, Sr, Y, and Zr were also determined by ICP-MS. Except Zr,and to a lesser extent Y, all ICP-MS data are comparable to the re-spective XRF values (see Zuleger et al., this volume, for a compari-son of XRF and ICP-MS data).

Previous studies of Hole 504B rocks revealed that chemical datasets of various laboratories differ significantly because of differentanalytical accuracies and precisions (Dick, Erzinger, Stokking, et al.,1992, and Boström and Bach, 1995, for a discussion). We thereforecarefully tested our data with the reference data from Govindaraju(1994) and with data from other laboratories to ensure proper analyt-ical accuracy.

Table 1 presents the analytical accuracy and precision of the XRFdata. Assuming the analytical error is close to twice the standard de-viation, the precision for the major elements is better than 1.5 rel%and the precision for the trace elements is better than 10 rel%, mostlybetter than 5 rel%.

41


Table 1. Precision and accuracy of XRF data.

Rec. valueDR-N

Mean(N = T) lσ

RelativeSD (%)

Major elements (wt%):SiO2

TiO,A12O,Fe2O,MnOMgOCaONa^OK 2 O

52.851.09

17.529.700.224.407.052.991.700.25

53.251.08

17.879.680.224.427.073.081.720.24

0.100.010.040.020.000.020.020.010.0040.00

0.180.700.250.230.000.500.240.440.220.00

Rec. valueBHVO-1

Mean(N=9) Iσ

RelativeSD (%)

Trace elements (ppm):CrNiCuZnGaRbSrYZrNb

2891211361052111

40327.6

17919

29711414210321.010.6

39826.2

17218.8

5.61.72.41.10.90.72.41.43.20.4

1.91.51.71.14.47.00.605.31.92.4

Notes: The recommended (Rec.) values are from Govindaraju (1989). The internationalstandard rocks DR-N and BHVO-1 were used for major elements and trace ele-ments, respectively. SD = standard deviation.

Table 2 compares our analyses of the international reference rockBIR-1 with three other laboratories using similar methods and instru-ments. The precision of our data is usually better than 10 rel%. Acomparison between our "BAS 140" XRF and ICP-MS data and the"BAS 140" data of Sparks (1995), Zuleger et al. (1995) and Zulegeret al. (this volume) is given in Table 3. The results of all elements arein very good agreement, indicating that a direct comparison of ourdata with the data set of Sparks (1995) and Zuleger et al. (1995, thisvolume) is allowed.

Chemical compositions of the samples under investigation aregiven in Table 4. Cs was measured using ICP-MS, but the concentra-tions are below the detection limit of 20 ppb. Ba and Pb ICP-MS datahave been found to be affected by solvent blanks and memory prob-lems at the low concentration level of Hole 504B rocks, and hence arenot reported. XRF data of Zr correlate excellently with ICP-MS Nb,REE, and Th data, whereas ICP-MS data of Zr do not. Zr ICP-MSvalues are systematically lower than Zr XRF values, possibly be-cause of incomplete dissolution of trace minerals. Because Hf doescorrelate only with the (poor) Zr ICP-MS data but does not correlatewith any other element, we assume that both Zr and Hf ICP-MS dataare poor and thus they are not reported. Interestingly, we did not havethese problems with the Leg 148 Hole 896A rocks (Teagle et al., thisvolume), suggesting that indeed it may be a problem of incompletedissolution of trace minerals (zircon?) that occur in the lower sheeteddikes but not in the lavas.

RESULTS

Petrography

All samples are sparsely to moderately phyric diabases with pla-gioclase, olivine, clinopyroxene, and rare spinel phenocrysts. Grain-size analyses and the distribution of chilled margins led Umino(1995) to develop a model of multiple dike intrusions with up to 8.5-m-thick dike packets. Grain size data display decreasing grain sizesfrom 2000 to 2042 mbsf, followed by overall increases in averagegrain-sizes downwards (Alt, Kinoshita, Stokking, et al., 1993).

Table 2. Comparison of trace-element data obtained by ICP-MS withcertified values of the international reference rock BIR-1 and data fromthree other laboratories.

BIR-1reference

This work(Mean) l σ

Univ. KielICP-MS

MPI MainzSSMS

Univ St. John'sICP-MS

LiScCrCoNiCuZnGaRbSrYZrNbLaCePrNdSmEuGdTbDyHoErTmYbLuHfTh

3.444

38251.4

166126

71160.25

1081615.50.60.621.950.382.51.10.541.850.362.50.571.70.261.650.260.60.03

3.3641.3

38053

16111773150.27

11013.213.40.490.671.910.42.41.080.561.730.372.60.581.660.251.660.250.610.037

0.321.4

364

14540.90.0870.80.90.040.070.060.040.090.110.050.080.030.150.050.040.020.090.020.040.002

41.9343

52.4162120

0.293.71512.70.490.541.740.332.1310.471.530.312.230.511.490.221.510.220.570.024

0.272109

16.812.10.490.5941.830.3412.191.070.4571.520.32.260.551.50.231.360.230.360.031

3.143

0.23111

15.616.30.590.6

1.870.362.31.060.561.730.362.690.611.70.241.630.250.600.036

Notes: BIR-1 data from Govindaraju (1994). University (Univ.) of Kiel data fromGarbe-Schönberg (1993). Max-Planck-Institut (MPI) Mainz data from Jochum et al.(1990). University of Saint John's data from Jenner et al. (1990). ICP-MS = induc-tively coupled plasma-mass spectrometry, SSMS = spark-source mass spectrome-try. Concentrations are in parts per million (ppm).

The rocks are hydrothermally altered to various extents. To quan-tify the extent of alteration we visually estimated the percentage ofsecondary minerals, which we refer to below as "degree of alter-ation." During Leg 148, thin sections were point-counted for alter-ation minerals. The deviation of our visual estimates from theshipboard point counts is less than 10%. The Shipboard ScientificParties of Legs 137, 140, and 148 (Becker, Foss, et al., 1992; Dick,Erzinger, Stokking, et al., 1992; Alt, Kinoshita, Stokking, et al.,1993) identified at least three different types of alteration. A frequen-cy distribution of the degree of alteration in the different rock typesis shown in Figure 3. First, dark gray diabases affected by pervasivebackground alteration up to 60%, but usually between 10% and 40%.The diabases are affected by a pervasive background alteration dur-ing which olivine was completely replaced by various secondaryminerals, clinopyroxene was partly replaced mostly by amphibole,and plagioclase remained virtually unaltered except albitizationalong cracks. Second, alteration halos along amphibole veins showdiscoloration and variable degrees of alteration (Fig. 3) but are usu-ally more extensively recrystallized than the host rock. The most ob-vious difference between the halos and the adjacent host rock is thehigher abundance of amphibole in the halos (35%-70%) as comparedto the host diabase (5%-35%). This amphibole replaces virtually allmagmatic clinopyroxene in the halos. Generally, veins and halos aremore abundant in the section drilled during Leg 148 (21 veins/m, 1.2vol% veins) as compared to Leg 140 (15 veins/m, 0.6 vol% veins)(Alt, Kinoshita, Stokking, et al., 1993). Third, patches of intense al-teration up to 80% recrystallized vary from centimeter to decimeterscale and occur as irregular to subrounded, dark green areas gradinginto host diabase. Within the patches, 45%-80% of the igneous min-erals are replaced by secondary phases, typically amphibole (rangingin composition from actinolite to actinolitic hornblende to magnesio-hornblende), chlorite, secondary magnetite, and titanite. The halos


Table 3. Comparison of chemical data of the laboratory standard BAS140.

Major elements (wt%):SiO 2

TiO2

A12O3

Fe^O,MnOMgOCaONa,OK2OP 2O 5

Trace elements (ppm):

CrNiCuZnGaSrYZr

LiScNbLaCePrNdSmEuGdTbDyHoBrTmYbl.uHfHiU

Sparks

(1995)

50.60.99

14.5811.080.198.14

12.411.85

ND0.065

19285848515.645.526.644.8

Sparks

(1995)

NDNDNDNDNDNDNDNDNDNI)NDNDNDNDNDNDNDNDNDND

Zuleger et al.

(1995)

50.61.00

14.611.30.198.21

12.51.78

ND0.08

19481837816472749

Zuleger et al.

