Monitoring of atmospheric polycyclic aromatic hydrocarbons through optimization of analytical techniques

Abstract

Atmospheric pollutant monitoring in low to medium-income countries can play a vital role in improving human welfare through reductions in environmental impacts and the global burden of disease. Polycyclic aromatic hydrocarbons (PAHs) have been widely used as markers of environmental pollutants in the context of ambient air, indoors, as well as in working environments. Sampling, extraction, sample clean up and analysis constitute the principal stages required for the determination of airborne PAHs. Most conventional methods utilize high volume samplers and high volume extraction methods for the analysis of atmospheric gaseous and particulate PAHs, which is time consuming and may introduce unreliable data for human exposure estimations. Additionally, large amounts of toxic solvents generate waste and could cause adverse environmental impacts.

In order to counteract these unfavorable environmental effects, recent studies worldwide have been focusing on the miniaturization of sampling and analysis techniques. However, most of these developed portable sampling devices for simple and efficient solvent minimized techniques are not easily available, especially in developing countries which have limited laboratory budgets. This research focused on multi-channel polydimethylsiloxane (PDMS) traps as portable denuder devices for the monitoring of semi-volatile organic compounds (SVOCs), with most attention being given to the United States Environmental Protection Agency (US EPA) priority PAHs. These portable PDMS samplers are currently being analyzed by direct thermal desorption using a commercial desorber, which is quite expensive to obtain, maintain and operate.

In response to challenges faced by laboratories in developing countries, a plunger assisted solvent extraction (PASE) method for multi-channel PDMS rubber trap samplers was developed as an alternative to direct thermal desorption for the monitoring of PAHs. Extraction parameters which are capable of influencing extraction efficiency of target analytes such as choice of extraction solvent, number of sequential extractions and PDMS trap orientation were investigated and optimized. The proposed method used minimum solvent (total 2 mL hexane) and was quicker (approximately 4 min needed for PASE extraction including weighing, compared to the 11.5 min desorption time). Additionally, the PASE technique was advantageous over thermal desorption (TDS) in that samples could be re-analyzed, as only 1 µL of the final extract was injected. However, limits of detection (LODs) for thermal desorption remained superior, ranging from 0.14 ng m-3 for acenaphthylene to 0.9 ng m-3 for benzo[g,h,i]perylene, as it analysed the whole sample. In comparison, PASE method LODs ranged from 13.6 ng m-3 for naphthalene to 227.1 ng m-3 for indeno[1,2,3-cd]pyrene in sampled air.

Two sequential extractions resulted in optimum overall extraction efficiencies of the target PAHs, and ranged from 76% for naphthalene to 99% for phenanthrene, with relative standard deviations (RSDs) below 6%. Although sequential extractions did not result in much improvement in extraction efficiencies of lighter PAHs (2-3 rings), it improved efficiencies for the heavier target analytes (4-6 rings). Notable quality control merits (mean for 15 PAHs) of the proposed PASE method included good repeatability based on intra-day variations, (n=5), which ranged from 0.5% for pyrene to 4.9% for naphthalene and acceptable recoveries (n=3) ranging from 76% for naphthalene to 99% for phenanthrene. Linearity of the calibration curves ranged from 0.983 for naphthalene to 0.999 for acenaphthene and acenaphthylene, which demonstrated the suitability of the method for routine monitoring of PAHs. The PASE procedure was firstly applied to the analysis of domestic outdoor fire (charcoal) air emission samples and naphthalene was the most dominant analyte with maximum concentrations of 9.5 µg m-3. Fluorene, anthracene, phenanthrene, fluoranthene and pyrene were also identified and quantified.

In the next phase of the study, the applicability of the PASE method was evaluated for the analysis of indoor and outdoor air samples from residential homes in rural and urban coastal Kenya (Taita Taveta and Mombasa counties). A concentration step was added to the PASE method, whereby extracts were blown down to 100 µL. Calibrations and quantification were based on 100 µL final extracts and LODs (based on 100 µL final extract, 1 µL injection and an air sampling volume of 5 L) ranged from 1.9 ng m-3 for 2-methylnaphthalene to 34.9 ng m-3 for indeno[1,2,3-cd]pyrene. A PDMS trap coupled to a portable sampling pump was used to sample combustion emissions (within the breathing zone) using a flow-rate of 500 mL min-1 for 10 min. Typical combustion devices in rural and urban residential areas of Kenya included gas, kerosene, charcoal (jiko), 3-stone and improved 3-stone stoves. In addition to the combustion device, other possible influencing factors for PAHs levels in indoor environments such as the type of dwelling, ventilation, geographical location and fuel used were also explored. Indoor area measurements during operation of cooking devices in rural areas resulted in higher total gaseous PAH concentrations per household (ranging from 11.37 to 85.20 µg m-3) compared to urban homes (ranging from <LOQ to 36.40 µg m-3). However, ambient PAH concentrations were higher in urban environments, likely due to traffic contributions. As evidenced by principal component analysis (PCA), the most pronounced variations were observed from wood burning emissions during use of traditional 3-stone stoves in rural households, with total PAH concentrations averaging 46.23 ±3.24 µg m-3 (n=6).

In order to evaluate the relative carcinogenic contribution of each individual PAH based on benzo[a]pyrene (BaP), BaP equivalent (BaPeq) values were calculated using selected Toxic Equivalence Factors (TEFs). From an overall analysis of household cooking fuel use in coastal Kenya, gas stoves were found to account for the least (0%) PAH emissions (no detected carcinogenic contribution) in indoor setups. Average BaPeq total concentrations for kerosene, jiko, 3-stone and improved 3-stone stoves were 43.31, 88.38, 309.61 and 453.88 ng m-3 respectively. These high variations in PAH concentrations and carcinogenic potencies indicate the importance of good combustion devices and well ventilated conditions to reduce possible health impacts in urban and rural communities.

Further laboratory controlled experiments were performed to shed more light on the principal operating mechanisms of the PDMS traps and quartz fiber filters in a denuder assembly. Efficient transmission of sample particles through the PDMS trap is of paramount importance otherwise gas phase SVOC concentrations would be over-estimated and particle phase SVOC concentrations would be underestimated if SVOCs were associated with these particles. In this light, experiments mainly addressed particle collection and transmission efficiencies for a wider size range of particles than had been previously studied, including more realistic ambient particulate matter for ambient measurements. Ambient aerosol and aqueous suspensions of monodisperse polystyrene latex (PSL) as test particles (0.3, 0.5 and 0.8 µm), were evaluated. A temperature controlled chamber consisting of an aerosol generator, dilution and mixing unit, a flow tube and the sampling apparatus was used to evaluate particle transmission and collection efficiencies. With regards to quartz fiber filters, collection efficiency for the tested PSL particles was high (>98%) with acceptable variation coefficients all less than 16%. However, the tested ambient particles were collected with lower efficiencies, ranging between 88-95%, most likely due to the higher diffusion coefficients of the smaller ambient particles and electrostatic effects. This data allowed for a better understanding of the denuder operating mechanisms and can provide useful guidelines as well as potential limitations which should be considered in future sampling applications and data interpretation.

In conclusion, the optimized multi-channel PDMS denuder device, extraction and analysis methods investigated in this study could find application in environmental laboratories, especially where cost considerations are important (such as Sub-Saharan Africa). Additionally, the application of these methods would facilitate the widespread monitoring of atmospheric PAHs in a cost effective manner and identify potential hotspots of regional pollution.