Previous   |   Next

doi:10.2204/iodp.pr.320T.2009

Paleomagnetism

Paleomagnetic analyses during Expedition 320T consisted of long-core measurements of the natural remanent magnetization (NRM) of split-core sections from Site U1330 before and after alternating-field (AF) demagnetization. During the expedition, the superconducting rock magnetometer (SRM), which is used to make these measurements, was being evaluated for its readiness following refurbishment of the JOIDES Resolution (see "New shipboard Paleomagnetism Laboratory," below). Initial measurements indicated that a significant source of magnetic field noise existed in the laboratory. The first 3 weeks of the 4 week expedition were mainly spent determining the origin of this noise and attempting to abate its impact on measurement quality. Once we reduced the magnetic noise to acceptable levels, the last week of the expedition was used to further test the SRM (track positioning and sample measurement accuracy and precision) and to measure a few but not all of the split-core sections. In this section, we highlight paleomagnetic results from one short interval. Additional data collected the last few days of the expedition are available in the LIMS database.

We also obtain rock magnetic information from susceptibility data and from coercivity and magnetic concentration constraints derived from NRM data. Susceptibility data were collected on the WRMSL, STMSL, and SHMSL.

Because Site U1330 is at the same location as Site 807, the paleomagnetism and rock magnetism results from sediments collected during Leg 130 (e.g., Kroenke, Berger, Janecek, et al., 1991; Musgrave et al., 1993; Kok and Tauxe, 1999) provide details beyond the shipboard measurements presented here, such as a magnetostratigraphy, a relative paleointensity record, and details on reduction diagenesis. They also provided us with background information that helped guide coring and measurement strategies as we tested the SRM and extracted new paleomagnetic information from the core.

Paleomagnetic results

Much of the paleomagnetic data from the expedition was collected while the SRM software was in a constant state of change. Units were being recorded in cgs rather than SI units until 2 March 2009, the sign of the y-axis moment was inverted for some measurements, and declination data were therefore in error for most of the data collected before 2 March. On 2 March, we measured Cores 320T-U1330B-1H and 2H after the software had matured sufficiently that all computations done internally were correct and all data were output in SI units.

We also used the raw moment data from Section 320T-U1330B-8H-7 to calculate the paleomagnetic direction in order to illustrate the capability of the magnetometer independent of the software. Progressive AF demagnetization of this section shows a smooth decrease in magnetization with an associated migration of the direction from a steep drill string overprint direction to a shallow direction, presumably related to the depositional remanent magnetization (Figs. F21, F22). The declination points fairly consistently to the north, which may be caused by a radial drilling overprint. Given the very weak magnetization of these sediments, the results illustrate that the magnetometer is capable of making accurate and precise measurements.

Kok and Tauxe (1999) observed the Brunhes/Matuyama reversal at 12.19 m depth in Hole 807A. Based on this information, we examined paleomagnetic data from this interval in Hole U1330B. The Brunhes/Matuyama is recorded by a change in declination of ~180° between 11.7 and 12.2 m depth, giving a mean depth of 11.95 m, which is very consistent with the Hole 807A Brunhes/Matuyama (Fig. F23).

Magnetic susceptibility

Raw susceptibility data from the three multisensor loggers contain several artifacts caused by drilling contamination (rust), drilling disturbance, and measurement practices. These dominate the susceptibility record available from the LIMS database (Fig. F24).

Drilling contamination appears to be a problem in the upper ~2 m of all cores from Hole U1330A and in the upper ~2 m of Cores 320T-U1330B-5H through 10H. The largest signal is very likely caused by rust from the pipe. Pyrite pieces that are not as easily washed out of the hole during drilling also accumulate at the top of each core. The top of Core 320T-U1330A-3R is particularly contaminated, with susceptibilities exceeding 400 x 10^–5 SI. We confirmed that this was not an artifact of the WRMSL measurement for Section 320T-U1330A-3R-1 by repeating the measurement on the STMSL. The results were virtually identical.

