Exp384 Paleomagnetics Technical Report

Gary Acton, Mark Higley, and Alex Roth

Scientists: NA

Summary

Expedition 384 was an engineering testing expedition in the North Atlantic which visited 2 sites, U1554 and U1555. The primary objectives for site U1554 was to test IODP’s various magnetic orientation tools (MOTs) and site U1555 was to test new bit performance in hard rock. Throughout the entire expedition, a secondary objective of the paleomagnetism (Pmag) lab was to use the entire Pmag lab for performance verification of all systems, equipment, and software as well as provide a thorough training opportunity for the two new Pmag technicians.

At Site U1554, 3 APC holes were cored to approximately 70 mbsf and a 4^th shallow hole was drilled down to approximately 14 mbsf. Four MOTs, 3 Icefield MI-5s and 1 Flexit, were deployed for testing purposes while coring these holes. The MOTs consistently gave values expected for the site, with the exception of one tool which gave values 180° off on average. The source of this error was fortunately discovered at the end of the expedition and once corrected, this tool also gave good results. This discovery may have larger implications for MOT data from previous expeditions that also have values that appeared to be 180° from anticipated values. Other, smaller sources of error affecting MTF angles were identified during the set up and the MOT and the core barrel on the rig floor. These errors could account for values which appear to differ by less than 180° from the expected value. Of the 25 cores recovered at site U1554, 24 provide an excellent, high-resolution record of Bruhnes age sediment including several geomagnetic excursions. Only one core was excluded due to significant coring disturbances.

A total of 7 holes were drilled at Site U1555 in order to test new bit performance; two of these holes were cored. Hole F was RCB cored from 0 mbsf to approximately 184 mbsf using a PDC bit which generally yielded excellent core samples. Hole G was drilled down to 168.6 mbsf then RCB cored from 168.6 mbsf to 309.5 mbsf. Cores from site U1555 consisted of basalt which generally recorded normal polarity.

COMMENTS AND ISSUES

General Lab

The Pmag lab was thoroughly exercised during Expedition 384. With no scientists on board, the technicians were required to collect and run all samples which provided excellent training opportunities for two new Pmag techs. Every instrument in the lab was used to analyze and measure samples. This also offered a chance to examine and evaluate all systems, from instrument and software performance to user guides and protocols, under supervision of a dark lord Pmag sith (Acton).

Discrete samples were collected from sediment cores at Site U1554 and from hard rock cores at Site U1555. Sampling methods for sediment included both push samples and extruded samples into either ODP cubes or Japanese Cubes (J-cube). Hard rock discrete samples were collected using the parallel saw, rock saw, and mini-corer.

The paleomagnetic standards page and user guide on confluence were both updated. When the standards were measured during expedition Exp378, declination and inclination appeared correct but the intensities that were measured were approximately an order of magnitude less than the listed intensity. During this expedition it was realized that the standard’s listed intensity was calculated using a volume of 1 cm³ not 7 cm³ as was assumed. When the measured intensity is calculated using 1 cm³, the measured intensity is in agreement with the listed intensity. The table which contains the standards values has been updated to have both values.

Technicians could not locate the J-cube sample guide used to keep J-cubes aligned when punching into sediment. A new guide was printed using the 3D printer as well as a guide for the J-cube extruder. These guides are kept in the drawer to the left of the SRM loading area. 