(this volume)

1.143

0.561.03.80.784.92.080.83.20.604.50.983.00.442.90.451.24

NDND

This study

(Mean)

50.40.99

14.611.2

0.188.15

12.51.78

ND0.076

19281827816473049

This study

(Mean)

1.1144

0.550.983.80.764.82.060.82

3.10.624.40.962.90.432.850.421.120.0210.007

lσ

0.050.0040.020.030.0040.040.030.01

0.005

1.70.510.50.60.60.90.6

lσ

0.0710.020.070.10.050.10.050.050.10.040.150.020.10.020.040.020.040.0060.003

Note: ND = not determined. Rare-earth elements in parts per million (ppm).

and the patches are thought to represent areas of extensive fluid-rockinteraction. Only one sample (Section 148-504B-247R-1, Piece 10)shows both patches and haloed veins, where the age relationshipappears to be such that the patch was formed before the vein (Alt,Kinoshita, Stokkingetal., 1993). Alt etal. (1985, 1989, 1995) report-ed millimeter-scale amygdules that apparently represent primaryvugs or pore space filled by amphibole, chlorite, laumontite, and epi-dote in the alteration patches. These vugs can make up 15% to 20%of the rock volume. Such amygdules were observed in one samplefrom Leg 148 (Section 148-504B-252R-1, Piece 5) and consist ofneedle-like amphibole. Thus, high primary porosity was present atleast locally in the Leg 148 section. Patches appear to be generallyrestricted to regions with large average grain sizes (i.e., the centerparts of dike packets). Rock type and alteration type of all samplesunder investigation are given in the heading of Table 4. It is importantto note that many samples are composed of different types of alter-ation (e.g., dark + patch), this information is also given in Table 4.The classification into groups (D = dark diabase, H = Halo, and P =patch) was made on the basis of the prevailing type of alteration (fora more complete petrographic documentation and photomicrographsof the three rock types, see Alt, Kinoshita, Stokking, et al. 1993; Altet al., this volume; and Vanko et al., this volume). The Leg 140 Ship-board Scientific Party (Dick, Erzinger, Stokking, et al., 1992) ob-served a fourth rock type of light gray color and more intensebackground alteration as compared to the dark diabases. This type ofrock was not sampled during Leg 148.

Chemical Composition

Rocks from the sheeted dike complex (SDC) display an overallremarkable homogeneity in chemical compositions. The composi-tional range (in wt%) of samples from Leg 148 is as follows: SiO2 =47.9-50.3, TiO2 = 0.54-1.03, A12O3 = 14.8-17.3, FeO* = 7.5-9.7,MgO = 7.8-10, CaO = 12.3-14.0, and Na2O = 1.57-2.11. These rang-es are similar to those previously reported for the diabases of Hole504B(Autioetal., 1989; Zuleger etal., 1995; Sparks, 1995). Basalticrocks from Hole 504B have been classified as olivine to quartz nor-mative tholeiites of relatively primitive composition (Hubberten etal., 1983; Emmermann, 1985). Trace-element compositions (in ppm)of Leg 148 rocks are also similar to those from previous legs: Cr =266-399, Ni = 86-167, Zr = 29-53, Y = 16-27, Sr = 50-66, La =0.54-1.53, Ce = 2.2-5.5, Sm = 1.1-2.3, and Yb = 1.7-2.5. Overall,the diabases display the trace-element characteristics of highly de-pleted and relatively primitive basalts. Evidence for the unusual de-pletion of the mantle source arises from the low (La/Sm)N (0.29-0.42) and high Zr/Nb (83-140) ratios. The values differ drasticallyfrom average depleted mid-ocean ridge basalts ([La/Sm]N = 0.65, andZr/Nb = 32; Hofmann, 1988). The generally low concentrations of in-compatible elements and the high MgO, Cr, and Ni values indicatethat basaltic melts must have been primitive to mildly evolved. Th/Uratios of the lower oceanic crust have been rarely reported so far butthey are important to construct mantle evolution models. Our data re-veal Th/U ratios of the lower SDC in Hole 504B of 3 ± 0.5. This isnot significantly different from the average Th/U of fresh MORBs(2.6; Hofmann, 1988).

All rocks of the SDC are hydrothermally altered to various extentsaccompanied with chemical changes as a result of varying mobilitiesof elements during recrystallization. On the other hand, basalt glassesfrom the volcanic section (Natland et al., 1983) show significantchemical variability, probably because of magmatic differentiationoccurring upon cooling of a magma reservoir.

DISCUSSION

Chemical Variation With Depth

The chemical variation of the lower SDC rocks including samplesfrom Leg 140 (Zuleger et al., (1995) and Leg 148 (this work) areshown in Figure 4. Depth intervals with abundant patches are markedby the shaded areas.

Generally, the degree of alteration ranges to higher values in thesections with alteration patches. TiO2, MnO, Zr, and CaO show low-est and water shows highest concentrations in these zones of highalteration. Two sections have MgO and Ni contents higher than usual,one from 1700 to 1730 mbsf, and a second one at the lowermost partof the drill hole (2080-2111 mbsf). In these intervals, MgO valuesare between 9 and 10 wt% and Ni varies from 125 to 190 ppm. Someelements display a crude sawtooth-like pattern in their variation withdepth. This is most clearly seen in the TiO2, Zr, and MnO data andthe reflected type of pattern is delineated by Ni and, less obvious,MgO. This type of pattern becomes less apparent in depth intervalswhere the sample spacing is wide (e.g., 1830-1900 mbsf). In thedepth interval 2000-2111 mbsf drilled during Leg 148 are trends ofdecreasing MgO and Ni and increasing TiO2, Zr, and SiO2 from 2000to about 2050 mbsf, followed by reversed trends from 2050 to thebottom of the hole. The water contents and degrees of alterationexhibit a similar trend of decreasing values down to 2050 mbsf andincreasing values below. Zuleger et al. (1995) discussed a possiblerelationship between such vertical variations and the intrusion of dikepackets with varying compositions. The new data support this possi-bility as the systematic trends in Ni and Zr, for example, are welldeveloped in the lower section and these trends correlate with sys-tematic decreases and increases in the average grain sizes (Alt, Ki-noshita, Stokking, et al. 1993). Because the dikes are steeply dipping


Table 4. Chemical compositions of Hole 504B diabases.

Core, section:Interval (cm):Piece no.:Depth (mbsf>:Lith. unit:Rock type:Alteration type:

Group:Alteration (%):

239R-145-51

14-HALO2000.85

271M-POD

Halo

H70

Major elements (wt%):SiO,Tic•;A12O,FeöFe2O,MnOMgOCaO‰ OK7OP2O5

H,OCO;Sum

Mg#FeO*Fe^/ZFe

48.60.87

16.56.471.740.137.82

12.92.11

<O.Ol0.061.760.25

99.2

65.88.040.20

Trace elements (ppm):SCrNiCuZnGaSrYZr

LiScRbNbLaCeNdSmEuGdTbDyHoErTmYbLuThUTh/UZr/NbLa/NbTh/Nb(La/Sm)N

(Sm/YbV

240R-141-45

1 ]

2007.31273

S-PPDDarkpatch

P45

48.90.63

17.36.251.770.138.79

13.01.880.010.041.65

100.4

68.97.840.20

<20371135

84415581731

0.88390.110.270.712.723.321.450.652.150.403.010.611.860.271.840.280.0160.0062.88

1162.630.0600.320.88

240R-195-99

222007.85

274M-OPCD

Patch+ vugs

P70

48.90.65

16.66.672.060.138.47

12.51.910.010.041.760.11

99.8

66.38.530.22

<20370126

84414582035

241R-I11-15

42016.61

276M-POCD

Dark(+ patch?)

D30

49.30.80

17.16.071.590.148.47

14.01.89

<O.Ol0.051.32

100.8

69.17.500.19

30339125144515572140

1.17410.090.310.943.494.2S1.850.712.610.503.810.752.310.332.220.320.0170.0062.91

1313.030.0530.33

241R-142-46

92016.92

276M-POCD

Darkact.veins

D30

47.90.79

16.85.951.880.138.73

13.71.80

<O.Ol0.061.550.07

99.3

69.37.640.22

70334134153615562643

241R-195-98

192017.45

276M-POCD

Halo + amph.veins

H65

48.70.76

16.45.962.030.148.47

13.51.82

<O.Ol0.051.590.1

99.6

68.37.780.23

<20330126

84015552339

24IR-1117-118

202017.67

276M-POCD

Patch

P70

48.20.67

17.36.201.760.138.52

13.01.900.010.041.62

99.3

68.47.780.20

<20383135

84113591632

0.78410.100.200.702.623.171.470.602.050.392.940.581.840.261.810.260.0150.0052.95

1232.710.0580.310.90

242R-I0-5

12025.90

276M-POCDPatch +

vugs + darkP60

49.00.71

16.26.512.440.149.32

12.81.830.010.051.720.09

100.8

68.08.700.25

35345145304113532337

245 R-l48-51

162043.28

281S-PPDDark

D50

49.70.90

15.66.622.190.138.24

13.12.030.010.061.130.1

99.7

65.58.590.23

8026696323316572747

246R-18-11

42052.28

283S-POD

Halo + amph.veins + dark

H30

48.70.83

16.06.712.460.169.13

12.71.79

<O.Ol0.061.340.08

100.1

67.08.930.25

310311137914816552643

(79°-84°), it is possible that a small number of dike packets was pen-etrated during Leg 148, although the thickness of such packets wasestimated to be less than 8.5 m (Umino, 1995). Another possibility isto assume that alteration caused the compositional trends in such away that chlorite formation increases MgO (Ni?) and H2O, decreasesSiO2, and the larger amounts of secondary minerals dilute incompat-ible elements like Ti and Zr. This dilemma makes a close inspectionof the possible chemical effects of magmatic differentiation and hy-drothermal alteration necessary.

Magmatic Differentiation

The effects of fractional crystallization of plagioclase and olivine(± clinopyroxene) on major-element compositions are well con-strained (e.g., Bryan et al., 1976). We used crystallization models ofNielsen (1990) and Weaver and Langmuir (1990) to assess the chem-

ical effects of fractional crystallization in rocks from Hole 504B. Adetailed discussion of the modeling is beyond the scope of this paper.We thus present a summary of the modeling results and their impli-cations.

