Previous   |   Next

doi:10.2204/iodp.proc.324.105.2010

Igneous petrology

Stratigraphic unit description

Hole U1348A was drilled on the north flank of Tamu Massif of Shatsky Rise (Fig. F1). The top of the drilled succession (Figs. F24, F25) is a Cenozoic–Cretaceous chert-chalk series (stratigraphic Unit I) and poorly recovered bioclastic sandstones partially cemented with a mixture of calcite and silica (stratigraphic Unit II). The bioclastic sandstones contain a small proportion of volcaniclastic material (e.g., basaltic fragments and individual feldspar crystals) and are intercalated with two well-developed clay horizons, most likely representing two tuffaceous ash horizons (see "Sedimentology" for details).

The "seismic basement" was reached at the top of stratigraphic Unit III (~200 mbsf) in Section 324-U1348A-14R-1, after which drilling continued for ~120 m into a succession of highly altered hyaloclastic material (stratigraphic Units III–VI). The hyaloclastic sequence has been further divided into four units based on macroscopic core descriptions, including differences in the modal clast size, degree of matrix or clast support, occurrence or absence of bedding features, and proportions of constituent components. The primary features of the hyaloclastite sequence, however, have been profoundly altered. From a purely volcanological perspective, the description of Units III–VI presents a challenge; in many instances, a volcanological interpretation is only possible through detailed microscopic examination. The following sections provide a general description of the hyaloclastite Units III–VI and their volcanic constituents. This overview should be considered in close association with descriptions given in "Sedimentology," which in addition contains comments regarding the sedimentology, primary origin, and reworking of the hyaloclastic material. Physical descriptions of the rocks and identification of the cementing minerals were assisted by XRD determinations.

Layered granular hyaloclastite (Unit III)

In general, this is a granular hyaloclastite succession mostly containing moderately sorted clast-supported granules but ranging up to matrix-supported breccias (e.g., interval 324-U1348A-17R-1, 20–60 cm) containing ~40% vitrous basaltic fragments. The vitric fragments retain pseudomorphs of acicular plagioclase, a low abundance of small vesicles, and a spongy texture conferred by the development of alteration spheroids in the original glass (for examples see Thin Section 157; Sample 324-U1348A-14R-1, 18–20 cm, and intervals 17R-3, 52–54 cm, and 17R-3, 93–95 cm). Grain size variations reveal reverse grading, passing upward from poorly sorted hyaloclastite siltstone and sandstone into occasional thin layers of granular hyaloclastite (Fig. F26). This may represent direct deposition of volcanogenic material into water, with larger vesicular hyaloclasts (lapilli size) (Fig. F26C, F26D) remaining more buoyant and falling through the water column more slowly than the smaller, denser fragmental material. Finer layers consist of glass shards (cuspate vitric fragments) with a significant clay matrix produced by alteration (Fig. F26E). All these fragments are now entirely replaced by secondary palagonite and cemented together by zeolite and calcite. Some finer horizons preserve small-scale cross lamination and ripple structures, indicating wave or current reworking (e.g., in Subunit IIIc; Thin Section 161; Sample 324-U1348A-15R-1, 55–57 cm). Determination of the primary and replacement mineralogy of different clast types contained within the coarser layers of this unit proved extremely difficult, both from core observation and petrographically.

Angular clasts as long as 3–5 cm occur in Cores 324-U1348A-14R (Subunit IIIa) and 15R through 17R (Subunit IIIc), and a number of these clasts are clearly volcanogenic and basaltic, containing acicular plagioclase crystals, chilled margin zones, and vesicular textures. Examples of these glassy clasts were analyzed by ICP-AES and confirm an original basaltic composition similar to that determined for basalts from Hole U1347A on Tamu Massif. Also present are fragments of very fine grained, highly altered "honeycomb" spheroid-rich material, for which immobile trace element concentrations also confirm a parental basaltic composition (see "Geochemistry").

Fossil-bearing hyaloclastite sandstone (Unit IV)

In terms of grain size, this stratigraphic unit is a volcanogenic sandstone succession containing horizons of much larger (pebble size) clasts. Sedimentary laminations are preserved in some intervals, but elsewhere these are completely destroyed by bioturbation. Profoundly altered dark gray and brown weathered angular clasts of vesicular basalt are scattered throughout this lithology. Honeycomb spheroid-rich fragments are numerous and represent reworked and profoundly altered volcanogenic vitric material. In addition, partially replaced cuspate glass shards, together with fragments of broken feldspar, can still be recognized in matrix material (interval 324-U1348A-20R-4, 77–81 cm). Small shell fragments are common within this succession, including sparse examples of entire ammonites and fragments of large bivalves. These shell fragments likely represent part of an allochthonous component, possibly brought in by turbidity currents sourcing material from shallower water environments.

