Environmental Science and Pollution Research

, Volume 21, Issue 8, pp 5628–5636

Evaluation of the reaction artifacts in an annular denuder-based sampler resulting from the heterogeneous ozonolysis of naphthalene

  • Mathieu Goriaux
  • Maryline Pflieger
  • Anne Monod
  • Sasho Gligorovski
  • Rafal S. Strekowski
  • Henri Wortham
Research Article

DOI: 10.1007/s11356-014-2503-x

Cite this article as:
Goriaux, M., Pflieger, M., Monod, A. et al. Environ Sci Pollut Res (2014) 21: 5628. doi:10.1007/s11356-014-2503-x
  • 117 Downloads

Abstract

The heterogeneous ozonolysis of naphthalene adsorbed on XAD-4 resin was studied using an annular denuder technique. The experiments involved depositing a known quantity of naphthalene on the XAD-4 resin and then measuring the quantity of the solid naphthalene that reacted away under a constant flow of gaseous ozone (0.064 to 4.9 ppm) for a defined amount of time. All experiments were performed at room temperature (26 to 30 °C) and atmospheric pressure. The kinetic rate coefficient for the ozonolysis reaction of naphthalene adsorbed on XAD-4 resin is reported to be (10.1 ± 0.4) × 10−19 cm3 molecule−1 s−1 (error is 2σ, precision only). This value is five times greater than the currently recommended literature value for the homogeneous gas phase reaction of naphthalene with ozone. The obtained rate coefficient is used to evaluate reaction artifacts from field concentration measurements of naphthalene, acenaphthene, and phenanthrene. The observed uncertainties associated with field concentration measurements of naphthalene, acenaphthene, and phenanthrene are reported to be much higher than the uncertainties associated with the artifact reactions. Consequently, ozone reaction artifact appears to be negligible compared to the observed field measurement uncertainty results.

Keywords

Polycyclic aromatic hydrocarbons Annular denuder Particulate matter Naphthalene Ozone Reaction artifact 

Introduction

Polycyclic aromatic hydrocarbons (PAHs) are one of the most widespread organic pollutants of great environmental concern. PAH’s environmental concern results from now well documented carcinogenic, mutagenic, and teratogenic properties. These toxic properties of PAHs have prompted the international community to establish guidelines to better understand the ubiquitous PAH pollution. In 1990, the United States of America has passed the Clean Air Act Amendment that identified PAHs as hazardous air pollutants that require future regulation. Further, in 1999, the European Commission created a work group with an aim to better understand and assess the ambient air PAH pollution under the Air Quality Framework Directive (96/62/EC). However, to date, no specific guideline value has been recommended for PAHs as such in air (WHO 2006). Difficulties arising from assigning a specific guideline value originate from PAH’s physico-chemical properties and atmospheric behavior. PAHs belong to a family of semi-volatile organic compounds (SVOCs) that partition themselves between the particulate phase and the gas phase. To date, the respective roles in the toxicity and adverse health effects of the PAHs themselves or the complex particulate matrix as such remain uncertain. To better understand the toxic effects of PAHs, it is necessary to better assess their atmospheric fate. The atmospheric fate of PAHs is strongly governed by the particle phase—gas phase partition. As a result, to better determine the atmospheric behavior of PAHs, it is necessary to better understand and quantify the relative gas phase and particulate phase PAH concentrations.

In their earlier work, Temime et al. (Temime et al. 2002; Temime-Roussel et al. 2004a) evaluated the gas phase collection efficiencies of selected PAHs using a commercially available eight-channel and 28.5-cm long annular denuder-based sampler (URG2000-30-CF) coated with XAD-4 resin. These authors developed a bench-top permeation system to generate a steady flux of selected PAHs to test the collection efficiencies of gaseous PAHs of their annular denuder-based sampler under a wide range of experimental conditions namely, temperature, relative humidity, PAH concentration, sampling flow rate, and sampling duration (Temime-Roussel 2002). Temime et al. reported gas phase collection efficiencies of naphthalene of more than 90 % under given experimental conditions employed (Temime-Roussel et al. 2004a).