1. Crystallization of primitive Hole 504B basalts starts either witholivine or plagioclase, depending on the initial composition. Becauseof the generally high Ca/Na ratios of the basalts, in many cases pla-gioclase is the first liquidus phase and the plagioclase compositionsare unusually calcic (Ang8_94; Natland et al., 1983).

2. The first liquidus phase is joined by the second within the first4 wt% of crystallization, after which olivine and plagioclase crystal-lize in a cotectic proportion of 2:1.

3. Clinopyroxene joins the fractionating mineral assemblage after15%—30% of the initial melt is crystallized. Spinel is also an early liq-uidus phase and crystallizes mostly along with plagioclase.

44


Table 4 (continued).

Core, section:Interval (cm):Piece no.:Depth (mbsf):Lith. unit:Rock type:Alteration type:


246R-126-30

92052.46

283S-PODDark+patch

P40

Major elements (wt%):SiO,TiO~2

A12O,FeOFe 2 O,MnOMgOCaONa2OK2OP2O,H,OCO;Sum

Mg#FeO*Fe3+/ZFe

49.40.87

15.56.492.780.168.86

12.91.85

<O.Ol0.061.270.09

100.2

66.18.990.28

Trace elements (wt%):SCrNiCuZnGaSrYZr

LiScRbNbLaCeNdSmEuGdTbDyHoErTmYbLuThUTh/UZr/NbLa/NbTh/Nb(La/Sm)N(Sm/Yb‰

3103701161045014562744

246R-146-50

142052.66

283S-PODDark

D20

50.10.88

15.86.942.660.178.90

12.91.87

<O.Ol0.061.14

101.4

65.49.330.26

560362123895515552342

0.6945

0.060.300.943.754.401.820.742.700.513.790.762.370.342.320.330.0150.0062.54

1403.110.0490.330.87

246R-166-70

192052.86

283S-PODDark

D20

48.90.86

15.26.652.860.178.70

12.91.83

<O.Ol0.061.320.09

99.5

65.19.220.28

3303961131056013542744

246R-1103-107

302053.88

284S-PODDark+patch

P45

48.80.73

16.06.582.110.159.13

13.01.74

<O.Ol0.051.54

99.8

68.18.480.22

<20379132224014541934

0.7945

0.090.280.752.863.471.600.632.350.413.020.621.900.271.820.270.0120.0052.37

1222.700.0410.300.98

247R-118-26

62056.88

285M-POD

Dark

D20

48.70.82

15.46.602.970.169.14

12.81.76

<O.Ol0.061.24

99.7

66.19.270.29

920376145985414542341

1.0542

0.050.2')0.983.674.311.750.722.5')0.503.620.752.320.332.280.330.0130.0043.29

1383.350.0450.360.S5

247R-130-33

82057.00

285M-POD

Dark

D35

48.40.76

16.06.032.780.159.15

13.21.71

<O.Ol0.051.780.11

100.1

68.08.530.29

5603991401154715522440

247R-146-50

122057.16

285M-POD

Halo + amph.vein +dark

H60

49.30.77

15.86.661.850.149.05

13.11.820.010.061.560.09

100.2

68.38.330.20

<20377124

103414532440

247R-176-79

182057.46

286M-OPD

Dark

DND

49.70.95

15.27.262 . 1 60 . 1 78.37

12.91.91

<O.Ol0.070.93

99.6

64.39.210.21

1010333102957215652450

0.9143

0.040.361.214.525.202.080.792.900.544.020.792.420.362.380.350.0250.0092.88

1393.380.0710.380.97

248R-155-58

162062.35

288M-CPD

Halo

HND

49.20.92

15.16.792.440.168.62

13.01.87

<O.Ol0.061.23

99.3

65.58.990.24

350336111765014

- 612347

1.0344

0.240.461.404.995.372.150.843.100.574.080.822.540.352.300.340.0270.0092.81

1013.050.0580.421.04

249R-14-8

22071.24

290M-POD

Dark

D30

49.61.03

14.87.342.590 . 1 9

8.3713.02.11

<O.Ol0.071.34

100.3

63.19.670.24

210267

86775315662753

1.0645

0.230.591.535.466.182.350.953.450.614.250.902.650.382.500.370.0310.0083.69

902.580.0530.421.04

4. A maximum of 50% crystallization of a parental melt, corre-sponding to cooling from 1260° to 1170°C, can account for the chem-ical variability of most elements (except very mobile elements like S,H2O, and Cu and highly incompatible elements like Th, LREEs, Nb,etc.). The larger variability of mobile elements is clearly an effect ofhydrothermal alteration, whereas the larger variability of highly in-compatible elements has been previously explained by heteroge-neous mantle (e.g., Tual et al., 1985) or multi-stage melting (e.g.,Autioetal, 1989).

5. The net chemical consequences of advanced fractional crystal-lization are decreases in MgO, A12O3, CaO, Cr, and Ni and increasesin SiO2, TiO2, FeO*, Na2O, Zr, REE, and other incompatible ele-ments. Concentrations of Cu, Sr, and Co are not drastically changedwhen the degree of fractional crystallization is small (< 50%).

6. Fractional crystallization cannot significantly change the ratiosof highly and moderately incompatible elements (e.g., Nb/Zr) andmoderately and mildly incompatible elements (e.g., Zr/Ti).

Chemical Effects of Alteration

Details of the diverse alteration mineralogy of the lower SDC arediscussed in Alt et al. (this volume) and Vanko et al. (this volume).The basic chemical changes that accompany the various types of al-teration are listed in Table 5.

One possible way to assess the degree of chemical alteration is todetermine H2O

+ concentrations, because water can only be gainedduring alteration in these rocks. The problem is that secondary phaseshave different water contents. For example, talc is formed at lowwater-rock interaction, but has high water contents and thus obscuresthe correlation between water contents and degree of alteration.

Figure 5 plots selected elements vs. the degree of alteration withthe different rock types represented by different symbols. The weakcorrelation observed between water content and the degrees ofalteration indicates a net gain of water because of the formation ofsecondary minerals (amphibole, chlorite, talc, etc.). Very high water

45


Table 4 (continued).

Core, section:Interval (cm):Piece no.:Depth (mbsf):Lith. unit:Rock type:Alteration type:


249R-171-76

242071.91

290M-POD

Dark

D30

Major elements (wt%):SiO2TiO2A12O3

FeOFe2O,MnOMgOCaONa2OK2OP2O,H2OCO,Sum

Mg#FeO*Fe^/IFe

50.30.84

16.46.432.280.158.80

13.41.85

<O.Ol0.061.35

101.8

67.38.480.24

Trace elements (ppm):SCrNiCuZnGaSrYZr

LiScRbNbLaCeNdSmEuGdTbDyHoErTinYbLuThUTh/UZr/NbLa/NbTh/Nb(La/Sm)N(Sm/Yb)N

10033212138

15

2243

0.9841

0.170.511.134.084.641.820.742.600.483.400.712.180.302.050.290.0300.0074.34

832.220.0590.400.99

249R-10-5

12072.68

291M-POCD

Dark

D10

49.40.87

16.16.822.070.168.25

13.11.88

<O.Ol0.060.88

99.6

65.38.680.21

970326109866915662146

0.9643

0.150.401.214.404.991.930.772.870.493.380.722.110.312.030.300.0240.0082.93

1163.060.0610.411.06

250R-114-21

42080.54

291M-POCD

Dark

D15

49.70.86

15.96.812.020.168.27

13.01.93

<O.Ol0.061.14

99.8

65.58.630.21

570325104826616612242

1.2642

0.130.371.184.314.912.000.832.900.533.610.732.270.322.100.29

<0.02<O.Ol

1133.19

0.381.06

250R-198-102

282081.38

291M-POCD

Halo + amph.vein + dark

H40

49.40.85

15.66.452.230.168.65

12.81.930.010.061.630.11

99.9

66.98.450.24

70386111

94416592545

251R-112-17

42090.02

293M-POCD

Patch+ vugs

P60

48.50.67

15.57.311.880.149.83

12.52.020.010.052.520.1

101.0

68.49.000.19

<2()355147

64913602134

1.8838

0.250.582.513.121.290.621.930.352.900.611.850.301.930.300.0130.0052.69

1362.300.0540.290.74

251R-172-76

112090.62

293M-POCD

Dark

D30

48.60.72

16.56.681.990.159.45

13.01.57

<O.Ol0.051.91

100.5

68.88.470.21

146036616093491451:o36

1.4444

0.130.320.963.524.011.670.692.450.463.120.662.080.301.940.28

<0.02<O.Ol

1123.03

0.370.96

251R-176-80

122090.66

293M-POCD

Green(patch?)

