Journal of Environmental Monitoring

Transcription

1 PAPER Journal of Environmental Monitoring Comparison of pressurized fluid extraction, Soxhlet extraction and sonication for the determination of polycyclic aromatic hydrocarbons in urban air and diesel exhaust particulate matter M. Rynö, a L. Rantanen, b E. Papaioannou, c A. G. Konstandopoulos, c T. Koskentalo d and K. Savela* e Received 8th November 2005, Accepted 21st February 2006 First published as an Advance Article on the web 9th March 2006 DOI: /b515882f In order to characterize and compare the chemical composition of diesel particulate matter and ambient air samples collected on filters, different extraction procedures were tested and their extraction efficiencies and recoveries determined. This study is an evaluation of extraction methods using the standard 16 EPA PAHs with HPLC fluorescence analysis. Including LC analysis also GC and MS methods for the determination of PAHs can be used. Soxhlet extraction was compared with ultrasonic agitation and pressurized fluid extraction (PFE) using three solvents to extract PAHs from diesel exhaust and urban air particulates. The selected PAH compounds of soluble organic fractions were analyzed by HPLC with a multiple wavelength shift fluorescence detector. The EPA standard mixture of 16 PAH compounds was used as a standard to identify and quantify diesel exhaust-derived PAHs. The most effective extraction method of those tested was pressurized fluid extraction using dichloromethane as a solvent. Introduction a Finnish Institute of Occupational Health, Helsinki, Finland b Neste Oil, Porvoo, Finland c Aerosol & Particle Technology Laboratory, CERTH/CPERI, PO Box 361, Thermi Thessaloniki, Greece d Helsinki Metropolitan Area Council (YTV), Helsinki, Finland e National Veterinary and Food Research Institute (EELA), PL 45 (Ha meentie 57), Helsinki, Finland. kirsti.savela@eela.fi; Fax: þ ; Tel: þ Diesel engine exhaust is an urban air pollutant containing several carcinogenic compounds in either gas or particle phase. Polycyclic aromatic hydrocarbons (PAHs) containing 2 5 benzene rings have shown to be in the gas phase and more than 5 rings have tendency to adsorb to the particles. Including molecular weight of PAHs also the meteorological and physical transformation processes influence in distribution between volatile and particulate phase. PAHs adsorbed onto the exhaust particles induce genotoxic and chronic effects in humans. 1 3 Diesel engine exhaust is considered probably (Group 2A) and gasoline engine exhaust possibly (Group 2B) carcinogenic to humans. 4,5 Diesel and gasoline engines emit particles and PAH compounds due to incomplete combustion of the fuel. The particulate mass and PAH emissions from diesel engines are generally higher than those of gasoline engines with a catalytic converter. 6 8 In this study we investigated three solvent extraction procedures with three separate solvents in order to evaluate the most effective method to analyze particulate PAH compounds in environmental air samples and in diesel exhaust samples generated either directly from a car 9 or from a bench engine. To characterize the chemical composition of the particulate matter, urban air and diesel exhaust were collected on the glass fibre and Teflon-coated glass fibre (PTFE) filters. The efficiency of extracting PAHs from the particulates was compared between Soxhlet extraction, ultrasonic agitation and pressurized fluid extraction (PFE) It should be noted that instead of PFE, PLE (pressurized liquid extraction) or ASE (accelerated solvent extraction) are used as abbreviations of the same extraction method in the literature. PFE is the liquid extraction of analytes from a solid matrix using an elevated temperature and pressure to enhance mass transfer. The system applies pressure so that elevated temperatures can be used during extraction without boiling the solvent system. The PFE method has been shown to be comparable to Soxhlet extraction for the determination of PAHs in a variety of environmental matrices. 11,15 This method has also been validated according to an extraction technique by NIST (National Institute of Standards and Technology) for the certification of natural matrix reference materials. 