It has been suggested that open grain boundaries, cavities, and depressions can form a network which allows fluid circulation in rocks 2,7,11,36−38. However, such cavities and pores in rocks are not uniformly distributed 8,11,36,39. During metamorphism, it is known that textural equilibration significantly modifies grain boundary geometry 30,40,41; however, there has been no systematic investigation as yet to determine if there is any correlation between the grain boundary width and the grade of metamorphism, which is one of the primary controls on rock texture, along with other factors such as the orientation of the grain boundary relative to local kinematics, stretching direction, or cavitation along specific boundaries, to name a few. Textural equilibration is accompanied by attainment of equilibrium dihedral angles at triple junctions of grains 41 with a concomitant change in crystallographic orientations across grain boundaries; these changes are commonly determined by Electron Backscatter Diffraction (EBSD). During metamorphism, crystallographic re-orientation may in some cases lead to the formation of ‘special’ boundaries between grains, known as Coincident Site Lattice or CSL boundaries. The mechanical process of migration of fluids along a grain boundary is controlled by the width of the channel, with a greater width facilitating easier fluid transport along the channel. If some of these boundaries are CSL boundaries, they offer enhanced resistance to fluid percolation along the grain boundary network 17,34,42,43. Thus, in metallurgical and material science studies, the width of the grain boundaries, in terms of wider general boundaries and narrower CSL boundaries, are taken into consideration to model percolation behaviour in metals and materials 42,44.
In this study, we estimated the width of the grain boundary domain using Atomic Force Microscopy (AFM), which also reveals other interesting morphological features along the boundaries. Comparison of grain boundaries in quartzites metamorphosed under conditions varying from the greenschist to granulite facies shows distinct differences. Quartzites metamorphosed under greenschist facies conditions have wider grain boundaries compared to those metamorphosed under granulite facies conditions. For all the dynamically recrystallized samples, the reduction of grain boundary width is correlatable with increasing metamorphic grade. This implies that lower grade rocks generally have wider grain boundary domains, with abundant pores along the grain boundary network, than high grade rocks, and are therefore more amenable to fluid flow. The wide grain boundary domains in one apparently discrepant sample in this study (RN 36) can be attributed to the effect of modification of statically recrystallized grain boundary domains during the later lower temperature shearing event in the Rengali Province. Apart from the grain boundary width, an important morphological feature observed across the grain boundaries are the bridge-type structures detected with the AFM under high magnification, suggesting enhanced bonding across these boundaries. These bridge-structures are exclusively observed along narrower grain boundary segments, and have not been previously documented with any alternative technique. The results of force-distance spectroscopy provide more insight into the significance of these bridges. On grain boundary segments that contain bridges, the bridge domains show significantly less deformation on interaction with the AFM tip compared to the domains between the bridges. This suggests that there is a spatial variation in bond strength along these grain boundaries, being significantly higher across the bridges than in inter-bridge segments. Additionally, force-distance spectroscopy on one low grade and one high grade sample also suggests that bonding across grain boundaries in high grade rocks are significantly stronger than in low grade rocks. The enhanced bonding strength, and the presence of bridges across grain boundaries in high grade rocks increase the bulk strength of the rock and improve resistance to fluid percolation; they may also be regarded as evidence for the existence of special (CSL) boundaries in rocks, similar to those reported in metals.
Since the present study attempts to relate grain boundary domains in metamorphic rocks of different grades to fluid flow, it is pertinent to discuss if the grain boundary widths measured under ambient pressure temperature conditions are representative of the in situ situation at depth. Studies on quartz 45–48 show that volume reduction associated with decompression is anisotropic. This implies that quartz grains would undergo volume reduction as they cool through the brittle-ductile transition during uplift, with associated widening of grain boundaries. However, the results of this study argue that for high grade and low grade rocks, the situation should be different. The enhanced degree of bonding (manifested as bridges under an AFM) in high grade rocks would imply that the constriction of grains on either side of the boundary is limited, and does not necessarily influence the boundary width itself. Additionally, strong bonds across a grain boundary would prevent the boundary from widening or forming voids due to decompression, to a much larger extent than boundaries across which the bonds are weak. The width of boundaries that contain bridge structures are therefore inferred to preserve their original widths from depth to the surface, and are not an artefact of exhumation. Thus, it appears that in quartzites, grain boundary widths reduce significantly with increasing metamorphic grade, with enhanced bonding at higher grades forming ‘bridge-type’ structures across the grain boundaries. Ultimately, this implies an increased resistance of channels in quartz-rich high grade rocks to fluid flow, consistent with the relatively anhydrous character of the granulite facies. Enhanced bonding across most grain boundaries in high grade rocks reduces the degree of opening of grain boundaries on cooling and decompression, whereas the more limited amount of bonding in lower grade rocks means these boundaries open more; therefore, difference in observed widths reflect the difference in bonding and its influence on grain boundary opening.