P60

48.40.72

15.86.821.910.159.46

13.01.60

<O.Ol0.051.640.1

99.6

68.78.540.20

20316140

183914532336

251R-1105-109

202090.95

293M-POCD

Dark

D20

48.70.78

16.36.312.330.159.23

13.41.70

<O.Ol0.061.470.06

100.5

68.58.410.25

3103331391714214542439

252R-14 - 8

22099.44

293M-POCD

Patch+ dark

P60

48.10.76

16.26.801.880.14

10.0112.9

1.61<O.Ol

0.052.070.07

100.6

70.08.490.20

120319167434215502438

252R-115-20

52099.55

293M-POCD

Patch+ vugs

P80

48.90.54

15.16.952.110.159.65

12.31.960.020.041.440.11

99.3

68.48.850.21

<20375147

64912611929

0.2737

0.230.542.172.911.120.681.790.342.550.571.630.231.740.250.0220.0082.60

1242.340.0940.310.72

Note: See text for sample descriptions and analytical precisions. Group letters H, P, and D explained in text. Lith. unit = lithologic unit. Rock type abbreviations are as follows: S-sparsely phyric, M- = moderately phyric, P = plagioclase, O = olivine, C = clinopyroxene, and D = diabase, act. = actinolite, and amph = amphibole. ND = not determined. Mg# =molar ratio (Mg/[Mg + Fe 0.9]) x 100.

concentrations (up to 6 wt%) as reported from rocks of Leg 140(Zuleger et al., 1995) have not been observed in Leg 148 diabases.

Following are other observations on chemical trends with percent-age of alteration:

1. Possible slight decreases of CaO and SiO2 and possible in-creases of MgO (chloritization, Table 5). However, thesetrends are poorly defined.

2. Variable enrichments and depletions of Na2O resulting fromthe formation of secondary albite-oligoclase (gain) and chlo-rite (loss).

3. Drastic decreases in TiO2, MnO, Zr, and Ce (and other REEs)in the more altered rocks, which are particularly obvious in thealteration patches.

4. Dramatic decreases in Cu and S, probably because of thebreakdown of primary sulfides (e.g., chalcopyrite), and locallyredistribution of Cu and S through precipitation of secondarysulfides (cf. Alt et al., 1995).

5. Overall decreases in Zn concentrations from 70 to 30 ppmwithin the dark diabases with 10%-50% alteration and a pos-sible slight increase in Zn contents within the most extensivelyaltered patches.

Most trends are more obvious if the alteration halos are excluded,suggesting that the particularly complex alteration mineralogy of thehalos obscures chemical trends. On Figure 6, many elements (e.g.,Ca, Ti, Na, S, Cu, Zn, Zr) show considerable compositional scatteronly above about 20% alteration. However, some of the composition-

46


ü

requ

en

u_

req

uen

cy

u_

^uen

cy

SLL

12

10

8

6

4

2

0

12

10

8

6

4

2

0

25

20

15

10

5

0

-

1 ,111111

1 1 1

Patches

-

- —i10 20 30 40 50 60 70 80 90 100

Alteration (%)

1 1 1

_

1

1 1 1 1

Halos

-

>i

10 20 30 40 50 60 70 80 90 100Alteration (%)

Dark gray

LJ !

10 20 30 40 50 60 70 80 90 100Alteration (%)

Figure 3. Histograms of degrees of alteration in the different rock types ofthe lower sheeted dike complex in Hole 504B. Shaded bars = Leg 140(Zuleger et al., 1995), and open bars = Leg 148 (this work). Note that thebackground alteration within the dark diabases is slightly higher in Leg 148.

al variability is probably the result of primary magmatic differencesin rock chemistry.

Nature of the Patches

One of the most intriguing results of studying the chemistry of theSDC drilled at Site 504 is the peculiar depletion of incompatible ele-ments in the alteration patches. In this paper, we discuss several pro-cesses that could be related to the formation and the depleted natureof the patches and try to evaluate the merits and limitations of differ-ent models.

Several authors studied the chemistry of alteration patches in thelower SDC in Hole 504B and their findings and conclusions differ inmany respects. Zuleger et al. (1995) reported that REEs (and Zr, Ti,Y) are increasingly depleted as the rocks become more altered, andthat the depletion is particularly high in alteration patches. Sparks(1995) stated that REEs, Zr, Ti, and Y are not depleted in rocks up to60% alteration and that the depletions are restricted to alterationpatches. Gorton et al. (1995) interpreted variations in REE and Zr asreflecting mantle source heterogeneities and observed significantlosses of Cu, Zn, and Ti and some loss of La. Kepezhinskas et al.

(1995) thought that the variable REE depletions represent originalmagmatic characteristics of the crust. In contrast to the findings ofZuleger et al. (1995), Kepezhinskas et al. (1995) reported slight Laand Ce increases with progressive degree of alteration and virtualimmobility of the other REEs.

Our results unequivocally support the finding of Zuleger et al.(1995) that REE concentrations are lower in highly altered rocks andare particularly low in alteration patches. This is demonstrated in Fig-ure 6A where REE patterns of increasingly altered rocks are plotted.A remarkable feature is the positive Eu anomaly displayed by themost altered samples. Figure 6B shows the average REE concentra-tions of dark diabases and alteration patches with filled areas repre-senting the standard deviations. On average, the alteration patcheshave lower REE concentrations and slightly positive Eu anomalies.There are more chemical differences between patchy and dark dia-bases. Table 6 summarizes a comparison of the mean chemical com-positions of these two rock types. Elements with significant changesare printed in italic style, whereby we defined differences exceedingthe standard deviations as significant. Beside the REEs, Nb, Zr, Ti,Cu, Zn, and S also show significant losses, whereas water shows sig-nificant gains. Figure 7 shows the compositional differences betweenpatch and dark diabase in the order of increasing degree of depletion.S and Cu are most extensively depleted in the patches followed by theLREEs, Nb, Th, MREEs, Zr, Zn, Eu, HREE, Ti, Mn, P, Fe, Sr, Ca,and Si. Enrichments are indicated only of Mg, Al, Cr, Ni, and H2O.The REEs (plotted as inset in Fig. 7) show larger depletions in LREE(La, Ce, Nd) as compared to the heavier REEs. Eu displays a smallerdegree of depletion than the neighboring elements.

Mobilization of Nominally Immobile Elements

Rare-earth elements and elements of high field strength (Zr, Nb,Ti) are generally considered as being immobile during most alter-ation processes (e.g., Pearce and Norry, 1979). Gillis et al. (1992)argued that REE concentrations either would not change or would in-crease during rock alteration at high temperatures. In contrast, Sparks(1995) suggested the mobilization of REEs (and Zr, Ti, and Y) duringintensive alteration of the patches in Hole 504B and discussed thepossibility that the alteration patches could be a source for the highREE concentrations in high-temperature hydrothermal fluids (e.g.,Michard, 1989).

Figure 8 depicts schematic patterns of such a hydrothermal fluidand a hypothetical fluid carrying the REEs lacking in the alterationpatches (cf. Fig. 7). In contrast to the marked positive Eu anomalytypical for all high-temperature hydrothermal vent fluids, the "patchfluid" would have a negative Eu-anomaly. As the hydrothermal flu-ids continuously react with the rocks such that calcic plagioclase,oligoclase, and albite (in order of decreasing fluid temperature) areformed (e.g., Alt et al., 1995), Eu would become more depleted in thefluid in the course of upwelling, thus leading to even more pro-nounced negative Eu anomalies.This sort of consideration makes asimple relationship between the REE depletion in the alterationpatches and the REE enrichment in hydrothermal fluids unlikely.

In general, a mobilization of such large amounts of (nominallyimmobile) elements is difficult to envision. Chloride complexeswould transport the LREEs more efficiently than the HREEs (Wood,1990, and references therein), but they are unlikely to leach largequantities of REEs unless the water/rock ratios are extremely high(>105, Michard and Albarède, 1986). Carbonate complexes as poten-tial carriers of REEs cannot account for this depletion because the hy-drothermal fluids are acidic and hence do not contain sufficientamounts of carbonate or bicarbonate ions. Fluoride complexes wouldpreferentially transport the heavy REEs (Walker and Choppin, 1967)which would cause the opposite trend of what has been observed.Sulfate complexes are important in acidic solution but show littlepreference between the LREEs and the HREEs (Wood, 1990, and ref-erences therein).

47


Alteration (%) SiO2 (wt%) MgO (wt%) TiO2 (wt%)

50 100 46 48 50 52 7 β 9 10 11 O.i

Na2O (wt%) H2O (wt%)

1.5 2 2.5 0 2 4 6

1000 2000 0 50 100 150 200 250 0.1 0.15 0.2 20 40

:

1700 • I > I D

Figure 4. Chemical variation of the lower sheeted dike complex with depth. Open symbols = Leg 140 (Zuleger et al. 1995), and solid symbols = Leg 148 (thiswork; Alt, Kinoshita, Stokking, et al., 1993). Squares = dark diabases, circles = patches, triangles = halos, and diamonds = light diabases (only Leg 140). Shad-ing marks regions with abundant patches.

Little is known about the mobilities of nominally immobile ele-ments when a rock reacts with very high-temperature fluids (500°C),as was the case for the alteration halos and the alteration patches. Itis, however, interesting that the patches are depleted but the halos arenot. In short, in view of these theoretical considerations, a profoundincompatible element depletion of the patches through alteration ishard to explain.

Mass-balance Calculations

To make a more quantitative approach to assess possible elementtransports during alteration, we used the mass-balance calculationmethods of Gresens (1967) and Grant (1986).