Layered granular hyaloclastite (Unit V)

Unit V is a layered granular hyaloclastite. Individual grains are completely replaced by zeolite and calcite but originally were either vitric clasts or glass shards. Some sections consist of moderately sorted clast-supported granules, probably lapilli, and contain a few fine-grained intercalations. Grain size variations reveal fining-upward packages interpreted as turbiditic reworking of an original volcaniclastic source (see "Sedimentology"). The layering in Unit V is inclined, which may represent core sections through original cross-bedding structures or else could represent overall dipping of this sequence at 5°–25° (see "Structural geology" and "Downhole logging" for more details).

Massive clast-supported hyaloclastite breccia (Unit VI)

Unit VI is a massive, granular-textured succession consisting predominantly of closely packed clasts, many larger than 2 cm. Individual smaller volcanogenic grains or glass shards are completely replaced by zeolite and calcite and occur in crushed layers of subspherical "shells" (<1 mm). These shells are unlikely to be primary because they are too regular in shape to represent altered vesicle fillings and, instead, most probably originated as alteration "spheroids" that developed from quench mineral spherulites in vitric shards during an earlier stage of alteration. Interestingly, in a ~26 cm interval in the top of Unit VI, these vitric clasts still contain ~50% of their original volcanic glass (Fig. F27), providing a unique record of the composition of these volcanic constituents before compaction and alteration (see "Petrology and igneous petrology" for a description of this interval). Postemplacement "ptygmatic veins" indicate differential compaction throughout the succession and are most accentuated in the coarser hyaloclastite layers.

Relation of unit division to downhole logging data

Downhole logging measurements (Fig. F28) indicate low resistivity throughout Hole U1348A. The lowest resistivity occurs throughout the sediments of stratigraphic Units I and II. Typically, higher values were recorded in the "seismic basement," with the highest values of the succession in Subunits IIIa and IIIc and Unit V and lower, intermediate values throughout Units IV and VI. Accordingly, variation in resistivity confirms the primary stratigraphic divisions based upon the core observations outlined above. A notable resistivity peak occurs in the top of Unit VI and most likely corresponds to the ~26 cm thick, calcite-cemented layer that also contains a high proportion of fresh volcanic glass (see above). Total NGR counts (Fig. F28) are low throughout Unit I and highest in Units III–IV. Isolated, narrow gamma ray peaks for Th correspond to layers in Unit II, possibly the altered ash horizons (Fig. F29). Gamma ray peaks for U most likely correspond to the fossil-rich sandstone layer of Subunit IIIb and two unrecovered interbeds in the top of Unit IV (~246 mbsf) and the middle of Unit V (~271 mbsf). Oscillating total gamma ray counts in both Units IV and V may correlate with the observed bedding and grain/clast-size variation.

Magnetic susceptibility is an order of magnitude lower than in basaltic basements of Holes U1347A and U1350A. Also, the log density is relatively low throughout, reflecting the high degree of alteration. Another reason is that the Hole U1348A materials are not hard rock but volcaniclastics (Fig. F30). Porosity is as high as 27%–55% throughout the succession but decreases systematically downhole in Unit VI as a result of compaction (see also "Physical properties").

Petrology and igneous petrology

Given the profound degree of alteration within the volcanogenic succession and the difficulty in recognizing primary igneous features in what apparently formed as a hyaloclastite, it is important to outline the process of alteration and transformation. Basaltic clasts can be recognized in all of the volcanogenic units from thin section, but with increasing uncertainty in the lower part of the succession (see description summaries in 324TS.XLS in LOGS in "Supplementary material"). Unit VI is particularly enigmatic. In this unit, the presence and recognition of clast material has been a source of debate; however, partially altered glass shards that retain many of their primary characteristics have been identified in a ~26 cm interval (Fig. F27). Other than this one example, the bulk of the succession presents itself as a zeolite- and calcite-cemented altered hyaloclastite mass containing variable quantities of palagonitized vitric and lithic remnants, with palagonite spheroids imparting a honeycomb texture.

