Hammer drill testing was carried out on the Atlantis II Bank, located east of the Atlantis II Fracture Zone which offsets the SWIR between 31º50'S and 33º40'S at 57ºE (Fig. F1). The platform is a flat-topped bench, ~9 km long and 4 km wide. This massif was successfully cored during Legs 118 and 176 to a depth in excess of 1500 meters below seafloor (mbsf). This location was chosen primarily as an area of opportunity, to coincide with a point in the hammer drill development when sea trials were in order. As an additional benefit, the shallow water (~700 m) was envisioned to facilitate efficiency in our early operations. The flanks of the massif offered additional targets at greater depth and with topographic slopes, should the engineering tests advance beyond the initial objectives.
Site 1104 is located at a water depth of 731 m on the east rim of the Atlantis II Fracture Zone, ~200 m northwest of Hole 735B (32º43.32'S, 57º15.85'E; Fig. F2). This site was selected based on a video survey beginning at Hole 735B, by which we sought to find a reasonably flat, large outcrop to initiate the first spud tests of the hammer drill. During the transit from Cape Town, the water hammer was successfully deck-tested. The initial assembly of the drill string including just the SDS hammer and a concentric arm bit to test the spudding capability of the system. A frequency analyzer for monitoring hammer operations was built and installed during the transit.
After a 4-hr video survey starting at the Hole 735B guide base, we selected a location with extensive, relatively flat-appearing outcrop and set the bit down on the outcrop to see how the hammer functioned without rotation. Several spud tests indicated the hammer was performing as expected, so we decided to recover the subsea camera and begin hammer drilling Hole 1104A. After ~45 min, it appeared from rig floor observation that we had made ~1.5 m of penetration, so we deployed the camera to inspect the hole. We had also noted excessive vibration of the stand pipe and derrick during hammer operations. A clean, circular hole was apparent on the video image, so we recovered the camera and initiated a second test hole (Hole 1104B). After ~2 hr rotation, we had made ~2 m of penetration but also noted increasing torque and slower rate of penetration (ROP). Another camera trip revealed a second clean, circular hole, but some apparent obliquity indicated the hole had been initiated on a small local slope. We recovered the camera and attempted to spud a third hole, but the hammer would not fire, so it was pulled to the surface.
Inspection of the concentric arm bit indicated that the reaming arms were damaged and that a valve had cracked in the hammer. The hammer was rebuilt, a new concentric arm bit installed, and we ran the assembly back to the seafloor. After a short video survey to inspect the site, we recovered the camera and spudded Hole 1104C. In <2 hr, although we noted ~2 m of penetration, there was also indication of high erratic torque and ROP effectively ceased. We attempted to initiate another hole (Hole 1104D) but made no advancement and experienced high and erratic torque, so this hammer test was terminated and we pulled the drill string. During the pipe trip, we deployed the two United States Geological Survey (USGS) ocean-bottom seismometers (OBS), 100 and 300 m, respectively, from our drill site to monitor the noise transmitted through the outcrop which was generated by the hammer.
When inspection of the second concentric arm bit indicated once again that the underreamer arms had experienced excessive wear, a bit was modified by trimming the concentric arms to match the outside diameter of the pilot bit. After this modification, however, the bit did not appear robust enough to cut through the hard rock so this modification was abandoned. A second modification removed the concentric arms, cut the bit shank, and welded the interval where the arms had been closed. We had hoped to test the drilling capability of the bit without the added challenge of attempting to ream out the hole. Unfortunately, during the modification of the bit, a crack developed and the bit was set aside. We then modified a third bit by welding the concentric arms closed. This bit was tripped to the seafloor and we initiated Hole 1104E. After ~1 hr, we had made ~1.5 m of penetration, but the bit stuck in the hole. We were able to free the bit with left-hand rotation, indicating that the arms had broken free and were causing the bit to stick. Having exhausted all the bits we had on board for hammer testing, and with the promise of delivery of a different bit design in a few days from a supply vessel, we chose to commit to conventional rotary coring while we waited for the equipment transfer. The OBS and positioning beacons were recovered, thus ending operations at Site 1104.
Because of continued deterioration of sea state, only the new bits were transferred from the supply vessel. After delivery of the new bits, we returned to the location of Site 1104 in anticipation of continued hammer testing. Despite the same coordinates as Site 1104 (32º43.32'S, 57º15.85'E; Fig. F2), because a new beacon was deployed and we wanted the record to indicate the next phase of hammer testing, this location was assigned as Site 1106. The hammer was tested on the deck in preparation for deployment.
The hammer was deployed, and after a
brief seafloor survey, Hole 1106A was initiated. After ~2 m of penetration, the hammer
ceased activity, and we tripped it back to the surface. Once again a valve had broken in
the hammer, potentially because of excessive heave during the continuing poor sea state.
Hole 1106B was initiated on the ensuing pipe trip, which included the second of the three
bits that we acquired on the transfer (coincidentally the last bit capable of drilling an
overgauge hole) because the first bit was worn after Hole 1106A. Only ~
m of penetration was realized before the hammer ceased activity again, necessitating
another pipe trip. Again the bit was worn, so it was replaced with the last of the bit
configurations we had available, a flat-faced drilling bit. Our decision to run this bit
was based on the assumption that if we could demonstrate the ability to make a hole, we
could use this information in future bit design.