In their subsequent and complementary work, Temime-Roussel et al. (2004b) tested the particulate phase transmission efficiencies of an annular denuder-based sampler coated with XAD-4 resin. Their results indicated average particle number transmission efficiencies of 96 and 95 % for sampling flow rates of 17 and 34 slm (standard liters per minute), respectively, for particles in 12-size categories with the aerodynamic diameter in the range of 0.04 to 8.24 μm (Temime-Roussel et al. 2004b). The particle mass transmission efficiencies were reported to be 84 (±14) % and 81 (±21) % for the sampling flow rates of 17 and 34 L min−1, respectively.

During sampling, the collected PAH concentration and chemical nature may change due to oxidation by ozone and a number of other photochemical oxidants associated with ozone (Atkinson et al. 1994; Finlayson-Pitts et al. 1997; Jacob 2000). However, atmospheric ozone is considered to be the main photochemical oxidant of sampled PAHs assuming that hydroxyl radicals are degraded in the sampling device (cyclone and nozzle among others) upstream to the annular denuder. It is of foremost importance to understand and quantify the degree of PAH degradation by ozone during sampling since a change in the PAH concentration profile may induce an overestimation or underestimation of the PAH’s toxic effects (Pitts-Jr. et al. 1981; Pöschl 2002). Recent field and laboratory studies showed a significant particulate phase PAH degradation during long sampling times at high ozone concentrations (Liu et al. 2006; Tsapakis et al. 2003). Further, Schauer et al. (2003) reported that the degradation of five-ring and six-ring PAHs on filters had a near-linear dependence on ozone volume mixing ratio and that a filter artifact can lead to a twofold underestimation of the actual atmospheric PAH concentration.

As a result, the objective of this work is to study the formation of artifacts resulting from sampling of PAHs using an annular denuder-based sampler. Naphthalene was used as a surrogate PAH since it is the lightest, the most volatile, and the most reactive PAH to obtain an upper limit value. Further, since naphthalene is the most volatile PAH, it is easy to manipulate and use under laboratory conditions.

Further, the denuder trapping efficiency was found to decrease with increased volatility of the compound under study (Gundel et al. 1995; Coutant et al. 1992). That is, less volatile PAHs have been observed to have higher trapping efficiencies and more volatile PAHs have been observed to have lower trapping efficiencies. Since naphthalene is the most volatile PAH, it was chosen in order to minimize sampling errors due to an overestimation of naphthalene losses due to poor denuder trapping.

Experimental

The experimental approach is similar to the one employed in previous studies where an annular denuder-based sampling method was used to determine atmospheric gas phase and particle phase PAH collection efficiencies (Temime et al. 2002; Temime-Roussel 2002; Temime-Roussel et al. 2004a; Temime-Roussel et al. 2004b). All experiments involved depositing a known quantity of naphthalene on the XAD-4 resin (particulate phase) and then measuring the quantity of solid naphthalene that reacted away after it was allowed to be exposed to a continuous flow of a known concentration of gaseous ozone (0.064 to 4.9 ppm in air) for a given amount of time. The PAH artifact apparatus is shown schematically in Fig. 1 (Goriaux 2006). As shown in Fig. 1, the PAH artifact apparatus consists of three parts, namely (1) the permeation chamber, (2) the ozone generation system, and (3) the two annular denuder tubes connected in parallel.
Fig. 1

Schematic representation of the PAH artifact apparatus used in this work. MCF Mass Flow Controller; BV on/off butterfly valve

The geometry of the PAH artifact apparatus was such that the permeation system was in parallel to the ozone generation system upstream of the two denuders placed in parallel. By way of control with two-way valves, the flow through the denuders could thus be directed to be naphthalene flow only (to coat the denuders) followed by flow containing either ozone in air or clean air. This configuration allowed first for a known quantity of naphthalene to be deposited on the XAD-4 resin coated on the walls of the two denuder tubes. The reaction of solid naphthalene with ozone was studied by following the naphthalene mass loss as a function of time. All experiments were carried out under laboratory conditions where the temperature and the mass flow rates were precisely controlled. The experimental details of the PAH artifact apparatus that are particularly relevant to this work are given below.

