Accepted Manuscript

Assessment of a passive sampling method and two on-line gas chromatographs

for the measurement of benzene, toluene, ethylbenzene and xylenes in ambient
air at a highway site

Robert M. Healy, Jonathan M. Wang, Nicholas S. Karellas, Aaron Todd, Uwayemi Sofowote,
Yushan Su, Anthony Munoz

PII: S1309-1042(18)30656-1

DOI: 10.1016/j.apr.2019.01.017

Reference: APR 511

To appear in: Atmospheric Pollution Research

Received Date: 31 October 2018

Accepted Date: 22 January 2019

Please cite this article as: Robert M. Healy, Jonathan M. Wang, Nicholas S. Karellas, Aaron Todd,
Uwayemi Sofowote, Yushan Su, Anthony Munoz, Assessment of a passive sampling method and
two on-line gas chromatographs for the measurement of benzene, toluene, ethylbenzene and
xylenes in ambient air at a highway site, Atmospheric Pollution Research (2019), doi: 10.1016/j.apr.
2019.01.017

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form.
Please note that during the production process errors may be discovered which could affect the
content, and all legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT

Assessment of a passive sampling method and two on-line gas

chromatographs for the measurement of benzene, toluene,
ethylbenzene and xylenes in ambient air at a highway site

Robert M. Healy*1, Jonathan M. Wang1, Nicholas S. Karellas1, Aaron Todd1, Uwayemi

Sofowote1, Yushan Su1, Anthony Munoz1

1Environmental Monitoring and Reporting Branch, Ontario Ministry of the Environment,

Conservation and Parks, Toronto, Canada

*Corresponding author contact details:

robert.healy@ontario.ca
Ontario Ministry of the Environment, Conservation and Parks
125 Resources Road, Toronto, ON M9P 3V6, Canada
 ACCEPTED MANUSCRIPT

1 Abstract

2 Accurate, continuous long-term measurements of aromatic volatile organic compounds with

3 known health impacts are needed for comprehensive assessments of air quality in the vicinity of

4 industrial and vehicular emission sources. Historically, evacuated canister sampling and on-line

5 gas chromatography (GC) have been the most popular approaches applied by government

6 institutions in North America to measure benzene, toluene, ethylbenzene and xylenes (BTEX) in

7 ambient air. Canister sampling is labour-intensive and typically does not provide the continuous

8 data needed for epidemiological studies, while on-line GCs are costly and require shelter and

9 power. In this work, the accuracy and suitability of three techniques for generating long-term

10 ambient BTEX datasets are assessed. The first is a passive sampling method involving sorbent-

11 packed thermal desorption tubes, the second is an on-line miniature gas chromatograph

12 (miniGC) coupled with a photoionization detector (PID) and the third is a more traditional on-

13 line gas chromatograph coupled with a flame ionization detector (GC-FID). These three

14 techniques were deployed at an air quality monitoring station located adjacent to a busy highway

15 for six weeks in summer 2018 to compare method performance and accuracy. Ambient mean

16 concentrations of BTEX determined from the three methods for the entire study period were

17 found to agree within 7% for benzene and within 30% for the other species. Interestingly, a

18 temperature dependent positive bias previously identified for 14-day passive tube sampling for

19 ethylbenzene and xylenes under cold wintertime conditions was found to be negligible under

20 summertime conditions.

21

22 Keywords: VOCs; benzene; passive sampling; continuous monitoring; air quality

2
 ACCEPTED MANUSCRIPT

23 1. Introduction

24 Exposure to aromatic volatile organic compounds (VOCs) in air is a concern for those living in

25 environments impacted by industrial, vehicular traffic and combustion emissions (Lau and Chan,

26 2003; Shine, 2007; Correa et al., 2012; Gilman et al., 2015; Shuai et al., 2018; McMullin et al.,

27 2018). Several aromatic VOCs are associated with negative health outcomes and identified as

28 hazardous air pollutants by the United States Environmental Protection Agency (USEPA, 2018).

29 Chronic exposure to benzene, in particular, is associated with increased cancer risk in humans

30 (Loomis et al., 2017). Effective epidemiological assessment of the risks associated with chronic

31 exposure to benzene and other aromatic VOCs in ambient air requires long-term continuous

32 monitoring data.

33 Under Title III of the Clean Air Act Amendments of 1990, the USEPA was required to develop

34 approaches to reduce toxic air pollutants, including VOCs, in ambient air. This led to the

35 development of EPA Method TO-15 (USEPA, 1999a), which involves the use of clean pre-

36 evacuated canisters to slowly sample a known volume of air over a defined time period.