16 The PFE apparatus consists of a stainless steel extraction cell which is kept in an electronically-controlled oven during the extraction. The extraction cell is first filled with organic solvent and then pressurized and heated up to the desired temperature. The sample is extracted for a certain time (static extraction step), after which the solvent is transferred to a collection vial. The sample and the connecting lines are then washed with fresh solvent and finally, using a suitable gas, the residual solvent is purged from the sample and the connecting lines to the collection vial. The performance of this apparatus for many different environmental samples is well described by Bjo rklund et al. 11 The enhanced performance of liquid solvents at an elevated temperature and pressure compared to sonication or Soxhlet extraction is due to the disruption of the surface equilibrium, solubility and mass transfer effects. The elevated temperature disrupts strong solute-matrix interactions, decreases the viscosity of the solvent allowing improved 488 J. Environ. Monit., 2006, 8, This journal is c The Royal Society of Chemistry 2006

2 penetration into the matrix, and increases the solubility and diffusion rates of analytes in the solvent. The high temperature also forces solvent into the matrix pores and hence into contact with the analytes. The effect of the different variables (i.e. temperature and pressure) on the performance of PFE has been studied by using reference materials by separate authors. 14,15 Experimental Chemicals and materials Methanol and acetonitrile were purchased from Lab Scan Ltd, Analytical Sciences, Dublin, Ireland. Dichloromethane, toluene and cyclohexane (all of analytical grade) were all from Merck Darmstadt, Germany. Standard mixture of 16 priority PAH cocktail (containing each PAH 1 mg ml 1 in acetonitrile) was purchased from Chiron AS, Norway. The corresponding individual PAHs were obtained from Ehrenstorfer (Augsburg, Germany). Three different samples were used to test the extraction procedures: (a) environmental samples (glass fibre, mm, 0.3 mm), (b) diesel exhaust samples from a passenger car (Pallflex Fiberfilm T60A20, PTFE-coated glass fibre, mm, 0.3 mm) indicated as diesel 1, and (c) diesel engine exhaust samples from a bench engine (A/E by Pall glass fibre filter + 47 mm, 0.3 mm) indicated as diesel 2. Environmental samples collected from Helsinki City on glass fibre filters and diesel exhaust samples collected on PTFE filters (Pallflex PTFE-coated) were selected for testing. Details of the particle sampling and modifications of the emission tests have been previously described. 9 The effect of different matrices, the extraction efficiency and the recovery of dieselderived PAH compounds in the soluble organic fraction of particulate matter was analyzed. In addition, three different filters were spiked with a PAH standard mixture and the extraction efficiency and recovery were tested. Twenty-four filters were obtained from the Aerosol & Particle Technology Laboratory of CPERI/CERTH in Greece, for the experiment. The extraction of 21 glass fibre filters was performed with dichloromethane for 60 min using an accelerated solvent extractor (ASE100). Six glass fibre filters were spiked with 250 ng of 16 PAH standard containing 10 ng ml 1 in acetonitrile in order to measure the percentage recovery of the extraction procedure. Loaded filters and filters spiked with the PAH standard mixture were extracted and analyzed using the same procedure. Sample extraction and analysis of PAH from particulate extracts Urban air samples collected on glass-fibre filters were used to optimise the conditions for three extraction methods and three solvents. Extraction techniques and solvents were tested in order to select the most suitable method for analyzing PAHs derived from urban air and diesel particles collected on glassfibre and Teflon-coated filters, respectively. The extraction efficiency of PFE with sonication and Soxhlet was compared in order to improve the recovery of PAH compounds from environmental samples. The comparison of manual and time consuming extraction methods (sonication and Soxhlet) with PFE was difficult due to the constant pressure and automatic solvent consumption. To check the contamination due to the extraction procedure (filter, apparatus, solvents, gases) a series of three blanks for each solvent and extraction method was used. No traces of contaminants were detected in blank samples neither using three different extraction methods nor solvents. Soxhlet extractions were conducted using 250 ml of dichloromethane (DCM), toluene methanol (1 : 1) and cyclohexane to determine the concentration of PAHs in urban air and diesel exhaust samples. The Soxhlet apparatus