The permeation chamber

The permeation system used in this study to produce naphthalene gas flows is the same as the one used in previous work where a PAH vapor generator was developed to test annular denuder-based samplers (Temime-Roussel et al. 2002; Goriaux 2006). Briefly, the Pyrex tube permeation cell filled with naphthalene crystals and sealed using a PTFE semi-porous membrane (Durieux PTFE membrane; pore, 10–15 μm; thickness, 0.0005 in.) was placed in a temperature controlled Pyrex chamber (Fig. 1). A mass flow of pure nitrogen (purity > 99.999 %) gas was allowed to flow through the temperature controlled chamber and over the PTFE tube to ‘pick up’ a known quantity of gaseous naphthalene that was allowed to diffuse through the semi-porous membrane as a function of temperature. The resulting naphthalene/nitrogen flow was then allowed to flow into the mixing chamber where it was diluted with clean and dry air. The air was dried using a condensation method and then purified by allowing it to flow first through an active carbon filter and then over the XAD-2 resin. Under typical experimental conditions employed, the temperature of the permeation system was kept constant at T (°C) = 40 ± 0.02 and the nitrogen and air flows were 0.02 and 4.5 slm (standard L min−1), respectively.

To assure a regular production of naphthalene and system stability, the Pyrex tube containing naphthalene crystals was weighed every 2 days using a Cahn 1000 electrobalance (Cahn Instruments, Cerritos, CA). The plot of the observed naphthalene production as a function of time is shown in Fig. 2. The observed naphthalene loss (production) was 2.24 mg day−1. It can be seen in Fig. 2 that naphthalene was produced in a steady and regular manner implying good permeation system stability.
Fig. 2

Plot of the amount of naphthalene sublimed at 40 °C as a function of time. Solid line fit was obtained from linear least squares analysis and the resulting naphthalene production rate was 2.24 mg day−1

The ozone generation system

The ozone was generated by allowing a clean and dry air to flow through the Aqua-Sander electrical discharge ozone generator (Aqua-Sander C25). Similar to the situation listed above, the air used in the system was first dried by condensation and then cleaned by allowing it to flow through an active carbon filter and the XAD-2 resin. Ozone concentration was measured using the UV Absorption Ozone Analyzer (Environment S.A. Model O3 41M). Under typical experimental conditions, the air mass flow that was allowed to flow through the ozone generator was varied between 1.5 and 4.5 slm. The resulting maximum ozone concentration used was 4,900 ppbv and the uncertainty in ozone concentration was ±10 %.

The annular denuder tubes

The annular denuder tubes used in this work are similar to the ones used by Temime et al. (Temime et al. 2002; Temime-Roussel et al. 2004a; Temime-Roussel et al. 2004b). These investigators employed such tubes to study PAH partition between the gas phase and the particulate phase in the natural atmosphere. Annular denuder-based samplers were home-made from sandblasted Pyrex tubes. The denuder geometry was such that it consisted of two 35-cm concentric tubes of 3.4-cm internal diameter and 3.2-cm external diameter, respectively. The denuder tube dimensions were calculated based on the theoretical work of Possanzini et al. (1983) and Winiwarter (1989). The inner tube was sealed at both ends resulting in the annular width of 1 mm. Under typical experimental conditions, the flow rate was 4.5 slm. Under these flow conditions the flow was laminar (NRe < 300). Prior to each experiment, the denuder walls were coated with XAD-4 resin, a styrene–divinylbenzene polymer (surface area, 725 m2 g−1; porosity, 1–150 Å; diameter, 1 μm) as described by Gundel et al. (Temime et al. 2002; Gundel et al. 1995). The XAD-4 resin was proven to be an effective sorbent trap for volatile organic compounds such as PAHs (Gundel et al. 1995; Loiselle et al. 1991). The XAD-4 resin was first finely ground for 6 h using a planetary grinder (Retsch GmbH) to obtain XAD-4 particles of about 1 mm in diameter. The resulting resin particles were then Soxhlet washed several times using acetone, acetonitrile, and dichloromethane to remove possible contamination by alkylated derivatives of benzene, styrene, biphenyls, and other lower molecular weight PAHs (Temime-Roussel 2002; Hunt et al. 1982). After washing, the solvent was separated from the XAD-4 powder in an ultracentrifuge (Hettich Universal, 3,200 revolutions min−1) and the resin was dried at 100 °C for 48 h. After, 50 mg of resin was added to 20 mL of hexane and the resulting slurry was placed in an ultrasonic bath for 5 min. The resulting resin was applied to the walls of the denuder in several application/drying cycles. After each application, the hexane solvent was removed by flushing the denuder with a dry flow of nitrogen.

