Comparative study of the linearity and sensitivity of electron capture and flame photometric detectors using a pesticide standard.

A study was undertaken to compare and evaluate the different types of 63Ni and 3H electron capture (ECD) and flame photometric detectors (FPD) as part of the Canadian Check Sample Program on Pesticide Residue Analysis. Twenty-seven laboratories were supplied with chlorpyrifos standard solutions for the determination of linear range and minimum detectable amounts (MDA). Each laboratory selected its own operating conditions except for the FPD study. Results for the FPD were compared with standardized flame conditions designed to optimize the exponential factor i, the S-mode. The MDA of the ECDs (pulsed and direct-current) ranged from 0.09 to 0.17 pg/sec for 63Ni and 3H sources respectively. The corresponding MDA for the linearized ECDs (pulse modulated or constant current) was 0.04 pg/sec and a linear range of 10(4) or greater with nitrogen as the carried gas. Use of argon/methane and a pulse width of 0.1 usec extended this range to 10(5) but the sensitivity was not as great as with nitrogen carrier gas. The practical limit of detector sensitivity was found to be instrument dependent depending primarily on the noise level. In the FPD study, operating in the P-mode gave an average MDA of 5.97 pg/sec and a linear range of 2 x 10(3). Similarly, the S-mode parameters were 81.1 pg/sec and 8 x 10(2) respectively with an average exponential factor of n 1.8. Attempts to apply optimized detector gas-flow conditions improved sensitivity and linearity in the S-mode only.     

A new gas chromatographic determination of ethylenethiourea residues without derivatization

Abstract

A new gas chromatographic method reliable in the microgram range, was developed to measure ETU residues in various crops. Contrary to previously published procedures, this new method does not require prior derivatization of the ETU. The limit of sensitivity was shown to be 0.01 ppm using the flame photometric detector, FPD (sulfur mode, 394 nm).     

Biogenic Sulfur Gas Emissions from Soils in Eastern and Southeastern United States

Data are presented for the first systematic measurements of biogenic sulfur gas flux from the major soil orders within the eastern and southeastern United States. Sulfur flux samples were collected and analyzed on-site during the fall of 1977, spring and summer of 1978 and summer of 1979. A total of 27 sampling locales in 17 states were examined. Eight additional sites were visited in 1980.

At some locales, two to four soils were examined, providing an even broader sampling of the soil orders. Three of the locales were revisited two or three times during the course of the study to establish the influence of seasonal climatology upon the measured emission rates and chemical composition of the sulfur flux mixtures.

The sulfur gas enhancement of sulfur-free sweep air passing through dynamic emission flux chambers placed over selected sampling areas was determined by combined cryogenic enrichment sampling and wall-coated, open tubular, capillary column, cryogenic gas chromatography (WCOT/GC) using a sulfur selective, flame photometric detector (FPD).

Sulfur gas mixtures varied with soil order, ambient temperature, insolation, soil moisture, cultivation, and vegetative cover. Statistical analyses indicated strong temperature and soil order relationships for sulfur emissions from soils.

Fluxes ranged from 0.001 g to 1940 g of total sulfur as S/m²/yr. The calculated mean annual sulfur flux, weighted by soil order, was 0.03 g S/m²/yr for the study land area, or 110,872 metric tons (mT). The estimated annual average sulfur flux increased from 65 mT per 6400 km² for the land grids in the northernmost east-west grid tier to an average 1800 mT for the land grids in the southern Florida grid tiers.

This systematic sampling of major soils provides a much broader data base for estimating biogenic sulfur flux than previously reported for isolated intertidal sites, and presents the first sulfur flux estimates for inland soils which make up approximately 93% of the land of the eastern United States.     

