- Email: [email protected]
Geochemical characterization of single atmospheric particles from the Eyjafjallajökull volcano eruption event collected at ground-based sampling sites in Germany
Atmospheric Environment 48 (2012) 113e121
Contents lists available at ScienceDirect
Atmospheric Environment journal homepage: www.elsevier.com/locate/atmosenv
Geochemical characterization of single atmospheric particles from the Eyjafjallajökull volcano eruption event collected at ground-based sampling sites in Germany Nina Schleicher a, *, Utz Kramar a, Volker Dietze b, Uwe Kaminski b, Stefan Norra a, c a b c
Institute of Mineralogy and Geochemistry, Karlsruhe Institute of Technology, Adenauerring 20b, 76131 Karlsruhe, Germany Air Quality Department, Research Center Human Biometeorology, German Meteorological Service, Stefan-Meier-Str. 4, 79104 Freiburg i. Br., Germany Institute of Geography and Geoecology, Karlsruhe Institute of Technology, Kaiserstr. 12, 76128 Karlsruhe, Germany
a r t i c l e i n f o
a b s t r a c t
Article history: Received 8 February 2011 Received in revised form 25 April 2011 Accepted 11 May 2011
Volcanic particles can be transported over long distances in the atmosphere and can cause severe problems for air traffic. This was the case over large areas of Europe in spring 2010 after the eruption of the Eyjafjallajökull (E15) volcano on Iceland. The scope of this work was to characterize these volcanic particles more in detail with regard to size and chemical composition in order to provide valuable information needed for a better estimation of the possible impact on airplane jet engines and cockpit windows. Another question of this study was which share of the overall atmospheric particles in Germany originated from the E15 eruption and whether this amount of volcanic particles could cause any adverse health effects to humans. To this end, single particle analysis by means of scanning electron microscopy with energy dispersive X-ray analysis (SEM-EDX) and synchrotron radiation based micro X-ray fluorescence analysis (mS-XRF) together with multivariate statistical methods were applied for samples collected on ground-level in Southwest Germany and Iceland. Based on the obtained chemical fingerprints combined with multivariate statistical methods it was possible to discrimate between the amount of volcanic particles from Iceland and other atmospheric particles from non-volcanic sources. This aspect distinguishes this single particle approach from most other studies. The results of the study showed that at least 40% of the analyzed particles between 2.5 and 10 mm size at the remote sampling sites in the Black Forest area and about 25% in the city of Freiburg were clearly of volcanic origin from the E15 volcano eruption event. Ó 2011 Elsevier Ltd. All rights reserved.
Keywords: Volcanic particles Eyjafjallajökull (E15) Single particle analysis SEM-EDX Microsynchrotron radiation based X-ray fluorescence analysis (mS-XRF) Geochemical fingerprint
1. Introduction and motivation The eruption of the Eyjafjallajökull (E15) volcano on Iceland starting in March 2010 distributed volcanic particles over large areas in Europe and caused severe problems for air traffic. The strong northwesterly winds over Iceland at the time of eruption carried fine-grained siliceous ash southeastwards into the crowded airspace of the UK and continental Europe and caused the largest aerial shutdown in Europe since World War II and affected at least ten million passengers worldwide (Petersen, 2010). Numerous flights were canceled in order to avoid dangerous encounters of airplanes with the ash plume, such as the famous incident of a Boeing 747 in 1982 after the eruption of the Galunggang volcano in Indonesia (Brooker, 2010; Durant et al., 2010; Guffanti et al.,
* Corresponding author. E-mail address: [email protected] (N. Schleicher). 1352-2310/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.atmosenv.2011.05.034
2010). However, at the time of the E15 eruption, no detailed scientific data were available in order to estimate the real danger for airplanes. Of most potential danger is volcanic ash that deposits in the fuel nozzles, combustor, and turbine of an airplane and restricts airflow through the engine, which leads to the loss of engine thrust (Brooker, 2010). For the investigation of such possible impacts of volcanic particles on airplane jet engines and the “sandblasting” effect on cockpit windows, the characterization of these particles in detail with regard to size, shape, and mass concentrations is of tremendous importance. Furthermore, the chemical composition of the particles helps to draw conclusions about their mineralogy and hardness. Sharp and hard ash fragments, such as volcanic glass of siliceous composition, produce the immediate damage in major incidents and damages external glass, plastic and metals of the aircraft (Brooker, 2010). For areas thousands of kilometers away from the volcano (e.g. Germany), the central question is the characterization of the longrange transported volcanic particles reaching ground. Beside the
114
N. Schleicher et al. / Atmospheric Environment 48 (2012) 113e121
mass concentration of these volcanic particles also their size distribution is of special interest. The focus of this work is a detailed geochemical and mineralogical characterization of single atmospheric particles collected in April 2010 at several ground-based sampling sites in Southwest Germany as well as the determination of the size distribution and the spatial distribution of the volcanic particles. Based on chemical fingerprints combined with multivariate statistical methods, the amount of volcanic particles will be separated from the non-volcanic overall dust load. Consequently, it is possible to quantify the actual amount of volcanic particles from the Eyjafjallajökull, which were transported to Germany. Most studies, which already investigated the E15 eruption event, applied remote sensing methods (Ajtai et al., 2010; Ansmann et al., 2010; Flentje et al., 2010; Colette et al., 2011; Gasteiger et al., 2011; Schäfer et al., 2011), measurements by aircrafts (Schumann et al., 2011), bulk analysis of near-source samples (Gislason et al., 2011; Ilyinskaya et al., 2011), or bulk analysis of PM10 filter samples (Colette et al., 2011). With these approaches it is, in contrast to our study, not possible to discriminate between individual volcanic particles from Iceland and the non-volcanic atmospheric particles. The German Meteorological Service (Deutscher Wetterdienst, DWD) is running a nationwide network of passive samplers for the coarse fraction of atmospheric particles (2.5e80 mm). Single atmospheric particles were collected on a weekly basis on transparent adhesive collection plates. Within this study, samples of the event week (16/04/2010e23/04/2010) were compared to the sampling weeks before and after the volcanic particles reached Southwest Germany. The area around Freiburg and the Black Forest can be regarded as a hot spot for the deposition of volcanic particles and was therefore chosen for more detailed investigations. The high deposition of volcanic particles is reflected by high mass concentrations of particles between 2.5 and 10 mm collected passively and also by the PM10 concentrations measured within the monitoring network of the German Federal States and the German
Federal Environment Agency (Umweltbundesamt, UBA). Fig. 1 shows a map of the calculated daily PM10 distribution over Southwest Germany at the 19th of April 2010. At this day, the main fall out of the volcanic particles occurred in the Southwest and, therefore, this study focuses on the high concentrations near Freiburg. With the scope to identify particles of volcanic origin and to calculate the amount of volcanic particles leading to the high particle concentrations during the event week, fingerprints of the chemical composition of volcanic particles collected directly on Iceland were established. The collection of individual atmospheric particles provides an excellent starting position for further analysis by optical and X-ray techniques (SEM-EDX, mS-XRF). Contrary to bulk analysis, which is regularly done by active PM10 sampling, a detailed knowledge about size, shape, and chemical composition can be gained without interfering influences by other non-volcanic particles. For the characterization of atmospheric aerosols, SEMEDX is an established method, which is often applied for source apportionment studies (e.g. Post and Buseck, 1984; Weinbruch et al., 1997; Sobanska et al., 2003; Ebert et al., 2004; Vester et al., 2007; Schleicher et al., 2010a). Also some studies applied SEMEDX analysis for volcanic particles from different volcanoes (Ersoy et al., 2007; Mills and Rose, 2010; Lautze et al., 2011). On the contrary, only very few studies already used mS-XRF methods for single atmospheric particle studies (Yue et al., 2004, 2006; Godelitsas et al., 2009; Schleicher et al., 2010b). The advantage of synchrotron radiation compared to SEM-EDX is the higher intensity and resolution and, thus, the lower detection limits within the range of sub to few ppm for most trace elements in particles with a diameter of 2e10 mm. This enables the application of trace element fingerprints to single aerosol particles. Generally, trace element fingerprints are more characteristic for the different sources than major element fingerprints. With regard to volcanic particles in general and those of the E15 eruption in particular, our approach to use single particles analysis of SEM-EDX and mS-XRF combined with multivariate statistical methods was applied for the first time for single particle samples from sampling sites at groundlevel within this study. 2. Experimental 2.1. Sampling Single atmospheric particles were collected on a weekly basis on transparent adhesive collection plates at ground-based sampling sites located in Southwest Germany. These sites belong to a monitoring network of the German Meteorological Service (DWD). The passive sampling technique with a Sigma-2 device was performed according to VDI guideline 2119 (VDI, 2011). Samples were collected during the event week (16/04/2010e23/04/2010) as well as during the weeks before and after the volcanic particles reached Germany. The six sampling sites selected for detailed analyses are located in the city of Freiburg (300 m MSL), and at four remote sites in the Black Forest, Baiersbronn (585 m MSL), Bad Märgen (895 m MSL), Hinterzarten (885 m MSL) and Wolfach (305 m MSL), and as well as at the Schauinsland mountain (1200 m MSL). Additionally, samples of volcanic particles from the vicinity of the Eyjafjallajökull volcano on Iceland were analyzed as reference material in order to get a geochemical fingerprint of the volcanic material from the respective eruption.
Fig. 1. Ambient air quality map of PM10 concentration in Southwest Germany at the 19th April 2010 (map after UBA, 2010: http://www.umweltbundesamt.de). Data based on measurements from the monitoring network of the German States and the Federal Environment Agency (UBA).
2.2. Optical microscopy A quantitative computer-aided automatic optical microscope technique was carried out at the German Meteorological Service
N. Schleicher et al. / Atmospheric Environment 48 (2012) 113e121
115
The synchrotron radiation at ANKA is produced by electrons of a 2.5 GeV storage ring. The source at the FLUO-beamline is a 1.5 T bending magnet with a critical energy of about 6 keV. A double multilayer monochromator (WeSi multilayers in 2.7 nm period) is used at the FLUO-beamline. As focusing optics x-ray lenses (CRL) were applied and the energy used was 22 kV. Measurement time for each particle was 1000 s. At Beamline-L at the DORIS III synchrotron ring HASYLAB/DESY, the synchrotron radiation originates from positrons with 4.5 GeV at a bending magnet with a radius of 12.12 m. The critical energy is 16.6 keV. The excitation energy was adjusted to 20.7 keV by a Ni/C, Mo/Si multilayer monochromator. The incident beam was focussed to a diameter of 15 mm by using a cross-slit system and a subsequent polycapillary. Currently in total 425 aerosol particles of a diameter between 1 and 10 mm have been measured by mS-XRF. Due to the very small particle size no concentrations but absolute element masses within the individual particles are obtained, e.g. the mass of a volcanic glass sphere with diameter between 1 mm and 10 mm is in the range of w1.3 1012 g to w1.3 109 g. A concentration of 100 mg g1 of an element is than equivalent to w1.3 1016 g to w1.3 1013 g or between 9 10þ05 and 9 10þ08 atoms for example for Sr. Fig. 2. Scanning electron microscope picture of particles collected during the event week (16/04/2010e23/04/2010) at the remote sampling sites in the Black Forest, Southwest Germany. At a magnification of 500-times, 65 particles were randomly chosen for SEM-EDX analysis.