consisted of a 500 ml round-bottomed flask, a condenser and extractor tube seated in a temperature-controlled heating mantle. DCM, toluene, methanol and cyclohexane of analytical-reagent grade were used in all extractions. Each filter sample was placed in an extractor tube and the extraction was carried out for h. After extraction, the samples were concentrated under nitrogen (38 1C) with an evaporator (TurboVap LV Evaporator, Zymark) and the solvent was gradually changed. Acetonitrile was added dropwise while evaporating the extraction solvents in order to prevent the samples drying out. The final volume was adjusted to 1 ml with acetonitrile and the extracts were filtered through a 0.45 mm syringe filter (GHP Bulk Acrodisk 13, Gelman Sciences) prior to HPLC analysis. Sonication was carried out at room temperature. The glass fibre filters ( mm) were cut into small pieces, placed in Erlenmeyer flasks and extracted in 110 ml of solvent (DCM, toluene methanol (1 : 1), cyclohexane) for 30 min. After extraction the samples were evaporated under nitrogen as described earlier, the solvent was changed to acetonitrile and the extract filtered prior to HPLC analysis. PFE was performed by placing the folded filter in the extraction cell. With small diameter filters the cell was first filled with an inert material (ASE Prep DE, Diatomaceous Earth, Dionex) to keep the filter from attaching to the cell walls. The extraction temperature was 100 1C and the static extraction time 60 min. The volume of solvents (45 60 ml) was determined automatically for each sample by the PFE instrument. The same three solvents as used with Soxhlet and sonication were tested. Filter extracts were concentrated with an evaporator, adjusted to a final volume of 1 ml with acetonitrile, and filtered prior to HPLC analysis. PAH analysis was carried out with an automated high performance liquid chromatography/multiple wavelength shift fluorescence detector system (HPLC-FLD) based on peak recognition (HP1100 series, Agilent, Karlsruhe, Germany). A reversed-phase HPLC column (ChromSpher 3 PAH column, mm, 3 mm particle size, ChromPack, Krotek, Finland) was used with a precolumn (Chromguard PAH column, 10 3 mm, ChromPack). The flow rate was 0.3 ml min 1 and the injection volume was 15 ml. Elution was performed with a gradient starting from an eluent composition of 40% acetonitrile in water. The acetonitrile concentration was increased to 100% over 16 minutes, and maintained at 100% for 13 min. Between the analyses the column was washed with 100% acetonitrile using a flow rate of 0.6 ml min 1 for 3 min. 17 A fluorescence detector with two emission channels (FLD1A, FLD1B) and an external calibration curve This journal is c The Royal Society of Chemistry 2006 J. Environ. Monit., 2006, 8,

3 were used in the quantification. The data were collected and processed with HP ChemStation software. Two emission channels were used to separate any close-eluting peaks at the optimum excitation and emission wavelengths by selecting one of the two emission channels for the quantification. The US Environmental Protection Agency (EPA) standard mixture of 16 PAH compounds (Naphthalene; Napht, Acenaphthene; Ace, Fluorene; Fluo, Phenanthrene; Phen, Anthracene; Anthr, Fluoranthene; Flrt, Pyrene; Pyr, Benz[a]anthracene; B[a]A, Chrysene; Chry, Benzo[b]fluoranthene; B[b]F, Benzo[k]fluoranthene; B[k]F, Benzo[a]pyrene; B[a]P, Dibenzo[a,h]anthracene; DB[ah]A, Benzo[ghi]perylene; B[ghi]Pe, Indeno[1,2,3-cd] pyrene; I[123cd]P) (16 Priority PAH, Cocktail, Chiron AS, Norway) in acetonitrile was used to quantify and identify PAHs in particulate extracts. The corresponding individual PAHs were obtained from Ehrenstorfer (Augsburg, Germany). Acenaphthylene was not analysed due to its weak fluorescence. Table 1A shows a list of 16 PAH compounds with maximum excitation and emission wavelengths. Wavelength program for the PAH standard mixture is presented in Table 1B showing the selected excitation and emission wavelengths for the quantification. The wavelength shifts were programmed based on the individual maximum excitation and emission wavelengths of each PAH compound. During 24 minutes and between 230 nm and 298 nm all monitored PAH compounds eluted from the Table 1 Determination of 16 PAH compounds using high performance liquid chromatography/multiple wavelength shift fluorescence detector system. (A) List of 16 PAH compounds with maximum excitation and emission wavelength. For the quantification the optimum