Evidence for magmatic vapor deposition of anhydrite prior to the 1991 climactic eruption of Mount Pinatubo, Philippines

Anhydrite (CaSO4) phenocrysts from Mount Pinatubo pumices show evidence of having responded dynamically to changing conditions prior to the June 15, 1991 climactic eruption. Micrometer-sized and smaller pyramidal surface growth features and lesser numbers of etch pits on anhydrite surfaces are documented by scanning electron microscopy. Chemical analyses indicate that the pyramids are a CaSO4 polymorph and electron backscatter diffraction patterns show conclusively that the pyramids are indeed orthorhombic anhydrite and not another Ca-sulfate phase. Unit-cell measurements of volcanic anhydrite are identical with evaporitic anhydrite, as determined from single-crystal X-ray diffraction patterns. The computer program SOLVGAS was used to identify conditions under which the pyramids may have precipitated. Thermodynamic modeling of a cooling magmatic gas (H2O-CO2-SO2) at 500 bars (maximum model pressure) and NNO +1.7 was performed. Assuming that the gas contained >10 mol% Ca and 4 mol% SO2, the program indicates that anhydrite will precipitate homogeneously at approximately 780 ∞C, whereas an isothermal drop in pressure would likely lead to dissolution. Pyramids located between a phenocryst and adjacent glass provide physical evidence that at least a portion of the pyramids nucleated and grew before the melt quenched. We propose a mechanism to account for these previously unrecognized surface growths, which is that the anhydrite pyramids precipitated from a fluid or vapor phase that had separated from the magma at depth. At least a portion of the Pinatubo anhydrite phenocrysts provided substrates for nucleation and epitaxial growth of anhydrite. Because the anhydrite pyramids resemble products of chemical vapor deposition of metals and ceramics, we propose that this previously unrecognized process be termed magmatic vapor deposition. Evidence of high-temperature, magmatic vapor deposition of anhydrite could be used as an indicator that a pre-eruptive gas phase was accumulating or stored at depth, especially at arc volcanoes where excess amounts of sulfur are vented into the atmosphere. JAKUBOWSKI ET AL.: MAGMATIC VAPOR DEPOSITION OF ANHYDRITE 1030 pumice (Bernard et al. 1991; Fournelle 1991) and at least 12 papers in the Mt. Pinatubo Fire and Mud monograph (Newhall and Punongbayan 1996) discussed this anhydrite. To date, no one has examined the crystal chemistry of the Pinatubo (or other volcanic) anhydrite. Virtually all the sulfur in the pumice is in the anhydrite, and some investigations have suggested that anhydrite breakdown is significant in the production of the erupted SO2. Additionally, gas modeling has assumed that the anhydrite in the pumice is orthorhombic b-CaSO4. This study focuses on a detailed examination of the Pinatubo anhydrite and the surface features discovered in the course of our research. Pallister et al. (1996, p. 726) summarized some important features of the Pinatubo anhydrite: “...anhydrite occurs primarily as subhedral to euhedral crystals in the June 12 scoria and June 15 pumice. It typically includes apatite microlites and, as at El Chichon (Luhr et al. 1984), it is found in growth contact with apatite phenocrysts. Anhydrite is only rarely found in growth contact with silicate phenocrysts.” They suggested that “these relations are consistent with [anhydrite] growth mainly from a separate fluid phase in the magma.” Pasteris et al. (1996) evaluated inclusions in the Pinatubo dacite phenocrysts, paying particular attention to the liquid and vapor phases. They suggested that the magma may have reached saturation with an H2O-CO2-SO2 supercritical fluid before the melt reached anhydrite saturation, stripping the melt of much of its sulfur, and that in most regions of the magma “anhydrite reached saturation only after quartz phenocrysts had begun to precipitate, meaning that anhydrite was a late phenocryst phase,” p. 885. Pumices from the dacitic pyroclastic-flow and -fall deposits from the June 15, 1991, Mt. Pinatubo eruption were classified by phenocryst abundance into two groups (Imai et al. 1996). Type 1 is white in color and phenocryst-rich (>~20%): whereas type 2 is yellowish in color and phenocryst-poor (<~ 20%). Even though the type 2 pumices are phenocryst-poor samples, they contain abundant microscopic crystal fragments (J. Pallister, personal communication 2001). The pumices sampled in this study, labeled P4 and P2, are type 1 (Fournelle et al. 1996). Phenocrysts are principally plagioclase and hornblende, with lesser amounts of cummingtonite, biotite, quartz, ilmenite, magnetite, apatite, anhydrite, and