(DWD) in Freiburg, Germany. Projected areas of individual particles were automatically measured in the size range between 2.5 and 80 mm (geometrical equivalence diameter). Mass concentrations were calculated as described in Dietze et al. (2006) and Schleicher et al. (2010b).
2.3. Scanning electron microscopy Scanning electron microscopy (SEM) was carried out for various single particles and their chemical composition was determined by energy dispersive X-ray analysis (EDX). Analyses were carried out at the Laboratory for Electron Microscopy (LEM) at the Karlsruhe Institute of Technology (KIT), Germany, with a FEI Quanta 650 FEG Instrument. About 60 particles in the size range between 2 and 10 mm were chosen randomly for each sampling site (currently in total 270 particles). The particle distribution on the adhesive plates collected during the event week and the selected particles for SEMEDX analysis at a magnification of 500-times can be seen in Fig. 2 for the remote Black Forest site as an example. Measurement time for EDX analysis of each particle was 60 s.
2.4. Synchrotron radiation based X-ray fluorescence Single particles were analyzed for their elemental composition by means of energy-dispersive synchrotron radiation based micro X-ray fluorescence analysis (mS-XRF). The advantage of mS-XRF compared to SEM-EDX is the higher intensity and resolution and, thus, the lower detection limits within the range of sub to few ppm for most trace elements in particles with a diameter of 2e10 mm. For this analysis, the FLUO-beamline at ANKA (Angström-Quelle Karlsruhe, Karlsruhe Institute of Technology, Germany) and the beamline-L at the light source DORIS at HASYLAB/DESY (Deutsches Elektronensynchrotron, Hamburg, Germany), were used. The transparent adhesive collection plates could be used directly without any further preparation. The measurements were carried out under atmospheric conditions.
2.5. Analysis of material collected on Iceland as reference samples Volcanic ash fall out directly from Eyjafjallajökull volcano was used as reference sample. These particles were collected directly on the surface on Iceland near to the volcano as bulk samples. Some of these reference particles were later placed onto blank transparent
Table 1 Bulk concentrations of the two reference samples (ref-01 and ref-02) collected directly on Iceland close to the Eyjafjallajökull volcano. (A) Results from HR-ICP-MS analysis after acid digestion; (B) results from EDX and WDX (bold values) analysis. (A)
ICP-MS ref-01 (B)
EDX/WDX ref-01 EDX/WDX ref-02
Li [mg kg1] Na [g kg1] Mg [g kg1] Al [g kg1] P [g kg1] (S) [mg kg1] K [g kg1] Ca [g kg1] Sc [g kg1] Ti [g kg1] V [mg kg1] Cr [mg kg1] Mn [g kg1] Fe [g kg1] Co [mg kg1] Ni [mg kg1] Cu [mg kg1] Zn [mg kg1] Ga [mg kg1] As [mg kg1] Rb [mg kg1] Sr [mg kg1] Y [mg kg1] Zr [mg kg1] Nb [mg kg1] Mo [mg kg1] Ag [mg kg1] Cd [mg kg1] Sn [mg kg1] Sb [mg kg1] Cs [mg kg1] Ba [mg kg1] Pb [mg kg1] U [mg kg1]
11 40 13 73 1.5 520 15 35 15 10 74 48 1.9 67 13 23 24 180 29 1.4 45 220 82 390 50 3.6 3.2 0.5 3.8 0.3 0.5 380 3.3 1.7
5.2 2.8 14 57 0.3 680 1.9 5.0 1.5 0.2 9.7 19 30 143 25 41 310 87 480 62 610 61 120 9.5
Na2O [%] MgO [%] Al2O3 [%] SiO2 [%] P2O5 [%] (S) [mg kg-1] K2O [%] CaO [%] TiO2 [%] MnO [%] Fe2O3 [%] Ni [mg kg-1] Cu [mg kg-1] Zn [mg kg-1] Ga [mg kg-1] Rb [mg kg-1] Sr [mg kg-1] Y [mg kg-1] Zr [mg kg-1] Nb [mg kg-1] Ba [mg kg-1] La [mg kg-1] Ce [mg kg-1] Pb [mg kg-1]
5.3 2.4 15 58 0.3 570 2.0 4.7 1.4 0.2 9.2 22 25 145 26 45 290 94 500 62 610 62 120 10
116
N. Schleicher et al. / Atmospheric Environment 48 (2012) 113e121
adhesive plates in order to replicate the measurement conditions of the airborne particles collected in Germany with the Sigma-2 passive sampler device. The samples prepared that way were also analyzed by SEM-EDX (see Section 2.3) and mS-XRF (see Section 2.4). Additionally, the bulk chemical composition of the reference material was determined by EDX/WDX analysis and by highresolution inductively coupled plasma mass spectrometry (HRICP-MS, Axiom, VG Elemental) after acid digestion. The volcanic material was digested in Teflon vessels with HClO4, HNO3, and HF (Merck, p.a.). 2.6. Cluster analysis Statistical analysis was carried out using the software package STATISTICA (Release 8.0) by StatSoft, Inc. (USA). Intensities of X-ray lines obtained by SEM depend on both the concentration of the respective element and on the shape of the surface at the measuring point. Therefore, the sum of the observed elements (O, Na, Mg, Al, Si, P, S, Cl, K, Ca, Ti, Mn, Fe, Cu, Zn and Sn) was rescaled to obtain a constant sum of 100% for each measured particle. Afterward the data set was z-transformed using median value and interquartile distance. In total, 208 particles were included in the cluster analysis. In order to group similar particles, K-means cluster analysis was performed in three consecutive steps. Beside the particles form Southwest Germany, several EDXRF-spectra of fall out particles, collected on Iceland close to the eruption site, were included into the data set as well as particles collected in Southwest Germany one week before the eruption. In a first step clustercenters and memberships were calculated for seven clusters. In this first step mainly anthropogenic particles, salts, phosphates, sulfates and carbonates were separated from silicate particles. In the second step the three cluster characterizing siliceous particles were subdivided and with the last step the group containing the potentially
volcanic particles was refined. After this step all Iceland particles clustered in one group together with the particles of probable volcanic origin. For the synchrotron EDXRF data the intensities for the analyzed elements strongly depend on the particle diameter to the power of three. In order to eliminate the particle size effect the element masses were rescaled by normalization to a constant sum of the determined elements weighted with the reciprocal concentrations of the respective elements in volcanic glass sampled at the Eyjafjallajökull eruption site. After this step K-means cluster analysis based on Euclidian distances was carried out using the median and inter-quartile distance normalized relative elemental values:
c*i ¼ ðci Medi Þ=ðQ 3i Q 1i Þ (e.g. Steinhausen and Langer, 1977) with: c*i –> normalized value of element (i); ci e value of element (i); Medi e Median of element (i); Q1i e 1st quartile (25%) of element (i); Q3i e 3rd quartile (75%) of element (i). For the trace element determinations, in the last step, particle agglomerations were considered by a one-step Fuzzy cluster (Abonyi and Feil, 2007; Kramar, 1995).
3. Results and discussion 3.1. Geochemical fingerprints of the volcanic particles from Iceland Most of the geogenic aerosol particles in the investigated grain size range (2e10 mm) normally consist of a single mineral. Therefore, the major and trace element chemistry of these particles is dominated by few elements, which are characteristic for the respective mineral. Most anthropogenic particles are expected to show high concentrations of few elements, which are characteristic
Fig. 3. Two characteristic volcanic particles (top) and their respective element spectra (bottom) obtained by SEM-EDX analysis. Particles were collected at the surface directly in the vicinity of the Eyjafjallajökull volcano on Iceland (reference material).
N. Schleicher et al. / Atmospheric Environment 48 (2012) 113e121
for a definite pollution source e.g. an alloy, pigment particles or others. On the contrary, the chemical fingerprint of volcanic glass shows a complex multi-element fingerprint, consisting of all major and trace elements typical for this kind of volcanism. Such a fingerprint may be difficult to identify in a bulk dust sample, but can be easily identified from the major and trace element composition of single particles. The Eyjafjallajökull volcano is situated in a propagating rift outside the main zone of plate spreading in Iceland, at the southern termination of the eastern rift zone (Sigmundsson et al., 2010). It has an alkaline composition, similar to other off-rift volcanoes on Iceland (Sturkell et al., 2010). The bulk composition of reference material collected on the surface close to the volcano (labeled as ref-01 and ref-02) is summarized in Table 1. Since element ratios are more robust than elemental concentrations, characteristic element ratios for the Eyjafjallajökull volcanic glass are given in the following. Additionally, the complete elemental patterns were used to assign the particles to their sources. Characteristic major element ratios of the E15 are Na/Mg of about 2.2, Al/Si of about 0.29, K/Ca of about 0.45, and Ti/Fe of about 0.13. Characteristic trace element ratios are for example Sr/Zr of about 0.65, Sr/Rb of about 6, and Zn/Cu of about 6. Apart from the bulk composition, also single particles from the reference material were analyzed by SEM-EDX and mS-XRF. Fig. 3 shows characteristic SEM images of two volcanic particles from the reference material and their respective EDX spectra.
117
3.2. Characterization of volcanic particles deposited in Southwest Germany 3.2.1. Size distributed particle concentrations Automated optical microscopy with subsequent calculation of the particle mass concentration showed high concentrations in the size range from 2.5 to 10 mm at the sampling sites in Southwest Germany during the week from the 16th to the 23rd of April 2010 (“event week”). Backward trajectories indicate the Eyjafjallajökull volcano as a possible particle source (Fig. 4). The trajectories shown in Fig. 4 are representative for the dust fall event week. During the sampling weeks before and after the particles from the E15 volcano eruption event reached Southwest Germany, average concentrations of particles from 2.5 to 10 mm were about 5.2 mg m3 (Fig. 5b). These concentrations were more than two times higher during the event week. For example, concentrations up to 13 mg m3 for particles from 2.5 to 10 mm were observed at the “Schauinsland” mountain station (1200 m MSL) near Freiburg during the event week (Fig. 5a). Similar concentrations were also observed at the other locations, such as Freiburg (Fig. 5a). In Fig. 5a, the size distribution of coarse atmospheric particles collected during this event week (CW 15/10) at Freiburg (dashed line) and at the Schauinsland (bold line), Germany, are compared to the average size distribution before and after the event (Fig. 5b). During the event week, particle concentrations for the size class between 2.5 and 5 mm were approximately 3-times higher in Freiburg and at the
Fig. 4. Backward trajectories arriving in Freiburg, Germany, at the 17th of April 2010.
