Orthogonal switching of AMS axes during type-2 fold interference: Insights from integrated X-ray computed tomography, AMS and 3D petrography

Abstract


Introduction
X-ray computed micro-tomography (µ-CT) is increasingly being applied in structural geology and ore petrology due to its ability to resolve the three-dimensional (3D) shape and spatial distribution of minerals and associated textures in metamorphic rocks (e.g., Sayab et al., 2015;Macente et al., 2017).
Sulfides and oxides yield brighter gray values than rock-forming silicates owing to high X-ray M A N U S C R I P T A C C E P T E D ACCEPTED MANUSCRIPT 2 attenuation or higher density.They can be segmented or separated, for example, using the threshold limits of the gray level histogram, and rendered and quantified in 3D volume (e.g., Sayab et al., 2016a;Hanna and Ketcham, 2017).The resulting 3D views provide a spatial distribution of high-density ore minerals (ranging from ca. 4 to 22 g/cm 3 ) inside a given rock specimen, which is otherwise difficult to determine using serial thin sectioning methods (Hayward, 1990;Aerden, 2003;Bell and Bruce, 2006).
The µ-CT technique is non-destructive and can be integrated with a number of conventional and unconventional microstructural methods to investigate regions of interest in extraordinary detail without destroying the whole sample (Kyle and Ketcham, 2015).As a holistic 3D visualization method of analyzing rock fabrics (Sayab et al., 2015), µ-CT is ideal for integrating with measurements of bulk rock anisotropy of magnetic susceptibility (AMS).
AMS measures the directional variations of magnetic susceptibility (K) present in a rock sample, and the AMS magnitude ellipsoid defines the bulk rock orthogonal principal susceptibility axes (K1: maximum, K2: intermediate, K3: minimum; Tarling and Hrouda, 1993).These axes can be correlated with the finite strain axes (X ≥ Y ≥ Z) and structural features, for example mineral stretching and intersection lineations, and, therefore, provide a mean to study deformation in 3D (e.g., Borradaile, 1988;Riller et al., 1996;Ferré et al., 2014;Parsons et al., 2016).However, since AMS measures the sum of different magnetic carrier minerals in a sample, its interpretation requires knowledge of amounts and relative abundances of these minerals, their magnetic susceptibilities, shape preferred orientation (SPO), lattice preferred orientation (LPO) and relationship with tectonic fabrics (e.g., Kruckenberg et al., 2010;Borradaile and Jackson, 2010).Thus, a good understanding of the internal microstructure behind AMS measurements is crucial.High-resolution µ-CT can fulfill this requirement by allowing 3D visualization and determination of statistically volumetric abundances of magnetic carrier minerals and their SPOs.We have explored this possible nexus using an oriented drill core of a pyrrhotite rich micaschist sample that preserves two tectonic foliations.Apart from applying AMS and µ-CT techniques, we studied the sample in three perpendicular thin sections oriented with reference to the matrix foliation.Through this integrated approach, the deformation history of the sample is reconstructed and considered in the regional tectonic framework as revealed by high-fidelity

M A N U S C R I P T A C C E P T E D ACCEPTED MANUSCRIPT
3 aeromagnetic image (Fig. 1a).The application of non-invasive (AMS and µ-CT) followed by invasive (thin sections) methods illustrated herein offers a means of determining the magnetic lineation and foliation in ore-bearing rocks, which is highly relevant to the exploration of structurally controlled mineral deposits as will be briefly discussed.