SRM

A significant bug was found in the SRM IMS software related to discrete measurements. After measuring a background for the discrete tray, the same tray was measured with no samples in it. These null samples were named Empty_003, Empty_013,…, Empty_153 and treated as a typical 7 cm3 volume cube from the Working half. The orientation in the IMS 10.2 software was set for the arrow on top of the cube to point out of (away from) the SRM. This is called TOP-AWAY in the lingo of the SRM. The goal was to see what the noise level is for the SRM for a typical cube. The values were quite high, with intensities >1E-04 A/m. From looking at the graphs and data from IMS, it is clear that the background correction is in error. Rather than subtracting the background, it was added for the X and Y moments and subtracted for the Z moment. The experiment was then repeated for the orientation in which the arrow of a cube sample would be on top and pointing into the SRM (i.e., TOP-INTO) and the sample was assumed to be from a Working Half. This resulted in the background being subtracted from the X and Y moments but added to the Z moment. The only setting in software in which the background is properly subtracted is if the sample is assumed to be from the Archive Half and the TOP-AWAY arrow orientation is used. It is clear from this experiment that the background correction is being made after the sample measurements have been converted into their orientation rather than before, which should be the case. For example, when the X-moment of the sample (Xs) is measured, the X-moment of the background (Xb) should be subtracted and then the coordinate transformation completed. The corrected moment would be = Xs-Xb, which then would be transformed into preferred coordinate system. For the Archive half orientation, the transformation matrix is (1 0 0, 0 1 0, 0 0 1). In other words, no changes to the axes are required. For the working half, the matrix is (-1 0 0, 0 ­1 0, 0 0 1). The X and Y moments are multiplied by -1. Because the software is completing this transformation prematurely, the moment it computes is = ­Xs ­ Xb, and so the background is not removed from Xs, it is instead doubled.

Given that background measurements are only done correctly for discrete samples in the Archive-Half Orientation, all discrete samples should be measured when the SRM is set for Archive-Half orientation. The software, however, will not allow working half samples to be measured while the orientation is set to archive half. For a typical sample collected by pushing a cube into the working half, the user would merely have to rotate the sample to have the arrow on the bottom to have the sample in the Archive half orientation system. Unfortunately, the software does not allow that option. The work around for getting the software to allow the samples to be entered with the orientation system set to working half is to start the measurement sequence but then abort it after the tray has started to move. Then the user starts the measurement again after using the “Recall samples” button and setting the orientation system to archive half. The software does not complain about the working half samples being measured, the background is removed correctly, and the data are output with the proper orientation transformation as long as the samples are place in the tray correctly. For a push sample, the sample should be placed arrow-side down into the tray and pointing away from the magnetometer (BOTTOM-AWAY) and for an extruded sample the arrow should be at the top and pointing away (TOP-AWAY).

The IMS DAFI U-Turn utility was tested during this expedition. The program appeared to make the corrections and the corrected data was shown in the IMS window as expected. The output .SRM file, however, does not have the corrected data. The data in the output file does not match what is displayed in the IMS window.

Upon completion of the previous expedition, it was noted that the volume correction for discrete samples was not being performed correctly. Further investigation during this expedition revealed that the volume correction for discrete sample is in fact being done correctly but the sample information can be misleading. In the IMS sample preset editor, the dimensions box displays the text ‘Sample Area’ and the units are area units. The number corresponds to the volume of a J-cube though. Despite this confusion, IMS seems to understand that if a discrete is being measured, then the dimension values are volumes and makes the correct calculations.

Offline treatments for discrete samples continue to be entered in the comments section. The lingering bug which puts IMS into an infinite loop if you try to enter offline treatments through the sequence editor still exists.

During port call a null field was trapped twice. In both instances, the trapped field was good but it was repeated for training purposes. The lab area was thoroughly cleaned to remove any remnants of dust from the previous dry dock.

JR-6A Spinner

The JR-6A spinner magnetometer was used to measure NRM, ARM, and IRM (acquisition and AFD) for select discrete samples (see appendix A for a table of discrete samples and treatments). Since both sediment (Japanese cubes) and hard rock (cubes and cylinders) samples were analyzed all measurements were run at slow spin speed. All samples were inserted with the split plane up arrow pointing up and to the left and the split plane surface out of the sample holder. Since extruded samples are flipped around the z-axis by 180°, the samples were insert with the split plane up arrow pointing up and to the left but with the split plane surface into the sample holder (Figure 1).

Figure 1 Sample Placement for Extruded Sample

Since cylinders need to be rotationally aligned in the sample holder, consistent results can be achieved by orienting the split plane up arrow to point up and left in the same manner as cube samples. In this case, a reference mark should be made on the sample holder for alignment purposes. Alternatively, the split plane up arrow could be aligned with one of the notches in the sample holder but then the proper orientation parameters would need to be determined and set in the Remasoft software.

D-2000 AF Demagnetizer

Selected discrete samples were subjected to AFD (stepwise up to 100 mT and even 200 mT in several cases). The D-2000 was also used to impart a 50 μT ARM on select samples. No issues were noted. 