A typical experimental study of the heterogeneous degradation of solid naphthalene by ozone in an annular denuder-based sampler involved first preparation of two denuder tubes as described above. Then, the gaseous naphthalene flow from the permeation system was allowed to flow through the two denuder tubes placed in parallel for 15 to 30 min (see Fig. 1). This allowed for a known and equivalent quantity of gaseous naphthalene to be deposited on the XAD-4 resin in the two denuder tubes. After the elapsed time the gaseous naphthalene flow was diverted toward a leak and a flow of dry and clean air was allowed to flow through one denuder tube and a separate and equivalent flow of ozone in dry and clean air was allowed to flow through the second denuder tube. The maximum ozone concentration used was 4,900 ppb and the maximum naphthalene exposition time to ozone was 6 h. The ozone/air flow ranged from 1.55 to 4.5 slm. All experiments were performed under dry conditions. That is, no water vapor was added to the system.

The quantity of the deposited naphthalene in the denuder tubes may have varied from one experiment to another. Consequently, to assure experimental reproducibility, “blank” experiments (without the presence of ozone) were run to test and verify the influence of the air flow on the two denuder tubes that have been coated with XAD-4 resin and naphthalene. As a result, air/naphthalene mix was first allowed to flow through both the reference denuder tubes, [Nap]ref, and the test denuder tube that would be normally exposed to a flow of ozone/air mix, [Nap]O3. The results of the effects of the air flow without ozone on the two denuder tubes, i.e., [Nap]ref and [Nap]O3, coated with naphthalene are summarized in Table 1. The observed [Nap]O3/[Nap]ref ratios are 1.07 ± 0.10 and 0.88 ± 0.12 (errors are ±2σ, precision only) for air flow exposition times of 3 and 6 h, respectively.
Table 1

Influence of the air flow without the presence of ozone on the two denuder tubes previously coated with naphthalene. The air flow was 4.5 slm

Total number of experiments

Exposition time to air (min)

Averaged ratioa ([Nap]O3/[Nap]ref)

T (°C)

4

0

1.02 ± 0.16

26–28

5

180

1.07 ± 0.10

26–28

4b

180

1.07 ± 0.10

26–28

4

360

0.88 ± 0.12

26–28

aUncertainties are ±2σ and represent precision only

bAir flow was 1.55 slm

The observed [Nap]O3/[Nap]ref ratio is lower for the air flow exposition time of 6 h, but it is well within the uncertainty values for both sets of experiments. The observed difference in ratio is most likely due to the inherent physical differences of the denuder tubes themselves since they were both home made. Nonetheless, to take into consideration the observed difference in the [Nap]O3/[Nap]ref ratios, same denuder tubes in the same configuration were always used to minimize systematic errors associated with the experimental technique. Moreover, all experiments were performed at exposition times to the ozone/air mix of 180 or 360 min, i.e., same experimental conditions that were used in the blank air flow experiments as described before.