37 Canisters are subsequently analyzed off-line by gas chromatography-mass spectrometry (GC-

38 MS). Common applications of the method include point-of-impingement testing in the vicinity of

39 industrial facilities and monitoring of VOCs in urban environments (Conley et al., 2005; WDNR,

40 2007; Ras et al., 2009; Mukerjee et al., 2018). A modified version of the method is also used by

41 the Canadian federal government for monitoring VOCs at sites across the country under the

42 National Air Pollution Surveillance (NAPS) program (Wang and Austin, 2006). However,

43 canister data are typically not continuous, often involving 24-hour deployments covering only

44 one out of every six days, for example.

3
 ACCEPTED MANUSCRIPT

45 Several other technologies exist for the measurement of VOCs in ambient air, including on-line

46 chemical ionization mass spectrometers (Warneke et al., 2003; Badjagbo et al., 2009; Karellas

47 and Chen, 2013; Healy et al., 2017a), membrane introduction mass spectrometers (Davey et al.,

48 2011; Krogh and Gill, 2014), on-line gas chromatographs (GC) with a variety of detectors

49 (Rappenglück and Fabian, 1999; Liaud et al., 2014; Nasreddine et al., 2016), actively pumped

50 sorbent-packed tubes that require off-line analysis (USEPA, 1999b; Durmusoglu et al., 2010),

51 and passive sampling devices that also require off-line analysis (McClenny et al., 2005;

52 Zabiegała et al., 2010; USEPA, 2015; McAlary et al., 2014).

53 The need for continuous, long-term aromatic VOC data for effective exposure assessment rules

54 out expensive research-grade instruments such as real-time mass spectrometers, and thus less

55 expensive, more robust passive samplers and on-line GC systems represent the most promising

56 alternatives to canister sampling.

57 Passive sampling thermal desorption (TD) tubes packed with Carbopack X have been deployed

58 for one-week periods as part of a summertime study in Michigan and the results were compared

59 with both on-line gas chromatograph-flame ionization detector (GC-FID) measurements and

60 traditional evacuated canister analysis for benzene, toluene, ethylbenzene and xylenes (BTEX).

61 While the passive tube BTEX measurements were found to be precise, compound-specific

62 positive and negative biases were observed for the tube method relative to the on-line GC and

63 evacuated canister methods (Mukerjee et al., 2009). More recently, the same one-week tube

64 deployment method was again found to be precise for BTEX, but statistically significant positive

65 biases were observed relative to canister analysis at a refinery site in Indiana in the fall, although

66 differences between the methods remained within 0.1 ppbv (Mukerjee et al., 2018). Two-week

67 passive tube deployments were demonstrated to agree quite well with on-line gas GC-FID

4
 ACCEPTED MANUSCRIPT

68 measurements during a year-round study at a refinery site in Texas, however only benzene was

69 investigated in that case and the two-week sampling duration was found to hinder source

70 apportionment applications (Thoma et al., 2011).

71 Traditionally, on-line GC-FID systems have been favoured for accurate continuous BTEX

72 monitoring at locations where air quality stations are available (Oliver et al., 1996; Brown et al.,

73 2007), however GC instruments featuring photoionization detectors (PID) are becoming more

74 prevalent, typically using pressurized nitrogen or purified air as a carrier gas (Liaud et al., 2014;

75 Marć et al., 2015; Nasreddine et al., 2016; Zhang et al., 2017). Recent advances in micro-electro-

76 mechanical systems (MEMS) have led to the development of miniature GC-PID (miniGC)

77 instruments that do not require any external pressurized gases, providing greater portability

78 (Zampolli et al., 2009). To the best of our knowledge, there has only been one field assessment

79 of the performance and accuracy of the miniGC system relative to routine evacuated canister

80 measurements for BTEX, with no statistically significant differences observed (Healy et al.,

81 2018).

82 In this work, the suitability of three methods for the continuous monitoring of BTEX in ambient

83 air are assessed: an on-line miniGC featuring a PID detector, an on-line GC-FID system, and a

84 passive sampling tube method involving off-line thermal desorption gas chromatography mass

85 spectrometry (TD-GC-MS). These methods were deployed concurrently in Ontario for six weeks

86 in summer 2018 to investigate their accuracy and their suitability for ambient air monitoring

87 applications requiring different temporal resolutions. Comparisons with a previous wintertime

88 study at the same site involving two of the three methods are also discussed.

