Introduction

Prolonged and heavy exposure to fine, crystalline silica-rich volcanic ash particles could potentially cause chronic, fibrotic disease, such as silicosis, in vulnerable individuals (Horwell and Baxter 2006). Cristobalite is the silica polymorph of prime concern as it may crystallise, post-extrusion, in volcanic lava domes, filling cracks and vesicles, so that it becomes a major mineral phase (up to 12 wt.%). This concern primarily relates to andesitic/dacitic eruptions where cristobalite crystallises in porphyritic domes by vapour-phase deposition and devitrification of glass (Baxter et al. 1999), although it may also be found as an alteration product in old volcanic edifices (Getahun et al. 1996). In rhyolitic, obsidian (glassy) domes, cristobalite is observed in spherulites which slowly nucleate from glass (Swanson et al. 1989; Watkins et al. 2009), but it is not known whether rapid, large-scale cristobalite formation is also possible through vapour-phase deposition in cracks, as with their andesitic/dacitic counterparts.

Rhyolitic eruptions are relatively rare, and the eruption of Chaitén volcano, Chile, (from 2 May 2008 to present) provides the first opportunity to determine whether active rhyolitic domes form substantial cristobalite and could therefore potentially pose a health hazard as in the andesitic eruption on Montserrat (Hincks et al. 2006; Horwell and Baxter 2006). Initial explosions, through a vent in an ancient (approximately 9.3 ka B.P. (Naranjo and Stern 2004)) rhyolitic lava dome, generated ash which fell across Patagonian Chile and Argentina with local towns being temporarily evacuated (e.g. Futaleufú) or abandoned (e.g. Chaitén village). By 21 May, lava dome growth had commenced and is ongoing (at time of press).

Sample collection

A suite of ash samples were collected, deposited from the initial explosive eruption (2–5 May 2008) and thereafter following collapses of the new dome (August 2008, January 2009 and February 2009; Fig. 1 and Table 1). All samples were dried in an oven at 80°C for at least 12 h. Samples were then sieved (Endecott stainless steel sieves) through 1- and 2-mm sieves to remove non-ash-grade particles. In all cases, there were no particles >1 mm in diameter.

Fig. 1
figure 1

Location of collection of samples in relation to Chaitén volcano

Full size image
Table 1 Information on the samples
Full size table

Crystalline silica quantification

A key parameter of silica toxicity assessment, in both occupational and environmental settings, is the accurate quantification of the silica polymorphs present in the dust. For heterogeneous ash samples, this has been challenging (Baxter et al. 1999; Horwell et al. 2003), primarily because of the overlap in the major cristobalite and plagioclase feldspar peaks in the X-ray diffraction (XRD) pattern, but recent innovations in XRD methodology now allow rapid, accurate quantification (Le Blond et al. 2009).

Using the Internal Attenuation Standard (IAS) method of quantification with XRD-static Position Sensitive Detection (sPSD; see Supplementary Information and Le Blond et al. 2009), we find that ash erupted explosively contained little cristobalite (approximately 2 wt.% except one sample with 7.4 wt.%), whereas ash erupted following collapse of the new dome contained 13.9–19.0 wt.% cristobalite (Fig. 2 and Table 2). We hypothesise that the ash from the initial explosions was generated from fresh magma and/or conduit material with minor incorporation of the ancient dome, with cristobalite therefore being sourced from the altered conduit or ancient dome material. The high cristobalite content measured in the ash erupted following onset of new dome growth suggests that the rhyolitic dome formed cristobalite in substantial quantities. By mid-August 2008, 3 months after dome growth began, cristobalite content was already approximately 16 wt.% of the ash, indicating that cristobalite crystallisation is rapid. Quartz observed in the samples is likely to have a magmatic origin, with little variation amongst the samples (0.6–2.3 wt.%); tridymite was not identified.

Fig. 2
figure 2

X-ray diffraction pattern for Chai_10, with the main peaks for cristobalite (c) and plagioclase feldspar (p, labradorite) labelled. The peak labelled by asterisk is a detector glitch (artefact from a damaged channel in this particular detector)

Full size image
Table 2 Quantification of crystalline silica by the Internal Attenuation Standard method with X-ray diffraction with static position sensitive detection and grain size results from Malvern Mastersizer analyses (average of three tests)
Full size table

For comparison, we also used the IAS XRD-sPSD method to test cristobalite contents in ash produced after two lava dome collapses, and a series of vulcanian explosions at the andesitic Soufrière Hills volcano, Montserrat. The dome-collapse samples contained 12–15 wt.% cristobalite, and the vulcanian explosion sample contained 8.6 wt.%, not dissimilar to Chai_02 (Table 2). Horwell et al. (2003) found that the respirable fractions of ash (<4 µm aerodynamic diameter), which can reach the alveolar region of the lung where pathogenic reactions would be triggered, were enriched in crystalline silica in comparison to the coarser fractions (e.g. for sample MRA5/6/99, 30% of <1 µm fraction (by number of particles) was crystalline silica compared with 10 num.% for the >4 µm fraction). The actual quantity of cristobalite in the respired ash may, therefore, be greater than the proportions from bulk samples reported here.