Hole 1106C was spudded and we drilled ~1 m in <1 hr before the hammer stalled again. On the ensuing pipe trip, the piston in the hammer was replaced and the flat-faced bit was run back to the seafloor. Hole 1106D was attempted, but the hammer would not start, so it was pulled and rebuilt once again. Hole 1106E was initiated, and the hammer drill system performed admirably, cutting an 8-m-deep hole in <2 hr. The pressure transducer on the stand pipe gave way at this time, so the pumps had to be shut down for repairs. Once the repair was completed, the driller noted no pressure buildup and was able to slowly lower the drill string to the total depth of the hole + 4 m, indicating that we had lost some of the bottom-hole assembly. The subsequent camera trip indicated that the bit and hammer were indeed missing, and because we could not see them on the seafloor, we assume they are still in Hole 1106E. Weather conditions had still not improved, and we did not have a clear idea which of the several holes within a few meters radius was Hole 1106E, so a fishing attempt was unrealistic. Given that we had exhausted all the bits and hammer spare parts, we declared the hammer test for Leg 179 complete, and got under way for Ninetyeast Ridge.
In summary, although a detailed summation of all the data relevant to the hammer testing awaits postcruise development, we do have some preliminary impressions. We are encouraged by the performance of the hammer and will be able to use this series of tests for optimal design improvements. Despite the less than desired performance of the bits, again we are optimistic, particularly based on the last test where we made 8 m of penetration in <2 hr, that bit design improvements will yield improved performance in the future. Finally, as with all our operations, sea state appears to be a primary control if not on the success of an operation, at least on its duration and ease of completion.
Because of delays in the resupply ship that inhibited resumption of hammer tests near Hole 735B, Hole 1105A was drilled during Leg 179 for a period of 6 days. The hole was located ~1.3 km east-northeast of Site 735 on the Atlantis platform along the eastern transverse ridge of the Atlantis II Transform (32º43.13'S, 57º16.65'E; Fig. F2). The site is along a near-ridge axial trend with respect to Hole 735B, but more distal from the north-south trending Atlantis II Transform that lies to the west. The site was chosen to avoid a duplication of Hole 735B efforts that might occur by drilling at proximal Site 1104. At the same time, we wanted to utilize Hole 735B as a reference section to attempt lateral correlation of large-scale igneous units, structural features, and geophysical characteristics over the broader distance represented by the offset in holes in the direction approximately parallel to the former ridge axis. In addition, the site was chosen to help constrain the overall structure of the massif exposed on the platform. If successful, the correlation experiment could yield a minimum measure of the dimensions of subaxial magma chambers and continuity of structure and processes along strike of the ridge axis at a very slow-spreading center. If correlations are unsuccessful we can limit the dimensions of igneous units, former magma chambers, or structures to be smaller than the scale of the experiment. Correlation will be attempted on the basis of detailed and integrated data sets including core descriptions and subsequent shore-based laboratory analyses to establish cryptic chemical and mineralogical variations, and the alteration and structural framework of the core. A full and highly successful logging program that was completed after the cessation of drilling will aid in the correlation attempts.
The hole penetrated to a depth of 158 m, and the cored interval measured 143 m, starting 15 mbsf. Core recovery included 118.43 m of gabbroic rock for a total recovery of 82.8%. Together with logging results, this recovery provides nearly complete coverage of the rock types and a comprehensive view of pseudostratigraphy in the gabbroic section cored (Fig. F3). Shipboard results now indicate a high probability that specific units, structures, and/or geophysical characteristics from Holes 735B and 1105A may indeed be correlated.
The cores recovered record a wide variety of rock types ranging from gabbro (Fig. F4), oxide gabbro with up to 20-25 modal% Fe-Ti oxides (Fig. F5), and olivine gabbro (Fig. F6) to scarcer troctolitic gabbro, gabbronorite, and felsic rocks such as trondhjemite. One hundred forty-one rock intervals have been described within the core and defined on the basis of distinct changes in mode, modal proportions, grain size, and/or texture. Well-defined igneous layer contacts or structural boundaries to these intervals are preserved in many sections of the core (Fig. F7). The highly layered nature of the gabbroic rocks documented within the core is supported by high-quality continuous Formation MicroScanner (FMS) logs of the borehole (Fig. F8), as well as other logs and whole-core magnetic susceptibility measurements (Fig. F9). The scale of the layering in the core varies from a few centimeters to meters. On a broader scale, the intervals define four basic units from top to bottom consisting of (1) a gabbroic unit characterized by more primitive rock types, a scarcity or lack of oxide gabbro, and a low (828 SI) average magnetic susceptibility, (2) a gabbroic unit characterized by a high abundance of interlayered oxide gabbro and oxide-bearing gabbro, and a high (3208 SI) average magnetic susceptibility, (3) a gabbroic unit characterized by more primitive rock types and a lack of oxide gabbro with low (780 SI) average magnetic susceptibility, and finally (4) another unit rich in oxide gabbro and oxide-bearing gabbro with high (3472 SI) average magnetic susceptibility. Rocks of all four units are crosscut by millimeter- to decimeter-sized veins of leucocratic gabbro, quartz diorite, trondhjemite, and irregular pegmatitic gabbro intrusions. Irregular veins and bands of oxide minerals have also been observed.