ACS Impulse Magnetizer

Both the IM10 and IMS10-30 impulse magnetizers were used to impart a 1000 mT IRM on select samples and complete 1000 mT IRM acquisition curves were measured for 2 of the samples. All samples treated with IRM and subsequent AFD were measured in either the SRM, the JR-6, or both for instrument comparison. Samples subjected to a field this high caused many large flux jumps when measured in the SRM. To help mitigate this, the SRM was slowed down to 1 cm/sec which helped reduce flux jumps on sediment samples to near zero.

The vendor manuals were used to create user guides for both the IM10 and IM10-30 since neither existed. Field versus voltage tables were generated from the vendor calibration data for the IM10 and the IM10-30 (coils #2, #3, and #4) as the user needs to know the required voltage for a desired field (tables were previously displayed in reverse). The files containing the tables have an editable table where the user can enter whatever fields desired and the new B vs V table will be automatically generated.

When measuring samples which were run in the IM10-30 impulse magnetizer, it was realized that the field is directed into the unit rather than out of the unit as is the case for the IM10. A label was placed on the IM10-30 noting this and it is noted in the user guides as well. Labeling on the IM10 and IM10-30 was updated to reflect this and unnecessary sample holder labeling was removed as well to avoid potential confusion.

Thermal Demagnetizer

The thermal demagnetizer was used for several samples (including sediment J-cubes and a hard rock cube and cylinder). The sediment cubes were given a 1000 mT IRM before the thermal demagnetization. Sediment J-cubes were heated up to 125°C safely then removed from their plastic cubes and the remainder of the heating was done out of the plastic cube. Sediment and hard rock samples were heated from 100°C up to 600°C generally in 50°C increments. After each heating cycle, the samples were measured in the SRM; the hard rock samples were measured with the JR-6 as well for comparison.

Kappabridge

The Kappabridge was used to measure both bulk susceptibility and magnetic anisotropy for 15 sediment samples. Both AMS spin and sufar were used for running the program. The user guide was used to walk through the steps for measuring samples however some steps were not clear of the information was incorrect. The AMS spin section of the user guide received numerous edits to bring it up to date. Sufar was not used extensively nor was the user guide heavily scrutinized since this software will be phased out with the new kappabridge. An output file from the vendor for the new unit was verified to open in the Anisoft5 program and the result communicated to shore.

CORE Orientation Tools

One of the primary objectives of this expedition was to determine if the MOTs were functioning properly. The tools give a magnetic tool face (MTF) angle which is then used to correct the core orientations and give declination values. There have been years of inconsistent of possibly incorrect declinations being reported and despite numerous testing in the past, there was still no consensus on what was causing the errors. There are 5 MOTs onboard the Joides Resolution: 3 Icefield MI-5 tools and 2 Flexit tools (2007, 2043, 2052 and 0936, 0937, respectively). During this expedition, the tools were used in orienting the APC cores as well as tested in various setting on board the ship and on the dock away from the magnetic influence of the ship.

During port call in Kristiansand, Norway, tools 2007, 2052, 0936, and 0937 were tested on the dock to verify they could record the correct magnetic tool face (MTF) angle which is the angle between magnetic north and the tools orientation face. Tool 2043 was not tested on the dock during port call as it was not apparent that it was in fact onboard. Despite some flaws in the experiment related to top vs bottom of the tool (in lay terms - the tools were upside down), all 4 tools that were tested demonstrated their ability to correctly measure MTF.

Further testing on all 5 tools was conducted onboard the ship while underway to site U1554. Due to the magnetic influence of the ship, it is not feasible to measure absolute magnetic angles. Instead, tools were tested to see if they recorded the correct relative change in MTF when rotated a given angle clockwise. Since onboard the ship each MOT is likely in a different magnetic environment, they were normalized to 0° using the initial position and the following rotations measurements being relative to this position. These experiments concluded that all 5 tools were capable of correctly measuring relative changes in orientation. A more detailed write up of these tests is available on confluence.

