IODP Proceedings    Volume contents     Search
iodp logo

doi:10.2204/iodp.proc.304305.103.2006

Metamorphic petrology

General observations

During Expeditions 304 and 305, two deep holes were drilled at Site U1309 (Atlantis Massif): Hole U1309B (to 101.8 mbsf) and the upper 401.3 m of Hole U1309D were drilled during Expedition 304, and Hole U1309D from 401.3 to 1415.5 mbsf was drilled during Expedition 305. In general, the mineral assemblages of the rocks recovered during both expeditions at Site U1309 record the cooling of mafic plutonic rocks from magmatic conditions to zeolite facies during the unroofing and uplift of Atlantis Massif. Individual samples generally display a range of superposed metamorphic conditions, but no single sample records the entire cooling history of the site. The assemblages encountered in any sample depend on at what point in its cooling history the rock underwent deformation or hydration. It is only by studying the assemblages recorded in a large number of samples from this site that the whole cooling history can be inferred.

Metamorphic history at Site U1309 can be summarized as follows:

  1. Granulite facies, probably near-solidus, mylonitic deformation, and recrystallization of plagioclase + clinopyroxene and brown amphibole.
  2. Amphibolite-facies replacement of pyroxene by green to brown hornblende in diabases, gabbros (especially oxide gabbros), and mylonite zones. The extent of this largely static event is hard to judge because of overprinting by greenschist-facies amphiboles and uncertainties over amphibole compositions in thin section.
  3. Widespread, largely static upper-greenschist-facies to lower-amphibolite-facies metamorphism manifested by the following:
      • Formation of secondary plagioclase and secondary amphibole and, below 384 mbsf, epidote growth that appears to be related to late magmatic leucocratic intrusions.
      • Replacement of pyroxene by actinolitic amphibole. This is the major effect of the greenschist event in most gabbroic rocks and diabases. In the upper 300 m of Hole U1309D, this alteration is pervasive and all samples are affected to a greater or lesser extent. At greater depths, the alteration is increasingly associated with the emplacement of amphibole-rich veins and accompanying halo alteration.
      • Development of tremolite-chlorite ± talc corona texture in all rocks containing both olivine and plagioclase. This may include some amphibolite-facies formation of cummingtonite and green hornblende. In the upper 300 m of Hole U1309D, this reaction went to completion in almost all samples, removing either olivine or plagioclase from the assemblage. At greater depths, the reaction commonly did not go to completion and is increasingly localized by amphibole veins and the margins of gabbroic dikelets. At shallower depths, most amphibole veins postdate corona formation.
  4. Mainly static lower-greenschist to subgreenschist metamorphism including the following:
      • Serpentinization of olivine in olivine gabbro, troctolite, and olivine-rich troctolite, with concomitant formation of prehnite and hydrogrossular in associated plagioclase. Above 300 mbsf, serpentinization is restricted to rocks where olivine was in excess over plagioclase and was therefore still present after the corona-forming reaction went to completion. At deeper levels, serpentine, prehnite, and hydrogrossular are often localized on closely spaced, variably oriented fractures (ladder veins).
      • Metasomatic talc-tremolite ± chlorite veins and irregular zones in ultramafic rocks, especially near contacts with mafic rock. These veins and zones overprint serpentine.
      • Talc-tremolite-chlorite schist with ultramafic protolith (upper 25 m of Holes U1309B and U1309D and fragments in Holes U1309E and U1309H). This is assumed to be the same age as the talc-tremolite-chlorite veins and zones deeper in the hole, although the relationship to serpentinization is not observed. It is inferred to reflect the main deformation on the detachment fault that forms the surface of the massif.
      • Sporadic talc-carbonate metasomatic alteration of olivine-rich rocks.
      • Relatively late emplacement of slip-fiber amphibole veins and associated local metasomatism.
      • Serpentinization of isolated grains in olivine gabbro and relict grains in coronas.
  5. Zeolite-facies metamorphism that includes replacement of plagioclase by zeolites throughout the core and emplacement of zeolite-bearing veins below 700 mbsf. Late, open, irregular fractures commonly contain a clay mineral that may be saponite, along with carbonate and (below 700 mbsf) zeolite minerals.

Hole U1309B

Three rock types with distinct chemical compositions were encountered in Hole U1309B. In order of decreasing abundance, these are

  1. Mafic igneous rocks,
  2. Peridotite, and
  3. Talc-tremolite-chlorite rocks of mainly ultramafic protolith.

Mafic igneous rocks

Mafic rocks varying from gabbro and troctolite through diabase to basalt form the major rock types within the core. All rocks evidence some degree of hydration, with the effects of hydration decreasing downhole. Traces of fresh augite appear in the gabbros at ~40 mbsf, and relics of fresh olivine appear in the troctolite at 82 mbsf. Assemblages indicative of conditions from amphibolite- to zeolite-facies hydration are found within the core (Fig. F112). Traces of brown pleochroic hornblende are found in some rocks (Fig. F112A), indicating that, in places, amphibolite-facies alteration preceded pervasive greenschist alteration. Some grains of brown hornblende may have formed during late-stage crystallization of fractionated gabbroic magma. Other grains are zoned to green margins or show metamorphic fabrics in ductile shear zones; these probably formed in response to postmagmatic hydration. Cummingtonite is found in a number of gabbro samples, where it probably formed by hydration of orthopyroxene or olivine. Cummingtonite is best recognized where it is adjacent to calcic amphibole because in these locations the difference in extinction angle between the two amphiboles is obvious (Fig. F112B). Although patches of albite with or without actinolite and chlorite are seen to replace plagioclase in a few samples, in most rocks the primary plagioclase appears to be fresh, even in breccias. Zeolite replaces feldspar locally (Fig. F112C) but was not detected in X-ray diffraction (XRD) spectra (Table T4) (e.g., Sample 304-U1309B-5R-2, 55–56 cm) and appears to be a minor phase. Saponite is present in some XRD spectra.

Although the metamorphism in Hole U1309B extends over a wide range of temperatures, the most abundant alteration is in greenschist facies. This is marked by the abundance of actinolite and tremolite in the mafic rocks. Chlorite and albite are present in many thin sections but are subordinate to amphibole. The amphibole shows a wide range of pleochroism, ranging from deep green actinolite in diabase and basalt to colorless tremolite in some gabbros. Because tremolite and actinolite have similar stability fields in mafic rocks, when we use the term “actinolite” in this report we refer to any presumed low-temperature amphibole, regardless of its pleochroism. We use the term “tremolite” to refer to colorless amphibole in ultramafic rocks and in tremolite-chlorite schists.

In most rocks, actinolite replaces pyroxene and olivine. In the basaltic rocks, fine-grained actinolite has replaced the matrix and, if it had been present, glass. In all rocks, the actinolite replacing pyroxene is texturally distinct from that replacing olivine. Single grains of actinolite may grow epitaxially on augite so that the host mineral and the pseudomorph share the same c-axis orientation. The amphibole produced by this process is coarse grained and can only be distinguished from the original augite by its small extinction angle. Actinolite pseudomorphs after pyroxene are less well developed in diabase, where the actinolite replacing pyroxene may form clusters rather than single grains (Fig F112D).

Amphiboles replacing olivine in gabbros and troctolites are texturally distinctive from those replacing pyroxene. In the plutonic rocks, the reaction between olivine and surrounding plagioclase produces a cluster of minerals with a distinct corona structure (Fig. F112E, F112F). In samples where olivine has been completely altered, the cluster has two zones. The outer rim consists of chlorite, which was probably formed from plagioclase. The core of the cluster, which probably marks the extent of the original olivine, contains a complex intergrowth of chlorite and amphibole. The amphibole appears to be mostly actinolite or tremolite, but the possible presence of cummingtonite in the core cannot be ruled out. In two samples (304-U1309B-15R-1, 110–112 cm, and U1309B-16R-2, 71–73 cm), fresh olivine remains and the cluster consists of three zones: a rim of chlorite, a zone of tremolite, and an inner zone of talc.

Amphibole replacing olivine in diabase is also texturally distinct. In diabases, olivine pseudomorphs form as rounded clusters of small, randomly oriented actinolite (rarely, cummingtonite is present) as well as chlorite (Fig. F112D). Commonly, the only chlorite found in an altered diabase is present in these structures. However, the distinct chlorite rims against plagioclase that are common in troctolitic rocks are rarely observed in the diabase.

