Seawater sulfate reduction and sulfur isotope ... - Science Direct

2 downloads 0 Views 2MB Size Report
May 19, 1981 - (SHANKS and BISCHOFF, 1977, 1980), and the recently discovered Zn and Cu sulfides of the East Pacific Rise at 21”N (HEKENIAN et al., ...

Seawater sulfate reduction and sulfur isotope fractionation in basaltic systems: Interaction of seawater with fayalite and magnetite at 200-350°C W. C. SHANKS III Department of Geology

and Geophysics,


of Wisconsin,


WI 53706. U.S.A

and JAMES L. BISCHOFFand ROBERTJ. ROSENBAUER U.S. Geological (Receiced



23 June 1980; accepted

Park, CA 94025, U.S.A in recised,form

19 May 1981)

Abstract+Sulfate reduction during seawater reaction with fayalite and with magnetite was rapid at 35O’C, producing equilibrium assemblages of talc-pyrite-hematite-magnetite at low water;rock ratios and talc-pyrite-hematite-anhydrite at higher water/rock ratios. At 25O”C, seawater reacting with fayalite produced detectable amounts of dissolved H,S, but extent of reaction of solid phases was minor after 150 days. At 2OO”C, dissolved H2S was not detected, even after 219 days, but mass balance calculations suggest a small amount of pyrite may have formed. Reaction stoichiometry indicates that sulfate reduchydrothermal systems is provided by Mg tion requires large amounts of H+, which, in subseafloor metasomatism. Seawater contains sufficient Mg to supply all the H+ necessary for quantitative reduction of seawater sulfate. Systematics of sulfur isotopes in the 250 and 350’C experiments indicate that isotopic equilibrium IS reached, and can be modeled as a Rayleigh distillation process. Isotopic composition of hydrothermally produced H,S in natural systems is strongly dependent upon the seawater/basalt ratio in the geothermal system, which controls the relative sulfide contributions from the two important sulfur sources, seawater sulfate and sulfide phases in basalt. Anhydrite precipitation during geothermal heating severely limits sulfate ingress into high temperature interaction zones. Quantitative sulfate reduction can thus be accomplished without producing strongly oxidized rocks and resultant sulfide sulfur isotope values represent a mixture of seawater and basaltic sulfur.


SULFATEreduction has often been cited as an explanation for the sulfur source for sulfide mineral deposits. Sulfur isotope values of sulfide deposits are often lighter than contemporaneous seawater sulfate, which forms a large, homogeneous, and readily available sulfur reservoir. Thus, the suggestion of sulfate-reductive processes is reasonable. Bacterially mediated sulfate reduction in the pore waters of reducing marine sediments is well established (GOLDHABERand KAPLAN, 1974) and produces iron sulfides with sulfur isotope values variably more negative by 1545 per mil with respect to the initial seawater sulfate. Such processes have likely been important in the formation of some metallic mineral deposits. Inorganic sulfate reduction during seawater hydrothermal processes is less well understood. However, sulfate reduction resulting from interaction of seawater with ferrous iron bearing minerals in submarine volcanic rocks has been suggested for a number of massive sulfide deposits, including the Kuroko deposits (KAJIWARA, 1973), the Cyprus copper deposits (HEATONand SHEPPARD, 1977), the Raul deposit (RIPLEY and OHMOTO, 1977), the Red Sea brine deposits (SHANKS and BISCHOFF, 1977, 1980), and the recently discovered Zn and Cu sulfides of the East Pacific Rise be more reactive than magnetite and the magnetitehematite buffer would only be operatlvc ufter the basalt was significantly altered and relatrvell oxidized. The small but persistent quantities of dissolved sulfate in the final samples from 77-2 ma] aiso he related to the sluggishness of magnetite oxidation. The persistence and abundance of fayahte in the reaction products of the 250 and 200FC experiments indicates very limited reaction progress. Reactions (1I

Seawater sulfate reduction and sulfur isotope fractionation


60 40







l+--u-o--o-o.o_ o-o-o-o-o-o-



IO 0 6 4 A-A-A-A-A-&

G 0 E 2

I 0.0 0.6 0.4



St08rting sea water

10 ‘at




Fig. 6. Solution composition and pH trends for experiment 77-12, fayalite-amorphous interaction at 2OO”C,500 bars, and a water/rock ratio of 130.

and (4) probably represent the processes adequately, but even magnetite production in the solids was very limited (Table 2). In 77-6, however, there was a fairly rapid production of small amounts of H,S, which requires the formation of minor magnetite. Anhydrite and talc also precipitated in all three of these lower temperature experiments (Table 2). Phaw


Phase diagrams for the system Fe-S-O-Ca-Si are expressed in terms off,, andf,, in Fig. 7 (a, b, and c) at 350, 250, and 2OOC, respectively, and 500 bars total pressure. Thermodynamic data for construction of these diagrams are derived from the data bank of H. Helgeson and his co-workers (HELGESON and KIRKHAM, 1974a, b, 1976; HELGESON et a/., 1978, 1980). Equilibrium constants of appropriate reactions were computed using the program SUPCRT (HELGESON er al., 1978). The most reducing assemblage expected in these experiments is fayaliteemagnetite-pyrrhotite (Fig. 7). However, neither pyrrhotite nor pyrrhotite pseudo-


morphs were detected in any run products. Apparently, a pyrrhotite assemblage cannot exist at equilibrium until all anhydrite and dissolved sulfate in the system are eliminated by sulfate reduction (Fig. 7). The high water/rock ratios of these experiments require complete destruction of fayalite long before equilibrium can be reached. Thus, in natural rock alteration processes where seawater is the aqueous fluid, we would expect to find pyrrhotrte only where low water/rock ratios prevail. Indeed, MOTTL et al. (1979) observed pyrrhotite in one of their low water/ rock experiments at 300°C and several higher temperature experiments. Phase relations in Fig. 7 (b and c) indicate that the coexistence of fayalite and anhydrite in the 200 and 250°C experiments is a non-equilibrium assemblage. Because the concentration of sulfate predominates over ferrous iron at the water/rock ratios of these experiments, fayalite must be considered the metastable phase. Fayalite was completely destroyed in experimennts 77-2 and 77-3. Thus, all three 350°C experiments are


W. C.


III et al



-3oI -60


m I \ b





1 6























Fig. 7. Phase relations in the system Pe-~S&Ca Si at 500 bars total pressure and (a) 350 C, (bl 250 t’ and (c) 2oi)‘C. Mineral stability fields in terms of oxygen and sulfur fugacities calculated using the data of HELGEKIN et ul. (1979). Anhydrite solubility lines constructed using solution chemistry of experiment, 77-3, 77-6, and 77-12 (see text for details). Numbered symbols represent seawater fugacity relations as ,I function of time for individual samples from specific experiments.

