Diffusion of Sr and Ba in plagioclase: Composition and silica activity dependencies, and application to volcanic rocks

• Sr and Ba diffusion rates in plagioclase are similar. • Source concentration and assemblage control the success of diffusion experiments. • Application of new diffusivities yields incubation rather than eruption timescales. Strontium and barium diffusion chronometry in plagioclase is routinely applied to mafic and felsic magmatic systems. This technique can be used to determine the timescales of magma reservoir assembly and the cooling rates of plutons and volcanic rocks, which has emerged as a useful method to assess volcanic hazards. Here we report diffusion experiments that aim to constrain the diffusivities of Sr and Ba in oligoclase and labradorite at 1 atm pressure, between 900 and 1,200 °C, and assessing diffusion in different crystallographic orientations. In all of the reported experiments, silica activity ( a SiO 2 ) is buffered by varying stable phase assemblages in the diffusant source powder. The experimental products were analysed by secondary ion mass spectrometry (SIMS) depth profiling and laser ablation inductively coupled mass spectrometry (LA-ICP-MS) line scanning. There is no resolvable dependence of Sr and Ba diffusion in plagioclase upon a SiO 2 or crystal orientation. However, Sr and Ba diffusivities are found to vary as functions of the plagioclase anorthite content. As such, we parameterise the diffusivity of Sr and Ba in plagioclase as a function of temperature and anorthite content as follows: log 10 D Sr ( m 2 s − 1 ) = − 1.65 ( ± 0.24 ) X An − 3.03 ( ± 1.16 ) − [ 368 , 142 ( ± 27 , 141 ) 2.303 R T ] , log 10 D Ba ( m 2 s − 1 ) = − 1.43 ( ± 0.20 ) X An − 4.65 ( ± 0.96 ) − [ 337 , 037 ( ± 22 , 969 ) 2.303 R T ] , where X An is the plagioclase anorthite content (mole fraction), T is the temperature (K), and R is the gas constant (J mol -1 K -1 ). The diffusion rate of Sr in plagioclase determined in this study is ∼1.5–2 orders of magnitude slower than previously determined, whereas Ba diffusion is similar to previous studies. This is most likely due to the Ba-feldspar stability at the experimental conditions employed by previous studies, whereas Sr-feldspar was absent from the source powder assemblage. By applying the diffusivities determined in this study to plagioclase crystals from the Cerro Galán ignimbrite (Argentina) and Santorini caldera (Greece), we find timescales of ∼10 5 years, with a good agreement between results from Sr and Ba diffusion modelling. Therefore, our data reconcile experimental diffusion data with measured Sr and Ba profiles in plagioclase and suggest that, at least regarding the Cerro Galán ignimbrite and Santorini caldera, plagioclase records the time needed to differentiate magma reservoirs and assemble large volumes of eruptible magma.