Tectonic setting and sample description
The Martimo metasedimentary belt forms a part of the Peräpohja region, which lies in the center of the Fennoscandian shield.The area is characterized by lack of topography, where structural field relations in the 3 rd dimension are difficult to reconstruct.Five incremental, but heterogeneously distributed deformation events have been proposed for the Martimo belt in a timeframe ranging from 1.92 to 1.77 Ga (Lahtinen et al., 2015).The belt hosts metapelites interbedded with metapsammites and preserves a sequence of near-orthogonal tectonic fabrics that are not fully developed elsewhere in the region (Lahtinen et al., 2015).In particular, the area is passing through a period of extensive geological research because of the recent discoveries of Rompas Au-U and Rajapalot Au (e.g., Nykänen et al., 2017;Ranta et al., 2015;Ranta et al., 2016).Metamorphism in the area is characterized by upper greenschist facies, however, the grade varies from low greenschist facies rocks in the southwest to kyanite bearing migmatites in the northeast (Hölttä and Heilimo, 2017).
An oriented sample (sample 57;Fig. 1;66° 16' 53.43" N,24° 29' 24.69" E) was drilled from the middle of the Martimo belt.The drill core sample is characterized by two near-orthogonal foliations that correspond to the D1 and D2 deformations of Lahtinen et al. (2015).S1 is subparallel to S0 (S0//S1) and on average strikes N-S.It is folded with E-W striking axial planes and gently east plunging axes, and overprinted by a steeply dipping, finely-spaced S2 foliation (Fig. 1b).The drillcore sample consists of pyrrhotite-pyrite-quartz mica schist with two distinct foliations (S1 and S2), and thus ideal for integrated µ-CT and AMS studies.Before extracting it from the bedrock using a portable mini-drill, the sample was marked with a north pointing groove on the top surface so that it could be reoriented in the lab.The sample has a 2.5 cm diameter and is 8 cm long.

Scanning configuration and processing
The sample was scanned using the high-resolution Phoenix|X-ray Nanotom 180 (GE) scanner at the University of Helsinki.Before scanning the drill core, two brass pins were attached, about 40° from north counting towards the east for reference.The tungsten target X-ray tube was operated with an acceleration voltage of 140 kV and tube current 175 µA.The radiation was filtered with 1.0 mm of Cu.Six hundred views per 360° were acquired, with 1.75 s total exposure time per view.The virtual 3D volume is composed of 1152 horizontal slices with a voxel size of 33 µm × 33 µm × 33 µm.The projections were captured with a Hamamatsu flat panel sensor C7942SK-05.The 3D reconstructions were computed with datos|x -reconstruction software provided by Phoenix|X-ray.

3D visualization
After reconstructing the 3D distribution of X-ray attenuations, the contrasting gray shades of pyrrhotite and pyrite allowed us to semi-automatically separate them from the rest of the silicate matrix by mapping gray levels to specific colors using Avizo software (http://www.fei.com/software/avizo -3d -for -materials-science/). Volume rendering further allowed to visualize the dataset in 3D (Figs. 2 and 3).Besides looking at the microtextures in 3D in all directions, it is possible to digitally cut the sample at any angle and to make certain surfaces transparent while rendering others opaque for better visual analysis.A virtual 3D tour inside the sample, at 33 µm voxel resolution, allowed us to closely examine petrographic details and the spatial distribution of sulfides.
To show the penetrative nature and spatial distribution of pyrrhotite and pyrite grains, the drill core was virtually chopped at different horizontal (Fig. 2) and vertical (Fig. 3) levels.The 3D images clearly show that pyrrhotite is present throughout the core but localized within millimeter-scale metasedimentary laminae or bands (S0).Pyrite porphyroblasts are mainly localized along layers in the center (Figs.3c,e).For better visualization from the top, two such layers were isolated and rendered in red and blue (Fig. 3g).The fold shape, defined by these layers, match the fold pattern observed in outcrop (cf.Fig. 1b).
Most of the pyrrhotite are standalone grains in the 3D visuals, whereas some pyrite grains are intergrown or touching each other (Figs.2a and 3e).In horizontal and E-W striking vertical gray slices (Figs.3a,f) as well as the 3D rendered images (Figs.3c,e), pyrrhotite grains mimic the bulk finite Turning the core (~90°) and looking west, the pyrrhotite grains appear flattened or oblate in the Y-Z plane (Figs.3i,j; see also 7f).