expected to correspond to the magnetite--hematite pyrite assemblage. Again, however, the presence of sulfates and the water/rock ratio must be considered. Sulfate is abundant enough in 77-3 and 77-11 to completely oxidize all magnetite, and thus the equilibrium assemblage in these experiments is hematite-pyriteanhydrite. The final mineralogy observed in 77-3 approaches this assemblage, but magnetite persists metastably, as in 77-l 1. Experiment 77-2, on the other hand, was conducted with a predominance of ferrous iron over sulfate. Thus the predicted, and observed, final assemblage is magnetite-hematite -pyrite (except for very minor anhydrite). In summary, examination of phase relations in the

system Fe- SO-Cam~Si emphasizes the importance of water/rock ratio. At very low ratios the expected assemblage is fayalite -magnetite--pyrrhotite and at intermediate ratios, hematite--magnetite pyrite. At higher water/rock ratios, sulfate and sulfide can coexist, and the stable assemblage is anhydritepyrite-hematite. sample

Distribution of’ species und mine4


Activities of dissolved species were calculated using the FORTRAN program SOLVEQ developed by REED (1977). This program uses the NewtonRaphson method to converge on the equilibrium distribution 01 species and calculates individual ion activity coehi-

Seawater sulfate reduction and sulfur isotope fractionation cients using modification

the Debye-Huckef limiting law with for swamping concentrations of NaCl (HELGESON, 1969). In situ pH is calculated using 25°C measurements and conserving mass of total ionizable hydrogen (BISCHOFFand SEYFRIED.1978). Dissociation constants of aqueous complexes are the same as described in BISCHOFFand SEYFRIED(1978), with the following exceptions. The constant for NaSO, is from STYRIIGJVICHrt td. (1968) and ferrous chloride complexes were added to the program using constants from CRERAR et ni. (1978). Dissociation constants for all aqueous species are for the vapor pressure of water at the tem~rature of interest and are uncorrected for the 500 bar pressure of the experiments. Solubility or hydrolysis constants for solid phases were computed using SUPCRT at the temperature of interest and 500 bars total pressure. In additionfo, and&> were calculated from sulfatesulfide equilibria in the solutions to test for agreements of phase relations (Fig. 7) and observed mineral assemblages. The following equations, from SUPCRT at T and 500 bars. were used to computejo, and,f& in the solutions and provide an additional check on the state of equilibrium of the system: H+ $ SO;- = HS- + 20,(g) 2HS + O,(g) + 2H+ = S,(g) + 2H20

(5) (6)

Use of these equations is limited to solutions with detectable amounts of HIS and SO,, and the results are greatly influenced by the computed in situ pH and ion activities from SOLVEQ. The 200°C experiments are excluded due to lack of measurable H,S. A number of sequential samples from experiments 77-3 and 77-6 (Fig. 7) indicate time-sequence trends within a given experiment. Changes from initial conditions indicate significant increase in fs, and slight decreases in ,fo, Thus, equations 5 and 6 are both driven to the right by pH decrease and sulfate reduction. Sulfur and oxygen fugacities for experiments 77-2, 77-3. 77-l I, and 77-6 (Fig. 7) center on the pyritehematite boundary but are shifted from the pyritemagnetite-hematite invariant point. This result is explained by the presence of anhydrite and/or dissolved sulfate in these experiments. Solubility boundaries for anhydrite precipitation are plotted from the reaction. 1:X,

+ 3:202

+ Ca2’

+ Hz0

= 2H+ + CaS04


anhydrite Activities of Ca’+ and HZ0 and in situ pH are from the speciation calculations. The actual values used in constructing the anhydrite lines are from the final aqueous samples of experiments 77”3,.77-6, and 77-12. However, use of values from other samples produces only insignificant shifts in the relative position of this boundary. That the equilibrium assemblage for experiments 77-3. 77-l 1. and 77-6 is. pyrite, hematite, and anhydrite is clearly supported by the oxygen and sulfur


fugacities as plotted in Fig. 7. The final mineral assemblage in 77-2 included magnetite, hematite, pyrite, and minor anhydrite. The solution fugacities for 77-2 (Fig. 7) are slightly more reducing than the other, higher water/rock, 350°C experiments but are still considerably removed from the pyrite magnetite hematite triple point. The reason for this is the persistence of a small amount of sulfate in solution (40 ppm). For sotutions of seawater composition, aqueous sulfate must be significantly below analytical detection limits before the pyrite- magnetite--h~rn~~tite invariant point is approached. Presumably, sulfate would have been further reduced had experiment 77-2 been continued. The slow, approximately exponential decrease in sulfate in the latter parts of this experiment (Fig. 1) may be attributed to slow reaction of magnetite or to slow dissolution of solid anhydrite, some of which may have been partially armored by other precipitates. We now examine the relation between ionic activities computed using SOLVEQ and solubility or hydrolysis reactions for specific minerals using eyuilibrium constants from SUPCRT. There are a number of computer codes designed to do the same thing as SOLVEQ---that is, calculate an equilibriu~n distribution of species. The choice of program is based mainly on convenience, but the results are critically dependent on the choice of dissociation constants for aqueous complexes (NORDSTROMer ul.. 1980). Unforiunately, ambiguity exists for many of the constants, particularly at temperatures above 300 C. The constants chosen for the present case are considered the best currently available. Thus, at the least, application of the species calculations to experiments such as these provides an evaluation of the validity of such distribution models, which are sometimes applied uncritically to natural systems. Results (Table 4) are presented in terms of the Saturation Index, which is here defined as the ratio of the log of the ion activity product (IAP) for a given reaction (from SOLVEQ) to the log of the equilibrium constant (from SUPCRT). In general. agreement between minerals formed and predicted (Table 4) is good but not exact, perhaps due to the unassessed effect of pressure on the dissociation constants for the aqueous species. Quartz, anhdydrite, pyrite. and fayalite (excepting fayalite at 200 C) are close to eyuilibrium with the solution. It is particularly gratifying to note that the two experiments (77-3 and 77-4) for which quartz undersaturation is predicted are the only ones without quartz in the reaction products. The speciation calculations predict talc undersaturation in every experiment except 77-4, even though talc precipitated in all (Table 2). However, the talc mineral formed is intermediate in composition between talc and minnesotaite, thus the solubility constant for this phase likely lies somewhere between these two. The fact that talc is consistently undersaturated and rninnesotalte is consistently oversaturated suggests equilibrium might exist. The exception is experiment 77-4.


