IODP

doi:10.2204/iodp.sp.345.2012

Scientific objectives

The principal objective for drilling at the Hess Deep Rift is to test competing hypotheses of magmatic accretion and hydrothermal processes in the lower ocean crust formed at the fast-spreading EPR. These hypotheses make predictions that can only be tested with drill cores, including the presence or absence of systematic variations in mineral and bulk rock compositions, presence or absence of modally layered gabbro, and the extent and nature of hydrothermal alteration and deformation. Specific scientific questions that address these predictions are outlined below.

How is melt transported from the mantle through the lower crust?

Melts erupting at mid-ocean ridges are almost never saturated in all mantle phases at plausible segregation depths (O’Hara, 1968; Stolper and Walker, 1980). The upper mantle and crustal processes responsible for the evolution of mantle primary melts into primitive MORB are the subject of a great many studies in the geochemical literature (e.g., Kelemen et al., 1997). Melts are transported and aggregated by porous flow at both mantle and crustal levels; the latter is a process that may be identified on the basis of textural and chemical evidence. The different mechanisms of igneous differentiation (e.g., fractional crystallization, equilibrium crystallization, assimilation, and porous reactive flow) strongly influence the chemical evolution of residual liquids and host cumulates. Melt transport through the lower crust is an important boundary condition of the crustal accretion models. Is melt transported largely via porous flow through the lower crust, or is it transported in dikelike brittle fractures (Kelemen and Aharanov, 1998)? Mineral and bulk chemical data for the core will provide tests for both the mechanisms of igneous differentiation and melt transport in the base of the ocean crust.

What is the origin and significance of layering?

Modally and compositionally layered gabbroic rocks are common in the lower crustal sections of ophiolites (e.g., Anonymous, 1972). A layered lower crust is thus one of the key and nearly ubiquitous features of all models of the fast-spreading lower crust. However, modal layering of the sort observed in major ophiolites has rarely been observed or sampled on the ocean floor so far. This may be because commonly used sampling and observation methods are not ideal to detect such layering readily, the right areas have not been sampled, or the layering is absent. If ophiolites indeed represent sections of normal mid-ocean-ridge crust, we expect to drill significant thicknesses of layered igneous rocks in this expedition. Their nature and extent is likely to have a strong bearing on the applicability of ophiolite-based accretionary models for the formation of the lower ocean crust.

How, and how fast, is heat extracted from the lower plutonic crust?

Hydrothermal fluids initially penetrate all levels of the plutonic crust along microfracture networks at high temperatures, with fractures sealing at 600°–800°C (Alt et al., 2010; Manning et al., 1996; Coogan et al., 2002a). Initiation of incipient cracking in the upper gabbros at Hess Deep overlaps the solidus temperatures of the most evolved lithologies, as recorded in magmatic amphibole (850°–925°C) (Gillis et al., 2003) and zircon (690°–790°C) (Coogan and Hinton, 2006). Whether this is the case in the lower plutonic crust, where more primitive lithologies dominate, is not known. Penetration of fluids at high enough temperatures could promote hydrous partial melting (Koepke et al., 2007); if and at what depth this process occurs is not known. Cooling rates for upper gabbro sections from fast-spreading crust and equivalent sections from the Oman ophiolite are rapid (4000°–5000°C/m.y.) (Coogan et al., 2002b, 2007a), indicative of significant convective cooling. Slower cooling rates in the Oman ophiolite suggest that heat flow was largely conductive (Coogan et al., 2002b, 2007a), cooling rates for lower gabbro sections from fast-spreading crust are not known.

Coring will allow us to determine the thermal history of the lower plutonic rocks using recently developed geospeedometric and thermometric methods. Temperature-time paths may be constructed by dating minerals with different thermal histories (e.g., John et al., 2004) and cooling rates may be calculated using Ca-in-olivine, Li-in-plagioclase, and other developing geospeedometers that quantify the diffusive exchange of elements between minerals (e.g., Coogan et al., 2005, 2007a). Using these and other approaches, it will be possible to determine the timing and rate of hydrothermal cooling in the lower plutonics, addressing questions such as, What is the rate of cooling with depth? Is hydrothermal flow along microfracture networks an effective mechanism to cool the lower crust? When is hydrothermal cooling initiated? Does fluid penetration occur at high enough temperatures to induce hydrous partial melting? Does the lower crust cool largely by conductive or convective heat transport?