Trace phthalate and adipate esters contaminated in packaged food

A method for trace analysis of two plasticizers, di-2-ethylhexyl phthalate (DEHP) and di-2-ethylhexyl adipate (DEHA), contaminated in packaged curry paste were investigated by gas chromatography with flame ionization detector (GC-FID). Curry paste samples were extracted by ultrasonic and solid phase extraction using Florisil® cartridge. Analysis by the GC-FID system provided limits of detection for DEHA and DEHP at 12 and 25 μ g L^{? 1} and a linear dynamic range between 25 μ g L^{? 1} to 60 mg L^{? 1} with a coefficient of determination (R²) greater than 0.99. High recoveries were obtained, ranged from 91 to 99% and 88 to 98% for DEHP and DEHA with RSD lower than 7 and 10% respectively. The method detection limit and limits of quantitation were ranged from 27 to 30 and 90 to 100 μ g L^{? 1}. The analysis of curry paste samples showed concentrations of DEHP and DEHA in the range of 4.0 ng g^{? 1} to 0.61 μg g^{? 1}.     

Determination of organochlorine and organophosphorus pesticide residues in low moisture,nonfatty products using a solid phase extraction cleanup and gas chromatography

Abstract

A multiresidue solid‐phase extraction (SPE) method for the isolation and subsequent gas Chromatographie determination of organochlorine and organophosphorus pesticide residues in low‐moisture, nonfatty products is described. Residues are extracted from samples with an acetonitrile/water mixture. Cleanup of the extract is performed using graphitized carbon black and anion exchange SPE columns, and analysis is performed by gas chromatography with Hall electrolytic conductivity and flame photometric detection. Recovery data was obtained by fortifying corn, oats and wheat with pesticides. The average recoveries were 79–123% for eight organochlorine and 51–122% for 28 organophosphorus pesticide residues. The limit of quantitation for chlorpyriphos was 0.05 ppm using the Hall electrolytic conductivity detector and <0.005 ppm using the flame photometric detector.     

Solid-phase extraction for multi-residue analysis of pesticides in honey

A fast and simple multi-residue method for the analysis of 15 organophosphorus (OP), 17 organochlorine (OC), 8 pyrethroids (PYR), 12 N-methyl-carbamate (NMC) pesticide residues and bromopropylate in honey is presented. Ready–to–use EXtrelut®NT 20 column, eluted with dichloromethane, was used to extract the pesticide residues from the aqueous-acetone honey sample, obtaining a clean extract directly analyzable. Determination was carried out by gas chromatography (GC) coupled with flame photometric detector (FPD) for OP compounds and by GC coupled with mass spectrometry detector (MSD) for OC and PYR pesticides and bromopropylate. The NMC pesticides were analysed by liquid chromatography-double derivatization coupled with spectrofluorimetric detector (LC/DD/Fl). This method allows the determination of the 53 pesticide residues at low concentrations (0.0005–0.074 mg/kg) and can be used to assess the compliance with the Maximum Residues Levels (MRLs) set by the European Union. The performance of the method was evaluated and specificity, linearity, recovery, repeatability, reproducibility, limit of quantification (LOQ) and limit of detection (LOD) were determined. A good linearity (r²? 0.99) was found in the range 0.0005–0.074 mg/kg for the majority of the compounds studied. Most of the pesticides had recoveries in the range 70–103 % and values of relative standard deviation (RSD) < 20 for repeatability and reproducibility, showing good accuracy and precision of the method. Aldicarb partially degraded in aldicarb sulphoxide during the analytical procedure, giving anomalous values. The LOQ for all pesticides investigated was from 0.0005 to 0.025 mg/kg while the LOD ranged from 0.0002 to 0.008 mg/kg.     