1988 Table 4.

111et d

Solubility relations for appropriate minerals. Calculated using species distribution in situ pH, fg2 from reduced/oxidized sulfur species, and equilibrium constants from SUPCRT (Helges~n et al., 1978). _-


Final Solids*

Predicted Saturation Index (log UP/K)





fay -0.81 ianh,mt, talc]



f&y, SiO (am) it,?,, anb,qtz, mt ,pvl

!).iY -0.4s



fay, qtz[anh, py,talc, nlL'

!!.35 - IJ. f+ .A -0.45











-ii.,r', i.3'!

.il.2 6 2 R 7




11.02 3.09





-1.26 2.78




. ..n7




.- _~~~.___ Tali \linn






h,mt, talc 77-11


qtz,anb G.26 py,hl,mt, talc

0. iH

*Bracketed minerals are indicated from solution-solid mass balance reiat~ons. Abbreviations: fay, fayalite; anh, anhydrite; mt, magnetite; qtz, quartz; py, pyrite; tun, hematite; minn, minnesotaite; SiO2(aml, amorphous silica.

where both phases are oversaturated due to much higher pH. However, there is only minor evidence of Mg removal and fayalite reaction in this experiment. Thus, reaction progress according to eqn (4) is and the interdependence of extremely limited, magnesium removal and H ’ production is again emphasized. The predicted Saturation Indices for magnetite. fayalite at 200°C. and perhaps hematite (Table 4) do not agree well with observation. In experiments 77-2, 77-3, and 77-11 magnetite is clearly present as an intermediate reaction product and yet the speciation calculations predict that it is oversaturated by 3.9 and 3.3 orders of magnitude, respectively. Similar results are calculated for 77-6. An even more severe contradiction is the apparent oversaturation of fayalite in the 200°C experiments. where it is clearly apparent from solution chemistry that fayalite is dissolving (Figs 5 and 6, Table 2). Phase relations do not preclude hematite being oversaturated as predicted (Table 4), but it seems unlikely that oversaturation by 3 orders could persist in very long duration experiments like 77-l 1. However, the generally good predictions for all other minerals suggest that the fundamental solubility data for hematite, magnetite, and fayalite (at 200°C) are in error. Pyrite, hematite, and magnetite solubility could not be calculated in the 200°C experiments because of the lack of measurable dissolved sulfide, which is needed to establish fo,. However, dissolved iron content in

77-12 is very stable, probably due to equilibrtum with one of these minerals. It seems likely that iron COW centration is controlled by pyrite saturation, as in the higher temperature experiments. If this is the case. then the predicted amount of dissolved sultide in equilibrium with the Fe concentration of 77-I.! would be about 0.08 ppm, which is below the detection limit. In addition, the sulfur budget (Table 2) suggests the presence of a small amount of pyrite. Accepiing this, it is important to note that sulfide was produced fairly rapidly in 77-12. However, the change in the solubility constant for pyrite is dramatic in this temperature range, and the amount of dissolved sulfide is minimal.

Sulfide was produced by sulfate reduction m the 25@‘C experiment (Fig. 4) and in all of the 350 C experiments (Figs 1, 2, and 3) within 10-54 days. At 200°C steady-state Fe concentrations occurred in experiment 77-12 (Fig. 6) after 1X days, possibly due to pyrite saturation. SAKAI and DICKSON (1978) studied the rate of sulfur isotope exchange between aqueous sulfate and sulfide. They measured a half-time of isotope exchange of 6.9 days at 300°C pH = 5.5, and total dissolved sulfur of 0.01 mol/l. In addition, they suggested that the rate of the istopic exchange reaction is first order with respect to both total dissolved sulfur and hydrogen ion activity. Half-times of exchange for the present experiments can be tentatively estimated from the treatment


sulfate reduction

and sulfur isotope

of SAKAI and DICKSON (1978) to be on the order of a few minutes at 350°C a few days at 250°C and a few years at 200°C. Thus, isotope exchange equilibriun would be expected for the 350 and 250°C experiments but not for those at 200°C. Evaluation of isotopic equilibria in these experiments is complicated because’ it was not possible to obtain sufficient dissolved sulfide for isotopic analysis. However, the question can be approached using the calculated final sulfate values (Table 3) a Rayleigh distillation model to account for the reservoir effect, and (in 77-3) the isotopic composition of precipitated pyrite. Isotope values for tinal aqueous sulfate prior to quenching were calculated from the following equation :




I where ;5 refers to b3‘% and m refers to concentration in moles. The subscripts quench. anhydrite, and final refer respectively to the total aqueous sulfate after quenching, sulfate redissolved from anhydrite during quenching, and final sulfate in the solution prior to quenching. This equation is written on the assumption that anhydrite, which precipitates initially with a sulfur isotope value of +20.5, does not continuously re-equilibrate isotopically with aqueous sulfate during the course of the experiments. The assumption is supported because solid-state diffusion in the anhydrite is expected to be slow enough to preclude any significant re-equilibration at the time scale of the experiments. ANDERSON (1969) measured self-diffusion of oxygen and carbon in calcite and found half-times of exchange at 350°C of approximately 5 yr. Sulfur dif-



I 8

0 .


\o -



* 5oo400




OL 0










a “4




1 60

“2S Fig. 8. Sulfur isotope fractionation between sulfate and H,S as a function of temperature from experimental studies (IGUMNOV et ul., 1977; SAKAI and DICKSON, 1978) and theoretical calculations (SAKAI, 1968).


0.2 OF







Fig. 9. Rayleigh distillation calculations for aeparatton of sulfur isotopes between SO, and H2S using lsotoptc equrlibrium fractionation factors of 1.018 at 350 C and I.0249 at 250 C (see Fig. 8). Closed circles represent sulfur Isotope values of residual aqueous sulfate for variou\ experrments. plotted on the appropriate distillation Itne. Arrows Indicate the actual range of ‘fraction of aqueous sulfate reduced’ for each’ experiment, as determined by solution chemistry. Agreement between rsotopic and solution chemical values indicates isotopic equilibrium is established.

fusion in anhydrite would likely be as slow or slower than this. Similarly, it is unlikely that precipitated pyrite continuously re-equilibrates with dissolved sulfate and sulfide. Pyrite precipitation is thus a batch process. and the effects on sulfur isotope distribution between residual aqueous sulfate and precipitated pyrite can be calculated using a Rayleigh distillation equation if instantaneous fractionation factors are known, In this case. it is initially assumed that the fractionation factor (a) conforms to the equilibrium values for sulfate-pyrite separation. Recent experimental studres of equilibrium isotopic exchange between aqueous sulfate and H2S (IGUMNOV er d.. 1977; SAKAI and DICKSON. 1978) showed that fractionation at 2500 C is somewhat less than the theoretically predicted value of SAKAI (1968). These results are summarized graphically in Fig. 8. Using these data. we choose equilibrium 1000 In Z~“J values of 18.1 at 350 C and 24.9 at 250°C. The Rayleigh distillation equation is 6 = looo(1 -,1“-



where (r is per mil difference in the isotope value of residual sulfate with respect to initial sulfate, and f is the fraction of unreduced sulfate remaining in the sys-