Analytical method

The analytical approach is similar to one employed in the previous laboratory and field work where an annular denuder-based sampler was used to determine atmospheric PAH concentrations (Temime-Roussel et al. 2004a). Briefly, the trapped PAHs were extracted immediately after sampling. Because of the high reversibility of the adsorption phenomenon of the organic compounds on polymeric resins, it was possible to extract PAH using liquid–solid extraction. To extract the PAHs from the XAD-4 resin, the denuder tube was completely immersed in cyclohexane at T = 40 °C and sonicated for 30 min. The extract was then filtered (FHLP Millipore filter, pore size of 0.45 mm) to remove the XAD-4 particles extracted from the denuder walls. Upon extraction, 9-methylanthracene was added to the mix as an internal standard. The resulting samples were then rotavapor concentrated to 10 mL at 40 °C and 235 mbar. Field samples were concentrated to 1 mL using a gentle flow of nitrogen gas. The obtained sample volumes were determined gravimetrically. The analyses were carried out using a reverse phase high-performance liquid chromatography (HPLC) equipped with a Supelco Supelcolsil C18 column (25 cm long; 4.6 mm id; pore size = 5 μm), and a Varian fluorescence detector 9075 using an internal standard (9-methylanthracene). A rheodyne 7010 sample injector equipped with a 20-μL loop was used. Elution was made using water and acetonitrile as solvent. The linear gradient sweep began immediately from 40/60 to 100/0 over a sweep time of 30 min. The column was kept at a constant temperature T = 32 °C. The Varian fluorescence detector excitation wavelength used for naphthalene and acenaphthene was 224 nm and the emission wavelength for the two compounds was 330 nm. Similarly, the excitation and emission wavelengths for phenanthrene were 248 and 370 nm, respectively. The internal standard was detected at 364 and 410 nm.

Reagents

The gases used in this study had the following stated minimum purities: N2 (Linde Gaz)—99.999 % and purified air—99.997 % (total hydrocarbon ≤ 0.001 %). Ozone was prepared by allowing purified air to pass through a commercial ozonizer. The reagents used for the analysis were naphthalene crystals (Prolabo), acenaphthene crystals (Prolabo), phenanthrene crystals (Prolabo), acetonitrile (HPLC grade, Acros organics), cyclohexane (HPLC grade, Acros organics), and n-hexane (95 %, HPLC grade, Acros organics).

Results and discussions

Laboratory reactivity study

The reaction of adsorbed naphthalene on the XAD-4 resin within the annular denuder tubes with gaseous ozone can be written as follows:
$$ {\mathrm{Nap}}_{\mathrm{ads}}+{\mathrm{O}}_3\left(\mathrm{g}\right)\to \mathrm{Products} $$
(1)
All experiments were performed under pseudo-first-order conditions with the ozone concentration in excess over naphthalene concentration, [O3] ≫ [naphthalene]. Under these conditions the naphthalene signal was governed by its reactions with O3. Since [O3] was much larger than [naphthalene], observed naphthalene signal profiles followed the following pseudo-first order rate law.
$$ \ln \frac{{\left[{\mathrm{Nap}}_{\mathrm{ads}}\right]}_0}{\left[{\mathrm{Nap}}_{\mathrm{ads}}\right]}=k\left({\mathrm{O}}_3\right)\times t $$
(2)
In Eq. (2), k is the bimolecular rate coefficient for the reaction of adsorbed naphthalene with gaseous ozone, [Napads]0 is the amount of naphthalene trapped in an annular denuder sampler without reaction artifact [Napads]ref, and [Napads]t is the actual amount of naphthalene measured. The above Eq. (2) can be rewritten in the following form:
$$ \frac{1}{t} \ln \frac{{\left[{\mathrm{Nap}}_{\mathrm{ads}}\right]}_0}{\left[{\mathrm{Nap}}_{\mathrm{ads}}\right]}= k\mathit{\prime} $$
(3)
In Eq. (3), k′ = k[O3(g)] and the units are s−1. The bimolecular rate coefficient for the reaction of adsorbed naphthalene with gaseous ozone was obtained from the variation of 1 / t × ln{[Napads]0/[Napads]} with [O3]. The obtained data are listed in Table 2 and shown in Fig. 3 and the figure shows plots of 1 / t × ln{[Napads]0/[Napads]} vs. [O3(g)]. The slope of the solid line shown in Fig. 3 gives the bimolecular rate coefficient for the heterogeneous ozonolysis of naphthalene adsorbed on XAD-4 resin k = (1.01 ± 0.04) × 10−18 cm3 molecule−1 s−1. Uncertainties in the kexp values listed in Table 2 and shown in Fig. 3 are absolute uncertainties and are calculated as shown below.
Table 2