The static greenschist (or possibly low-amphibolite)-facies alteration described above occurs to a greater or lesser degree in all the mafic rocks examined. Alteration involves influx of water but appears to be otherwise isochemical and predominantly affects pyroxene and olivine. Except where it was originally in contact with olivine, primary plagioclase is largely unaffected in many basalts and diabases and some gabbros. Also prominent in the core are zones of brecciation, cataclasis, or intense veining (Fig. F113) where the host rock, usually gabbro, is cut by yellow-green veins rich in calcic amphiboles. In one instance, the breccia contains a clast of chromite-bearing tremolite schist inferred to be of ultramafic origin (Fig. F113A; see “Talc-tremolite-chlorite rocks,” below). Breccia matrix varies from weakly (Fig. F113B, F113C) to intensely (Fig. F113E, F113F) foliated. In most breccias, the amphibole is a weakly pleochroic actinolite, but in some, brown amphibole is present (Fig. F113F). In some intensely deformed areas, the amphibole is a colorless tremolite, suggesting that the influx of Mg-enriched fluids accompanied deformation. In most instances, these “breccia veins” show halos of milky secondary plagioclase, turbid in thin section, containing abundant solid and fluid inclusions (Fig. F113D). Deeper in the hole the breccia veins become less common, but intense hydrothermal alteration is seen in Core 304-U1309B-14R in the vicinity of a late leucocratic magmatic dikelet (Fig. F114). Dark green patches as long as 8 cm are filled almost entirely with fine-grained actinolite and have milky margins as wide as 2 cm, in which turbid secondary plagioclase net-veined by actinolite is developed. One patch appears to contain clinopyroxene relics.

Serpentinized peridotite

Approximately 1.5 m of serpentinite was recovered from Core 304-U1309B-11R. The rock is variably hydrated, producing characteristic ribbon and mesh textures. In rocks where serpentinized pseudomorphs (i.e., bastites) indicate the presence of original orthopyroxene, both the pyroxene and the olivine are completely serpentinized. Locally, where the major pyroxene is diopside, substantial amounts of both olivine and diopside remain. Also present in the serpentinite are small amounts (up to a few percent) of chromite, magnetite, and magnesite. Primary chromite grains are invariably rimmed by magnetite. In many grains, magnetite also grows on fractures cutting the chromite grains. These textures indicate that chromite was out of equilibrium with the serpentinizing fluid. Magnetite forms both as rims on chromite and as trains of fine crystals in the serpentinite. The metaperidotite is cut by veins at all scales (Fig F115). Fine-grained carbonate, probably magnesite, is commonly associated with serpentine veins and also partially replaces serpentinite ribbon textures (Fig. F115E). Calcite (or aragonite) veins are also present locally, particularly at contacts with gabbro.

Talc-tremolite-chlorite rocks

Pale green tremolite ± talc ± chlorite rocks were recovered as small clasts sampled from just beneath the carbonate ooze in Section 304-U1309B-1R-3 (Pieces 2 and 3) and from a clast in basaltic breccia in Section 3R-1 (Fig. F113A). These samples are variably schistose with isotropic to brecciated and intensely foliated fabrics. Included in these samples is one piece that was marked by the bit and, therefore, was possibly drilled in situ. These rocks generally contain chromite grains and “cuspate textures” inferred to be relict orthopyroxene pseudomorphed by talc and magnetite and are inferred to have an ultramafic protolith. This is supported by shipboard geochemical analysis of Section 304-U1309B-1R-3 (Piece 3) (see “Geochemistry”). Some samples from Section 304-U1309B-1R-3 contain rutile, suggesting that they have mafic components. Similar rocks have been described from the detachment fault zone on the Southern Ridge of the massif (Schroeder and John, 2004) and from a detachment fault in the 15°45′N area of the MAR (Escartín et al., 2003). We infer that the top of the central dome of Atlantis Massif includes poorly recovered talc-tremolite schists from a detachment fault that caps the massif. This inference is supported by recovery of tremolite schist in a short core (Core 304-U1309H-1R) from the top 4 m of the central dome, 1.6 km to the east (see “Hole U1309H”).

Similar rocks are present in a steeply dipping sheared vein cutting peridotite in Section 304-U1309B-11R-2 (Fig. F115A, F115B) and give a clue to the geochemical origin of talc-tremolite rocks. This zone roots in a poorly recovered interval containing altered gabbroic fragments and tremolite schists. The zone tapers upward and has an outer talc-rich zone with an isotropic fabric replacing serpentinite (Fig. F115C, F115D) and an inner tremolite-rich zone with highly schistose talc bands. Dark grains within the talc alteration are bastites containing serpentine + magnetite assemblages. In contrast, several former orthopyroxene grains beyond the talc front are completely pseudomorphed by talc and magnetite, giving a pseudoschistose texture. We infer that these were pyroxenes that had not been serpentinized and, therefore, had Mg:Si ratios more suitable for talc replacement, whereas the bastites within the band were already serpentinized when the talc alteration occurred.

The tremolite-talc rocks are clearly metasomatic and require introduction of a fluid with higher Si/Mg and Ca/Mg activities than that in equilibrium with serpentine. The most likely local source for such a fluid is mafic layers or veins within the peridotite. Another interesting observation is the lack of magnetite in the talc rock. Reduction of Fe3+ to Fe2+ is required, suggesting reducing fluids (perhaps CH4 bearing). Apparently, alteration of serpentine to tremolite is reversible, because in Section 304-U1309B-11R-3, we find a tremolite vein that has been cut by serpentine veins (Fig. F115F, F115G). Adjacent to the serpentine veins, the tremolite vein has been altered to serpentine.

Metamorphic conditions and degrees of alteration

The main hydration event occurred in the greenschist facies, as indicated by the presence of actinolite in the gabbros and diabases throughout the core. The preservation of early brown hornblende in some gabbro and diabase samples indicates that, locally, hydration occurred at temperatures corresponding to amphibolite facies. Similarly, local zeolite replacing plagioclase and the presence of yellow to brown saponite veins cutting the gabbroic rocks indicates that, in places, hydration occurred under conditions at or below those of zeolite facies. Unlike the clay- and zeolite-bearing assemblages, which are present around isolated veins and fractures, the greenschist assemblages are pervasive. This indicates that the whole sequence of rocks was open to pervasive fluid flow (probably low flux) at temperatures ~350°–450°C but has been relatively impermeable since.

Although there is a tendency for increased survival of primary pyroxene and olivine downhole, there appears to be no systematic gradient in the intensity of alteration through Hole U1309B (Fig. F116). Figure F116A shows that the total degree of alteration depends largely on rock type. Where olivine is abundant (in the peridotite and the troctolites), the degree of alteration can reach 100% because of the reaction between olivine and plagioclase. Because primary plagioclase was stable during the greenschist hydration event and because plagioclase generally makes up 50% of the rocks, most olivine-poor gabbro and diabase intervals show no more than 50% alteration. Gabbros and diabases that show >50% alteration invariably have undergone deformation or intense veining. This is shown in Figure F116B, which estimates the degree of vein-related alteration (veins and halos). The most intense alteration is associated with zones of brecciation and veins in gabbro and diabase above 60 mbsf and in the serpentinite, which is cut by talc-tremolite veins. Below 60 mbsf, the extent of vein-related alteration is much less and is mainly associated with late magmatic leucocratic dikelets. Diabase intrusions away from zones of brecciation generally show little or no vein-related alteration, suggesting that they were intruded after the hydrothermal event that produced actinolite-rich veins.

Hole U1309D

Gabbroic rocks (gabbro, olivine gabbro, oxide gabbro, gabbronorite, and troctolite) constitute >90% of the rocks recovered from Hole U1309D. Several large diabase intrusions occur in the upper 130 m, but diabase is uncommon below this depth. Ultramafic and near ultramafic rocks (wehrlite, dunite, and olivine-rich troctolite) are present sporadically, with large concentrations between 300 and 350 mbsf and 1100 and 1225 mbsf and narrower occurrences near 175 and 700 mbsf. Talc-chlorite schists with an ultramafic protolith are restricted to the upper 25 m of Hole U1309D. Although restricted in abundance and occurrence, these rocks are important for the interpretation of the structural history of Site U1309.

Although some types of metamorphism in Hole U1309D are restricted to certain rock types (serpentinization, for example), by and large all rocks experienced the same range of metamorphic conditions. In particular, the dominance of vein-related alteration in the lower 1100 m of the core lends itself to description by metamorphic type as opposed to rock type. In contrast to Hole U1309B, the discussion of metamorphism in Hole U1309D is organized according to metamorphic type.

Hole U1309D was drilled ~20 m north of Hole U1309B. Not surprisingly, the alteration seen in the upper ~100 m of Hole U1309D is similar to that encountered in Hole U1309B. Alteration is intense with alteration >90% common. Greenschist-facies metamorphism in which calcic plagioclase is stable and where actinolite is the main metamorphic mineral predominates. The major difference between the holes is that the rocks in the upper portion of Hole U1309D record a high-temperature deformation and metamorphism event that is poorly represented in cores from Hole U1309B.

XRD analyses

Selected vein material and all samples chosen for whole-rock ICP-AES analysis were analyzed by XRD (Tables T4, T5, T6; Figs. F117, F118, F119). Diabase and basalt from the upper 400 m of the hole show actinolite, plagioclase, and, sometimes, chlorite as discernable peaks (the main peaks for clinopyroxene coincide with some amphibole peaks, so a separation is often difficult). In addition, a number of late veins and alteration patches were sampled and consist mainly of amphibole + clays (possibly saponite in most cases). A few more unusual spectra are shown in Figure F117, some of which remain to be identified. No zeolites were identified in any spectra from the upper 400 m of the hole.