89

5
 ACCEPTED MANUSCRIPT

90 2. Methods

91 2.1 Sampling site

92 All measurements were performed at an air quality station operated by the Ontario Ministry of

93 the Environment, Conservation and Parks (MECP) located adjacent to a major highway

94 (HWY401) in Toronto, Ontario, Canada. This site has been described in detail previously (Healy

95 et al., 2017b), and this section of highway is characterized by very high daily traffic volumes

96 (MTO, 2017). The passing vehicles represent a near-continuous source of BTEX suitable for

97 investigating traffic-related emissions. The on-line GC-FID was rack-mounted within the air

98 quality station, the miniGC system was housed in a temperature-controlled portable enclosure

99 beside the station and the passive sampling tubes were housed in outdoor shelters on the roof of

100 the station. The three methods were calibrated using independent certified BTEX gas standards

101 and independent gas dilution systems. Measurements were performed for a six-week period from

102 10 July - 21 August 2018. Mean 24-hour ambient temperatures were in the range of 20-29 °C

103 during this study.

104 2.2 Passive tubes and TD-GC-MS system

105 The passive tube method is based on EPA Method 325 and has been described in detail

106 previously (Healy et al., 2018). Briefly, stainless steel TD tubes packed with Carbopack X

107 sorbent are fitted with diffusion endcaps and placed in basic outdoor shelters to provide

108 protection from precipitation. In this study, tubes were deployed for 24-hour periods and 14-day

109 periods. BTEX species diffuse to the sorbent and are retained; published uptake rates (McClenny

110 et al., 2005) and ambient temperature data are then applied to back-calculate ambient

111 concentrations from the mass of each analyte measured on the TD tubes. For each 14-day

6
 ACCEPTED MANUSCRIPT

112 deployment triplicate samples and a single field blank were used, and for 24-hour deployments

113 one duplicate sample and one field blank were deployed once every seven days in line with

114 Method 325 requirements (USEPA, 2015). Where duplicate or triplicate samples were deployed,

115 the final reported concentration is the mean of the individual sample concentrations, and

116 precision for replicates was less than 13%. Calibrations were performed by doping TD tubes

117 with 600 mL of a 1 ppmv VOC standard (Praxair) diluted in nitrogen using an Environics Series

118 2014 computerized VOC gas dilution system. BTEX signals in the field blanks were subtracted

119 from the corresponding field sample signals, but were below the method detection limit in every

120 case. A laboratory blank and a calibration check were also included in each TD-GC-MS

121 sequence.

122

123 2.3 miniGC system

124 The Airpointer miniGC BTEX analyzer (mlu-recordum Environmental Monitoring Solutions

125 GmBH, Austria) has been described in detail previously (Healy et al., 2018). Briefly, the miniGC

126 samples VOCs in ambient air onto a preconcentrator at 300 mL min-1 for 9.7 minutes. The

127 preconcentrator temperature is increased from 23 to 105 °C at ~27 °C s-1 and held at 105 °C for 4

128 minutes. The MEMS chip GC column is ramped from 32 to 45 °C at ~7 °C s-1, held at 45 °C for

129 0.5 minutes, then ramped from 45 to 175 °C at 24 °C min-1, and held at 175 °C for 2.5 minutes.

130 BTEX species are quantified upon elution using a PID (10.6 eV). Air is used as the carrier gas

131 with a flow rate of 6 mL min-1 and no pressurized cylinders are required for this instrument,

132 providing mobility and reducing siting requirements. The instrument was calibrated immediately

133 prior to the study period using a 1 ppmv BTEX standard (Praxair) diluted in zero air using an

7
 ACCEPTED MANUSCRIPT

134 Environics 6100 gas dilution system. Internal permeation tubes containing benzene and toluene

135 were used for concentration span checks at midnight every 24 hours. The miniGC provides

136 BTEX concentration data at 15-minute resolution, but data were averaged to hourly, 24-hour and

137 14-day resolution for direct comparison with the other methods.

