6.6 Origin of the leucogranites and melt migration

Radiogenic and stable isotopic composition of the leucogranites, their common occurrence within the high-grade regional metamorphic terranes, the lack of spatial and temporal association with basalt magmatism strongly suggest that the leucogranites are the product of pure crustal melt, uncontaminated by mantle material.

The source of the Himalayan Leucogranites is generally ascribed to the aluminous schists and gneisses of the HHCS, partly because of structural relation and partly because of isotopic systematics for Rb, Sr, Nd, O (France-Lanord et al., 1988; Deniel et al., 1987; Ferrara et al. 1991, Harris and Massey, 1994). The monazite and zircon upper intercept ages around 500 Ma of the Zanskar leucogranite clearly indicate that they must be partly derived from the melting of the Cambro-Ordovician Granites. As these old granites are relatively poor in muscovite, the leucogranite must also have had a more pelitic source. On the basis of Sr and Nd isotope ratios, Ferrara et al. (1992) found two different isotopic signatures for the Gumburanjun Leucogranite. They attribute these differences to variations in the composition of the source rocks, one yielding a characteristic signature of a metapelitic protolith and the other of an igneous (granitic) protolith.

The origin of Himalayan leucogranite was believed to be related to water-saturated melting. Fluid advection from the footwall of the Main Central Thrust (Lesser Himalaya) into the relatively hot hanging wall (HHCS) was invoked as the driving mechanism for the production of water-saturated melt in the upper structural levels of the HHCS (Le Fort et al. 1987). There is however little evidence to support the notion of pervasive aqueous fluids during metamorphism of the mid-lower crust and recent trace element studies argue against fluid present melting in the formation of crustal melts (Harris et al. 1993). Also in Zanskar, metamorphic conditions and textural evidences observed in the migmatitic zone rather indicate that melts were derived from the incongruent melting of muscovite.

Lately, Himalayan leucogranitic magmas are thus increasingly thought to be initially water-undersaturated, indicating either that a fluid phase with aH2O<< 1 was present during melting or that the melting reactions were fluid-absent. In the absence of free water, melting depends on the availability in the source region of hydrous minerals like muscovite or biotite which may release their water during anatexis. The amounts of water released by the breakdown of these minerals is usually not sufficient to saturate the magma and is dissolved in the melt without formation of a vapour phase (Vilzeuf and Holloway, 1988). This process is referred to as «fluid-absent melting» (Clemens, 1984), «dehydration melting (Thompson, 1982)» or «vapor-absent melting (Grant, 1985)». Aluminous schists and gneisses are generally considered to be the likely source for peraluminous granites.

The breakdown of muscovite occurs through the reaction: 22 Muscovite + 7 plagioclase + 8 quartz = 5 K-feldspar + 5 Al₂SiO₅ + 2 biotite + 25 melt (Patino Douce, 1998), which is also known as the second sillimanite isograd (Spear, 1993). This reaction can however only produce 10-15 vol. % melt between 750° and 850°C and at 10 kbar (Vilzeuf and Holloway, 1988). Up to 50 vol. % melt can however be produced at higher temperatures (850°-900°C) with the breakdown of biotite through the reaction: biotite + Al₂SiO₅ + plagioclase + quartz = Garnet + K-feldspar + melt. (Vilzeuf and Holloway, 1988). The amount of partial melting required before melts begin to segregate and forms plutons is called critical melt percentage. The value of this critical melt percentage is generally believed to be of the order of 25% (Spear, 1993).

A possible objection to fluid-absent melting of muscovite as the source for the leucogranites might thus be that the low melt fraction obtained through this reaction (F < 0.15) would preclude a critical melt percentage being obtained (F > 0.25) and hence prevent the melt from leaving its source. The concept of critical melt percentage does, however, assume that the melt will migrate through diapirism. We have described above that abundant feeder dikes are rooted into the migmatitic zone, which indicates that the ascent of the melts occurred through fracture propagation and was driven either by buoyancy or injection pressure, and thus diapiric arguments are irrelevant for the present case. Moreover, we have seen that the xenoliths within the leucogranitic plutons have also undergone incongruent melting and thus that the injection zone must also have contributed to the global amount of melt. Under such circumstances it is difficult to place a minimum constraint on the melt fraction that can be extracted from the source.

Patino Douce and Harris (1998) have conducted a melting experiment on two types of metapelitic rocks from the HHCS. Their results indicate that dehydration melting begins at 750°-800°C and produces melts that are virtually identical to the himalayan leucogranites. Adding extra water to the starting material lowers the melting point < 750°C, which is pretty logical, but the resulting melt has a trondhjemitic composition, which is different from most Himalayan leucogranites. For these authors, the Himalayan leucogranites were generated by fluid-absent melting at temperatures around 750°C and 6-8 kbars during adiabatic decompression and are solely the result of the breakdown of muscovite, which is in contradiction with Spear's (1993) critical melt percentage.

The dP / dT slope of the biotite dehydration curve is steep and is expected to intersect the flatter muscovite dehydration curve at high pressure. The actual pressure of the crossover is however uncertain as the exact temperature of the breakdown of biotite is highly dependant on the mg-number of biotite and the Ti and F contents of both biotite and muscovite as well as on the plagioclase composition of the source rocks. Estimation range from 9 kbar (Patino Douce and Harris, 1998) to 14 kbar (Spear, 1993).

Finally, the recognition that melting occurs essentially in the upper structural levels of the HHCS argues against the correlation of anatexis through fluid influx from the regional thrusting on the MCT, but is consistent with an association between granite formation and exhumation along the Zanskar Shear Zone.