absorption spectrum of ambient aerosol and its correlation ... · 532 nm, 355 nm and 266 nm...

Aerosol and Air Quality Research, 13: 49–59, 2013 Copyright © Taiwan Association for Aerosol Research ISSN: 1680-8584 print / 2071-1409 online doi: 10.4209/aaqr.2012.04.0078 Absorption Spectrum of Ambient Aerosol and Its Correlation with Size Distribution in Specific Atmospheric Conditions after a Red Mud Accident Ágnes Filep1, Tibor Ajtai1*, Noémi Utry1, Máté D. Pintér1, Tünde Nyilas2, Szabolcs Takács3, Zsuzsanna Máté3, András Gelencsér4, András Hoffer4, Martin Schnaiter5, Zoltán Bozóki1, Gábor Szabó1 1 Department of Optics and Quantum Electronics, University of Szeged, Dóm tér 9, 6720, Szeged, Hungary 2 Department of Mineralogy Geochemistry and Petrology, University of Szeged, Egyetem Street 2, 6722, Szeged, Hungary 3 Department of Public Health, University of Szeged, Dóm tér 10, 6720, Szeged, Hungary 4 Air Chemistry Group of the Hungarian Academy of Sciences, University of Veszprém, P.O. Box 158, 8201 Veszprém, Hungary 5 Institute of Meteorology and Climate Research, Karlsruhe Institute of Technology, P.O. Box 3640, 76021 Karlsruhe, Germany ABSTRACT

Approximately one month after a red sludge industrial accident, a high concentration of atmospheric aerosol was reported in the affected area. This increased aerosol load originated mainly from the fugitive dust emitted from the red sludge sediment, representing a serious threat to public health. In a study presented here, a recently developed four-wavelength photoacoustic spectrometer (4λ-PAS), which simultaneously measures optical aerosol absorption in the near infrared, visible and ultra-violet spectral ranges with high temporal resolution, was operated in the centre of the sludge hit area. Based on the results of the photoacoustic absorption measurements of temporal changes in light absorption and their correlations with other aerosol variables, including size distribution, were carried out in special atmospheric conditions (including with the presence of “atypical” red mud particles). Correlations between segregated absorption Ångström exponents (AAEs) and specific size ranges were tabulated, and the resulting data were used to draw conclusions about the reasons for the temporal changes to the measured quantities. The absorption behavior of red mud particles was determined in UV, VIS, and near IR spectral regions under laboratory conditions. The absorption spectrum of red mud aerosol was found to be atypical (i.e., it did not follow a power law), and this was used as a selective parameter for its identification in ambient bulk. Each measurement day, at around noon, dust from the red sludge sediment was identified; while during the rest of the day atmospheric aerosol composition was dominated by carbonaceous aerosols, originating most probably from home heating and/or from traffic. In-situ number concentration and size distribution measurements confirm this interpretation of the measurements obtained with 4λ-PAS. Keywords: Ångström exponent; Photoacoustic spectroscopy; SMPS. INTRODUCTION

Hungary’s largest ecological disaster took place on October 4, 2010 at 1:30 p.m. when the western dam of a toxic red sludge reservoir, located about 100 km south-west of Budapest ruptured. Due to the ruptured dam 600–700 thousand m3 of dilute red sludge flooded 800 hectares of surrounding areas killing 10 persons. The highly alkaline red sludge, by-product of alumina manufacturing, is a corrosive * Corresponding author. Tel.: +3662-544-285; Fax: +3662-5444-658 E-mail address: [email protected]

substance. The most extreme devastation was caused in the villages of Devecser and Kolontár, which are located near the reservoir (http://redsludge.bm.hu). After the disaster the area was closed, partial traffic ban was implemented but the inhabitants were not evacuated.

Even after the immediate critical phase of red sludge had passed, the potential public health and environmental hazard of red sludge and its constituent had raised increasing particulate scientific concern. Three of by-product disposal types are most common in alumina product, namely (i) lagoon type, (ii) disposal in the sea and (iii) dry mud stacking. Disposal of red mud in the sea has been appreciably reduced over the last decades because of the inestimable environmental hazard both of the remaining two disposal solutions involve the risk of air pollution through the dusting

Filep et al., Aerosol and Air Quality Research, 13: 49–59, 2013 50

of disposal site (Paramguru et al., 2005). Moreover, the health professionals are in the common platform that the most relevant adverse health effect is associated to inhalation of wind-dispersed red mud particles. Under specific meteorological conditions these particles could appear far-off from its source through long range transport (Prospero et al., 1999). Therefore, the motivation to characterize the physical-chemical features of red mud particles as well as its health effects are not restricted to this disaster or localities of alumina industry based on the Bayer process but more broadened. Certainly, despite of its importance there are no reliable methods for on-line determination of red mud content of ambience to alarm when the concentration of red mud particles in the atmosphere increases. Most commonly used method for composition identification is based on off-line chemical analysis (Sandradewi et al., 2008a; Genberg et al., 2011). Physical quantities of red mud particles, which can be measured on-line such as absorption features or size distribution, are none, or poorly characterized in the literature. Consequently, spectral behavior of red mud particles, as atypical atmospheric aerosol needs to be characterized in details.