Segmentation of pyrrhotite and pyrite
Segmentation of pyrrhotite and pyrite grains was performed with ImageJ and in-house written software as follows.In order to reduce imaging noise, the reconstructed tomographic image was filtered with variance weighted mean filter (noise variance = 500, radius = 2 pixels; Gonzalez and Woods, 2002).After filtering, contrast to noise ratio between mineral crystals and background was high (>~8), and thus the crystals could be segmented using simple thresholding.The threshold value was determined using Tsai's moment preserving thresholding method (Tsai, 1985).After thresholding, the image contained pyrrhotite and pyrite grains as white regions, and additionally some small white regions caused by imaging noise that is typically found in tomographic images and caused by various processes such camera noise and variance in the count of X-ray photons received by the camera (Stock, 2008).Regions corresponding to imaging noise were suppressed by removing all white regions whose volume was less than 20 voxels, a cut-off that seemed to remove noise regions efficiently but preserve even the smallest crystals.Finally, pyrite and pyrrhotite grains were separated manually based on the characteristic cubic or parallelepipedic shape of pyrite.
Pyrrhotite grains were separated from each other using watershed segmentation (Meyer and Beucher, 1990) seeded with local maxima of distance map.In order to avoid over-segmentation near locations with multiple nearby local maxima, only the largest of them was taken to be a seed point.
The resulting segmentation was checked by visually comparing the segmented regions to the original data.The quality of the segmentation was found to be satisfactory.Each segmented grain or crystal was analyzed separately.Grain volumes were calculated as the number of voxels in the grain.The orientations of long (X), intermediate (Y) and short (Z) axes were determined by counting voxels along three mutually perpendicular principal axes within the grain (Pearson, 1901).The aspect ratio was determined as the ratio between the lengths of the X and the Z axes, and the Flinn parameters were calculated as ratios between the X and Y, and Y and Z, respectively.

M A N U S C R I P T A C C E P T E D ACCEPTED MANUSCRIPT
6 Finally, the orientation data for all grains were plotted in rose diagrams for the horizontal plane and for vertical planes (Figs. 4a,b and 5a,b).Pyrite grains were separated from each other and analyzed similarly as pyrrhotite grains, but applying a closing operation with radius of 5 pixels before watershed segmentation.A flowchart summarizing the image processing operations is shown in Appendix 1, and provided as an electronic supplement.

3D quantitative analysis
We collected quantitative data for volume, surface area, aspect ratio (X/Z and Y/Z), and orientation of pyrrhotite and pyrite grains.The volume fraction in the whole rock sample obtained through full segmentation of pyrite and pyrrhotite is 3.5±0.3%and 4±2%, respectively, where the uncertainty estimates have been made according to Fusseis et al., (2012).A total of 25,920 pyrrhotite and 155 pyrite grains were encountered but grains touching the edges or boundary of the sample were discarded due to incomplete size and axes attributes.
Pyrrhotite X-axes plunge shallowly (~ 22°) towards the east, whereas metasedimentary layers (S0) dips about 45° in the same direction b).The mean aspect ratio and volume of the pyrrhotite grains are ~14 and 0.019 mm 3 , respectively (Figs. 4c,e).A small fraction of pyrrhotite grain (about 4%) have a much higher aspect ratio as others reaching a maximum value of 292 (Fig. 4c).The Flinn diagram portrays dominantly prolate shapes of pyrrhotite grains, although with a significant fraction of oblate shapes (Fig. 4d).We interpret the latter as representing pyrrhotite microboudins (see micro-textural description below).Aspect ratio increases with increasing grain volume (Fig. 4e).
The horizontal rose diagram for the long axes of pyrite grains (Fig. 5a) shows a wide range of trends with a main E-W maximum and weaker N-S maximum (Fig. 5a) that can be correlated with L 1 2 and L 0 1 , respectively.The rose diagram of Fig. 5b shows that the long axes (X) of pyrite grains generally pitch shallowly in both EW and N-S vertical sections.The 3D images show that the larger pyrite grains are confined to a single layer located in the middle of the sample (Figs.3c,e).The mean aspect ratio of pyrite grains is about 4 with a smaller spread as observed for pyrrhotite grains.The Flinn diagram for pyrite grains indicates a mixture of highly prolate to highly oblate shapes (Fig. 5d), whereas the volume to aspect ratio plot reveals a slightly inverse relationship between grain volume and aspect ratio (Fig. 5e).We now proceed to compare the above described 3D data for the shape and orientation of the two main sulfide minerals in the studied sample with AMS fabrics and microstructural observations in thin section.