What are the fluid and geochemical fluxes in the EPR lower plutonic crust?

Our understanding of the extent of fluid flow and reaction in the lower crust is presently very limited. Thermal models developed to test the crustal accretion models predict that advective cooling of the lower plutonic crust at or very close to the EPR would lead to a progressive decrease in the fluid flux and attendant fluid-rock interaction with depth, whereas more gradational conductive cooling over a broader time frame would likely lead to lower fluid fluxes and more limited fluid-rock interaction. Bulk rock Sr isotope data have constrained the time-integrated fluid fluxes for the upper crust (Bickle and Teagle, 1992; Gillis et al., 2005; Teagle et al., 2003; Kirchner and Gillis, in press); application of this approach will allow us to constrain fluid fluxes in the lower crust. Thermodynamic modeling predicts that high-temperature fluid flow and reaction would leave little macroscopic evidence (McCollum and Shock, 1998), which is supported by δ18O data for petrologically fresh samples (Alt and Bach, 2006; Gao et al., 2006). Thus, it will be critical to combine petrological and geochemical data to assess the extent of fluid-rock interaction in the lowermost plutonic crust.

Well-established petrological and geochemical techniques may be used to characterize the extent, nature, and timing of chemical exchange between the lower plutonic crust and seawater. Stable isotope compositions of minerals and geothermometers may be used to determine the temperature of hydrothermal replacement (e.g., Alt et al., 2010; Früh-Green et al., 1996; Manning et al., 1996). Major and trace elements and stable and radiogenic isotope compositions of minerals and whole rocks will provide constraints on the chemical evolution of fluid compositions and their origin (e.g., seawater-derived versus magmatic fluids) (e.g., Gregory and Taylor, 1981; Teagle et al., 1998; Gillis et al., 2003). Using these and other approaches, it will be possible to address questions, such as, How does the extent of alteration vary with depth? How does fluid flux vary with depth? What is the extent of chemical exchange between the lower crust and seawater? At what temperature is fluid-rock interaction initiated? What is the role of magmatic fluids?

In order to address the questions listed above, it will be important to distinguish between structures and hydrothermal alteration that formed at and near the EPR versus those associated with Cocos-Nazca rifting. This may be achieved by using standard petrographic techniques, in combination with Formation MicroScanner and structural data, as was done successfully on cores from Hole 894G (e.g., MacLeod et al., 1996b).

Drilling strategy and operations plan

The highest priority for drilling at Hess Deep will be to sample one or more 100 to Š250 m long sections of primitive gabbroic rocks. Three primary drill sites have been identified; however, if coring is proceeding well in the first or second of these sites, it will be continued as long as possible in order to capitalize on good drilling conditions and obtain the longest possible continuous sample.

The primary drill sites (proposed Sites HD-01B, HD-02B, and HD-03B) are situated on a flat-lying, east–west-trending, sedimented bench at ~4850 mbsl along the southern slope between the intrarift ridge and Hess Deep, in an area dominated by primitive gabbros (Figs. F5, F8). A ~200 m wide, flat-lying bench is covered with ~15 m of pelagic sediment mixed with lithic debris. The slope north of the bench is the footwall of a steeply dipping, southward-facing normal fault. The bench itself has a series of 5–15 m high, northward-striking narrow ridges that are attributed to the combined effects of variation in sediment thickness, sediment draping over small-scale basement structures, and/or relief caused by westward-dipping, northward-striking normal faults. The three drill sites are carefully located to be as far removed from these surface features as possible and to maximize the lengths of unfaulted sections that we will drill. The drill holes should encounter contiguous vertical sections on the order of 90–290 m in the footwall (Fig. F5). Our drill sites are spaced from 300 to 400 m apart along the bench to maximize coverage of the area where primitive gabbros (Mg# = 75–85) crop out within 150–200 m to the north and south (see distribution of olivine gabbros in Figs. F5, F8). As described above, an east–west transect is not intended, rather drilling will proceed at one or more sites as long as possible. In case drilling conditions at these sites prove impossible, we are requesting permission to relocate sites anywhere on the bench. These strategies are intended to maximize our chances of meeting our principal scientific objectives, which are to determine how melt is transported from the mantle through the crust, where melts fractionate and crystallize, and the extent and nature of hydrothermal alteration and deformation.