Strict application of mass-balance calculations to the problem ofelement mobilization in basaltic rocks has shown that in many casesvolume changes during metamorphism can account for the changesin REE contents rather than mobility of REEs (Bartley, 1986).

On the other hand epidosites in ophiolites sometimes exhibit near-ly quantitative removal of Zr and Y (Schiffman and Smith, 1988).These epidosites are different from other metabasites: they are totallyrecrystallized into granoblastic textures. This texture and the deple-

tions of Mg, Na, and sometimes Al argue for very high water/rockratios of about 500 (Schiffman et al., 1990; Nehlig et al., 1994).

Gorton et al. (1995) declined the application of mass-balance cal-culations to Hole 504B rocks, because of too large primary chemicaldifferences of the dikes. We therefore restrict our mass-balance con-sideration to one section (Section 140-504B-193R-1) where Zulegeret al. (1995) analyzed a piece of dark diabase adjacent to a virtuallycompletely recrystallized patch. A Grant plot showing the gains andlosses in various elements is depicted in Figure 9. If the chemicalcomposition had not been changed, the data would plot along the sol-id lines, assuming no volume changes or mass changes.

In Figure 9, most elements plot below the solid lines (i.e., wouldhave been mobilized during alteration). The plot indicates gains inH2O, Rb, and Cr, and virtually no changes in Si, Mg, Al, and Fe. Ex-cept Cu and S, elements like Zr, REEs, Nb, and Th would be the mosteffectively mobilized ones. Another approach to mass transport phe-nomena is to assume that certain elements are immobile. As dis-cussed earlier, REEs, Zr, Nb, and so on, are commonly thought to beimmobile. These elements plot along the dashed line in Figure 9 andthe slope of this line is proportional to the volume change of the al-teration process. All elements plotting above the dashed line must


Table 5. Simplified chemical effects of various types of alteration.

UralitizationCpx—>Amph

ChloritizationOl/Cpx/Plg Chl

AlbitizationPlg Ab

EpidotizationPlg Epi

Ilmenite-TitaniteIlm Tit

Loss Gain Loss Gain Loss Gain Loss Gain Loss Gain

SiO2

CaO

HoO SiO2

CaOA12O3

MgO

H,O

CaO Na2O

A12O3 SiO2

SiO, CaO

H 2O

FeO CaO

TiO2(?) SiO2

Note: Cpx = clinopyroxene, Amph = amphibole, Ol = olivine, Pig = plagioclase, Chi = chlorite, Ab = albite, Epi = epidote, Ilm = ilmenite, and Tit = titanite.

•5- 0.8

J!• Φ

• * m

—i—i—i—

o

I-1—h

,;.. Φ

Δ 0 Δ

6 Δ *0w y

Φ " *

1 1 ^H %

20 40 60 80 100 20 40 60 80 100

14 -

13 -

12 .

11

#

•™':::::: ft:*-::

• # ÷• •~•• A

' ' 1 ' 1

Δ .

| |* Φ 0 O

P *

-1 1 l—l—1 1

20 40 60 80 100

(Ul

a.

N

70

60

50

40

3 0

20

10

0 -

m *** +

- S:-: :•*:ji O A Δ é

1 1 1 1 1 1 1 1 1

20 40 60 80 100

Alteration (%)

20 40 60 80 100

Alteration (%)

20 40 60 80

Alteration (%)

Figure 5. Variation of rock chemical compositions with degree of alteration. Symbols as in Figure 4.


A iθη

Φ 5 "

T3

O

ü

2 -

249R-2, 0-5 cm, 10% alt., dark247R-1,18-26 cm, 20% alt., dark249R-1, 71-76 cm, 30% alt., dark246R-1, 103-107 cm, 45% alt., patch251R-1, 12-17 cm, 60% alt., patch241 R-1, 117-118 cm, 70% alt., patch252R-1, 15-20 cm, 80% alt., patch

La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

B 20π

1 0 -

ë sIuoDC

1

• Dark

• Patch

La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

Figure 6. Rare-earth element (REE) patterns of selected samples from Leg148. A. Various extents of REE depletion within the dark diabases and thepatches. B. Mean REE contents of dark diabases and patches, respectively.The stippled areas represent standard deviation of the mean. Note thatpatches are significantly depleted in REEs and have positive Eu anomalies.

have been added to the rock. This approach yields volume changes ofaround 100% which are unrealistic as primary magmatic textures arepreserved in many patches. Bartley (1986) considered volume chang-es of up to 30% as a maximum for rocks showing preservation of pri-mary textures. Gresens's calculations also yield unrealistically highvolume factors of up to 2. This holds true not only for the dark/patchpair of Section 140-504B-193R-1 but for many other dark/patch pairsthat have been taken into consideration.

We conclude that alteration might be capable of producing deple-tions in REEs, Y, Ti, Zr, Nb, etc., to some extent, but that theobserved dramatic depletion of these elements in the patches throughalteration is improbable, especially because the major elements arevirtually unchanged. Moreover, alteration halos around amphiboleveins have similar alteration mineralogy and some are also intensive-ly altered, but halos do not show the same sort of depletion, suggest-ing that the chemical depletion of the patches might be of primarynature.

Primary Depletion of the Patches

Zuleger et al. (1995) proposed a primary depletion of the patchesthrough expulsion of late-stage, highly differentiated interstitial melt,

leaving a rock depleted in incompatible elements.This process musthave been somehow related to the formation of high porosities thatled subsequently to high fluid flows through these more permeableareas. The high fluid flow resulted in intense water-rock interactionand high degree of recrystallization. Evidence for the presence ofsuch highly evolved interstitial melts are provided by small pocketsof granophyric quartz and albite with minor clinopyroxene, apatite,and Fe-Ti oxides (Dick, Erzinger, Stokking, et al., 1992). The diffi-culty with this concept is to find a physically reasonable process forthe separation of these melts from the solid residue (i.e., a drivingforce for the liquids). Exsolution and expansion of volatiles in mag-mas has been discussed as a possible mechanism in Zuleger et al.(1995). Indeed, volume expansion of volatile-rich magmas and coin-ciding segregation of gas vesicles has been documented recently inbasalts from the Lau Basin (Bloomer, 1994). However, the Lau Basinexample may not be strictly comparable to the situation at Site 504,because the primary water contents of these back arc basin basaltswere high and the eruption depth relatively shallow. Taking the prim-itive and depleted nature of Hole 504B basalts into account, the pri-mary water contents must have been very low, maybe in the order of0.1 or 0.2 wt% (cf. Michael, 1988). The solubility of water in basalticmelts at 1200°C and 1 kbar is as high as 4 wt% (Burnham, 1975). Itwould thus need at least 95-98 wt% fractional crystallization to get amelt supersaturated in water.

Exsolution of CO2 could be an alternative mechanism, since thesolubility of CO2 in basaltic melts is low (0.44 ppm/bar; Stolper andHolloway, 1988) and nucleation and growth of CO2 bubbles starts al-ready at great depths (e.g., Bottinga and Javoy, 1990). However, tak-ing into account the molar volume of CO2 at 1200°C and 1 kbar, itwould take 0.25 to 2.5 wt% CO2 exsolved to account for a porosity of2 to 20 vol% (calculated assuming a perfect gas). The primary CO2

contents are probably lower than 400 ppm (Fine and Stolper, 1986)and therefore such high CO2 concentrations are unrealistic. More-over, high degrees of supersaturation result in drastically lower nu-cleation rates of the CO2 bubbles, thus inhibiting exsolution andsegregation of CO2 bubbles particularly because ascent velocities ofbasaltic melts in dikes are high (1 m/s; Delaney and Pollard, 1982).

The amygdules found in many patches are thought to representprimary pore space ranging from 2% to 20%. The subrounded shapesof many amygdules are similar to gas vesicles and exsolution of gascould potentially drive out interstitial melt leading to the depleted na-ture of the patches. However, because of the simple calculationsmade above we feel that volatile-rich magma expansion is not verylikely to occur in dikes of the oceanic crust. We therefore favor thepossibility that the amygdules represent primary pore space producedby shrinkage during crystallization of the melt and subsequent cool-ing of the rock. The idea of late-stage melt expulsion must not neces-sarily be rejected, if these interstitial liquids were just squeezed outmechanically (e.g., during dike compaction).

This puzzling question remains: Why are the patches depleted inincompatible elements and highly altered at the same time? We donot favor alteration to account for the depletion and we hardly canthink of a reasonable physical process that may have expelled orsqueezed out late-stage interstitial melts.

Yet another problem with the late-stage melts as an explanation isthat they would be enriched in Zr and REEs but not in Ti, whichwould decrease after FeTi-oxides join the fractionating mineralassemblage (Fig. 10). The observed simultaneous depletion of Ti andZr, REEs, Nb, etc. is therefore difficult to explain by removal of late-stage melts. A third possibility would be to regard the patches asxenocrystic material that was incorporated in the melt upon dike in-trusion. If the material were highly altered, water-rich diabases orgabbros, this could account for both the high alteration and the chem-ically distinct (depleted) compositions. Petrographic studies of thepatches, however, do not support a xenocrystic nature of the patches,since clear textural changes across the patch margins have never beenobserved.


Table 6. Comparison of compositional range and average compositions of dark diabases and patchy diabases.

Min.