Exceptionally low (negative) susceptibility values are observed for all data at the ends of each core owing to the WRMSL measurement practice in which the density standard is measured directly after the last section of each core. The density standard is made of aluminum and has a minimum (peak negative) value of about –8000 x 10^–5 SI. This standard needs to be measured at least 40 cm below the last section to avoid this artifact. Similarly, the susceptibility values from the STMSL are shifted by about –14 x 10^–5 to –20 x 10^–5 SI (Fig. F25). This is related to the plastic bars on which the split-core sections sit and the braces on which the plastic bars sit. This problem could be easily overcome by measuring the susceptibility of the track without a sample present and then using this "track signal" from each measurement as a background correction.

Once these artifacts are removed, the magnetic susceptibility measurements give low but positive values (~3 x 10^–5 to 10 x 10^–5 SI) for the upper four cores from Hole U1330B (above 32 m depth) (Figs. F26, F27). Given these low susceptibilities, measurement precision could have been improved by taking more than one measurement at each depth. By only taking one measurement, which was recorded as an integer raw meter value, the data plot with steplike precision from one observation to the next (e.g., the WRMSL data in Fig. F25). Below Core 320T-U1330B-4H the sediment is diamagnetic, resulting in slightly negative susceptibility values (about –1 x 10^–5 SI). This pattern mimics that observed for Site 807, with the drop to negative susceptibility occurring at ~32 m depth (Kroenke, Berger, Janecek, et al., 1991), which is in the coring gap between Cores 320T-U1330B-4H and 5H.

New shipboard Paleomagnetism Laboratory

As part of getting the new Paleomagnetism Laboratory ready for future expeditions, we conducted a variety of tests of the paleomagnetism equipment and did magnetic field surveys of the laboratory and inside the instruments. This included testing and calibrating the KappaBridge (KLY 4S) susceptibility meter using vendor-supplied standards and testing the Schonstedt TSD-1 shielded oven, DTECH AF demagnetizer, portable fluxgate magnetometer, and ASC IM-10 impulse magnetizer for general functionality. All of these instruments were functional, as was the SRM, by the end of the expedition.

Software issues aside, the SRM functioned poorly during the first 3 weeks of the expedition owing mainly to magnetic noise. Initial measurements of the carbonate ooze split-core samples were plagued by large flux counts on the y-axis of the magnetometer, which could occur abruptly even when the sample was not moving nor even fully in the SRM sensor region. Given the relatively weak magnetizations of carbonate ooze sediments, these large flux counts were unexpected even when the sample was in the sensor region.

The rapid flux counts, which were clearly unrelated to the magnetic remanence of the carbonate ooze samples, have been observed before on past expeditions and in shore-based laboratories but have not been well explained. They are commonly referred to as flux jumps and are suspected to arise from a number of factors. Most commonly flux jumps are associated with the inability of the superconducting quantum interference device (SQUID) sensors to count (i.e., measure) the number of flux quanta associated with highly magnetized materials as the material moves through the sensor region. For example, mafic igneous rocks are notoriously difficult to measure accurately in SRMs owing to their high magnetizations. Similarly, flux jumps occur frequently when large gradients occur along a long-core sample, as would be the case when a weakly magnetized carbonate is juxtaposed against a highly magnetized ash or clay layer. Flux jumps may also be caused by random power surges that generate radio-frequency (RF) magnetic fields (generally those magnetic fields with frequencies from a few hundred kilohertz to a few gigahertz). RF magnetic field surges are usually relatively rare and can often be traced to specific equipment, like the saws used in the core splitting room and arc welders used during ship maintenance. As noted in the 2G Enterprises guidebook (Applied Physics Systems, 1995), SQUID sensors are very sensitive to RF interference.

In order to track down the source of the magnetic noise, we conducted a number of tests and measurements. First we mapped the static (long wavelength) magnetic field in the laboratory and within the SRM by doing magnetic surveys with a three-axis fluxgate magnetometer. The frequency response of the fluxgate magnetometer is flat from direct current to 250 Hz (Applied Physics Systems, 2003).

The magnetic field in the laboratory met with expectations, with field values generally comparable within a factor of 2–3 to Earth's magnetic field (Fig. F27; Tables T11, T12, T13, T14). Highs (80,000–130,000 nT) were associated with some of the air-conditioning vents, the corner of a wall near the elevator, the SRM cold head, and the region near the forward wall, which has the air-conditioning unit behind it. The two support columns produced lows near the ceiling of ~20,000 nT. The position and size of the magnetic anomalies are, of course, dependent on geographic position, ship orientation, and the relative positions of electronic equipment and metallic objects. Mainly, the survey showed that there are no anomalously high magnetic fields within the Paleomagnetism Laboratory.