Alternate proposed Site HD-04B is located a few hundred meters north of Hole 894G, at the summit of the western end of the intrarift ridge (Figs. F4, F6). The summit is covered by <10 m of flat-lying pelagic ooze mixed with lithic debris (Gillis, Mével, Allan, et al., 1993). At alternate proposed Site HD-04B, a borehole would start in evolved upper gabbros but there is potential of recovering more primitive lithologies at several hundred meters depth because the gabbroic rocks become more mafic along the slope of the intrarift ridge immediately to the south.

Operations plan

Although we present an example operations plan in Table T1, the single primary goal of our expedition is to establish a hole that can be reentered multiple times, allow formation stability to be maintained, and core/log that hole as deeply as possible. When this occurs, all remaining available time may be spent on this hole and other holes/sites will not be established.

We anticipate that drilling, coring, and logging operations in the Hess Deep Rift may be quite challenging. We are preparing a range of potential operational approaches that we might apply to address challenges that include very thin sediment cover, initiating/maintaining a hole through the basement contact, and coring deeply through potentially unstable basement.

Given all this, it is unlikely that the expedition operations will unfold as presented in Table T1.

Operational approach(es)

To maximize the potential to achieve the operational and scientific objectives of initiating and advancing a deep hole, we plan to provide the hardware and supplies to implement a variety of operational approaches. Maintaining maximum operational flexibility will provide the highest chance of success in this challenging environment. There are many combinations and permutations of operational scenarios possible which will continue to be evaluated leading up to the expedition and will all be under consideration as the expedition progresses. Sediment thickness, nature of upper basement drilling/coring conditions, and hole stability will determine which approach may be the most viable. The operational scenarios can basically be categorized into four general approaches depending upon the actual conditions encountered. At each site, a series of pilot holes will be drilled/cored to verify formation conditions.

Standard reentry cone approach

Deployment of a standard reentry cone is the most favored operational approach. This approach is shown in Figure F9 and Table T1. If successful, this approach allows the best chance of success at achieving the deeper objectives because it allows the best chance to combat anticipated unstable hole conditions and provides a possible opportunity to case off upper basement formation if required. This approach starts with jetting in a standard reentry cone with a short section of 20 inch casing. Then an 18.5 inch hole is drilled into uppermost basement into which 16 inch casing is installed and cemented. Rotary core barrel (RCB) coring and logging would extend below the base of the casing. If necessary to stabilize the hole, a third casing string (10.75 inch) may also be deployed. The thinnest sediment cover in which a standard reentry cone has been successfully deployed is ~38 m during Expedition 336. Attempting to do this in less sediment, as expected at the drill sites, may be problematic (see below). One concern is possible undermining of the reentry cone base plate from drilling circulation while drilling the 18.5 inch hole into basement.

Free-fall funnel approach

Another possible operational option is to deploy a free-fall funnel (FFF) or “nested” FFFs in an attempt to facilitate multiple reentries of an existing hole. There is a spectrum of possible operational options that entail using FFF or “modified” FFF technology. A couple of potential options are shown in Figure F10. This approach may take a bit less operational time to deploy than a standard reentry cone; however, it does not provide an option to deploy casing to stabilize any unstable basement below nor does it provide as robust a seafloor structure to remediate an unstable hole or problematic drilling conditions. Historically there can be problems with FFFs being pulled out of the hole when pulling the core bit out of the hole. FFFs also can occasionally interfere with reentering the basement portion of the hole because of washing of the sediment overlying the basement contact and offsetting of the bottom of the FFF casing from the hole in basement. One potential way to address this would be to cement the FFF into the basement as shown in Figure F10.

Bare rock reentry approach

This operation would be to rely on the ability make multiple reentries into small diameter (10–20 inch) bare rock holes without installation of a seafloor structure. This is shown in Figure F11. Once again, this approach has many potential operational combinations and permutations. This approach is likely to be the quickest at getting to RCB coring and could be attempted at any of the pilot holes. Although bare rock reentries have been successful in the past, this approach will likely require the most amount of time for subsequent reentries; is more impacted by changes in environment such as variable current, changing sea states, and so on; and is the least capable of supporting remedial actions caused by unstable hole problems. A larger diameter hole near the surface to facilitate reentry and potentially accommodate subsequent seafloor structure installation could also be made with an underreamer or bi-center reamer positioned above the main bit (e.g., create a countersunk hole; Figure F11, scenario 3); however, its likely that no coring can be undertaken when running these reamers.

