IODP Proceedings    Volume contents     Search



Petrographic results

Petrographic details of the samples are shown in Table T2. They represent granoblastic hornfels with a groundmass composed of microgranular plagioclase, clinopyroxene, orthopyroxene, and Fe-Ti oxides. In some samples, secondary amphibole was identified, always in small amounts (<5 vol%). The goal of this study was to characterize the compositions of those minerals in typical granoblastic dikes, showing specific petrographic features observed on board during Expedition 335 (Teagle, Ildefonse, Blum, and the Expedition 335 Scientists, 2012):

  1. All samples except one (Sample 335-U1256D-Run12-RCJB-Q) contain phenocrysts of plagioclase integrated into the granoblastic matrix (Figs. F1, F2, F3, F4). These structures were previously interpreted as relict phenocrysts from the primary magmatic stage that survived the granoblastic overprint (Teagle, Ildefonse, Blum, and the Expedition 335 Scientists, 2012).

  2. Two samples (Run12-RCJB-B and Run15-EXJB) contain plagioclases that display very patchy zoning as a result of the presence of dusty ghost cores (Fig. F5).

  3. Four samples (Run12-RCJB-B, Run12-RCJB-Q, Run12-RCJB-S, and Run15-EXJB) display domains where clinopyroxene and orthopyroxene are arranged in vein- or bandlike structures or in monomineralic clusters with much larger grain size than the granoblastic matrix (Fig. F6). Some of these samples are foliated, whereas the majority of the samples are not.

Microanalytical results

Compositions of the analyzed phases (plagioclase, clinopyroxene, orthopyroxene, amphibole, and Fe-Ti oxides) in the investigated granoblastic dikes are shown in Table T3. To obtain a general overview of the mineral compositions, several measurements were collected from all minerals present in individual granoblastic dikes. Whenever possible, individual phases were analyzed in core and rim areas to detect zonation. If an individual sample contained specific domains like veins, bands, or clusters, analyses of these domains were obtained separately and listed in Table T3. In the following, we focus on those features highlighted in the previous section.


Composition of matrix and phenocrysts

The cores of matrix plagioclases show anorthite (An) contents varying in a small interval between 50.1 and 54.7 mol% (Fig. F7). Only one sample, 335-1256D-235R-1W-19, contains matrix plagioclase with core compositions significantly higher in An (68.5 mol%; Table T3), due to relics of primary An-rich plagioclase. Rim compositions are very similar to those of the cores, varying between 47.5 and 56.4 mol%, and confirming the petrographic result that most granoblastic dikes bear matrix plagioclases without significant zoning. Only a few samples contain matrix plagioclases with characteristic patchy zoning due to the presence of ghost cores (see below). In general, the compositions of the matrix plagioclase overlap with those from granoblastic dikes recovered during Expedition 312 (Koepke et al., 2008) (Fig. F7).

In contrast to the matrix plagioclases, the plagioclase phenocrysts have much higher An contents between 65.5 and 72.7 mol% (core analyses). They show compositions typical of plagioclases from the fresh pillow basalts and dikes from the upper part of Hole 1256D (Fig. F7; data from Dziony et al., 2008), implying that these phenocrysts correspond to relics inherited from the primary magmatic stage. The inherited magmatic cores can be clearly identified in backscattered electron (BSE) images by their euhedral shapes showing zoning with very a sharp boundary against the granoblastic matrix (Figs. F1, F2, F3, F4). To determine the variation in An content throughout the phenocryst zonation, analytic profiles were measured through the phenocrysts including the adjacent matrix for four samples. The results are presented in Figures F1, F2, F3, and F4. By obtaining very accurate CaAl-NaSi concentration profiles at the rims of such phenocrysts, Zhang et al. (2014) applied diffusion modeling techniques and extracted cooling rates for the magma chamber roof rocks at Site 1256. Their results show that cooling from the peak thermal overprint at 1000°C–1050°C to 600°C are yielded within ~10–30 y as a result of very effective hydrothermal circulation above the melt lens during a phase of magma starvation.

Composition of the ghost cores

Matrix plagioclases with typical patchy zoning were analyzed in two samples (Table T3). The BSE images reveal the presence of cores enriched in An content with a sharp boundary against an interstitial network composed of plagioclase lower in An (Fig. F5). The dusty appearance of these cores is due to millions of micrometer-sized oxide inclusions. A microprobe profile of An content through a typical core is also shown in Figure F5. The plagioclase cores show An contents varying between 49.4 and 53.9 mol%, corresponding to the composition of matrix plagioclases in granoblastic dikes (see above). The composition of the interstitial network is strongly enriched in albite component (~30 mol% on average), implying that these compositions are due to a secondary hydrothermal overprint at lower temperatures. This hypothesis is supported by a relatively high amount of secondary amphibole in one sample (335-U1256D-Run12-RCJB-B).


