13C,
18O, and 87Sr/86Sr.
At Site 897, calcite was precipitated within the serpentinites prior to their incorporation in the mass flow unit of late Hauterivian to late Aptian age (Shipboard Scientific Party, 1994a), whereas at Site 899, fracturing and calcitization postdate deposition of the serpentinite breccia and the underlying mass flow unit of Barremian to Aptian age (Shipboard Scientific Party, 1994b; Gibson et al., this volume).
Comparison with the Sr isotopic curve for Cretaceous seawater (Jones et al., 1994) reveals that 87Sr/86Sr values at both sites are consistent with the timing inferred from stratigraphy (Fig. 5). Early to middle Jurassic times for alteration fall within the observed range of 87Sr/86Sr in the calcite, but are deemed implausible because such ages predate the time of rifting. The earliest possible time for pre-massflow calcitization at Site 897, as determined from the observed 87Sr/86Sr, is somewhat older than the inferred time of rifting (approximately 130 Ma; Sawyer, Whitmarsh, Klaus, et al., 1994); calcitization essentially concomitant with deposition of the mass-flow unit is also possible. The youngest possible age of the post-mass-flow calcite precipitation at Site 899 (determined from the highest observed 87Sr/86Sr value of 0.70754) can be constrained only as older than Campanian. Younger ages for calcitization are not possible because the 87Sr/86Sr of seawater did not fall below 0.7076 in the post-Cretaceous. This is consistent with biostratigraphic data that constrain the sediment immediately overlying the altered zone at the top of the breccia to a Campanian-Maastrichtian age (Shipboard Scientific Party, 1994b). Further analysis of stratigraphic variation of 87Sr/86Sr in calcite and other authigenic phases (e.g., clays) at Sites 897 and 899 may provide more details about the timing and sequences of basement alteration.
Fe and Mn in the calcites are plausibly interpreted to have been mobilized into the pore fluid during dissolution of silicate components in the
serpentinites. Derivation of these elements from the overlying sediment column is precluded by the Sr isotopic evidence discussed above. Petrographic evidence for intermittent precipitation of Fe oxides in highly weathered serpentinite and serpentinite breccias
(Milliken et al., this volume) suggests a plausible local sink for dissolved oxygen, which could have led to fluctuations in the oxidation state of the fluid and its dissolved Fe and
Mn. However, the range and uniformity of isotopic values in calcite suggest that fluid/rock interactions did not cause a detectable shift in fluid composition away from normal marine values for
13C,
18O, and 87Sr/86Sr.
Figure 6 shows a range of temperature and
that are
consistent with the range of 18Ocalcite observed. Precipitation of these calcites from unmodified seawater (0‰, standard mean ocean water
[SMOW]) requires temperatures in the range of 10° to 17°C. The
18Ocalcite values are about 2‰-3‰ lighter than those observed in
calcites associated with mid-ocean-ridge ophicarbonates (Bonatti et al., 1974). Values observed by Bonatti et al. are consistent with temperatures of precipitation very near 0°C.
Unfortunately, uncertainties in the geologic history of mantle exposure at Sites 897 and 899, especially with regard to the earliest stages of rifting, allow wide latitude for interpreting the conditions of calcite precipitation. Based on the tectonic setting alone, fluids ranging from -5 ‰ to +7‰ can be postulated; temperatures ranging from near 0° to >100°C are possible.
Precipitation of calcite from strictly meteoric fluids can probably be ruled out on the basis of the apparent marine 87Sr/86Sr signature together with whole-rock rare-earth element
(REE) patterns suggesting a strong marine diagenetic overprint (Milliken et al., this volume). However, fluids other than a 0‰ seawater still can be postulated under several scenarios. Glacial cycles shift
18Oseawater by about ±1‰ (e.g., Arthur et al., 1983). Modification of fluids in submarine diagenetic environments can shift
into the range of negative values by several possible mechanisms. Precipitation of copious quantities of authigenic phases at very low temperatures (e.g., Lawrence and
Gieskes, 1981), recharge of meteoric fluids into submarine aquifers (e.g., Manheim and
Paull, 1981), and expulsion of unmodified, relict meteoric water from rocks buried in a marine environment (e.g., Gieskes et al., 1994) are examples of mechanisms for producing 18O-depleted pore fluids in marine settings. Interpreted excursions of
to values below about -3‰ at 0° to 5°C can be excluded for our data set because such values lead to the prediction of improbably low temperatures. Any 18O-depleted (negative) fluid at temperatures >17°C is inconsistent with observed
18Ocalcite, which again probably rules out any direct (i.e.,
surficial) meteoric exposure. A slightly 18O-depleted fluid in the temperature range of 5° to 17°C cannot be excluded as a possible agent of calcite precipitation. Expulsion of depleted fluids from underlying (or adjacent) basinal sediments seems unlikely for these basement rocks, especially given the volume of fluids necessary to account for the large amount of calcite that is
present. Thus, 18O-depleted fluids, if once present, would be most plausibly interpreted as the result of authigenic mineral precipitation, as suggested by Lawrence and Gieskes (1981), especially in view of the Sr isotopic values that are consistent with Cretaceous seawater.
Values of
in the positive range can be produced by evaporative concentration of 18O in surface fluids (Lloyd, 1966), an alternative reasonably excluded in this tectonic setting. "Rock-dominated" fluid/rock interaction, in which the fluids acquire an isotopic signature that is shifted toward that of the rock by mineral dissolution or replacement, is a more likely mechanism for producing 18O-enriched fluids for calcite precipitation in this setting. Possible excursion of
into the positive range is a reasonable consideration in view of the Fe and Mn concentrations in the calcites reported here
and the extreme degree of fluid/rock interaction evidenced by secondary porosity, Mg loss from the whole rock, and clay mineral precipitation
(Milliken et al., this volume). However, at least three lines of evidence suggest that significantly positive values of
can also be ruled out. First, although temperatures consistent with 18O-enriched fluids (up to 53°C for a
in the range of +7) are
possible in this rift setting, the uniformity of 18Ocalcite argues against this sort of "hydrothermal" system because both temperature and
would have to have been maintained within narrow limits (or have covaried precisely) for the duration of precipitation of the multiple fracture-filling calcite generations. Second, two-phase fluid inclusions have not been observed among the abundant inclusions in the vein-filling calcite, which suggests a very low temperature of precipitation that is incompatible with 18O-enriched fluids over the observed range of
18Ocalcite. Finally, the absence of Mg-bearing carbonate phases (dolomite or
magnesite) and 87Sr/86Sr values that range narrowly about the values of Early to middle Cretaceous seawater (e.g., Burke et al., 1982; Koepnick et al., 1985; Ingram et al., 1994; Jones et al., 1994) suggest that dissolution of the serpentinite did not strongly overprint the composition of the fluid that mineralized the fractures.
The lack of systematic compositional trends (for both trace elements and isotopic values) across petrographic variations in calcite morphology and sequence of precipitation is an unexpected observation. Apparently, variation in calcite composition was not influenced by the same factors that affected calcite morphology. Rather, it seems likely that morphological variation in the crack-filling calcite was controlled by some factor—precipitation rate, for example—that is not strongly manifested in calcite composition. In addition, the lack of correlation between the composition and temporal sequences of calcite precipitation observed suggests that neither temperature nor fluid composition evolved in any systematic or significant way during vein filling. 87Sr/86Sr values that span nearly the total observed range for both vacuolized and clear calcites suggest that the sequence from "early" vacuolized calcite to "late" clear spar developed repeatedly during vein filling.