Major elements (wt%)SiO2TiO2

A12O3FeO*MnOMgOCaONa2OP2O,H2O

47.880.72

13.787.500.138.24

12.481.570.030.30

Trace elements (ppm):SCrNiCuZnGaSrYZrLiScRbNbLaCcNdSmEuGdTbDyHoErTmYbLuThU

20

861431134819360.69

340.040.290.943.494.011.670.692.150.463.240.662.110.302.030.280.0130.004

Dark (TV

Max.

50.331.03

17.149.750.199.88

14.032.110.071.91

1538399160171

7216662953

1.44470.230.591.535.466.182.350.953.450.614.250.902.650.382.500.370.0310.009

= 28, 9)

Mean

49.420.84

15.448.750.168.87

13.031.820.051.01

745312120805015572345

1.0641

0.120.381.124.134.771.920.772.750.513.680.752.310.332.220.320.0220.007

SD

0.660.070.780.530.020.410.300.130.010.45

53849193111

15250.2130.060.100.190.630.650.200.080.350.040.320.060.170.030.170.030.0070.002

Min.

48.110.54

14.857.780.128.47

12.331.600.040.68

20247101

627125016290.27

370.090.230.542.172.911.120.601.790.342.550.571.630.231.740.250.0120.005

Patch (N = 12,5)

Max.

50.340.87

17.309.000.16

10.0113.352.020.062.52

3083831671045016612748

1.88480.110.280.752.863.471.600.682.350.413.020.621.900.301.930.300.0220.008

Mean

48.860.71

16.048.450.149.13

12.871.830.051.58

1183431362541145621370.92

410.100.260.662.583.201.390.642.050.382.880.601.820.271.830.270.0160.006

SD

0.590.090.750.430.010.530.290.140.010.47

106431727

713360.5940.010.020.090.260.210.190.030.210.030.190.020.100.020.070.020.0040.001

Notes: Numbers in italic typeface indicate elements that are significantly different in patches. N = number of XRF and ICP-MS analyses, respectively. Min. = minimum, Max. = max-imum, and SD = standard deviation.

Zn, Cu, and S Depletions

Dick, Erzinger, Stokking, et al. (1992) reported a generally de-creasing trend of Zn concentrations in the lower part of the SDC in504B (>1600 mbsf), the so-called "reaction zone." At the same timeCu and S show various extents of depletion on the meter scale andhave a crude bimodal distribution. This was explained by Zuleger etal. (1995) and Sparks (1995) as evidence for silicate-controlled be-havior of Zn and sulfide-controlled behavior of Cu and S. In rocksfrom Leg 148, Zn and Cu show a poor covariation and Cu exhibits abimodal distribution (Fig. 11). Most of the dark diabases have highCu concentrations around 80 ppm, which is close to the Cu concen-trations of fresh lavas from Hole 504B (Hubberten et al., 1983). Mostof the patches and halos have low Zn and Cu concentrations, with Cubeing more extensively leached than Zn.

Simultaneous Cu and Zn depletion are known from many ophio-lites and such base-metal depleted zones have been referred to asreaction or root zones of hydrothermal fluids (e.g., Richardson et al.,1987; Schiffman et al., 1990; Nehlig et al., 1994). The base-metaldepletion in ophiolites, however, occur in epidosites which are ex-tremely altered and are often intercalated with more normal green-schist diabases. These epidosites are thought to be conduits forextremely high fluid flow, often parallel to dike margins (e.g., Nehliget al., 1994). A general Zn depletion as observed in the lower SDC ofHole 504B has been reported by Harper et al. (1988) from the basalSDC of the Josephine ophiolite.

Alt et al. (1995) proposed that sulfide dissolution, sulfide precip-itation and sulfate formation occurring all over the lower SDC areresponsible for the perplexing pattern of depth variation of Cu and S.In the section drilled during Leg 148, Cu and S show similar behavior

to that in the upper sections. The general trend of progressive Zn de-pletion with depth from 1600 to 2000 mbsf changes somewhat in theLeg 148 section (2000-2111 mbsf). Although low Zn contents simi-lar to the overlying dikes are still present, higher Zn contents also oc-cur, particularly below 2050 mbsf. This general trend is displayed byour data and the "shipboard data set" of Leg 148 (Alt, Kinoshita,Stokking, et al., 1993). On the diagram of Zn vs. degree of alteration(Fig. 5), Zn contents of dark diabases and halos generally decrease,but increase in the patches. In our opinion, two explanations for sucha behavior of Zn are possible: first, a phase in the patches retains Zn,or more likely, takes up Zn (lost during alteration of the dark diabas-es) from hydrothermal fluids, and second, the patches are altered athigh temperatures where the solubility of Zn is decreased. The latterpossibility was also proposed by Sparks (1995) on the basis of resultsof Seyfried and Janecky's (1985) leaching experiments, which indi-cated a maximum solubility of Zn and Cu between 375° and 400°Cand markedly decreased solubility at higher temperatures. This couldmean that the rocks of the SDC were altered by pervasive fluids ofoptimal temperature for Zn leaching (375°-400°C) in depths wherethe Zn depletion is most obvious (1700-2000 mbsf). Trends of in-creasing 18O depletion continue to the final depth of Leg 148, sug-gesting that temperatures of pervasive alteration are indeed higherthan in the cores recovered during Leg 140 (Alt et al., this volume).

CONCLUSION

Rocks of the lower SDC drilled during Leg 148 (2000-2111mbsf) exhibit similar chemical characteristics to those previouslyrecovered from Hole 504B. They are also highly depleted in incom-

51


CDQ

1α.

1 I I I I I I I I I I I I I I I I I I I I I 1 I I I I 1 I I I I I I I I

Figure 7. Bar chart, showing average element concentrations in alteration patches divided by the average concentrations in dark diabases in order of increasingdepletion. Numbers < 1 indicate depletion (loss), whereas numbers > 1 indicate enrichment (gain). The inset shows the same sort of diagram for the REEs.

High-temperaturehydrothermal ventfluids

Hypotheticluids carrying REEs

mobilized from patches

CO

Figure 8. Schematic sketch, demonstrating the difference in REE pattern ofhydrothermal fluids and a hypothetical fluid that transports the REEs mobi-lized from patches. Seawater pattern is shown for reference. Cone. = concen-tration, and arb. units = arbitrary units. See text for discussion.

patible elements and show a small range in chemical compositions,most of which can be explained by moderate degrees of magmaticdifferentiation.

A crude sawtooth pattern in the depth variation of Zr, Ti, and Ni,which is most obvious between 1900 and 2111 mbsf, could be inter-

preted as magmatic trends suggesting intrusions of multiple dikepackets.

Hydrothermal alteration in the lowermost section is generallymore intense than in the section above. The effects of this alterationon the rock chemistry are slight gains of MgO and H2O and losses ofCaO, TiO2, and SiO2.

Alteration halos and alteration patches are widespread in the low-er SDC. Alteration patches are most abundant from 2000 to 2030mbsf and between 2085 and 2111 mbsf, where average grain sizes arelarge, which promotes high alteration of the interior parts of dikepackets. The alteration halos show no systematic chemical trends andno clear relationship to the degree of alteration.

In contrast to alteration halos, patches of intense alteration are sig-nificantly depleted in REE, Zr, Nb, Ti, etc., and exhibit positive Euanomalies. Theoretical considerations and mass-balance calculationsindicate that this depletion is probably not solely the consequence ofintensive alteration, but rather reflects primary differences betweenthe patches and the neighboring diabases. It remains unclear howsuch differences could be related to magmatic differentiation pro-cesses within the dikes.

The alteration patches are significantly enriched in H2O and high-ly depleted in Cu (Zn) and S. This makes the lower SDC a possiblesource for base-metal deposits in the oceanic crust.

Regardless of whether or not the REE depletions of the patchesare a matter of alteration, the patches cannot be simply regarded assources for the REEs in high-temperature hydrothermal fluids. Ifhydrothermal fluids inherited their REEs inventory solely during for-mation of the patches, they would have negative Eu anomalies in-stead of positive ones.

The questions of the formation of alteration patches, their miner-alogical (but not chemical) similarity to alteration halos, and the pe-

52


σ>

it

o

tooó

CCCOσ>

t>u "

5 0 -

4 0 -

3 0 -

2 0 -

1 0 -

0 -

- Patch • H2O

Rb / /

/y A I

T j Th -×

f

•Cr

A

' Nb

' Sc

• N a

• L i. Z r ^ - "

•Ce

D a r / c1 • 1 • • • • 1

0 10 20 30 40 50 60

193R-1, 44-46 cm, P =3.08 g/cc

Figure 9. Grant plot, after Grant (1986), showing the difference between apatch (Sample 140-504B-193R-1, 58-61 cm) and a neighboring dark diabase(Sample 140-504B-193R-1, 44-46 cm). V = volume, m = mass, and p = pyc-nometric density. See text for details.

