Physical speciation of arsenic, mercury, lead, cadmium and nickel in inhalable atmospheric particles

Depósito de investigación de la Universidad de Sevilla https://idus.us.es/ “This is an Accepted Manuscript of an article published by Elsevier in: ANALYTICA CHIMICA ACTA on 2004, available at: https://doi.org/10.1016/j.aca.2004.02.004” Physical speciation of arsenic, mercury, lead, cadmium and nickel in inhalable atmospheric particles F. Fernández Álvarez, M. Ternero Rodríguez, A. J. Fernández Espinosa*, A. Gutiérrez Dabán Department of Analytical Chemistry, Faculty of Chemistry, University of Seville, C./ Profesor García González, s/n. E-41012, Seville, Spain Abstract. Physical speciation of metals (size distribution), in airborne particles has been studied in 24 areas of Seville (Spain). Urban particles were collected with high-volume sampling systems equipped with cascade impactor and PM10 inlets. Particle samples were analysed by acid digestion assisted by microwave using a HNO3:HClO4 (3:1) mixture. Arsenic was determined by hydride generation atomic absorption spectrometry (HGAAS), mercury by atomic absorption spectrometry and cold vapour technique (CV-AAS) and inductively coupled plasma atomic emission spectrometry (ICP-AES) was employed to determine lead, cadmium and nickel. Analyses of PM10 samples revealed that the highest mean concentration was obtained for Pb, 14.04 ng m-3. Values of 3.42 ng m-3 and 2.88 ng m-3 were obtained for As and Ni, respectively. The lowest concentrations were obtained for Cd, 0.49 ng m-3, and mainly for Hg, 0.07 ng m-3. Their physical speciation in air concentrations (ng m-3) showed that the toxic elements As and Pb were accumulated in the fine particles less than 0.6 µm (PM0.6), as well as the metals Cd and Ni, with high percentages of 74.8, 69.0, 61.3 and 57.5, respectively. Hg constituted a lower percentage of 40.4%. The results in mass concentrations (mg Kg-1) demonstrated that Pb and Cd contents were clearly higher in fine particles between 1.3 and 0.6 µm than in PM0.6. The Corresponding autor: Tel.: + 34-954554368; fax: + 34-954557168. E-mail address: [email protected] (A.J. Fernández Espinosa). size interval where Hg contents were higher than in PM0.6 (4.9 to 0.6 µm) corresponded to a fine and also a coarse mode, meanwhile Ni presented similar percentages between all size ranges. However, As is mainly accumulated in PM0.6 in both air and mass concentrations, revealing that it constitute the element more hazardous in relation to its size distribution. Keywords: Size distribution; Arsenic; Mercury; Airborne particles; PM10; Speciation; Microwave. 1. Introduction Atmospheric pollution by heavy metals in particles is one of the most worrisome problems nowadays in the scientific community. Particulate matter has a direct incidence on the health through the respiration because of their small size and the heavy metals composition, which have influence in biochemical mechanisms, especially because they are bioaccumulable. Fine particles are more easily transported to large distances and they can stay in the air in a prolonged manner depending on the meteorological situations and size distribution [1]. Size distribution of particles and metals give information for studies of transport process, effect of meteorology and sources study. For instance, it well known that elements associated with natural sources, such as soil and ocean, are usually found in coarse particles, while elements emitted from anthropogenic sources are associated with fine particles [2]. The relation between meteorology and particle size distribution has also been the subject of numerous recent and earlier works, as in Manchester [3], Helsinki [4] and in a Mediterranean zone of France [5]. The degree of bioaccumulation and bioavailability depends on the physical and chemical composition of particles. Conventional metal determinations are still carried out today by methods that lead to total concentrations. These total concentrations are nowadays determined only in a few cases in TSP (total suspended particles). These determinations should be done with particles speciated by size [6, 7], such as PM10 or PM2.5 and, less frequently, by impactor systems [8, 9]. Therefore, knowledge of particle size is vital in understanding the effects on human health. The degree of respiratory penetration and retention is a direct function of the aerodynamic particle (aed) diameter [10]. In general, only particles less than 10 or 15 µm in size reach the lower respiratory tract from the nose to the alveoli [11]. Particles with a diameter higher than 5 µm are filtered in the nose for the most