Approximately one month after the disaster high aerosol concentration was measured in the surrounding area. These measurement results caused serious concerns because the aerosols re-suspended from red sludge sediment were thought to pose high risk to human health. Several measurement campaigns were performed in the area in order to assess the level of potential health and environmental hazard (Gelencsér et al., 2011; Ruyters et al., 2011; Czövek et al., 2012). These works are mainly focused on the quantitative characterization of chemical feature and health hazard based on the off-line sampling.

This study reports the first laboratory investigation of wavelength dependent optical absorption coefficient (OAC) of re-dispersed red mud particles in UV, VIS, near IR spectral region as well as for ambient aerosol measured at disaster (red mud passed) area using a novel multi-wavelength photoacoustic spectrometer (Ajtai et al., 2011). This instrument could measure OAC without any sampling artifact and with high reliability (below 6% depending on applied wavelengths). In one of our recent study we demonstrated that the instrument selectivity is the highest among the recently available on-line instruments for measurement of absorption behavior of particles (Ajtai et al., 2011). Main scientific goal of this study was to investigate photoacoustic response of atmospheric aerosol in solar spectral region at specific ambient conditions including atypical (red mud) particles in relation to concurrently measured other aerosol variable such as size distribution. Absorption spectrum of re-dispersed red mud aerosol was determined under laboratory conditions. We have found that spectral response of red mud particles is characteristic to its constituent, shows extremely high absorption in the UV and differs significantly from typical constituents of urban aerosol such as soot from traffic or heating (Sandradewi et al., 2008b). Second motivation was to investigate the possibility of on-line identification of red mud aerosol in the atmosphere from the measured ambient absorption spectra

using the laboratory data. Since neither absorption nor number size distribution are unique indicators of constituents, we have used segregated absorption Ångström exponent (AAE) deduced from absorption measured by photoacoustic spectroscopy at different wavelength pairs and the correlations between measured quantities to fulfill these measurement purposes. EXPERIMENTAL

The measurements reported here took place between 03 and 08 November 2010 at the schoolyard of the primary school of Devecser (GPS position: 47.106397, 17.43727) having special permission from the Hungarian National Directorate General for Disaster Management. The instruments were operated inside a temperature and humidity controlled mobile measurement station of Hilase Ltd installed in a three-ton van. All instruments used in the campaign were connected to PM2.5 cut-off inlets placed approximately 2 meters above the roof of the van. The Four-Wavelength Photoacoustic Spectrometer (4λ-PAS)

Photoacoustic spectroscopy i.e., energy transfer from a modulated light source to an acoustic wave by optical absorption is a unique method to measure the light absorption coefficient of aerosols directly without the necessity of complicated filter sampling and/or site-specific calibration procedures (Roessler and Faxvog, 1980). Photoacoustic (PA) signal is linearly proportional to light absorption and it is completely insensitive to light scattering. A PA system, after being calibrated with a light absorbing gas with known optical absorption cross section, can be used to determine OAC in real time with high reliability even under field conditions (Hamasha and Arnott, 2009; Ajtai et al., 2011).

The four-wavelength PA spectrometer (4λ-PAS) applied in this study is the upgraded version of the laboratory instrument described in details elsewhere (Ajtai et al., 2010a). It has a dimension of 65 cm × 65 cm × 130 cm, and approximately 100 kg in weight. It measures OAC simultaneously at four different wavelengths: 1064 nm, 532 nm, 355 nm and 266 nm spanning the NIR, VIS and UV wavelength regions. It has four detection PA cells and each is illuminated with one of the light beams. Its operation is controlled by a built-in programmable electronic unit which makes its operation fully automated and unattended. The gas handling system is built in a way to eliminate possible measurement interferences from gaseous compounds in the atmosphere (most importantly NO2 and ozone). From the acoustic signals measured in the detection cell (representing an acoustic resonator), the electronic unit calculates OACs with the help of calibration factors determined during an independent gas phase calibration of the instrument (Ajtai et al., 2010a). Repeatability of the measurements are 0.2 Mm−1 at 1064 nm and 7 Mm−1 at 266 nm (Mm−1 = 10-6 m−1), corresponding to a minimum detectable black carbon (BC) mass concentration of about 50–400 ng/m3 calculated with optical absorption cross section values given in A.D.A. Hansen 2005 depending on


the measurement wavelength. The instrument has already operated successfully in several laboratory and field measurement campaigns reported elsewhere (Schnaiter et al., 2003, Wagner et al., 2009, Ajtai et al., 2010a, b, 2011; Wagner et al., 2012). Supplementary Instrumentation of the Field Campaign

In addition to the 4λ-PAS, the optical absorption of aerosol was also measured with a seven wavelength (370 nm, 470 nm, 520 nm, 590 nm, 660 nm, 880 nm and 950 nm) Aethalometer (Magee-scientific AE31) possessing a restricted wavelength coverage compared to that of the 4λ-PAS. The Aethalometer was operated in an automated mode, including filter tape advance whenever light attenuation at 370 nm decreased to 75% of the value measured on an aerosol free filter due to deposited aerosols. Measurement artifacts of the Aethalometer induced by shadowing and multiple scattering effects were minimized by using correction factors suggested by Weingartner et al. (2003). OAC determined by the Aethalometer was calibrated against OAC measured by the 4λ-PAS at 532 nm by interpolating the readings of the Aethalometer to this wavelength. The combined operation of 4λ-PAS and Aethalometer was established in a previous measurement campaign (Ajtai et al., 2011).