Summary of the kinetic data for the reaction of naphthalene adsorbed on XAD-4 resin with ozone

Exp No.

[Nap] (μg)

[O3] (ppm)

[O3] (molec. cm−3)

[O3]/[Nap]

Flow (slm)

Exposition time (minutes)

Napt/Nap0

1/t × ln{[Napads]0/[Napads]} a (10−6 s−1)

1

5.34

0.064

1.6 × 1012

98

4.4

360

0.98

0.94 ± 2.7

2

10.03

0.129

3.2 × 1012

110

4.4

370

0.95

2.3 ± 2.7

3

13.23

0.123

3.0 × 1012

79

4.4

360

0.93

3.4 ± 2.9

4

4.41

0.127

3.1 × 1012

241

4.4

362

0.95

2.4 ± 2.8

5

4.25

0.179

4.4 × 1012

333

4.4

340

0.93

3.6 ± 3.0

6

10.63

2.876

7.0 × 1013

614

2.4

180

0.47

70 ± 10

7

7.89

3.086

7.6 × 1013

548

1.5

180

0.41

83 ± 11

8

5.53

4.896

1.2 × 1014

1,302

1.5

189

0.26

120 ± 17

9

7.21

0.992

2.4 × 1013

196

1.5

188

0.74

27 ± 6

10

4.53

0.559

1.4 × 1013

174

1.5

180

0.82

18 ± 6

11

4.59

0.205

5.0 × 1012

67

1.5

190

0.93

6.4 ± 4.8

aStated errors are absolute uncertainties in the kexp value. \( \varDelta {k}_{exp}^{\prime }=\frac{1}{t}\cdot \frac{\varDelta x}{x} \)

Fig. 3

Plot of k′ versus [O3] for the reaction of naphthalene with ozone. Experimental conditions: T = 26–30 °C; open circles, flow = 1.5 L min−1, exposure time = 3 h; open square, flow = 2.4 L min−1, exposure time = 3 h; and filled circles, flow = 4.4 L min−1, exposure time = 6 h. Fit is obtained from least squares analysis and the slope gives the bimolecular rate coefficient for the reaction of solid naphthalene adsorbed on XAD-4 resin with ozone, (1.01 ± 0.04) × 10−18 cm3 molecule−1 s−1. Horizontal error bars are 10 % uncertainty in ozone concentration. Vertical error bars (Δkexp) are absolute uncertainties in the kexp value. \( \varDelta {k}_{exp}^{\prime }=\frac{1}{t}\cdot \frac{\varDelta x}{x} \)

$$ d{k}_{exp}^{\prime }=d\left\{\frac{1}{t} \ln \left(\frac{{\left[{\mathrm{Nap}}_{\mathrm{ads}}\right]}_0}{{\left[{\mathrm{Nap}}_{\mathrm{ads}}\right]}_t}\right)\right\} $$
(4)
If \( x=\frac{{\left[{\mathrm{Nap}}_{\mathrm{ads}}\right]}_0}{{\left[{\mathrm{Nap}}_{\mathrm{ads}}\right]}_t} \), then Eq. (4) may be rewritten in the following form:
$$ d{k}_{exp}^{\prime }=d\left(\frac{1}{t} \ln x\right)=\frac{1}{t}\frac{\partial \ln x}{\mathit{\partial x}} dx=\frac{1}{t}\frac{ dx}{x} $$
(5)
Assuming that the variations are very small, then
$$ \varDelta {k}_{exp}^{\prime }=\frac{1}{t}\frac{\varDelta x}{x}. $$
(6)

