Curie temperature data in combination with the shapes of thermomagnetic curves are characteristic of particular magnetic minerals (e.g., Dunlop and Hale, 1977; Moskowitz, 1981). Thermal demagnetization curves of low-T IRM can indicate the presence of minor amounts of stoichiometric magnetite (as the Verwey transition) and other magnetic phases (e.g., Özdemir et al., 1993; Moskowitz et al., 1998). Therefore, our characterization of magnetic phases in the samples is mainly based on features of thermomagnetic and low-T IRM demagnetization curves. The heating curves of pillow basalt samples generally show an apparent Curie temperature ranging 250°-400°C. When heated above ~300°C, the pillow basalts show one or two maxima, which probably indicate formation of new magnetic phases and remagnetization of samples (Fig. F2A). Further heating gives leads to a Curie temperature of ~580°C. These features of thermomagnetic curves are characteristic of titanomaghemite, which becomes unstable and transforms to magnetite at elevated temperature (e.g., O'Reilly, 1984). The cooling curves are completely irreversible relative to the heating curves for the pillow basalt samples. In contrast, the massive basalt samples show almost reversible thermomagnetic curves and a Curie temperature of 170°-230°C (Fig. F2B), indicating titanomagnetite as a dominant magnetic carrier. The metadiabase sample shows approximately irreversible thermomagnetic curves with a slight maximum at 250°-300°C, which might be due to the presence of small amounts of maghemite or titanomaghemite (Fig. F2C). However, the overall heating curve gives a Curie temperature of ~580°C, indicating a predominance of magnetite in this sample.
Low-temperature magnetic properties are distinguishable among the different rock types. The pillow basalts show smooth and monotonous low-T IRM demagnetization curves, decreasing from 5 K to room temperature without any significant phase transition (Fig. F2A). In contrast, the massive basalts show phase transition at ~55 K (Fig. F2B). We interpret the transition at ~55 K as a possible indication of titanomagnetite (Moskowitz et al., 1998). The metadiabase sample shows phase transition at ~120 K, which is characteristic of the Verwey transition of magnetite (Fig. F2C) (Özdemir et al., 1993).
The NRM intensities of Leg 187 basalts show a wide range (0.12-16.3 A/m). The average NRM of the basalt samples for each site (or hole) ranges from 1.16 to 8.71 A/m and is equivalent to the NRM of typical ocean-ridge pillow basalts (e.g., Johnson and Hall, 1978; Beske-Diehl, 1990; Gee and Kent, 1994, 1997). This NRM range is consistent with that of oxidized pillow basalts. Although the NRM variations within one site or one hole are very large, when plotted vs. crust ages the NRM intensities show a general decreasing trend with increasing age (Fig. F3). A few sites, such as Sites 1161, 1162, and Hole 1164A have very low average NRM intensities. However, the specimens from these three sites show the highest degree of alteration among the sites (Shipboard Scientific Party, 2001) and their relatively low NRM values could be a result of pervasive alteration that has significantly modified and/or reduced concentrations of magnetic minerals. When the NRM is plotted vs. sample depth in igneous basement, it does not vary systematically with depth.
Figure F4 is a logarithmic plot of NRM vs. Curie temperature using data obtained from 43 samples. The plot shows a systematic trend of decreasing NRM with increasing Curie temperature. As Curie temperatures generally increase with increasing degrees of low-temperature oxidation (maghemitization) for oceanic basalts (e.g., O'Reilly, 1983; Zhou et al., 2001), the trend shown here implies that the degrees of maghemitization systematically increase with age in these basalts during this ~15-m.y. time span. However, a plot of Curie temperature vs. age (not shown here) does not show a clear trend of increasing Curie temperature with age. This apparent inconsistency further implies that the degree of maghemitization varies not only with age but may vary locally with the permeability of the basalts.
Table T1 contains NRM and magnetic susceptibility data for all of the analyzed specimens. Figure F5 shows two plots of NRM vs. susceptibility, with data grouped according to mantle province and tectonic zonation (Zone A or B). The plots show that both NRM and susceptibility from the different groups vary within a considerable range and neither the Pacific-type samples nor the Zone A samples have, on average, higher NRM than the Indian-type or the Zone B samples. Although the amplitudes of the magnetic anomaly east of the AAD eastern boundary (Zone A) are generally much higher than those west of the AAD eastern boundary (Zone B) at the ridge area, this contrast is seen in regions only up to 46°S (e.g., Marks et al., 1990). The Leg 187 drill sites are within an area from 41° to 46°S, where the magnetic anomaly amplitudes have no significant variations between Zone A and Zone B. Our observations show no significant differences of the NRM intensities between Zone A and Zone B samples, and NRM values are consistent with the distributions of the magnetic anomaly amplitudes in this area.
