The oddity: glass preservation in slowly cooled rocks
Cesare et al. (2009) showed that glass is preserved in some of the inclusions found in the garnets from the KKB granulites, in the same clusters containing the nanogranites and interspersed with them. The presence of silicic melt inclusions had also been reported in zircons from the migmatites of the Ulten Zone, Italy (Braga and Massonne, 2008). After these first studies, glass has also been found in inclusions within the garnets from the Ronda migmatites (Figure 14A). The amorphous nature of this phase has been proven by Raman spectroscopy, although the optical isotropy, SEM-homogeneous appearance and rhyolitic composition (see below) of this phase would suffice to demonstrate that it is a glass.
Figure 14. SEM microstructures of MI and remelted nanogranites
A: glassy inclusion (right), apparently larger than coexisting nanogranite (left) in the same cluster, hosted by garnet in a migmatite from Ronda. B: partially crystallized melt inclusions in garnet from Ronda. C: the product of remelting with an ambient-pressure heating stage of a nanogranite inclusion from the KKB rocks: decrepitation and some interaction with the host are visible (white arrows), along with thin overgrowths of peritectic mineral on inclusion walls (red arrows). D: the product of remelting in a piston cylinder of nanogranite inclusion from the Ronda migmatites: no decrepitation or incongruent melting of garnet walls is observed.
Recent work on the Ronda migmatites indicates that along with totally crystalline or completely glassy MI, partially crystallized inclusions are also common (Figure 14B): these contain variable proportions of glass, and the crystallized phases are generally quartz, biotite, muscovite and, more rarely, Na-rich plagioclase. Plagioclase is absent also in the few observed partially crystallized inclusions from the KKB.
While glassy inclusions are expected in phenocrysts from extrusive rocks, their xenoliths and enclaves, or pyrometamorphic rocks, all of which undergo sudden cooling from suprasolidus temperatures, the preservation of glass (quenched anatectic melt) in regionally metamorphosed and slowly cooled migmatites and granulites is a geological oddity. For example, even though migmatisation of metapelites around the Ronda peridotite could be modelled or considered in terms of a "regional contact metamorphism" that might have lasted only a few m.y., the available geochronological record (Cenki et al., 2004) indicates that the KKB granulites took >60 m.y. to cool to <350°C from the peak UHT conditions. These timescales are generally considered incompatible with the solidification of melt to glass.
Moreover, considering that the presence of temperature gradients across the mineral hosts can be ruled out due to their small volume (a few mm3), the above problem has two (possibly connected) aspects: (i) the presence of glass, and (ii) the coexistence, a few µm apart, of glassy and (partially) crystallized inclusions. This translates into two questions: (1) what is the cause of the lack of melt crystallization? (2) Why did crystallization occur only in some inclusions?
Concerning the second question, there may be several possible answers, such as: i) differences in composition among melts in different inclusions, or ii) heterogeneous distribution of nucleation sites among inclusions (e.g., irregularities on inclusion walls' surface, trapped minerals, etc.), or iii) heterogeneous microfracturing/decrepitation that promoted nucleation, or iv) difference in inclusion size, with inhibition of nucleation in the smaller ones. Measurements of the diameters of MI in garnets from the KKB rocks led Cesare et al. (2009) to the preferred explanation that inclusion size was the main parameter affecting the behaviour on crystallization, and that, as already shown in acqueous solutions (Putnis et al., 1995), crystallization was inhibited in the smaller inclusions (see also Holness and Sawyer, 2008).
The pore size effect, supported by the KKB samples, seems to find less evidence in the MI from Ronda, where glassy inclusions appear to have the same range of size (Figure 9) or even be smaller than fully crystallized inclusions (Figure 14A). Based on the currently available data we can suggest that ˜5-10 µm may represent a threshold diameter under which crystallization is inhibited, and that in each inclusion additional factors, such as surface irregularities or preexisting nuclei, may play an additional role.
More data and case studies are required to obtain a definitive solution to the paradox of glass preservation, but regardless of such uncertainty, and albeit rare, the glassy MI constitute an extremely important feature: since their composition is comparable with that of crystallized nanogranites (see below), glassy MI are the most primary examples of anatectic melts that can be found in migmatites and granulites, and their composition can be analyzed – within the analytical limitations imposed by their small size – without having to manipulate the inclusion, e.g. by re-melting the nanogranite.