Application Notes
Published: 07 Dec 2021 · Last updated: 07 Dec 2021
Tags: EBSD, EDS
Geological research is frequently focused on understanding the formation and history of rocks now present at the surface of the Earth. A significant branch of geology involves the study of metamorphic rocks: these are rocks that have been affected by deformation, heat and high pressure, usually as a result of larger scale tectonic processes. In order to reconstruct the history of such rocks it is necessary to identify the presence of key metamorphic minerals and to characterise their relationship with each other and with microstructures that have been modified by deformation processes. In this way geologists can reconstruct the pressure (P), temperature (T) and deformation (D) history of rocks and improve understanding of past tectonic processes. Such an understanding can help with the search for economically important ore bodies as well as improve our knowledge about modern day plate tectonics-related events such as earthquakes and volcanoes.
Conventional metamorphic petrology (the study of metamorphic rocks) typically uses a combination of optical microscopy and chemical analysis techniques (such as energy dispersive X-ray spectrometry – EDS – or electron probe micro-analysis – EPMA). The examination of polished thin sections enables polarised light microscopy (for rapid mineral identification) to be coupled with EPMA (for quantitative chemical analyses), providing a comprehensive petrological characterisation of metamorphic rock samples. However, some key metamorphic minerals are polymorphs: they exist in different crystallographic forms with identical chemistry, with their stability linked to pressure and temperature conditions. For example, the mineral coesite (SiO2) is a high-pressure polymorph of quartz (also SiO2) and so the presence of coesite in a rock indicates that it has been subjected to pressures exceeding 2.5 GPa (equivalent to depths > 70 km below the Earth's surface). Similarly, the minerals sillimanite, andalusite and kyanite are all polymorphs of Al2SiO5, but are stable at different temperatures and pressures (see figure 1). In samples where these minerals are coarse-grained, they can be relatively simple to identify using well-established polarised light microscopy techniques. However, in cases where such key minerals are finer grained or are wholly enclosed within other minerals, it can be very challenging to correctly identify them using light microscopy and, having identical composition, chemical analysis techniques cannot help.
In this study of a complex metamorphic rock sample, the combined power of electron backscatter diffraction (EBSD) and EDS is demonstrated, illustrating how EBSD can assist not only in the correct identification of key minerals, but also in determining the complex relationships between deformation episodes and the formation of metamorphic minerals.
Fig 1. P-T chart showing the stability fields of the Al2SiO5 polymorphs, andalusite, kyanite and sillimanite.
The sample is a metamorphosed pelitic gneiss collected from the Leverburgh Belt, South Harris, in Northwest Scotland. These basement rocks have been dated using geochronology techniques as ~2.8 Ga old, with a complex metamorphic history indicating multiple episodes of deformation and metamorphism between 2.5 and 1.7 Ga before present [1–3].
The sample was prepared into a polished thin section, with a final polishing step using colloidal silica suspension. A thin (~10 nm) carbon coat was applied prior to analysis in a field emission gun scanning electron microscope (FEG-SEM).
The sample was characterised with combined EDS and EBSD, using a Symmetry S2 EBSD detector and an UltimMax 170 EDS detector. Data were acquired using the AZtec v5.0 software, and additional data processing was carried out using AZtecCrystal v2.1. Initial targeted EDS and EBSD maps were used to identify the main constituent minerals and to determine a suitable region of interest. Subsequently a large area map covering ~25 x 25 mm was collected using a measurement step size of 4 μm: the analysis, comprised of 480 individual maps, was carried out using a beam current of 24.9 nA and at an analysis speed of 250 patterns per second. EBSD patterns (622 x 512 pixel resolution) were indexed using 12 mineral phases, and EDS data were collected using process time 3 at an input count rate of ~140k counts per second.
The large sample size necessitated working at a long working distance (WD) in the SEM (~29 mm): to maintain optimum EBSD geometry, the Symmetry S2 detector was lowered 10 mm from its standard geometry using its software-controlled elevation control. This ensured high quality EBSD patterns as well as minimal occlusion of the EDS detector.
The area of analysis displayed a coarse gneissic foliation and included 7–8 large garnet grains that appeared significantly fractured, surrounded by mica-rich regions and separated by bands of quartz and plagioclase feldspar (see figure 2). In this application note we focus primarily on one of the garnet grains, as highlighted in figure 2.
Figure 2. Left: montaged Z-contrast BSE image collected using the upper forescatter detectors. Right: EBSD phase map, with major phases including garnet (red), quartz (blue) and plagioclase feldspar (orange). Field of view ~25 mm. Yellow box marks area displayed in greater detail in figure 3.
Fig 3. EBSD phase map showing the garnet cluster highlighted in figure 2. Major mineral phases are labelled as follows: Grt – garnet, Pl – plagioclase feldspar, Qz – quartz, Bt – biotite, Ilm – ilmenite, Ky – kyanite, And – andalusite, Sil – sillimanite. Field of view ~10 mm.