138 2.4 GC-FID system

139 An AMA GC5000 series on-line GC-FID (AMA Instruments GmbH, Germany) was also

140 deployed consisting of two separate GC-FID modules; one for the measurement of more volatile

141 C2-C6 VOCs, and a second for the measurement of less volatile C6-C12 VOCs. We will report

142 data from the second module only as this study is focused exclusively on BTEX; the only four

143 species that can be measured using all three methods. The instrument is described in detail

144 elsewhere (An et al., 2014). Briefly, the instrument samples ambient air at a flow rate of 12 mL

145 min-1 for 40 minutes of each hour onto an integrated preconcentrator. The preconcentrator is then

146 rapidly thermally ramped and the desorbed VOCs are separated using a capillary column with

147 the following temperature program: 50 ºC held for 15 minutes, ramped to 80 ºC at 6 ºC min-1 and

148 held for 0.05 minutes, ramped to 100 ºC at 10 ºC min-1 and held for 7 minutes, then finally

149 ramped to 180 ºC at 10 ºC min-1 and held for 8 minutes. Eluting VOCs are quantified by FID,

150 with the detector held at 180 ºC. The rack-mounted system requires pressurized gases for

151 operation, and includes an on-board hydrogen generator and a gas-dilution system used to dilute

152 a 100 ppv VOC mixture for three-point calibrations. Calibrations were performed immediately

153 prior to, and two months after, the study period with minimal differences observed between the

154 concentration-response profiles. While the GC-FID system has a larger footprint and poorer

155 temporal resolution than the miniGC, it offers the advantage of a much wider compound list. The

8
 ACCEPTED MANUSCRIPT

156 instrument provides BTEX concentration data at hourly resolution, and these data were averaged

157 to 24-hour and 14-day resolution for comparison with the other methods.

158

159 3. Results and Discussion

160 Air quality at the sampling location is strongly influenced by the local continuous traffic source.

161 As shown in the hourly resolution average diurnal trends from the two on-line GC systems

162 (Figure 1), BTEX concentrations peak during the morning rush hour period (05:00-09:00),

163 decrease during the afternoon due to increased atmospheric mixing and dispersion, before

164 increasing again in the evening with increasing atmospheric stability. This temporality is

165 consistent with other primary traffic-related pollutants measured at the same location, including

166 black carbon (Healy et al., 2017b).

9
 ACCEPTED MANUSCRIPT

1.2
Toluene GC-FID
Toluene miniGC
1.0 m,p-Xylene GC-FID
m,p-Xylene miniGC
Benzene GC-FID
Benzene miniGC
0.8 Ethylbenzene GC-FID
Ethylbenzene miniGC
ppbv

0.6

0.4

0.2

0.0
0 2 4 6 8 10 12 14 16 18 20 22
167 Hour of Day

168 Figure 1: Average hourly resolution diurnal trends for BTEX measured using the on-line GC-

169 FID and miniGC systems. Data for 00:00-01:00 are excluded from this comparison due to the

170 automated daily internal miniGC span check.

171 The 14-day passive tube deployment BTEX concentrations and the mean concentrations from the

172 two online GC instruments for the three 14-day periods of the study are shown in Table 1. The

173 overall six-week mean concentrations determined using the three methods agree within 7%,

174 15%, 14% and 8% for benzene, toluene, ethylbenzene and m,p-xylene, respectively. Statistically

175 significant differences between the three datasets were investigated using the Kruskal-Wallis test

176 (Kruskal and Wallis, 1952), a non-parametric one-way ANOVA, coupled with a post-hoc Dunn

177 multiple comparison test (Dunn, 1964), to compare pair-wise differences between the methods

10
 ACCEPTED MANUSCRIPT

178 (error rate, α = 0.05). No statistically significant differences were observed between the three

179 datasets using this approach. 14-day resolution temporal agreement was also good between all

180 three methods (R2 = 0.84-0.99, slope = 0.91-1.27). Thus, for continuous monitoring under

181 summertime ambient conditions the three methods provide comparable results.

182

183 Table 1: Concentrations determined for each 14-day period and overall study mean

184 concentrations (± 1σ) determined using the three methods. All values are in units of ppbv.

Period 1 Period 2 Period 3 Entire study

(14 days) (14 days) (14 days) (6 weeks)
Benzene
Passive Tubes 0.13 0.16 0.18 0.16 ± 0.03
miniGC 0.14 0.16 0.19 0.16 ± 0.03
GC-FID 0.14 0.17 0.20 0.17 ± 0.03
Toluene
Passive Tubes 0.44 0.50 0.55 0.50 ± 0.06
miniGC 0.49 0.56 0.65 0.57 ± 0.08
GC-FID 0.50 0.58 0.65 0.58 ± 0.08
Ethylbenzene
Passive Tubes 0.06 0.07 0.08 0.07 ± 0.01
miniGC 0.07 0.08 0.10 0.08 ± 0.02
GC-FID 0.06 0.07 0.08 0.07 ± 0.01
m,p-Xylene
Passive Tubes 0.22 0.26 0.29 0.26 ± 0.04
miniGC 0.19 0.24 0.30 0.24 ± 0.06
GC-FID 0.22 0.26 0.31 0.26 ± 0.05