118
N. Schleicher et al. / Atmospheric Environment 48 (2012) 113e121
Schauinsland, compared to the average concentrations of eight weeks before and after this event week. An increased concentration was also observed for the particles between 5 and 10 mm of about 2.5-times at Freiburg and 3-times at the Schauinsland, respectively. The relatively strong increase for particles smaller than 10 mm during the event week can be explained by the long-range transport from Iceland since the residence time of finer particles in the atmosphere is longer (Seinfeld and Pandis, 2006). Beside the fractions between 2.5 and 20 mm, the concentrations of particles between 20 and 40 mm were also slightly higher during the event week (Fig. 5). These increased concentrations in the fraction larger than 20 mm can be explained by local resuspension processes favored by dry weather conditions during the sampling period. This assumption is supported by the observation that the increased concentration for this particles size fraction was more pronounced in Freiburg, where the traffic intensity and other anthropogenic activities are higher, compared to the Schauinsland site (Fig. 5a). Particles in this size range are not likely to be transported over long distances since they would have settled in approximately 24e48 h residence time from the source (Jaenicke, 1978).
Fig. 5. Calculated mass concentrations of coarse atmospheric particles in the size range from 2.5 to 80 mm collected at Freiburg (dashed line) and Schauinsland (bold line), Germany. (a) Sampling week from 16/04/2010e23/04/2010 (event week), (b) average for 17 weeks from 19/02/2010e25/06/2010.
3.2.2. Geochemical characterization of single aerosol particles collected in Southwest Germany As mentioned in Section 3.2.1, the high concentrations of particles between 2.5 and 10 mm and the backward trajectories indicated that high concentrations of volcanic particles were transported to Germany. However, as described above, only the investigation of the chemical and mineralogical composition of the particles collected during the event week provides a reliable tool to discriminate between volcanic and other particles and to calculate the concentration of volcanic particles arriving in Germany in April 2010. Therefore, the chemical composition of individual particles measured by SEM-EDX analysis was compared to the element composition of reference samples collected on Iceland near the
Fig. 6. Two characteristic volcanic particles (top) and their respective element spectra (bottom) obtained by SEM-EDX analysis. Particles were collected at the Schauinsland, Germany, during the event week (16/04/2010e23/04/2010).
N. Schleicher et al. / Atmospheric Environment 48 (2012) 113e121
119
Fig. 7 shows mS-XRF spectra measured as trace element fingerprints of non-volcanic particles (Fig. 7c) and an ash fall out (Fig. 7b), both collected at the remote sampling site in the Black Forest. Additionally, an ash fall out particle collected on Iceland close to the eruption site is shown in Fig. 7a.
Fig. 7. mS-XRF spectra measured as trace element fingerprints of (A) an ash fall out particle collected in Iceland close to the eruption site, (B) an ash fall out particle collected in the Black Forest, (C) a non-volcanic particle collected in the Black Forest.
volcanic source. With this approach, a characteristic fingerprint for the volcanic particles from the E15 volcano eruption was obtained (Section 3.1). Fig. 6 shows SEM images and EDX spectra of two typical particles collected in Southwest Germany and identified as originating form the Eyjafjallajökull volcano.
3.2.3. Estimation of the amount of volcanic particles and of particles from other sources From SEM-EDX and mS-XRF analysis it was estimated that at least 40% of the analyzed particles within the size range of 2.5e10 mm in Southwest Germany were volcanic particles with the characteristic Eyjafjallajökull fingerprint. Cluster analysis helped to group the respective particle types. The average relative element concentrations for each of the eleven clusters are listed in Table 2. Cluster Cl3-1 represents the volcanic particles. All four particles from the reference material included in the cluster analysis clearly plot in this cluster. Cl3-2 represents siliceous particles, which are similar to the volcanic particle composition, but this cluster includes also some pre-event particles. The other clusters represent chlorides (NaCl and KCl, Cl8þ9), anthropogenic particles (e.g. color pigments, rust, metal abrasives Cl10, gypsum Cl11) and several silicate minerals and/or mineral agglomerates (Cl1, 2, 4, 5, 6). About 30e40% of the analyzed particles could be clearly identified as volcanic particles from the Eyjafjallajökull volcano. Highest concentrations were found at the Schauinsland with about 60% volcanic particles. The remote sites in the Black Forest showed lower contribution of volcanic particles but in general comparable results. The particles analyzed from Freiburg showed a much higher load of regional dust particles (w23%). This leads to a relative decrease of the amount of particles from volcanic origin. However, the volcanic cluster (cluster Cl3-1 in Table 2) only includes the particles with the typical siliceous glass composition. Some of the other particles might also be transported from the E15 eruption and, thus, the share of particles from the Eyjafjallajökull would be higher. In this regard especially more information about the material of the volcanic cone itself and the sediments around the volcano (below the glacier), which are also carried into the atmosphere by the eruption, would be needed.
Table 2 Average relative concentrations of all eleven clusters obtained after three-step Cluster analyses. Before clustering the relative element concentrations were normalized (ztransformed) using median and inter-quartile distance. The cluster-centers (means of all samples belonging to the respective cluster) are rescaled to the relative concentrations. Srel: relative standard deviation. The characteristic elements of each cluster are marked in bold. Cluster-center Cl1 Particle number srel O Na Mg Al Si P S Cl K Ca Ti Cr Mn Fe Cu Zn Sn
Cl2
Cl3-1 Volc!
Cl3-2
Cl4
Cl5
Cl6
35
4
87
27
8
18
5
17%
50%
10%
19%
35%
24%
45%
40.2 0.88 3.78 6.23 19.55 0.32 0.75 0.38 0.82 2.02 1.73 0.23 0.42 21.38 0.56 0.48 0.27
32.6 0.54 0.24 10.93 40.72 0.13 0.36 0.14 11.15 0.66 0.25 0.15 0.24 0.89 0.48 0.32 0.18
40.2 1.65 2.02 8.50 30.42 0.29 0.60 0.29 1.46 1.95 0.94 0.2 0.3 10.0 0.5 0.4 0.2
39.8 1.90 1.92 9.21 31.70 0.31 0.58 0.27 2.31 1.54 0.69 0.2 0.2 8.4 0.4 0.3 0.2
36.5 0.78 1.98 3.41 16.85 0.28 0.89 0.22 0.54 5.73 10.80 0.18 0.41 20.31 0.51 0.34 0.31
35.4 1.22 2.82 7.96 28.33 0.35 2.08 0.88 0.87 10.71 0.84 0.19 0.49 6.16 0.62 0.54 0.48
31.8 1.32 1.63 3.37 13.01 0.79 0.57 0.21 0.65 1.13 29.51 0.16 0.69 13.90 0.53 0.50 0.20
Cl7 1
Cl8 3
Cl9 1
58% 21.6 0.68 7.57 0.35 0.45 27.91 8.82 0.36 9.24 11.18 0.53 0.14 5.46 0.71 1.59 1.76 1.63
6.8 38.46 0.85 0.17 0.36 0.22 6.08 41.73 1.35 0.05 0.36 0.36 0.71 0.39 0.96 0.46 0.71
5.2 0.51 0.63 0.21 0.40 0.09 0.14 26.11 61.12 0.21 0.92 0.88 0.42 0.12 1.87 0.67 0.52
Cl10
Cl11
2
17
71%
24%
16.5 0.26 0.27 0.42 0.60 0.61 8.28 0.81 0.36 0.45 0.37 0.67 0.48 0.22 59.81 1.77 8.10
29.5 3.74 5.94 1.33 2.80 1.73 13.67 1.13 1.42 29.14 0.86 0.89 1.03 1.05 2.33 1.86 1.59
120
N. Schleicher et al. / Atmospheric Environment 48 (2012) 113e121
Germany. Therefore, further single particle analyses including coarser particles are additionally planned in order to verify whether even these coarser volcanic particles reached Southwest Germany. For a better estimation of the negative effects on airplane jet engines, it is recommended that experiments on turbines should be carried out with similar particle sizes as a function of distance from the eruption site, shape and hardness like the volcanic particles identified by the chemical fingerprints within this study. Based on the relatively low amount of volcanic particles determined on the ground in the Black Forest area and the short time of exposure, no adverse health effects on humans can be expected in Southwest Germany.
Fig. 8. E15-rationed values of the centers of the 6 clusters (trcl1etrcl6) obtained from the trace element data measured by mS-XRF.
