Petrogenesis and Experimental Petrology of Granitic Rocks

Modern granite petrology based on experimental results began with the investigations of Tuttle and Bowen (1958) on the haplogranite system Qz-Ab-Or-H₂O. The results demonstrated that granite magmatism can be a viable crustal process, and stimulated many petrologists, geochemists, structural geologists, and geophysicists to work on problems related to granite magmatism. Many important ideas and findings are the results of these efforts.

Natural cooling of granitic magmas is a slow process, usually sufficiently slow to attain equilibrium compositions in the crystallizing minerals. Most of the rock-forming parameters, like temperature, pressure, water activity, oxygen fugacity, etc., can be very well simulated in experiments; but the span of time available in nature, and the slow and more or less continuous change in the parameters given above cannot be attained. It is still problematic to reach equilibrium at low temperatures and low water activities, and geoscientists intending to apply experimental data to interpret the formation of crystalline rocks have to be aware that in many cases (even in carefully performed experiments) only near-equilibrium compositions could be attained. This is especially true for experiments involving plagioclase feldspars. Plagioclase poses a special problem due to slow chemical diffusion in the crystal structure. Problems inherent to plagioclase in low-temperature (<800°C) experiments are discussed in Chapters 6 and 7.

Water is generally recognized as the most important magmatic volatile component, both for its abundance and for its effects on physical and chemical properties of the melt. The fundamental role of water was understood very early by Goranson (1931), who emphasized that “many debatable problems of volcanology ... may be readily apparent if we know how water dissolved in magmas behaves at different temperatures and pressures”. The dramatic effects of the change of water content of the melt on phase relations (as outlined in Chap. 2) is due to the variation of physical properties of melts with changing water content (e.g., viscosity, density, diffusivity, solubility of other elements). Therefore, the role of water in melting processes (collecting and segregation of melts), migration of melts, and crystallization of magmas is fundamental.

After silicon and oxygen, aluminum is the most abundant element in silicate melts. As numerous leucogranites and rhyolites formed by partial melting of crustal rocks are peraluminous¹ (e.g., White and Chappell 1977; Clarke 1981; Clemens and Wall 1981), it is of great interest to know experimentally the effect of normative A1₂O₃ on phase relations and on properties of alumino-silicate melts.

In the past 30 years, many experimental studies have been performed in which components other than those forming quartz and feldspars have been added to the haplogranite or haplogranodiorite systems. Iron and magnesium are the two most important additional elements in granites since the presence and amount of these elements will control the mafic assemblage. Different types of experiments have been conducted: investigations of phase relationships, of the composition of coexisting ferromagnesian phases and melt, and of some individual reactions such as the crystallization or breakdown of garnet, biotite, cordierite, amphibole, or pyroxene in presence of melt. However, only a few of these studies were performed on synthetic systems. This is particularly true in the case of the determination of phase relationships in Fe-bearing systems because of the two possible oxidation states of Fe (Fe²⁺ and Fe³⁺). Investigating the stability of Fe-bearing phases implies that oxygen fugacity (\( {f_{{{O_2}}}} \)) has to be controlled (or at least known) in pressure vessels, a condition which is not always satisfied in the experimental studies. In this section, only the role of \( {f_{{{O_2}}}} \), the amounts of Fe and Mg which can be incorporated in granitic melts, and some of the reactions involving ferromagnesian minerals and melt will be discussed. Because of the lack of data obtained in synthetic systems, the discussion is based on results obtained from both natural and synthetic starting materials.

Tonalites are quite widespread in the Archean and in the lower continental crust, and are considered to be products of partial melting of basic protoliths, and important source rocks for the generation of granodiorites and granites (Rutter and Wyllie 1988; Skjerlie and Johnston 1992). In spite of their enormous significance and widespread occurrence, there have been relatively few experimental investigations on the tonalite system.

In addition to the haplogranite components Qz, Ab and Or, the granite system contains also the component CaAl₂Si₂O₈ (An). Two feldspars may be present, a plagioclase (Na_1-xCa_xAl_1+xSi_3-xO₈) and an alkali feldspar (K, NaAlSi₃O₈). Due to the melting loop of plagioclase (distribution of plagioclase components between melt and coexisting plagioclase, Bowen 1913), this quarternary system is noneutectic. At a given pressure and bulk composition, there is always a temperature interval between solidus and complete melting. The beginning of melting temperature at a given pressure is different for different Ab/An ratios of the system.

The first melting experiments on granitic and basaltic rocks were carried out by Hall (1805) at the end of the 18th century. Hall’s pioneering attempts were aimed to provide experimental support for Hutton’s theory on the origin of granites by crystallization of magmas. During the last decades of the 19th century and the first of this one, new attempts were made to determine the melting temperatures of natural rocks of granitic and basaltic composition by heating small specimens until they softened in laboratory furnaces. Melting temperatures for granites were reported in the range 1100–1300 °C. These are values as high or even higher than melting temperatures reported for basaltic rocks (Fouqué and Lévy 1882; Douglas 1907).

The hydrous rock-forming minerals have long been recognized as a source for the formation of water-bearing granitic magmas (Brown and Fyfe 1970; Robertson and Wyllie 1971 a, b; Wyllie 1971; Kerrick 1972; Eggler 1972; Grant 1973; Huang and Wyllie 1973; Huang et al. 1973 a, b). Lambert et al. (1969), Brown and Fyfe (1970), Lambert and Wyllie (1970, 1972), Huang and Wyllie (1973, 1981), and Millhollen and Wyllie (1974) were some of the earliest to investigate pressure-temperature conditions for the melting of rocks containing at least one hydrous mineral. However, these investigations were mainly restricted to deducing dehydration melting conditions from water-saturated melting experiments. Burnham (1979b) presented solidi for dehydration melting of mineral assemblages containing muscovite, biotite, and hornblende, respectively, based on hydrothermal relations.