W. C.


tern. HIS in equilibrium with aqueous sulfate at a givenf has an isotope value computed by

111ef u/

in pyroxenes or in basaltic glass. In addition. about 0.1 wtu/, S occurs in pyrrhotite and minor pentlandite or in glass. Thus, a number of ferrous iron bearing s,, = 6m4 - 1000 In a$& (IO) reducing agents are available and the possibility t)f Results of the Rayleigh distillation calculations are direct mobilization of basaltic S must be considered. plotted in Fig. 9 for both 250 and 350°C. In general, A great deal of attention has been directed at the residual sulfate becomes dramatically heavier as a sig- relation between hydrothermal metamorphism of basalts and the formation of seafloor sulfide deposit\. nificant fraction of H2S is produced by reduction. Instantaneous H2S isotopic values ‘track’ SO4 values Seawater circulation through the oceanic lithosphere and, for any given.f, are always lighter than the sulfate near divergent plate boundaries is apparentlq quite by the equilibrium fractionation value. The average important and, indeed, is required to balanu: oceanic crustal heat budgets (LISTER. 1972; WCUKY and isotope value of H2S produced is thus the weighted SLEEP. 1976). In general, metamorphisnl b> heated mean of the instantaneous values. In order to test the equilibrium assumption, the seawater produces greenstones with spilitic itssemisotope values of residual sulfate (Table 3) for experi- blages consisting mainly of albite, epidote. chlorite, and actinolite. Oxygen isotopic studies of submarmr: ments 77-3, 77-11, and 77-6 are plotted on the appropriate &SO4 line in Fig. 9. If equilibrium is estab- greenstones and ophiolitic rocks confirm that sealished, then thefvalues indicated in this manner from water is the metamorphic fluid (MUEHLF~%U11s and the isotope data should agree with the fraction of CLAYTON,1972; SPOONERrt ul.. 1974). Very high geusulfate reduction calculated from solution chemistry. thermal gradients prevailed (3OO”C;‘km)as evidenced by the presence of actinolite within I km ui‘ the scaThis comparison is not completely straightforward because successive aqueous samples of 5-15 ml each floor (COLEMAN,1977). Detailed mineralogical and removed significant quantities of sulfate from the sys- chemical studies of greenstones dredged from the seatems. Thus, on Fig. 9, a range offvalues from early to floor (HU~HRIS and THOMPSON,1978) and from ophiolites (COISH, 197’7; MUNH~ and KERRKX. t%O: late for each experiment is indicated. Nonetheless, SPOONERand FYFE, 1973) indicate that sucl~ hydroagreement is good, especially for periods near the end thermally metamorphosed rocks are not strongly oxiof experiments 77-3 and 77-11. and the equilibrium dized, except in hematite bearing near-surface zones assumption is supported. where low temperature oxygenated seawater has penAn additional check on equilibrium can be applied using the isotope value of pyrite from 77-3 of 10.34 etrated. In general, hematite is not present and is little affected during (Table 3). For equilibrium, this should represent the primary titanomagnetite greenschist facies metamorphism (ADE-HALL ~‘1 ai.. average value of sulfide produced. The calculated average value of H2S from 77-3 for anf of 0.57 is 9.15 1971; BANERJEE,1980). However, there is a general and pyrite is expected to be heavier than H2S by 1.03 increase in ferric iron content and ferric/ferrous ratto (CORSH,1977; MUNHAand KERRICI-1.1980: H~.~MPHKIS at 350°C [OHMOTO and RYE, 1979). Again, agreement is excellent and sulfur isotopic equilibrium seems very and THOMSON,1978) which may be attributed to sullikely for seawater hydrothermal systems. at least at fate reduction reactions during hydrothermai altcration. Presumably, this ferric iron is in magnetite. temperatures above 250°C. chlorite, and epidote but definitive studies of mineral Naturul basalt-seawater systems chemistry in such rocks have not been carried out. Pyrite is a common accessory in greenstone terrains The results of these experiments confirm the observations of previous studies (SAKAZ and DICKSON, and pyritic Cu-Zn massive sulfide deposits arc important in many ophiolite complexes. Relevant sulfur iso1978; KIYOSU, 1980; IGUMNOV,1976) that inorganic tope studies have been done on massive sulfide deposulfate reduction is geologically rapid at temperatures sits from Cyprus (CLARK,1971: JOZ~NSON, 1972). Newabove 25O”C, at least in acidic solutions. In addition, we have shown that complete sulfate reduction can foundland (BACHINSK~,1977). and the Iberian pyrite belt (MUNHAand KERRICH,1980). The Iberran pyrite occur in certain cases where the amount of oxidizable ferrous iron stoichiometrically exceeds the amount of belt is, of course, not recognized as an ophictlitic ter-rain, but is included here because of clear oxygen issulfate available (Ex~riment 77-2). Both fayalite and magnetite oxidation can cause sulfate reduction, but tope evidence for seawater hydrothermal metamorphism of basaltic rocks (MUNHAand KERR~C 1-i. J9801. magnetite reaction is apparently much slower. Where In all three cases the sulfur isotope values of sulfide sulfate and sulfide coexist in solution, sulfur isotopic minerals are intermediate between primary basaltic equilibrium is attained at 250 and 350°C. Natural oceanic tholeiites, however, are much more sulfide with values close to 0 (SCHNEIDER.1970) and contemporaneous seawater sulfate (CI.A~KX LJ~cl/.. complex than the simple systems we have studied. For 1980). These data strongly suggest the participation ot example, fayalite is only a minor component (12-20 mol %) in forsteritic olivine. Oxides are often seawater sulfate in the sulfide forming process. Moreover, the sulfides generally give a fairly narrow range complex combinations of ilmenohematites and titanomagnetites of variable compositions. The bulk of the of isotopic values which average 13-16 per mil lighter ferrous iron which averages about IO wt% Fe0 occurs than the contemporaneous seawater sulfate. This snp-


