Estimation of Polycyclic Aromatic Hydrocarbon Variability in Air Using

Apr 6, 2015 - Organic films and leaves provide a medium into which organic contaminants, such as polycyclic aromatic hydrocarbons (PAHs), can accumula...
0 downloads 7 Views 736KB Size
Download PDF
Article pubs.acs.org/est

Estimation of Polycyclic Aromatic Hydrocarbon Variability in Air Using High Volume, Film, and Vegetation as Samplers Elisa Terzaghi,†,‡ Marco Scacchi,† Bruno Cerabolini,‡ Kevin C. Jones,§ and Antonio Di Guardo*,† †

Department of Science and High Technology, University of Insubria, Via Valleggio 11, 22100 Como, Italy Department of Theoretical and Applied Sciences, University of Insubria, Via J. H. Dunant 3, 21100 Varese, Italy § Lancaster Environment Centre, Lancaster University, Lancaster LA1 4YQ, United Kingdom ‡

S Supporting Information *

ABSTRACT: Organic films and leaves provide a medium into which organic contaminants, such as polycyclic aromatic hydrocarbons (PAHs), can accumulate, resulting in a useful passive air sampler. In the present work, the temporal variability (weekly) in PAH concentrations and the fingerprint of films developed on window surfaces were investigated. Moreover, films and leaves of two tree species (Acer pseudoplatanus and Cornus mas) collected at the same time were used to derive PAH air concentrations and investigate their short-term variability. In general, the most abundant chemicals found in films were phenanthrene and pyrene (22%), followed by perylene (21%) and fluoranthene (16%), but the fingerprint (in contrast to leaves and air) changed over time. Leaf derived air concentrations were within a factor of 2 to 9 from measured values, while air concentrations back-calculated from films were within a factor of 2 to 53. This happened because predicted air concentrations using films and vegetation samplers (especially for low KOA chemicals) generally reflect only the last few hours (due to the fast equilibrium) of the weekly integrated samples obtained employing the high-volume sampler. This means that films and leaves can be usefully employed for predicting the short-term variability of low KOA organic contaminant air concentrations.

1. INTRODUCTION

mass, could greatly vary from urban to rural location and differ depending on seasonal and diurnal variation.12,14,20 The capability of organic films to collect contaminants was compared to that of leaf cuticles:19 both films and leaves, thanks to their physical and chemical properties, provide a medium into which gas-phase and particle-phase organic compounds can partition and be trapped. Leaves can accumulate organic contaminants with different degrees of efficiency depending on plant species,21 and they can be used to investigate organic contaminant air concentration levels. However, leaves have different disadvantages that limit their usage, such as variability in their properties, depending on species and time in their growth cycle.9 The exchange of low molecular weight organic contaminants between air and synthetic polymer films characterized by a thickness of about 100 nm was shown to be very rapid since chemicals in film reach equilibrium within a short time (hours to days).3 This is also applicable to urban films, which have the ability to rapidly respond to organic contaminant air concentration changes:20 they could release contaminants

Polycyclic aromatic hydrocarbons (PAHs) concentrations in air are generally monitored using high-volume samplers. This method is powerful but expensive, especially when data on short time scales are needed, limiting spatial and temporal coverage. An alternative and cheaper approach may be the use of passive air samplers such as polyurethane foams (PUFs),1,2 polymer-coated glass (POGs),3,4 ethylene vinyl acetate (EVA) samplers,5 sorbent-impregnated PUF (SIP),6 or matrices that could act as passive air samplers such as plant leaves,7,8 artificial plant leaves9,10 and organic films which grow on most impervious surfaces.11−13 Our urban and suburban areas are characterized by the presence of different impervious surfaces such as windows, walls, and roadways. Different studies11,12,14−18 demonstrated that window glass and other impervious surfaces accumulate a variety of organic compounds. This happens because an organic film develops on such impervious surfaces, capturing PAHs and other organic pollutants, such as polychlorinated biphenyls (PCBs) and n-alkanes19 thanks to both gaseous and particle deposition from air.11,14,19 This organic film can range from 11 to 1000 nm in thickness and from 5% to 20% in organic matter (OM) content11,12,14,20 and generally develops linearly with time.20 However, film growth rate, as well as thickness and © XXXX American Chemical Society

