RESULTS AND DISCUSSION

All of the 12 sediment samples under examination experienced distinct redox changes within the 6 months between core retrieval and the sampling of the ODP archive core. Whereas the Fe(II)/Fe(III) ratio was near 50/50 in the initial material, between 12% and 24% of the total Fe in the sediment--that is, between 24% and 45% of the initial Fe(II)--passed through a Fe(II)-Fe(III) redox transition in that time period (Fig. F2). This result, obtained by Mössbauer spectroscopy at 77 K, already documents the extent to which ODP sediment cores may undergo chemical alteration during storage. Moreover, a hematite (-Fe₂O₃) fraction could be clearly determined at 77 K (Fig. F1). This mineral is present in seven of the 12 samples; its contribution to the total Fe content varies between 6% and 11% and remains unchanged (within analytical precision) after 6 months of archive storage (Table T2). In the hematite-rich sediment layers, known as "red lutites," the hematite is of allochthonous nature, being derived from the Permian-Carboniferous deposits of the Canadian Maritime provinces (Shipboard Scientific Party, 1998c).

Both siderite and vivianite are below their Mössbauer detection limits in all of the 12 samples, although the method is very sensitive with respect to those two minerals (see König and Hollatz, 1990). This implies that <1.5% of the bulk sediment iron can possibly be contained in siderite or vivianite. The ability to detect even minor siderite contents in sediments is of particular interest in the context of Leg 172 with respect to the gas hydrate that has been found in the Blake Ridge area (e.g., Paull, Matsumoto, Wallace, et al., 1996). Whereas one of the scientific objectives of Leg 172 was to assess the lateral distribution of gas hydrate and its related geochemical signatures within the continental rise (Shipboard Scientific Party, 1998b), close relations between these gas hydrate occurrences and carbonate diagenesis are a well-established fact (e.g., Lancelot and Ewing, 1972). Since siderite found in marine sediments is an indicator of carbonate authigenesis, this mineral might serve as a sensitive tracer.

Mössbauer spectroscopic examination of one of the samples (94 mbsf) at a large range of temperatures (300.0-0.3 K) revealed that the major part of the total Fe is structural Fe(II) or Fe(III) in the aluminosilica fraction (A. Lougear et al., unpubl. data). Whereas goethite (-FeOOH) was not detected, and its maximum possible contribution to the total iron, therefore, is <1.5%, small amounts (~3% ± 1% of the total iron) of pyrite (FeS₂) could be identified (Fig. F1B).

The Fe(II) subspectrum depicted in Figure F1A can clearly be assigned to structural Fe(II) in silicate minerals. Separation of the Fe(II) subspectrum at 0.3 K (Fig. F1B) must result from different lattice Fe(II) neighborhoods. Hence, virtually the entire Fe(II) to Fe(III) oxidation process occurred within silicate mineral lattices. This implies significant modifications in both the sediment color (Lyle, 1983) and the geochemical redox reactivity (König et al., 1999), whereas the mineral contents remained unchanged.

Markedly contrasting examples are known especially from freshwater systems, such as lake and river sediments that contain considerable amounts of siderite and vivianite (e.g., König and Hollatz, 1990). These two Fe(II) minerals are essentially transformed to Fe(III) oxides and oxyhydroxides upon contact with atmospheric oxygen (König et al., 1988). Even Blake Ridge-area sediments--other than the small series reported here--might contain important quantities of siderite originating from early diagenesis related to gas hydrates. On archive storage of the sediments, that portion may perhaps soon be transformed, as can be deduced from the pronounced oxidation processes observed in the present study.

Both findings, (1) small pyrite and negligible goethite contents in the presence of detrital hematite and (2) considerable amounts of the structural Fe(II) in the aluminosilica fraction that are oxidized to Fe(III) upon contact with atmospheric oxygen, imply that pronounced postdepositional diagenesis has occurred in the seafloor. Although the stage of iron reduction was obviously reached, there is only a minor indication of sulfidic conditions. Again, this may be different for Blake Ridge-area sediments outside of the series investigated and reported here.

Inasmuch as the deposition of the sediments happened in oxygenated bottom waters, the potential of the structural silicate Fe(II) to be oxidized to Fe(III) so easily (during storage) indicates that the inverse reaction (i.e., chemical Fe(III)-Fe(II) reduction) must have occurred to at least the same extent in the seafloor. Fe(III) oxides and oxyhydroxides would have been affected by the Fe(III)-Fe(II) reduction even more than the silicate lattice Fe(III) because they are more directly exposed to the redox environment than the comparatively shielded structural iron. Chemical reduction of Fe(III) oxides and oxyhydroxides, however, implies iron mobilization and upward transport, as well as reoxidation and precipitation above the iron redox cline.

Therefore, reversible shifts of the Fe(II)/Fe(III) ratio within the silicate mineral content of sediments may be interpreted in terms of possible diagenetic dissolution and redistribution of iron oxide minerals in the seafloor. At Site 1062, this is important with respect to the lithostratigraphic value of the hematite-rich red lutites and their significance as advective proxies for deep circulation (Shipboard Scientific Party, 1998c).

Likewise, the oxidizable silicate Fe(II) content in a given sediment horizon may be interpreted as an indicator of possible chemical overprint on the primary magnetic signal in the seafloor, which principally resides in iron oxide minerals, especially magnetite. High magnetic susceptibility layers should be associated with little or no oxidizable silicate Fe(II) since they experienced iron oxide precipitation instead of dissolution. As a matter of fact, both a spike in the concentration of solid-phase iron (i.e., goethite) and enhanced sediment magnetic intensities (attributable to magnetic mineral authigenesis associated with early diagenesis) have been found in the near-surface sediments of the Blake/Bahama Outer Ridge sediments (Schwartz et al., 1997).

Also Torii (1997) observed manifest downcore changes in the magnetic properties (i.e., Verwey transition) across the iron redox boundary in deep-sea sediments (ODP Leg 161) from the western Mediterranean Sea. From his findings, he hypothesized in situ formation of a maghemite (-Fe₂O₃) coating on primary magnetite grains above the redox boundary and dissolution of that coating, followed by progressive dissolution of the magnetite core itself, below the boundary.

Thus, early diagenetic processes that affect the iron mineral assemblage in the sediment sequence are generally of interest with regard to the reliability of the magnetic record. This especially applies to the Leg 172 cores, which contain an extraordinarily detailed and complete record of the Earth's magnetic field variability for the past 1.2 m.y. (Shipboard Scientific Party, 1998c). The postdepositional shift of the Fe(II)/Fe(III) ratio within the silicate mineral content, as well as the other data on the sediment iron that are obtainable by Mössbauer spectroscopy, can in doubtful cases, together with rock magnetic methods, help to judge the authenticity of magnetic field excursion records. However, that information is getting lost during storage of the sediment cores.