Number concentration and size distribution of the atmospheric aerosol were measured with a Wide Range Aerosol Spectrometer (Grimm Aerosol Technik GmbH, composed of a SMPS+C system consisting a Condensation Nucleus Counter, Model #5.400 and a Classifier “Vienna”-Type DMA, Model #5.500 and a Portable Aerosol Spectrometer, Model #1.108) in a wide particle diameter interval between 5 nm–2.5 μm. A scan was completed in every 4 minutes.

Besides the online measurements, atmospheric aerosols were collected on quartz fiber filters for the laboratory analysis described in Laboratory measurements section. Aerosol collection on a filter lasted for 24 hours resulting in a total sampled ambient air volume of 24 m3 per filter.

Meteorological parameters such as wind speed, wind direction, temperature, humidity, total radiation, and pressure were measured by the Middle Transdanubian Environment Inspectorate next to our van. Laboratory Measurements

The field measurements were supplemented with various

laboratory measurements. One of them targeted the determination of the wavelength-dependent OAC of the red sludge aerosol. Experimental set-up of these laboratorial measurements is shown in Fig. 1. The re-dispersion of red sludge aerosol was achieved by blowing synthetic air with a flow rate of about 10 L/min above a Petri dish in which the dried and pulverized red mud sediment – collected from the red sludge hit area – was stored. Outlet of the aerosol generating system was divided into 4 lines by using polished stainless steel tubes with an outer diameter of 6 mm and special 45° splitters. Length and bends of the tubes were made as short and small as possible in order to keep the aerosol losses at minimum. The first line was connected to SMPS, the second to 4λ-PAS, the third to OPC and the fourth line was left open to maintain ambient pressure in the PA cells.

In another set of laboratory measurements the mass concentrations of major aerosol components were determined from 24 h aerosol samples collected on quartz fiber filters. Mass of total carbon (TC) collected on a filter was provided by Astro Model 2100 laboratory TC analyzer (Zellweger Analytics) in solid module operating mode. In this instrument the determination of TC mass is based on catalytic combustion of the sample in oxygen at 680°C and the detection of the evolving carbon dioxide by a non-dispersive infrared detector. The measurement of levoglucosan (Bari et al., 2010) was performed after derivatization with bis(trimethylsilyl)triflouroacetamide (BSTFA): trimethylchlorosilane (TMCS) 99:1 in pyridine (3 h, 70°C) by gas chromatography-mass spectroscopy (Agilent 6890N, Agilent 5973N).

Elemental composition of the red mud was analyzed by using X-ray fluorescence analysis with a Horiba Jobin-Yvon XGT 5000 apparatus implemented on pressed powder pellets from the sediment with 5 parallel measurements (30 kV, Rh ray-source, 100 μm ray-pile, 900 sec time period). RESULTS Results of the Laboratory Measurements

The chemical analysis of the samples collected on the quartz fiber filters demonstrated quite high total carbon and levoglucosan ratio to the total particle mass concentration (that was measured by the Middle Transdanubian Environment Inspectorate during our campaign) (respectively 1.5 and 3 times higher as at the same time in Budapest).

Fig. 1. Experimental set-up of the laboratory measurements.


This enhancement could be explained also by differences in emission rate or in heating technologies predominately used in Budapest and Devecser (Bari et al., 2010) as well as differences in meteorological conditions.

It was demonstrated by Gelencsér et al. (2011) that the red mud dust which is released to the atmosphere has identical chemical composition as particle samples that were collected directly from the sludge sediment. Relying on this fact we have made XRF analysis on our own sediment samples to get information on the mineralogical composition of the aerosol particles. As it can be seen in Fig. 2 major elements in our sample are aluminum, silicon, calcium, titanium and iron. Chemical composition of bauxite is highly variable, but according to Bárdossy et al. (1982) major mineral component of the Hungarian karst bauxite (and hereby also the by-product of the so called Bayer process of bauxite namely of red mud) is hematite (Fe2O3), and could include some boehmite (AlO(OH)), gibbsite (Al(OH)3), perovskite (CaTiO3), rutile (TiO2), and calcite (CaCO3). Paramguru et al. (2005) observed 15.2% Al2O3 38.45% Fe2O3, 10.15% SiO2 and 4.6% TiO2 and 8.12% Na2O in Hungarian red mud samples. From those components only iron-oxide has optical absorption in the visible range (Querry et al., 1985). The high amount of Si and Ca in our XRF spectra might result from soil-derived minerals (quartz, albite and muscovite) which are normally not present in the atmosphere because of their large particle diameter. Taking our results and those from Gelencsér et al. (2011) we propose that the major light absorbing (in the visible range) component of airborne red mud dust is most likely hematite.