This value is five times greater than the current recommended value for the homogeneous gas phase rate coefficient of the reaction of naphthalene with ozone (2.0 × 10−19 cm3 molecule−1 s−1) (Atkinson et al. 1984). As a result, it seems that the heterogeneous reaction of naphthalene with ozone is favored compared to the homogeneous reaction. At this point, we cannot give any convincing explanation regarding a possible mechanism (catalytic for example) and to explain the observed reactivity enhancement. However, it is known that the heterogeneous kinetics of PAHs with ozone can be enhanced by the nature of the particles (Alebic-Juretic et al. 1990; Perraudin et al. 2005; Perraudin et al. 2007a, 2007b; Perraudin et al. 2007a, 2007b; Pflieger et al. 2009a, 2009b; Pflieger et al. 2009a, 2009b; Net et al. 2010). As a result, this phenomenon seems common to different compounds adsorbed on different surfaces. Nevertheless, these previous studies investigating substrate-dependent heterogeneous kinetics using different particulate matrices suggest that the importance of the enhancement depends strongly on the solid support and/or the individual compounds under study. It is possible that the heterogeneous reactivity depends not only on the nature of the particles, but also on the chemical properties of the semi-volatile organic compound (SVOC) studied. Thus, the measured rate constant in this study may be specific to naphthalene adsorbed to XAD-4 substrate.

The obtained rate coefficient was then used to study the reaction artifacts due to reactivity with ozone during field sampling using an annular denuder-type sampler.

Field studies

In addition to laboratory experiments, six field campaigns (three in the summer and three in winter) carried out in three different locations in southeastern France (Chamonix valley, Maurienne valley and Marseille) from 2003 to 2005 were revisited and the obtained results reanalyzed to evaluate the importance of the ozone degradation artifact. The goal of this reevaluation was to compare the uncertainties resulting from the PAH sampling method itself namely, analytical uncertainties and blank contamination among others, and the uncertainty resulting from artifact reactivity. The experimental details of the six field campaigns and the obtained results are described elsewhere (Goriaux et al. 2006). In the present work, the newly determined rate coefficient for the reaction of solid naphthalene with gaseous ozone is used to study the reaction artifacts.

The ratio [Napads]0/[Napads]t was calculated by doing some simple algebra and rewriting Eq. (2) listed above in the following form:
$$ \frac{{\left[{\mathrm{Nap}}_{\mathrm{ads}}\right]}_0}{\left[{\mathrm{Nap}}_{\mathrm{ads}}\right]}= \exp \left(k\left[{\mathrm{O}}_{3\left(\mathrm{g}\right)}\right]t\right) $$
(7)
Using the above Eq. (7), the reactivity of naphthalene toward ozone was then calculated by taking the difference in the initial naphthalene concentration, [Napads]0, and its final concentration at the end of sampling time, [Napads]t.
$$ \frac{{\left[{\mathrm{Nap}}_{\mathrm{ads}}\right]}_0-{\left[{\mathrm{Nap}}_{\mathrm{ads}}\right]}_t}{{\left[{\mathrm{Nap}}_{\mathrm{ads}}\right]}_0}=\frac{ \exp \left(k\times \left[{O}_3\right]\times t\right)-1}{ \exp \left(k\times \left[{O}_3\right]\times t\right)} $$
(8)
Certain gas phase PAHs are also likely to be collected in the annular denuder-based sampler namely, acenaphthene and phenanthrene. As a result, these two compounds were measured in addition to naphthalene during the six PAH field campaigns. The homogeneous gas-phase rate constants of acenaphthene and phenanthrene toward an attack by ozone are 5 × 10−19 cm3 molecule−1 s−1 (Atkinson et al. 1988) and 4 × 10−19 cm3 molecule−1 s−1 (Zielinska 2005), respectively (see Table 3). These two values are similar to the literature value for the homogeneous gas-phase reaction rate constant of naphthalene with ozone (3 × 10−19 cm3 molecule−1 s−1) (Atkinson et al. 1984; Atkinson et al. 1986). Consequently, it may be assumed that the degradation kinetics of adsorbed acenaphthene and phenanthrene on XAD-4 resin in an annular denuder-type sampler may be similar to that of naphthalene.
Table 3

