Previous   |    Next

doi:10.2204/iodp.proc.335.204.2016

Methods

Plagioclase and clinopyroxene CPOs were measured using the electron backscatter diffraction (EBSD) technique (e.g., Prior et al., 2009) using two scanning electron microscopes (SEM) at the University of Montpellier, France: a JEOL JSM-5600 and CamScan 500XE crystal probe. Both systems are equipped with Oxford/Nordlys EBSD detectors; the diffraction patterns were collected and processed using the Channel 5 suite of programs. The JSM-5600 and crystal probe were used at accelerating voltages of 15 and 20 kV, respectively, and a working distance of 25 mm. Crystallographic orientation maps were obtained for each sample, scanning the thin section by moving the stage with step sizes ranging from 21 to 60 µm. The step sizes are at least 10 times smaller than the grain size of the phases of interest. The indexing rate (fraction of patterns that are automatically indexed during mapping) ranges from ~50% to 80% in the raw maps. We use the HKL reference data files to index olivine, clinopyroxene, and orthopyroxene and an in-house (Géosciences Montpellier, France) reference data file to index plagioclase. The raw data contain all indexed pixels with a mean angular deviation (MAD; i.e., the angle between the acquired diffraction pattern and the indexing solution proposed by the software) of <1.3°. The first stage of postacquisition data processing was done using the Tango software of the Channel 5 suite to increase the quality of the maps, consisting of removing isolated pixels that are either nonindexed or indexed as a given phase and surrounded by pixels indexed for another phase and filling nonindexed pixels that have a minimum of 5 neighboring pixels with the same orientation.

EBSD data sets were then processed using MTEX (version 4.0.23), a free Matlab toolbox for analyzing and modeling crystallographic orientation (http://mtex-toolbox.github.io) (Hielscher and Schaeben, 2008; Bachmann et al., 2010). We used MTEX to identify grains and produce maps from the EBSD data, calculate pole figures of the plagioclase and clinopyroxene preferred orientation, analyze the crystallographic misorientations within grains, and calculate CPO strength and shape indexes.

Grains were identified from the EBSD data by choosing a 10° threshold. If the misorientation between two adjacent pixels of the same phase is >10°, then it is assumed a grain boundary is present. Grains that have a surface <5 pixels could be erroneous measurements and were removed from the data set. Twins in both plagioclase and clinopyroxene were distinguished from grain boundaries by filtering out the 178°–180° misorientations in grain boundary identification. Pole figures were calculated using both the grid data set from EBSD mapping (grid data) and the average crystallographic orientation for each grain (grain data). The second option is preferred to avoid the over-representation of larger grains when the grain size distribution is heterogeneous at the thin section scale.

The CPO strength for each phase is determined using both the orientation distribution function (ODF) J-index exclusively based on crystallographic orientations (e.g., Bunge, 1982; Mainprice and Silver, 1993), and the M-index (Skemer et al., 2005) based on the misorientation angle distribution across a sample. J-index values vary between 1 (for a uniform distribution) to infinity (for a single crystal); M-index values vary from 0 to 1 (see Mainprice et al., 2014, for the details of J- and M-index calculations and for a comparison between these two indexes). The ODF was calculated using the “de la Vallee Poussin” kernel with a half-width of 10° (Schaeben, 1999; Mainprice et al., 2014).

Symmetry of the CPOs is determined using the BA- and BC-indexes, which are calculated from the point (P), girdle (G), and random (R) indexes, which are calculated from eigenvalues (λ₁ ≥ λ₂ ≥ λ₃, with λ₁ + λ₂ + λ₃ = 1) of the orientation tensor for each pole figure (Vollmer, 1990; Satsukawa et al., 2013; Mainprice et al., 2014):

and

In a plagioclase CPO resulting from magmatic flow, foliation is classically marked by a preferred orientation of [010] planes and lineation by a preferred orientation of [100] axes (Satsukawa et al., 2013). The BA-index is 0 for a perfect axial-B CPO, an oblate (planar) fabric, and is 1 for a perfect axial-A CPO, a prolate (linear) fabric. With pyroxene, the results are similar, except that magmatic lineation is marked by the preferred orientation of [001] axes. Hence, we use the BC-index to characterize the variations between a perfect oblate (BC = 0) and a perfect prolate (BC = 1) CPO.

In EBSD maps produced using MTEX, the misorientation within the grains is quantified using two parameters, the misorientation to the mean (M2M), and the kernel average misorientation (KAM). The M2M gives, for each pixel within a grain, the angle between the crystallographic orientation of that pixel and the average crystallographic orientation of the grain. The M2M allows for the visualization of the misorientation between domains separated by subgrain boundaries or the progressive misorientation related to undulose extinction in optical microscopy. The KAM is, for each pixel, the average misorientation (<10°) of the nearest N pixel neighbors within a crystal (N = 4 here; see Wright et al., 2011, for a review of misorientation parameters).

Top of page   |    Previous   |    Next