Received: November 21, 2014 Revised: April 3, 2015 Accepted: April 6, 2015

A

DOI: 10.1021/es5056929 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

2. MATERIALS AND METHODS 2.1. Sampling Site. The sampling site was located adjacent to the center of Como (a highly urbanized town of about 85 000 inhabitants in Lombardy region, in the north of Italy). Due to its conformation and in agreement with the European Environmental Agency (EEA) air quality monitoring and information network,32 the location can be classified as an urban background site due to its distance from the nearest street of about 150 m. Organic film samples were obtained from the windows of the Local Health Unit (ASL Como) building placed in proximity of a small broadleaf wood where leaf samples were collected. The high-volume air sampler was placed in the adjacent clearing area which consisted of a meadow with sporadic trees. More details about leaf and air sampling area are reported in Terzaghi et al.27 All sampling heights were between 1.5 and 2.5 m. 2.2. Sampling. The sampling campaign was performed weekly from March to June 2007, except for the last two samples which were collected after a longer exposure period. For window organic films, twenty-two samples were collected from window surfaces by vigorously wiping with four dichloromethane (DCM) wetted Kleenex tissues. Kleenex tissues were not precleaned since blank contamination was not significant. Windows were sampled up to 10 cm of the window’s edge to minimize contamination from exterior caulking, paints, etc.11 During each sampling event, two different windows were sampled and the respective tissues were stored in glass jars and then in a freezer at a temperature of −30 °C until extraction. A surface of about 1 m2 was sampled for each window. Before the first sampling, the chosen window glasses were cleaned using DCM, giving a sample that reflects the previous contamination of an unknown period and which was not considered when elaborating data. Window samples were collected approximately between h 13:00 and h 15:00 depending on the day, between March 15 and June 20. The sampled windows were partially sheltered and therefore little exposed to rain events and, to a lesser degree, to sunlight. Leaf and air samples were collected as described in Terzaghi et al.27 Briefly, air sampling was performed with a high-volume sampler, separately analyzing gas (PUFs) and particle (GFF) phase. Samples were continuously collected over weekly intervals (from March 29 to June 7) yielding 500−850 m3 samples. Leaves (about 100 g) of cornel (Cornus mas) and maple (Acer pseudoplatanus) were weekly sampled from March 29 to June 7 between h 11:00 and h 13:00 depending on the day. Maple leaves were not available during the first two sampling weeks because bud burst had not occurred yet. A total of 19 leaf samples and 10 air samples were collected. Table S1, Supporting Information, reports sampling times for each matrix. The leaf concentrations expressed on a fresh weight were converted to a dry weight basis assuming water content of 75% and normalized employing the specific leaf area (SLA) reported in Terzaghi et al.27 in order to obtained values expressed as ng/m2 comparable with those of film. Film concentrations are reported as the average of the two sampled windows. 2.3. PAH Analysis. PAH analysis was performed for acenaphthylene (ACY), acenaphthene (ACE), fluorene (FLUO), phenanthrene (PHE), anthracene (ANTH), fluoranthene (FLUOT), pyrene (PYR), banzo[a]anthracene (B[a]ANTH), chrysene (CHR), benzo[e]acephenanthrylene (common name benzo[b]fluoranthene) (B[e]ACEP), benzo-