11
 ACCEPTED MANUSCRIPT

185 In a previous evaluation of the passive tube method under wintertime conditions using evacuated

186 canister samples for reference, 14-day concentrations derived using the passive tubes were found

187 to be accurate for benzene and toluene, but biased high (by up to 90%) for ethylbenzene and m,p-

188 xylene, necessitating the development of temperature-dependent uptake rates (Healy et al.,

189 2018). This was attributed to direct deposition of the less volatile VOCs on the tube surfaces

190 under low-temperature sampling conditions (<7 °C). In that study the miniGC was found to be

191 accurate under low temperature conditions (the system is housed in a temperature-controlled

192 enclosure). The GC-FID system is also expected to be unaffected by wintertime conditions

193 because it is also housed in a temperature-controlled environment. Thus, for continuous, long-

194 term year-round monitoring, the three methods are expected to be equally valid for monitoring

195 benzene and toluene; however the passive tube method may be biased high for ethylbenzene and

196 m,p-xylene when ambient temperatures are lower than approximately 7 °C and changes in uptake

197 rates are not accounted for (Healy et al., 2018).

198 24-hour passive tube samples were also collected on 22 days of the six-week study. A

199 comparison of the distribution of results for the three methods for days when 24-hour passive

200 tubes were deployed is shown in Figure 2. The mean tube and on-line GC concentrations for the

201 22 days when 24-hour tube samples were deployed are also provided in Table 2. The mean

202 concentrations for all three methods agree within 6% for benzene, 23% for toluene, 27% for

203 ethylbenzene and 30% for m,p-xylene. Again, no statistically significant differences were

204 observed between the three methods. Temporal agreement is high between the two online GCs

205 for all four species at 24-hour resolution (R2 = 0.80-0.96, slopes = 0.85-1.08). Temporal

206 agreement is also high between the 24-hour passive tubes and the two online GCs for benzene

207 (R2 = 0.80-0.90, slope = 0.85-1.14) and toluene (R2 = 0.83-0.96, slope = 0.83-1.13), but poorer

12
 ACCEPTED MANUSCRIPT

208 for ethylbenzene (R2 = 0.59-0.62, slope = 0.46-0.57) and m,p-xylene (R2 = 0.42-0.68, slope =

209 0.51-0.62). The poorer temporal agreement observed for these two species suggest that positive

210 bias, although minimal, may be occurring for 24-hour passive tube sampling for less volatile

211 aromatic compounds under summertime conditions. However, any potential positive bias

212 observed for these two species in the summertime 24-hour samples is not statistically significant,

213 and if present would be minor (≤30%) relative to the positive bias previously observed for 24-

214 hour mean concentrations under low-temperature wintertime conditions (185-272%) (Healy et

215 al., 2018).

0.30 1.4
Benzene Toluene
0.25 1.2

1.0
0.20
0.8
ppbv
ppbv

0.15
0.6
0.10
0.4

0.05 0.2

0.00 0.0
Passive miniGC GC-FID Passive miniGC GC-FID

0.20 0.7
Ethylbenzene m,p-Xylene
0.6

0.15
0.5

0.4
ppbv

ppbv

0.10
0.3

0.2
0.05

0.1

0.00 0.0
Passive miniGC GC-FID Passive miniGC GC-FID

216

13
 ACCEPTED MANUSCRIPT

217 Figure 2: Box plots of 24-hour concentrations determined using the three methods. Horizontal

218 lines represent the median, boxes represent the 75th percentile and vertical lines represent the

219 90th percentile.

220

221

222

223 Table 2: Mean concentrations (± 1σ) determined using the three methods for the 22 days when

224 24-hour tube samples were deployed. All values are in units of ppbv.

Benzene Toluene Ethylbenzene m,p-Xylene

Passive Tubes 0.16 ± 0.04 0.49 ± 0.22 0.09 ± 0.04 0.35 ± 0.16

miniGC 0.16 ± 0.05 0.62 ± 0.27 0.09 ± 0.03 0.26 ± 0.09

GC-FID 0.17 ± 0.05 0.62 ± 0.23 0.07 ± 0.03 0.28 ± 0.10

225

226 4. Conclusions

227 The three methods investigated here are all suitable for accurate long-term continuous

228 monitoring of BTEX in ambient air, and differences between the methods during this six-week

229 study are not statistically significant. The on-line GC-FID and miniGC instruments require

230 power and temperature-controlled environments, and are therefore limited to fixed site or

231 portable air quality monitoring stations with access to electricity. However, these instruments

