The application of ICP-MS methods to tephrochronological problems

The accurate recognition of tephra deposits is of great value to Earth scientists because they facilitate stratigraphic correlation. The most useful tephra deposits form from violent volcanic eruptions; they are isochronous and widespread. Most are dacitic and rhyolitic in composition, and can be difficult to identify unequivocally using major element chemistry alone. Distal tephras are typically thin and are prone to contamination and thus are awkward to analyse by bulk methods. Here, the authors review their previous work in the development of analytical techniques for the analysis of small volumes of glass separates from tephra deposits, both by solution nebulisation and by laser ablation (LA) inductively coupled plasma mass spectrometry (ICP-MS), placing particular emphasis on the precision and accuracy of the various methods. In solution nebulisation ICP-MS, accurate data can be obtained from samples as small as 0.025 g. LA-ICP-MS methods are described for the analysis of small bulk samples and single glass shards as small as 40 μm in diameter. Accurate and reproducible analyses can be achieved by ICP-MS by both solution and laser ablation methods on homogeneous materials. Solution analyses are normally accurate to ±5% and have typical precisions (1 σ) of around ±4% for abundant trace elements (e.g. Zr, Rb) but this can deteriorate to about ±20% for rare elements in small samples (e.g. HREE in a 25 mg sample). Laser ablation methods are slightly less accurate (typically ±5–10%) and precision decreases from about ±3% at concentrations of a few hundred ppm, to about ±10% at 1 ppm and about ±30% at 0.05 ppm. An apparent lack of precision in the bulk analysis of small volumes of glass shards by LA-ICP-MS often represents within sample heterogeneity (and not analytical error), inter-shard variation becoming abundantly clear in some tephra deposits when individual glass shards are analysed. Single grain analysis on shards as small as 40 μm can provide an accurate analysis of the pure glass phase, which may not be achieved in solution or bulk sample LA-ICP-MS methods. Analyses affected by micro-phenocryst phases, such as feldspar or zircon can be easily removed following careful inspection of the data. Single shard LA-ICP-MS also allows any compositional variation within the parental magma to be defined