1050

Temperature (°C)

1100 1150 1200

5 -•

4

n

ü oOoC oβ) ^

W1

InterstitialA _. melts in dike°_. melts in dike

° solidification zone o La (ppm)

A Zr/50 (ppm)

α TiO2 (wt%)

H h4 6

MgO (wt%)

10

Figure 10. Trends of magmatic differentiation within a solidification zone,calculated using the program of Nielsen (1990). The model assumes a paren-tal melt composition of 10 wt% MgO, 0.7 wt% TiO2, 30 ppm Zr, and 0.8ppm La. Note that Ti concentrations would decrease after a certain amount ofcrystallization (about 80%), whereas Zr and REEs would continuouslyincrease. The model assumes the amount of fractionation occurring withinthe solidification zone before recharge to be 70% (cf. Nielsen, 1990). Sym-bols represent compositions of interstitial melts in a hypothetical solidifica-tion zone. Solid lines represent compositions of the melts in the dike interiorwhere no crystallization takes place. The kink in the Ti trend is at 80 wt% ofcrystallization.

EQ.α

80

70 -–

60 -–

50

40 -–

N 30

20 -–

10 -–

50 100

Cu (ppm)

150 200

Figure 11. Zn vs. Cu for lower SDC rocks of Leg 148. Symbols as in Figure4. Shading marks the approximate primary range of Hole 504B lavas.

culiar depletion in nominally immobile elements remains largely

unanswered and needs further attention.

ACKNOWLEDGMENTS

We thank M. Grünhàuser, H.-G. Plessen, H. Rothe, and R. Zierfor their cordial assistance in doing the chemical analyses. The nu-merous discussions with E. Zuleger have been particularly helpful.The paper benefitted from a review by P. Schiffman. This study wasin part supported by Grant No. Er 123/7-1 of the Deutsche For-schungsgemeinschaft (DFG) to J.E., and by grants from JOI-USSACto J.C.A.

REFERENCES

Alt, J.C., Anderson, T.F., Bonnell, L., and Muehlenbachs, K., 1989. Mineral-ogy, chemistry, and stable isotopic compositions of hydrothermallyaltered sheeted dikes: ODP Hole 504B, Leg 111. In Becker, K., Sakai,H., et al., Proc. ODP, Sci. Results, 111: College Station, TX (OceanDrilling Program), 27-40.

Alt, J.C., and Emmermann, R., 1985. Geochemistry of hydrothermallyaltered basalts: Deep Sea Drilling Project Hole 504B. In Anderson, R.N.,Honnorez, J., Becker, K., et al., Init. Repts. DSDP, 83: Washington (U.S.Govt. Printing Office), 249-262.

Alt, J.C., Honnorez, J., Laverne, C, and Emmermann, R., 1986. Hydrother-mal alteration of a 1 km section through the upper oceanic crust, DeepSea Drilling Project Hole 504B: mineralogy, chemistry, and evolution ofseawater-basalt interactions. /. Geophys. Res., 91:10309-10335.

Alt, J.C., Kinoshita, H., Stokking, L.B., et al., 1993. Proc. ODP, Init. Repts.,148: College Station, TX (Ocean Drilling Program).

Alt, J.C., Laverne, C, and Muehlenbachs, K., 1985. Alteration of the upperoceanic crust: mineralogy and processes in Deep Sea Drilling ProjectHole 504B, Leg 83. In Anderson, R.N., Honnorez, J., Becker, K., et al.,Init. Repts. DSDP, 83: Washington (U.S. Govt. Printing Office), 217-247.

Alt, J.C., Zuleger, E., and Erzinger, J.A., 1995. Mineralogy and stable isoto-pic compositions of the hydrothermally altered lower sheeted dike com-plex, Hole 504B, Leg 140. In Erzinger, J., Becker, K., Dick, H.J.B., andStokking, L.B. (Eds.), Proc. ODP, Sci. Results, 137/140: College Station,TX (Ocean Drilling Program), 155-166.


Autio, L.K., and Rhodes, J.M., 1983. Costa Rica Rift Zone basalts: geochem-ical and experimental data from a possible example of multistage melt-ing. In Cann, J.R., Langseth, M.G., Honnorez, J., Von Herzen, R.P.,White, S.M., et al., Init. Repts. DSDP, 69: Washington (U.S. Govt. Print-ing Office), 729-745.

Autio, L.K., Sparks, J.W., and Rhodes, J.M., 1989. Geochemistry of Leg 111basalts: intrusive feeders for highly depleted pillows and flows. InBecker, K., Sakai, H., et al., Proc. ODP, Sci. Results, 111: College Sta-tion, TX (Ocean Drilling Program), 3-16.

Barrett, T.J., 1983. Strontium- and lead-isotope composition of some basaltsfrom Deep Sea Drilling Project Hole 504B, Costa Rica Rift, Legs 69 and70. In Cann, J.R., Langseth, M.G., Honnorez, J., Von Herzen, R.P.,White, S.M., et al., Init. Repts. DSDP, 69: Washington (U.S. Govt. Print-ing Office), 643-650.

Bartley, J.M., 1986. Evaluation of REE mobility in low-grade metabasaltsusing mass-balance calculations. Nor. Geol. Tidsskr., 66:145-152.

Becker, K., Foss, G., et al., 1992. Proc. ODP, Init. Repts., 137: College Sta-tion, TX (Ocean Drilling Program).

Bloomer, S.H., 1994. Origin of segregation vesicles in volcanic rocks fromthe Lau Basin, Leg 135. In Hawkins, J., Parson, L., Allan, J., et al., Proc.ODP, Sci. Results, 135: College Station, TX (Ocean Drilling Program),615-623.

Boström, K., and Bach, W., 1995. Trace element determinations by X-rayfluorescence analysis: advantages, limitations, and alternatives. InBatiza, R., Storms, M.A., and Allan, J.F. (Eds.), Proc. ODP, Sci. Results,142: College Station, TX (Ocean Drilling Program), 61-68.

Bottinga, Y., and Javoy, M., 1990. Mid-ocean ridge basalt degassing: bubblenucleation. J. Geophys. Res., 95:5125-5131.

Bryan, W.B., Thompson, G., Frey, F.A., and Dickey, J.S., 1976. Inferredgeologic settings and differentiation in basalts from the Deep Sea Drill-ing Project. J. Geophys. Res., 81:4285^304.

Burnham, C.W., 1975. Waters and magmas: a mixing model. Geochim. Cos-mochim. Acta, 39:1077-1084.

Delaney, P.T., and Pollard, D.P., 1982. Solidification of basaltic magma dur-ing flow in a dike. Am. J. Sci., 282:856-885.

Detrick, R., Collins, J., Stephen, R., and Swift, S., 1994. In situ evidence forthe nature of the seismic layer 2/3 boundary in oceanic crust. Nature,370:288-290.

Dick, H.J.B., Erzinger, J., Stokking, L.B., et al., 1992. Proc. ODP, Init.Repts., 140: College Station, TX (Ocean Drilling Program).

Emmermann, R., 1985. Basement geochemistry, Hole 504B. In Anderson,R.N., Honnorez, J., Becker, K., et al., Init. Repts. DSDP, 83: Washington(U.S. Govt. Printing Office), 183-199.

Etoubleau, J., Corre, O., Joron, J.L., Bougault, H., and Treuil, M, 1983.Costa Rica Rift: variably depleted basalts in the same hole. In Cann, J.R.,Langseth, M.G., Honnorez, J., Von Herzen, R.P., White, S.M., et al., Init.Repts. DSDP, 69: Washington (U.S. Govt. Printing Office), 765-773.

Fine, G., and Stolper, E., 1986. Dissolved carbon dioxide in basaltic glasses:concentrations and speciation. Earth Planet. Sci. Lett., 76:263-278.

Garbe-Schoenberg, C.-D., 1993. Simultaneous determination of thirty-seventrace elements in twenty-eight international rock standards by ICP-MS.Geostand. Newsi, 17:81-97.

Gillis, K.M., Ludden, J.N., and Smith, A.D., 1992. Mobilization of REE dur-ing crustal aging in the Troodos Ophiolite, Cyprus. Chem. Geol., 98:71-86.

Gorton, M.P., Schandl, E.S., and Naldrett, A.J., 1995. Platinum group ele-ment (PGE) concentrations in a sheeted dike complex and geochemicalchanges during alteration of the lowermost part of a 2-km-thick oceaniccrust, Hole 504B. In Erzinger, J., Becker, K., Dick, H.J.B., and Stokking,L.B. (Eds.), Proc. ODP, Sci. Results, 137/140: College Station, TX(Ocean Drilling Program), 199-206.

Govindaraju, K., 1994. 1994 compilation of working values and sampledescription for 383 geostandards. Geostand. Newslett., 18.

Grant, J.A., 1986. The isocon diagram—a simple solution to Gresens's equa-tion for metasomatic alteration. Econ. Geol., 81:1976-1982.

Gresens, R.L., 1967. Composition-volume relationships of metasomatism.Chem. Geol., 2:47-65.

Harper, G.D., Bowman, J.R., and Kuhns, R., Jr., 1988. A fluid, chemical, andstable isotope study of subseafloor metamorphism of the Josephine Ophi-olite, California-Oregon. J. Geophys. Res., 93:4625^656.

Hey, R., Johnson, G.L., and Lowrie, A., 1977. Recent plate motions in theGalapagos area. Geol. Soc. Am. Bull, 88:1385-1403.