232 offer the advantage of finer temporal resolution, one hour and 15 minutes for the GC-FID and

233 miniGC, respectively, and are thus well suited to monitoring short-lived or intermittent emission

14
 ACCEPTED MANUSCRIPT

234 events. The passive tube method offers the advantage of greater spatial coverage at minimal cost,

235 and 14-day deployments generate only 26 samples for analysis per site per year. However, the

236 tube method has much poorer temporal resolution than the on-line GCs, with 24 hour

237 deployments likely representing the shortest viable sampling time possible due to the associated

238 method detection limits. While 14-day deployment periods are optimal for passive sampling, 24-

239 hour deployments can also be accurate, provided that measurements are limited to the warmer

240 seasons (temperatures above 7 °C). 24-hour passive tube measurements are expected to be

241 particularly useful for air quality compliance and source apportionment applications.

242 References

243 An, J., Zhu, B., Wang, H., Li, Y., Lin, X., Yang, H., 2014. Characteristics and source
244 apportionment of VOCs measured in an industrial area of Nanjing, Yangtze River Delta,
245 China. Atmos. Environ. 97, 206-214.
246 Badjagbo, K., Picard, P., Moore, S., Sauvé, S., 2009. Direct atmospheric pressure chemical
247 ionization-tandem mass spectrometry for the continuous real-time trace analysis of
248 benzene, toluene, ethylbenzene, and xylenes in Ambient Air. J. Am. Soc. Mass Spectr.
249 20, 829-836.
250 Brown, S.G., Frankel, A., Hafner, H.R., 2007. Source apportionment of VOCs in the Los
251 Angeles area using positive matrix factorization. Atmospheric Environment 41, 227-237.
252 Conley, F.L., Thomas, R.L., Wilson, B.L., 2005. Measurement of Volatile Organic Compounds
253 in the Urban Atmosphere of Harris County, Texas. J. Environ. Sci. Heal. A 40, 1689-
254 1699.
255 Correa, S.M., Arbilla, G., Marques, M.R.C., Oliveira, K.M.P.G., 2012. The impact of BTEX
256 emissions from gas stations into the atmosphere. Atmos. Pollut. Res. 3, 163-169.
257 Davey, N.G., Krogh, E.T., Gill, C.G., 2011. Membrane-introduction mass spectrometry (MIMS).
258 Trends Anal. Chem. 30, 1477-1485.
259 Dunn, O.J., 1964. Multiple Comparisons Using Rank Sums. Technometrics 6, 241-252.

15
 ACCEPTED MANUSCRIPT

260 Durmusoglu, E., Taspinar, F., Karademir, A., 2010. Health risk assessment of BTEX emissions
261 in the landfill environment. J. Hazard. Mater. 176, 870-877.
262 Gilman, J.B., Lerner, B.M., Kuster, W.C., Goldan, P.D., Warneke, C., Veres, P.R., Roberts, J.M.,
263 de Gouw, J.A., Burling, I.R., Yokelson, R.J., 2015. Biomass burning emissions and
264 potential air quality impacts of volatile organic compounds and other trace gases from
265 fuels common in the US. Atmos. Chem. Phys. 15, 13915-13938.
266 Healy, R.M., Bennett, J., Wang, J.M., Karellas, N.S., Wong, C., Todd, A., Sofowote, U., Su, Y.,
267 Di Federico, L., Munoz, A., Charland, J.-P., Herod, D., Siu, M., White, L., 2018.
268 Evaluation of a passive sampling method for long-term continuous monitoring of volatile
269 organic compounds in urban environments. Environ. Sci. Technol. 52, 10580-10589.
270 Healy, R.M., Chen, Q., Bennett, J., Karellas, N.S., 2017a. A multi-year study of VOC emissions
271 at a chemical waste disposal facility using mobile APCI-MS and LPCI-MS instruments.
272 Environ. Pollut.
273 Healy, R.M., Sofowote, U., Su, Y., Debosz, J., Noble, M., Jeong, C.H., Wang, J.M., Hilker, N.,
274 Evans, G.J., Doerksen, G., Jones, K., Munoz, A., 2017b. Ambient measurements and
275 source apportionment of fossil fuel and biomass burning black carbon in Ontario. Atmos.
276 Environ. 161, 34-47.
277 Karellas, N.S., Chen, Q.F., 2013. Real-time air monitoring of trichloroethlyene and
278 tetrachloroethylene using mobile TAGA mass spectrometry. J. Environ. Prot. 4, 99-105.
279 Krogh, E.T., Gill, C.G., 2014. Membrane introduction mass spectrometry (MIMS): a versatile
280 tool for direct, real-time chemical measurements. J. Mass Spectrom. 49.
281 Kruskal, W.H., Wallis, W.A., 1952. Use of Ranks in One-Criterion Variance Analysis. J. Am.
282 Stat. Assoc. 47, 583-621.
283 Lau, W.-L., Chan, L.-Y., 2003. Commuter exposure to aromatic VOCs in public transportation
284 modes in Hong Kong. Sci. Total Environ. 308, 143-155.
285 Liaud, C., Nguyen, N.T., Nasreddine, R., Le Calvé, S., 2014. Experimental performances study
286 of a transportable GC-PID and two thermo-desorption based methods coupled to FID and
287 MS detection to assess BTEX exposure at sub-ppb level in air. Talanta 127, 33-42.
288 Loomis, D., Guyton, K.Z., Grosse, Y., El Ghissassi, F., Bouvard, V., Benbrahim-Tallaa, L.,
289 Guha, N., Vilahur, N., Mattock, H., Straif, K., 2017. Carcinogenicity of benzene. Lancet
290 Oncol. 18, 1574-1575.

