RESULTS

Iron concentrations and isotope values are reported in Table T2. Results are plotted against stratigraphic depth in Figure F1 and compared to selected shipboard organic geochemical data.

The amount of elemental Fe obtained by HNO₃-HF (Fe_HF) digestion ranges between 0.6 and 1.3 wt%. The average sample mass digested during this procedure was 82% (range = 73%–87%), indicating that much of the lithogenic material was consumed and that Fe_HF essentially represents the total amount of Fe within the sediment. The residue from these extractions should consist mostly of organic matter. The total organic carbon content in the Cenomanian–Turonian organic-rich levels that were sampled in Hole 1260B varies ~10 wt% (Erbacher, Mosher, Malone, et al., 2004), which corresponds to a proportionally higher percentage of organic matter when other elements (H, O, and S) are taken into account. The AR digestion extracted between 0.2 and 0.5 wt% Fe (mean = 0.34%) or between 25% and 46% of the total Fe in the sample. Iron-sulfides, Fe oxides, and any Fe carbonates are dissolved during this procedure. Shipboard analyses of smear slides showed that pyrite, which is the main component expected in Fe_AR extracts, occurs in variable concentrations downhole to the base Core 207-1260A-44R but is rare or absent lower in the black shale unit. The low abundance of microscopically identifiable pyrite agrees with the relatively low Fe_AR values that were obtained. At this stage in our studies, we cannot evaluate the relative importance of Fe carbonate; however, the minor proportion of loosely bound Fe and that contained in the oxide fraction would have been consumed.

With the exception of two samples, the amount of Fe extracted by the CBD method yielded low to very low Fe concentrations (between 0.009% and 0.16%, mean = 0.07%, n = 8), such that Fe_CBD concentrations range from 2% to 49% of Fe_AR. The two anomalously high Fe_CBD yields that were obtained are the result of failure in the anion exchange procedure, manifested by abnormal discoloration, and behavior of the AG MP-1 resin bed during Fe separation. A negligible difference in sample mass before and after CBD extraction is found; in some samples, a small increase in mass was measured, presumably because of precipitation of Na that is present in the CBD mixture. From the negligible mass difference, it is assumed that the CBD extraction did not dissolve any Fe carbonate. Calcite is a major component of the sediment, and any dissolution of carbonates would have resulted in more significant losses from sample mass. The low Fe_CBD concentrations indicate that Fe oxides are rare, as would be expected under the anoxic conditions that predominated during the mid-Cretaceous at Demerara Rise.

The ⁵⁶Fe values for Fe_AR extracts range between 0.02 and –0.77 and, except for the three lightest extracts (⁵⁶Fe = –0.47 ± 0.06, –0.43 ± 0.11, and –0.77 ± 0.05), correspond to Fe concentration (r = 0.96). This correlation may reflect increasing amounts of Fe leached from the silicate fraction of the sediment, which results in the isotopic composition corresponding more closely to the global average isotopic composition of Fe (⁵⁶Fe = 0). There is no correlation between ⁵⁶Fe and weight percent dissolved sample or between ⁵⁶Fe and weight percent Fe_CBD. The three lightest extracts may represent significant reduced Fe in another mineral component not represented in the other seven extracts or in a different proportion of Fe sulfide and Fe carbonate relative to each other. Unfortunately, the isotopic composition of only two Fe_CBD extracts was measured. Both Fe_CBD measurements are significantly heavier compared with the isotopic composition of the corresponding Fe_AR extracts, yielding positive ⁵⁶Fe values of 0.73 (Sample 207-1260A-46R-2, 31–32 cm) and 0.63 ± 0.04 (Sample 46R-5, 53–54 cm). These values are ~0.7 heavier than the global average Fe isotope composition of ⁵⁶Fe = 0 (Beard et al., 2003). Therefore, it is unlikely that the three lightest Fe_AR extracts are the result of a Fe oxide component but instead may represent a diagenetic component precipitated from reduced Fe(II)-bearing pore waters or biologically mediated Fe that is not present in the other AR extracts.

Comparison of the Fe_AR with Fe_CBD extractions for Samples 207-1260A-46R-2, 31–32 cm, and 46R-5, 53–54 cm, reveals a difference in ⁵⁶Fe of 0.72 and 1.10, respectively, between ⁵⁶Fe_CBD and ⁵⁶Fe_AR. These values qualitatively represent the difference between the Fe oxide and AR-extractable mineral components in the sediment. A mass balance correction for the presence of Fe oxides in the Fe_AR extract can be performed assuming that all CBD-extractable Fe is also extracted by AR:

a x ⁵⁶Fe_CBD + (1 – a) x ⁵⁶Fe_AR–CBD = ⁵⁶Fe_AR,

in which a is the percentage of Fe_CBD relative to Fe_AR. Resulting values for ⁵⁶Fe_AR–CBD are –0.002 in Sample 207-1260A-46R-2, 31–32 cm, and –0.49 in Sample 46R-5, 53–54 cm. This results in a corrected Fe_AR–CBD that is only ~0.02 lighter than the measured Fe_AR because of the low concentration of Fe_CDB in these two samples. This indicates an isotopic difference between the Fe oxide (Fe[III]) and Fe sulfide + Fe carbonate (Fe[II]) components in the sediment of between 0.75 and 1.12.

The ⁵⁶Fe variation between different mineralogical components may reflect redox cycling in the sediments of Demerara Rise, which may include reduction by interaction with dissolved H₂S and bacterial processing. The reduction of Fe by bacteria produces isotopically light Fe(II)_aq, and the equilibrium fractionation between this and Fe(III)_aq in low-pH environments has been shown by experiment to be negative (Johnson et al., 2002, 2004; Welch et al., 2003). Severmann et al. (2002) measured Fe(II)_aq as low as ⁵⁶Fe = –5.0 in pore fluids from C_org-rich sediments. Experimental studies have indicated that Fe isotope fractionation between FeCO₃–Fe(II)_aq and FeS–Fe(II)_aq partitions the lighter isotope into the solid phase (Wiesli et al., 2003; Butler et al., 2003), suggesting that diagenetically precipitated minerals will be isotopically light in reducing environments. It is likely that the Fe isotopic compositions measured in the Demerara Rise sediments reflect partitioning between different mineral components and Fe(II)_aq under changing redox conditions and that there is indeed a change in redox conditions within the section is indicated by shipboard C/N ratios and oxygen index (OI) values. Two of the three samples with light ⁵⁶Fe_AR values are located below 442 mcd, at which level there is a change in organic matter preservation. On average, high C/N ratios and OI values are found below 442 mcd, indicating that organic matter is more degraded and contains a higher portion of inert material than in the upper part of the section (Erbacher, Mosher, Malone, et al., 2004); hydrogen index values are high also, which indicates that the organic matter is of marine origin, not terrigenous. The shift in organic matter preservation points to a change in bacterial action, which may have affected the partitioning of Fe within the sediment.