part, while those less than 1-2 µm predominantly gets deposited in the alveolar regions of the lung, where the absorption efficiency for trace elements is 60-80% [12]. They could affect lung physiology, especially if the particles contain biologically available toxic elements, such as heavy metals. Knowledge of the size distribution of these metals using cascade impactors (physical speciation) and inlets (PM10) is the aim of the current research. This way, an objective of this work was to assess an analysis methodology for determine heavy metal concentrations in airborne particles based on physical speciation techniques. These techniques are also useful in characterising and identifying sources of heavy metals [13]. The other objective of this work is to propose a cut-point for the future PM inlets manufactured. Actually, the American Environmental Protection Agency [6] and the European regulatory laws will create soon mandatory measurements on PM2.5, breathable particles [14], which shows a clear inclination to study smaller particles. Actually, PM10 inlet is the only reference method that is exists. PM10 inlets used by us and the cascade impactor systems used in our physical speciation techniques give wide information about different size fractions of particles and their chemical composition. The evaluation of the size distribution with the impactor system device allows establish the relationship between the different size fractions and the different cut-points of inlets (PM10, PM2.5 or PM1) along a single sampling, without needing different inlets for each size fraction. The comparison between PM10 results and physical speciation of heavy metals constitute an objective of the current investigation in order to propose a different cut-point for the future. Besides, making demands on the new European Framework Directive of Evaluation and Management of Ambient Air Quality [15]. Because of these demands our research group set up a study on speciation of arsenic, mercury and other heavy metals to conform to the law. Various daughter directives were published, such as that concerning to PM10 (particles less than 10 micrometers) and lead [16]. Actually there is a recent proposal for the Fourth Daughter Directive relating to arsenic, mercury, cadmium, nickel and polycyclic aromatic hydrocarbons in ambient air [3]. For example, no later than 31 December 2003 the European Commission will submit a report, based on knowledge acquired in the application of the Directive 99/30/EC (article 10), to the European Parliament and the European Council. The Junta de Andalucia, our regional government, through its environmental office (Consejeria de Medio Ambiente) finances research for the Andalusian Universities, focusing the results on the annual reports they present to the European Commission. Our particular group is one of those collaborating with the Junta for this mission, offering analytical methods to the European Community to measure pollutants and in return being able to investigate new analytical strategies to feed into the scientific community. In relation to these two engagements, this work presents the results of composition of airborne particles by heavy metals, focusing the interest of results on arsenic and mercury. Both metals were no measured before using our physical speciation techniques. Even the metal lead, measured previously by us when leaded gasoline was still used in cars [2], present a special interest because actually it cannot come from a traffic source due to the use of unleaded gasoline. 2. Experimental 2.1. Sampling network design A sampling network of sampling stations was designed at the city of Seville (37°23’ N, 5°58’ W, Fig. 1 and Table 1), the largest and most air polluted city by traffic of Andalusia (Spain) [17]. The network was spanned including 24 sampling sites, twelve more than in our previous network [2], covering all city districts and some peripheral zones and towns. Two samplers were rotated and located near the typical emission sources of the city, such as highly trafficked streets, some big local industries and zones nearby to unpaved soils, as well as in locations with cleaner air. Consequently, the different sampling points were classified into three categories: sites under influence of industrial emissions, station under the influence of vehicular emission and peripheral station under the influence of resuspended soils. This way, information about a higher number of sampling sites than the number of sampling stations of our Local Air Monitoring Network of Seville was obtained in the city. 2.2. Samples collection system Atmospheric particulate