Supplementary material A: Methods
Mineral chemistry
The used SEM is equipped with an energy dispersive X-ray spectrometer (EDS) from Oxford Instruments, using INCA software. Running conditions were 20 kV accelerating voltage, 3.5 nA specimen current and 40 s livetime. Cobalt was used as a reference standard and to monitor the system drift. X-ray counts were translated into concentrations using stored calibrations based on mineral, pure element and oxide standards (e.g. jadeite for Na, Fe-metal for Fe and Al2O3 for Al). Physical corrections for atomic number, absorption and enhancement were performed using a ZAF model (phi-rho-Z). From measurements of Smithsonian Standard Diopside, accuracy was within 5% for all major and minor oxides reported in Table 2. Precision was estimated by ten repeated analyses of plagioclase, giving coefficients of variation (i.e. relative standard deviation) of 0.80 for SiO2 and Al2O3, 1.77 for CaO and 4.14 for Na2O for 5 s measurement time. These figures were improved by 50% when increasing measurement time to 40 s.
PT Calculations
In Domino, a Gibb’s free energy minimization approach is used in order to calculate mineral reactions in a given system using the chemical composition of the system and a thermodynamic database. Here, the JUN92 data base was used. The foundation of JUN92 is Berman et al. (1988), but expanded with other data (see Domino software for full references). In the absence of reliable alternatives we chose for epidote-clinozoisite and chlorite ideal mixing models; however, for white mica (Keller et al. 2005) and plagioclase (Fuhrman and Lindsley 1988) non-ideal were chosen. Equation of state for water and carbon dioxide was after Kerricks and Jacobs (1981).
Electron Backscatter Diffraction analysis (EBSD)
Analyses were performed using a Phillips XL-30 FEG-ESEM equipped with a Nordlys detector and Channel 5 analysis suite from HKL Technology (Oxford instruments). Thin sections were chemically polished and left uncoated during automatic beam scans over gridded areas. A step size of 1 or 2 µm was used so that even the smallest grains (less than 9 µm) could be reliably characterized. EBSD settings were 20 keV accelerating voltage and 0.4 torrvaccum. Andesine (C; Fitz Gerald et al. 1986) served as theoretical model to index plagioclase. The raw EBSD data was processed following the procedures of Prior et al. (2002) and Bestmann and Prior (2003).
Angles of longest grain axis to foliation, Y-X, used for shape preferred orientation (SPO) diagrams were extracted from the EBSD data. The distributions of misorientation axes are plotted to assess possible relationships between large grains and adjacent small grains and also with respect to the structural XYZ reference frame (Wheeler et al. 2001). Relative variations in misorientation from a chosen reference orientation (i.e. a selected point on an EBSD map) are displayed in cumulative orientation maps. Grey scale pattern quality maps illustrated the quality of the indexed points where the brighter grey is higher quality. These maps are used to represent boundaries (subgrain-, twin-, and grain boundaries) as these appear dark due to the low data quality at boundaries.
References
Berman RG (1988) Internally-consistent thermodynamic data for minerals in the system Na2O-K2O-CaO-MgO-FeO-Fe2O3-Al2O3-SiO2-TiO2-H2O-CO2. J Petrol 29:445-522.
Bestmann M, Prior DJ (2003) Intragranular dynamic recrystallization in naturally deformed calcite marble: diffusion accommodated grain boundary sliding as a result of subgrain rotation recrystallization. J StructGeol 25:1597-1613.
Fitz Gerald JD, Parise JB, Mackinnon IDR (1986) Average structure of an An48 plagioclase from the Hogarth Ranges.Am Mineral 71:1399-1408.
Fuhrman ML, Lindsley DH (1988) Ternary feldspar modeling and thermometry. Am Mineral 73:201-215.
Keller LM, de Capitani C, Abart R (2005) A quaternary solution model for white micas based on natural coexisting phengite-paragonite pairs. J Petrol 46:2129-2144.
Kerrick DM, Jacobs GK (1981) A modified Redlich-Kwong equation of state for H2O, CO2, and H2O-CO2 mixtures at elevated pressures and temperatures. Am J Sci 281:735-767.
Prior DJ, Wheeler J,Peruzzo L, Speiss R, Storey C (2002) Some garnet microstructures: an illustration of the potential of orientation maps and misorientation analysis in microstructural studies. J StructGeol 24:999-1011.
Supplementary Figure 1
Rose-diagrams of new plagioclase grains showing the orientation of long grain axis compared with the orientations of foliation, fractures and twins. Long grain axis orientations are extracted from the EBSD data collected over areas shown in Fig. 3. For Area 2, data for(a) is from left square in Fig. 3c, data for(b) is from the right. Data for diagrams of Area 3 are from left (a) and right (b) half of box in Fig. 3e. Stippled line is trace of most prominent twin boundaries. Structural reference frame is shown. Arrows point to the average value.
Supplementary Figure 2
Inverse pole figures (equal area and lower hemisphere) depicting the distribution of misorientation axes in the crystal and sample (structural) coordinate system for host clast and new grains by low angle boundaries (<10°) and high angle boundaries (10-180°). Clusters encircled by stippled red circles only occur in areas with a chemical change (i.e. lower An-content). Complete data sets from whole grains.
Supplementary Figure 3
Crystallographic orientations of selected minerals from Areas 2 and 3a.Foliation is horizontal line and lineation is at X. Shown as one point one grain; equal area, lower hemisphere projection. N is number of grains measured.
Supplementary Figure 4
Pattern of misorientation change in amphibole in the central part of domain 3; inset displays the position of the profile across the grain. Colour coding of inset represents orientation differences in degrees from the starting position (red star). An un-strained/un-bent grain would thus only be blue, abrupt colour changes represent subgrain boundaries.