In thin section, spherulites were obviously a ubiquitous quench feature in the volcanogenic fragments; these have been subsequently converted to alteration spheroids, in particular throughout Units V and VI. The largest of the resulting spheroids are of the order of ~1–2 mm in diameter. Some spheroids are open and others are clay-lined, whereas some are partially to completely filled with white calcareous or siliceous minerals. In hand specimen, the spheroids impart a honeycomb-like texture to the fragments. Individual elements of this texture are discernible as numerous, rounded, evenly distributed, and consistently sized particles that can be mistaken for infilled vesicles, constituents of a biological origin (e.g., infilled and/or recrystallized radiolarian infusoria), or, because of their clay-rich composition, particles of detrital origin. However, recognition of volcanic attributes in thin sections (e.g., basaltic textures, clasts with angular/irregular shapes, and pseudomorph "outlines" of glass shards) leads us to dismiss these interpretations. Accordingly, the sediments of Units III–VI are interpreted as an unusual variety of palagonitized (altered) hyaloclastite consisting almost entirely of profoundly altered basaltic volcanic glass. Within these hyaloclastites, the original fragmental or shard structures have been broken down into fine angular fragments ranging in size from micrometers to a few centimeters and always containing a high abundance of spheroids.

Examples of hyaloclastite formation

Hyaloclastites (Rittman, 1958) are glassy fragmental volcanic rocks. The presence of volcanic glass indicates extremely high cooling rates that are obtained only by contact with water, and, for significant quantities of glass to be formed, it requires that the lava either flowed into or erupted under water. Pillow lavas forming at spreading ridges experience extremely high cooling rates, yet never produce glassy margins thicker than 2–3 cm (Kirkpatrick, 1979). They therefore do not normally spall into thick piles of purely glassy material, which may be a function of eruption into deep water (>2500 mbsl) at pressures above the triple point of water. At shallower depths, the formation of water vapor facilitates rupture of the glassy chilled crust along fractures induced by contraction of the quenched pillow rim, and as a result hyaloclastites are generated more abundantly in water shallower than ~500 mbsl. It is thought that the great majority of hyaloclastites dredged from seamounts probably formed by this kind of spalling of pillow surfaces. Submarine "lava fountaining" is also known to produce glass shards called "limu o Pele," which were first described in Hawaii around the propagating margins of submarine lava flows (Clague et al., 2009).

Fragmentation of the surfaces of pillow lavas in shallow water often forms breccia deposits that are mixtures of glass shards and pillow fragments, called "aquagene tuffs" (Carlisle, 1963). However, the finer glass fractions of such breccias can become entrained in the water column and separate to form a cloud of suspended material that eventually is deposited on top of the broken pillow mass. Carlisle (1963) termed these "laminated aquagene tuffs" and, in British Columbia where he observed them, they can be several meters thick and up to hundreds of meters wide in outcrop. However, possibly the most dramatic example of this type of deposit is two beds of altered and cemented hyaloclastites, each several meters thick, that were drilled at ODP Site 1223 on the crest of the Hawaiian Arch (Stephen, Kasahara, Acton, et al., 2003). These hyaloclastites were derived from the great Nu'uanu landslide, which carried away almost half of the Ko'olau Volcano of the island of Oahu, Hawaii, ~2 m.y. ago (Moore and Clague, 2002). Presumably, the glass was generated by submarine eruption(s) when the whole side of the volcano fell away and the Ko'olau magmatic rift system became exposed to seawater. The glass shards were then entrained in turbidity currents that reached as far as 200 km from their source and deposited hyaloclastites that now are moderately to extensively altered to palagonite and cemented by clays and zeolite in a similar fashion to the alteration observed in Hole U1348A.

Background on palagonite formation in the sea

A classic work on the petrography of palagonite is that of Peacock (1926), who described altered Icelandic tuffs that formed in melted pools of water beneath glaciers. Palagonite is a combination of iron oxyhydroxides and clays replacing basaltic glass. Two types of palagonite were recognized by Peacock (1926): (1) gel-palagonite, which is orange in thin section, has no birefringence, and often forms rims around hyaloclastite glass shards; and (2) fibro-palagonite, similarly orange in color but in which the constituent clay minerals are clearly discernible and which shows birefringence under the microscope. Fibro-palagonite typically forms distinctive rims around glass or gel-palagonite. It also projects radially around tiny spheroidal objects that are lined or filled with clays and develop during palagonitization of glass. Peacock (1926) called these "spherulites," but in modern usage this term is now most commonly applied to fibrous radiating crystal masses of plagioclase and clinopyroxene produced at pillow margins during quenching of lava. In this volume, to avoid confusion, we use the term "spheroid" in place of Peacock's spherulite to denote these alteration products. Both gel- and fibro-palagonite are developed in the tuffs of Site 1223, which could well be considered the prototype for the altered rocks recovered at Site U1348.