excitation and emission wavelengths were selected (Emission A and B) to separate the close-eluting peaks as shown in Table 1A. (B) Wavelength program for the PAH standard mixture A. Wavelength maxima of excitation and emission for each PAH compound Compound Excitation, l max /nm Emission, l max /nm Naphthalene Acenaphthene Fluorene Phenanthrene Anthracene Fluoranthene Pyrene Benz[a]anthracene Chrysene Benzo[b]fluoranthene Benzo[k]fluoranthene Benzo[a]pyrenev Dibenzo[a,h]anthracene Benzo[ghi]perylene Indeno[1,2,3-cd]pyrene HPLC column. Acenaphthene and fluorene, fluoranthene and pyrene were co-eluting and two-channel detection solved the problem. Benzo[b]fluoranthene, benzo[k]fluoranthene, benzo [a]pyrene, dibenzo[a,h]anthracene, benzo[ghi]perylene and indeno[1,2,3-cd]pyrene were analysed using two channels according to their maximum emission wavelengths and thus a better sensitivity of the quantification compared to onechannel detection was obtained. Linear calibration curves (R = ) were analysed from the standard mixture of 16 PAHs in triplicate using different concentrations (2, 4, 8, 16, 32, 64, 128 ng ml 1 ). Results and discussion The final results of our study demonstrated that naphthalene, acenaphthene, fluorene, phenanthrene, anthracene and fluoranthene in environmental samples were partly lost during sonication, whereas PFE also yielded values for these PAHs. The range of concentrations of weak (benz[a]anthracene, chrysene, benzo[ghi]perylene) and strong (benzo[b]fluoranthene, benzo[k]fluoranthene, benzo[a]pyrene, dibenzo[a,h] anthracene, indeno[1,2,3-cd]pyrene) carcinogenic PAHs was large, showing the lowest and highest concentration for DB[ah]A and B[ghi]Pe, respectively. 1 A summary of the results obtained from the analysis of PAHs in urban air particulate matter using three solvents and three extraction techniques revealed that Soxhlet and PFE had equal efficiencies in extracting a total of 15 PAHs (Fig. 1). Sonication was clearly the weakest extraction method when comparing the PAH levels with PFE and Soxhlet extraction, whereas no differences between solvents were observed when using PFE. PFE was selected as the best extraction method due to the good reproducibility of extraction and the simple and rapid handling of the samples. DCM was selected as the solvent for the extraction, since it gave the cleanest background in the HPLC/FLD analysis and, being more volatile than toluene methanol, it quickly evaporated during the concentration step, making the analysis less time-consuming. Toluene methanol has the disadvantage of leaving a hard soluble residue and cyclohexane evaporated slowly. Fig. 2 presents HPLC chromatograms for the PAH standard solution (Fig. 2A) together with the particulate extract of urban air B. Wavelength program used for the PAH detection Time/min Excitation Emission A Emission B Fig. 1 Total concentration of 15 PAHs analyzed from a soluble organic fraction extracted from urban air samples with dichloromethane, toluene methanol (1 : 1) and cyclohexane by using pressurized fluid extraction (PFE), sonication and Soxhlet apparatus. 490 J. Environ. Monit., 2006, 8, This journal is c The Royal Society of Chemistry 2006

4 Fig. 3 PAH compounds analyzed in a soluble organic fraction extracted by three different methods using DCM as a solvent. Urban air samples were collected on glass fibre filters and extracted by sonication (12 filters), Soxhlet (12 filters) and PFE (17 filters). collected on filters seven years prior to analysis and the filters were stored at room temperature. However, measurements confirmed that PAHs are very strongly adsorbed to the particles. DCM was the most effective solvent, yielding higher levels of PAHs than toluene methanol and cyclohexane. Soxhlet extraction Fig. 2 Panel (A) shows an example of a chromatogram obtained from the PAH standard solution analyzed by HPLC with the multiple wavelength program as indicated in Table 1(B). In panel (B) a chromatogram of an urban air sample from Helsinki City collected on a glass fibre filter, and (C) shows an example of HPLC chromatogram from a particulate extract of diesel exhaust collected from an emission test. (Fig. 2B) and diesel exhaust (Fig. 2C). Our study on extraction procedures showed that further work is still needed to reduce the matrix effect in analysing urban air and diesel exhaust samples in order to obtain a satisfactory separation of PAH mixture. Including HPLC/FLD also mass spectrometric methods (LC/GC) could be useful identifying