Number (and volume) size distributions and absorption spectra of the re-dispersed red mud dust and hematite (Fe2O3) particles (Reanal Private Ltd., Hungary, degree of purity > 95%) – which were thought to be the major constituent having significant absorption in the UV-VIS spectral region – were measured in the laboratory. Number and volume size distribution of the re-dispersed samples can

be seen in Fig. 3. Geometric median diameters of the red mud dust and hematite aerosols were found to be 86 nm and 83 nm, respectively. In both samples the shoulders on the right side of number size distribution modes could be attributed to the agglomeration of the individual particles (Zelenyuk and Imre, 2007). Sizes of artificially generated and ambient aerosols may differ significantly, partly due to the dissimilarity of the particle re-dispersion way in nature and in laboratory. Therefore, we used the number (and volume) size distribution spectra measured in laboratory only for model calculation. The measured size spectrum of re-dispersed red mud particle agrees well with earlier study (Gelencsér et al., 2011). Larger particles (Dp > 500 nm) dominated in volume in both cases.

OACs of both samples at the operation wavelength of the applied photoacoustic instrument as well as the modeled absorption spectra of hematite aerosol in the wavelength coverage of the photoacoustic instrument at three different characteristic sizes are shown in Fig. 4. For theoretical simulation of the hematite aerosol absorption Mie theory was used by assuming spherical particles. Although assuming spherical particles is a limitation for accurate modeling the optical absorption of real particles at a given wavelength, the spectral behavior of hematite in different spectral regions could be characterized with high reliability with this approach. Measured and calculated data agree well with each other and also with others results (Mogili et al., 2007; Wagner et al., 2009), which further confirms the reliability of this approximation. The optical constants of hematite aerosol were published by Querry (1985). The suggested method to model the aerosol absorption in case of anisotropic (birefringent) material is to calculate the individual spectra for both ordinary (o-ray) and extraordinary (e-ray) refractive indices by Mie theory and take average of the spectra weighted by the o-ray to e-ray ratio. Since hematite is a birefringent material having two equivalent o- and one e-ray axes, we used 2/3 o-ray + 1/3 e-ray average of the optical constant for calculations. The

Fig. 2. XRF spectrum of the red mud dust sample identifies iron as one of the major elements.


Fig. 3. Number (full squares refer to SMPS and empty squares to OPC data) and volume (solid line refer to SMPS and dashed line to OPC data) size distribution of re-dispersed hematite and red mud dust samples measured by SMPS and OPC (black lines and symbols refer to red mud dust and grey lines and symbols to hematite sample).

Fig. 4. Measured and calculated OACs of re-suspended hematite and red mud dust samples. Measured values are represented by symbols. Black squares: the red mud dust and grey triangles: the hematite powder. Lines represent modeling results for hematite particle size distributions with different median diameters. Grey line: GMD 110 nm, black line: GMD 300 nm and light grey line: GMD 900 nm (note the different scaling on the right axis for this size).

absorption coefficient measured by 4λ-PAS is the sum of the absorption contributions from all particles that reside in the photoacoustic cell at a given moment. The calculated absorption cross section refers to a single particle at a given size. We have integrated the calculated absorption cross sections over the measured number size distribution in order to get the modeled absorption coefficient then it can be compared with the coefficients measured by the PAS. We have also calculated hematite absorption coefficients for

two additional median particle sizes (300 nm and 900 nm), but with the same number concentration and distribution width as measured for our re-dispersed sample. Consequently, it has been demonstrated that possible differences between the size distributions of hematite and red mud aerosols measured in ambience and in the laboratory could not affect the spectral habitus of the absorption coefficient. The hematite aerosol possesses strong and featureless absorption spectrum in the UV and the lower visible as well as a low


or even negligible absorption in the upper visible and the near IR spectral region. Red mud dust shows similar spectral absorption attribute as the hematite aerosol except at mid-UV wavelength 266 nm where the red mud sample exhibit relatively high OAC. We therefore conclude from Fig. 4 that the spectral absorption behavior of the red mud dust aerosol is mainly governed by its hematite content except at 266 nm where the absorption of other components probably soil-derived minerals becomes dominant (Sokolik and Toon, 1999). Results of the Field Measurements

Average temperature during the field measurement was quite stable, with a minimum of 5°C and a maximum of 15°C. There was precipitation only at the night of 3 November, and the average relative humidity was 74%. Average wind speed was below 2 m/s until the night of 7 November, when it increased to 8 m/s. Prevailing wind direction was south-southwest (200°).

In Fig. 5 temporal variations of OAC measured by 4λ-PAS and Aethalometer are compared. Readings of the 7λ-Aethalometer were inter- and extrapolated to the measurement wavelengths of the 4λ-PAS as follows. AAEs were fitted to the OAC readings of the Aethalometer in 370–520 nm and 590–950 nm range, respectively, and the

former one was used to estimate bλ at 266, 355 and 532 nm while the latter one was used to estimate OAC at 1064 nm.