Onsite at U1554, 25 cores were collected using four different magnetic orientation tools in Holes U1554A-D. Eight cores were collected with Flexit Tool 0937 and the remaining 17 cores were collected with three Icefield Tools (2007, 2043, and 2052). Following magnetic cleaning using progressive alternating field demagnetization, the mean paleomagnetic direction was estimated for each core. The resulting core mean paleomagnetic declination was used to determine the known orientation of the core, because Brunhes age (0-780 ka) sediments that are good paleomagnetic recorders, like those at Site U1554, will have mean declinations of approximately 0°. The difference between true north and the paleomagnetic declination gives a paleomagnetically-determined reorientation (PDR) angle that can be compared directly with the MTF angle. If the magnetic orientation tool is accurately measuring the core orientation, the difference between the PDR angle and MTF angle should be negligible relative to the errors in the method, which were expected to be roughly ±15°. 

The orientation angles (PDR–MTF) for 20 of the 25 cores differed by <28°, with a mean difference of 8.7° and a standard deviation of 13.9°. One core (U1554B-5H) had significant coring disturbance throughout and was not used in the assessment. The paleomagnetic directions were clearly disturbed in this core, and likewise the PDR–MTF was somewhat larger (34.8°) than observed for the undisturbed cores. The four results obtained with tool 2043 were all anomalous, with PDR–MTF differences of 155.3°, 183.6°, 192.6, and 189.0°, with an average of be 180°. which lead technicians to believe that the tool probably has the sign on a couple of the fluxgate magnetometers or accelerometers backwards.

This theory was further tested on shore in Kristiansand when the ship returned to port. Tools 2007, 2043 and 0937 were tested on the dock away from the magnetic influence of the ship.

All three tools correctly measured the correct absolute MTF. The only difference between this test and every other test performed was that the tools were tested without their pressure barrel. This suggested that the source of error was related to the pressure barrel. Further inspection of the pressure barrels revealed that the pressure barrel snubber used for Icefield 2043 during this expedition was out of alignment by 180 degrees (Figure 2). This misalignment would account for the incorrect MTF angles in cores 1 through 4 of Hole U1554B. The snubber was realigned. It is conceivable that the mis-aligned snubber and/or pressure barrel could have been used with different tools in the past which would account for instances of other tools recording incorrect MTF angles which are 180° off.

Figure 2 Snubber Alignment (Middle snubber is 180 degrees out of alignment)

Moving forward, technicians who set up the tools should record the pressure barrel/snubber part numbers in the orientation tool log as a means to troubleshoot suspicious data.  It is further suggested that each tool be used with the same pressure barrel/snubber each time. See Table 1 for suggested Icefield tool and corresponding pressure barrel scheme.

Table 1: Icefield Part Numbers

Flexit tool 0936 was not tested downhole due to a weakly soldered battery terminal wire. The terminal soldered connection broke during setup of the tool. The connection was re-soldered but due to the short length and thin gauge wire coupled with small connection terminals, it was difficult make a strong connection. Confidence in the soldered connection was not high enough to send it down hole with the possibility of not receiving any data.

Other factors related to how the MOTs are set up on the drill rig floor could have significant influence on the accuracy of the orientation data returned. Prior to arriving onsite at U1554, technicians were given a thorough walk through on how the MOTs are set up and connected to the core barrel which is summarized in the paragraphs below:

The orientation tool connects to the top of a sinker bar assembly via the keyed ‘T-slot’ fitting (Figure 3). The orientation tool can only fit into this fitting one way. The T-slot fitting on top of the sinker bar is threaded onto the sinker bar and the alignment is set using shims to limit the distance the T-slot fitting can be threaded on. These shims (Figure 3), which are essentially washers of varying thickness, are adjusted so that when the T-slot fitting has been threaded on and tightened, the orientation point of the T-slot (the center of the T opening) is in alignment with the orientation point on the sinker bar, which corresponds to the apex of the curve in the D-pin receiver (discussed later). Getting the shim spacing right could be a difficult process and care must be taken to ensure the alignment is correct. However, once the spacing has been set and the alignment verified to be correct, no further adjustments should be needed until the T-slot fitting needs to be changed. Technicians should double check the orientation of the T-slot prior to coring to ensure that it has been set correctly. The T-slot fittings can take a lot of abuse and may wear out over time. These fittings should be inspected periodically to ensure they are in good shape and replaced if necessary. Due to the difficulty in getting the shim spacing correct, there may be a reluctance on the rig floor to change the T-slot fitting. 