Our thermal demagnetization data reveal that a stable direction persists in a majority of the pillow basalts up to ~540°C (Fig. F6A), clearly indicating that the unblocking temperatures of NRM are much higher than the Curie temperatures (250°-400°C) of titanomaghemite. We also note that a significant portion of the NRM intensity (~30%-40%) remained above the Curie temperatures. Irving (1970) suggested that this phenomenon of anomalously high unblocking temperatures relative to the Curie temperatures is caused by an artificial magnetic phase, formed as a reaction product of titanomaghemite during thermal demagnetization. However, by studying thermal demagnetization properties of very young oceanic basalts from the East Pacific Rise, Kent and Gee (1994) demonstrated that the NRM with much higher unblocking temperatures than Curie points is original and is carried by naturally occurring magnetite. Although our observations of thermal demagnetization properties for the Leg 187 pillow basalts are very similar to those of Kent and Gee (1994), our low-T IRM results do not support any significant Verwey transition of magnetite in the pillow basalts (Fig. F2A). Further high-resolution petrographic studies should be carried out to confirm this proposition.
During AF demagnetization, most of the pillow basalts reveal a very stable high-coercivity remanence (Fig. F6B). On the contrary, the massive basalts have a very low coercivity remanence, showing a rapid decrease in intensity and highly scattered directions at >10-mT AF demagnetization (Fig. F6C). These two types of AF demagnetization curves are typical of oceanic oxidized pillow basalts and unoxidized massive flows, respectively (Dunlop and Hale, 1976). Our results thus indicate that the majority of the Leg 187 pillow basalts are characterized by a single stable component of remanence, whereas the coarse-grained massive basalts readily acquire components of viscous remanence, consistent with the results of previous rock magnetic studies on oceanic basalts (e.g., Dunlop and Hale, 1976; Johnson and Hall, 1978). The AF demagnetization results also suggest that the grain sizes of magnetic carriers in the pillow basalts should be much smaller than those in the massive basalts (see below).
The granulometry of magnetic minerals can generally be inferred from the hysteresis loops and Day plot (Mrs/Ms vs. Bcr/Bc; Day et al., 1977). The hysteresis loops obtained from the specimens of three major rock types are shown in Figure F7. Hysteresis parameters of all specimens are listed in Table T1, and their Day plots are shown in Figure F8. The hysteresis loops of the pillow basalt samples generally show wasp-waisted or constricted shape (Fig. F7A), suggesting that the magnetic mineral is possibly composed of a mixture of superparamagnetic (SP) and stable single-domain (SSD) grains (Dunlop and Özdemir, 1997; Gee and Kent, 1999). Most of the Leg 187 specimens have Mrs/Ms = 0.40-0.70 and Bcr/Bc = 1.2-2.0. It should be noted that more than one-half of the Mrs/Ms ratios plot above 0.5, which is a theoretical limit for uniaxial anisotropy (Dunlop and Özdemir, 1997). Our data indicate that the magnetic minerals in the Leg 187 basalts are predominantly of cubic anisotropy, which is the same as that demonstrated for a broad collection of oceanic basalts by Gee and Kent (1999). On the basis of the Day plot and hysteresis loops, the titanomaghemite contained in the dominant pillow basalt samples ranges from SSD to pseudosingle-domain (PSD) grains (Figs. F7, F8). Magnetic grains are larger in the massive basalts from Hole 1160B than those in the pillow basalts because their data points on the Day plot are closer to the multidomain category (larger PSD). The massive basalt samples have low coercivity (Bc) but relatively high susceptibility and saturation magnetization (Ms) (Table T1). The metadiabase sample from Hole 1162A also plots in the larger PSD category (Fig. F8).
The petrographic observations of polished thin sections under SEM are consistent with the results inferred from magnetic properties. Most pillow basalt samples contain very fine grained skeletal or subhedral titanomagnetite (a few micrometers in diameter to ~10 µm in long dimension), which occurs as an interstitial phase between plagioclase and clinopyroxene grains (Fig. F9A). The massive basalt samples from Hole 1160B contain subhedral to euhedral titanomagnetite up to tens of micrometers in size (Fig. F9B). These titanomagnetite grains do not show any distinguishable internal textural features that might be caused by alteration or oxidation. Therefore, the dominant magnetic carrier in the massive basalts is primary titanomagnetite with relatively larger grain sizes, consistent with their magnetic properties, such as low coercivity and low Curie temperature. The metadiabase sample from Hole 1162A contains subhedral titanomagnetite pseudomorphs (up to a few hundred micrometers across) with well-oriented ilmenite lamellae (Fig. F9C). This lamellar texture is characteristic of "oxidation-exsolution" or high-temperature oxidation, to which the titanomagnetite was subjected and which, together with pervasive hydrothermal alteration, gave rise to end-member magnetite as the magnetic carrier in this rock type (e.g., Shau et al., 1993, 2000). The effective grain size of the magnetite is significantly reduced due to the crosscutting of primary titanomagnetite by ilmenite lamellae.