From the number size distribution data measured by SMPS and by OPC, and the calculated volume size distribution (by assuming spherical particles) hourly averages were calculated and the results were analyzed in details. Two distinct peaks were found both in the number and volume size distributions with altering relative intensities (Fig. 6). The two identified modes are GMD20 (geometric median diameter 19 ± 6 nm) and GMD100 (98 ± 20 nm). As it can be seen in Fig. 6 contribution of particles with sizes above 500 nm to the total aerosol number concentration was found to be negligible. Since light absorption by aerosols is driven by particle volume, contribution of these (GMD > 500) particles to the total aerosol volume needs to be quantified as well. We found that these particles contribute to the total aerosol volume even by 50% in some cases. No characteristic mode was found in this size range similarly to the results of Gelencsér et al. (2011). Main source of red mud dust in the atmosphere is supposed to be Aeolian process, because of the extremely high re-suspensation potential (200 times higher than at curbside observed by Gelencsér et al., (2011)) in this area. The rate of wind-dispersion increases sharply with particle diameter, therefore in later analysis we have sum the total number of particles

Fig. 5. Temporal variation of OAC measured by 4λ-PAS and 7λ-Aethalometer at the operational wavelengths of 4λ-PAS. Grey lines represent the readings of 4λ-PAS and black lines represent OAC calculated from the Aethalometer readings (see text for details).


Fig. 6. Average number (a) and volume (b) size distribution in the two most representative periods; data from 08:30–09:30 are labeled with black solid line and from 20:30–21:30 with grey line and symbols (note the different scaling on the right axis).

in the characteristic size ranges 15–25 nm and 85–115 nm, as well as above 500 nm. DISCUSSION

Correlations between identified size modes and measured OAC and AAE values are collected in Table 1. Although neither absorption nor size distribution of atmospheric aerosol is unique indicator for emission sources, many papers have demonstrated that composition of atmospheric aerosol could be identified by measuring these parameters with high reliability (Cappa et al., 2009; Moosmüller et al., 2011; Salma et al., 2011). In urban air the size mode around 20 nm refers to traffic emission and the mode around 100 nm refers to residential heating (Bond et al., 2002; Wehner and Wiedensohler, 2003). Similarly, absorption in red to near IR spectral region is mainly governed by inorganic refractory carbon (i.e., traffic emission) while towards shorter wavelengths organic carbon aerosol fraction (e.g., from residential heating) becomes more dominant in absorption strength (Hand et al., 2005; Favez et al., 2009).

Correlations between OAC measured at 1064 nm and GMD > 500, and GMD100 are negligible but becomes significant in case of GMD20 (upper row in Table 1). High correlation between OAC at 1064 nm and the 20 nm mode increases the reliability of the conclusion that this mode refers to local traffic sources (Fig. 7(a)). Likewise, the overriding

correlation between OAC at 355 nm and strength of the 100 nm mode (R = 0.81) compared to other specified size region (third row in Table 1), supports the conjecture that particles in this mode are mainly governed by organic carbon aerosol emission from residential heating (Fig. 7(b)). Slope of the correlations summarized in Table 1 could indicate the relative sensitivity of the measured quantities to a given change in all cases investigated in this study. The slopes in Fig. 7(a) and 7(b) may indicate that the optical responses are more efficient at indicating the changes of relative aerosol fraction from traffic and heating than the size distribution.

OAC is not unique indicator of the composition mainly because of its sensitivity to the concentration (Roessler and Faxvog, 1980). Wavelength dependence of AAE is not affected by concentration changes and is more characteristic to the sample (Sandradewi et al., 2008b). AAE is determined as the slope of log-log representation of the measured OAC spectra. Furthermore, due to the wide wavelength coverage of 4λ-PAS and the fact that aerosols often show different wavelength dependencies in the different regimes of UV-VIS-NIR spectral range, wavelength dependence of AAE itself has to be taken into account when describing spectral behavior of measured OACs. Alternatively, this wavelength dependence has been already proved to increase the applicability of our PA system for aerosol source apportionment (Ajtai et al., 2011). Three AAEs


Table 1. Correlation between optical parameters and number concentration values.

GMD20 GMD100 GMD > 500

Correlation coefficient

Slope Correlation coefficient


Slope

OAC@1064 0.81 641.69 0.39 517.37 –0.40 289.41 OAC@532 0.70 239.41 0.48 192.17 –0.32 107.07 OAC@355 0.51 82.4 0.81 77.7 –0.25 40.74 OAC@266 0.49 83.47 0.66 76.51 –0.27 40.81

GMD20/Ntot GMD100/Ntot GMD > 500/Ntot

Correlation coefficient



Slope

AAE@1064–532 0.33 8.12 × 10 0.18 7.46 × 10–3 0.21 1.87 × 10–3 AAE@532–355 0.03 8.64 × 10 0.64 8.87 × 10–3 –0.13 3.15 × 10–3 AAE@355–266 0.17 9.08 × 10 0.37 9.15 × 10–3 0.65 4.31 × 10–3

Fig. 7. Correlation between the measured OACs and number concentrations - a) Number concentration of GMD20 mode as function of OAC at 1064 nm; b) Number concentration of GMD100 mode as function of OAC at 355 nm.

were calculated from OACs measured by the 4λ-PAS for neighboring wavelength pairs 1064–532, 532–355 and 355–266 nm. Similarly, two AAEs were calculated from OACs measured by the Aethalometer. One exponent refers to the upper (950–590 nm), and the other refers to the lower (520–370 nm) spectral region of the instrument. AAEs, calculated from the measurements of 4λ-PAS and Aethalometer for the NIR (Fig. 8(a)) and visible (Fig. 8(b)) spectral regimes, show very good agreement. Especially the exponent of the visible region shows distinct temporal changes (Fig. 8(b)). The UV exponent (Fig. 8(c)) also shows temporal fluctuation but with different phase compared to the visible. Values determined for the visible spectral region have maxima in evening hours and stays high during night, whereas AAE of the UV spectral range exhibit sharper peaks around noon, i.e., from 11:00 to 13:00 o’clock Central European time.