The whole-rock spectra from the lower 1000 m of the core are dominated by plagioclase and clinopyroxene. Locally, olivine, serpentine, amphibole, and ilmenite are also found. Orthopyroxene was found in fewer samples than expected, probably because of peak overlap problems. The most significant finding from whole-rock XRD was the detection of brucite in one serpentinized olivine-rich troctolite (Fig. F118). Significant findings from XRD examination of veins include the identification of anhydrite in Sample 305-U1309D-150R-3, 22–23 cm (Fig. F119A), zeolites (analcime and thomsonite) (Fig. F119B, F119C), and a green, waxy, translucent clay mineral (possibly saponite).

Metamorphism

Granulite-facies metamorphism associated with deformation

This is the highest temperature metamorphism recognized in the core and is directly related to plastic deformation (see “Structural geology”). It is characterized by recrystallization of pyroxene and plagioclase (and, more rarely, brown hornblende) into an equigranular, mosaic texture (for example, Samples 304-U1309D-13R-1, 68–70 cm, 30R-1, 26–28 cm, 44R-4, 16–18 cm, and 305-U1309D-110R-3, 46–49 cm) (Fig. F120). Less common is the presence of narrow (~0.2 mm) undeformed zones of recrystallized pyroxene and plagioclase (see Sample 304-U1309D-50R-1, 52–56 cm). Granulite-facies shear zones are most common at igneous contacts in the upper 150 m in the core but do occur at greater depths.

Amphibolite-facies replacement of pyroxene by green and brown hornblende

Textural evidence from mafic rocks in Hole U1309D indicates that amphibolite-facies recrystallization accompanied ductile deformation, brittle deformation, and static hydration (Fig. F121A, F121B). Brown hornblende surrounded by brown hornblende neoblasts in the mylonitic portion of Sample 304-U1309D-8R-2, 26–28 cm, indicates that hornblende was stable in this rock during ductile deformation. In an entirely different texture, green-brown hornblende in Sample 304-U1309D-4R-2, 108–110 cm, fills fractures formed during brittle deformation. In an unusual sample (304-U1309D-16R-5, 1–4 cm), brown hornblende has replaced pyroxene in the matrix of a brecciated diabase and has also grown in the deformed zones between the clasts.

The replacement of pyroxene by green and/or brown hornblende postdates granulite-facies recrystallization, as some deformed pyroxenes in the mylonites are replaced by hornblende. Brown hornblende has been found growing as tails on deformed pyroxene (Fig F121B), as well as forming adjacent to narrow felsic dikelets. Brown hornblende is commonly associated with upper-amphibolite-facies and granulite-facies conditions. The extent of amphibolite-facies recrystallization is uncertain and may have been underestimated in shipboard studies. Many thin sections contain small amounts of brown or green-brown hornblende, often overprinted by much more abundant green amphiboles. Initial shore-based electron microprobe work shows that some of these green amphiboles are hornblende and others are actinolite.

Late magmatic leucocratic dikelets, breccia zones, and associated formation of secondary plagioclase, secondary amphibole, and epidote

A “bleaching” alteration in which plagioclase is converted to albite and pyroxene is converted to amphibole, commonly in the vicinity of late magmatic leucocratic dikelets (also referred to as “magmatic veins”), was first observed in Hole U1309B (Fig. F114) and is common below 140 mbsf in Hole U1309D (Fig. F122A, F122B). The largest and most voluminous individual veins in Hole U1309D are these magmatic intrusions, some of which have thicknesses of tens of centimeters. As such, they can locally dominate modal “vein” abundance. Several of the most conspicuous alteration zones in Hole U1309D are associated with late magmatic leucocratic dikelets. Most of the vein-related alteration between 60 and 400 mbsf is attributable to these veins, as are several prominent zones of bleaching farther downhole. Typically, the late magmatic leucocratic dikelets contain euhedral plagioclase, often zoned, surrounded by fine-grained actinolite mats or amorphous patches (Fig. F122A, F122B). Brown hornblende is generally present but is extensively replaced by actinolite. Brown hornblende and secondary white (turbid in thin section) plagioclase are commonly developed in the wallrocks of the vein, together with extensive actinolite mats. Epidote is an important phase in the alteration assemblage between 380 and 950 mbsf but is scarce both below and above this interval. Bleaching alteration is rare below 1000 mbsf. Many of the late magmatic leucocratic dikelets are overprinted by intense amphibole veining and brecciation (Figs. F121D, F123) and in some cases are recognizable only by the presence of zircon and apatite in breccia zones (Fig. F113F).

This alteration is linked to the intrusion of late, relatively leucocratic (and rarely quartz-bearing) melts. Difficulties in ascertaining the exact nature of the link arise because the primary beneficiary of the alteration is usually the intrusion itself. In most instances (i.e., where the intrusion lacks quartz), it is difficult to distinguish between the bleached intrusion and the bleaching in the surrounding gabbro. In some cases, the contacts of the leucocratic areas with the adjacent gabbro are gradational, either as alteration fronts or as zones of microbreccia with or without recrystallization of the breccia fragments (e.g., Fig. F123). On the other hand, minerals that appear igneous, especially zoned plagioclase, brown hornblende, and zircon, are commonly associated with the bleached zones.

We suggest that the late magmatic veinlets exsolved significant amounts of fluid in the latter stages of crystallization, forming brown hornblende. As cooling occurred, remaining magmatic fluid in fractures and fluid inclusions led to lower effective stress, localizing brittle failure. Further influx of fluid produced greenschist-facies alteration. The “magmatic” veins are almost always associated with (and crosscut by) the green amphibole veins described below.

Corona texture formation by reaction of plagioclase and olivine

In olivine-bearing gabbro, olivine gabbro, and troctolite, the most widespread alteration products are the aggregates of tremolite, talc, chlorite, and (tentatively identified) cummingtonite that form corona texture or coronitic pseudomorphs after primary olivine and plagioclase (Fig. F124A). Corona textures are not seen in olivine diabase. In thin section, the corona texture in Hole U1309D is very similar to that recorded in Hole U1309B (Fig. F112E, F112F). The most common texture consists of partial or complete replacement of olivine by tremolite (and/or cummingtonite) ± talc and replacement of the edges of neighboring plagioclase by chlorite. Chlorite is also present as fracture filling within plagioclase, often forming radial crack networks where olivine grains were completely surrounded by plagioclase (Fig. F124C, F124D). These textures are indicative of volume increase during the reaction. Tremolite may either entirely replace olivine or be restricted to its margins and may be intergrown with chlorite or talc. Cummingtonite (tentatively identified on the basis of twinning) has the same paragenesis as tremolite. Where relatively fresh olivine cores are present, an annulus of talc may separate it from the outer tremolite zone. Complex coronas containing carbonate, serpentine, and sulfides have also been recognized (Fig. F124E, F124F). Serpentine is generally present only as relics within fractures in olivine, as a late replacement of olivine cores (usually “oxidized serpentine”), or, rarely, as a replacement of tremolite. In a few cases, calcite is also found associated with serpentine, talc, or tremolite (Fig. F125A, F125B). Magnetite and sulfides (pyrite, pyrrhotite, and chalcopyrite) are often dispersed within the olivine pseudomorphs and locally form concentric patterns. Orthopyroxene, where present in the gabbros, alters to talc (Fig. F126B). In some samples, chlorite and actinolite form a reaction zone between orthopyroxene and plagioclase (Fig. F126C). This is analogous to the corona textures found between olivine and plagioclase, but it is usually not as well developed around orthopyroxene as it is around olivine. Coronas of tremolite and chlorite are not produced between clinopyroxene and plagioclase (Fig. F127B, F127C). In coronitic rocks, clinopyroxene is overgrown by actinolite or partially replaced by brown amphibole and appears to be uninvolved in the corona-forming reactions.

Corona texture is observed in hand specimen and thin section throughout the hole. Above 280 mbsf, the corona-forming reactions commonly went to completion, with all olivine being replaced by tremolite with or without small amounts of chlorite and talc. In olivine-rich troctolites, plagioclase is completely replaced by chlorite, leaving relict olivine which may be serpentinized (Fig. F127A). Below this depth, relict olivine becomes increasingly common in coronitic alteration zones. Below 350 mbsf, rocks appear where olivine and plagioclase are in direct contact (Fig. F126A). Rocks where the corona-forming reaction did not go to completion can often be recognized because characteristic “ladder vein” textures are developed in which olivine and plagioclase are replaced by serpentine and prehnite/​hydrogarnet, respectively (Fig. F124B). In the lower part of the hole (below 350 mbsf), there is a correlation between focused fluid flow and corona formation. Most coronas are associated either with dark green amphibole veins or with igneous contacts between gabbro and olivine gabbro (Fig. F128). Above this depth, coronitic alteration is pervasive and the coronitically altered rocks are notable for their lack of veining.

Actinolitic amphibole replacement of pyroxene

In most rocks cored in Hole U1309D, the principal metamorphism is a greenschist-facies event manifested by the replacement of pyroxene by green fibrous to acicular amphibole tentatively identified as actinolite. The replacement may be complete (i.e., pseudomorphic), but in less altered rocks, partial replacement along grain margins and cleavage or exsolution planes is more common (Fig. F112B). This amphibole is petrographically distinct from the colorless tremolitic amphibole that is common in the coronas developed during plagioclase-olivine reaction. In diabases, green amphibole, some of which is hornblende, replaces both pyroxene and olivine (Fig. F112D).