Bare rock installation of a reentry cone with short casing pup

This scenario would allow a cased hole on bare or thinly sedimented rock (<10 m). After a large-diameter hole is drilled a few meters into basement, a short section of casing with casing hanger and reentry cone is lowered into the hole. The hole then would be able to be continued deeper with casing as described above (see “Standard reentry cone approach”). However, lowering such a seafloor structure into the hole is probably not possible with the standard camera system, which cannot see around and below the reentry cone. This likely would require a through-pipe camera, which is not part of the normal operational equipment available.

Tentative planned sequence of operations

  1. Locate site per coordinates provided, deploy camera/sonar system, and conduct limited seafloor survey to confirm acceptability of proposed drilling area.

  2. Conduct a series of probes into the seafloor to basement at multiple locations to determine approximate variation in sediment thickness and estimate slope of seafloor and basement contact. This can be done by “jetting” the drill string through the sediment until the basement contact is reached. This should be a fairly rapid operation given the drill string only has to be pulled clear of the seafloor between probes and the camera can remain deployed.

  3. Based upon the camera survey, probe information, and other available data, identify a location to attempt drilling a pilot hole.

  4. Spud and deepen a pilot hole 50–100 m into basement using the RCB. Ultimate depth of the pilot hole will be determined by hole stability, core recovery, and science evaluation of the core data. Note that multiple pilot holes may be required before a suitable drilling location is identified.

  5. Based on pilot hole drilling conditions, three options are suggested:

    1. If the pilot hole is going well and there is reasonable belief that the hole can be successfully deepened and valuable core data obtained, then the coring operation should be terminated based upon a “conservative” estimate of rotating bit hours (40–45 h) and the drill string round-tripped for a bit change. A FFF can be deployed prior to clearing the seafloor, or plans can be made to reenter the bare rock hole with a second RCB core bit.

    2. If the pilot hole determines that this location is not suitable for attempting to achieve the scientific objectives, then the area can be abandoned and another location selected for evaluation.

    3. If the pilot hole is not considered suitable for deepening but the location is considered adequate for an attempted deeper “reentry” hole, then the operational approaches identified above can be discussed and an operational approach decided upon using the hardware and personnel resources available onboard.

Recent improvements in drilling unstable hard rock formations

Several improvements in drilling/coring fractured hard rock formations have been made recently. We now use “blended” cement made up of standard American Petroleum Institute (API) Class G cement blended with 0.25 lb of Cello Flake lost circulation material (LCM) per 94 lb sack of cement. We have had much better success cementing casing in fractured formations using this blended cement.

Standard API Class G cement has also recently been used successfully during Expedition 335 to cement problematic zones within a borehole, allowing further successful advancement of the hole.

Also, we have determined that significantly larger mud sweeps are required to achieve adequate hole cleaning in these environments. In the past, typical mud sweeps ranged from 15 to 25 bbl. Recently, during Expedition 327 (Juan de Fuca), mud sweeps in excess of 100 bbl were used with good success.

Finally, in the past there was concern that “working” a hole too much could in effect do more damage than good. Whereas this is always a possibility, we have found that in many cases tenacity and multiple wiper trips actually eventually paid dividends in cleaning up the hole and allowing successful advancement.

All of these improvements will be kept in mind as we tackle the challenges of successful drilling/coring operations in the Hess Deep environment.

Operational challenges

Drilling and coring operations are anticipated to be challenging during the Hess Deep expedition. Water depths in excess of 4400 m will impact routine operations such as pipe tripping, reentry, and wireline coring/logging, making these operations more time consuming than they would be in shallower water. Because of the anticipated limited sediment cover on location, the installation of a reentry structure may be more problematic than normal. Following installation of a reentry system, basement drilling/coring operations are expected to be difficult, with unstable hole conditions and difficulty performing effective hole cleaning. If a sustainable hole is achieved and deepened, core recovery in basement is typically fairly low and deployment of wireline logging tools may not be possible in unstable holes. Despite these challenges, the ultimate operational goal is to establish a hole capable of sustaining multiple reentries (allowing multiple bit changes), with “adequate” hole stability, deepen that hole as far as possible with continuous RCB coring, and recover a full suite of wireline logs.