Comparison of the rate coefficients for the homogeneous gas phase reactions of ozone with selected PAHs and the obtained rate coefficient for the reaction of solid naphthalene with ozone

Compound

ko3 (10−19 cm3 molecule−1 s−1)

Reference

Acenaphthene

4

Atkinson et al. 1988

Phenanthrene

5

Zielinska 2005

Naphthalene

2.01

Atkinson et al. 1984

 

3.01

Atkinson et al. 1986

Naphthalene adsorbed on a XAD-4 resin

10.1 ± 2

This work

Ozone artifact errors are calculated using Eq. (8). The sampling uncertainties are based on the complete assessment of the analytical protocol, that is, the PAH/solvent extraction, the PAH/solvent evaporation, and the uncertainties in the HPLC measurements. The calculated ozone artifact errors during the field campaigns range between 0.1 and 6.9 %. The importance of the artifact is proportional to the ozone concentration because the sampling times are more or less equivalent for all the field campaign data. As a result, the higher artifact (6.9 %) is obtained for the higher ozone concentration (131 μg m−3 on 27th Jun 2003). These ozone artifacts have to be compared to the sampling errors. The two data sets (sampling errors and ozone artifact errors) are shown in Fig. 4. As shown in this figure, the observed uncertainty associated with field measurements is much higher than the uncertainty generated by the artifact reactions with ozone. As a result, based on the observed field measurement uncertainty results, it can be stated that ozone artifact is weak or sometimes negligible compared to the observed field campaign ozone concentrations.
Fig. 4

Comparison of the measured relative errors during a field campaign and the calculated ozone induced degradation for naphthalene. [O3]field campaign < 80 ppbv (Goriaux et al. 2006). Solid line is y = x

Conclusions

The kinetic rate coefficient for the reaction of adsorbed naphthalene on XAD-4 resin in an annular denuder-type sampler with ozone is reported to be (10.1 ± 0.4) × 10−19 cm3 molecule−1 s−1 (error is 2σ, precision only). The measured kinetic rate coefficient for the reactions of solid naphthalene with ozone is five times greater than the current recommended value for a homogeneous gas phase-rate coefficient of the reaction of naphthalene with ozone (2.0 × 10−19 cm3 molecule−1 s−1) (Atkinson et al. 1984; Atkinson et al. 1986). At this point, we cannot give any convincing explanation regarding a possible mechanism and explain the observed enhanced reactivity of adsorbed naphthalene toward ozone.

Given typical tropospheric ozone mixing ratios and the experimental sampling times, the artifact resulting from the heterogeneous ozonolysis of naphthalene adsorbed on XAD-4 resin seems negligible compared to the overall sampling error.

Acknowledgments

Financial support by the Provence-Alpes-Côte d’Azur (PACA) Region and the Agence de l’Environnement et de la Maitrise de l’Energie (ADEME) is gratefully acknowledged.

Copyright information

© Springer-Verlag Berlin Heidelberg 2014

Authors and Affiliations

  • Mathieu Goriaux
    • 1
  • Maryline Pflieger
    • 1
    • 2
  • Anne Monod
    • 1
  • Sasho Gligorovski
    • 1
  • Rafal S. Strekowski
    • 1
  • Henri Wortham
    • 1
  1. 1.Aix-Marseille Université, CNRS, LCE, FRE 3416MarseilleFrance
  2. 2.Laboratory for Analytical ChemistryNational Institute of ChemistryLjubljanaSlovenia

Personalised recommendations