Pyrrhotite, AMS and tectonic fabric
Pyrrhotite (density 4.6 g/cm 3 ) is a magnetically ordered (ferromagnetic) mineral that, at ambient temperature, has its atomic magnetic moments arranged at the lattice scale (e.g., Borradaile and Jackson, 2010).Because of this magneto-crystalline property, a correlation between the AMS and rock deformation fabrics has been proposed as a proxy for 3D strain geometry (e.g., Parsons et al., 2016).
The shape of the AMS ellipsoid has been shown to be a function of the deformation history in a rock and the response of minerals to deformation in terms of deformation mechanisms (e.g., Borradaile and Jackson, 2004).
It has been shown that AMS is mainly controlled by ferromagnetic and paramagnetic minerals in a rock, whereas diamagnetic minerals have only a minimal effect (e.g., Borradaile and Jackson, 2010;Ferré et al., 2014).Strong SPO and LPO of ferromagnetic minerals produce a magnetic foliation and lineation that reflect the bulk rock fabric.The principal susceptibility axes (K1, K2, K3) may or may not correspond to finite strain axes (X ≥ Y ≥ Z), and depending on crystallographic properties and 8 deformation mechanism, an inverse or blended relationship can result other than the expected strain axes (Rochette et al., 1999;Ferré, 2002;Borradaile et al., 2012).
To identify the origin of AMS in a given rock sample, both ferromagnetic and paramagnetic minerals can be decoupled by measuring and plotting their thermomagnetic or susceptibilitytemperature (K/T) curves and Curie points (Hrouda et al., 1997).The thermomagnetic curves represent the variation with temperature (heating and cooling effects) of magnetic susceptibility (Hrouda, 1994) and can be used to constrain the mineral species that are the magnetic carriers in a sample (e.g.Skyttä et al., 2010;Karell et al., 2014).A detailed account on the AMS and petrofrabrics of deformed and metamorphosed rocks can be found in seminal reviews by Borradaile andJackson, (2004, 2010).

Analytical methods
For this study, AMS and temperature variation of the magnetic susceptibility (K/T) experiments were performed at the Research Laboratory, Geological Survey of Finland in Espoo.The 2.5 cm × 8 cm (diameter × length) drill core (sample 57) that was partly used for the µ-CT analysis was cut into three pieces of 2.1 cm length each.AMS was analyzed with a KLY-3S Kappabridge, operating at a frequency of 875 Hz and with a field intensity of 300 Am -1 , while rotating each subsample about three mutually perpendicular axes to construct its magnetic susceptibility axes and determine the corresponding magnetic parameters.Thermomagnetic measurements were performed using CS-3 furnace apparatus (Agico, Inc.).The AMS and thermomagnetic data were processed with the Anisoft (version 4.2) and Cureval (version 8) programs, respectively.

AMS results
The mean magnetic susceptibility (Km) of the 3 subsamples has a high value ranging from 17703 to 23308 µSI, whereas the corrected degree of anisotropy (P') varies from 1.81 to 1.94 (Fig. 6).The shape of the AMS ellipsoid (T) is clearly prolate with T values of -0.14 to -0.23.The thermomagnetic curve shows a typical pattern of pyrrhotite with an abrupt drop in magnetization at ~325°C (Hrouda et al., 1997), suggesting that AMS is mainly controlled by this mineral.The orientation of the magnetic foliation matches that of the east-west striking and subvertically dipping S2 foliation (Fig. 6d; cf.Fig. 1b), whereas the short axis (K3) is oriented perpendicular to this plane.The orientation of the triaxial 9 magnetic ellipsoid also coincides with the principal X, Y and Z axes of pyrrhotite determined through the µ-CT (cf.Figs. 3 and 4).