As mentioned above AAE could also be an indicator of aerosol composition (Park et al., 2006; Ajtai et al., 2010b). AAE around 1 indicates high amount of BC, while increase of AAE might be a hint of enhanced brown carbon (BrC) aerosol concentration (Kirchstetter et al., 2004; Sandradewi et al., 2008a). Here BrC is defined as light absorbing organic matter in atmospheric aerosol with steeply increasing

absorption towards the UV (Andreae and Gelencsér, 2006). The slight temporal changes of AAE532–1064 nm could be explained by the fact that BrC fraction has low or negligible absorption value in this spectral region (Sandradewi et al., 2008b). Change in composition is one possible explanation for more relevant temporal fluctuations of AAE calculated from measured absorption at shorter wavelengths (Favez et al., 2009). Similarity of volume size distributions measured at the two representative periods (Fig. 6) further confirms that variations in AAE were not caused by variation in particle size, but by changes in composition. The authors have never observed and also not found in the literature such phase shift in temporal changes of AAEs calculated in different spectral region. For understanding this phenomenon we have investigated the correlation between relative number concentration of particles in the characteristic modes and AAEs associated to different spectral regions (Figs. 9(a)–9(b) and Table 1).

Reliable and distinct correlation (R = 0.64) between AAE355–532 nm and relative number concentration associated to 100 nm mode was found compared with other size ranges (6th row in Table 1). This may also indicate that spectral changes in this wavelength region are governed mostly by the aerosols associated with residential heating


Fig. 8. Temporal variations of wavelength dependent AAEs, calculated from the records of the 4λ-PAS (black solid line) for the wavelength ranges 1064–532 nm (Fig. 8(a)), 532–355 nm (Fig. 8(b)), 355–266 nm (Fig. 8(c)) and the Aethalometer (grey line with symbols) for 950–590 nm (Fig. 8(a)) and 520–370 nm (Fig. 8(b)).

(Fig. 9(a)). Although particles above 500 nm did not show any characteristic modes neither in number nor in volume concentrations, relative fraction of these particles showed remarkable correlation with AAE266–355 nm (R > 0.65) (Fig. 9(b) and 7th raw in Table 1). Taking into account that red mud aerosol has significant absorption at 266 nm and it is dominant in higher size range in ambience (Gelencsér et al., 2011) it would mean that this fraction (Dp > 500 nm) may contain red mud aerosol. By mean of this assumption the phase shift in temporal variation between AAE266–355 nm and AAE355–532 nm values could also be explained. SUMMARY AND CONCLUSION

Novel multi-wavelength photoacoustic instrument and SMPS + OPC were used to study aerosol light absorption and aerosol size distribution after the disaster at the red sludge hit area. Spectral response of the red mud aerosol was also characterized under laboratory conditions. Measured absorption spectrum of red mud particles is governed by its hematite content in the visible region and shows relatively high absorption at 266 nm. Model calculation was used to demonstrate that the shape of absorption spectrum is not affected by size, so it is

characteristic to the sample. Under ambient conditions two characteristic modes in number size distribution were identified. Temporal variation of identified modes and absorption coefficients measured at different wavelengths were recorded. Correlation between modes and absorption coefficients were deduced and tabulated (Table 1). Wavelength dependent AAEs were also calculated from the measured absorption coefficients in different spectral regions. Phase shift between the temporal variation of AAE266–355 nm and AAE355–532 nm has been identified and explained by revealing the correlation with variation of relative amount of particles in different size regions.

Unfortunately, the red sludge accident gave an extreme opportunity to test the selectivity of our novel multi-wavelength photoacoustic instrument under real and special ambient conditions. Since spectral behavior of red mud particle differs significantly from that of typical urban aerosol originating from traffic or heating, it might be used as composition selective parameter in on-line identification. The correlation coefficients calculated in all possible combination are remarkable and reveal a novel possibility of more efficient on-line investigation of the ambient aerosol via the concurrently measured absorption and size spectrum, even though the relatively short measurement period. For quantitative determination of the efficiency of the presented on-line investigation for composition identification further measurements are needed with supplemented chemical analysis.

We experimentally demonstrated that the identification of composition or the changes in composition through measurement of absorption spectra is more efficient at shorter wavelengths especially in the UV wavelength region. These results further confirmed our last result that the reliability of the estimated absorption in the UV from the measured absorption at longer wavelengths is strongly depending on the composition of ambience (Ajtai et al., 2011). We also demonstrated that in some cases, especially when atmosphere includes “atypical” aerosol such as red mud particle, the correction of such estimation can not be scaled by AAE as was done in our previous work (Ajtai et al., 2011). Finally, we experimentally demonstrated that the reliability of the composition identification through the aerosol absorption measurement can be further increased by using number concentration and size distribution data together. ACKNOWLEDGEMENT

This work has been supported by the Hungarian National Development Agency (JEDLIK_AEROS_EU and TÁMOP-4.2.2/B-10/1-2010-0012) and OTKA foundation from the Hungarian Scientific Research Fund (project number: CNK 78549, K 101905 and K 81181).