matter was collected with two standard high-volume samplers MCV, model CAV-A/HF [18], one of them was equipped with a five-stage standard cascade impactor plus a backup filter MCV, model IC/CAV, and other was coupled with a PM10 inlet (MCV, model PM10/CAV). PM10 inlet was used to collect all particles less than 10 micrometers (aed) and the MCV mark was inter-compared with the reference PM10 inlet of the EN 12341 European Norm. Cascade impactor system separates the particulate matter following aerodinamic equivalent cut-off diameters (aed) at 50% efficiency: Stage 1, > 10 m, stage 2, 10-4.9 m, stage 3, 4.0-2.7 m, stage 4, 2.7-1.3 m, stage 5, 1.3-0.6 m and stage 6, < 0.6 m (0.6-0.01) (Backup). Both sampling systems were set at a flow rate of 68 m-3h-1, according manufacturing specifications. The set of filters used as collection media (Whatman GF/A) were constituted of five cut filters (14.2 x 14.2 cm) for stages of cascade impactor and a complete non-cut filter (24.5 x 20.3 cm) for the backup and PM10 inlet was used. PM10 inlet gives an only value of particle and metal concentration in each sampling. Impactor separator gives different particle and metal concentrations in each sampling. These different concentrations could be expressed simply as percentage regarding the total concentration, that is, the sum [8], or expressed as various types of normalised concentration [18]. This way, the analytical data of size distribution of particles and physical speciation of metals was obtained with the impactor system. A total of 93 samples were collected, being 48 samples collected with cascade impactor and 45 samples collected with PM10 inlet. Sampling period was from January 2000 to June 2001, during 18 month, so, samples were usually taken on different days each week, obtaining samples of every day of the week, therefore, in different meteorological and environmental conditions. Sampling periods were from 0:00 a.m. to 23:59 p.m. 2.3. Analytical procedure Before the analysis of samples, filters used for particle collection were dried, before and after weighing, in a high-volume desiccator under 5% R.H. for a 48-h period. Operations of conservation, drying and weighing took place in a dark room at 505 % R.H, a temperature of 201 ºC was controlled by a heating-cooling system [2]. Care was taken in handling the sampled filters in order to avoid contamination problems using a vertical laminar airflow cabinet with a HEPA filter Indelab, model IDL-48V, thereby ensuring air cleanliness standards of class 100 according to the US-Federal Standard 209E [19]. Filters were placed in closed teflon vessels with 15 ml of HNO3-HClO4 v/v (3:1) and digested in a microwave oven Millestone, model Ethos 900 with a specific digestion programme. After digestion and cooling, sample solutions were transferred into sterile tubes. Residual particles became colourless and filters were completely whites, with regard to the previous system used: heating with water bath [2]. Since individuals blanks were not available for each filter used for sampling, a set of unexposed filters for each type of filter was analysed as blanks using the same procedure used for samples. The mean unexposed filters value were determined and subtracted from each sample to obtain the best estimate of each element in the particulate matter. Table 2 shows the limits of detection (LOD) of the method (obtained from blanks values and their standard deviations) and the instrumental limits of detection (obtained from analytical calibrations). The accuracy of the method was previously checked when this was validated by analysis of a NIST reference material SRM 1648 [21,22,25]. The standard addition method was used to validate the mercury determination, since this element was not present in the SRM 1648 certificate. The results of recovery correspond with 18 different test applied to the SRM (Table 3) showing that recoveries of all elements were into the intervals recommended by the A.O.A.C. A gold solution was used as internal standard to spike samples, blanks and reference material and to check the correct quantification of the elements. Previously, it was studied that gold was not presented in urban air and its presence within the teflon vessel had no influence on analytical determinations of the other metals. All ultrapure reagents and standards were supplied by Merck and distilled water was of Milli-Q grade obtained from a Waters-Millipore apparatus, model Plus. Acid solutions were analysed for lead, cadmium and nickel by inductively coupled plasma atomic emission spectrometry (ICP-AES), using a Fisons-ARL 3410 sequential multi-element instrument. The