Petrographic and SEM analyses
Three thin sections were cut from the sample normal and parallel to the main S2 matrix foliation and L 1 2 intersection lineation.The horizontal thin section shows a penetrative, highly differentiated and finely-spaced S2 foliation striking east-west (Fig. 7a).The long axes (X) of most pyrrhotite grains are well aligned with this fabric (Fig. 7b).Pyrite grains are wrapped by S2 suggesting pre-or syn-D2 growth (Zwart, 1962;Passchier and Trouw, 2005).No inclusions of pyrrhotite were found inside pyrite, at least in these three thin sections.Some pyrrhotite grains appear microboudinaged in east-west direction (Fig. 7c) with the boudin necks filled with silicate minerals.As mentioned earlier, oblate pyrrhotite grains in the Flinn diagram of Fig. 4d probably represent microboudins of originally larger prolate grain, or alternatively, grains weakly deformed during D2 and retaining their original (syn-D1) oblate shape (Figs.7e,f).
The vertical N-S section cut normal to S2 and the mineral stretching lineation reveals a high-angle relationship between S2 overprinting S1 (Figs. 7d-f) and a stage 4 crenulation cleavage morphology in the scheme of Bell and Rubenach (1983).An important observation in this section is that the Y-axes of pyrrhotite grains are parallel to S1 but truncated, deflected or even folded by S2 (Figs.7e,f) indicating that this mineral grew prior to D2. Close examination of pyrrhotite grains reveals variable shapes ranging from flattened parallel to S1, to oval shaped parallel to S2, to microfolded with axial planes parallel to S2 (Fig. 7f).Pyrrhotite grains with Y-and Z-axes parallel and perpendicular to the S1 foliation, respectively, are relatively less deformed.Whereas, highly deformed pyrrhotite grains have Y-and Z-axes perpendicular and parallel to the S1 in the same thin section.
The section cut parallel to S2 (Fig. 8) shows the finely laminated layers already seen in the µ-CT images (Fig. 3), wherein pyrrhotite grains plunge shallowly (25°) towards the east (Fig. 8d).Pyrite grains have euhedral outlines and are associated with symmetric fibrous quartz strain fringes more or less aligned E-W with the stretching lineation and L 1 2 intersection lineation.

M A N U S C R I P T A C C E P T E D ACCEPTED MANUSCRIPT
10 In order to verify the chemical composition of sulfides, Energy-dispersive X-ray Spectroscopy (EDS) analyses were performed at the Research Laboratory, Geological Survey of Finland in Espoo, using an Oxford Instruments EDS-spectrometer X-Max 80 mm 2 (SDD) attached to a JEOL JSM 7100F Schottky field emission scanning electron microscope (representative FE-SEM images and EDS spectra are provided in Appendix 2 as an electronic supplement).The analytical conditions were as follows: high vacuum mode, a COMPO back-scattered signal (BSE), 20 kV accelerating voltage and 0.5 nA probe current.FE-SEM and EDS analyses showed that almost all pyrite grains are partially to completely pseudomorphed by pyrrhotite (Fig. 9), indicating prograde heating resulting in sulphur release from pyrite (Craig and Vokes, 1993).Skeletal textures mark this pyrrhotite to pyrite replacement (Fig. 9b).Besides, we detected inclusions of euhedral arsenopyrite and anhedral sphalerite around the median and rim regions of pseudomorphed pyrite (Figs.9a,b).EDS analyses, conducted in an FE-SEM, of both intact and microboudinaged grains confirmed their pyrrhotite composition (Figs.9c,d).Some of the cubic-looking grains have rhombic shape possibly due to replacement of diagenetic anhydrite by pyrrhotite (Hall, 1982).