Actinolite replacement of pyroxene occurs in all pyroxene-bearing rock types and often occurs in the same thin section as corona texture. The intensity of pyroxene replacement in gabbros is often related to the presence of amphibole veins, particularly below 400 mbsf. In diabase, pyroxenes are often completely replaced without any macroscopic veining.

Green amphibole veins

Green amphibole veins are the most common vein type in Hole U1309D. Above ~200 mbsf, the most important vein set (Type 2 veins in the “Expedition 304 vein log” in “Supplementary material”) contains yellow-green actinolite and is often associated with breccia zones and late magmatic leucocratic dikelets. These veins are commonly braided and typically associated with an alteration halo 5 mm to 2 cm wide. The halo is usually much wider than the vein itself. At greater depths, dark green veins logged by the Expedition 305 scientists are interpreted as the same vein set. In some cases, dark green veins form dense networks (Fig. F129A), but they are also present in relative isolation. In other cases, the veins can be seen in thin section to be a zone of deformation that has been permeated by fibrous green amphibole. True amphibole veins commonly extend outward from these deformation zones.

In thin section, most dark green veins consist of 90%–100% colorless, pale brown, or pale green amphibole (probably actinolite). This amphibole is generally acicular to fibrous and is present as mats of needles or as needles aligned parallel or perpendicular to the vein. In many thin sections, colorless to pale green amphibole fills numerous small veins and fractures. It is especially conspicuous in fractures in plagioclase (Fig. F129B). Chlorite is the common accessory phase in the dark green veins, but it is not found in all veins and is almost always subordinate to amphibole. Turbid plagioclase containing abundant fluid and solid inclusions is commonly associated with actinolite veins (Fig. F121D).

Green amphibole veins show different relationships to other forms of alteration at different depths in the core. Above 330 mbsf, the coronitic reaction mostly went to completion, and amphibole veins appear to postdate this reaction. The main new alteration in vein halos is alteration of plagioclase to turbid plagioclase (probably albite) filled with solid and fluid inclusions. Corona-textured intervals are often notably free of macroscopic veins, but in some cases, amphibole in the coronas shows yellow-green discoloration in the vicinity of amphibole veins. Below 330 mbsf, the dark green amphibole veins are closely associated with localized development of corona textures and replacement of pyroxene by actinolite (Fig. F129B). Plagioclase close to the veins is altered to a variety of minerals that may include chlorite, albite, and prehnite. Orthopyroxene, when present, shows alteration to both amphibole and talc.

Our interpretation is that the pervasive alteration in the upper part of the hole predated the main phase of vein development. At deeper levels, the veining event introduced hydrous fluid into rocks containing unreacted igneous minerals and therefore caused localized occurrences of the reactions that had already occurred above. An important conclusion is that any or all alteration assemblages in the core may be diachronous in their development.

Serpentinization in olivine-rich rocks

The main mode of alteration of the most olivine-rich rocks (harzburgite, dunite, and olivine-rich troctolite) is serpentinization. The degree of serpentinization varies widely from >90% to <10% of original olivine. In the simplest case, serpentinization proceeds via the development of kernel texture (O’Hanley, 1996) (Fig. F130A). Olivine grains are penetrated and surrounded by serpentinite veinlets, leaving isolated fragments of olivine. In hand sample, areas of more extensive serpentinization form a conspicuous but generally irregular foliation. In places, this foliation reflects an original magmatic foliation defined by aligned olivine. The veinlets that define the foliation appear black because of included opaque phases (mostly magnetite, but some sulfides [pyrrhotite and, more rarely, pyrite]). Some brownish serpentine (possibly oxidized) is also locally found replacing olivine (Fig. F130B). Wehrlitic rocks generally show lower degrees of serpentinization than dunites and harzburgites at similar levels in the core. In a few samples, tremolite has grown on the margin of clinopyroxene (Fig. F127B). In general, however, clinopyroxene oikocrysts are fresh and have sharp boundaries against olivine and serpentine (Fig F127C).

Plagioclase in the serpentinized rocks ranges from virtually unaltered to partially altered (usually to prehnite) to complete replacement by hydrogarnet and/or prehnite (Fig. F127E). This is a rodingitic assemblage normally interpreted as being due to introduction of Ca into a mafic rock, although it can equally well be modeled by removal of silica. The coupled serpentinization and rodingitization are spectacularly developed in ladder veins seen below 330 mbsf (Fig. F124B). Black seams of serpentine contain white “bridges” of prehnite and hydrogarnet replacing plagioclase. In thin section, it is clear that hydrogarnet postdates prehnite in the local metasomatic sequence and that serpentine, prehnite, and hydrogarnet all postdate a chlorite rim that formed in an earlier corona-forming event (Fig. F127F). Serpentine is also seen to replace tremolite (Fig. F127D). Ladder veins can vary rapidly in orientation over a few meters of core and sometimes occur parallel to gabbroic veins, where they continue as prehnite veins. Hydrogarnet appears to be most common in highly serpentinized rocks, whereas the plagioclase is essentially fresh in the one known brucite-bearing sample. One commonly observed feature in serpentinized rocks is the development of fracture sets that radiate or extend into plagioclase from neighboring serpentinized olivine grains (Fig. F131). These fractures appear to be filled with prehnite and/or hydrogarnet and demonstrate that volume increase accompanied the reaction. Above 330 mbsf, the coronitic reaction has mostly gone to completion and prehnite and hydrogarnet are rarely developed even in plagioclase-rich rocks. This is further evidence for the coupled nature of the serpentinization and rodingitization reactions.

Veins

Talc-tremolite-chlorite rocks

Talc-tremolite-chlorite rocks are a minor but important component of the rocks encountered in Hole U1309D. A pale green tremolite + talc + chlorite rock was recovered as a 10 cm thick horizon in Section 304-U1309D-1R-3 (Fig. F132A), and veins with similar assemblages were found cutting peridotitic rocks sporadically throughout the core. Two thin sections from interval 304-U1309D-1R-3, 0–11 cm (Samples 1R-3, 0–4 cm, and 1R-3, 9–11 cm), display relations that are important for understanding the origin of these rocks. Sample 304-U1309D-1R-3, 0–4 cm, contains abundant tremolite and talc with ~20% chlorite. This talc-rich composition, along with the presence of large chromite grains (Fig. F132B, F132C), indicates that this rock was derived, at least in part, from metaperidotite. Sample 304-U1309D-1R-3, 9–11 cm, in contrast, contains far more chlorite (up to 50%), as well as tremolite pseudomorphs after pyroxene and minor amounts of brown hornblende (Fig. F132D). All these factors suggest that this rock was formed to a large extent from gabbro (Fig. F132E). These samples may correlate with the piece of tremolite schist found in a breccia at about the same level in Hole U1309B (Fig. F113A, F113B). Similar rocks were recovered from Hole U1309H and collected from the detachment fault on the Southern Ridge of the massif (Schroeder and John, 2004). A likely explanation for their present location is that the detachment fault was intruded by basaltic and diabasic intrusions.

Clues to the origin of the talc-tremolite schists are found in tremolite ± chlorite ± talc veins that cut peridotitic rocks in intervals 304-U1309D-10R-1, 90–100 cm, 31R-1, 130–138 cm, and 63R-3, 1–4 cm, and talc-tremolite assemblages that are commonly seen at contacts between ultramafic and mafic rocks. Extensive talc metasomatism associated with tremolite in two narrow zones below 500 mbsf also deserves mention. The first zone is observed in interval 305-U1309D-111R-2, 79–103 cm (~550 mbsf), and the second one is in interval 140R-2, 110–113 cm (~691 mbsf). These two zones are evidence of extensive replacement of the initial troctolitic gabbro (intervals 305-U1309D-111R-2, 79–103 cm, and 140R-2, 110–113 cm, respectively) by a secondary mineral assemblage dominated by talc. In Section 305-U1309D-111R-2, the metasomatism is present as 1–10 cm wide alteration halos around amphibole-chlorite veins (Fig. F133A). In Section 305-U1309D-140R-2, olivine-rich troctolite exhibits a striking banded appearance and talc metasomatism is developed at several contacts between olivine-rich troctolite and gabbro. All these cases represent greenschist-facies metasomatism of peridotite by fluids rich in Si and Ca that have passed through mafic rocks. The reciprocal metasomatic reaction in mafic rocks leads to complete chloritization of plagioclase and replacement of actinolite by tremolite. A very similar vein from Hole U1309B (Fig. F115) has been described in detail and consists of a tremolite core and talc margins replacing serpentine and olivine. The tremolite-talc rocks are clearly metasomatic and require introduction of a fluid with higher Si/Mg and Ca/Mg activity ratios than that in equilibrium with serpentine. The most likely local source for such a fluid is adjacent mafic rocks. We believe that the same metasomatic reactions produced the talc-tremolite-chlorite schists found in the uppermost cores of Holes U1309B, U1309D, and U1309F and on the Southern Ridge of Atlantis Massif (Schroeder and John, 2004).

