The geology of ambient aerosols: characterising urban and rural/coastal silicate PM_(10-2.5) and PM_(2.5) using high-volume cascade collection and scanning electron microscopy

摘要

Atmospheric PM_(10-2.5) and PM_(2.5) (fine) samples were obtained, using a high-volume impaction collector, from two very different sites in the UK: rural/coastal (Lizard Peninsula, Cornwall) and urban/road traffic (Central London). Samples were characterized directly on the collecting substrate using SEM (2000 analyses: 500 from each coarse and fine sample). The London PM_(10-2.5) mostly comprises iron, carbon, nitrate, and silicate particles, with minor amounts of sulphate and chloride. The Lizard PM_(10-2.5) is predictably much richer in silicates and chlorides, and relatively low in iron and carbon. The London PM_(2.5) fraction is dominated by fine/ultrafine carbon from road traffic, followed by nitrate and sulphate particles (both in similar numbers to PM_(10-2.5)), but is depleted in iron and, especially, in silicate and chloride. The Lizard PM_(2.5) is also dominated by carbon, sulphate and nitrate, and has much lower amounts of silicate and chloride. The silicate particles in the coarse London PM fraction include many more felsie (mostly quartz and clay minerals) than mafic (Fe-Mg-rich) particles, whereas the reverse is true for the Lizard where the most common silicate PM_(10-2.5) is Mg-rich and derived from the local geology. Silicate PM_(2.5) is virtually absent from the London sample but occurs in appreciable amounts in the Lizard (mostly felsic, especially Al-rich clay minerals) indicating that such silicates in the fine, more deeply inhalable, fraction arise mainly from natural rather than technogenic sources. The distinctive silicate particle chemistry of the Lizard samples provides an excellent example of how local geology will be recorded in the composition of ambient aerosols, and contrasts with the London site which lies in an entirely different geological setting and is extremely contaminated by anthropogenic aerosols.