Thin sections indicate typical cumulate textures in the majority of samples that range from adcumulate to orthocumulate and show variable amounts of core-to-rim zoning in plagioclase. Poikilitic textures are also common with pyroxene as the oikocryst phase and plagioclase as the chadocryst phase (Fig. F10). Igneous laminations were observed in several samples but are generally scarce or may be overprinted by crystal-plastic fabrics in deformed parts of the core. Preliminary bulk rock geochemical results show a wide range in the chemistry of gabbroic rocks with Mg# varying from ~0.80-0.23, Fe2O3 from ~3.5-24.0 wt%, P2O5 from ~0.01-4.1 wt%, Y from 7-192 ppm, Nb from 1-10 ppm, and Cr from 1-1066 ppm.
Alteration of the primary igneous mineralogy in the core is generally low, but varies on the scale of a thin section to meters. Alteration of olivine to chlorite, tremolite-actinolite, and talc is the most common manifestation of alteration, whereas plagioclase and clinopyroxene tend to be less altered. It is common for clinopyroxene, where altered, to be partially replaced by patchy brown amphibole, but such alteration generally does not exceed 1%-2%. A portion of this brown amphibole is likely to be of magmatic origin. Where alteration is extensive, clinopyroxene is replaced by assemblages of actinolite and chlorite. Plagioclase is generally not strongly affected by alteration.
Actinolite and chlorite are also the most common vein assemblages. Scarce high-temperature brown amphibole and low-temperature smectite and carbonate veins have also been observed.
The structure of the core is complex, and structural styles range from brittle to ductile. Most of the gabbroic samples cored possess igneous textures, but there are several parts of the core that display crystal-plastic fabrics. Mylonitic zones characterized by high oxide-mineral content were observed at ~53 and 71 mbsf (Fig. F11). Coarser grained centimeter- to decimeter-thick zones of ductile shear are present in the upper 90 m of core, whereas thicker zones of ductile deformation with weak to strong crystal-plastic fabrics become more prevalent at depths in excess of 90 mbsf. Intervals of penetrative ductile deformation in the lower portion of the core locally exceed 2 m in thickness. Zones of ductile deformation are commonly oxide rich, as are the contact regions between undeformed and ductilely deformed rocks. Oxide gabbro-rich zones appear to be strain localizers because many, although not all, of the crystal-plastic shear zones are rich in oxide minerals. Inclination of the ductile foliations on the core face vary from ~18º to 75º in the cored intervals and averages ~30º-35º. Thin sections show a range of textures from strictly igneous to slightly deformed igneous to dynamically recrystallized metamorphic textures with crystal-plastic fabrics. As deformation intensity increases, the effect can be most easily observed in plagioclase, where a progression from strain-free plagioclase to plagioclase with deformation twins, undulose extinction, kink bands, and dynamic recrystallization to neoblasts along grain margins progresses to porphyroclastic textures with small neoblasts of plagioclase and highly strained and kinked plagioclase, pyroxene, or olivine porphyroclasts. Olivine appears to have fully recrystallized to neoblast grain sizes prior to pyroxene, which tends to be preserved as the dominant large porphyroclastic phase unless the intensity of deformation is very severe. Brittle fractures are generally filled with vein material such as actinolite and chlorite, but no large faults zones were recovered in the core. There were several regions of low recovery that could correspond to fault zones based on temperature, sonic, resistivity, and porosity logs. These regions of poor recovery generally sampled little intact core, although gabbroic rocks that were recovered were altered to smectite and contained carbonate veins.
Preliminary analysis of the downhole geophysical measurements from core and logging data yields a wide variety of information. Magnetic data indicate that the core possesses a single coherent magnetic direction with an average inclination of ~67º. This is compared with an inclination of -52º expected for the site. As in Hole 735B, these results indicate a consistently reversed polarity for the section and may indicate a significant block rotation of the massif similar in magnitude to rotations interpreted from Hole 735B (15º-20º). The consistency of the magnetic inclination downhole suggests that any relative rotations along ductile shear zone in the section must have occurred before cooling below the blocking temperature and are necessarily high temperature in nature. Magnetic susceptibility measurements clearly define zones of oxide gabbro and oxide-bearing gabbro documented in the core. Likewise, it provides a direct downhole comparison for the FMS logs, which measure resistivity. Oxide-rich zones are conductive whereas oxide-free zones have high resistivity. Magnetic intensity on split cores ranges from ~0.2-5 A/m and provides a significant signal range.
Lastly, an SWD experiment was conducted at Hole 1105A using two USGS OBS. These data, together with accelerometer data from the drill rig, will be employed to test the feasibility of SWD during drilling operations of the JOIDES Resolution.