standard operation condition of this instrument are summarised as follows: the carrier gas, coolant gas, and plasma gas is argon at 80 psi of pressure, the carrier gas and plasma gas flow rate are 0.8 l min-1, the coolant gas flow rate is 7.5 l min-1 and the integration time is 1 s. One mini-torch consumes argon gas at a radio frequency power of 650 W. Consequently, it is capable of consuming a few millilitres of sample at a flow rate of 2.3 ml min –1. In order to avoid errors in results due to the effect produced by the different viscosity and density of the acid matrix of samples on the ICP nebulisation, calibration curves have been made with the same acid matrix. Arsenic was analysed by hydride generation atomic absorption spectrometry (HGAAS) and mercury by cold vapour technique (CV-AAS) using a Perkin-Elmer 3100 atomic absorption spectrometer with hydride generator Perkin-Elmer model MSH-10. The standard operation condition of this instrument are summarised as follows: the carrier gas is nitrogen at 36 psi of pressure, stock solution of reductant containing 3% m/v NaBH4 and 1% m/v NaOH assuming the reduction of As(V) species to As(III) and the generation of hydrides corresponding to As(III) formed; reaction medium containing 1.5% HCl m/v [20-24]. The atomiser used was a quartz T-shaped tube, heated with an air-acetylene flame for arsenic and flameless for mercury determination. The detection wavelength was 193.7 nm for arsenic and 253.6 nm for mercury. 3. Results 3.1. Particle levels in air Mean values of PM10 samples in the 24 sampling sites of Seville were shown in Fig. 2. Average of PM10 levels over the sampling period (18 month) was 41.8 µg m-3. This [23] J. Moreda-Piñeiro, C. Moscoso-Pérez, P. López-Mahia, S Muniategui-Lorenzo, E. Fernández-Fernández, D. Prada Rodríguez. Talanta 53 (2001) 871. [24] J.L. Gómez Ariza, D. Sánchez Rodas, I Giraldes, E. Morales. Talanta 51 (2000) 257. [25] F. Fernández Álvarez, A. Gutiérrez Dabán, M. Ternero Rodríguez, A. J. Fernández Espinosa, D. Trigo García. VIII International Symposium on Analytical Methodology in the Environmental Field. A Coruña (2003) 508. [26] X. Jinhui, F.S.C. Lee. Analytica Chimica Acta 416 (2000) 111. [27] P.L. Melgarejo, M. Ternero, I. Gracia. International Journal of Environmental Analysis 24 (1986) 283. [28] Commission of the European Communities (COM), 2003. Proposal for a directive of the european parliament and of the council relating to arsenic, cadmium, mercury, nickel and polycyclic aromatic hydrocarbons in ambient air. Brussels, July 16, COM 423 final, 2003/0164 Cod. [29] WHO, 2000. World Health Organization. Guidelines for Air Quality. Cluster of Sustainable Development and Healthy Environment (SDE), Department of Protection of the Human Environment (PHE), Occupational and Environmental Health Programme (OEH). Geneva. Table 1 Description of the 24 sampling stations corresponding to the sampling network of Fig.1 Number Site Number Site 1 Art High School 13 School of Informatics Engineering 2 ‘Macarena’ Public School 14 Air Force Clinic 3 ‘Esperanza’ Public School 15 Polytechnic High School 4 Office of ‘Hermandades y Cofradias’ 16 ‘I. S. Mejias’ Public School 5 The Andalusian Ombudsman 17 ‘S. Angel de la Guarda’ Public School 6 School of Applied Arts 18 The SE-30 Road Maintenance Centre 7 Chemical Research Institute (IIQ - CSIC) 19 Transfer Centre of Urban Residues (LIPASAM) 8 ‘Llanes’ Secondary School 20 ‘Tartessos’ Secondary School 9 ‘Baltasar de Alcazar’ Public School 21 ‘Camas’ Secondary School 10 ‘Borbolla’ Public School 22 ‘Carlos V King’ Public School 11 ‘Esperanza’ Neighbours Association 23 ‘Cortijo de Cuarto’ School of Agricultural Engineering 12 Faculty of Chemistry 24 ‘A. M. Mateo’ Public School Table 2 Limits of detection of method and instrumental limits of detection for an air volume of 1632 m3 Element LOD of method (ng m -3 ) Instrumental LOD (ng m -3 ) As 0.025 0.002 Hg 0.027 0.002 Pb 0.940 0.283 Cd 0.199 0.112 Ni 0.305 0.240 Table 3. Mean values of recovery obtained from the accuracy study with the SRM 1648. Elements Recovery (%) (n=18) R.S.D. (%) (n=18) Confidence Interval (P=0.05) A.O.A.C. Criterion Arsenic 99.9 2.6 3.3 90 - 107 Mercury 99.1 3.5 2.1 80 - 110 Lead 95.3 2.6 1.3 95 - 105 Cadmium 110.0 3.7 1.8 80 - 110 Nickel 86.0 4.2 2.1 80 - 110 Fig. 1. Sampling network designed in Seville: location of the 24 sampling stations. Fig. 2. Values of PM10 obtained at the different sampling stations of Seville: horizontal line represents the average value. Fig. 3. Particle size distribution corresponding to the sampling period expressed as percentages of the total particles collected.