Fig. 8 shows ratios of the E15 normalized trace element concentration values. The cluster trcl1 with high loads of Ca and Sr is indicative for calcite. The cluster trcl2 with high loads of K and Rb is indicative for orthoclase, whereas trcl3 with high loads of Cr, Ni and Pb can be interpreted as anthropogenic. In trcl4 nearly all ratios are plotting close to 1, thus the volcanic particles can be assigned to this cluster. Cluster trcl5 shows high ratios for Cu, Zn and Pb. This cluster can be interpreted either as anthropogenic metals or as particles from sulphide mineralization in the Black Forest and the high Ti content in trcl6 may be an indication for the presence of pigment particles possibly from wall or car paintings. 4. Conclusions and outlook It could be shown that volcanic particles from the E15 volcanic eruption reached the ground in Southwest Germany. Single particle analysis by means of SEM-EDX and mS-XRF provided a suitable fingerprint for the composition of the volcanic particles and enables the identification of probable mineral phases of the different particles and their type of sources (volcanic event, anthropogenic e.g. traffic, local geology and possible mineralization). The study showed that the combination of single particle major and trace element analysis together with data evaluation by modern multivariate statistical methods, which were applied in this context for the first time, proved to be a good tool for the identification of the volcanic particles. The deposition of volcanic particles from the E15 in Southwest Germany was proven by the recovery of the fingerprints from the reference samples collected directly on Iceland. Thus, the assumption from the PM10 distribution map (Fig. 1) that around Freiburg and the Black Forest a hot spot of E15 particle deposition occurred was verified by geochemical analyses in this study. In the remote Black Forest sites, at least 40% of the analyzed particles between 2.5 and 10 mm were identified as volcanic glass originating clearly from the Eyjafjallajökull. Due to the higher anthropogenic dust load in the city of Freiburg, the relative amount of volcanic particles is reduced to approximately 23%. Further analysis are planned in order to increase the amount of analyzed particles and, thus, the statistical reliability, and to obtain a more detailed picture of the spatial distribution of the fall out in Southwest Germany. In this regard, the focus on particles with grain sizes between 2.5 and 10 mm is in good accordance with a study by Flentje et al. (2010) in which the authors found particles up to 7.5 mm size in the volcanic plume over Germany. However, another study by Ansmann et al. (2010) reported even coarser particles larger than 20 mm in the ash layers over Leipzig and Munich,
Acknowledgments SEM-EDX analyses were carried out at the Laboratory for Electron Microscopy (LEM), Karlsruhe Institute of Technology (KIT), Karlsruhe, Germany. The authors express their special gratitude to Volker Zibat for his support with SEM-EDX analyses. Synchrotron radiation based XRF measurements were carried out at ANKA (Angström-Quelle Karlsruhe), KIT, and DORIS at HASYLAB/DESY, Hamburg. In this regard, the authors want to thank David Batchelor for assistance with setting up the FLUO-beamline at ANKA and Karen Appel and Manuela Borchert for assistance with beamline-L at HASYLAB/DESY. Furthermore, the authors want to thank Andreas Holbach and Sarah Perktold, KIT, for support with measurements at ANKA and LEM, respectively. Reference material collected on Iceland was provided by Natalia Garcia Villora, ITS, KIT, Karlsruhe, Germany, and by staff of SRI International, NSF Sondrestrom Research Facility, Greenland. References Ajtai, N., Stefanie, H., Stoian, L., Oprea, M., 2010. The volcanic ash and its impact on European air transport industry. A case study on the detection and impact of the Eyjafjallajökull volcanic ash plume over North-Western Europe between 14th and 21st April 2010. Advances in Environmental Sciences 2010 (2), 57e68. Abonyi, J., Feil, B., 2007. Cluster Analysis for Data Mining and System Identification. Birkhäuser, Basel, Boston, Berlin. Ansmann, A., Tesche, M., Groß, S., Freudenthaler, V., Seifert, P., Hiebsch, A., Schmidt, J., Wandinger, U., Mattis, I., Müller, D., Wiegner, M., 2010. The 16 April 2010 major volcanic ash plume over central Europe: EARLINET lidar and AERONET photometer observations at Leipzig and Munich, Germany. Geophysical Research Letters 37, L13810. Brooker, P., 2010. Fear in a handful of dust: aviation and the Icelandic volcano. Significance 7, 112e115. Colette, A., Favez, O., Meleux, F., Chiappini, L., Haeffelin, M., Morille, Y., Malherbe, L., Papin, A., Bessagnet, B., Menut, L., Leoz, E., Rouïl, L., 2011. Assessing in near real time the impact of the April 2010 Eyjafjallajökull ash plume on air quality. Atmospheric Environment 45, 1217e1221. Dietze, V., Fricker, M., Goltzsche, M., Schultz, E., 2006. Luftqualitaetsmessungen in deutschen Kurorten Teil 1: Methodik und Absicherung. Gefahrstoffe e Reinhaltung der Luft 66, 45e53. Durant, A.J., Bonadonna, C., Horwell, C.J., 2010. Atmospheric and environmental impacts of volcanic particulates. Elements 6, 235e240. Ebert, M., Weinbruch, S., Hoffmann, P., Ortner, H.M., 2004. The chemical composition and complex refractive index of rural and urban influenced aerosols determined by individual particle analysis. Atmospheric Environment 38, 6531e6545. Ersoy, O., Gourgaud, A., Aydar, E., Chinga, G., Thouret, J.C., 2007. Quantitative scanning-electron microscope analysis of volcanic ash surfaces: application to the 1982e1983 Galunggung eruption (Indonesia). Geological Society of America Bulletin 119, 743e752. Flentje, H., Claude, H., Elste, T., Gilge, S., Köhler, U., Plass-Dülmer, C., Steinbrecht, W., Thomas, W., Werner, A., Fricke, W., 2010. The Eyjafjallajökull eruption in April 2010 e detection of volcanic plume using in-situ measurements, ozone sondes and lidar-ceilometer profiles. Atmospheric Chemistry and Physics 10, 10085e10092. Gasteiger, J., Groß, S., Freudenthaler, V., Wiegner, M., 2011. Volcanic ash from Iceland over Munich: mass concentration retrieved from ground-based remote sensing measurements. Atmospheric Chemistry and Physics 11, 2209e2223.
N. Schleicher et al. / Atmospheric Environment 48 (2012) 113e121 Gislason, S.R., Alfredsson, H.A., Eiriksdottir, E.S., Hassenkam, T., Stipp, S.L.S., 2011. Volcanic ash from the 2010 Eyjafjallajökull eruption. Applied Geochemistry. doi:10.1016/j.apgeochem.2011.03.100. Godelitsas, A., Toli, K., Simon, R., 2009. Elemental Distribution in Saharan Dust from Red Rain Precipitated over Athens, Greece, ANKA Annual Report 2009, pp. 178e179. Guffanti, M., Casadevall, T.J., Budding, K., 2010. Encounters of Aircraft with Volcanic Ash Clouds; a Compilation of Known Incidents, 1953e2009: U.S. Geological Survey Data Series 545, Ver. 1.0. 12 , plus 4 appendixes including the compilation database, available only at. http://pubs.usgs.gov/ds/545 pp. Ilyinskaya, E., Tsanev, V., Martin, R., Oppenheimer, C., Le Blond, J., Sawyer, G., Gudmundsson, M., 2011. Near-source observations of aerosol size distributions in the eruptive plumes from Eyjafjallajökull volcano, MarcheApril 2010 Atmospheric Environment 45, 3210e3216. Jaenicke, R., 1978. Über die Dynamik atmosphärischer Aitkenteilchen. Berichte der Bunsen-Gesellschaft für Physikalische Chemie 82, S1198e1202. Kramar, U., 1995. Application of limited fuzzy-clusters for anomaly recognition in complex geological environment. Journal of Geochemical Exploration 55, 81e92. Lautze, N.C., Taddeucci, J., Andronico, D., Cannata, C., Tornetta, L., Scarlato, P., Houghton, B., Castro, M.D.L., 2011. SEM-based methods for the analysis of basaltic ash from weak explosive activity at Etna in 2006 and the 2007 eruptive crisis at Stromboli. Physics and Chemistry of the Earth. doi:10.1016/ j.pce.2011.02.001. Mills, O.P., Rose, W.I., 2010. Shape and surface area measurements using scanning electron microscope stereo-pair images of volcanic ash particles. Geosphere 6. doi:10.1130/GES00558.1. Petersen, G.N., 2010. A short meteorological overview of the Eyjafjallajökull eruption 14 Aprile23 May 2010. Weather 65, 203e207. Post, J.E., Buseck, P.R., 1984. Characterization of individual particles in the phoenix urban aerosol using electron-beam instruments. Environmental Science and Technology 18, 35e42. Schäfer, K., Thomas, W., Peters, A., Ries, L., Obleitner, F., Schnelle-Kreis, J., Birmili, W., Diemer, J., Fricke, W., Junkermann, W., Pitz, M., Emeis, S., Forkel, R., Suppan, P., Flentje, H., Wichmann, H.E., Gilge, S., Meinhardt, F., Zimmermann, R., Weinhold, K., Soentgen, J., Münkel, C., Freuer, C., Cyrys, J., 2011. Influences of the 2010 Eyjafjallajökull volcanic plume on air quality in the northern Alpine region. Atmospheric Chemistry and Physics Discussions 11, 9083e9132. Schleicher, N., Norra, S., Chai, F., Chen, Y., Wang, S., Stüben, D., 2010a. Anthropogenic versus geogenic contribution to total suspended atmospheric particulate matter and its variations during a two-year sampling period in Beijing, China. Journal of Environmental Monitoring 12, 434e441. Schleicher, N., Kramar, U., Norra, S., Dietze, V., Kaminski, U., Cen, K., Yu, Y., 2010b. mscale variations of elemental composition in individual atmospheric particles by means of synchrotron radiation based m-XRF analysis. X-RAY OPTICS AND
121
MICROANALYSIS: proceedings of the 20th international congress. AIP Conference Proceedings 1221, 172e180. Schumann, U., Weinzierl, B., Reitebuch, O., Schlager, H., Minikin, A., Forster, C., Baumann, R., Sailer, T., Graf, K., Mannstein, H., Voigt, C., Rahm, S., Simmet, R., Scheibe, M., Lichtenstern, M., Stock, P., Rueba, H., Schäuble, D., Tafferner, A., Rautenhaus, M., Gerz, T., Ziereis, H., Krautstrunk, M., Mallaun, C., Gayet, J.F., Lieke, K., Kandler, K., Ebert, M., Weinbruch, S., Stohl, A., Gasteiger, J., Olafsson, H., Sturm, K., 2011. Airborne observations of the Eyjafjalla volcano ash cloud over Europe during air space closure in April and May 2010. Atmospheric Chemistry and Physics 11, 2245e2279. Seinfeld, J., Pandis, S., 2006. Atmospheric Chemistry and Physics: From Air Pollution to Climate Change. John Wiley & Sons, Inc, Hoboken, New Jersey. Sigmundsson, F., Hreinsdóttir, S., Hooper, A., Árnadóttir, T., Pedersen, R., Roberts, M.J., Óskarsson, N., Auriac, A., Decriem, J., Einarsson, P., Geirsson, H., Hensch, M., Ófeigsson, B.G., Sturkell, E., Sveinbjörnsson, H., Feigl, K.L., 2010. Intrusion triggering of the 2010 Eyjafjallajökull explosive eruption. Nature 468, 426e430. Sobanska, S., Maenhaut, W., Coeur, C., Adams, F., 2003. SEM-EDX Characterisation of tropospheric aerosols in the Negev Desert (Israel). Journal of Atmospheric Chemistry 44, 299e322. Steinhausen, D., Langer, K., 1977. Clusteranalyse e Einführung in Methoden und Verfahren der automatischen Klassifikation. Walter de Gruyter, Berlin, New York. Sturkell, E., Einarsson, P., Sigmundsson, F., Hooper, A., Ófeigsson, B.G., Geirsson, H., Ólafsson, H., 2010. Katla and Eyjafjallajökull volcanoes. Developments in Quarterny Sciences 13, 5e21. VDI, 2011. VDI Guideline 2119: Partikelprobenahme mit einem Grobstaubsammler (Sigma-2) zur Charakterisierung von Einzelpartikeln und Berechnung der Konzentration. Vester, B.P., Ebert, M., Barnert, E.B., Schneider, J., Kandler, K., Schütz, L., Weinbruch, S., 2007. Composition and mixing state of the urban background aerosol in the Rhein-Main area (Germany). Atmospheric Environment 41, 6102e6115. Weinbruch, S., Wentzel, M., Kluckner, M., Hoffmann, P., Ortner, H.M., 1997. Characterization of individual atmospheric particles by element mapping in electron probe microanalysis. Mikrochimica Acta 125, 137e141. Yue, W., Li, Y., Li, X., Yu, X., Deng, B., Liu, J., Wan, T., Zhang, G., Huang, Y., He, W., Hua, W., 2004. Source identification of PM10, collected at a heavy traffic roadside, by analyzing individual particles using synchrotron radiation. Journal of Synchrotron Radiation 11, 428e431. Yue, W., Li, X., Liu, J., Li, Y., Yu, X., Deng, B., Wan, T., Zhang, G., Huang, Y., He, W., Hua, W., Shao, L., Li, W., Yang, S., 2006. Characterization of PM2.5 in the ambient air of Shanghai city by analyzing individual particles. Science of the Total Environment 368, 916e925.