The titanomagnetite grains in the pillow basalt samples are indeed pseudomorphs, which have presumably been subjected to different degrees of low-temperature oxidation and changed into titanomaghemite. That the titanomagnetite underwent low-temperature oxidation is not indicated only by the relatively high Curie temperatures and irreversible thermomagnetic curves for these pillow basalt samples, but also by the shrinkage curvature cracks that are commonly seen in titanomagnetite pseudomorphs from the oldest samples from Leg 187 (e.g., samples from Sites 1154 and 1152) (Fig. F10A). The curvature cracks are a diagnostic feature for the presence of titanomaghemite, which formed after titanomagnetite via maghemitization and with a volume change (e.g., Johnson and Hall, 1978). The formation of shrinkage cracks might divide the titanomagnetite grains into submicroscopic grains. The shrinkage cracks are most abundant in titanomagnetite pseudomorphs from the oldest samples of Leg 187 but are not clearly visible in those of the younger samples. This observation is consistent with the above argument that the degree of maghemitization systematically increases with age. However, the younger basalt samples still contain titanomaghemite, as shown by their Curie temperatures and irreversible thermomagnetic curves. Under higher magnification, we found that many submicroscopic Fe-Ti oxide mineral grains are present within the interstitial glass of many pillow basalts (Fig. F10B). The sizes of these submicroscopic magnetic mineral grains are much smaller than 0.5 µm and are consistent with the relatively large Mrs/Ms ratios and SSD grains inferred from the Day plot.
Table T2 summarizes the Leg 187 core samples with respect to rock types, magnetic minerals, magnetic properties, and alteration features. Most core samples are pillow basalts that contain titanomaghemite as the main magnetic carrier and are characterized by a single stable component of remanence. Three massive basalt samples contain nearly unoxidized titanomagnetite, and a metadiabase sample contains secondary magnetite as its dominant magnetic carrier.
The NRM intensities of MORB may vary in relation to several factors such as geomagnetic field intensity (e.g., Ravilly et al., 2001), concentration of magnetic minerals (mainly titanomagnetite), which is in turn related to rock chemistry (e.g., Marshall and Cox, 1972; Gee and Kent, 1997), and grain size of magnetic minerals (e.g., Marshall and Cox, 1971; Ryall and Ade-Hall, 1975; Gee and Kent, 1997). Significant variations of NRM vs. distance from pillow margins have been demonstrated in relatively young and zero-age MORB (e.g., Marshall and Cox, 1971; Ryall and Ade-Hall, 1975; Gee and Kent, 1997). Ryall and Ade-Hall (1975) showed significant radial variations of NRM intensity in 0- to 286-ka pillows and very uniform NRM intensities vs. distance from margin of a 740-ka pillow basalt except in the first few millimeters of the pillow rinds. As the MORB drilled during Leg 187 is 14-28 Ma, we expect that the radial variations of NRM intensities will be much less significant. However, in order to decrease the possible radial variations and compare the NRM intensities among samples from different sites and holes from Leg 187, we collected samples for our magnetic measurements in a similar depth range at ~0.5 cm from pillow margins for the rim samples and ~1-3 cm from pillow margins for the core samples.
The Leg 187 basalt samples show a general decreasing trend of NRM with increasing crust ages from 14 to 28 Ma. This trend might indicate an increasing degree of low-temperature oxidation (e.g., Johnson and Pariso, 1993; Zhou et al., 2001) or reflect decreasing concentration of magnetic minerals due to increasing degree of alteration (e.g., Xu et al., 1997). Our data show that NRM intensities correlate reciprocally with the Curie temperatures (Fig. F4). In addition, when considering a larger time span, the Curie temperatures (250°-400°C) of the Leg 187 pillow basalts are significantly higher than those from the very young MORB. For example, Gee and Kent (1997) showed Curie temperatures ranging 100°-240°C for most of the zero-age basalts collected from the southern East Pacific Rise. This contrast in the Curie temperatures indicates that the older Leg 187 basalts were subjected to much higher degrees of low-temperature oxidation. We therefore infer that the general trend of decreasing NRM intensities with age in the Leg 187 basalts reflects an increasing degree of low-temperature oxidation. However, as mentioned previously, a plot of Curie temperature vs. sample age does not unambiguously show a positive relationship between the degree of low-temperature oxidation and crust ages. This led us to believe that, in addition to the factor of timing (crust ages), the large variations of NRM are mainly controlled by local variations of the extent of low-temperature oxidation, which is closely dependent on the permeability of the basalts. The presence of nearly unoxidized titanomagnetite in the massive basalts also indicates that the permeability of rocks is a main controlling factor to maghemitization. This is consistent with the observations by Johnson and Hall (1978), in which they showed that the impermeable massive flows from the Nazca plate can remain unoxidized for as long as 40 m.y.
The NRM of the dominant pillow basalt samples from Leg 187 probably originated from a primary TRM acquired during basaltic subsolidus cooling and from a CRM that modified the TRM during low-temperature oxidation and alteration. Recent studies have shown that the NRM of MORB can be partially contributed from unoxidized, very fine grained titanomagnetite in the interstitial glass (Zhou et al., 1997, 1999, 2000, 2001). Although we have also observed very fine grained Fe-Ti oxide minerals in the interstitial glass of the Leg 187 pillow basalt samples (Fig. F10B), the resolution of the SEM is not high enough to characterize their composition and texture. Further study on the magnetic minerals in these basalt samples will be carried out using a high-resolution transmission electron microscopy (TEM).