Nevertheless, palagonitization in rocks of Hole U1348A is far more extensive than observed at Site 1223 because prolonged alteration has significantly modified even the palagonite itself. The possible origin of this transformation will be discussed later, but the transformation to clays and the formation of both open spheroids and zeolites involves both uptake of water and an increase in porosity of every glass shard so affected. The origin of the alteration spheroids is obscure, and they may readily be confused with clay-lined and zeolite-filled vesicles that originally may have been present in the shards. Quite often, however, the alteration spheroids are so densely packed in individual grains that they would imply an improbably high vesicularity in the original lava fragment. However, in thin section, alteration spheroids are more readily distinguished because they often touch, are almost always nearly circular in plan, and never overlap or coalesce. By contrast, vesicle textures often reveal flattened, distorted, or coalesced forms.

Process of volcanic glass shard transformation

The proposed sequence of transformations that generates the spheroids from the volcanic glass shards (and basaltic clasts) is complex but may be summarized as shown in Figure F31. The original volcanic glass is first replaced through a process of palagonitization, which progresses from the outside inward, and along existing fractures, producing a rim of orange or brownish orange clay materials around the glass shard (or volcanic fragment). During this process, spheroids of clay materials begin to form in radial aggregates within the palagonitized regions of the shard, thus conferring a "cellular structure" within. As noted, these incipient spheroids can be easily confused with vesicles but are entirely secondary in origin. By this stage, the alteration has produced a series of clay boundaries both around the shard and around the spheroids inside. Further alteration entirely replaces all of the remaining primary glass as zeolite (and calcite). The material remaining consists of the pseudomorphed outlines of the original glass fragments, preserved as thin clay rinds set within a zeolite matrix. In addition, within these pseudomorphs, a honeycomb texture is produced from the continued development and formation of spheroid margins that also are preserved in clay. By this stage, virtually all of the primary glass shard mineralogy and physical structure are obliterated. The next possible stage in this evolution is caused by compaction through increasing overburden and/or through mechanical weakening caused by the mineral replacement processes described above. Compaction crushes the rims and distorts the shape of the spheroids. Continuation of this process results in a fine-grained material consisting almost entirely of curved fragments with clay rims and spheroids set in a zeolite/calcite matrix. At this stage most characteristics that might identify the original volcanic parentage of the rock have been entirely erased.

A series of photomicrographs reveals this transformation process in detail, beginning with the example of unaltered glass cemented by calcite in the top of Unit VI (Figs. F27, F32). The well-preserved glass fragments range in size from ash- to lapilli-size angular particles and contain crystallites of primary igneous silicate minerals (i.e., plagioclase needles) and occasionally vesicles. Spheroids are well developed in only one fragment, at the lower left of Figure F32A, which is a region nearly entirely replaced by matrix cement. These spheroids are defined by rims of clay, some touching but never merging. The once contiguous glass shards in the upper part of Figure F32A, form a fragmented arrangement separated by alteration products. The cementing matrix of calcite forms about half of the view in Figure F32A, and about a third in Figure F32B. Many fragments are defined by curving outlines clearly related to the tendency of glass to break along conchoidal fractures and are independent of the original placement of vesicles. The contiguous nature of adjacent fragments indicates that glass fragmentation in this sample occurred in situ, after deposition, with little or no subsequent movement of the depositional material. Moreover, cementation with calcite must have occurred soon afterward and acted to seal the freshness into this unique portion of the rock.

Figure F33 is a composite of photomicrographs covering the entirety of the large fragment at the top of Figure F32B. Here, the upper edge of the fragment and a roundish patch toward its middle are partially transformed to a darker brown palagonite, and fibro-palagonite rims surround two oblate vesicles on the left. The glass is broken into segments ~0.5–1 mm long by cracks, some of which were avenues for alteration fluids to penetrate the interior. Incipient round spheroids containing clays but retaining the pale brown color of the glass occur at the lower right. Details of the cracking pattern in the vitric clast are shown in Figure F34A. Alteration to fibrous clays is enhanced along some cracks but not along others; the development of fibropalagonite spheroids is clearly controlled by the crack porosity structure. Although the glass contains primary plagioclase and clinopyroxene (crystals clustered on the right in Fig. F34B) and even fresh olivine (on the left), there is no spherulitic crystallinity of the type that forms near the margins of rapidly cooled pillow lava. Plagioclase-clinopyroxene intergrowths also occur (Fig. F34C). The spheroids in Figure F34D and F34E have almost the same color as the basaltic glass; their rims are composed of radiating aggregates of tiny clay minerals, tiny acicular plagioclase, and feathery dendritic clinopyroxene.

Figure F34D and F34E show how spheroids form during incipient alteration of basaltic glass along microscopic fractures in the glass and that they are not vesicles filled with clays. In all the rest of the palagonitized glass of Hole U1348A, this process proceeded to an advanced degree so that very little fresh glass remains.