Contents lists available at ScienceDirect
Atmospheric Environment journal homepage: www.elsevier.com/locate/atmosenv
Geochemical characterization of single atmospheric particles from the Eyjafjallajökull volcano eruption event collected at ground-based sampling sites in Germany Nina Schleicher a, *, Utz Kramar a, Volker Dietze b, Uwe Kaminski b, Stefan Norra a, c a b c
Institute of Mineralogy and Geochemistry, Karlsruhe Institute of Technology, Adenauerring 20b, 76131 Karlsruhe, Germany Air Quality Department, Research Center Human Biometeorology, German Meteorological Service, Stefan-Meier-Str. 4, 79104 Freiburg i. Br., Germany Institute of Geography and Geoecology, Karlsruhe Institute of Technology, Kaiserstr. 12, 76128 Karlsruhe, Germany
a r t i c l e i n f o
a b s t r a c t
Article history: Received 8 February 2011 Received in revised form 25 April 2011 Accepted 11 May 2011
Volcanic particles can be transported over long distances in the atmosphere and can cause severe problems for air traffic. This was the case over large areas of Europe in spring 2010 after the eruption of the Eyjafjallajökull (E15) volcano on Iceland. The scope of this work was to characterize these volcanic particles more in detail with regard to size and chemical composition in order to provide valuable information needed for a better estimation of the possible impact on airplane jet engines and cockpit windows. Another question of this study was which share of the overall atmospheric particles in Germany originated from the E15 eruption and whether this amount of volcanic particles could cause any adverse health effects to humans. To this end, single particle analysis by means of scanning electron microscopy with energy dispersive X-ray analysis (SEM-EDX) and synchrotron radiation based micro X-ray fluorescence analysis (mS-XRF) together with multivariate statistical methods were applied for samples collected on ground-level in Southwest Germany and Iceland. Based on the obtained chemical fingerprints combined with multivariate statistical methods it was possible to discrimate between the amount of volcanic particles from Iceland and other atmospheric particles from non-volcanic sources. This aspect distinguishes this single particle approach from most other studies. The results of the study showed that at least 40% of the analyzed particles between 2.5 and 10 mm size at the remote sampling sites in the Black Forest area and about 25% in the city of Freiburg were clearly of volcanic origin from the E15 volcano eruption event. Ó 2011 Elsevier Ltd. All rights reserved.
Keywords: Volcanic particles Eyjafjallajökull (E15) Single particle analysis SEM-EDX Microsynchrotron radiation based X-ray fluorescence analysis (mS-XRF) Geochemical fingerprint
1. Introduction and motivation The eruption of the Eyjafjallajökull (E15) volcano on Iceland starting in March 2010 distributed volcanic particles over large areas in Europe and caused severe problems for air traffic. The strong northwesterly winds over Iceland at the time of eruption carried fine-grained siliceous ash southeastwards into the crowded airspace of the UK and continental Europe and caused the largest aerial shutdown in Europe since World War II and affected at least ten million passengers worldwide (Petersen, 2010). Numerous flights were canceled in order to avoid dangerous encounters of airplanes with the ash plume, such as the famous incident of a Boeing 747 in 1982 after the eruption of the Galunggang volcano in Indonesia (Brooker, 2010; Durant et al., 2010; Guffanti et al.,
* Corresponding author. E-mail address: [email protected] (N. Schleicher). 1352-2310/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.atmosenv.2011.05.034
2010). However, at the time of the E15 eruption, no detailed scientific data were available in order to estimate the real danger for airplanes. Of most potential danger is volcanic ash that deposits in the fuel nozzles, combustor, and turbine of an airplane and restricts airflow through the engine, which leads to the loss of engine thrust (Brooker, 2010). For the investigation of such possible impacts of volcanic particles on airplane jet engines and the “sandblasting” effect on cockpit windows, the characterization of these particles in detail with regard to size, shape, and mass concentrations is of tremendous importance. Furthermore, the chemical composition of the particles helps to draw conclusions about their mineralogy and hardness. Sharp and hard ash fragments, such as volcanic glass of siliceous composition, produce the immediate damage in major incidents and damages external glass, plastic and metals of the aircraft (Brooker, 2010). For areas thousands of kilometers away from the volcano (e.g. Germany), the central question is the characterization of the longrange transported volcanic particles reaching ground. Beside the
114
N. Schleicher et al. / Atmospheric Environment 48 (2012) 113e121
mass concentration of these volcanic particles also their size distribution is of special interest. The focus of this work is a detailed geochemical and mineralogical characterization of single atmospheric particles collected in April 2010 at several ground-based sampling sites in Southwest Germany as well as the determination of the size distribution and the spatial distribution of the volcanic particles. Based on chemical fingerprints combined with multivariate statistical methods, the amount of volcanic particles will be separated from the non-volcanic overall dust load. Consequently, it is possible to quantify the actual amount of volcanic particles from the Eyjafjallajökull, which were transported to Germany. Most studies, which already investigated the E15 eruption event, applied remote sensing methods (Ajtai et al., 2010; Ansmann et al., 2010; Flentje et al., 2010; Colette et al., 2011; Gasteiger et al., 2011; Schäfer et al., 2011), measurements by aircrafts (Schumann et al., 2011), bulk analysis of near-source samples (Gislason et al., 2011; Ilyinskaya et al., 2011), or bulk analysis of PM10 filter samples (Colette et al., 2011). With these approaches it is, in contrast to our study, not possible to discriminate between individual volcanic particles from Iceland and the non-volcanic atmospheric particles. The German Meteorological Service (Deutscher Wetterdienst, DWD) is running a nationwide network of passive samplers for the coarse fraction of atmospheric particles (2.5e80 mm). Single atmospheric particles were collected on a weekly basis on transparent adhesive collection plates. Within this study, samples of the event week (16/04/2010e23/04/2010) were compared to the sampling weeks before and after the volcanic particles reached Southwest Germany. The area around Freiburg and the Black Forest can be regarded as a hot spot for the deposition of volcanic particles and was therefore chosen for more detailed investigations. The high deposition of volcanic particles is reflected by high mass concentrations of particles between 2.5 and 10 mm collected passively and also by the PM10 concentrations measured within the monitoring network of the German Federal States and the German
Federal Environment Agency (Umweltbundesamt, UBA). Fig. 1 shows a map of the calculated daily PM10 distribution over Southwest Germany at the 19th of April 2010. At this day, the main fall out of the volcanic particles occurred in the Southwest and, therefore, this study focuses on the high concentrations near Freiburg. With the scope to identify particles of volcanic origin and to calculate the amount of volcanic particles leading to the high particle concentrations during the event week, fingerprints of the chemical composition of volcanic particles collected directly on Iceland were established. The collection of individual atmospheric particles provides an excellent starting position for further analysis by optical and X-ray techniques (SEM-EDX, mS-XRF). Contrary to bulk analysis, which is regularly done by active PM10 sampling, a detailed knowledge about size, shape, and chemical composition can be gained without interfering influences by other non-volcanic particles. For the characterization of atmospheric aerosols, SEMEDX is an established method, which is often applied for source apportionment studies (e.g. Post and Buseck, 1984; Weinbruch et al., 1997; Sobanska et al., 2003; Ebert et al., 2004; Vester et al., 2007; Schleicher et al., 2010a). Also some studies applied SEMEDX analysis for volcanic particles from different volcanoes (Ersoy et al., 2007; Mills and Rose, 2010; Lautze et al., 2011). On the contrary, only very few studies already used mS-XRF methods for single atmospheric particle studies (Yue et al., 2004, 2006; Godelitsas et al., 2009; Schleicher et al., 2010b). The advantage of synchrotron radiation compared to SEM-EDX is the higher intensity and resolution and, thus, the lower detection limits within the range of sub to few ppm for most trace elements in particles with a diameter of 2e10 mm. This enables the application of trace element fingerprints to single aerosol particles. Generally, trace element fingerprints are more characteristic for the different sources than major element fingerprints. With regard to volcanic particles in general and those of the E15 eruption in particular, our approach to use single particles analysis of SEM-EDX and mS-XRF combined with multivariate statistical methods was applied for the first time for single particle samples from sampling sites at groundlevel within this study. 2. Experimental 2.1. Sampling Single atmospheric particles were collected on a weekly basis on transparent adhesive collection plates at ground-based sampling sites located in Southwest Germany. These sites belong to a monitoring network of the German Meteorological Service (DWD). The passive sampling technique with a Sigma-2 device was performed according to VDI guideline 2119 (VDI, 2011). Samples were collected during the event week (16/04/2010e23/04/2010) as well as during the weeks before and after the volcanic particles reached Germany. The six sampling sites selected for detailed analyses are located in the city of Freiburg (300 m MSL), and at four remote sites in the Black Forest, Baiersbronn (585 m MSL), Bad Märgen (895 m MSL), Hinterzarten (885 m MSL) and Wolfach (305 m MSL), and as well as at the Schauinsland mountain (1200 m MSL). Additionally, samples of volcanic particles from the vicinity of the Eyjafjallajökull volcano on Iceland were analyzed as reference material in order to get a geochemical fingerprint of the volcanic material from the respective eruption.
Fig. 1. Ambient air quality map of PM10 concentration in Southwest Germany at the 19th April 2010 (map after UBA, 2010: http://www.umweltbundesamt.de). Data based on measurements from the monitoring network of the German States and the Federal Environment Agency (UBA).
2.2. Optical microscopy A quantitative computer-aided automatic optical microscope technique was carried out at the German Meteorological Service
N. Schleicher et al. / Atmospheric Environment 48 (2012) 113e121
115
The synchrotron radiation at ANKA is produced by electrons of a 2.5 GeV storage ring. The source at the FLUO-beamline is a 1.5 T bending magnet with a critical energy of about 6 keV. A double multilayer monochromator (WeSi multilayers in 2.7 nm period) is used at the FLUO-beamline. As focusing optics x-ray lenses (CRL) were applied and the energy used was 22 kV. Measurement time for each particle was 1000 s. At Beamline-L at the DORIS III synchrotron ring HASYLAB/DESY, the synchrotron radiation originates from positrons with 4.5 GeV at a bending magnet with a radius of 12.12 m. The critical energy is 16.6 keV. The excitation energy was adjusted to 20.7 keV by a Ni/C, Mo/Si multilayer monochromator. The incident beam was focussed to a diameter of 15 mm by using a cross-slit system and a subsequent polycapillary. Currently in total 425 aerosol particles of a diameter between 1 and 10 mm have been measured by mS-XRF. Due to the very small particle size no concentrations but absolute element masses within the individual particles are obtained, e.g. the mass of a volcanic glass sphere with diameter between 1 mm and 10 mm is in the range of w1.3 1012 g to w1.3 109 g. A concentration of 100 mg g1 of an element is than equivalent to w1.3 1016 g to w1.3 1013 g or between 9 10þ05 and 9 10þ08 atoms for example for Sr. Fig. 2. Scanning electron microscope picture of particles collected during the event week (16/04/2010e23/04/2010) at the remote sampling sites in the Black Forest, Southwest Germany. At a magnification of 500-times, 65 particles were randomly chosen for SEM-EDX analysis.