The talc-tremolite veins show ambiguous relationships to serpentinization. Figure F115D shows that talc replaces serpentine on the margin of veins, whereas Figure F115G shows serpentine replacing tremolite. The corona textures, which reflect a similar reaction, clearly predate serpentinization. We have tentatively placed the talc-tremolite alteration in both the schists at the top of the massif and in ultramafic layers within the core after the main serpentinization event, recognizing that both alteration processes are almost certainly diachronous, may occur at different rates in different lithologies, and may overlap in time.

Talc-carbonate veining and alteration in olivine-rich rocks

Some carbonate veins are associated with talc-carbonate alteration of the peridotite in the upper 350 m of the core (Fig. F134) (Sample 304-U1309D-56R-1, 20–23 cm). Talc-carbonate alteration is characterized by the replacement of olivine by carbonate (presumably magnesite) and talc. This type of alteration is very rare below 350 mbsf. Carbonate veins and replacement zones are also present in serpentinites close to gabbro contacts and may be related to Lost City hydrothermal vent field–type fluids.

Slip-fiber amphibole veins and associated metasomatism

Some amphibole veins consist of white to brownish tremolite/​actinolite in a slip-fiber configuration. Where crosscutting relationships are observed, these veins invariably cut the dark green amphibole veins. The core commonly breaks along slip-fiber veins, revealing, in many cases, a strong lineation in hand sample (Fig. F133A) (see “Structural geology” for a detailed discussion of this lineation). Some slip-fiber veins in more olivine-rich lithologies contain talc and/or serpentine. A few veins also contain carbonate. Slip-fiber veins are usually pale green, pale bluish green, or white in hand sample. Serpentine and talc-rich examples may be darker or brighter green. In thin section, the amphibole may be pale green, colorless, brown, or, in many cases, very dark brown. Other light green amphibole veins have the optical peculiarity of appearing pale green or bluish green in hand sample and dark honey-brown in thin section (Fig. F133C). These veins can contain a substantial amount of isotropic and/or very fine grained material that may include hydrogarnet. At shallower levels in the core, isolated veins (Type 3 in the Expedition 304 vein log in “Supplementary material”) containing actinolite ± chlorite ± titanite occur in diabase and gabbro and postdate the green amphibole veins described above. These may be the same age as the slip-fiber veins described by the Expedition 305 scientists.

There are some large alteration halos, as wide as 10 cm, associated with slip-fiber veins. Many of these veins, however, pass through the rock with little apparent effect except for local brecciation (Fig. F133C). The most spectacular halos tend to be present in olivine-rich rocks. In several sections (e.g., Sections 305-U1309D-140R-2 and 111R-2) (Fig. F133A, F133C), slip-fiber veins are associated with wide zones of talc metasomatism, where rocks for as much as 10 cm on either side of the vein are replaced almost entirely by talc. Other metasomatic zones are more complex and resemble rodingites. In the latter areas, olivine is completely replaced by tremolite and/or talc, pyroxene is completely replaced by amphibole (occasionally, clinopyroxene remains unaltered), and plagioclase is completely replaced by chlorite, albite, prehnite, zeolites, or hydrogarnet for up to 10 cm around a 2 mm vein. As with the more abundant dark green veins, the signature of this alteration seems to be the transport of silica into olivine-rich lithologies.

Quartz veins

Quartz veins are dispersed sparsely through the core but are more abundant in a few intervals between 200 and 300 mbsf. They disappear from the vein assemblage below 800 mbsf. They are sometimes associated with trondhjemite intrusives but also occur as isolated veins, sometimes containing chlorite. Quartz shows unusual sector extinction and contains fluid inclusions (Fig. F121E, F121F).

Zeolite and other late veins

These veins postdate the greenschist-facies metamorphism. They have variable mineralogy that, at least in part, correlates with depth. Late carbonate veins are common in ultramafic rocks from the upper 350 m of Hole U1309D. These commonly parallel the contacts between the ultramafic rock and late gabbroic and dioritic dikes (Fig. F134A). Although some carbonate veins are associated with talc-carbonate alteration of the peridotite (see Fig. F134B, F134C) (Sample 304-U1309D-56R-1, 20–23 cm), others (Fig. F134D) leave the serpentinite completely unaltered and appear to be replacive. Carbonate and, possibly, zeolite (commonly with small amounts of sulfide minerals) are constituents of the white veins in the upper 400 m of Hole U1309D. Confirmed zeolites are absent between 0 and 700 mbsf. They become increasingly common below 700 mbsf. Zeolites identified from the lower part of the hole include analcime and natrolite-thomsonite. In some veins, the zeolites are associated with prehnite (Fig. F23) and, in one case, anhydrite. Note that anhydrite solubility in seawater is retrograde (Mills et al., 1998); it dissolves in seawater with a temperature of <160°C. Thus, the hydrothermal system from which the anhydrite precipitated must have been closed to direct seawater input at temperatures >160°C. Alteration associated with the white veins varies widely from little or no discernable halo to irregular halos in which plagioclase seems especially susceptible to alteration (usually to albite, prehnite, and/or zeolites).

Clay (saponite) veins

The youngest vein set in Hole U1309D consists of a waxy, translucent, pale to dark green clay mineral (easily mistaken for talc) that may be saponite. The combination of the friable, hydroscopic nature of the clay vein material and the propensity of the core to break along clay-filled fractures makes shipboard thin sectioning impossible, so identification is based on XRD data. A more complete characterization of the clay will be a shore-based activity. The clay is present along with calcite and/or zeolites (especially thomsonite) in broad, uneven fractures in the rock. In the few cases where an alteration halo is present, it is likely that the halo is left over from an earlier vein fluid that passed through the same conduit. The clay veins seem to be most commonly associated with olivine-rich rocks and are notably abundant in gabbroic rocks intercalated with olivine-rich troctolite between 1100 and 1230 mbsf.

Although we have no firm temperature constraints on the formation of the clay veins, bottom hole temperatures measured during Expedition 305 (120°C) (see “Downhole measurements”) provide a minimum temperature constraint.

Oxide and sulfide mineralogy

The main oxide phases are ilmenite and magnetite (Fig. F135A, F135C, F135D). Rutile is present in association with titanite in some altered ilmenites. Ilmenite is found in all lithologies, with the highest amount (up to 8 vol%) observed in oxide gabbros and olivine gabbros. Ti-rich magnetite in the oxide gabbros has recrystallized to a mixture of ilmenite and magnetite. Magnetite is also found as a secondary phase resulting from the serpentinization of olivine in most olivine-bearing rocks.

The sulfides present as primary and secondary phases are pyrrhotite, pyrite, chalcopyrite, and, less commonly, bornite and pentlandite (Fig. F135). Sulfides occur both as a primary phase in the gabbros and as a metamorphic mineral. Pyrrhotite and pyrite are found in alteration products after both serpentinized and corona-textured olivine and in pyroxene altered to amphibole. Some pyrrhotite is present in veins associated with calcite. Chalcopyrite generally occurs with igneous pyrrhotite (Fig. F135B–F135D). In a few cases, “flame textured” pentlandite was observed as exsolutions in pyrrhotite (Figs. F135B, F135C). One possible occurrence of a Ni-Fe alloy (awaruite) was noted in the olivine-rich troctolite (interval 305-U1309D-247R-3, 97–98 cm) associated with serpentinized olivine. The sulfide phases in the gabbro, oxide-gabbro, troctolite, and olivine-rich troctolite are mainly low-sulfur assemblages. They suggest low fS2 and low fO2 for the rocks recovered in Hole U1309D.

Downhole variation of modal alteration intensity

Hole U1309D shows a heterogeneous alteration profile, ranging from fresh (<2%) to highly altered (up to 100%); however, most of the cores below 350 mbsf show a low to moderate alteration. The overall alteration, shown in Figure F136, is the estimate from visual examination of the core for the modal abundance of secondary hydrous minerals. An expanded version of this figure from the top 400 m of Hole U1309D (Fig. F137A) shows a general decrease of alteration downhole, with peaks corresponding to olivine-rich horizons.

The degree of alteration of rocks from the uppermost ~350 mbsf is relatively high (as much as 100%) and reflects the pervasive alteration of all rock types high in the core (Fig. F137A). At depths less than ~100 mbsf, all rocks in the core display some degree of hydration, and hydration and metamorphism are essentially pervasive to a depth of 350 mbsf. Traces of fresh clinopyroxene appear at ~40 mbsf and become more abundant downhole. Fresh gabbro with essentially no alteration is found below ~135 mbsf, and some fresh pyroxene is found in most samples below that depth. Fresh olivine appears in the troctolites at ~150 mbsf, and fresh troctolite with only minimal alteration of olivine appears at 380 mbsf. Prehnite alteration of plagioclase occurs locally below 300 mbsf, where it is most commonly associated with serpentinization of neighboring olivine. Epidote-bearing veins were found in intervals 304-U1309D-75R-2, 77–80 cm, and 77R-3, 112–118 cm.