Analysis and fate of acephate and its metabolite,methamidophos, in pepper and cucumber

Abstract

Levels of acephate (Orthene^R) and its principle metabolite, methamidophos, in/on greenhouse‐grown pepper and cucumber fruits and leaves in relation to the applied methamidophos were monitored. Dislodgeable and total residues of acephate and methamidophos were determined by gas‐liquid chromatography equipped with a flame ionization detector (GC‐FID) and were confirmed by nitrogen phosphorus detector (GC‐NPD). The dissipation curves of the residues followed first‐order kinetics (R²> 0.96). Initial residues of acephate on fruits varied between pepper (15.12 ppm) and cucumber (2.16 ppm) . Total residues in fruits and leaves determined at intervals following application revealed the greater persistence of acephate on pepper fruits (half‐life [t_1/2] of 6 d) than on cucumber fruits (t_1/2 was 3.7 d) . T_1/2 values for the applied methamidophos were 4.7 and 5.3 d on pepper and cucumber fruits, respectively. Deacety‐lation of acephate (formation of its metabolite) was detectable 1 d following acephate treatment and reached a maximum of 2.05% of initial acephate residues 3 d after application on pepper fruits. On cucumber fruits, acephate metabolite reached a maximum of 2.12% one wk following application. No acephate residues were detected above the limit of detection of 0.001 ppm in pepper fruits 50 d following acephate application while its metabolite was detectable at that time (detectability limit was 0.0001 ppm).     

Monitoring of pesticide residues in a cotton crop soil

Abstract

This paper reports on the residues of methyl parathion (O,O‐dimethyl O‐4‐nitrophenyl phosphorothioate), trifluralin (α, α, α‐trifluoro‐2, 6‐dinitro‐N, N‐dipropyl‐p‐toluidine), endosulfan [(1, 4, 5, 6, 7, 7‐hexachloro‐8, 9, 10‐trinorborn‐5‐en‐2, 3‐ylenebismethylene) sulfite] and dimethoate (O, O‐dimethyl S‐methylcarbamoylmethyl phosphorodithioate) in a cotton crop soil. Soil samples (0–15 cm) were collected at different periods from the cotton crop farm and subjected to Soxhlet extraction. The extracted material was analysed after clean‐up by a HP5890 II gas Chromatograph equipped with a ⁶³Ni electron‐capture detector (ECD‐⁶³Ni) and fitted with a 25m x 0,2mm i.d. fused silica capillary column [Ultra‐2 (5% phenylmethyl polysiloxane)]. The recoveries of the pesticide residues from the spiked control soil were determined after Soxhlet extraction and C₁₈ cartridges clean‐up by using radiotracer techniques with the corresponding ¹⁴C‐pesticides. The results show that in the cotton crop soil the pesticide residues under study were present in the range of 0.1 to 0.4 mg ? kg^‐1. Endosulfan was found to be rapidly degraded in the soil and formed a sulfate metabolite.     

Validation of QuEChERS method for the determination of 36 pesticide residues in fruits and vegetables from Ghana,using gas chromatography with electron capture and pulsed flame photometric detectors

In this study, “Quick, Easy, Cheap, Effective, Rugged and Safe” ‘QuEChERS’ method was modified for the determination of 36 pesticides fortified at (0.01–1.0) mg kg^?1 in three vegetables and a fruit (lettuce, carrot, tomatoes and pineapples respectively) from Ghana. The method involved extraction with acetonitrile, phase separation with primary secondary amine and magnesium sulfate; the final injection solution was reconstituted in ethyl acetate. Organochlorine and synthetic pyrethroids residues were detected with electron capture detector whereas organophosphorus, pulsed flame photometric detector was used. The recoveries at different concentration levels (0.01, 0.1 and 1.0 mg kg^?1) were in the range of 83% and 93% with relative standard deviation ranging from 2% to 10% (n = 5) and the coefficient of determination (R²) was greater than 0.99 for all the 36 pesticides. The method was successfully tested on 120 real samples from Accra markets and this proved to be useful for monitoring purposes particularly in laboratories that have no gas chromatography-mass spectrometry and liquid chromatography-mass spectrometry.     