(DWD) in Freiburg, Germany. Projected areas of individual particles were automatically measured in the size range between 2.5 and 80 mm (geometrical equivalence diameter). Mass concentrations were calculated as described in Dietze et al. (2006) and Schleicher et al. (2010b).
2.3. Scanning electron microscopy Scanning electron microscopy (SEM) was carried out for various single particles and their chemical composition was determined by energy dispersive X-ray analysis (EDX). Analyses were carried out at the Laboratory for Electron Microscopy (LEM) at the Karlsruhe Institute of Technology (KIT), Germany, with a FEI Quanta 650 FEG Instrument. About 60 particles in the size range between 2 and 10 mm were chosen randomly for each sampling site (currently in total 270 particles). The particle distribution on the adhesive plates collected during the event week and the selected particles for SEMEDX analysis at a magnification of 500-times can be seen in Fig. 2 for the remote Black Forest site as an example. Measurement time for EDX analysis of each particle was 60 s.
2.4. Synchrotron radiation based X-ray fluorescence Single particles were analyzed for their elemental composition by means of energy-dispersive synchrotron radiation based micro X-ray fluorescence analysis (mS-XRF). The advantage of mS-XRF compared to SEM-EDX is the higher intensity and resolution and, thus, the lower detection limits within the range of sub to few ppm for most trace elements in particles with a diameter of 2e10 mm. For this analysis, the FLUO-beamline at ANKA (Angström-Quelle Karlsruhe, Karlsruhe Institute of Technology, Germany) and the beamline-L at the light source DORIS at HASYLAB/DESY (Deutsches Elektronensynchrotron, Hamburg, Germany), were used. The transparent adhesive collection plates could be used directly without any further preparation. The measurements were carried out under atmospheric conditions.
2.5. Analysis of material collected on Iceland as reference samples Volcanic ash fall out directly from Eyjafjallajökull volcano was used as reference sample. These particles were collected directly on the surface on Iceland near to the volcano as bulk samples. Some of these reference particles were later placed onto blank transparent
Table 1 Bulk concentrations of the two reference samples (ref-01 and ref-02) collected directly on Iceland close to the Eyjafjallajökull volcano. (A) Results from HR-ICP-MS analysis after acid digestion; (B) results from EDX and WDX (bold values) analysis. (A)
ICP-MS ref-01 (B)
EDX/WDX ref-01 EDX/WDX ref-02
Li [mg kg1] Na [g kg1] Mg [g kg1] Al [g kg1] P [g kg1] (S) [mg kg1] K [g kg1] Ca [g kg1] Sc [g kg1] Ti [g kg1] V [mg kg1] Cr [mg kg1] Mn [g kg1] Fe [g kg1] Co [mg kg1] Ni [mg kg1] Cu [mg kg1] Zn [mg kg1] Ga [mg kg1] As [mg kg1] Rb [mg kg1] Sr [mg kg1] Y [mg kg1] Zr [mg kg1] Nb [mg kg1] Mo [mg kg1] Ag [mg kg1] Cd [mg kg1] Sn [mg kg1] Sb [mg kg1] Cs [mg kg1] Ba [mg kg1] Pb [mg kg1] U [mg kg1]
11 40 13 73 1.5 520 15 35 15 10 74 48 1.9 67 13 23 24 180 29 1.4 45 220 82 390 50 3.6 3.2 0.5 3.8 0.3 0.5 380 3.3 1.7
5.2 2.8 14 57 0.3 680 1.9 5.0 1.5 0.2 9.7 19 30 143 25 41 310 87 480 62 610 61 120 9.5
Na2O [%] MgO [%] Al2O3 [%] SiO2 [%] P2O5 [%] (S) [mg kg-1] K2O [%] CaO [%] TiO2 [%] MnO [%] Fe2O3 [%] Ni [mg kg-1] Cu [mg kg-1] Zn [mg kg-1] Ga [mg kg-1] Rb [mg kg-1] Sr [mg kg-1] Y [mg kg-1] Zr [mg kg-1] Nb [mg kg-1] Ba [mg kg-1] La [mg kg-1] Ce [mg kg-1] Pb [mg kg-1]
5.3 2.4 15 58 0.3 570 2.0 4.7 1.4 0.2 9.2 22 25 145 26 45 290 94 500 62 610 62 120 10
116
N. Schleicher et al. / Atmospheric Environment 48 (2012) 113e121
adhesive plates in order to replicate the measurement conditions of the airborne particles collected in Germany with the Sigma-2 passive sampler device. The samples prepared that way were also analyzed by SEM-EDX (see Section 2.3) and mS-XRF (see Section 2.4). Additionally, the bulk chemical composition of the reference material was determined by EDX/WDX analysis and by highresolution inductively coupled plasma mass spectrometry (HRICP-MS, Axiom, VG Elemental) after acid digestion. The volcanic material was digested in Teflon vessels with HClO4, HNO3, and HF (Merck, p.a.). 2.6. Cluster analysis Statistical analysis was carried out using the software package STATISTICA (Release 8.0) by StatSoft, Inc. (USA). Intensities of X-ray lines obtained by SEM depend on both the concentration of the respective element and on the shape of the surface at the measuring point. Therefore, the sum of the observed elements (O, Na, Mg, Al, Si, P, S, Cl, K, Ca, Ti, Mn, Fe, Cu, Zn and Sn) was rescaled to obtain a constant sum of 100% for each measured particle. Afterward the data set was z-transformed using median value and interquartile distance. In total, 208 particles were included in the cluster analysis. In order to group similar particles, K-means cluster analysis was performed in three consecutive steps. Beside the particles form Southwest Germany, several EDXRF-spectra of fall out particles, collected on Iceland close to the eruption site, were included into the data set as well as particles collected in Southwest Germany one week before the eruption. In a first step clustercenters and memberships were calculated for seven clusters. In this first step mainly anthropogenic particles, salts, phosphates, sulfates and carbonates were separated from silicate particles. In the second step the three cluster characterizing siliceous particles were subdivided and with the last step the group containing the potentially
volcanic particles was refined. After this step all Iceland particles clustered in one group together with the particles of probable volcanic origin. For the synchrotron EDXRF data the intensities for the analyzed elements strongly depend on the particle diameter to the power of three. In order to eliminate the particle size effect the element masses were rescaled by normalization to a constant sum of the determined elements weighted with the reciprocal concentrations of the respective elements in volcanic glass sampled at the Eyjafjallajökull eruption site. After this step K-means cluster analysis based on Euclidian distances was carried out using the median and inter-quartile distance normalized relative elemental values:
c*i ¼ ðci Medi Þ=ðQ 3i Q 1i Þ (e.g. Steinhausen and Langer, 1977) with: c*i –> normalized value of element (i); ci e value of element (i); Medi e Median of element (i); Q1i e 1st quartile (25%) of element (i); Q3i e 3rd quartile (75%) of element (i). For the trace element determinations, in the last step, particle agglomerations were considered by a one-step Fuzzy cluster (Abonyi and Feil, 2007; Kramar, 1995).
3. Results and discussion 3.1. Geochemical fingerprints of the volcanic particles from Iceland Most of the geogenic aerosol particles in the investigated grain size range (2e10 mm) normally consist of a single mineral. Therefore, the major and trace element chemistry of these particles is dominated by few elements, which are characteristic for the respective mineral. Most anthropogenic particles are expected to show high concentrations of few elements, which are characteristic
Fig. 3. Two characteristic volcanic particles (top) and their respective element spectra (bottom) obtained by SEM-EDX analysis. Particles were collected at the surface directly in the vicinity of the Eyjafjallajökull volcano on Iceland (reference material).
N. Schleicher et al. / Atmospheric Environment 48 (2012) 113e121
for a definite pollution source e.g. an alloy, pigment particles or others. On the contrary, the chemical fingerprint of volcanic glass shows a complex multi-element fingerprint, consisting of all major and trace elements typical for this kind of volcanism. Such a fingerprint may be difficult to identify in a bulk dust sample, but can be easily identified from the major and trace element composition of single particles. The Eyjafjallajökull volcano is situated in a propagating rift outside the main zone of plate spreading in Iceland, at the southern termination of the eastern rift zone (Sigmundsson et al., 2010). It has an alkaline composition, similar to other off-rift volcanoes on Iceland (Sturkell et al., 2010). The bulk composition of reference material collected on the surface close to the volcano (labeled as ref-01 and ref-02) is summarized in Table 1. Since element ratios are more robust than elemental concentrations, characteristic element ratios for the Eyjafjallajökull volcanic glass are given in the following. Additionally, the complete elemental patterns were used to assign the particles to their sources. Characteristic major element ratios of the E15 are Na/Mg of about 2.2, Al/Si of about 0.29, K/Ca of about 0.45, and Ti/Fe of about 0.13. Characteristic trace element ratios are for example Sr/Zr of about 0.65, Sr/Rb of about 6, and Zn/Cu of about 6. Apart from the bulk composition, also single particles from the reference material were analyzed by SEM-EDX and mS-XRF. Fig. 3 shows characteristic SEM images of two volcanic particles from the reference material and their respective EDX spectra.
117
3.2. Characterization of volcanic particles deposited in Southwest Germany 3.2.1. Size distributed particle concentrations Automated optical microscopy with subsequent calculation of the particle mass concentration showed high concentrations in the size range from 2.5 to 10 mm at the sampling sites in Southwest Germany during the week from the 16th to the 23rd of April 2010 (“event week”). Backward trajectories indicate the Eyjafjallajökull volcano as a possible particle source (Fig. 4). The trajectories shown in Fig. 4 are representative for the dust fall event week. During the sampling weeks before and after the particles from the E15 volcano eruption event reached Southwest Germany, average concentrations of particles from 2.5 to 10 mm were about 5.2 mg m3 (Fig. 5b). These concentrations were more than two times higher during the event week. For example, concentrations up to 13 mg m3 for particles from 2.5 to 10 mm were observed at the “Schauinsland” mountain station (1200 m MSL) near Freiburg during the event week (Fig. 5a). Similar concentrations were also observed at the other locations, such as Freiburg (Fig. 5a). In Fig. 5a, the size distribution of coarse atmospheric particles collected during this event week (CW 15/10) at Freiburg (dashed line) and at the Schauinsland (bold line), Germany, are compared to the average size distribution before and after the event (Fig. 5b). During the event week, particle concentrations for the size class between 2.5 and 5 mm were approximately 3-times higher in Freiburg and at the
Fig. 4. Backward trajectories arriving in Freiburg, Germany, at the 17th of April 2010.
