Dissolved Volatile Concentrations in Melt Inclusions

One of the primary goals of most modern studies of MI has been to determine the amounts of dissolved volatiles present within crystallizing magmas. Most studies have concentrated on volcanic rocks, as they are often pristine and unaffected by post-entrapment phenomena such as leaking and crystallization. Table 2 lists studies that have determined the concentrations of dissolved volatiles in MI from silicic magmas such as dacites and rhyolites, compositions typically associated with hydrothermal ore-forming systems.

Note: Data presented as X000-Y000 indicates range of values in original publication. Data presented with '±' shows 1s as listed in original publication. Dash indicates data not reported.

# CO2 by IR spectroscopy (FTIR). In studies with CO2 data, H2O data are also by FTIR.

ÝPersonal communication (1994).

*Peralkaline rocks: Molar (Na2O+K2O)/Al2O3 >> 1.0

---

Though not all silicic melts have been shown to contain high H2O concentrations, most seem to have at least 4 wt.% dissolved H2O in the melt. Some magmas are zoned in H2O before eruption (Bishop and Bandelier tuffs), whereas others are homogeneous (1912 Valley of 10,000 Smokes rhyolite, Taupo). Carbon dioxide has been detected in several systems (Pine Grove, Bishop tuff, Pinatubo), though others contain virtually none (Crater Lake, 1912 Valley of 10,000 Smokes rhyolite). Chlorine and F are highly variable in concentration, and their abundances appear to be correlated with tectonic setting (higher Cl/F in arcs relative to continental settings). Sulfur is generally very low (<100 ppm) in high-silica rhyolitic melts with < 1.0 wt.% FeO, although there is experimental evidence that oxidized, high-temperature silicic melts can contain higher concentrations of S as dissolved sulfate (Luhr 1990). Table 3 summarizes current understanding of volatile concentrations in silicic melts, based on analyses of silicate MI.

Data on magmatic volatiles have the following applications. 1) They provide constraints on the amounts of magma-derived volatiles likely in ore-forming systems and on the general compositions of exsolving fluids. 2) They can demonstrate the relative importance of decompression-related degassing versus crystallization-induced degassing on the fluid-release history of a magma (Lowenstern 1994b). For example, if a magma is shown to have been volatile saturated at a given depth, the mass of exsolved fluid produced during further ascent can be calculated (Shinohara & Kazahaya 1995). 3) They give reliable estimates of the dissolved concentrations of ore-metals in magmas prior to degassing. Prior to the study of MI, economic geologists had to depend solely on analyses of degassed intrusions and extrusive rocks to characterize the metal and volatile composition of potential ore-forming magmas. 4) They may be used as geobarometers to determine the minimum pressures at which crystals grew (Anderson et al. 1989; Lowenstern 1994b).

Care must be used when interpreting data from MI, as they represent only a single phase within the magma reservoir: the melt phase. To estimate the total amount of volatiles in a magma reservoir, one must have constraints on the relative amount of melt, crystals and exsolved fluid (Gerlach et al., accepted; Wallace et al. 1994). Though MI yield constraints on the composition of fluids in equilibrium with the magma, one must also recognize that the composition of exsolved fluid may change as it ascends from the magma, cools and interacts with the surrounding environment (Hedenquist 1995).