Liquid Entrainment from a Mobile Bed Scrubber

Liquid entrainment rate and drop size distribution were measured in the exhaust gas stream from a mobile bed scrubber. The pilot plant scrubber was 46 cm (18 in.) square and was packed with 3.8 cm (1.5 In.) diameter hollow polyethylene spheres to a static depth of 25 cm (10 in.). Entrainment flow rate depends on both gas and liquid rates. At a liquid/gas ratio of 6.7 l/m³ (50 gal/Mcf) and a superficial gas velocity of 2.6 m/sec (8.5 ft/sec) the entrainment flow rate was 0.0064 l/m³ (0.05 gal/Mcf) and at 3.75 m/sec (12.3 ft/sec) it was 0.031 l/m³ (0.23 gal/Mcf). The mass median drop diameter was about 400 nm at a liquid/gas ratio of 6.7 l/m³. The drop size distribution appears to be bimodal. Dye impregnated paper and cascade impactor techniques were used to measure drop size.     

Simultaneous determination of eight metabolites of organophosphate and pyrethroid pesticides in urine

A simultaneous method for quantifying eight metabolites of organophosphate pesticides and pyrethroid pesticides in urine samples has been established. The analytes were extracted using liquid–liquid extraction coupled with WCX solid phase extraction (SPE) cartridges. Eight metabolites were chemically derivatized before analysis using gas chromatography–tandem mass spectrometry (GC–MS–MS). The separation was performed on a HP-5MS capillary column (30 m × 0.25 mm × 0.25 µm) with temperature programming. The detection was performed under electro-spray ionization (ESI) in multiple reaction monitoring (MRM) mode. An internal standard method was used. The extraction solvent, types of SPE cartridges and eluents were optimized by comparing the sample recoveries under different conditions. The results showed that the calibration curves of the five organophosphorus pesticides metabolites were linear in the range of 0.2–200 μg/L (r² ≥ 0.992) and that of the three pyrethroid pesticides metabolites were linear in the range of 0.025–250 μg/L (r² ≥ 0.991). The limits of detection (LODs, S/N ≥ 3) and the limits of quantification (LOQs, S/N ≥ 10) of the eight metabolites were 0.008–0.833 μg/L and 0.25–2.5 μg/L, respectively. The recoveries of the eight metabolites ranged from 54.08% to 82.49%. This efficient, stable, and cost-effective method is adequate to handle the large number of samples required for surveying the exposure level of organophosphorus and pyrethroid pesticides in the general population.     

Effect of Water and Carbon Dioxide in Chemiluminescent Measurement of Oxides of Nitrogen

ABSTRACT

This paper reports results of studies using a biotrickling filter with blast-furnace slag packings (sizes = 2–4 cm and specific surface area = 120 m²/m³) for treatment of ethylether in air stream. Effects of volumetric loading, superficial gas velocity, empty bed gas retention time, recirculation liquid flow rate, and biofilm renewal on the ethylether removal efficiency and elimination capacity were tested. Results indicate that ethylether removal efficiencies of more than 95% were obtained with an empty bed retention time (EBRT) of 113 sec and loadings of lower than 70 g/m³/hr. At an EBRT of 57 sec, removal efficiencies of more than 90% could only be obtained with loadings of lower than 35 g/m³/hr. The maximum elimination capacities were 71 and 45 g/m³/hr for EBRT = 113 and 57 sec, respectively. The maximum ethylether elimination capacities were 71 and 96 g/m³/hr, respectively, before and after the renewal at EBRT = 113 sec. With an EBRT of 113 sec and a loading of lower than 38 g/m³/hr, the removal efficiency was nearly independent of the superficial liquid recirculation velocity in the range of 3.6 to 9.6 m³/m²/hr. From data regression, simplified mass-transfer limited, and reaction- and mass-transfer limited models correlating the contaminant concentration and the packing height were proposed and verified. The former model is applicable for cases of low influent contaminant concentrations or loadings, and the latter is applicable for cases of higher ones. Finally, CO₂ conversion efficiencies of approximately 90% for the influent ethylether were obtained. The value is comparable to data reported from other related studies.     