118
N. Schleicher et al. / Atmospheric Environment 48 (2012) 113e121
Schauinsland, compared to the average concentrations of eight weeks before and after this event week. An increased concentration was also observed for the particles between 5 and 10 mm of about 2.5-times at Freiburg and 3-times at the Schauinsland, respectively. The relatively strong increase for particles smaller than 10 mm during the event week can be explained by the long-range transport from Iceland since the residence time of finer particles in the atmosphere is longer (Seinfeld and Pandis, 2006). Beside the fractions between 2.5 and 20 mm, the concentrations of particles between 20 and 40 mm were also slightly higher during the event week (Fig. 5). These increased concentrations in the fraction larger than 20 mm can be explained by local resuspension processes favored by dry weather conditions during the sampling period. This assumption is supported by the observation that the increased concentration for this particles size fraction was more pronounced in Freiburg, where the traffic intensity and other anthropogenic activities are higher, compared to the Schauinsland site (Fig. 5a). Particles in this size range are not likely to be transported over long distances since they would have settled in approximately 24e48 h residence time from the source (Jaenicke, 1978).
Fig. 5. Calculated mass concentrations of coarse atmospheric particles in the size range from 2.5 to 80 mm collected at Freiburg (dashed line) and Schauinsland (bold line), Germany. (a) Sampling week from 16/04/2010e23/04/2010 (event week), (b) average for 17 weeks from 19/02/2010e25/06/2010.
3.2.2. Geochemical characterization of single aerosol particles collected in Southwest Germany As mentioned in Section 3.2.1, the high concentrations of particles between 2.5 and 10 mm and the backward trajectories indicated that high concentrations of volcanic particles were transported to Germany. However, as described above, only the investigation of the chemical and mineralogical composition of the particles collected during the event week provides a reliable tool to discriminate between volcanic and other particles and to calculate the concentration of volcanic particles arriving in Germany in April 2010. Therefore, the chemical composition of individual particles measured by SEM-EDX analysis was compared to the element composition of reference samples collected on Iceland near the
Fig. 6. Two characteristic volcanic particles (top) and their respective element spectra (bottom) obtained by SEM-EDX analysis. Particles were collected at the Schauinsland, Germany, during the event week (16/04/2010e23/04/2010).
N. Schleicher et al. / Atmospheric Environment 48 (2012) 113e121
119
Fig. 7 shows mS-XRF spectra measured as trace element fingerprints of non-volcanic particles (Fig. 7c) and an ash fall out (Fig. 7b), both collected at the remote sampling site in the Black Forest. Additionally, an ash fall out particle collected on Iceland close to the eruption site is shown in Fig. 7a.
Fig. 7. mS-XRF spectra measured as trace element fingerprints of (A) an ash fall out particle collected in Iceland close to the eruption site, (B) an ash fall out particle collected in the Black Forest, (C) a non-volcanic particle collected in the Black Forest.
volcanic source. With this approach, a characteristic fingerprint for the volcanic particles from the E15 volcano eruption was obtained (Section 3.1). Fig. 6 shows SEM images and EDX spectra of two typical particles collected in Southwest Germany and identified as originating form the Eyjafjallajökull volcano.
3.2.3. Estimation of the amount of volcanic particles and of particles from other sources From SEM-EDX and mS-XRF analysis it was estimated that at least 40% of the analyzed particles within the size range of 2.5e10 mm in Southwest Germany were volcanic particles with the characteristic Eyjafjallajökull fingerprint. Cluster analysis helped to group the respective particle types. The average relative element concentrations for each of the eleven clusters are listed in Table 2. Cluster Cl3-1 represents the volcanic particles. All four particles from the reference material included in the cluster analysis clearly plot in this cluster. Cl3-2 represents siliceous particles, which are similar to the volcanic particle composition, but this cluster includes also some pre-event particles. The other clusters represent chlorides (NaCl and KCl, Cl8þ9), anthropogenic particles (e.g. color pigments, rust, metal abrasives Cl10, gypsum Cl11) and several silicate minerals and/or mineral agglomerates (Cl1, 2, 4, 5, 6). About 30e40% of the analyzed particles could be clearly identified as volcanic particles from the Eyjafjallajökull volcano. Highest concentrations were found at the Schauinsland with about 60% volcanic particles. The remote sites in the Black Forest showed lower contribution of volcanic particles but in general comparable results. The particles analyzed from Freiburg showed a much higher load of regional dust particles (w23%). This leads to a relative decrease of the amount of particles from volcanic origin. However, the volcanic cluster (cluster Cl3-1 in Table 2) only includes the particles with the typical siliceous glass composition. Some of the other particles might also be transported from the E15 eruption and, thus, the share of particles from the Eyjafjallajökull would be higher. In this regard especially more information about the material of the volcanic cone itself and the sediments around the volcano (below the glacier), which are also carried into the atmosphere by the eruption, would be needed.
Table 2 Average relative concentrations of all eleven clusters obtained after three-step Cluster analyses. Before clustering the relative element concentrations were normalized (ztransformed) using median and inter-quartile distance. The cluster-centers (means of all samples belonging to the respective cluster) are rescaled to the relative concentrations. Srel: relative standard deviation. The characteristic elements of each cluster are marked in bold. Cluster-center Cl1 Particle number srel O Na Mg Al Si P S Cl K Ca Ti Cr Mn Fe Cu Zn Sn
Cl2
Cl3-1 Volc!
Cl3-2
Cl4
Cl5
Cl6
35
4
87
27
8
18
5
17%
50%
10%
19%
35%
24%
45%
40.2 0.88 3.78 6.23 19.55 0.32 0.75 0.38 0.82 2.02 1.73 0.23 0.42 21.38 0.56 0.48 0.27
32.6 0.54 0.24 10.93 40.72 0.13 0.36 0.14 11.15 0.66 0.25 0.15 0.24 0.89 0.48 0.32 0.18
40.2 1.65 2.02 8.50 30.42 0.29 0.60 0.29 1.46 1.95 0.94 0.2 0.3 10.0 0.5 0.4 0.2
39.8 1.90 1.92 9.21 31.70 0.31 0.58 0.27 2.31 1.54 0.69 0.2 0.2 8.4 0.4 0.3 0.2
36.5 0.78 1.98 3.41 16.85 0.28 0.89 0.22 0.54 5.73 10.80 0.18 0.41 20.31 0.51 0.34 0.31
35.4 1.22 2.82 7.96 28.33 0.35 2.08 0.88 0.87 10.71 0.84 0.19 0.49 6.16 0.62 0.54 0.48
31.8 1.32 1.63 3.37 13.01 0.79 0.57 0.21 0.65 1.13 29.51 0.16 0.69 13.90 0.53 0.50 0.20
Cl7 1
Cl8 3
Cl9 1
58% 21.6 0.68 7.57 0.35 0.45 27.91 8.82 0.36 9.24 11.18 0.53 0.14 5.46 0.71 1.59 1.76 1.63
6.8 38.46 0.85 0.17 0.36 0.22 6.08 41.73 1.35 0.05 0.36 0.36 0.71 0.39 0.96 0.46 0.71
5.2 0.51 0.63 0.21 0.40 0.09 0.14 26.11 61.12 0.21 0.92 0.88 0.42 0.12 1.87 0.67 0.52
Cl10
Cl11
2
17
71%
24%
16.5 0.26 0.27 0.42 0.60 0.61 8.28 0.81 0.36 0.45 0.37 0.67 0.48 0.22 59.81 1.77 8.10
29.5 3.74 5.94 1.33 2.80 1.73 13.67 1.13 1.42 29.14 0.86 0.89 1.03 1.05 2.33 1.86 1.59
120
N. Schleicher et al. / Atmospheric Environment 48 (2012) 113e121
Germany. Therefore, further single particle analyses including coarser particles are additionally planned in order to verify whether even these coarser volcanic particles reached Southwest Germany. For a better estimation of the negative effects on airplane jet engines, it is recommended that experiments on turbines should be carried out with similar particle sizes as a function of distance from the eruption site, shape and hardness like the volcanic particles identified by the chemical fingerprints within this study. Based on the relatively low amount of volcanic particles determined on the ground in the Black Forest area and the short time of exposure, no adverse health effects on humans can be expected in Southwest Germany.
Fig. 8. E15-rationed values of the centers of the 6 clusters (trcl1etrcl6) obtained from the trace element data measured by mS-XRF.