Implications for the D1 related structures
Microstructures commonly exhibit complex overprinting relationships that are only partially reflected by outcrop-scale structures (Aerden, 1998;Sayab et al., 2016b).This counts in particular for Proterozoic belts, where glacially eroded bedrock and sedimentary cover exacerbate field observations in the 3 rd dimension.Under such limited exposure conditions, oriented drill cores are useful to understand multiple fabric overprinting relationships and microstructures (Sayab, et al., 2015).As described earlier, sample 57 was extracted from a D2 low-strain zone of a flat-lying folded surface, where the east-west striking pervasive fabric marks the axial planar S2 foliation.High-resolution tomographic 3D images accompanied by a vertical east-west oriented thin section revealed that metasedimentary layers (S0) dips ~23° more steeply than the S1 foliation, whereas the intersection lineation (L 0 1 ) is north-south trending and shallowly plunging (Fig. 10).This structural relationship suggests that the sample comes from the normal limb of an F1 recumbent fold, in which deformation (Fig. 5b).Pyrrhotite X-axes do not retain a primitive N-S direction, which is expected because as a much weaker mineral, it would have experienced much stronger deformation during D2 as pyrite (cf.Craig and Vokes, 1993).
The shape and spatial distribution of pyrite and pyrrhotite grains could also be influenced by the preferential growth in S1 microlithons or low-strain lenses, given the control of deformation partitioning on metamorphic reactions in general (Bell and Hayward, 1991).Additionally, the growth of sulfide minerals was concentrated in alternate metasedimentary layers, probably because of a chemically more favorable composition (Figs.3c,d,e).
Microtextural relationships show that arsenopyrite and sphalerite are cogenetic with pyrite, whereas the latter was partially to completely replaced by pyrrhotite.Thermodynamic modelling of the arsenopyrite-sphalerite-pyrite and pyrrhotite assemblage in closed-system metapelites (e.g., Lynch and Mengel, 1995) indicate upper-greenschist facies conditions at intermediate pressures of 5.5-6.9kbars, and these P-T estimates can potentially be applied to D1 in the study area.

Implications for the D2 structures
The north-south oriented vertical thin section, cut perpendicular to S2 and normal to L 1 2 reveals a high-angle relationship between S1 and S2 (Figs.7e,f).We already showed that pyrrhotite and pyrite porphyroblasts formed syn-D1 because pyrrhotite grains are aligned parallel to S1, but truncated and deformed by S2 (Fig. 7f), whereas pyrite grains are wrapped by S2 and developed syn-D2 strain fringes.We, therefore, infer two possible tectonic scenarios: 1) that the K3 axes of the AMS ellipsoid (pyrrhotite porphyroblasts) changed from subvertical (normal to a thrusting-related S1) to gently

M A N U S C R I P T A C C E P T E D ACCEPTED MANUSCRIPT
south-dipping normal to S2.In other words, the K2 and K3 axes interchanged positions during D2, while K1 was maintained constant (Figs.11a,b), 2) alternatively, the K1, K2 and K3 axes all interchanged positions during D2, similar to the pyrite porphyroblasts where all three principal axes of the shape ellipsoid switched from D1 to D2 (Figs. 5a and 11).Thus, the SPO of pyrrhotite during D1 depends on the geometry of the AMS axes (Fig. 11b).
We have argued that the magnetic foliation reflects an originally north-south striking and shallowly east dipping orientation of S1 and of axial planes of F1 recumbent folds (Figs. 10 and 11b).We further propose that progressive bulk inhomogenous shortening associated with foliation development (Bell and Rubenach, 1983) controlled the growth, orientation and AMS properties of pyrrhotite and pyrite porphyroblasts (Fig. 11a).Both minerals grew during D1 in low-strain microlithons bounded by S1 cleavage planes as manifested by the preferred orientation of pyrrhotite parallel to S1.With the development of S2, porphyroblast growth ceased and symmetric strain fringes grew off pyrite grains parallel to the E-W maximum extension direction.
Aeromagnetic image and field observations around the Martimo area show a type-2 fold interference pattern, where the east-west striking upright F2 folds deform east-vergent, shallowly dipping F1 folds locally resulting in crescent-mushroom or arrowhead surface patterns (Fig. 1a).The general fold pattern is characterized by a major synform in the south and antiform in the north (Lahtinen et al., 2015).This regional scale geometry fits well with our interpretation of fabrics in the studied sample, which comes from the core of the synformal D2 structure (cf.Figs.1a, 10 and 11c).A more extensive description of the deformation sequence can be found in Lahtinen et al. (2015).Our integrated approach and workflow of combining µ-CT, AMS and microstructures can be highly useful and applied to unfold the complex structural history, especially for the mineralized bedrock such as the Peräpohja region.