The authors also gratefully acknowledge the supplementary data of the Middle Transdanubian Environment Inspectorate, the entry permit of the Hungarian National Directorate General for Disaster Management and the technical support of the Hilase Developing, Producing, Servicing and Trading Ltd.


Fig. 9. Correlations between AAE and total amount of the given modes; (a) Correlation between AAE@355–532 nm and amount of the particles having Dp ~100 nm; (b) Correlation between AAE@266–355 nm and amount of the particles having Dp > 500 nm.

REFERENCES Ajtai, T., Filep, Á., Schnaiter, M., Linke, C., Vragel, M.,

Bozóki, Z., Szabó, G. and Leisner, T. (2010a). A Novel Multi-wavelength Photoacoustic Spectrometer for the Measurement of the UV-vis-NIR Spectral Absorption Coefficient of Atmospheric Aerosols. J. Aerosol Sci. 41: 1020–1029.

Ajtai, T., Filep, Á., Kecskeméti, G., Hopp, B., Bozóki, Z. and Szabó, G. (2010b). Wavelength Dependent Mass-Specific Optical Absorption Coefficients of Laser Generated Coal Aerosols Determined from Multi-Wavelength Photoacoustic Measurements. Appl. Phys. A 103: 1165–1172.

Ajtai, T., Filep, Á., Utry, N., Schnaiter, M., Linke, C., Bozóki, Z., Szabó, G. and Leisner, T. (2011). Inter-Comparison of Optical Absorption Coefficients of Atmospheric Aerosols Determined by a Multi-Wavelength Photoacoustic Spectrometer and an Aethalometer under Sub-Urban Wintry Conditions. J. Aerosol Sci. 42: 859–866.

Andreae, M.O. and Gelencsér, A. (2006). Black Carbon or Brown Carbon? The Nature of Light-Absorbing Carbonaceous Aerosols. Atmos. Chem. Phys. 6: 3131–3148

Bárdossy, G. (1982). Karst Bauxites: Bauxite Deposits on Carbonate Rocks, Elsevier, Amsterdam, p 441.

Bari, Md. A., Baumbach, G., Kuch, B. and Scheffknecht (2010). Temporal Variation and Impact of Wood Smoke Pollution on a Residential Area in Southern Germany. Atmos. Environ. 44: 3823–3832.

Bond, T.C., Covert, D.S., Kramlich, J.C., Larson, T.V. and Charlson, R.J. (2002). Primary particle Emissions from Residential Coal Burning: Optical Properties and Size Distributions. J. Geophys. Res. 107: 8347, doi: 10.1029/

2001JD000571. Cappa, D.D., Bates, T.S., Quinn, P.K. and Lack, D.A. (2009).

Source Characterization from Ambient Measurements of Aerosol Optical Properties. Geophys. Res. Lett. 36: L14813, doi: 10.1029/2009GL038979.

Czövek, D., Novák, Z., Somlai, C., Asztalos, T., Tiszlavicz, L., Bozóki, Z., Ajtai, T., Utry, N., Filep, Á., Bari, F. and Peták, F. (2012). Respiratory Consequences of Red Sludge Dust Inhalation in Rats. Toxicol. Lett. 209: 113–120.

Favez, O., Cachier, H., Sciare, J., Sarda-Estéve, R. and Matinon, L. (2009). Evidence for a Significant Contribution of Wood Burning Aerosols to PM2.5 during the Winter Season in Paris, France. Atmos. Environ. 43: 3640–3644.

Gelencsér, A., Kováts. N., Turóczi, B., Rostási, Á., Hoffer, A., Imre, K., Nyírő-Kósa, I., Csákberényi-Malasics, D., Tóth, Á, Czitrovszky A., Nagy, A., Nagy, Sz., Ács, A., Kovács, A., Ferincz, Á., Hartyáni, Z. and Pósfai, M. (2011). The Red Mud Accident in Ajka (Hungary): Characterization and Potential Health Effects of Fugitive Dust. Environ. Sci. Technol. 45: 1608–1615.

Genberg, J., Hyder, M., Stenström, K., Bergström, R., Simpson, D., Jönsson, J.Å. and Swietlicki, E. (2011). Source Apportionment of Carbonaceous Aerosol in Southern Sweden. Atmos. Chem. Phys. 11: 11387–11400.

Hamasha, K.M. and Arnott, W.P. (2009). Photoacoustic Measurements of Black Carbon Light Absorption Coefficients in Irbid City, Jordan. Environ. Monit. Assess. 166: 485–494.

Hand, J.L., Malm, W.C., Laskin, A., Day, D., Lee, T., Wang, C., Carrico, C., Carrillo, J., Cowin, J.P., Collett Jr., J. and Iedema, M.J. (2005). Optical, Physical, and Chemical Properties of Tar Balls Observed during the Yosemite Aerosol Characterization Study. J. Geophys.


Res. 110: D21210, doi: 10.1029/2004JD005728. Hansen, A.D.A. (2005). The Aethalometer, Magee Scientific

Company Berkeley, California, USA. http://redsludge.bm.hu. Kirchstetter, T.W., Novakov, T. and Hobbs, P.V. (2004).