Most of the rocks, whether cross-bedded or clast-supported, originally must have consisted of abundant angular fragments of altered glass that now contain, or are overprinted by, abundant alteration spheroids (Fig. F35A), and later by zeolite and calcite. Many samples contain pseudomorphs of very irregular fragments that originally may have been fragments of glassy bubble walls. These are now rimmed with fibropalagonite and have interiors replaced by brownish palagonite (most probably gel-palagonite) and are completely transformed to dark brown clays (Fig. F35B). In the case shown, the shard is partially encased by secondary rhombic calcite. Other fragments have stretched or riblike internal structures (Fig. F35C), evidently a consequence of stretching while still plastic during their original formation; these also retain their well-defined rims of fibropalagonite and have interiors completely replaced by dark brown clays. In some fragments, the density of alteration spheroids approaches 100% (Fig. F35D). However, original angularity and interior vesicle structures of fragments are often clearly retained by the former fibro-palagonite rims (Fig. F35E), even though the interiors are almost completely replaced by cementing calcite. In this case, all fragments are cemented by zeolite. If this arrangement reflects an original packing condition, then the original bed form was clearly both very porous and loosely packed.

Figure F36 shows details of the cementation of the vitric clasts. Figure F36A depicts an example of an angular, originally vesicular, fragment retaining its fibrous rim but with an interior completely replaced by cement. Figure F36B reveals calcite replacement of the interior of a fragment surrounded by almost nonbirefringent zeolite. Figure F36C shows a fibrous calcite infilling surrounded by clays. Figure F36D and F36E shows palagonite fragments now altered to green clay, and the whole appearing to be matrix-supported by cementing zeolite. In the cement, two or more zeolite types may be present; the most abundant one forms needlelike radiating arrays around fragments with preserved rims of fibro-palagonite (Fig. F36F). The other zeolite morphology is blocky and irregular even within spheroids and filled-in vesicles (Fig. F36G).

Preliminary assessment

Interpretation of the ~120 m of volcaniclastic rocks in Hole U1348A on Tamu Massif presents an array of scientific challenges. Even though the volcanogenic origin of Units III–VI is not in dispute, initial interpretation as the cores arrived on deck was difficult. The main issues were whether the deposits were derived from proximal or distal sources, whether they were wholly or only partly produced by submarine volcanic action or whether any of the material was pyroclastic (that is, traveled through air after explosive eruption), and, finally, whether they contain one or more varieties of lava. Based on the shipboard mineralogy of fresh glass discovered in the ~26 cm interval in Core 324-U1348A-23R at the top of Unit VI (Fig. F27), we conclude that the protolith for the altered hyaloclastites was submarine volcanic glass saturated in olivine, plagioclase, and clinopyroxene. ICP-AES analyses on some rare basaltic clasts (see "Geochemistry") confirm that compositionally the protolith was no different from basalts drilled at other sites on Tamu Massif, especially the fresh glassy basalts of Hole U1347A.

Accepting this origin, then all igneous material in the cores was once rapidly quenched basaltic volcanic glass. The origin of this basaltic glass is a matter for debate; no evidence was found that unequivocally attributes the formation of Hole U1348A volcaniclastic rocks to the process of disaggregation of basaltic lava resulting from lava entering the sea from land, or to derivation from other types of subaerial explosive eruptions. However, the observation that the overwhelming percentage of volcanic material in the cores was originally clear, brown, sharply angular, and relatively nonvesicular basaltic volcanic glass might indicate a submarine volcanic source. The fragmentation mechanism must have been spallation assisted by expansion of heated seawater in the form of vapor that expanded in cracks in the congealing magma (and not in the sparse vesicles), as the glass fracturing was mainly conchoidal. Also, the few fragmental pieces of originally holocrystalline basalt that are present could be sourced from the chilled margins of lava pillows (Fig. F37).

The great majority of these vitric clasts and glass shards have now been completely altered by a prolonged process of palagonization to spheroid-bearing hyaloclastite. The finer grained portions are well bedded, preserving attributes of particle sorting, grading, and scouring consistent with turbiditic deposition (Units III and IV). Given that their constituents were originally mainly vitric clasts, they can be termed laminated hyaloclastites (Carlisle, 1963) and may, therefore, indicate transport in suspensions that moved rapidly across the seafloor near the base of the local water column. Other, more coarsely bedded material was found toward the base of the hole (Units V and VI) and was almost certainly deposited closer to the putative submarine volcanic source(s).

Top of page   |   Previous   |   Next