Figure 3 T-Slot Fitting and Alignment Shims on Sinker Bar

The sinker bar is then connected to a D-shaped pin on top of the core barrel. The D-pin on top of the core barrel has one side flattened so that it mates with the receiving end of the sinker bar (Figure 4) in only one orientation.

Figure 4 D-Pin Receiver on Bottom of Sinker Bar

Although there is a small amount of play in the entire system, so long as care is taken to align the T-slot fitting, there is little chance for significant mis-alignment issues in the setup of the MOT, sinker bar, core barrel assembly.

Larger errors in orientation are possible in the alignment of the core liner and were indeed witnessed during this expedition. A small hole is drilled in the top of the core liner between the double (working) lines using a guide made for this purpose. The core liner is then inserted into the core barrel and this hole is intended to line up with a corresponding hole at the top of the core barrel. A small set screw is then screwed through both holes to hold the liner in place. After a core has been fired, the core barrel is brought back to the surface. While the core barrel is hanging from the wireline, the set screw is removed. The drill crew was asked to orient the core barrel in the rack so that the set screw hole is facing upwards. This is a step that is not normally done. The reason this step was taken was to ensure that the double (working)lines on the liner were still aligned with the set screw; if the set screw hole is facing up, the double lines should be facing up as well. Because the only place that the liner is held in place is at the top, by the set screw, there was concern that there could be some torsional rotation of the liner near the bottom. Once the core barrel was laid down in the rack and clamped in place, the core catcher was unthreaded and removed, exposing the liner to be checked for proper orientation. At this point, a vertical line with directional arrows pointed to the double (working) lines on the liner were scribed into the bottom of the core (Figure 5). Next, the vertical scribed line was observed as the seal sub was unthreaded. The objective of the vertical scribe line was to ensure that once the seal sub was removed, the core was still in the same alignment with the liner after unthreading the seal sub.

Figure 5 Core Liner Double Lines and Scribe Line

Once the seal sub has been removed, the core liner is then pulled out of the core barrel. The set screw hole on the liner was inspected for each core. There were instances where it appeared that the screw did not make it into the liner or the screw was displaced from the drilled hole at some point. This was evidenced  by other indentations surrounding the drilled holed which did not fully penetrate the liner and gouges which seemed to indicate some sort of motion of the liner as the set screw was against it (Figure 6). Indentations were also observed on the liner top edge (Figure 7). A possible explanation for this was that the liner was never fully inserted into the core barrel at the time the set screw was put in place. In this case, the set screw would not go into the drilled hole of the liner but could be in place some distance above it. If the liner were then to shift upward, its upward motion would eventually be halted by set screw contacting the top edge of the liner (leaving the indentation). Subsequent discussions with SIEM Ops and the core tech determined that the most plausible cause for the indentations and gauges is due to the high impact force on the liner as it is shot causing momentary deformation in the liner and allowing the pin jump out of the hole.

Figure 6 Core Liner Set Screw Displacement

Figure 7 Core Liner Set Screw Displacement

MUT upload

Uploading of SRM data was done manually as measurements were completed. This was due to the frequency of taking ‘empty_tray’ measurements which were not uploaded to LIMS.  A python script was written to check for duplicate SRM measurements as LIVE is not useful since everything appears as a duplicate. The script is on the MAC computer (Documents/python_scripts/SRM Duplicates Search) in the PMag lab and runs through the command line. JR-6 and Kappabridge data was uploaded manually as well.

Orientation data was not uploading correctly through MUT. The first core in an uploaded file would be uploaded as the correct core. The remaining cores in that same file would be uploaded as a different core. It was strange that one core would upload as Hole C for example, while the remainder would be uploaded as Hole B, because the header line of the file is the only place that specifies the hole. The developer determined the issue was a statically defined variable. The issue would only appear if data from more than one hole is uploaded at the same time. The short holes and rapid turnover of orientation tools is likely why this issue arose this expedition and not before. A new version of MUT was rolled out and the data was re-uploaded. Old data that was incorrect was cancelled.

Appendix

A.1 Discrete Sample Information

Table A.1 Discrete Sample Information