zircon. Overall, the bulkchemical composition and Fe-Ti oxide temperatures of the two pumice types are similar (Pallister et al. 1996). The S contents of the pumices are variable. Fournelle et al. (1996) attributed this to the variable distribution of anhydrite phenocrysts within the magma. The whole-rock S content of sample P4 is 900 ppm, lower than the 1200 ppm in pumice P2. Chris Newhall (personal communication 1991) collected sample P2 from an upper layer of a pyroclastic-flow deposit one month after the eruption. It was thus exposed to a month of rainfall, and may have originally had more S (Fournelle et al. 1996). Whole-rock S values for both samples are lower than similar samples reported by Bernard et al. (1991) and Pallister et al. (1996), i.e., 1500–2400 ppm. Anhydrite phenocrysts are sparse in the June 15 Mt. Pinatubo pumice, ~0.1 vol% according to Bernard et al. (1996). Over 97 wt% of the S in the pumice is in the form of sulfate (Fournelle et al. 1996). Euhedral anhydrite phenocrysts are parallelepiped (450 ¥ 550 ¥ 200 to 250 ¥ 250 ¥ 100 mm) or rectangular in section (100 ¥ 600 ¥ 50 to 100 ¥ 300 ¥ 50 mm). Subhedral to euhedral anhydrite crystals are surrounded by sharp contacts with vesicular matrix glass (Figs. 1a and 1b) (Bernard et al. 1991; Pallister et al. 1996; Fournelle et al. 1996). Several examples of anhydrite trapped within silicate minerals, specifically hornblende and plagioclase phenocrysts, were found (Figs. 1a, 1c, and 1d) and cited by Fournelle et al. (1996) as evidence that anhydrite was a primary magmatic phase at depth. Identical Sr/Sr ratios (~0.7042) of anhydrite and bulkrock confirm that the anhydrite was not sedimentary (Hattori 1996). The ion microprobe study of McKibben et al. (1996) found a unimodal frequency of dS values averaging near +7 per mil in the June 15 eruption anhydrite, indicating that the (SO4) component in the melt was isotopically uniform and well mixed. On the other hand, anhydrite phenocrysts from the June 12, 1991 Plinian eruption have a bimodal distribution in dS values (+6.5 and +10.5 per mil). The Sr and S isotope data suggest that most of the anhydrite in the June 15 pumices was not derived from hydrothermal or evaporitic anhydrite deposits in the crust below the volcano. In their evaluation of the remotely sensed SO2 emitted by Mt. Pinatubo, as well as melt inclusion data and other constraints, Westrich and Gerlach (1992), Wallace and Gerlach (1994) and Gerlach et al. (1996) suggested that most magmatic S had partitioned into a water-rich vapor phase at depth some time prior to the climactic eruption. A similar conclusion was reached by Pasteris et al. (1996). Gerlach et al. (1996) estimated that S preferentially partitioned into the vapor compared to the melt by a factor of ~720, and Wallace (2000) calculated ~950. Apatite within and adjacent to anhydrite has been observed in thin section, and in microprobe and SEM images [Figs. 1a–1d; Fournelle et al. (1996)]. Most apatite inclusions occur as welldeveloped hexagonal pinacoids and hexagonal prisms. Objective of this study A previous scanning electron microscope (SEM) examination of Pinatubo clasts indicated small hexagonal crystals on the surface of plagioclase in a pumice vesicle (Fournelle et al. 1996; their Fig. 9). Energy-dispersive spectrometry (EDS) Ka peaks of Ca and S suggested that the surface features were a Ca-sulfate phase. The hexagonal shape seemed to conflict with the orthorhombic symmetry of anhydrite. T. Gerlach (personal communication 1992) suggested that X-ray diffraction analysis of Mt. Pinatubo volcanic anhydrite was needed to verify that the appropriate thermodynamic properties were being used in gas modeling. An early observation in our current study was the unexpected existence of micrometer-scale pyramids upon the anhydrite phenocrysts. The goal here is to describe and identify these surface features on the crystal faces of anhydrite phenocrysts from two June 15, 1991, Mt. Pinatubo pumice samples. We consider whether the pyramids are orthorhombic (b) anhydrite, and not another Ca-sulfate phase such as a-CaSO4, g-CaSO4, CaSO4· JAKUBOWSKI ET AL.: MAGMATIC VAPOR DEPOSITION OF ANHYDRITE 1031 0.5H2O (hemihydrate or bassanite), or CaSO4·2H2O (gypsum) (Table 1). This is the first SEM study to examine anhydrite phenocryst surfaces. The computer program SOLVGAS is used here to model possible magmatic conditions for homogeneous anhydrite precipitation or resorption involving vapor. This has important implications for the question of whether or not the Pinatubo system contained a pre-eruptive gas phase prior to the climactic eruption. Overall, this study addresses the Pinatubo S budget, taking into account the observed Ca-S surface features.