改进GC/FID法连续观测大气中CO浓度

改进装配有氢火焰离子化检测器(即FID)的气相色谱仪(GC),可连续监测大气中痕量气体CO浓度.本系统采用单阀双柱反吹进样技术,优选前置柱能更有效地剔除杂质,提高了分析柱效率,保持色谱基线平稳,提高分辨率和定量分析的准确率.优化后的气路设计与色谱柱的改进,使GC/FID对CO的最低检出限达到10×10-9、精密度误差小于2%,准确度在±2%之内.将气相色谱与动态气体稀释仪耦合使系统能够自动进行工作曲线校准,系统自动采样、分析和标定,无需人员职守.对北京大气CO连续观测结果表明,北京大气CO浓度变化受气象要素与排放源双重控制.     

Surface Flux Measurements of CO2 and N2O from a Dried Rice Paddy in Japan during a Fallow Winter Season

Abstract

The CO₂ and N₂O soil emissions at a rice paddy in Mase, Japan, were measured by enclosures during a fallow winter season. The Mase site, one of the AsiaFlux Network sites in Japan, has been monitored for moisture, heat, and CO₂ fluxes since August 1999. The paddy soil was found to be a source of both CO₂ and N₂O flux from this experiment. The CO₂ and N₂O fluxes ranged from -27.6 to 160.4μg CO₂/m²/sec (average of 49.1 ± 42.7 μg CO₂/m²/sec) and from -4.4 to 129.5 ng N₂O/m²/sec (average of 40.3 ± 35.6 ng N₂O/m²/sec), respectively. A bimodal trend, which has a sub-peak in the morning around 10:00 a.m. and a primary peak between 2:00 and 3:00 p.m., was observed. Gas fluxes increased with soil temperature, but this temperature dependency seemed to occur only on the calm days. Average CO₂ and N₂O fluxes were 27.7 μg CO₂/m²/sec and 13.4 ng N₂O/m²/sec, with relatively small fluctuation during windy days, while averages of 69.3 μg CO₂/m²/sec and 65.8 ng N₂O/m²/sec were measured during calm days. This relationship was thought to be a result of strong surface winds, which enhance gas exchange between the soil surface and the atmosphere, thus reducing the gas emissions from soil surfaces.     

Presence of nitro-chlorinated dioxins on fly ash from municipal incinerators and from laboratory production through reactions between NO2 and T4CDD on fly ash

Treatment of tetrachlorodibenzo-p-dioxin (T₄CDD) on municipal incinerator fly ash at 30 to 150°C for 10 minutes using 5% (Vol/Vol) NO₂ in air resulted in production of nitro-T₄CDD. Percent conversion was between 40% at 50°C to 100% at 150°C. Nitro-T₄CDD produced in the laboratory was unstable in hexane/methanol or hexane/acetone solutions and decomposed at ?5°C with half-lives of 8–10 days^?1 to the original T₄CDD.Fly ash from a municipal incinerator in Toronto, Ontario was extracted using toluene/methanol for 12 hrs in a Soxhlet extractor. Condensate of this extract was analyzed using capillary gas chromatography with flame ionization, nitrogen-selective, and mass spectrometric detectors. Retention times for nitro-T₄CDD, multi-ion selected ion monitoring, and nitrogen detector response were used as supporting evidence for the presence of nitro-chlorinated dioxins as naturally occurring in this sample.     

