Fuel-type characterization of carbonaceous fly-ash particles using EDS-derived surface chemistries and its application to particles extracted from lake sediments

Abstract

Carbonaceous particles produced from the high temperature combustion of fossil fuels are found in high concentrations (e.g. more than 50 000 per gram dry mass, gDM -1 ) in the upper levels of lake sediment cores taken from regions of high acid deposition. The sediment record of these particles correlates well with the record of acidification as indicated by diatom analysis. The temporal distribution of particle concentrations in sediments follows consistent trends in lakes throughout Europe and has been used for indirect dating purposes. However, more information is available from particles extracted from lake sediments when their fuel-type is known. For example, potential sources of associated airborne pollutants can be determined and additional dates for sediment cores may possibly be obtained. This paper describes the development of a carbonaceous particle characterization for oil and coal fuel-types, using surface chemistries determined by energy dispersive spectroscopy (eds). A reference dataset was produced using ashes obtained from power stations in the UK and Ireland. Linear discriminant function analysis of the eds generated data produced a discrimination rule which allocated over 95% of the particles to the correct fuel-type under cross-validation. The technique was applied to carbonaceous particles extracted from a 210 Pb-dated sediment core taken from a lake in North London. The results showed good agreement with known combustion history of coal and oil. The spatial distribution of characterized particles was also studied using surface sediments from lakes in Scotland, northern England and Northern Ireland. Above average concentrations of oil particles occurred around Glasgow, the east coast near Aberdeen, and Galloway in the southwest. This suggests that the acidification of lochs in the Galloway area may partly relate to pollutants originating from oil-fired power stations in Northern Ireland. Attempts to use the characterization to identify power station sources was only partially successful. The range of chemistries in the present dataset do not include the full range produced from each individual power station as the fuels burnt at each station vary through time. Particles from peat combustion were included in a principal components analysis and show that the technique could be extended to include other fuel-types. This technique is applicable not only to particles stored in lake sediments, but also to particles collected in any environment, e.g. building stones, deposition gauges, soils, leaves, etc., and could be applied to a range of environmental questions in Britain, Europe and on a global scale.