16
 ACCEPTED MANUSCRIPT

291 Marć, M., Tobiszewski, M., Zabiegała, B., Guardia, M.d.l., Namieśnik, J., 2015. Current air
292 quality analytics and monitoring: A review. Anal.Chim.Acta 853, 116-126.
293 McAlary, T., Groenevelt, H., Nicholson, P., Seethapathy, S., Sacco, P., Crump, D., Tuday, M.,
294 Hayes, H., Schumacher, B., Johnson, P., Gorecki, T., Rivera-Duarte, I., 2014.
295 Quantitative passive soil vapor sampling for VOCs- part 3: field experiments. Environ.
296 Sci. Process. Impacts 16, 501-510.
297 McClenny, W.A., Oliver, K.D., Jacumin Jr, H.H., Daughtrey Jr, E.H., Whitaker, D.A., 2005. 24
298 h diffusive sampling of toxic VOCs in air onto Carbopack X solid adsorbent followed by
299 thermal desorption/GC/MS analysis-laboratory studies. J. Environ. Monitor. 7, 248-256.
300 McMullin, T., Bamber, A., Bon, D., Vigil, D., Van Dyke, M., 2018. Exposures and health risks
301 from volatile organic compounds in communities located near oil and gas exploration and
302 production activities in Colorado (U.S.A.). Int. J. Environ. Res. Public Health 15, 1500.
303 MTO, 2017. Ministry of Transportation Ontario: Provincial highways: Traffic volumes 2016.
304 Mukerjee, S., Oliver, K.D., Seila, R.L., Jacumin Jr, H.H., Croghan, C., Daughtrey Jr, E.H., Neas,
305 L.M., Smith, L.A., 2009. Field comparison of passive air samplers with reference
306 monitors for ambient volatile organic compounds and nitrogen dioxide under week-long
307 integrals. J. Environ. Monitor. 11, 220-227.
308 Mukerjee, S., Smith, L.A., Caudill, M.P., Oliver, K.D., Whipple, W., Whitaker, D.A., Cousett,
309 T.A., 2018. Application of passive sorbent tube and canister samplers for volatile organic
310 compounds at refinery fenceline locations in Whiting, Indiana. J. Air Waste Manag.
311 Assoc. 68, 170-175.
312 Nasreddine, R., Person, V., Serra, C.A., Le Calvé, S., 2016. Development of a novel portable
313 miniaturized GC for near real-time low level detection of BTEX. Sens. Actuators B
314 Chem. 224, 159-169.
315 Oliver, K.D., Adams, J.R., Hunter Daughtrey, E., McClenny, W.A., Yoong, M.J., Pardee, M.A.,
316 1996. Technique for monitoring ozone precursor hydrocarbons in air at photochemical
317 assessment monitoring stations: Sorbent preconcentration, closed-cycle cooler
318 cryofocusing, and GC-FID analysis. Atmos. Environ. 30, 2751-2757.
319 Rappenglück, B., Fabian, P., 1999. An analysis of simultaneous online GC measurements of
320 BTEX aromatics at three selected sites in the Greater Munich Area. J. Appl. Meteorol.
321 38, 1448-1462.