Modeling the Reduction of Vapor Phase Emissions from Surface Soils Due to Soil Matrix Effects: Porosity/Tortuosity Concepts

ABSTRACT

A major route for transport of volatile organic compounds within porous media is vapor phase diffusion. The diffusion rate through a porous medium is less than that through free-air due to the decreased cross-sectional area available for gas movement and the increased path length due to pore tortuosity. Numerous empirical expressions have been published that relate the diffusion coefficient in porous media to the diffusion coefficient in free-air (unobstructed gas phase). Published measurements of relative diffusivity and air-filled porosity were combined into a database. Empirical expressions available in the literature, including the popular Millington-Quirk equation, were evaluated along with a fourth-degree polynomial expression developed by the authors to determine the best type of equation to predict relative diffusivity as a function of air-filled porosity over the domain of values for porosity ranging from 0.071 to 1 for different types of materials. Mean square deviations were used as the statistical test to compare equations. The polynomial expression developed in this project produced a significantly different effective diffusion coefficient (1.3 x 10^-6 m²/sec) compared to values of 9.2 x 10^-6 m²/sec and 3.1 x 10^-6 m²/ sec predicted by forms of the Millington-Quirk equation for a specific case.     

Improved fluorescence peak integration in the Tekran 2537 for applications with sub-optimal sample loadings

The Tekran 2537 is widely used for monitoring atmospheric mercury. Although the instrument was designed for sample volumes in excess of 7.5 L, some recent research applications (e.g. aircraft) have used the instrument with significantly smaller collection times and sample volumes – and therefore smaller Hg loadings per cycle – than for which the instrument was designed. We have noticed a potential for non-linear (low) response in the fluorescence peak integration scheme, and thus the reported concentrations when the Hg loading (per cycle) is less than about 10–15 pg, e.g. at around 1 pg loading, the sensitivity is 25% lower than at 10 pg. We determined that although the atomic fluorescence detector was fundamentally linear down to at least 1 pg, the default peak integration scheme appeared to be optimized for > 10–15 pg cycle^?1 and so could introduce non-linearity in smaller peaks (i.e. lower mass loadings). For research applications where achieving maximum accuracy and precision of individual, high-time resolution (<5 min) points is crucial, users can mitigate this behavior by modifying the integration parameters or recording the full fluorescence peak and processing the data offline. Two offline methods of quantifying the peak also improved the precision and thus suggest an improvement in the detection limit is possible.     

Concentrations,atmospheric partitioning,and air–water/soil surface exchange of polychlorinated dibenzo-p-dioxin and dibenzofuran along the upper reaches of the Haihe River basin,North China

Polychlorinated dibenzo-p-dioxin and dibenzofuran (PCDD/PCDF) were overall measured and compared in ambient air, water, soils, and sediments along the upper reaches of the Haihe River of North China, so as to evaluate their concentrations, profiles, and to understand the processes of gas–particle partitioning and air–water/soil exchange. The following results were obtained: (1) The average concentrations (toxic equivalents, TEQs) of 2,3,7,8-PCDD/PCDF in air, water, sediment, and soil samples were 4,855 fg/m³, 9.5 pg/L, 99.2 pg/g dry weight (dw), and 56.4 pg/g (203 fg TEQ/m³, 0.46 pg TEQ/L, 2.2 pg TEQ/g dw, and 1.3 pg TEQ/g, respectively), respectively. (2) Although OCDF, 1,2,3,4,6,7,8-HpCDF, OCDD, and 1,2,3,4,6,7,8-HpCDD were the dominant congeners among four environmental sinks, obvious discrepancies of these congener and homologue patterns of PCDD/PCDF were observed still. (3) Significant linear correlations for PCDD/PCDF were observed between the gas–particle partition coefficient (K _p) and the subcooled liquid vapor pressure (P _L ⁰) and octanol–air partition coefficient (K _oa). (4) Fugacity fraction values of air–water exchange indicated that most of PCDD/PCDF homologues were dominated by net volatilization from water into air. The low-chlorinated PCDD/PCDF (tetra- to hexa-) presented a strong net volatilization from the soil into air, while high-chlorinated PCDD/PCDF (hepta- to octa-) were mainly close to equilibrium for air–soil exchange.     