Magnetic surveys along the SRM track inside the shielded region (Fig. F28) indicate that a very low null field (1 nT while the ship was stationary in Guam and 7 nT while transiting toward Hawaii) had been trapped by the marine technician (M. Hastedt) during the Guam port call. This is comparable to the field trapped within the SRM that resides inside a shielded room on the Chikyu. This is also the lowest field region within any of the shielded instruments on the JOIDES Resolution (SRM, Schonstedt oven, or D-Tech AF demagnetizer). Elsewhere within the SRM shield, the field is <400 nT, which is ~2–4 times higher than similar regions in the Chikyu's SRM. Exceptionally high fields occur within the shielded region of SRM at the joints in the magnetic shield, on either side of the in-line AFd magnetometer. This is true for the SRM on both the JOIDES Resolution and the Chikyu. The maximum field is ~1400 nT for the JOIDES Resolution and is <500 nT for the Chikyu. During the expedition, we took the three main sections (the main SRM section, the in-line AF unit, and the back slim-diameter shielded region) apart and attempted to reinstall them with tighter connections between sections. A much tighter connection was established, shortening the track by ~2.5 cm. Unfortunately, the magnitude of the magnetic field at these joints did not change appreciably.

We next ran a series of samples consisting of split-core sections from Hole U1330B, a concrete split-core sample (153 cm long), and empty tray samples. Flux jumps were absent when measuring the concrete sample or empty tray but occurred on virtually all of the Hole U1330B carbonate ooze sections. We noted that the flux jumps were common when the back of the 150 cm long sample spanned the two joints in the shield and the front of the sample was near or in the sensor region. The number of flux jumps increased significantly when the servo motor that drives the track was engaged and were notably higher when the power supply was plugged into the ship's unregulated power versus the regulated power. The power supply for the servo motor was near one of the shield joints and had a power cable that was not magnetically shielded. With the help of the ET (J. Kotze), we rectified this by moving the power supply to the very front of the magnetometer and by shielding its cables and any other cables in the vicinity of the magnetometer. This reduced the RF magnetic field significantly but not completely.

We continue to track the origin of the remaining RF magnetic fields and to understand why the position and type of sample controlled the occurrence of flux jumps. The best two indicators of when flux jumps were prone to occur or were occurring were to watch the digital readout on the SQUID electronics (in count mode rather than analog mode) or to watch the oscilloscope for degradation of the tuned SQUID signal. The SQUID electronics digital readout (SEDR) allows for all three axes to be monitored, whereas the oscilloscope monitors only one axis at a time. We connected the oscilloscope to the y-axis because it was the most prone to extraneous flux counts. An antennae was constructed by the ET that would receive or transmit RF magnetic fields of a few gigahertz. Stray RF magnetic fields were noted when the antennae was in receiver mode and moved around the laboratory, but no clear new sources of RF magnetic fields were identified.

In contrast, thousands of flux counts were measured by the SQUID when the antennae was in transmit mode and a carbonate ooze sample was in the magnetometer, in a position as described above, or when it was basically anywhere near the SQUID sensors. No flux counts were noted when the sample was removed. With the sample in the magnetometer, the number of flux counts increased dramatically when the antennae was pointed at the joints in the shield.

This is very conclusive evidence that the RF magnetic fields are being transmitted through the poorly shielded joints. Furthermore, the sample is acting like an antennae or wire in that it is transmitting the RF magnetic fields along its length from the joints in the shield up to the SQUID sensors, in a process called RF magnetic induction (e.g., see Wikipedia).

We worked diligently the rest of the expedition to track down other sources. A significant source arose when the new 220 V outlet was installed at the fore end of the laboratory. When the crest amplifier was plugged into this outlet, the SRM recorded a large number of flux jumps during measurements. This problem was solved by plugging the amplifier into a transformer, which was then plugged into the ship's regulated 110 V power. With these changes, we managed to reduce flux jumps that occur during measurements to a manageable level.

Top of page   |   Previous   |   Next