17
 ACCEPTED MANUSCRIPT

322 Ras, M.R., Borrull, F., Marcé, R.M., 2009. Sampling and preconcentration techniques for
323 determination of volatile organic compounds in air samples. Trends Anal. Chem. 28, 347-
324 361.
325 Shine, B., 2007. EPA Docket No. EPA-HQ-OAR-2003-0146. Technical memorandum: Potential
326 low bias of reported VOC emissions from the petroleum refining industry.
327 Shuai, J., Kim, S., Ryu, H., Park, J., Lee, C.K., Kim, G.-B., Ultra, V.U., Yang, W., 2018. Health
328 risk assessment of volatile organic compounds exposure near Daegu dyeing industrial
329 complex in South Korea. BMC Public Health 18, 528.
330 Thoma, E.D., Miller, M.C., Chung, K.C., Parsons, N.L., Shine, B.C., 2011. Facility Fence-Line
331 Monitoring Using Passive Samplers. J. Air Waste Manag. Assoc. 61, 834-842.
332 USEPA, 1999a. Air Method, Toxic Organics-15 (TO-15): Compendium of methods for the
333 determination of toxic organic compounds in ambient air. Second Edition: Determination
334 of volatile organic compounds (VOCs) in air collected in specially-prepared canisters and
335 analyzed by gas chromatography/mass Spectrometry (GC/MS).
336 USEPA, 1999b. Air Method, Toxic Organics-17 (TO-17): Compendium of methods for the
337 determination of toxic organic compounds in ambient air. Second Edition: Determination
338 of volatile organic compounds in ambient air using active sampling onto sorbent tubes.
339 USEPA, 2015. Petroleum refinery sector risk and technology review and new source
340 performance standards; Final Rule, 40 CFR Parts 60 and 63, Federal Register, Vol. 80,
341 No. 230.
342 USEPA, 2018. Initial list of hazardous air pollutants with modifications.
343 Wang, D.K.W., Austin, C.C., 2006. Determination of complex mixtures of volatile organic
344 compounds in ambient air: canister methodology. Anal. Bioanal. Chem. 386, 1099-1120.
345 Warneke, C., de Gouw, J.A., Kuster, W.C., Goldan, P.D., Fall, R., 2003. Validation of
346 atmospheric VOC measurements by proton-transfer- reaction mass spectrometry using a
347 gas-chromatographic preseparation method. Environ. Sci. Technol. 37, 2494-2501.
348 WDNR, 2007. Wisconsin Department of Natural Resources: Evaluation of Passive Sampling for
349 Monitoring Roadway and Neighborhood Exposures to Benzene and Other Mobile Source
350 VOCs.

18
 ACCEPTED MANUSCRIPT

351 Zabiegała, B., Urbanowicz, M., Szymanska, K., Namiesnik, J., 2010. Application of passive
352 sampling technique for monitoring of BTEX concentration in urban air: Field comparison
353 of different types of passive samplers. J. Chromatogr. Sci. 48, 167-175.
354 Zampolli, S., Elmi, I., Mancarella, F., Betti, P., Dalcanale, E., Cardinali, G.C., Severi, M., 2009.
355 Real-time monitoring of sub-ppb concentrations of aromatic volatiles with a MEMS-
356 enabled miniaturized gas-chromatograph. Sens. Actuators B Chem. 141, 322-328.
357 Zhang, W.-q., Li, H., Zhang, Y.-j., Bi, F., Meng, L.-s., Zhang, X.-m., Mao, J.-y., Cheng, N.-l.,
358 Fang, B., Yang, Y., Chen, C., Guo, K.-x., Zhan, G.-e., Sha, J., Wang, X.-z., 2017. Fast
359 Determination of Monocyclic Aromatic Hydrocarbons in Ambient Air Using a Portable
360 Gas Chromatography–Photoionization Detector. Chromatographia 80, 1233-1247.

361

19
 ACCEPTED MANUSCRIPT

Assessment of a passive sampling method and two on-line gas

chromatographs for the measurement of benzene, toluene,
ethylbenzene and xylenes in ambient air at a highway site

Robert M. Healy*1, Jonathan M. Wang1, Nicholas S. Karellas1, Aaron Todd1, Uwayemi

Sofowote1, Yushan Su1, Anthony Munoz1

1Environmental Monitoring and Reporting Branch, Ontario Ministry of the Environment,

Conservation and Parks, Toronto, Canada

Highlights

One passive sampling method and two online methods for the measurement of benzene, toluene,

ethylbenzene and m,p-xylene (BTEX) in ambient air were evaluated

Measurements were performed at a busy highway site for six weeks in summer 2018

All three methods exhibit comparable accuracy for BTEX, with no statistically significant

differences observed

The passive sampling method provides accurate continuous BTEX monitoring data with high

spatial coverage at minimal cost