400
ka, occurs just over 5 half-lives of 230Th
ago and, therefore, represents the approximate limit of the U/Th approach in
closed-system samples.Eighteen
samples were selected from highstand aragonite-rich sediment packages from Sites
1003, 1005, 1006, and 1007 (Table 1; Fig.
1). In general, samples were taken from successive sediment packages
down to a depth assessed to be marine isotope Stage 11 in preliminary age models
(Eberli, Swart, Malone, et al., 1997). Stage 11, at 400
ka, occurs just over 5 half-lives of 230Th
ago and, therefore, represents the approximate limit of the U/Th approach in
closed-system samples.
Existing U/Th work has been most successful with coral samples (Bard et al., 1990; Broecker et al., 1968; Edwards et al., 1987). Corals are an ideal material for U/Th dating as they contain almost no detrital material and incorporate essentially no initial Th on growth, thus yielding initial (230Th/238U) ratios of zero (where parentheses signify an activity ratio). The increase of this ratio toward secular equilibrium therefore provides a measure of the coral's age, provided it has remained a closed system. The aragonite-rich sediments of the Bahamas have U concentrations similar to those of corals but, while their initial (230Th/238U) ratios are low, they are not zero (Slowey et al., 1996; G.M. Henderson, N.C. Slowey, and M.Q. Fleisher, unpubl. data). This is because of the presence of small amounts of detrital material in the sediment and the presence of Th scavenged from the water column. These sources of Th add both 232Th and 230Th to the sediment and can, therefore, be recognized from the pure 230Th produced from U decay. However, detrital and scavenged Th have different 232Th/230Th ratios from one another, so if the 232Th is to be used to correct for a nonzero initial (230Th/238U), either the relative proportion of detrital versus scavenged Th must be assessed or one of these sources of Th must be eliminated prior to analysis.
In this study, we eliminate the detrital fraction from the sediments prior to analysis. Sediment is first sieved to acquire a 63- to 250-µm fraction. Removal of the <63-µm fraction eliminates clay minerals, which are the majority of the detrital fraction. Removal of the >250-µm fraction reduces the number of foraminifers and pteropods that may be filled with clay minerals and that have low U concentrations (Henderson and O'Nions, 1995). Bulk sediment samples of 6-10 g yielded 0.7-3.0 g of 63- to 250-µm material after sieving.
Each 63- to 250-µm sample then undergoes density separation to give a >2.8 g/cm3 fraction. This stage concentrates the aragonite, which is the only major mineral constituent of these sediments with a density >2.8 g/cm3. This stage also removes any silicate minerals surviving the sieving stage, including most clays, quartz, and opal. The 63- to 250-µm fractions were placed in 10 mL of sodium polytungstate solution (SPT) with a density of 2.80 (± 0.01) g/cm3 and centrifuged for 10 min. The dense fraction in the bottom of the centrifuge tubes was frozen with liquid nitrogen and the light fraction poured off. Further density separations were conducted on this light material to yield a high-Mg calcite (HMC) and low-Mg calcite (LMC) fraction, but these samples have not been analyzed for this study. Thorough washing of all SPT-treated samples with multiple water rinses is required to fully remove the heavy liquid. Effective removal can be checked by X-ray diffraction (XRD) analysis because dried SPT causes many spurious peaks throughout the XRD spectrum. Heavy-liquid separation yielded between 0.09 and 1.9 g of >2.8 g/cm3, 63- to 250-µm material.
In addition to cleaning the samples of any detrital Th, the sieving and heavy-liquid procedures decrease the importance of scavenged Th because the coarse grain size decreases the surface area to volume ratio. The removal of low-U grains and the concentration of high-U aragonite also increases the U content of the analyzed sample to a maximum. A final reason for these pretreatment steps is that the coarser grain size should make the effect of alpha recoil less significant.
The composition of representative final >2.8 g/cm3 63- to 250-µm separates was assessed by XRD (Fig. 2; Table 2). Aragonite percentages of significantly less than 100 in bulk sediment were increased to within error of 100% in each case. These final 63- to 250-µm aragonite fractions are expected to consist predominantly of small pieces of bank-derived algae such as Halimeda and also appear to include a high U organic material, probably as a coating (Henderson et al., 1999).