sulfate reduction

gests that either temperatures of sulfate reduction were very high, between 380 and 450°C (Fig. 8) or that the fraction of sulfate reduced was fairly large (0.2-0.4) and temperature was somewhat lower. Either of these explanations may be satisfactory if the bulk of the sulfide is derived by seawater sulfate reduction. However, if basalt-derived sulfide constitutes an important fraction of the reduced sulfur in the system, then unreasonably high temperatures or very large fractions of sulfate reduction are required for the seawater sulfur component. The lack of highly oxidized rocks in hydrothermally metamorphosed ophiolitic terrains suggests that this is not the case. Experimental studies of hydrothermal basalt-seawater interaction have contributed substantially to our knowledge of reaction phenomena (MOTTL et al., 1979; BISCHOFF and SEYFRIED, 1978; SEYFRIED and MOTTI., 1978). In particular, these studies have shown that aqueous sulfate is precipitated as anhydrite or magnesium-hydroxysulfate. At low water/rock ratios, where calcium is leached from the rock, sulfate removal from solution by anhydrite precipitation is nearly quantitative. Conversely, basaltic sulfide is soluble in hydrothermal seawater and, at higher water/ rock ratios, is quantitatively transferred to the fluid phase. At lower water/rock ratios aqueous sulfide concentration is limited by pyrite (or pyrrhotite) solubility. Thus, it is unlikely that sulfide derived from sulfate reduction is quantitatively dominant in ophiolitic sulfide deposits, and the sulfur isotopic data remain enigmatic. The recently discovered hot vents and sulfide deposits at 21”N on the East Pacific Rise (HEKINIAN et al., 1980) offer the possibility of direct study of such sulfide forming processes. Indeed, the hottest fluids are apparently sulfide rich and contain little or no sulfate (EDMOND, 1980). Sulfur isotope values of sulfide deposits are 2.54 (HEKINIAN et al., 1980) averaging 17 per mil lighter than contemporaneous seawater. However, currently known seafloor sulfide deposits are not closely analogous to ophiolitic massive sulfides, being substantially smaller and enriched in sphalerite, barite, and anhydrite. In addition, mineral precipitation processes are as yet poorly understood. Sulfur isotope sqstematics

in basalt-seawater


The results of these and other experimental studies, and the relevant observations of ophiolitic massive sulfides allow construction of a simplified model for sulfide formation during basalt/seawater interaction. We present this model with full realization that it is not unique or comprehensive. However, it is consistent with present information and, hopefully, will help define critical variables in further studies of seafloor and ophiolitic deposits, The enigmatic lack of anhydrite in ophiolitic terrains and greenstones dredged from the ocean floor has been discussed by MOTTL et al. (1979) and BrsCHOFF and DICKSON (1975). Anhydrite is known to

and sulfur isotope



precipitate upon heating of seawater above 150°C in experimental studies and in the Reykjanes geothermal system. However, the absence of sulfate minerals in ophiolites and greenstones cannot be attributed to quantitative reduction of seawater sulfate because the altered rocks are generally not highly oxidized. Therefore, the most reasonable explanation, proposed by MOTTL et a/. (1979) is that anhydrite is originally deposited in the upper, lower temperature portions of the system and is redissolved by cooler seawater as the hydrothermal system retrogresses. Thus, only occasional undissolved sulfate mineral remnants are expected, and these have thus far been described in one core from the Deep Sea Drilling Project (DRIXR et al., 1979). What concentration of seawater sulfate might be available for sulfate reduction in the zones of highest temperature seawater-basalt interaction? The amount of sulfate remaining in hydrothermal seawater after anhydrite (and magnesium-hydroxysulfate) precipitation can best be inferred from seawater and basalt experiments. As an upper bound, BISCHOFFand SEYFRIED (1978) found 11 millimolal sulfate at 3OO’C and 9 millimolal at 350°C upon simple heating of seawater. In the present study, we find very similar values in the high water/rock 77-3 experiment at 350°C. If basalt is present, then Ca leached from the rock leads to smaller Sulfate values due to increased anhydrite precipitation. For example, SEVFRIED and BISCHOFF(1981) measured 0.5 millimolal in a diabasee seawater experiment and 0.2 millimolal for basalt glass/seawater, both at 300°C and water/rock mass ratios of 10. Even at 15OC and water/rock ratio of 10, basalt glass/seawater reactions result in sulfate removal to the 3 millimolal level (SEYFRIED and BISCHOFF,1979). No sulfate removal occurs, however. for similar experiments at 7O’C. Thus, anhydrite precipitation during ingress of seawater is an effective filter which limits the amount of seawater sulfate entering the high-temperature portion of the system. It seems likely, then, that the following sequence of events is common and, perhaps, requisite to the generation of sulfide deposits related to basalt-seawater systems. Evolved seawater enters the high temperature portion of the system with only a small fraction of its original 27 millimolal sulfate. Sulfate reduction is quantitative, and only a small amount of the ferrous iron in the rock is utilized. Direct mobilization of basaltic sulfide is still important, and is dominant at most geologically reasonable water/rock ratios. A quantitative assessment of the relative importance of basaltic sulfide and seawater derived sulfide at various water/rock ratios is presented in Figure 10. This diagram is constructed on the assumption that sulfate reduction is rapid and is limited only by the amount of oxidizable ferrous iron in tholeiitic basalt which is taken to be lOwt% FeO. Oxidation is assumed to proceed only to magnetite so that, in effect, only 213 of the ferrous iron can be oxidized. Stoichiometry of sulfate reduction is according to the

W. C.








Fig. 10. Relative abundance of sulfur species for basalt-seawater interaction at tempera~urcs in cxces’r oi Original sulfur concentrations are assumed to be 31.2 mmol/kg in basalt and 1 mmoPkg 11) evolved seawater. Approximate pyrite saturation boundary is empirically derived front basalt seawatrf experiments. 250°C.



8Fei,‘,,,l + 10H’ + SOi-

= 8Fe:,:,,,

f H2S


Figure 10 is specifically constructed for seawater sulfate contributions of 1 millimolal and basaltic sulfide of 3 1.2 millimolal (0.1 wt”i,). For these concentrations, sulfate is quantitatively reduced at all water/rock ratios below 140. Finally, dissolved sulfide concentration is limited by approximate pyrite saturation (about 1 millimolal sulfide in numerous basalt-seawater experiments at 300 and 350°C) and the sulfide in solution is assumed to be homogeneously derived from the two sulfur sources. Several interesting conclusions can be drawn from these calculations. First, the amount of sulfide in solution at the approximate pyrite saturation boundary (SEYFRIEDand BISCHOFF, 1981) is sufficient to form a significant sulfide deposit and would exceed total base metals in solution in most instances. For example, EDMOND (1980) reports Zn values of 110 micromolal and Cu of about 1.5 micromolal for the end-member 350°C solution at 21”N. A second important conclusion is that the fluid is sulfate deficient at all reasonable water/rock ratios and, thus, sulfate-sulfide isotopic equilibrium is not a factor. Finally, it is obvious (Fig. 10) that basaltic sulfide increasingly dominates at all water/rock ratios below 35. Is this proposal consistent with the sulfur isotopic values of natural sulfide deposits? In order to evaluate this we calculate sulfur isotope values for bulk H2S based on the assumptions developed above. The resultant sulfide sufur isotope values are presented in Fig. 11 for a range of water/rock ratios, for initial sulfur isotope values of basaltic sulfide at 0 and of modern seawater at 20. The range of seawater sulfate concentrations (Fig. 11) is chosen to represent the range observed after sulfate mineral precipitation in experimental studies to date. In general, a smaller sea-