both parent PAHs, and alkyl and nitro-substituted PAHs in real samples. Fig. 3 shows the levels of individual PAH compounds extracted with DCM using the three extraction methods. Sonication Sonication at room temperature for 30 min gave 10-fold lower levels for volatile PAHs compared to particle-bound PAHs. It was not expected that volatile PAHs would still be detected in the particulate extracts, since the urban air samples were Extraction using Soxhlet apparatus was carried out for h using the same three solvents as in the sonication experiment. Toluene methanol yielded slightly higher levels of volatile PAHs than extraction with DCM and cyclohexane. The Soxhlet apparatus was more effective than sonication in extracting PAH compounds from particles. The concentrations of volatile PAHs varied from to ng m 3 and particulate PAHs varied from to ng m 3 using Soxhlet extraction. Concentrations of volatile PAHs for sonication were between and and for particulate PAHs between and Toluene methanol (1 : 1) showed slightly higher recovery for the volatile and particulate Pyr, B[a]A, Chry, B[b]F, B[k]F, B[a]P and DB[ah]A when compared to B[ghi]Pe and I[123cd]P but the difference was not statistically significant. Cyclohexane gave the lowest recovery of the three solvents used in Soxhlet extraction. Pressurized fluid extraction (PFE) Extraction of PAHs from urban air particles by PFE using DCM did not seem to improve the recovery when increasing the number of cycles from 2 10 min and 4 10 min to 5 10 min (Fig. 4). According to the extraction time there was little difference between using one or cycles mentioned above in extracting the particles. The three solvents used in this study did not differentiate much between extraction cycles, however, showing slightly higher levels of PAHs when DCM was used. Differences in solvent strength were observed when one extraction of 20 min, 40 min and 60 min was performed, showing a better efficiency for DCM during 20 min and 40 min and a lower efficiency during 60 min when compared to toluene methanol and cyclohexane. The concentrations of volatile and particulate PAH compounds analysed in urban air samples This journal is c The Royal Society of Chemistry 2006 J. Environ. Monit., 2006, 8,

5 Table 2 Recoveries (%) and relative standard deviation (RSD) for volatile and particulate PAHs extracted by PFE from 8 filters spiked with standard mixture (250 ng) of 16 PAHs Volatile % RSD Particulate % RSD Fig. 4 The effect of extraction time and the number of extraction cycles on recoveries. (n = 52) by PFE for 60 min using three solvents are shown in Fig. 5. Concentrations of volatile PAHs were 2-fold lower than those of particle-bound PAHs, being approximately on the same level when compared to those analyzed by Soxhlet extraction. Particulate PAHs analyzed by PFE and Soxhlet extraction were in the same range. Recoveries for volatile and particulate PAHs were determined from 8 filters spiked with 250 ng of PAH mixture (Table 2). Recoveries of volatile and particulate PAH compounds varied from % (relative standard deviation, RSD, %) and 76 94% (RSD %), respectively. Naphthalene, acenaphthene and fluorene showed very low recoveries, probably due to their low molecular weight allowing them easily to escape during the evaporation of DCM after extraction from the sample. The overall recovery for the total of 15 PAHs analyzed was about 70%. Napht Pyr Ace B[a]A Fluo Chr Phen B[b]F Anthr B[k]F Flrt B[a]P DB[ah]A B[ghi]Pe I[123cd]P Total 15 PAHs Fig. 6 A typical profile of 15 PAH compounds measured in particulate extracts of urban air sample collected on glass-fibre filters (n = 19), diesel exhaust collected PTFE-filters (diesel 1; passenger car, n = 11), and glass-fibre filters (diesel 2; bench engine, n = 10). PAH concentrations are presented as mg per g of the soluble organic fraction (SOF). Fig. 6 shows a typical profile for the 15 PAH compounds analyzed in urban air and diesel exhaust collected on PTFEcoated glass-fibre filters from passenger car (diesel 1) and glassfibre filters from bench engine (diesel 2). Filters were extracted with PFE using DCM as a solvent and further analyzed by HPLC-FLD. In comparison with glass-fibre and PTFE-coated glass-fibre filters, PAH concentrations in diesel particle extracts from glass-fibre filters were lower than those on PTFE-coated filters. This is probably due to different collection procedures (passenger car vs. bench engine). Numerous sources can cause variation in the results. When