Above 400 mbsf, vein-related alteration is generally superimposed on more pervasive alteration. Vein-related alteration (Fig. F137B) peaks from 0 to 50 m, where it is associated with breccia zones, and between 150 and 300 mbsf, where it is mostly associated with late leucocratic magmatic dikelets. Below 400 mbsf, virtually all alteration appears to be vein related, and hence this has not been plotted separately. From ~350 to 840 mbsf, the rocks recovered show moderate alteration. Zones 0.2–3 m wide with low alteration (total alteration = ~10 vol%) alternate with zones 0.4–3 m wide with moderate alteration (total alteration = >40 vol%). Higher alteration intensity in this interval is mainly in coarse-grained gabbro, later intrusions (leucocratic magmatic rocks and oxide gabbro), and zones of high deformation (cataclasite or mylonite). Two peaks showing a total alteration of >80 vol% (Cores 305-U1309D-152R and 162R at 746 and 789 mbsf, respectively) correspond to zones of highly altered gabbro with a cataclastic fabric. In both places, alteration is characterized by near-complete replacement of pyroxene by fibrous green amphibole. Core 305-U1309D-162R at 789 mbsf also shows a strong bleaching alteration with formation of secondary plagioclase and epidote. The highest peak of alteration (>90 vol% at 551.7 mbsf) corresponds to serpentinized olivine-rich troctolite, where olivine is completely replaced by serpentine.

From 841 to 1094 mbsf, the recovered rocks (mostly gabbro) generally show very low overall alteration (~2–3 vol%). Extensive alteration is restricted to zones with a high density of veins and to contacts between gabbro and diabase (Cores 305-U1309D-205R and 206R). There are some local peaks of higher alteration (50 vol%) corresponding to zones of high deformation (Section 305-U1309D-173R-1; 842 mbsf). Zones with total alteration amounts between 20 and 30 vol% are related to alteration halos around veins and fractures.

From 1095 to 1415.15 mbsf, the alteration is heterogeneous, showing increased degree of alteration between 1095 and 1200 mbsf. This high alteration zone (locally >90 vol%) corresponds to highly deformed zones of protomylonite (see “Structural geology”) and to zones of rodingitization within the olivine-rich troctolite (Sections 305-U1309D-231R-2, 234R-2, 237R-1, and 248R-2). The rodingites are strongly altered gabbroic rocks that appear to have undergone metasomatic reaction, possibly with serpentinizing fluids derived from nearby olivine-rich troctolites. The former gabbros now consist largely of chlorite, tremolite, and talc. The lowermost 200 m of Hole U1309D shows decreased alteration (~30 vol%), similar to the interval 841–1094 mbsf. The last high peak of alteration corresponds to zones of intense veining associated with bleaching alteration (green amphibole replacing pyroxene and crystallization of secondary plagioclase).

Most of the rock types exhibit the entire range of alteration, from 0% to 100%, except for the diabase, which mainly is from 0% to 40% altered (Fig. F138A, F138B). Gabbro is the main lithology recovered in Hole U1309D and exhibits a range of alteration from 0% to 100%, but most rocks (~90% of the recovered gabbros) are altered within the range 0–40 vol%. The oxide gabbros and the olivine and troctolitic gabbros show the same moderate alteration range as the gabbros, within the 0%–40% range. The troctolites and olivine-rich troctolites show variation in alteration from 0% to 100%; ~10% have alteration >60%, and 1% of them have alteration >80% that is related to the serpentinization process of these olivine-rich rocks.

Downhole variation in alteration mineralogy

The secondary mineralogy for cores collected during Expedition 305, as identified in thin section, is given in Table T7. Total alteration estimated in thin section (Fig. F139B) mirrors the hand specimen–based alteration log but contains a built-in bias toward altered samples deeper in the core. We have sought to remove this bias by plotting only the mineral abundances in thin sections from geochemical samples (Fig. F139C), which were generally selected to be the least altered parts of the core. Plots of the abundances of the secondary minerals in thin section show some systematic variations downhole. The second diagram in Figure F139 includes data only from the geochemical samples and clearly shows the presence of almost unaltered rock between 700 and 1100 mbsf. Green to colorless (actininolitic) amphibole (Fig. F139E) is the most abundant secondary phase in the rocks in thin section. It replaces pyroxene, occurs in corona textures, and is present in veins; thus, the percentage of actinolitic amphibole correlates closely with total alteration except in serpentinized horizons. Brown amphibole is present in minor amounts as patches in pyroxene throughout the core. The abundance of brown amphibole (Fig. F139D) is greatest high in the hole and decreases markedly downhole from 400 mbsf, with peaks in oxide gabbro layers where it may be of primary origin. Talc (Fig. F139J) and chlorite (Fig. F139F) are involved in the formation of corona texture and are present in varying amounts, but they are most common in the upper part of the hole and in intervals of higher intensity of veining. Peaks in talc also correspond to some ultramafic horizons. Serpentine (Fig. F139K) is present at least as a very minor component of almost all rocks containing olivine below 330 mbsf, although it may make up only a very small percentage of secondary minerals (as fine veinlets along fractures in olivine grains). Serpentine abundance peaks in several intervals that correlate with some higher degrees of alteration of olivine-rich rocks. Prehnite (Fig. F139L) and hydrogarnet (Fig. F139M) are almost absent above 330 m because of the corona-forming reaction going to completion. Prehnite is most common between 330 and 430 mbsf, where it is present in ladder veins, whereas the main occurrence of hydrogarnet is below 1100 mbsf. Epidote (Fig. F139H) and titanite (Fig. F139I) appear in association with zones of leucocratic alteration, but epidote is also observed as zoned, euhedral crystals in veins. Epidote is rare in cores shallower than 380 mbsf and deeper than ~910 mbsf. Zeolite (Fig. F139N) appears below 751 mbsf. Albite and titanite appear in zones of leucocratic alteration throughout the cores, but albite is most common above 300 mbsf.

Downhole variation in vein composition and abundance

Veins containing dark green and slip-fiber amphibole are the most common type observed in hand sample (Fig. F140). Other vein types include secondary plagioclase, serpentine, clay, and white secondary minerals (usually zeolite + carbonate). The vein data recorded during visual inspection of the core and plotted in the following figures are available in the vein log (see “Site U1309 core descriptions” in “Core descriptions”). Vein mineralogy constrained by XRD and petrography is given in Table T8. Although there are several distinct maxima in the number and volume of veins per core, there is no consistent correlation between depth or rock type and number or volume of veins (Fig. F141). There is a correlation between both number and volume of veins and the location of faults in the core. In particular, major maxima in vein count are present in close association with faults at 695, 746, 989, 1070, and 1337 mbsf. Smaller maxima are associated with faults at 785 and 960 mbsf. Of particular note is the lack of correlation between magmatic veins and faults. Only one magmatic vein maximum, at 695 mbsf, is present in conjunction with a fault, but there is also a peak in nonmagmatic veining at this point. No faults have been recognized that correspond to peaks at 440, 550, 1190, and 1230 mbsf.

Faults also correlate with abundance of the two major amphibole-bearing vein groups: dark green amphibole veins and slip-fiber amphibole veins (Fig. F142). Serpentine veins are present, not surprisingly, at depths dominated by olivine-rich rocks (Fig. F142). Magmatic veins are most abundant near the top of the hole (above 700 mbsf) (Fig. F143). They are particularly abundant, as indicated by the peak in secondary plagioclase, between 150 and 350 mbsf (Figs. F142, F144). A major increase in vein abundance is associated with the olivine-rich troctolites between 1100 and 1230 mbsf. Although veins in the olivine-rich troctolites themselves are mostly restricted to veining/​foliation associated with the early serpentinization, intercalated gabbroic rocks are highly fractured and intensely veined. Slip-fiber and clay (saponite?) veins, in particular, show a strong spike associated with alteration of gabbros intercalated with the olivine-rich troctolites between 1100 and 1230 mbsf (Figs. F141, F142, F143).

The components of the “white” veins vary significantly with depth. Above 400 mbsf, sulfide-bearing veins are abundant and quartz-bearing veins occur (Fig. F145). Quartz is rare below 400 mbsf, and little sulfide is found in veins below 800 mbsf. Zeolite appears at ~700 mbsf and increases in abundance downhole to the deepest penetration of Hole U1309D (Fig. F145). Carbonate is present in veins throughout the hole. A peak between 700 and 850 mbsf reflects the presence of calcite in many slip-fiber veins through that interval (Fig. F144).

Metamorphic conditions at Site U1309

Granulite-facies recrystallization

Granulite-facies shear zones occur mainly above 350 mbsf. In one shear zone studied during preliminary shore-based electron microprobe studies (Fig. F146A; Table T9; see “Glass” in “Supplementary material”), pyroxene neoblasts have similar Al and Ti contents to large grains in the host gabbro and brown amphibole grains in the shear zone contain 3.5% TiO2. The texture and mineralogy are both indicative of granulite facies and likely reflect formation at near-solidus conditions (probably 800°–900°C). Undeformed zones of recrystallized pyroxene and plagioclase occur locally and may mark zones of high-temperature fluid flow (Maeda et al., 2002). Pyroxene has also been observed filling microveins in igneous pyroxenes (Fig. F146B; Table T9), although the temperature at which this occurred is uncertain.