water sulfate contribution ~111 be involved In lowe water/rock situations. In natural systems. however. this will be controlled by a number of comphcated factors. Interestingly, comparison of Figs 10 and 1I mdlcates that the bulk sulfur isotope values of the sulfide are very sensitive to even small contribution\ of H2S from sulfate reduction. This is 3 result of the Gmple mass balance between basaltic sulfide and seawater. derived sulfide which has an isotope value of 20 for the quantitative reduction model. Thus. bulk sulfide isotopic values are totally insensitlve to temperature, as long as it is above 250°C so that reduction IS geelogically rapid. Application of this model to 21‘ N and to ophiohtic massive sulfides yields some interesting suggestions. At 21”N, numerous chemical and isotopic indicators (EDMOND, 1980; CRAIG et al., 1980) suggest low water, rock ratios in the l&5 range. Sulfur isotope values average about 3 (HEKINIAN rt ui., 1980) and. thus. this model predicts that about 4 millimolal sulfate entered the system and was reduced. This value cannot be tested directly but it is quite close to the valurs measured during basalt glass seawater mtcraction at 150°C (SEYFRIEDand BISCHOFF, 1979) which may be typical of the anhydrite precipitation zone in the upper, lower temperature portion of the hystem (MOTTL et al., 1979). The ophiolitic massive sulfides have sulfur Isotopic: values which are (13-16 per mil lighter than contemporaneous seawater, whereas the 2l“N sulfides are 17 per mil lighter than seawater. This suggests (Fig. II) that either the water/rock ratios were somewhat higher in the high temperature zones of the ophiolires or that more seawater sulfate was allowed to entct these systems. The former suggestion seems more likely. Indeed, HEATON and SHEPPARD (1977) have suggested that water-rock mass ratios in the stock-


sulfate reduction

and sulfur isotope


Fig. 11. Sulfur isotope values of total dissolved H,S for various values of water/rock ratio. Seawater sulfate which enters the system is quantitatively reduced and homogeneously mixed with basaltic aullide

work zones underlying massive sulfide deposits at Cyprus have water/rock values greater than l-5. SEYFRIED et al. (1979b) suggested values of 10-30. In addition, SEYFRIED and BISCHOFF (1977) have emphasized the importance of higher water/rock ratios in metal transport and this may be a fundamental difference between the small sulfide accumulations at 21”N and the much larger economic deposits in the ophiolitic terrains.



Fayalite reacts rapidly with seawater at 35O”C, to produce magnetite, hematite, iron-bearing talc, anhydrite, pyrite, and, in some cases, quartz and magnesium-hydroxysulfate. During reaction, seawater gains iron, silica, hydrogen sulfide, and hydrogen ion and loses magnesium, calcium, and sulfate. Magnesium removal from solution due to talc or magnesiumhydroxysulfate formation produces hydrogen ion, which is of critical importance in sulfate reduction reactions. I 1 moles of ferrous iron and 2 moles of hydrogen ion are required for every mole of sulfate reduced (eqn 1). Thus water/rock ratio is the pivotal variable which determines whether the final system will be sulfatebearing or sulfate-free. Fayalite/seawater interaction at 350-C produces a nearly sulfate-free system at a 40/l ratio (Table 2), whereas reaction at 130/l produces sulfate-bearing fluids and anhydrite. Magnetite oxidation to hematite also produces sulfide, but at a much slower rate than fayalite oxidation. At 25O’C, fayalite-quartz/seawater interaction produces detectable amounts of H2S, but fayalite oxidation is quantitatively minor. At 2oO”C, dissolved H2S is always below detection but mass balance calculations indicate that a small amount of pyrite may have formed (Table 2).

Phase diagrams for the system Fe S O-Ca Si in terms offo, andf,, variations (Fig. 7) indicate that the equilibrium mineral assemblage will also vary according to water/rock ratio. With increasing water/rock ratios the expected assemblages are fayalite pyrrhotite-magnetite, pyrrhotite-pyrite-magnetite, pyrite magnetite-hematite, and pyrite- hematites anhydrite. respectively. Fayalite/seawater reaction at 350 C at 40/l (77-2) produced the pyrite-magnetite hematite assemblage, and at 130/l (77-3) produced the pyrite hematite-anhydrite assemblage. The other experiments did not reach limiting assemblages. However. solution Jo, and fs, values calculated from dissolved sulfate, sulfide, and pH approach the pyrite- hematite anhydrite invariant point for experiments 77-l I and 77-6. Calculation of distribution of aqueous species and solubility of pertinent minerals (Table 4) indicates good agreement between predicted and observed occurrences of quartz, anhydrite, pyrite, and fayalite. Solubility constants are not available for iron-bearing talc [FeMg,Si,O,,,(OH),]. but oversaturation with mmnesotite and undersaturation with talc suggest equilibrium with the intermediate phase. Magnetite and hematite solubility predictions consistently suggest oversaturation and do not agree well with observation. Sulfur isotope values of dissolved aqueous sulfate indicate that isotopic equilibrium was established between aqueous sulfate and sulfide in the 350 and 250°C experiments. Sulfide produced by sulfate reduction is precipitated as pyrite in a batch process, and isotopic values of aqueous sulfate can be accurately modeled using a Rayleigh distillation equation. Sulfur isotope values of sulfide deposits related to natural basalt-seawater systems can be explained using a dual sulfur source, total sulfate reduction model. The key facet of this model is sulfate precipi-

W. C.



tation as anhydrite, which occurs at all temperatures above 150°C and which allows only a small fraction of seawater sulfate to enter the hotter portions of geothermal systems. This sulfate is quantitatively reduced to sulfide by interaction with ferrous iron bearing minerals and is homogeneously mixed with sulfide leached directly from the rock to produce bulk sulfur isotope values which are dependent only on water/ rock ratio. A~kno~lrdyements---BILL SEYFRIED and Roe Z~ERENBERG provided many helpful discussions. We thank W. MELSON and A. RADTKE for providing the fayalite and magnetite samples, respectively. Critical reviews by RANDY KOSKI. M. ARNOLD, MIKE MOTTL,and DAVECLAGUE were most helpful. JIM O’NEIL and GENE PERRYgenerously opened their

laboratories to us for isotopic analyses. These studies were supported by the National Science Foundation under Grant OCE-7805481 to WCS. Sulfur isotope studies at Northern Illinois University were partially supported by N.S.F. Grant EAR-7911334 to GENE PERRY.