Hobart, M.A., Langseth, M.G., and Anderson, R.N., 1985. A geothermal andgeophysical survey on the south flank of the Costa Rica Rift: Sites 504

and 505. In Anderson, R.N., Honnorez, J., Becker, K., et al., Init. Repts.DSDP, 83: Washington (U.S. Govt. Printing Office), 379^04.

Hofmann, A.W., 1988. Chemical differentiation of the Earth: the relationshipbetween mantle, continental crust, and oceanic crust. Earth Planet. Sci.Lett, 90:297-314.

Hubberten, H.-W., Emmermann, R., and Puchelt, H., 1983. Geochemistry ofbasalts from Costa Rica Rift Sites 504 and 505 (Deep Sea DrillingProject Legs 69 and 70). In Cann, J.R., Langseth, M.G., Honnorez, J.,Von Herzen, R.P., White, S.M., et al., Init. Repts. DSDP, 69: Washington(U.S. Govt. Printing Office), 791-803.

Jenner, G.A., Longerich, H.P., Jackson, S.E., and Fryer, B.J., 1990. ICP-MS:a powerful tool for high-precision trace-element analysis in earth sci-ences: evidence from analysis of selected U.S.G.S. reference samples.Chem. Geol, 83:133-148.

Jochum, K.P., Seufert, H.M., and Thirlwal, M.F., 1990. Multi-element analy-sis of 15 international standard rocks by isotope-dilution spark sourcemass spectrometry. Geostand. NewsL, 14:469^173.

Kempton, P.D., Autio, L.K., Rhodes, J.M., Holdaway, M.J., Dungan, M.A.,and Johnson, P., 1985. Petrology of basalts from Hole 504B, Deep SeaDrilling Project, Leg 83. In Anderson, R.N., Honnorez, J., Becker, K., etal., Init. Repts. DSDP, 83: Washington (U.S. Govt. Printing Office), 129-164.

Kepezhinskas, P.K., Sorokina, N.A., Mamontova, S.A., and Savichev, A.T.,1995. Rare-earth and large-ion lithophile (Sr and Ba) element geochem-istry of diabase dikes, Hole 504B, Costa Rica Rift, Leg 140. In Erzinger,J., Becker, K., Dick, H.J.B., and Stokking, L.B. (Eds.), Proc. ODP, Sci.Results, 137/140: College Station, TX (Ocean Drilling Program), 107-116.

Kusakabe, M., Shibata, T., Yamamoto, M., Mayeda, S., Kagami, H., Honma,H., Masuda, H.,and Sakai, H., 1989. Petrology and isotope characteristics(H, O, S, Sr, and Nd) of basalts from Ocean Drilling Program Hole 504B,Leg 111, Costa Rica Rift. In Becker, K., Sakai, H., et al., Proc. ODP, Sci.Results, 111: College Station, TX (Ocean Drilling Program), 47-60.

Laverne, C, Vanko, D.A., Tartarotti, P., and Alt, J.C, 1995. Chemistry andgeothermometry of secondary minerals from the deep sheeted dike com-plex, Hole 504B. In Erzinger, J., Becker, K., Dick, H.J.B., and Stokking,L.B. (Eds.), Proc. ODP, Sci. Results, 137/140: College Station, TX(Ocean Drilling Program), 167-190.

Marsh, N.G., Tarney, J., and Hendry, G.L., 1983. Trace element geochemis-try of basalts from Hole 504B, Panama Basin, Deep Sea Drilling ProjectLegs 69 and 70. In Cann, J.R., Langseth, M.G., Honnorez, J., VonHerzen, R.P., White, S.M., et al., Init. Repts. DSDP, 69: Washington(U.S. Govt. Printing Office), 747-763.

Michael, P.J., 1988. The concentration, behavior and storage of H2O in thesuboceanic mantle: implications for mantle metasomatism. Geochim.Cosmochim. Acta, 52:555-566.

Michard, A., 1989. Rare earth element systematics in hydrothermal fluids.Geochim. Cosmochim. Acta, 53:745-750.

Michard, A., and Albarède, F., 1986. The REE content of some hydrothermalfluids. Chem. Geol, 55:51-60.

Natland, J.H., Adamson, A.C., Laverne, C, Melson, W.G., and O'Hearn, T.,1983. A compositionally nearly steady-state magma chamber at theCosta Rica Rift: evidence from basalt glass and mineral data, Deep SeaDrilling Project Sites 501, 504, and 505. In Cann, J.R., Langseth, M.G.,Honnorez, J., Von Herzen, R.P., White, S.M., et al., Init. Repts. DSDP,69: Washington (U.S. Govt. Printing Office), 811-858.

Nehlig, P., Juteau, T., Bendel, V., and Cotten, J., 1994. The root zone of oce-anic hydrothermal systems: constraints from the Samail ophiolite(Oman). J. Geophys. Res., 99:4703^713.

Nielsen, R.L., 1990. Simulation of igneous differentiation processes. InNicholls, J., and Russell, J.K. (Eds.), Modern Methods of Igneous Petrol-ogy: Understanding Magmatic Processes. Rev. Mineral., 24:65-105.

Pearce, J.A., and Norry, M.J., 1979. Petrogenetic implications of Ti, Zr, Yand Nb variations in volcanic rocks. Contrib. Mineral. Petrol, 69:33-47.

Richardson, C.J., Cann, J.R., Richards, H.G., and Cowan, J.G., 1987. Metal-depleted root zones of the Troodos ore-forming hydrothermal systems,Cyprus. Earth Planet. Sci. Lett., 84:243-253.

Schiffman, P., Bettison, L.A., and Smith, B.M., 1990. Mineralogy andgeochemistry of epidosites from the Solea Graben, Troodos Ophiolite,Cyprus. In Malpas, J., Moores, E., Panayiotou, A., and Xenophontos, C.(Eds.), Ophiolites: Oceanic Crustal Analogues. Proc. Symp. "Troodos1987," Min. Agr. Nat. Resour., Nicosia, Cyprus, 673-684.

Schiffman, P., and Smith, B.M., 1988. Petrology and oxygen isotopegeochemistry of a fossil seawater hydrothermal system within the Solea

54


Graben, northern Troodos ophiolite, Cyprus. J. Geophys. Res., 93:4612-4624.

Seyfried, W.E., Jr., and Janecky, D.R., 1985. Heavy metal and sulfur trans-port during subcritical and supercritical hydrothermal alteration of basalt:influence of fluid pressure and basalt composition and crystallinity.Geochim. Cosmochim. Acta, 49:2545-2560.

Shimizu, H., Mori, K., and Masuda, A., 1989. REE, Ba, and Sr abundancesand Sr, Nd, and Ce isotopic ratios in Hole 504B basalts, ODP Leg 111,Costa Rica Rift. In Becker, K., Sakai, H., et al., Proc. ODP, Sci. Results,111: College Station, TX (Ocean Drilling Program), 77-83.

Sparks, J.W., 1995. Geochemistry of the lower sheeted dike complex, Hole504B, Leg 140. In Erzinger, J., Becker, K., Dick, H.J.B., and Stokking,L.B. (Eds.), Proc. ODP, Sci. Results, 137/140: College Station, TX(Ocean Drilling Program), 81-98.

Stolper, E., and Holloway, J.R., 1988. Experimental determination of the sol-ubility of carbon dioxide in molten basalt at low pressure. Earth Planet.Sci. Lett, 87:397^108.

Tual, E., Jahn, B.M., Bougault, H., and Joron, J.L., 1985. Geochemistry ofbasalts from Hole 504B, Leg 83, Costa Rica Rift. In Anderson, R.N.,Honnorez, J., Becker, K., et al., Init. Repts. DSDP, 83: Washington (U.S.Govt. Printing Office), 201-214.

Umino, S., 1995. Downhole variations in grain size at Hole 504B: implica-tions for rifting episodes at mid-ocean ridges. In Erzinger, J., Becker, K.,

Dick, H.J.B., and Stokking, L.B. (Eds.), Proc. ODP, Sci. Results, 137/140: College Station, TX (Ocean Drilling Program), 19-34.

Walker, J.B., and Choppin, G.R., 1967. Thermodynamic parameters of fluo-ride complexes of the lanthanides. Adv. Chem., 71:127-140.

Weaver, J.S., and Langmuir, C.H., 1990. Calculation of phase equilibrium inmineral-melt systems. Comput. Geosci., 16:1-19.

Wood, S.A., 1990. The aqueous geochemistry of the rare-earth elements andyttrium, 1. Review of available low-temperature data for inorganic com-plexes and the inorganic REE speciation of natural waters. Chem. GeoL,82:159-186.

Zuleger, E., Alt, J.C., and Erzinger, J.A., 1995. Primary and secondary varia-tions in major and trace element geochemistry of the lower sheeted dikecomplex: Hole 504B, Leg 140. In Erzinger, J., Becker, K., Dick, H.J.B.,and Stokking, L.B. (Eds.), Proc. ODP, Sci. Results, 137/140: CollegeStation, TX (Ocean Drilling Program), 65-80.

Date of initial receipt: 23 August 1994Date of acceptance: 29 January 1995Ms 148SR-114

4. CHEMISTRY OF THE LOWER SHEETED DIKE COMPLEX, HOLE …

Documents

Transcript of 4. CHEMISTRY OF THE LOWER SHEETED DIKE COMPLEX, HOLE …