In some gabbros, especially oxide gabbro, there is substantial static replacement of clinopyroxene by green and brown amphibole. In general, such textures are interpreted as magmatic, having formed by incongruent, vapor-absent reaction between hydrous melt and pyroxene (e.g., Beard et al., 2004).

clinopyroxene + Fe-Ti oxides + hydrous melt ±

plagioclase (calcic) = hornblende +

SiO2 (solid or melt) ± plagioclase (sodic).
(1)

The similarity of igneous brown amphibole in the oxide gabbros and metamorphic amphibole in the mylonites suggests that igneous activity overlapped deformation and metamorphism.

Amphibolite-facies recrystallization

Statically grown brown to green-brown amphibole has been found partially replacing pyroxene at all levels in Holes U1309B and U1309D, and amphibole neoblasts are observed in some shear zones. Electron microprobe analysis of one diabase sample (Fig. F147; Table T9; see “Glass” in “Supplementary material” for analyses) reveals a wide range of amphibole compositions ranging from fluorine-rich ferro-edenite to magnesiohornblende and actinolite, accompanied by a range of plagioclase compositions including labradorite, andesine, oligoclase, and albite. These observations indicate that hydration of the diabases, and probably the gabbros, occurred in both the amphibolite and greenschist facies and that the extent of amphibolite-facies metamorphism may have been underestimated on board ship.

A model for the formation of hornblende from pyroxenes is Reaction 2:

plagioclase + augite + orthopyroxene + H2O = hornblende. (2)

Because the composition of hornblende changes as a function of temperature and pressure, the exact stoichiometry of Reaction 1 is difficult to determine. Despite this, the location of the reaction in pressure-temperature space (Fig. F148A) is fairly well known (Spear, 1981). The experimental location of this curve shows the conditions at which hornblende begins to break down during heating of a typical metabasite. For a retrograde situation, this reaction marks the conditions at which pyroxene in a metabasite would be completely altered to amphibole if the rock were open to the movement of H2O. Because the substitution of H and Cl for OH and of Ti and Fe3+ into hornblende stabilize it to higher temperature, the temperature at which hornblende would start forming during cooling of a metabasite may be significantly higher than that at which reaction 1 takes place.

Corona-forming reactions

In the simplified system CaO-MgO-Al2O3-SiO2-H2O, reactions involving the phases olivine, anorthite, tremolite, and chlorite are divariant. Reactions among these phases can only be written assuming that one component is mobile. We have chosen to use CaO and SiO2 as mobile components because Al2O3 has limited mobility in most low-temperature environments and because oceanic fluids have notoriously low Mg contents. Using mobile components of SiO2 and CaO, we have calculated the stability relations between olivine, tremolite, talc, anorthite, and chlorite for 1 kbar and 450°C (Fig. F149). We have used α-quartz and lime as standard states for SiO2 and CaO. It is important to note that, if one chose different standard states (such as aqueous species) for the calculations, only the numbers on the coordinates would change; the topology would stay the same.

Figure F149 shows that the coronas around olivine could have been produced by a gradient in the activity of CaO and SiO2 between plagioclase and olivine. Because Al2O3 has low mobility, chlorite produced by this reaction will be concentrated adjacent to plagioclase, whereas tremolite will form adjacent to olivine (see arrow A in Fig. F149). Where the fluid gradient was relatively richer in SiO2 and poorer in CaO (arrow B in Fig. F149), a zone of talc will appear between olivine and tremolite.

There are two ways to write the reaction between olivine and plagioclase, depending on whether silica or CaO is the mobile component:

4CaAl2Si2O8 + 15Mg2SiO4 + 5SiO2 + 18H2O =

plagioclase       olivine       silica in fluid     fluid

2Ca2Mg5Si8O22(OH)2 + 4Mg5Al2Si3O10(OH)8,

amphibole                chlorite
(3)

and

11CaAl2Si2O8 + 35Mg2SiO4 + 47H2O =

plagioclase          olivine         fluid

3Ca2Mg5Si8O22(OH)2 + 11Mg5Al2Si3O10(OH)8 + 5CaO.

amphibole                          chlorite                Ca in fluid
(4)

Reactions 3 and 4 both involve large volume increases for the solid phases—24% and 19%, respectively. This explains why intense fracturing is commonly associated with the corona-forming reactions. In feldspar-dominated rocks, chlorite-filled cracks radiate from pseudomorphs of olivine and across the surrounding plagioclase and follow curved paths from one olivine pseudomorph to another (Fig. F124C, F124D).

Because of the multivariant nature of Reactions 3 and 4, we cannot calculate the breakdown temperature of the assemblage olivine + anorthite directly; however, natural assemblages place clear constraints on the occurrence of this assemblage. Frost (1976) noted that plagioclase never appears in the presence of chlorite in peridotitic hornfelses; spinel is the aluminous phase instead. From this, he concluded that plagioclase appears in peridotitic rock a few tens of degrees above the terminal reaction for chlorite. This means that, during hydration, plagioclase + olivine will react out of a rock near the upper limits for amphibolite facies (see heavy gray line on Fig. F148A). Of course, if the reaction between olivine and plagioclase occurred at such high temperatures, the product would be hornblende, rather than chlorite + actinolite. We do see traces of brown hornblende in some coronas, but most consist of chlorite + actinolite. In mafic rocks, chlorite + actinolite is a typical greenschist assemblage, indicating that the coronas probably formed under the same conditions as the actinolite alteration of the gabbros—approximately 400°C. Note that, if the presence of cummingtonite is confirmed, it is likely that at least some corona-forming reaction occurred at substantially higher amphibolite-facies temperatures.

Serpentinization and prehnite-hydrogrossular assemblages

Serpentine can form in the presence of pyroxene via

Mg2SiO4 + MgSiO3 + H2O = Mg3Si2O10(OH)2,

olivine         orthopyroxene        fluid          serpentine
(5)

and

6Mg2SiO4 + Ca2Mg5Si8O22(OH)2 + 9H2O =

olivine              tremolite              fluid

5Mg3Si2O5(OH)4 + 2CaMgSi2O6.

serpentine                diopside
(6)

Brucite was detected by XRD (Fig. F118) in one serpentinized olivine-rich troctolite (Sample 305-U1309D-111R-3, 131–138 cm). This suggests the straightforward serpentinization reaction:

2Mg2SiO4 + 3H2O = Mg3Si2O5(OH)4 + Mg(OH)2.

olivine          fluid             serpentine          brucite
(7)

Although we were unable to find brucite in a thin section (305-U1309D-111R-4, 22–25 cm; thin section number 328) near the XRD sample, this thin section contains no tremolite or talc, both of which are common in most olivine-bearing rocks in Hole U1309D and are normally unstable in the presence of brucite. Brucite is commonly finely intergrown with serpentine and is notoriously hard to determine in thin section. Because olivine is typically more Fe rich than coexisting serpentine, magnetite typically forms as a consequence of serpentinization Reaction 5 with the iron end-member (fayalite) of the olivine (O’Hanley, 1996; Frost, 1985).

Serpentine can also form directly from olivine in the presence of a Si-rich fluid via the reaction:

3Mg2SiO4 + SiO2 + 4H2O = 2Mg3Si2O5(OH)4.

olivine        silica in fluid          fluid             serpentine
(8)

It is possible that serpentines in rocks that contain prehnite and hydrogrossular formed via reactions of this sort, possibly with Si derived from the reaction of plagioclase to the rodingitic assemblage.

Reaction 5 marks the highest temperature that serpentine can form in a rock that has not previously been hydrated (i.e., one that lacks talc) (Fig. F148B). Using the compositions of phases usually found in peridotite (Trommsdorff and Evans, 1974), we calculate that, at 1 kbar, Reaction 5 takes place at ~460°C (see Fig. F148B). Reaction 6 marks the lowest temperature stability for tremolite in serpentinites; at 1 kbar this is 415°C. Reaction 7 is the lowest temperature at which olivine is stable in the presence of pure water. At 1 kbar, this temperature is 354°C (Fig. F148B).

In a number of rocks from Hole U1309D, we find tremolite growing from clinopyroxene in the presence of olivine (Fig. F127B). Tremolite is too silica rich to form by direct hydration of clinopyroxene in peridotite; its formation requires the input of silica or the expulsion of CaO. Because of this, we argue that textures in which tremolite grows on clinopyroxene must have formed in the presence of free water, and, therefore, this texture must have formed at temperatures above those of Reaction 6. Otherwise, the influx of water would have hydrated the olivine, not the clinopyroxene. In contrast, there are many serpentinites in which the clinopyroxene is still fresh (Fig. F127C). In these rocks, hydration could have occurred at temperatures below those of Reaction 6 or even at temperatures below those of Reaction 7.

The association of prehnite, hydrogrossular, and chlorite (Fig. F127F) constrains the formation of the prehnite alteration to temperatures below Reaction 9:

26Ca2Al2Si3O10(OH)2 + Ca3Al2Si3O12 + 3Mg5Al2Si3O10(OH)8 =

prehnite                    garnet                    chlorite

15CaMgSi2O6 + 20Ca2Al3Si3O11(OH) + 28H2O.

diopside             clinozoisite             water
(9)

At 1 kbar, this reaction in the pure system CaO-MgO-Al2O3-SiO2 occurs at 350°C (Fig. F148B). The prehnite-forming reaction must have occurred at temperatures below this; thus, fluid must have been flowing through the olivine-rich troctolites at least to temperatures <350°C.