ADL-HALLJ. M., PALMERH. C. and HUBBARDT. P. (1971) The magnetic and opaque petrological response of basalts to regional hydrothermal alteration. Geophys. J. R. As-on. Sot. 24, 131--174. ANDERSON T. F. (1969) Self-diffusion of carbon and oxygen in calcite by isotope exchange with carbon dioxide. J. Geophys. Rex 74, 3918~-3932.

BANERJEES. K. (1980) Magnetism of the oceanic crust: Evidence from ophiolite complexes. J. Geophys. Res. 85. 3557 -3566.

BIXHOFFJ. L. and DICKSONF. W. (1975) Seawater-basalt interaction at 200°C and 500 bars: Implications for origin of sea-floor heavy-metal deposits and regulation of seawater chemistry. Eurth Planet. Sci. Lett. 25, 385-397. BISCHOFFJ. L. and SEYFRIEDW. E. (1978) Hydrothermal chemistry of seawater from 25^ to 35@C. Am. J. Sci. 278, 838-860.


R. (-. (1979)

Gypsum and halite from the Mid-Atlantic Ridge, DSDP Site 395. Earth Planet. Sci. Lett. 42, 98~-103. EDMOND J. M. (1980) The chemistry of the 3% C‘ hot springs at 21 N on the East Pacific Rise. Eo.\ ‘/).:!tt\. .l!li Geophys. Union 61, 992.

GOLDHABER M. B. and KAPLANI. R. (lY74) fhc ,cJunentary sulfur cycle. The Sea, 5. Marim Chemlstr\ i&. I‘ 1) Goldberg), pp. 56%655. Wiley. HEATONT. H. E. and SHEPPARDS. M. F. (1977) likdroeen and oxygen isotope evidence for sea-wa;cr-hyd,-l;thzr;nal alteration and ore deposition. Troodos (‘omplex. Cyprus. Volcanic Processes in Ore Genexf\, 1,‘p. 42 57 Geol. Sot. London Spec. Publ. No. 7 HEKINIANR., FEVRIERM., BISCHOFFJ. L.. PIC,CJI1’. ami SHANKSW. C. (1980) Sulfide deposits from the East Pacific rise near 21 ‘N: A mineralogical and geochemicai study. Science 207, 1433-1444. HELGESON H. C. (1969) Thermodynamics of hydr~~thernt‘~l systems at elevated temperature and pressure. .l!t~ .I. SC1. 267, 729%804. HELGESONH. C. and KIRKHAMD. H. (lY74a) 1heoretlca! prediction of the thermodynamic behavior of aqueous electrolytes at high pressures and temperatures. 1. Summary of the thermodynamic/‘electrostatic prapcrties i>f the solvent. Am. J. Sri. 274, 1089 1198. HELUESONH. C. and KIRKHAMD. H. (1974b) li~eoret~cwl prediction of the thermodynamic behavior of aqueous electrolytes at high pressures and temperatures. II. Debye-Huckel parameters for activity coefficients and relative partial molal properties. .4n1. .I .Sir. 274, 1199-1261. HELGESONH. C. and KIRKHAMD. H. (1076) ‘Theoratlcal prediction of the thermodynamic behavior vt’ aqueous electrolytes at high pressures and temperatures. III Equation of state for aqueous species at infinltt. diluti~)n Am. J. Sci. 276, 97-240.

HELGESON H. C., DELANYJ. M.. NESBI~I H. W and BIHI) D. K. (1978) Summary and critique of the thcrmod?namic properties of rock-forming mineral4 .lii~ .1. vi, 278A, 1~299. HELGESONH. C., KIRKHAMD. H. and FLOMI~KSC;. (.‘. (1981) Theoretical prediction of the thermodynamic behavior of aqueous electrolytes at high pressures and tcmperatures. IV. Calculation of activity coefficients. nsmotic coefficients, and apparent molal and standard and relative partial molal properties to 5kb and 6iK) C. .-~/II. J. Sci. 280, in press. HUMPHRISS. E. and THOMPSONG. t lY78) Hvdrothermai alteration of oceanic basalts by seawater. Gr,oci!ij?r.(‘OX-

BACHINSKID. J. (1977) Sulfur isotopic composition 01 ophiolitic cupriferous iron sulfide deposits, Notre Dame Bay, Newfoundland. Econ. Geol. 72, 243.-257. CLARK L. A. (1971) Volcanogenic Ores: Comparison of mochim. Acta 42, 107-125. cupriferous pyrite deposits of Cyprus and Japanese KurIGUMNOV S. A. (1976) Sulfur isotope exchange between huoko ores. Sot. Min. Gee/. Jpn Spec. lssue 3, 206-215. CLAYP~OLG. E., HOLSERW. T., KAPLAN 1. R., SAKAI H. fide and sulfate in hydrothermal solutions. Geo~&n~. frzr. and ZAK I. (1980) The age curves of sulfur and oxygen 13, 113-124. isotopes in marine sulfate and their mutual interpretaIGUMNOVS. A., GRINENK~V. A. and P~NEK N. H (tY771 Temperature dependence of J4S for H,S and dtssolved tion. Chem. Geol. 28, 199-260. sulfate over the range 26@-400 C Grocllrn7 /!I( 14. C~ISH R. A. (1977) Ocean floor metamorphism in the Betts 81-82. Cove Ophiolite, Newfoundland. Contrib.Mineral. Petrol. JOHNSONA. E. (1972) Origin of Cyprus pyrite dep~~>~ts._‘+‘tii 60, 255-m270. COLEMANR. G. (1977) Ophiolites, Ancient Oceamc LithoInt. Geol. Conyr., Sect. 4, pp. 291 298. sphere? Rocks and Minerals, Vol. 12, 229 pp. Sp:inger. KAJIWARAY. (1973) Chemical composition of olc-iormmg solution responsible for the Kuroko type mincrali7ati~ln CRAIGH., WELHANJ. A., KIM K., POREDAR. and LUPTON in Japan. Geochem. J. 141.-149. J. E. (1980) Geochemical studies of the 21”N EPR hydroKEEFERK. D., HOCHELLAM. F. JR and IX JoNCi B. H. W. thermal fluids. Eos Trans. Am. Geophys. Union 61, 992. CRERARD. A., SUSAKN. J., BORCSIKM. and SCHWARTZS. S. (1980) The crystal structure of magnesium hydroxyaul(1978) Solubility of the buffer assemblage pyrite + pyrfate, MgSO, 1/3Mg(OH)* 1,‘3H,O. Proc. 3rd Int Swap. on W&r-Rock It&-act& pp. %51. rhotite + magnetite in NaCl solutions f;om 2od -to Krvosu Y. (1980) Chemical reduction and sulfur-lsotow 350°C. Geochim. Cosmochim. Acta 42. 1427 -1437. DICK~N F. W., BLOUNTC. and TUNNELLF. (1963) Use effects of sulfate by organic matter under hydrothermal conditions. Chem. Geol. 30, 47-56. of hydrothermal solution equipment to determine the LISTERC. R. B. (1972) On the thermal ballince clt’a Midsolubility of anhydrite in water from 100°C to 275°C ocean ridge. R. Astron. Sot. GeophJs. J. 26, 51s 535. and from 1 bar to 1000 bars pressure. Am. J. Sri. 261, LONSDAI.E P. F., BISCHOFF J. L.. BIIRNSV. M., K~.\I\IK M 61-78.


