Rare sulfur and triple oxygen isotope geochemistry of volcanogenic sulfate aerosols

Abstract

We present analyses of stable isotopic ratios 17O/ 16O, 18O/ 16O, 34S/ 32S, and 33S/ 32S, 36S/ 32S in sulfate leached from volcanic ash of a series of well known, large and small volcanic eruptions. We consider eruptions of Mt. St. Helens (Washington, 1980, ∼1 km 3), Mt. Spurr (Alaska, 1953, <1 km 3), Gjalp (Iceland, 1996, 1998, <1 km 3), Pinatubo (Phillipines, 1991, 10 km 3), Bishop tuff (Long Valley, California, 0.76 Ma, 750 km 3), Lower Bandelier tuff (Toledo Caldera, New Mexico, 1.61 Ma, 600 km 3), and Lava Creek and Huckleberry Ridge tuffs (Yellowstone, Wyoming, 0.64 Ma, 1000 km 3 and 2.04 Ma 2500 km 3, respectively). This list covers much of the diversity of sizes and the character of silicic volcanic eruptions. Particular emphasis is paid to the Lava Creek tuff for which we present wide geographic sample coverage. This global dataset spans a significant range in δ 34S, δ 18O, and Δ 17O of sulfate (29‰, 30‰, and 3.3‰, respectively) with oxygen isotopes recording mass-independent ( Δ 17O > 0.2‰) and sulfur isotopes exhibiting mass-dependent behavior. Products of large eruptions account for most of‘ these isotopic ranges. Sulfate with Δ 17O > 0.2‰ is present as 1–10 μm gypsum crystals on distal ash particles and records the isotopic signature of stratospheric photochemical reactions. Sediments that embed ash layers do not contain sulfate or contain little sulfate with Δ 17O near 0‰, suggesting that the observed sulfate in ash is of volcanic origin. Mass-dependent fractionation of sulfur isotopic ratios suggests that sulfate-forming reactions did not involve photolysis of SO 2, like that inferred for pre-2.3 Ga sulfates from Archean sediments or Antarctic ice-core sulfate associated with few dated eruptions. Even though the sulfate sulfur isotopic compositions reflect mass-dependent processes, the products of caldera-forming eruptions display a large δ 34S range and exhibit fractionation relationships that do not follow the expected equilibrium slopes of 0.515 and 1.90 for 33S/ 32S vs. 34S/ 32S and 36S/ 32S vs. 34S/ 32S, respectively. The data presented here are consistent with modification of a chemical mass-dependent fractionation of sulfur isotopes in the volcanic plume by either a kinetic gas phase reaction of volcanic SO 2 with OH and/or a Rayleigh processes involving a residual Rayleigh reactant—volcanic SO 2 gas, rather than a Rayleigh product. These results may also imply at least two removal pathways for SO 2 in volcanic plumes. Above-zero Δ 17O values and their positive correlation with δ 18O in sulfate can be explained by oxidation by high- δ 18O and high- Δ 17O compounds such as ozone and radicals such as OH that result from ozone break down. Large caldera-forming eruptions have the highest Δ 17O values, and the largest range of δ 18O, which can be explained by stratospheric reaction with ozone-derived OH radicals. These results suggest that massive eruptions are capable of causing a temporary depletion of the ozone layer. Such depletion may be many times that of the measured 3–8% depletion following 1991 Pinatubo eruption, if the amount of sulfur dioxide released scales with the amount of ozone depletion.