As expected, the EBSD phase map (figure 3) of this single garnet cluster shows 2 fractured garnet grains surrounded by bands of quartz and plagioclase, with biotite (mica) within both the fractures and the pressure shadows on either side of the garnet. However, the phase map also highlights the presence of all 3 Al2SiO5 polymorphs in association with the garnet – these Al-rich phases are clearly visible in the EDS element maps (figure 4) and verification of the EBSD identification of these phases is provided in figure 5, showing reliable indexing of EBSD patterns from all 3 Al2SiO5 phases within this area. The sillimanite grains are generally elongate, showing some preferred grain alignment, and are commonly enclosed within the garnet. The kyanite grains wrap around the garnets, with the andalusite grains generally found within the fractures between the fragments of garnet grains.
It is important to note that previous metamorphic petrological analyses of similar rock samples from the Leverburgh belt have never recorded the presence of andalusite, showing the advantage of the more rigorous phase identification permitted by the EBSD technique, especially on small grains and between phases with identical chemical characteristics. The spatial associations between these key minerals allow us to determine their order of crystallisation, with sillimanite forming first (possibly replacing earlier andalusite), followed by growth of the garnet grains (which enclose the pre-existing sillimanite grains), then kyanite and finally, as the garnet fractures and starts to decompose, the andalusite.
Many of the minerals in this sample show evidence for extensive deformation. The elongate sillimanite grains have a preferred alignment (figure 6), prior to being enclosed within the later-growing garnets, indicative of a pre-existing foliation from an earlier deformation episode. It is also noticeable that both the kyanite and sillimanite have very strong preferred rotation axes across their low angle boundaries (figure 7) caused by deformation in a dislocation creep regime. These rotation axes likely result from dominant slip on (010)[100] in sillimanite and (100)[001] in kyanite, suggesting that both these minerals were present during a later deformation episode [4, 5]. In contrast, the andalusite grains show no preferred rotation axis orientation and exhibit only a few low angle boundaries, suggesting their crystallisation in the latter stages of or following this late deformation episode.
The chemical data from simultaneous EDS spectra collection enable detailed element distribution maps to be plotted, as in figure 4. As each point in the map contains the full X-ray spectral information, high quality spectra can be extracted from the dataset for more rigorous quantitative analyses. The Mg distribution map suggests that the individual garnet fragments may be zoned and a transect across one such garnet has been quantitatively processed, with the results shown in figure 8.
Figure 4. Corresponding EDS element maps, showing Fe, Al, Mg and Na distributions. The Al element map is annotated to show the locations of the different Al2SiO5 polymorphs.
Figure 5. Indexed EBSD patterns from the 3 polymorphs of Al2SiO5, all collected from the area shown in figure 3: kyanite (left), sillimanite (centre) and andalusite (right).
There is a clear enrichment of Fe near the rims of the garnet, with a corresponding depletion of Mg; since each garnet fragment shows this chemical trend, the zoning must have occurred after the initial garnet growth and also after the subsequent fracturing and decomposition of the garnet. The presence of the andalusite within the fracture zones suggests that the rock was still at a relatively high temperature (but low pressure) at this stage, which would enable chemical modification via lattice diffusion processes (probably enhanced by the presence of Fe-enriched fluids).
Figure 6. Chart showing the orientation of the long-axes of sillimanite grains (with an aspect ratio >3) across the whole large map area.
Figure 7. Localised grain relative orientation deviation (GROD) maps and low angle disorientation rotation axis plots (1–2°, 2–5° and 5–10°) for kyanite (top) and sillimanite (bottom). Both minerals exhibit extensive plastic deformation (up to ~15° orientation change) and strongly clustered rotation axes (about <010> in kyanite and <001> in sillimanite).
Figure 8. Zoomed section of the Mg element map (see figure 4), highlighting chemical zoning within the remnants of the garnets. The yellow line marks the location of a transect of extracted EDS spectra across one garnet grain. Below: quantitative EDS results (in oxide %) across the transect, showing the clear chemical zoning of MgO and FeO.
Figure 9. Proposed P-T evolution based on the EBSD/EDS analysis of this pelitic gneiss sample. Note the peak metamorphic conditions (between steps 3 and 4) cannot be constrained from the results presented here.
A single large-area combined EDS and EBSD analysis can provide an extraordinary amount of information about the pressure, temperature and deformation history of a complex metamorphic rock, such as the pelitic gneiss examined in this study. In this example, EBSD has identified the extensive presence of andalusite, undetected in numerous previous studies of rocks from the same metamorphic complex. In addition, the quantitative information regarding phases, orientations, grain shapes and sizes, deformation processes and chemical zoning permits a comprehensive interpretation of the metamorphic evolution of the rock. This has been summarised in figure 9.
Further detailed information relating to the P-T conditions and the dates of different stages in the evolution of this rock can be extracted by an in-depth (ion) microprobe analysis of key accessory minerals, such as zircon, rutile and monazite. This single dataset has identified 39 zircon grains, 34 monazites and 56 rutiles (all over 25 μm diameter, large enough for an accurate chemical analysis), further demonstrating the value of a single high-quality EBSD and EDS analysis to metamorphic petrology research.
Oxford Instruments thanks Professor Sandra Piazolo (University of Leeds) for providing the sample and assisting with interpretations of the data.
© Oxford Instruments 2026