collecting particulate matter from different sources, factors such as the sampling process, filter material and temperature affect the particulate composition. Large variations in PAH levels can also be expected, depending on the method of particulate extraction, the fuel used and the emission test carried out. Conclusions Fig. 5 The average of total volatile (A) and particle-derived (B) PAH concentrations in extract of urban air samples (n = 52) analyzed by PFE for 60 min using three solvents. After comparison of the extraction techniques, PFE showed a marked advantage over Soxhlet extraction and sonication. Soxhlet extraction is time-consuming and laborious, which is a clear disadvantage when compared with PFE or sonication. Sonication is a vigorous extraction method and therefore it 492 J. Environ. Monit., 2006, 8, This journal is c The Royal Society of Chemistry 2006

6 unbinds some of the filter material and detaches collected particles during the extraction unlike with Soxhlet extraction or PFE, where only PAH compounds are extracted from the particles and the extract remains clear and can be directly analyzed by HPLC. HPLC/FLD is a useful and sensitive method for the analysis of PAHs derived from environmental air or diesel exhaust samples. PFE is a closed system, enabling the detection of volatile PAHs in the filter extract, but the concentrations of volatile PAH varied due to their low molecular weight. Particulate PAHs showed no large differences between the solvents and techniques, revealing the highest concentrations when using PFE. The concentrations of different polyaromatic compounds varied between matrices. With diesel engine exhaust, high concentrations of phenanthrene, fluoranthene and pyrene were detected, while concentrations of high molecular weight, carcinogenic PAH compounds were low. With urban air samples, PAH levels were low but the highest concentrations were detected for benzo[ghi]perylene, indeno[1,2,3-cd]pyrene, benzo[b]fluoranthene and pyrene. Highly volatile naphthalene, acenaphthene and fluorene were low in concentration, regardless of the matrix. Further work for the development of separating methods is needed on the chromatographic analysis of PAHs derived from volatile, semi-volatile and particle fractions of environmental air samples. In this study, PFE was a preferred extraction procedure for diesel exhaust and urban air samples due to its extraction efficiency, time span and usefulness. Acknowledgements The EU MAAPHRI QLK4-CT is acknowledged for financial support. References 1 IARC, Polynuclear Aromatic Compounds. Part 1. Chemicals, Environmental and Experimental Data, International Agency for Research on Cancer, Lyon, France, IARC, Polynuclear aromatic hydrocarbons, Part 2, Carbon blacks, mineral oils (lubricant base oils and derived products) and some nitroarenes, International Agency for Research on Cancer, Lyon, France, HEI, Diesel Exhaust: A Critical Analysis of Emissions, Exposure, and Health Effects (A Special Report of the Institute s Diesel Working Group), Health Effects Institute, Cambridge, MA, IARC, Overall evaluations of carcinogenicity: an updating of IARC Monographs volumes 1 to 42, International Agency for Research on Cancer, Lyon, France, IARC, IARC monographs on the evaluation of carcinogenic risks to humans. Diesel and gasoline engine exhausts and some nitroarenes, International Agency for Research on Cancer, Lyon, France, H.-H. Mi, W.-J. Lee, S.-J. Chen, T.-C. Lin, T.-L. Wu and J.-C. Hu, Chemosphere, 1998, 36, H.-H. Mi, W.-J. Lee, C.-B. Chen, H.-H. Yang and S.-J. Wu, Chemosphere, 2000, 41, S. K. Pohjola, M. Lappi, M. Honkanen, L. Rantanen and K. Savela, Mutagenesis, 2003, 18, J. Kokko, L. Rantanen, J. Pentikäinen, T. Honkanen, P. Aakko and M. Lappi, SAE Tech. Pap. Ser., 2000, , L. Ramos, J. J. Vreuls and U. A. T. Brinkman, J. Chromatogr., A, 2000, 891, E. Bjo rklund, T. Nilsson and S. Bowadt, Trends Anal. Chem., 2000, 19, S. B. Hawthorne, C. B. Grabanski, E. Martin and D. J. Miller, J. Chromatogr., A, 2000, 892, O. P. Heemken, N. Theobald and B. W. Wenclawiak, Anal. Chem., 1997, 69, B. E. Richter, B. A. Jones, J. L. Ezzell, N. L. Porter, N. Avdalovic and C. Pohl, Anal. Chem., 1996, 68, M. M. Schantz, J. J. Nichols and S. A. Wise, Anal. Chem., 1997, 69, D. L. Poster, M. M. Schantz, S. A. Wise and M. G. Vangel, Fresenius J. Anal. Chem., 1999, 363, M. Mäkelä and L. Pyy, J. Chromatogr., A, 1995, 699, This journal is c The Royal Society of Chemistry 2006 J. Environ. Monit., 2006, 8,