Fig. 8 shows ratios of the E15 normalized trace element concentration values. The cluster trcl1 with high loads of Ca and Sr is indicative for calcite. The cluster trcl2 with high loads of K and Rb is indicative for orthoclase, whereas trcl3 with high loads of Cr, Ni and Pb can be interpreted as anthropogenic. In trcl4 nearly all ratios are plotting close to 1, thus the volcanic particles can be assigned to this cluster. Cluster trcl5 shows high ratios for Cu, Zn and Pb. This cluster can be interpreted either as anthropogenic metals or as particles from sulphide mineralization in the Black Forest and the high Ti content in trcl6 may be an indication for the presence of pigment particles possibly from wall or car paintings. 4. Conclusions and outlook It could be shown that volcanic particles from the E15 volcanic eruption reached the ground in Southwest Germany. Single particle analysis by means of SEM-EDX and mS-XRF provided a suitable fingerprint for the composition of the volcanic particles and enables the identification of probable mineral phases of the different particles and their type of sources (volcanic event, anthropogenic e.g. traffic, local geology and possible mineralization). The study showed that the combination of single particle major and trace element analysis together with data evaluation by modern multivariate statistical methods, which were applied in this context for the first time, proved to be a good tool for the identification of the volcanic particles. The deposition of volcanic particles from the E15 in Southwest Germany was proven by the recovery of the fingerprints from the reference samples collected directly on Iceland. Thus, the assumption from the PM10 distribution map (Fig. 1) that around Freiburg and the Black Forest a hot spot of E15 particle deposition occurred was verified by geochemical analyses in this study. In the remote Black Forest sites, at least 40% of the analyzed particles between 2.5 and 10 mm were identified as volcanic glass originating clearly from the Eyjafjallajökull. Due to the higher anthropogenic dust load in the city of Freiburg, the relative amount of volcanic particles is reduced to approximately 23%. Further analysis are planned in order to increase the amount of analyzed particles and, thus, the statistical reliability, and to obtain a more detailed picture of the spatial distribution of the fall out in Southwest Germany. In this regard, the focus on particles with grain sizes between 2.5 and 10 mm is in good accordance with a study by Flentje et al. (2010) in which the authors found particles up to 7.5 mm size in the volcanic plume over Germany. However, another study by Ansmann et al. (2010) reported even coarser particles larger than 20 mm in the ash layers over Leipzig and Munich,
Acknowledgments SEM-EDX analyses were carried out at the Laboratory for Electron Microscopy (LEM), Karlsruhe Institute of Technology (KIT), Karlsruhe, Germany. The authors express their special gratitude to Volker Zibat for his support with SEM-EDX analyses. Synchrotron radiation based XRF measurements were carried out at ANKA (Angström-Quelle Karlsruhe), KIT, and DORIS at HASYLAB/DESY, Hamburg. In this regard, the authors want to thank David Batchelor for assistance with setting up the FLUO-beamline at ANKA and Karen Appel and Manuela Borchert for assistance with beamline-L at HASYLAB/DESY. Furthermore, the authors want to thank Andreas Holbach and Sarah Perktold, KIT, for support with measurements at ANKA and LEM, respectively. Reference material collected on Iceland was provided by Natalia Garcia Villora, ITS, KIT, Karlsruhe, Germany, and by staff of SRI International, NSF Sondrestrom Research Facility, Greenland. References Ajtai, N., Stefanie, H., Stoian, L., Oprea, M., 2010. The volcanic ash and its impact on European air transport industry. A case study on the detection and impact of the Eyjafjallajökull volcanic ash plume over North-Western Europe between 14th and 21st April 2010. Advances in Environmental Sciences 2010 (2), 57e68. Abonyi, J., Feil, B., 2007. Cluster Analysis for Data Mining and System Identification. Birkhäuser, Basel, Boston, Berlin. Ansmann, A., Tesche, M., Groß, S., Freudenthaler, V., Seifert, P., Hiebsch, A., Schmidt, J., Wandinger, U., Mattis, I., Müller, D., Wiegner, M., 2010. The 16 April 2010 major volcanic ash plume over central Europe: EARLINET lidar and AERONET photometer observations at Leipzig and Munich, Germany. Geophysical Research Letters 37, L13810. Brooker, P., 2010. Fear in a handful of dust: aviation and the Icelandic volcano. Significance 7, 112e115. Colette, A., Favez, O., Meleux, F., Chiappini, L., Haeffelin, M., Morille, Y., Malherbe, L., Papin, A., Bessagnet, B., Menut, L., Leoz, E., Rouïl, L., 2011. Assessing in near real time the impact of the April 2010 Eyjafjallajökull ash plume on air quality. Atmospheric Environment 45, 1217e1221. Dietze, V., Fricker, M., Goltzsche, M., Schultz, E., 2006. Luftqualitaetsmessungen in deutschen Kurorten Teil 1: Methodik und Absicherung. Gefahrstoffe e Reinhaltung der Luft 66, 45e53. Durant, A.J., Bonadonna, C., Horwell, C.J., 2010. Atmospheric and environmental impacts of volcanic particulates. Elements 6, 235e240. Ebert, M., Weinbruch, S., Hoffmann, P., Ortner, H.M., 2004. The chemical composition and complex refractive index of rural and urban influenced aerosols determined by individual particle analysis. Atmospheric Environment 38, 6531e6545. Ersoy, O., Gourgaud, A., Aydar, E., Chinga, G., Thouret, J.C., 2007. Quantitative scanning-electron microscope analysis of volcanic ash surfaces: application to the 1982e1983 Galunggung eruption (Indonesia). Geological Society of America Bulletin 119, 743e752. Flentje, H., Claude, H., Elste, T., Gilge, S., Köhler, U., Plass-Dülmer, C., Steinbrecht, W., Thomas, W., Werner, A., Fricke, W., 2010. The Eyjafjallajökull eruption in April 2010 e detection of volcanic plume using in-situ measurements, ozone sondes and lidar-ceilometer profiles. Atmospheric Chemistry and Physics 10, 10085e10092. Gasteiger, J., Groß, S., Freudenthaler, V., Wiegner, M., 2011. Volcanic ash from Iceland over Munich: mass concentration retrieved from ground-based remote sensing measurements. Atmospheric Chemistry and Physics 11, 2209e2223.
N. Schleicher et al. / Atmospheric Environment 48 (2012) 113e121 Gislason, S.R., Alfredsson, H.A., Eiriksdottir, E.S., Hassenkam, T., Stipp, S.L.S., 2011. Volcanic ash from the 2010 Eyjafjallajökull eruption. Applied Geochemistry. doi:10.1016/j.apgeochem.2011.03.100. Godelitsas, A., Toli, K., Simon, R., 2009. Elemental Distribution in Saharan Dust from Red Rain Precipitated over Athens, Greece, ANKA Annual Report 2009, pp. 178e179. Guffanti, M., Casadevall, T.J., Budding, K., 2010. Encounters of Aircraft with Volcanic Ash Clouds; a Compilation of Known Incidents, 1953e2009: U.S. Geological Survey Data Series 545, Ver. 1.0. 12 , plus 4 appendixes including the compilation database, available only at. http://pubs.usgs.gov/ds/545 pp. Ilyinskaya, E., Tsanev, V., Martin, R., Oppenheimer, C., Le Blond, J., Sawyer, G., Gudmundsson, M., 2011. Near-source observations of aerosol size distributions in the eruptive plumes from Eyjafjallajökull volcano, MarcheApril 2010 Atmospheric Environment 45, 3210e3216. Jaenicke, R., 1978. Über die Dynamik atmosphärischer Aitkenteilchen. Berichte der Bunsen-Gesellschaft für Physikalische Chemie 82, S1198e1202. Kramar, U., 1995. Application of limited fuzzy-clusters for anomaly recognition in complex geological environment. Journal of Geochemical Exploration 55, 81e92. Lautze, N.C., Taddeucci, J., Andronico, D., Cannata, C., Tornetta, L., Scarlato, P., Houghton, B., Castro, M.D.L., 2011. SEM-based methods for the analysis of basaltic ash from weak explosive activity at Etna in 2006 and the 2007 eruptive crisis at Stromboli. Physics and Chemistry of the Earth. doi:10.1016/ j.pce.2011.02.001. Mills, O.P., Rose, W.I., 2010. Shape and surface area measurements using scanning electron microscope stereo-pair images of volcanic ash particles. Geosphere 6. doi:10.1130/GES00558.1. Petersen, G.N., 2010. A short meteorological overview of the Eyjafjallajökull eruption 14 Aprile23 May 2010. Weather 65, 203e207. Post, J.E., Buseck, P.R., 1984. Characterization of individual particles in the phoenix urban aerosol using electron-beam instruments. Environmental Science and Technology 18, 35e42. Schäfer, K., Thomas, W., Peters, A., Ries, L., Obleitner, F., Schnelle-Kreis, J., Birmili, W., Diemer, J., Fricke, W., Junkermann, W., Pitz, M., Emeis, S., Forkel, R., Suppan, P., Flentje, H., Wichmann, H.E., Gilge, S., Meinhardt, F., Zimmermann, R., Weinhold, K., Soentgen, J., Münkel, C., Freuer, C., Cyrys, J., 2011. Influences of the 2010 Eyjafjallajökull volcanic plume on air quality in the northern Alpine region. Atmospheric Chemistry and Physics Discussions 11, 9083e9132. Schleicher, N., Norra, S., Chai, F., Chen, Y., Wang, S., Stüben, D., 2010a. Anthropogenic versus geogenic contribution to total suspended atmospheric particulate matter and its variations during a two-year sampling period in Beijing, China. Journal of Environmental Monitoring 12, 434e441. Schleicher, N., Kramar, U., Norra, S., Dietze, V., Kaminski, U., Cen, K., Yu, Y., 2010b. mscale variations of elemental composition in individual atmospheric particles by means of synchrotron radiation based m-XRF analysis. X-RAY OPTICS AND
121
MICROANALYSIS: proceedings of the 20th international congress. AIP Conference Proceedings 1221, 172e180. Schumann, U., Weinzierl, B., Reitebuch, O., Schlager, H., Minikin, A., Forster, C., Baumann, R., Sailer, T., Graf, K., Mannstein, H., Voigt, C., Rahm, S., Simmet, R., Scheibe, M., Lichtenstern, M., Stock, P., Rueba, H., Schäuble, D., Tafferner, A., Rautenhaus, M., Gerz, T., Ziereis, H., Krautstrunk, M., Mallaun, C., Gayet, J.F., Lieke, K., Kandler, K., Ebert, M., Weinbruch, S., Stohl, A., Gasteiger, J., Olafsson, H., Sturm, K., 2011. Airborne observations of the Eyjafjalla volcano ash cloud over Europe during air space closure in April and May 2010. Atmospheric Chemistry and Physics 11, 2245e2279. Seinfeld, J., Pandis, S., 2006. Atmospheric Chemistry and Physics: From Air Pollution to Climate Change. John Wiley & Sons, Inc, Hoboken, New Jersey. Sigmundsson, F., Hreinsdóttir, S., Hooper, A., Árnadóttir, T., Pedersen, R., Roberts, M.J., Óskarsson, N., Auriac, A., Decriem, J., Einarsson, P., Geirsson, H., Hensch, M., Ófeigsson, B.G., Sturkell, E., Sveinbjörnsson, H., Feigl, K.L., 2010. Intrusion triggering of the 2010 Eyjafjallajökull explosive eruption. Nature 468, 426e430. Sobanska, S., Maenhaut, W., Coeur, C., Adams, F., 2003. SEM-EDX Characterisation of tropospheric aerosols in the Negev Desert (Israel). Journal of Atmospheric Chemistry 44, 299e322. Steinhausen, D., Langer, K., 1977. Clusteranalyse e Einführung in Methoden und Verfahren der automatischen Klassifikation. Walter de Gruyter, Berlin, New York. Sturkell, E., Einarsson, P., Sigmundsson, F., Hooper, A., Ófeigsson, B.G., Geirsson, H., Ólafsson, H., 2010. Katla and Eyjafjallajökull volcanoes. Developments in Quarterny Sciences 13, 5e21. VDI, 2011. VDI Guideline 2119: Partikelprobenahme mit einem Grobstaubsammler (Sigma-2) zur Charakterisierung von Einzelpartikeln und Berechnung der Konzentration. Vester, B.P., Ebert, M., Barnert, E.B., Schneider, J., Kandler, K., Schütz, L., Weinbruch, S., 2007. Composition and mixing state of the urban background aerosol in the Rhein-Main area (Germany). Atmospheric Environment 41, 6102e6115. Weinbruch, S., Wentzel, M., Kluckner, M., Hoffmann, P., Ortner, H.M., 1997. Characterization of individual atmospheric particles by element mapping in electron probe microanalysis. Mikrochimica Acta 125, 137e141. Yue, W., Li, Y., Li, X., Yu, X., Deng, B., Liu, J., Wan, T., Zhang, G., Huang, Y., He, W., Hua, W., 2004. Source identification of PM10, collected at a heavy traffic roadside, by analyzing individual particles using synchrotron radiation. Journal of Synchrotron Radiation 11, 428e431. Yue, W., Li, X., Liu, J., Li, Y., Yu, X., Deng, B., Wan, T., Zhang, G., Huang, Y., He, W., Hua, W., Shao, L., Li, W., Yang, S., 2006. Characterization of PM2.5 in the ambient air of Shanghai city by analyzing individual particles. Science of the Total Environment 368, 916e925.