The solubility of H₂O-CO₂ fluids in a synthetic analogue of a phono-tephritic lava composition from Alban Hills (Central Italy) was experimentally determined from 50 to 500 MPa, at 1200 and 1250 °C. Contents of H₂O and CO₂in experimental glasses were determined by bulk-analytical methods and FTIR spectroscopy. For the quantification of volatile concentrations by IR spectroscopy, we calibrated the absorption coefficients of water-related and carbon-related bands for phono-tephritic compositions. The determined absorption coefficients are 0.62 ± 0.06 L/(mol·cm) for the band at ~4500 cm⁻¹ (OH groups) and 1.02 ± 0.03 L/(mol·cm) for the band at ~5200 cm⁻¹ (H₂O molecules). The coefficient for the fundamental OH-stretching vibration at 3550 cm⁻¹ is 63.9 ± 5.4 L/(mol·cm). CO₂ is bound in the phono-tephritic glass as CO₃²⁻ exclusively; its concentration was quantified by the peak height of the doublet near the 1500 cm⁻¹ band with the calibrated absorption coefficient of 308 ± 110 L/(mol·cm). Quench crystals were observed in glasses with water contents exceeding 6 wt% even when using a rapid-quench device, limiting the application of IR spectroscopy for water-rich glasses. H₂O solubility in the ultrapotassic melts (7.52 wt% K₂O) as a function of pressure is similar to the solubility in basaltic melts up to 400 MPa (~8 wt%) but is higher at 500 MPa (up to 10.71 wt%). At 500 MPa and 1200 °C, the CO₂ capacity of the phono-tephritic melt is about 0.82 wt%. The high CO₂ capacity is probably related to the high K₂O content of the melt. At both 200 and 500 MPa, the H₂O solubility shows a non linear dependence on XH₂O^f in the whole XH₂O^f range. The variation of CO₂ solubility with XCO₂^f displays a pronounced convex shape especially at 500 MPa, implying that dissolved H₂O promotes the solubility of CO₂. Our experimental data on CO₂ solubility indicate that the interaction between phono-tephritic magma and carbonate rocks occurring in the Alban Hills magmatic system may result in partial dissolution of CO₂ from limestone into the magma. However, although the CO₂ solubility in phono-tephritic melts is relatively high compared to that in silicic to basaltic melts, the capacity for assimilation of limestone without degassing is nevertheless limited to <1 wt% at the P-T conditions of the magma chamber below Alban Hills.

Hydrothermal experiments were conducted at ca. 1 to 7000 bars and 700 to 1250 °C in 121 rhyolitic to basaltic systems to determine Cl solubility in silicate melts, i.e., the maximum Cl concentration in melts that are saturated in a hydrosaline liquid with or without an aqueous or aqueous-carbonic vapor. The Cl concentration of melts increases with the Cl contents of the fluid unless the melt coexists with vapor plus hydrosaline liquid at fixed pressure and temperature; this phase assemblage buffers the Cl content of each phase with increasing Cl in the system. The Cl content of fluid(s)-saturated melts is independent of the CO₂ concentration of the saline liquid ± vapor with up to 21 wt% CO₂ in the fluid(s). The experiments show that Cl dissolution in aluminosilicate melts increases with temperature and pressure. Chlorine solubility is also a function of melt composition; it increases with the molar ([Al^1/2+Ca^1/2+Mg^1/2+Na]/Si) of the melt. These experimental data have been integrated with results involving 41 other experiments (Webster and De Vivo 2002) to develop a broadly expanded model that supports calculation of Cl solubility in 163 aluminosilicate melts. This empirical model applies to Cl dissolution in melts of most silicate magmas at depths as great as 25 km. It determines the exsolution of hydrosaline liquid, with or without a coexisting vapor, as magmas ascend from depth, cool, crystallize, and differentiate from mafic to felsic compositions. In combination with H₂O solubility models, our model supports determination of H₂O-Cl solubility relations for most aluminosilicate magmas and is useful for barometric estimations based on silicate melt inclusions containing low CO₂ and moderate to high-Cl concentrations. The model is applied to the phase relations of fluids in volatile-enriched magmas of Augustine volcano, Alaska. The Cl and H₂O concentrations of melt inclusions from 14, basaltic to dacitic eruptive units are compared with modeled solubilities of Cl and H₂O in Augustine melts. The majority of these eruptions involved magmas that first exsolved aqueous to aqueous-carbonic vapors when the melts were dacitic in composition (i.e., before the residual melts in these magmas had evolved to felsic compositions) and well prior to the eruptions. Hydrosaline liquid with or without a vapor phase exsolved from other, more-felsic fractions of Augustine melts at low, near-surface pressures of several tens of bars.