Estimation and characterization of unintentionally produced persistent organic pollutant emission from converter steelmaking processes

Unintentionally produced persistent organic pollutants (UP-POPs) including polychlorinated dibenzo-p-dioxins, and dibenzofurans (PCDD/Fs), polychlorinated biphenyls (PCBs), and polychlorinated naphthalenes (PCNs) were characterized and quantified in stack gas and fly ash from the second ventilation systems in five typical converters in five different steelmaking plants. The 2378-substituted PCDD/Fs (2378-PCDD/Fs) and dioxin-like PCB (dl-PCBs) toxic equivalents (TEQs) were 1.84–10.3 pg WHO-TEQ Nm^?3 in the stack gas and 5.59–87.6 pg WHO-TEQ g^?1 in the fly ash, and the PCN TEQs were 0.06–0.56 pg TEQ Nm^?3 in the stack gas and 0.03–0.08 pg TEQ g^?1 in the fly ash. The concentrations of UP-POPs in the present study were generally lower than those in other metallurgical processes, such as electric arc furnaces, iron ore sintering, and secondary metallurgical processes. Adding scrap metal might increase UP-POP emissions, indicating that raw material composition was a key influence on emissions. HxCDF, HpCDF, OCDF, HpCDD, and OCDD were the dominant PCDD/Fs in the stack gas and fly ash. TeCB and PeCB were dominant in the stack gas, but HxCB provided more to the total PCB concentrations in the fly ash. The lower chlorinated PCNs were dominant in all of the samples. The 2378-PCDD/F, dl-PCB, and PCN emission factors in stack gases from the steelmaking converter processes (per ton of steel produced) were 1.88–2.89, 0.14–0.76_, and 229–759 μg t^?1, respectively.     

Seasonal variations of particle-associated nitrosamines by gas chromatography–mass spectrometry in the atmospheric environment of Zonguldak, Turkey

A gas chromatography–mass spectrometry method has been proposed for the determination of low-level mutagenic and carcinogenic nitrosamines in particulate matter. The method includes the collection of particulate matters (PM_2.5 and PM₁₀) using a dichotomous Partisol 2025 sampler and extraction of the compounds from aqueous solution with dichloromethane/2-propanol after sonication with a slightly basic water solution prior to their GC-MS analysis in electron impact mode. The obtained recoveries of nitrosamines ranged from 92.4 to 99.2 %, and the precision of this method, as indicated by the relative standard deviations, was within the range of 0.95–2.46?%. The detection limits obtained from calculations using the GC-MS results based on S/N?=?3 were found within the range from 4 to 22 pg/m³. The predominant nitrosamines determined in particulate matter were N-nitrosodimethylamine, N-nitrosodiethylamine, N-nitrosodibutylamine and N-nitrosomorpholine. Furthermore, N-mono- and dinitrosopiperazine and N-nitrosoethylbutylamine were also determined. N-dinitrosopiperazine was detected in PM_2.5 samples at the highest concentrations of up to 22.85 ng/m³ and in PM_2.5–10 samples at concentrations up to 7.60 ng/m³ in winter, whereas it was found in PM_2.5 samples up to 5.15 ng/m³ and in PM_2.5–10 samples up to 3.12 ng/m³ in summer. The total concentrations of nitrosamines were up to 161.4 ng/m³ in fine and 53.90 ng/m³ in coarse fractions in winter, whereas in summer were up to 35.24 and 12.60 ng/m³, respectively. The concentration levels of nitrosamines fluctuated significantly within a year, with higher means and peak concentrations in the winter compared to that in the summertime. The seasonal variations of particle-associated nitrosamine concentrations were investigated together with their relationships with meteorological parameters using Pearson’s correlation analysis in the winter and summer periods. Analysis of variance was used to determine which concentrations of nitrosamines were statistically different from one another and, together with meteorological parameters and discriminant analysis, was used to classify the particle samples by particle size according to seasons. The classification results of the particle samples in different seasons were very satisfactory, allowing 99.5 % of cases to be correctly grouped.