FC/ZFC, where is given by
We carried out rock magnetic analyses that probed the composition, particle size, and concentration of the magnetic mineral assemblage. Many magnetic minerals undergo magnetic order-disorder transitions and crystallographic phase transitions at temperatures ranging from 10 to 1000 K that can be used as diagnostic indicators of a mineral's presence or absence (see Dunlop and Özdemir, 1997, for full discussion). We used low-temperature (20 to 300 K) measurements to determine the composition of the magnetic mineral assemblage. Low-temperature analyses consisted of a pair of measurements. First, a sample was cooled in zero applied field from 300 to 20 K, given a 2.5-T saturation remanence (MR) at 20 K, and MR was then measured during warming to 300 K. The sample was then cooled back down to 20 K, this time in the presence of a 2.5-T applied field, and MR was remeasured during warming to 300 K. This zero field cooled (ZFC) and field cooled (FC) pair of measurements, described in detail in Moskowitz et al. (1993) has proven useful in evaluating magnetic domain state and particle size and detecting the presence of intact chains of bacterial magnetite. All of the samples measured displayed a Verwey transition (Fig. F1), indicative of magnetite. We did not observe any salient features consistent with magnetic iron sulfides. However, greigite (Fe3S4) has no low-temperature transitions. Although magnetite is present throughout Site 1096, we observed several differences in the samples taken from above and below 18 mbsf. Samples collected from above 18 mbsf display the Verwey transition at ~103-108 K. MR decreases by 10%-15% across the Verwey transition, and the field cooled curve is higher than the zero field cooled curve (Fig. F2A). These samples show a continuing loss of remanence from 120 K up to room temperature, which could indicate particles with blocking temperatures at or near room temperature. Such particles sizes would be near the transition from superparamagnetic to stable single-domain (SSD) behavior at room temperature. The Verwey transition temperature observed in these samples is lower than values reported for stoichiometric magnetite. Previous studies have reported that oxidation of magnetite lowers, broadens, and suppresses the Verwey transition, as does nonstoichimometry resulting from the substitution of impurities such as Ti, Al, and Mn (Aragón et al., 1993; Özdemir et al., 1993; Kakol et al., 1994). Our observations may be indicative of some degree of cation substitution or nonstoichiometry. Samples from below 18 mbsf display the Verwey transition at 116-120 K (Fig. F1). MR decreases by 50%-60% across the Verwey transition. Further, the field cooled curve is lower than the zero field cooled curve (Fig. F2B). This is unexpected because cooling in the presence of a field should bias one of the (100) easy axes of magnetization upon cooling through the Verwey transition. However, this behavior has been observed in several natural and synthetic magnetite samples, and it appears to be a characteristic of multidomain magnetite larger than ~10-14 µm (Brachfeld et al., submitted [N1]).
Moskowitz et al. (1993) defined parameter FC/ZFC, where is given by
= (MR [80 K] - MR [150 K])/ MR [80 K].
Moskowitz et al. (1993) observed that intact chains of bacterial magnetite yielded FC/ZFC values >2. Mixed assemblages and assemblages containing nonbiogenic SSD particles had FC/ZFC values between 1 and 2. Synthetic pseudosingle domain (PSD) particles had FC/ZFC values near 1. Samples from above 18 mbsf have FC/ZFC values of 1.4-1.5, whereas those from below 18 mbsf are <1. We cannot conclusively identify bacterial magnetite by this test alone. However, our observations are consistent with a shift in the magnetic mineral assemblage at 18 mbsf. Hysteresis parameters from Site 1096 form a diffuse cluster that spans the boundary between the PSD and multidomain (MD) region of a Day plot (Day et al., 1977) (Fig. F3). Samples from Cores 1H and 2H from each of Holes 1096A and 1096B form a subcluster that plots slightly higher and to the left within the main cluster. However, samples with higher than average ratios of saturation remanence to saturation magnetization (MR/MS) occur approximately every 50-100 m down to the base of Hole 1096C.
The average value of the coercivity of remanence (HCR) shifts abruptly at 18 mbsf (Fig. F4; Table T1). Values of HCR for samples above 18 mbsf are higher than those reported for SSD stoichiometric magnetite (Heider et al., 1996, and references therein). HCR values for titanomagnetites may be elevated by high defect concentrations that pin domain walls or by cation substitution that increases crystalline anisotropy (Day et al., 1977; Readman and O'Reilly, 1972). Very high values of HCR (80-100 mT) have also been reported for SSD greigite (Fe3S4), which has high magnetocrystalline anisotropy (Roberts, 1995; Snowball, 1997).
The location of this shift in magnetic granulometry does not correspond to lithologic boundaries. There may be a correlation with geochemical boundaries. The location of the iron reduction boundary is not confidently identified at Site 1096. Iron concentrations were consistently zero in the interstitial water samples. The concentration of dissolved manganese increases from 0 to 15 mbsf, likely resulting from the dissolution of Mn oxides (Fig. F4). The bases of the Mn and sulfate reduction zones are at 60 mbsf. It is possible that we are observing the dissolution of fine-grained iron oxides, leaving a coarser residual assemblage below 18 mbsf. Alternatively, we may be seeing the presence of authigenic iron oxides and/or iron sulfides, possibly biogenic in origin, which form in the interval between 0 and 18 mbsf.
