**4. Implications**

Fine-grained samples can be mapped for their mineralogy with EDX on an SEM with a step size that is considerably smaller than the diameter of the interaction volume of the primary electron beam. Whilst the BSE photomicrograph provides a higher resolution display of the region of interest, it is clearly visible that there is not enough average atomic number contrast available between the mineral phases present to clearly differentiate and quantify the mineralogy only based on the BSE image. Therefore, mineral mapping at a very high resolution is necessary to obtain most of the information on the composition of the fault gouge sample. To be able to do this, a basic understanding of the mineralogy needs to be gathered from the sample (e.g., rock type, metamorphic facies, major minerals). Building a mineral library based on the measured compositions is the initial starting point to base classifications on for a finer pixel spacing (i.e., 200 nm) and the quantification at the nanoscale mineralogy. Whilst analyses at this resolution provided mixed pixels/signals, the ability to quantify the wt.% contribution of the elements allows for the identification and deconvolution of the mixtures and the successful classification of the mineralogy at this scale. Even for minerals with a low Z and thus a high interaction volume, which is several times larger than the size of the smallest particles and larger than the step size, a large part of the chemical information comes from the area where the beam hits the sample. Figure 3 shows that for all the investigated cases, 50% of the electron trajectories remain within a 75–160-nm wide interaction volume area and reach depths of maximum ca. 200–300 nm. Therefore, the mixed signals could successfully be identified and translated into classifications in the mineral library that enable the correct classification of those particles (Figure 4). The latter figure shows that even for 200-nm pixels, and most of the pixels at the grain boundary between two phases, the correct mineral could be determined from the mixed signal. Only one pixel away from the grain boundaries, the signal for each mineral is very near its expected composition. Compare e.g., the two yellow boxes for quartz in Figure 4.

The small step size applied for this analysis allows for an investigation of the fault gouge mineralogy at the nanometer scale with a quantification of the mineralogy to complement the BSE imaging. As EDX analyses are based on chemistry only, and the identification of the smallest grains is dependent on a beam interaction volume of a size larger than the smallest grains in the sample, X-ray di ffraction or electron backscatter di ffraction analyses might need to be performed in order to establish the crystallography of the minerals with an independent characterization method. This is especially important for clay minerals, which may show interstratifications at a nanometer scale, which might remain unnoticed with EDX analyses only.

The minerals in the amphibolite (anorthite, hornblende) were more intensively comminuted than the minerals in the orthogneiss (quartz, anorthite, and feldspar). Large grains of anorthite and especially quartz are still present in the sample, even though the majority of the grains were reduced to very small grain sizes. Illite is more fine-grained than quartz, which suggests that feldspar was more intensively comminuted than quartz, and that the illite grains did not grow after their formation. This brittle behavior of quartz and feldspar suggests greenschist facies temperature conditions during comminution [20,21]. Therefore, the fault gouge formed after the uplift of the area to upper-crustal levels (greenschist facies conditions) during a late stage of the geological history of the area. This is also confirmed by the observed mineral reactions of hornblende and anorthite to epidote, biotite to chamosite, and feldspar to illite, which all are typically occurring at lower greenschist facies conditions [22–24]. No large veins and no stacks of mica grains (chamosite, muscovite) or illite were observed, suggesting that these minerals grew after deformation, as a result of diagenetic processes (compare to [24]).