Evidence that the Spectral Dependence of Light Absorption by Aerosols is Affected by Organic Carbon. J. Geophys. Res. 109: D21208, doi: 10.1029/2004JD004999.

Mogili, P.K., Yang, K.H., Young, M.A., Kleiber, P.D. and Grassian, V.H. (2007). Environmental Aerosol Chamber Studies of Extinction Spectra of Mineral Dust Aerosol Components: Broadband IR-UV Extinction Spectra. J. Geophys. Res. 112: D21204, doi: 10.1029/2007JD008890.

Moosmüller, H., Chakrabarty, R.K., Ehlers, K.M. and Arnott, W.P. (2011). Absorption Ångström Coefficient, Brown Carbon, and Aerosols: Basic Concepts, Bulk Matter, and Spherical Particles. Atmos. Chem. Phys. 11: 1217–1225.

Paramguru, R.K., Rath, P.C. and Misra, V.N. (2005). Trends in Red Mud Utilization – A Review. Miner. Process. Extr. Metall. Rev. 26: 1–29.

Park, K., Chow, J.C., Watson, J.G., Trimble, D.L. and Doralswamy, P. (2006). Comparison of Continuous and Filter-Based Carbon Measurements at the Fresno Supersite. J. Air Waste Manage. Assoc. 56: 474–491.

Prospero, J.M. (1999). Long-Range Transport of Mineral Dust in the Global Atmosphere: Impact of African Dust on the Environment of the South-Eastern United States. Proc. Nat. Acad. Sci. U.S.A. 96: 3396–3403.

Querry, M.R. (1985). Optical Constants, U.S. Army CRDEC, Contractor Report, Aberdeen Proving Ground, Md.

Roeesler, D.M. and Faxvog, F.R. (1980). Photoacoustic Determination of Optical Absorption to Extinction Ratio in Aerosols. Appl. Opt. 19: 578–581.

Ruyters, S., Mertens, J., Vassilieva, E., Dehandschutter, B., Poffijn, A. and Smolders, E. (2011). The Red Mud Accident in Ajka (Hungary): Plant Toxicity and Trace Metal Bioavailability in Red Mud Contaminated Soil, Environ. Sci. Technol. 45: 1616–1622.

Salma, I., Borsós, T., Weidinger, T., Aalto, P., Hussein, T., Dal Maso, M. and Kulmala, M. (2011). Production, Growth and Properties of Ultrafine Atmospheric Aerosol Particles in an Urban Environment. Atmos. Chem. Phys. 11: 1339–1353.

Sandradewi, J., Prevot, A.S.H., Szidat, S., Perron, N., Alfarra,

M.R., Lanz, V.A., Weingartner, E. and Baltansperger, U. (2008a). Using Aerosol Light Absorption Measurements for the Quantitative Determination of Wood Burning and Traffic Emission Contributions to Particulate Matter. Environ. Sci. Technol. 42: 3316–3323.

Sandradewi, J., Prevot, A.S.H., Weingartner, E., Schmidhauser, R., Gysel, M. and Baltensperger, U. (2008b). A Study of Wood Burning and Traffic Aerosols in an Alpine Valley Using a Multi-Wavelength Aethalometer. Atmos. Environ. 42: 101–112.

Schnaiter, M., Horvath, H., Möhler, O., Naumann, K.H., Saathoff, H. and Schöck, O.W. (2003). UV-VIS-NIRspectral Optical Properties of Soot and Soot-Containing Aerosols. J. Aerosol Sci. 34: 1421–1444.

Sokolik, I.N. and Toon, O.B. (1999). Incorporation of Mineralogical Composition into Models of the Radiative Properties of Mineral Aerosol from UV to IR Wavelengths. J. Geophys. Res. 104: 9423–9444.

Wagner, R., Linke, C., Naumann, K.H., Schnaiter, M., Vragel, M., Gangl, M. and Horvath, H. (2009). A Review of Optical Measurements at the Aerosol and Cloud Chamber AIDA. J. Quant. Spectrosc. Radiat. Transfer 110: 930–949

Wagner, R., Ajtai, T., Kandler, K., Lieke, K., Linke, C., Müller, T., Schnaiter, M. and Vragel, M. (2012). Complex Refractive Indices of Saharan Dust Samples at Visible and near UV Wavelengths: A Laboratory Study. Atmos. Chem. Phys. 12: 2491–2512.

Wehner, B. and Wiedensohler, A. (2003). Long Term Measurements of Submicrometer Urban Aerosols: Statistical Analysis for Correlations with Meteorological Conditions and Trace Gases. Atmos. Chem. Phys. 3: 867–879.

Weingartner, E., Saathoff, H., Schnaiter, M., Streit, N., Bitnar, B. and Baltensperger, U. (2003). Absorption of Light by Soot Particles: Determination of the Absorption Coefficient by Means of Aethalometers. J. Aerosol Sci. 34: 1445–1463.

Zelenyuk, A. and Imre, D. (2007). On the Effect of Particle Alignment in the DMA. Aerosol Sci. Technol. 41: 112–124.

Received for review, April 4, 2012 Accepted, September 6, 2012

absorption spectrum of ambient aerosol and its correlation ... · 532 nm, 355 nm and 266 nm...

Documents