The distribution of epidote in the core presents an interesting problem. Reaction 10 gives the upper limit for clinozoisite stability in the presence of chlorite (~350°C at 1 kbar; Fig. F148). Epidote is clearly incompatible with the corona-forming reaction, where Fe contents of minerals are low. High Fe contents expanding the stability field may account for the formation of epidote at intermediate levels of the core:

10CaAl2Si2O8 + Ca2Mg5Si8O22(OH)2 + 6H2O =

plagioclase                   tremolite                     fluid

6Ca2Al3Si3O12(OH) + Mg5Al2Si3O10(OH)8 + 7SiO2.

epidote                  chlorite                  quartz
(10)

Movement of late magmatic and hydrothermal fluids

Above 330 mbsf, an early pervasive fluid flow event under low-amphibolite to upper-greenschist conditions affected most rocks in Holes U1309B and U1309D. Metamorphic reactions appear to have played an important role in creating permeability in these rocks. Focused fluid flow in amphibole veins and breccia zones and along late magmatic leucocratic dikelets was superimposed on this event. At greater depths, most of the alteration is related to veins. The gabbroic rocks in Hole U1309D show only minor evidence for prehnite-pumpellyite-facies alteration. Prehnite below 330 mbsf is clearly related to serpentinization and formed above 350°C. This relation suggests that, following the last greenschist hydration event, the rocks in the upper part of the core became closed to the movement of fluids until they had cooled below temperatures of ~100°–200°C. At these temperatures, fracturing formed the clay-bearing veins that cut many of the sections of the core. The main clay mineral detected by XRD is saponite (Table T4; Fig. F117). It is likely that much of the clay and zeolite observed in fractures below 700 mbsf formed under essentially ambient conditions because temperatures at the bottom of the hole were at least 120°C. The lack of zeolite above this level may reflect rapid cooling through the appropriate temperature interval during unroofing, rather than a lack of fluid flow.

In contrast to the upper parts of the hole, there is a close correspondence between veining, zones of cataclasis, and fluid movement in the lower 1000 m of Hole U1309D. The correspondence ranges from thin section–scale observations of the relationship between cataclasis and veining to an overall correlation of downhole alteration with major faults (see “Structural geology”).

Although some zones of cataclasis are unaltered at the thin section scale, more commonly they are pervaded by secondary minerals, especially fibrous amphibole, to the extent that they resemble dark green veins on visual inspection of the core. In some cases, there is an apparent history of fluid flow in these zones. Early high-temperature fluids, probably hydrous silicate melts, crystallize euhedral zircon, titanite, zoned plagioclase, and, more rarely, blocky green amphibole that is locally zoned. Later, and presumably at lower temperature, fibrous amphibole has grown as intergranular stringers and patches and as a replacement of pyroxene (Fig. F150A). Dark green amphibole veins may branch out from the zone of cataclasis into surrounding undeformed rock. Alteration of plagioclase in the zone of cataclasis at this stage may produce albite, prehnite, and, especially lower in the core, zeolites. The latest fluids crystallize, thus filling void space into which earlier higher temperature minerals have grown (Fig. F150B).

The composite talc-tremolite veins seen in some ultramafic layers (Fig. F115) may be a manifestation of the same greenschist-facies fluid flow event as the actinolite breccia veins. The intense metasomatic alteration is due to the movement of fluid rich in Si and Ca from mafic to ultramafic rocks. Similar zones where plagioclase is completely altered to chlorite (Fig. F132E) reflect flow from ultramafic into mafic rocks.

Fluid flow also seems to have been localized by gabbroic intrusions. Below 300 m, coronitic alteration of olivine and plagioclase in olivine gabbros or troctolites is often associated with zones within a few centimeters of late pegmatitic gabbros (Fig. F128) This could be a result either of exsolution of magmatic fluid or localization of fracturing and fluid infiltration into more competent gabbros. The fluid influx that produced the serpentine-prehnite alteration was also localized. Although the reasons for this are not entirely clear, the alteration is clearly fracture related and serpentine-prehnite ladder veins often appear to follow stress trajectories (e.g., rotating to be perpendicular to lithological boundaries), so a tectonic control appears probable.

Larger scale structures also influence the movement of metamorphic fluids. In particular, there is a close correspondence between veining, alteration, and the location of identified faults. Most major maxima in vein abundance are associated with these faults (Fig. F141). In addition, the overall downhole decrease in alteration is interrupted twice, once between 600 and 800 mbsf and once between 1100 and 1200 mbsf. In the first case, there are major faults near the bottom of the highly altered section. These faults may have acted as conduits, releasing metamorphic fluids into the overlying rocks, leading to increased veining and alteration.

In the second case, things may be somewhat more complex. There are faults beneath this interval, but much of the fluid flow may have been focused, not by fault zones but by relatively impermeable intervals of fine-grained, partially serpentinized, olivine-rich troctolite.

Summary

Rocks in Hole U1309D experienced metamorphism at grades ranging from granulite to zeolite facies, reflecting the cooling and deformation of a gabbro body. Granulite-facies metamorphism is restricted to mylonite zones in which plagioclase and pyroxene are recrystallized to an equigranular mosaic texture. This deformation may have been syn–late magmatic. Amphibolite-facies metamorphism includes hornblende-forming reactions in diabase, oxide gabbro, and mylonites. Some amphibolite-facies metamorphism may have been coincident with the latest stages of magmatism in Hole U1309D. In particular, late leucocratic dike injection and crystallization and late crystallization in the oxide gabbros appear to be continuous across the transition between fluid-rich late magmatism and amphibolite (and, ultimately, greenschist)-facies hydrothermal metamorphism.

The dominant metamorphic event in Hole U1309D is greenschist-facies hydration of gabbroic, ultramafic, and diabasic rocks. In the upper 350 m of the hole, this alteration is nearly pervasive and does not generally reflect focused fluid flow. With increasing depth, greenschist-facies hydration is almost always related to veins, igneous contacts, and other fluid conduits.

In all pyroxene-bearing rocks, the greenschist-facies event is reflected by partial to complete replacement of pyroxene by actinolitic amphibole, sometimes with subsidiary formation of sodic plagioclase and chlorite. This replacement event is sometimes static, especially in diabase and in the uppermost core, and sometimes clearly related to cataclasis and/or veining (gabbros and the core below 350 mbsf).

In olivine gabbros and troctolites, the greenschist event is most prominently manifested by corona-forming reactions between olivine and plagioclase. In their simplest guise, the coronas consist of a rim of chlorite around plagioclase and a rim of tremolite around (formerly) adjacent olivine. In more complex manifestations, talc and other minerals may form within the coronas. Cummingtonite has been tentatively identified in some coronas, suggesting some form at amphibolite-facies conditions. Unlike other greenschist-facies rocks in Hole U1309D, the formation of coronas appears to be nonisochemical, requiring addition of silica and/or removal of lime. Talc-tremolite-chlorite alteration zones formed at ultramafic/gabbro contacts suggest that these metasomatic rocks may simply reflect exchange of fluids between gabbros and ultramafic rocks.

In ultramafic and near-ultramafic rocks (dunites and olivine-rich troctolites), the dominant low-temperature alteration event is serpentinization. This is generally accompanied by local alteration of plagioclase to prehnite and hydrogarnet. Relict olivine in corona-textured rocks, as well as isolated olivine grains in gabbros and troctolites, may also be serpentinized.

The lowest temperature of metamorphism is manifested by zeolite veins in the lower (below 700 mbsf) core and by various carbonate and clay veins throughout the core. These may be forming under ambient conditions (e.g., ~120°C measured at bottom of hole).

The degree of alteration downhole changes significantly, decreasing between ~375 and 400 mbsf in the last few cores recovered during Expedition 304. Alteration increases again with depth until 841 mbsf. The interval of freshest rock recovered during Expedition 305 is from 841 to 1094 mbsf. Between 1095 and 1200 mbsf, a zone of high alteration reflects highly deformed, localized zones of protomylolite and rodingitization within the olivine-rich troctolites. Below 1094 mbsf, the gabbros and troctolites again are characterized by low to moderate degrees of alteration.

The overall trends in alteration and the changes in secondary mineralogy suggest that there may be two separate secondary processes at work in Hole U1309D. In the upper 841 m, we suggest that a seawater–rock interaction may pervade the sequences, finding downward access along high-angle structural pathways. If an “alteration front” is present in the in the rocks at Site U1309, we suggest that it reaches a depth of 841 mbsf. Below that depth, the nature of and the fluctuations in degree and style of metamorphism are related to fluids of a different composition percolating along fault zones and zones of deformation. We suggest that the fluid composition records an extensive history of rock–fluid interaction, may include magmatic fluids, and is reflected in the secondary mineralization of the rocks. In essence, alteration below 841 mbsf records the cooling of a relatively isolated magma body.