Composition of matrix and bands/clusters

Clinopyroxenes of the granoblastic matrix show relatively low TiO2 and Al2O3 contents (0.3–0.6 and 0.8–1.4 wt% for core analyses, respectively) and correspond to compositions typical of granoblastic dikes recovered during Expedition 312 (Fig. F7). They strongly contrast with compositions from the primary magmatic stage analyzed in fresh lavas and dike of Hole 1256D, which show high Al2O3 contents (Fig. F7; data from Dziony et al., 2008), and also high Cr2O3 contents, whereas the Cr2O3 concentrations in clinopyroxenes from the granoblastic dikes are always below the detection limit. All analyzed orthopyroxenes vary only over a small range, with relatively low TiO2 and Al2O3 contents (0.2–0.4 and 0.6–0.9 wt%, respectively) that are very similar to orthopyroxenes analyzed in granoblastic dikes recovered during Expedition 312 (Fig. F7).

Clinopyroxenes and orthopyroxenes analyzed in special domains like veins, bands, and clusters (Fig. F6), or as inclusions in other minerals, show compositions overlapping widely with those of the matrix pyroxenes (Fig. F7), implying that a high degree of chemical equilibrium has been achieved in these samples during the metamorphism. This supports the hypothesis, based on microscopic observation, that these special domains are derived from precursor structures generated by pervasive hydrothermal alteration, such as hydrothermal veins and patches in the basalts. These were subsequently transformed by the contact metamorphic overprint into the granoblastic paragenesis, occurring as textural domains such as bands and clusters, but with mineral compositions identical to the granoblastic matrix, as a consequence of achieving a global chemical equilibrium in these rocks.


All analyzed amphiboles show compositions with Na + K on A position below 0.5 corresponding to magnesiohornblendes and actinolites, which is typical for an equilibration within the amphibolite facies. Thus, the conditions of formation do not correspond to those of the peak metamorphism in the pyroxene hornfels facies (see below), implying that the amphiboles were formed during a later hydrothermal overprint after the high-temperature metamorphism. This in agreement with the observation that most of the amphiboles correspond to overgrowth or late veins, and that the amphibole-bearing samples show a relatively high amount of secondary, albite-rich plagioclases, as well as coexisting titanite.


To calculate equilibrium temperatures, two different geothermometry techniques were used. We applied the two-pyroxene geothermometer using QUILF (Andersen et al., 1993) and equation 36 of Putirka (2008), as well as the semiquantitative Ti-in-amphibole geothermometer of Ernst and Liu (1998). The application of the latter is possible due to the general presence of a Ti phase coexisting with the amphibole (mostly Fe-Ti oxide and/or titanite).

The results of these calculations are presented in Table T4. Both two-pyroxene geothermometers revealed nearly identical results in the six samples for the peak metamorphism equilibrium temperatures of the granoblastic overprint, varying only within a relatively small temperature range of 930°C–990°C for QUILF and 930°C–980°C for the Putirka (2008) method. These temperatures are characteristic of the pyroxene hornfels metamorphic facies, and are very similar to equilibrium temperatures estimated by QUILF for the granoblastic dikes recovered during Expedition 312 ranging from 930°C to 1050°C (Koepke et al., 2008). These temperatures, together with the specific geological setting of these rocks within the gabbro–dike transition of fast-spreading crust, imply that the granoblastic hornfels can be regarded as preserved fragments of the CBL.

Alt et al. (2010) presented slightly lower equilibrium temperatures in granoblastic dikes recovered during Expedition 312, ranging from 800°C to 920°C. However, they used the Brey and Köhler (1990) geothermometer, which results in systematically lower temperatures compared to the QUILF two-pyroxene thermometer (see discussion in Alt et al., 2010). Our estimated temperature range is also in agreement with temperature estimations for hornfels from the gabbro–dike transitions for the EPR crust at Pito and Hess Deeps and in the Oman and Troodos ophiolites, where estimated peak temperatures vary from 700°C to 1000°C (Gillis, 2008).

Distinctly lower equilibrium temperatures were derived from the Ti-in-amphibole geothermometer of Ernst and Liu (1998), ranging between 510°C and 630°C for hornblende and actinolite veins or overgrowth within the granoblastic dike. We interpret this as the record of a secondary metamorphic overprint under hydrous conditions in the amphibolite facies, probably related to hydrothermal alteration caused by seawater ingress. One hornblende forming blebs in orthopyroxene in Sample 335-U1256D-Run12-RCJB-Q revealed a distinctly higher equilibrium temperature of 853°C (Table T4). We attribute this temperature to a retrograde cooling stage under hydrous conditions within the hornblende hornfels facies.