Complementary character of AMS and µ-CT techniques
µ-CT quantitative data provide 3D information for metamorphic textures that significantly complements and allow a more rigorous interpretation of AMS results.More specifically, µ-CT data add statistical mineral abundances, spatial distribution, size and shape of different mineral species with 13 a precision that cannot be achieved with other methods.In the case that concerns us here, the magnetic lineation (K1) was shown to coincide with the X-axis of pyrrhotite grain volumes (cf.Figs.4a,b and 6d), whereas a Flinn plot for these grains shows predominant prolate shapes in agreement with the AMS shape parameter (T) and anisotropy (P').In addition, our µ-CT data reveal that both pyrrhotite and pyrite grains are localized along alternating metasedimentary layers suggesting that the sulfide growth was chemically controlled by individual laminae, and were formed during D1 to form metamorphic porphyroblasts.The study of thin sections is still indispensable for revealing microstructural relationships between deformation fabrics and different minerals, as shown herein for pyrite and pyrrhotite grains with respect to S1 and S2 foliations, the truncation, shape and microfolding of pyrrhotite by S2, or the microboudinage textures.

Conclusions
a.By adding the 3D shape and preferred orientations of sulfide and magnetic minerals in a rock, µ-CT can significantly enhance the interpretation of bulk rock AMS, particularly in multiply deformed rocks.In addition, integrating µ-CT and AMS methods are highly useful in interpreting metamorphic textures in 3D.
b.In this study, we have shown that the AMS principal susceptibility axes experienced major changes between two successive deformation phases associated with a horizontal (S1) and subvertical (S2) foliations.K3 changed its orientation from normal to S1 to normal to S2 foliations, while K1 was maintained.Alternatively, the long axes of both pyrite and pyrrhotite porphyroblasts changed from N-S to E-W orientations during D1 and D2, respectively.The sample scale interpretations fit well with the regional scale type-2 fold interference pattern.
their long (X) and short (Z) axes being aligned E-W and N-S, respectively.
top-to-the west shearing components and perpendicular shortening components (Fig.10), consistent with the east-vergent D1 thrusting and folding event ofLahtinen et al. (2015).The secondary north-south preferred orientation seen in the rose diagram of Fig.5asuggests that pyrite porphyroblasts were aligned N-S during D1.During D2 they would have developed their main E-W preferred orientation (Figs.5a and 11), probably associated with pressure solution as that is the dominant deformation mechanism of pyrite in greenschist facies conditions(McClay and Ellis, 1983) c. Quantitative measurements of individual pyrrhotite grains demonstrate a predominance of highly prolate shapes, whereas pyrite grains have more equant geometry reflecting the higher strength of this less-deformed mineral.Our µ-CT images highlighted the localization of sulfide minerals along alternating metasedimentary layers that crystallized as porphyroblasts during the D1 metamorphism.petrography using multiple oriented thin sections was crucial for unravelling the deformation history of the studied sample.We conclude that conventional thin section microstructural analysis cannot be substituted, but only complemented by more sophisticated techniques such as µ-CT and AMS.