Morphology, grain size and composition

Grain size analysis (using the methodology of Horwell 2007 on the Malvern Mastersizer 2000 with Hydro MU attachment at the Department of Geography, Cambridge University, UK) confirmed that the Chaitén samples have abundant respirable material (8.8–17.7 vol.% <4 µm diameter; Table 2) with sample Chai_09 being particularly fine. Horwell (2007) noted that ash generated in dome-collapse eruptions seems to be as fine, if not more so, than ash generated in plinian explosive eruption columns. X-ray fluorescence (XRF; PANalytical Axios Advanced XRF spectrometer at the Department of Geology, University of Leicester, UK) confirmed that all Chaitén samples are rhyolitic (72.6–74.9 wt.% SiO2).

Field emission scanning electron microscopy (FE-SEM; Philips XL-30 at the Natural History Museum, London, UK) observations confirmed that the general morphology of the ash was similar to previously observed volcanic ash samples (as discussed in Horwell and Baxter 2006). The ash particles are angular, with fractured surfaces, being composed of glass and fragmented crystals. Figure 3a shows a typical view of very fine-grained particles in sample Chai_04. Many of the particles are <1 µm in diameter. In addition to the angular particles, we occasionally observed fibre-like particles (e.g. Fig. 3b). The dimensions of these particles conform to the WHO definition for fibres (>5 µm length, <3 µm diameter with 3:1 aspect ratio; World Health Organisation 1986), but they are needle-like (acicular) single crystals rather than bundles of fibres which separate into individual fibrils (asbestiform or fibrous habit). The largest needle imaged was approximately 15 μm along the longest axis, and the smallest was approximately 2 μm. The short axis was usually between <0.2 and 0.8 μm in diameter. Most of the fibre-like particles were lath-shaped with occasional, longer, cylindrical particles. The samples did vary in the number of fibre-like particles that could be found by SEM. It was easiest to locate the needles in Chai_08 sample, although they were still scarce. Elemental analysis (SEM-energy dispersive spectroscopy (EDS) on a LEO 1455VP with EDS Oxford Inca software at NHM, London, UK) showed that the majority were composed of Si > Al > Na, K and Ca, indicating that they are glass or feldspars (Fig. 3b inset; in this spectrum the Al peak is relatively large, indicating a feldspar; other needles produced spectra with reduced Al and raised K, which is interpreted here, and in the SEM-EDS spectra of Horwell et al. (2003), as volcanic glass).

Fig. 3
figure 3

Scanning electron microscopy (SEM) images of (a) Chai_04 showing a typical view of fine-grained and larger particles. (b) Fibre-like ash particles sitting on a larger ash particle in sample Chai_08. Inset: SEM-energy dispersive spectroscopy spectrum of elemental composition of a typical Chai_08 fibre-like particle

Full size image

Discussion

The similarity in cristobalite content of the Chaitén and the Soufrière Hills dome-collapse ash suggests that the mechanism of cristobalite formation may be similar. This is backed up by visual observations of the Chaitén obsidian which is glassy, crystal-poor, vesicle free and fractured (Luis Lara, SERNAGEOMIN, personal communication), indicating that cristobalite generation through vapour-phase deposition on the walls of degassing pathways is a viable mechanism and that spherulitic formation, for the bulk of the cristobalite, is unlikely.

Recently, Reich et al. (2009) suggested that cristobalite nano-fibres, which they observed by transmission electron microscopy in the Chaitén explosive ash, may be formed through high-temperature vapour-phase reactions in the explosion column. Reich et al. did not quantify the nano-fibres, but here, we did not observe any nano-fibres (by FE-SEM), and the micro-fibres observed were not cristobalite and were sparse. We suggest an alternative mechanism for cristobalite nano-fibre formation, through break up of small spherulites in the ancient dome during the initial explosions in May 2009. Within spherulites, cristobalite grows as axiolitic intergrowths with feldspar, thereby forming acicular structures on fragmentation.

There is some concern as to whether single, acicular crystals and fibre-like particles pose a similar hazard to asbestos fibres. The fibre-like particles observed here, and in Reich et al., are not asbestos minerals (they do not contain any Mg or Fe). Fibre dimension, biopersistence (bio-solubility and breakability) and composition are key factors in determining fibre pulmonary pathogenicity. In addition, biopersistence is affected by the ability of the particles to break longitudinally into fibrils. There is no evidence of fibril formation in the Chaitén ash, with fibre shape indicating transverse breakage giving brittle particles which are more easily cleared by macrophages (Zoltai 1981). No data exist on the toxicity of volcanic glass, feldspar or cristobalite fibres, so we cannot rule out that these fibres may be harmful, but the scarcity of the individual fibre-like particles means that they are unlikely to increase the potential respiratory hazard of the ash.

Summary

The presence of elevated levels of respirable crystalline silica in the form of cristobalite in the dome-related samples from Chaitén volcano indicates that the potential health hazard of the ash from Chaitén volcano increased as the eruption switched style from explosive to dome growth and collapse. The findings justify the air monitoring effort (set up by public health officials in Chile and Argentina in 2009) to assess exposure to ash in the population at risk from ash fallout whilst the eruption continues. Together with the fact that the ash contains significant levels of respirable material, exposure assessment of the local population is vital so that the hazard can be considered with respect to the duration and extent of exposure of populations to these particles.