It has been frequently suggested that the bias in the data toward the "0°" declination in the ODP coordinate system (i.e., upward in the archive half, or radially inward) may be an instrument problem. In particular, it has been suggested that this is the case in weak samples, which is consistent with the familiar observation that the tray magnetization often appears to be so oriented.
One explanation that has been considered is the lack of symmetry in the half-core measurement. This was, however, investigated by J. Gee (pers. comm., 1998) and shown not be an important effect.
A second possibility, which we investigated during Leg 189, is that the problem is caused by drift in the x-component of the direct-current superconducting quantum interference device (SQUID). This would give a bias toward the "0°" declination direction if the drift were positive and toward the 180° declination if the drift were negative. Conversely, if the principal drift were in the SQUID's y-component, then the bias would be toward the 90° and 270° declination, or, if in the SQUID's z-component, vertically up or down.
A related problem, which can be acute with very weak samples such as carbonates, is the contribution of magnetization of the core liner, which cannot be separated from the magnetization of the sediments in the standard measurement. A number of liners were measured as described below.
The principal preparation that the cores are subjected to on board is core splitting. After core has been brought to the catwalk from the rig floor, it is cut into 1.5-m-long sections. It is then split into working and archive halves. For all APC cores and some extended core barrel (XCB) cores, wire splitting is used. Some XCB cores and all rotary core barrel (RCB) cores are split with a saw. It has been suggested by Witte and Kent (1988) and J. Ali (pers. comm., 2000) that the act of splitting the cores is a source of magnetic overprints. Examples of all three core types have been measured as whole-section cores before splitting and then compared with archive-half core measurements.
There is a soft moment that can usually be removed from almost all ODP recovered sediments by alternating-field (AF) demagnetization to 20 mT. It has often been suggested that the magnetism of the drill pipe is responsible for this overprint. However, because the core is inside a very long magnetic pipe, this is not very likely. The core should effectively be within a magnetic shield, apart from possible leakage at the joints. Indeed, we know the fields in the pipe are not very high because they are measured by the Tensor tool fluxgate every time the Tensor tool is used on a piston core. The strongest fields seen are ~0.1 mT.
The soft moment changes sign sometimes when the BHA is changed, which suggests that it is more likely picked up as the core passes through the BHA and, in particular, through the field of the drill bit. This suggestion was investigated during Leg 174B (Fuller and Garrett, 1998). A wash core was measured, demagnetized, remeasured, and then reassembled for tripping. The wash core was tripped in an experimental APC assembly, passing through the BHA. The experimental assembly consisted of two sections of nonmagnetic core barrel and one standard barrel above them. The standard barrel gave a relatively strong field at the joint with the nonmagnetic barrels. On its return the reassembled wash core was measured, demagnetized, and measured again. It had picked up a relatively strong moment in the same sense as the soft moment in cores drilled at the site. The field at the joint had also been recorded. The acquired magnetization was demagnetized by 20 mT, reversing the inclination to its value before tripping.
The processes involved in piston coring and rotary drilling obviously have considerable potential to affect the magnetization of recovered sediments. These processes can be broken down into a number of different aspects for each type of coring or drilling.
In APC coring, the passage of the cutting shoe through the sediment causes deformation, as does the passage of the APC core barrels into the sediment and the core barrel's slowing to come to rest in the sediment. Earlier experiments showed that the nature of the cutting shoe can have an effect. As noted above, it was shown during Leg 182 that the use of a nonmagnetic cutting shoe markedly reduced the "0°" declination effect on occasion. However, on other occasions, it seemed to make little difference. It appeared that the physical properties of the sediments were involved in determining the effect of the cutting shoe. During Leg 174B, it was demonstrated that the magnetization of the APC barrels was reflected in the magnetization of the sediments. This was consistent with an earlier observation that alternate cores had a similar pattern of magnetization reflecting the magnetization of the alternately used APC assemblies (Schneider et al., 1987). Our experiments suggested that the magnetization was acquired as the barrels came to rest in the sediment. We did not see the hard magnetization when the wash core was tripped, at which time it passed up and down the pipe. However, we did see this magnetization in the experimental coring, when the core barrel could have provided a field as it came to rest in the sediment. This field could then have been recorded by the sediment as it recovered from mobilization during coring. In XCB and RCB coring, there are the additional problems of the effect of the rotary cutting in the drilling process, but our main emphasis is on the APC coring, which is the method used for the sediments carrying the most recent records of the field.
In addition to magnetization processes, which are field dependent, there is also clearly the possibility of field independent deformation remagnetization. This problem has been addressed by M. Okada (pers. comm., 1998). He has developed a correction for the turning down of the sediments at the margin of the core. This was applied successfully to improve the inclination record. As noted above, this work has been extended by G. Acton (pers. comm., 2000).