sulfate reduction

and S~IXNEY R. E. (1980) A high-temperature hydrothermal deposit on the seabed at a Gulf of California spreading center. Et& Plnnet. Sri. Left. 49, 8-20. MALIKIN S. D. and KHITAROV N. I. (1969) Reduction of sulfate sulfur by hydrothermal conditions. Geochem. Inr. 6, 102~1025. MOTTI M. J.. HOLLANU H. D. and CORR R. F. (1979) C’hemical exchange during hydrothermal alteration of basalt by seawater-II. Experimental results for Fe, Mn and sulfur species. G~~~~~j~7. C~.~~~~e~t~~z. Acta 43, 869- 884. MI!E.HI.ENRATHSK. and CLAYTON R. N. (1972) Oxygen isotope geochemistry of submarine greenstones. Can. J. Earth Sci. 9, 471-478. MI’uH,~ J. and KERRICH R. (1980) Sea water basalt interaction in spilites from the Iberian pyrite belt. Contrih. Minerul. Petroi. 73, 191 -200. N~~l~STR~)~ D. K. et cd.(1980) A comparison of computerited chemical models of equilibrium calculations in aqueous systems. In Chemical Modeling in Aqueous S)‘ctems. Speciation, Sorption, Soluhilitp, and Kinetics (ed. E. A. Jenne). Vol. 93. pp. 857-892. Am. Chem. SOC. Spmp. ser. OHMWIO H. and RYE R. 0. (1979) Isotopes of sulfur and carbon. In Ge(~~~~~~i,sf~~qf ~~drotbermu~ Ore Deposits. (ed. H. L. Barnes), 2nd edn, pp. 509-567. Wiley. RE~~I)M. (1977) Geologic and Geochemical environment of deposition of the massive copper-iron-sulfide ore deposits, West Shasta District, California. Ph.D. dissertation, Univ. of California (Berkeley). REFS C. E.. JENKINS W. J. and MONSTER J. (1978) The sulphur isotopic composition of ocean water sulphate. Geochurn C~).s~z~~~~~~?~. Acra 42, 377-382. RIPLE> E. M. and OH~MOTOH. (1977) Mineralogic, sulfur isotope, and fluid inclusion studies of the stratabound copper deposits at the Raul Mine, Peru. Econ. Geol. 72, 1017 1041. ROSENWA~!~RR. J., BISCHOFF J. L. and SEYFRIED W. E. JR (1979) Determination of sulfate in seawater and natural brines by ‘33Ba and membrane dialysis. Limnoi. Ucel~~z(~~r. 24, 393. 396. S~KAI H. (1968) Isotopic properties of sulfur compounds in hydrothermal processes. Geochem. J. 2, 29-49. SAKA~ H. and DICKSON F. W. (1978) Experimental determination of the rate and equilibrium fractionation factors of sulfur Isotope exchange between sulfate and sulfide in slightly acid solutions at 3oo-‘C and 1M30 bars. Errrth ~;/~l~f~f.sci. trtl, 39, ISl-161. S(.rrur:rnen A. (1970) The sulfur isotope composition of hasahic rocks. Contrih. Miner&. Petrol. 25, 95-124.

and sulfur isotope



SEYFRIED W. E. JR and BISTHOE~ J. L. (1977) Hydrothermal transport of heavv metals by seawater: the role of seawater basalt ratio. Earth P/uf2t,t. Sci. Letr. 34, 71-77. SEYFRIED W. E. JR and MOTTL M. J. (1978) Origin of submarine metal-rich hydrothermal solutions: Experimental basalt-seawater interaction in a seawater-dommated system at 3OO’C, 500 bars. In Proc. 2nd Int. Conf. on Water-Rock Interaction. Univ. LOUIS Pasteur. Sect. IV, pp. 173-l 80. SEYFRIEDW. E. JR and BISCHofF J. L. (1979) Low temperature basalt alteration by seawater: an experimental study at 70 C and 150 C. Geochinl. Co,smoc/~inr. .AUII 43, 1937-1947. SEYFRIED W. E. JR and BIS(.HOFF J. I.. (1981) Experimental seawater-basalt interaction at 300 C. 500 bars. chemical exchange, secondary mineral formation and implication for the transport of heavy metals. ~~~0~~~~)~. ~~)~~f~~~,~~j~}i. Acta 45, 135-147.

SEYFRIED W. E. JR. GORDOK P. C. and DICKSOX F. W. (1979a) A new reaction cell for hydrothermal solution equipment. Am. Minerul. 64, 64tc649. SEYFRIED W. E. JR. MULTI, M. J. and JA%L(.LY D. K. (1979hj Origin of metalhfernus deposits of the Troodos ophiolite: Experimental seawater-basalt interaction in a seawater domin~~ted system. Gao!. Snc.. Arts. -Ibarr. Programs. SHANKS W. C. III and BISCHOFF J. L. (lY77) Ore transport and deposition in the Red Sea geothermal system: A geochemical model. Geochim. Co.smoi~llim .4cro 41, 1507-1519. SHANKS W. C. III and BIS(.HOFF J. L. (1980) Geochemistry. sulfur isotope composition, and accumulati(~n rates of Red Sea geothermal deposits. Ecctn. (&I/. 74, 445 459. SPOONER E. T. C. and FYFE: W. S. (1973) Suh-sea-floor metamorphism, heat and mass transfer. Conrrrh. .%f;,lcro/. Petrol.

42, 287. 304.

SPOONER E. T. C., BE(.KINSALE R. D., F~H: W. S. and SMEWING J. D. (1974) 0” enriched ophi~~litic metabasic rocks from E. Liguria (Italy). Pindos (Greece). and Troodos (Cyprus). Contrih. illinenii Perrol. 47, 41-62. STYRIKOVICH M. A.. NARTYN~VA 0. I.. BF,LOVA Z. S. and MENSHILOVA V. L. (1968) Electrolytic properties of sodium sulfate in high temperature water. D&l. .+t/,trcl. Naufi SSSR 182, 644-646. WOI.ER~ T. J. and SLEEP N. H. (1976) Hydrothermal circulation and geochemicai flux at mid-ocean ridges. .I. (;cc~l. 84, 249 275.

Suggest Documents