Fig. 1 .
Fig. 1.(a) Aeromagnetic map of the study area showing the location of sample 57.The area shows distinct mushroom-shape or arrowhead patterns typical of type-2 fold interference.A main antiform and synform pair can be distinguished.Inset on the top left shows a schematic 3D model.(b) Field photograph and line diagram of S0/S1 overprinted by S2, and drill core location.Fig. 2. Three-dimensional microstructures of sample 57 based on X-ray computed microtomography (µ-CT).The drill core is oriented and digitally chopped at three different levels to show the penetrative pyrrhotite fabric.Images on the left (a,c,e and g) are 3D side views, whereas those on the right (b,d,f, and h) are horizontal top-view slices.S denotes south.Pyrrhotite (po) and pyrite (py) are rendered yellow except for (b).Fig. 3. Different 2D and 3D representations of µ-CT data.(a) Horizontal 2D grayscale slice showing east-west aligned (X-axis) pyrrhotite (po) grains.(b) Volume rendering of the same image as shown in (a).(c,d,e) Side views, looking to the north and perpendicular to the long axes (X) of the pyrrhotite grains.Prolate pyrrhotite grains and cubic pyrite (py) are localized in alternating mm-scale metasedimentary layers (S0).The latter dips 45° east, whereas pyrrhotite long axes plunge more shallowly (~22°) in the same direction.(f) Grayscale µ-CT slice showing textural characteristics of pyrrhotite, pyrite and S0 and S1 relationship.(g) Folded metasedimentary layers are segmented (isolated) in red and blue.(h) Top 3D view of the drill core showing pyrrhotite grains in the X-Z axes.(i) 3D rendered side view along the Y-Z plane.(j) The same Y-Z plane is shown as a grayscale µ-CT slice.Pyrrhotite and pyrite are rendered yellow in (b), (c), (e), (g), (h) and (i).Fig. 4. Quantitative geometric analyses of pyrrhotite grains.(a) Rose plot showing the preferred eastwest alignment of pyrrhotite long axes (X).(b) Plunge values of X-axes of pyrrhotite averaging 22°.(c) Aspect ratio plot with a mean ratio of about 14.(d) Flinn diagram showing the tendency of grains towards the prolate shape, with a minor population of oblate shape (see text for discussion).(e) Aspect ratio vs. grain volume plot showing a trend between logarithms of grain volume and aspect ratio, where volume increases with increase in the aspect ratio.Fig. 5. Quantitative geometric analysis of pyrite grains.(a) Rose diagram for the long axes (X) of pyrite grains showing large spread with an E-W maximum, and a weaker N-S maximum.(b) Plunges of pyrite long axes are predominantly shallow (20-30°).(c) Aspect ratio diagram with an average value of about 4. (d) Flinn diagram for pyrite showing highly variable shapes of this mineral, but much lower aspect ratios than the pyrrhotite.(e) Aspect ratio vs. grain volume plot showing a minor volume increase with decreasing aspect ratio.Fig. 6.(a) Mean susceptibility (Km) vs. anisotropy degree (P′) plot showing typical values of pyrrhotite.(b) Shape parameter (T) vs. anisotropy degree (P') suggesting prolate shapes of pyrrhotite grains.(c) Heating (red) and cooling (blue) thermomagnetic curves showing a drop at about 325 °C, typical of pyrrhotite.(d) AMS plot showing east-west striking magnetic foliation (K1-K2) and shallowly plunging magnetic lineation (K1) carried by pyrrhotite.Fig. 7. Oriented photomicrographs showing microtextures in sample 57.(a) Horizontal thin section showing S2 deflecting around a pyrite (py) porphyroblast, crossed polarized light.Inset shows the µ-CT image marking the actual section plane.(b-c) Close ups from the horizontal thin section in (a) showing (b) intact, and (c) microboudinaged pyrrhotite (po) porphyroblasts elongated along S2, reflected light photomicrographs.(d) N-S vertical thin section showing change in the shape of pyrrhotite porphyroblasts from oblate to prolate.Strain fringes developed on the upper and lower faces of the pyrite porphyroblasts could be a cut effect (cf.Fig. 8b), crossed polarized light.(e) Close-up of the same thin section as 'd', showing pyrrhotite porphyroblasts truncated, deflected or microfolded by S2, plane polarized light.(f) Photomicrograph, plane polarized light image digitally rendered to