during periods of reduced air concentrations and accumulate them during periods of elevated air concentrations, therefore acting as a source or a sink. Plant leaves were also shown to behave as a dynamic compartment,22,23 contributing to the diurnal variation of the concentration of organic contaminants, which deposit or volatilize from their surfaces in response to changes in environmental conditions. As for film, the air−leaf surface exchange is believed to be the major short-term source and sink of many persistent organic compounds.22,23 Since films and leaves act as passive air samplers, PAH air concentration (gas phase) can be estimated if the partition coefficients between the film and air (KF‑A) and between plant and air (KP‑A) and the concentrations in film and plant are known, assuming that the two samplers are in equilibrium with air. Different relationships were developed in order to estimate film/air partition coefficients24,25 and plant/air partition coefficients21 and could be adopted in the calculation; however, film and leaves concentrations should be determined simultaneously if the comparison between the two types of samplers has to be investigated. Different studies have been conducted in the laboratory with synthetic polymer films such as EVA samplers and POGs, to investigate their use as passive air samplers over short time periods (hours/days).3,5 Many studies have been performed with “natural” films in field conditions, with the aim of measuring the accumulation of different organic chemicals (PAHs, PCBs, PBDEs) throughout the film growth stage (generally months)17,26 and to derive air concentrations13,15,16 but without comparing them to simultaneous air measurements. The objective of this study was to assess the temporal variability in PAH concentrations and fingerprint of films developed on window surfaces weekly collected from March to June, 2007 in an urban area in Northern Italy and compare film concentrations with those of air (active sampling) and leaves sampled simultaneously at the same site and reported in a companion paper.27 Moreover, PAH gas phase air concentrations were estimated from those of films and leaves and compared with those measured by the high-volume sampler. Since both films and leaves are considered dynamic exchangers of organic contaminants with air,22,23,25,28 their concentrations should be representative of atmospheric conditions at the time of sampling, allowing one to predict the short-term variability of PAHs in air. Obtaining this kind of data is of particular interest since they could be employed to define a more realistic exposure scenario29 to protect sensitive targets when performing environmental risk assessment. Additionally, air concentrations are generally available at a low temporal resolution (e.g., daily/weekly/monthly averaged values) leading to a possible misrepresentation of the extent of change. This has been reported by several authors22,23,30,31 who measured air concentration variation at short time intervals of different chemicals showing that, although concentrations follow a daily pattern, they are influenced by the hourly variability of different parameters such as temperature, planet boundary layer height (PBL), and OH radical presence. Further calculations were performed to define the time to reach equilibrium required by film and leaves (considering PAHs of different physical and chemical properties), both in “steady” and “unsteady” state environmental conditions, for evaluating the most suitable model to be employed to predict air concentrations. B

DOI: 10.1021/es5056929 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

most commonly detected in both film and leaf samples during the sampling period. A thickness of about 60 nm equivalent to 1 week of exposure considering a film growth rate of 8.5 nm/ d26 (average value for a 7 day period) was used for film, while leaf thickness was set to 162 μm for cornel and 125 μm for maple as reported by Castro-Diez et al.41 for Cornus sanguinea and Acer platanoides. An air side mass transfer coefficient of 15.4 and 23 m/h for film and leaves, respectively, was assumed as reported in Diamond et al.39 These values refer to a wind speed of 0.55 m/s and a film and leaf length of 7 cm. KPS‑A was calculated at 25 °C according to eq S4 and eqs S6 and S7, Supporting Information, for film (KF‑A) and leaf (KP‑A), respectively. KP‑A expressed on a mass/volume basis was converted to a volume/volume basis considering a dry leaf density of 214 000 g/m3 for cornel and 291 000 g/m3 for maple as reported by Castro-Diez et al.41 for Cornus sanguinea and Acer platanoides. Unsteady State Environmental Conditions. Some of the parameters involved in the sampler uptake rate equation could vary over time, being dependent on meteorological parameters. For example, KPS‑A depends on air temperature,16,42 and kA varies with wind speeds;25,43 film thickness could change during the whole exposure period being influenced by precipitation events20 and by dry particle deposition,25 while leaf thickness varies diurnally and seasonally.44 In order to evaluate the importance of such dynamic features in influencing the time to reach equilibrium during each sampling period, the parameters kU, t25, and t95 were recalculated for each period of exposure (Table S1, Supporting Information), updating the KPS‑A with the average air temperature of the hours of sampling and considering the actual film thickness (see Text S3, Supporting Information, for more details). Moreover, kA for both film and leaves was calculated considering wind speed of the hours of sampling and the correct passive sampler surface length according to Nobel45 (see Text S2, Supporting Information, for more details). Tables S4−S7, Supporting Information, summarize the values used to estimate the t25 and t95 for PHE, FLUOT, PYR, and CHR, respectively, for each exposure period. Here, t25 and t95 were compared to the actual exposure time of the passive air sampler, ranging from 6 d (144 h) to 24 d (576 h) for film and from 7 d (168 h) to more than two months (1680 h) for leaves. As explained in Terzaghi et al.,27 during each week, a mixture of older and fresh leaves characterized by different exposure time might have been collected, but here, it was assumed that all leaves of each sampling period had the same exposure time (the maximum). 2.6. Prediction of Air Concentrations. In order to calculate PAH gas phase air concentration based on the concentrations found in film and leaves, two equations were employed. If the two samplers had reached equilibrium, air concentrations were estimated using the film/air and plant/air partition coefficients as follows:

[a]pyrene (B[a]PYR), perylene (PER), benzo[ghi]perylene (B[ghi)PER), dibenzo[ah]anthracene (DB[ah]ANTH), indeno[cd]pyrene (I[cd]PYR), and coronene (COR), following the analytical procedure reported in Terzaghi et al.33 and summarized in Text S1, Supporting Information. 2.4. Meteorological Parameters. Measurements of air temperatures were carried out with a Testo temperature logger (Testo AG, Lenzkirch, Germany, mod. 174). The logger was sheltered to protect it from direct sunlight and precipitation. Temperature was measured at 1 h intervals for the entire sampling period at the clearing site. Other meteorological parameters such as precipitation, wind speed, and solar radiation were obtained from the Regional Environmental Protection Agency34 for Como city. Hourly PBL heights were obtained as described in Morselli et al.35 2.5. Evaluation of Passive Sampler Equilibrium Conditions. The estimation of atmospheric concentrations from passive sampling data is usually based on kinetic or equilibrium models, assuming constant air concentrations.36,37 To choose the appropriate model, it is relevant to understand if the chemical of interest, at a certain time of uptake, is in the linear, curvilinear, or equilibrium stages of the uptake curve for a particular sampler. To verify if equilibrium conditions were satisfied at the time of sampling, the uptake rate of the two samplers (film and leaves) was calculated as reported in the literature5,7,38 for some passive air samplers such as SPMD (semipermeable membrane device), PUFs, POGs, and EVA:

kU =

1 kA δ KPS‐A

(1)

where kU is the uptake rate (1/h), kA is the air-side mass transfer coefficient (m/h), δ is the passive sampler thickness (m), and KPS‑A is the passive sampler-air partition coefficient (dimensionless). Please note that, when using this equation for leaves, kU has to be multiplied for 2 to consider both leaf sides. This implies that high KPS‑A and δ, together with low kA, will result in a longer time to reach equilibrium. As in eq 1, the transfer of the chemical to the passive air sampler is assumed to be air-side controlled only and a one compartment model for both film and leaves was applied. The time to reach the 25% (t25) and 95% (t95) of equilibrium conditions was derived from kU as follows: t 95 =

ln(0.05) kU

(2)

t 25 =

ln(0.75) kU

(3)

When exposure time (texp) of the passive sampler was higher than t95, equilibrium conditions applied; when t25 < texp < t95, the chemical in the passive air sampler could be considered in the curvilinear phase, while if texp < t25 it was in the linear phase. Steady State Environmental Conditions. In order to evaluate if film and leaves had reached equilibrium after 1 week of exposure (texp = 168 h), t95 and t25 were first calculated using “static” and “not compound specific” air side mass transfer coefficients already adopted in the literature25,39,40 and ambient temperature passive sampler/air partition coefficients. Table S3, Supporting Information, summarizes parameters values used to estimate the t25 and t95 for four chemicals (PHE, FLUOT, PYR, and CHR) of different physical and chemical properties. These compounds were selected since they were

CA =

CF KF‐A

(4)

CA =

CP KP‐A

(5)

where CA is the gaseous PAH air concentration (ng/m3), CF is the PAH concentration in the window film (ng/g), KF‑A is the film/air partition coefficient (m3/g), CP is the PAH plant concentration (ng/g dw), and KP‑A is the plant/air partition C

DOI: 10.1021/es5056929 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

Figure 1. (a) PAH fingerprint in air (A), film (F), cornel (C), and maple (M) leaves; (b) PAH trend (on a logarithmic scale) in air (light blue), film (gray), cornel (light green), and maple (dark green) leaves.

coefficient (m/h), δ is the passive sampler (plant or film) thickness (m), and texp is the exposure time (h). As mentioned above, air concentrations were estimated for PHE, FLUOT, PYR, and CHR since they were mostly detected in both film and leaf samples during the sampling period.

coefficient (m3/g). KF‑A and KP‑A were calculated using the log KOA of the chemicals according to two existing approaches,21,25 and they were temperature corrected as reported in Text S3, Supporting Information. Where equilibrium was not reached within the exposure time of the passive sampler, the following equations5 (which describe the complete uptake profile) were employed to predict air concentrations: CA =

CF

(

KF‐A × 1 − exp − CA =

3. RESULTS AND DISCUSSION 3.1. Comparison of Air, Film, and Leaf Fingerprints. Figure 1a shows PAH fingerprints of air (A), film (F), cornel (C), and maple (M) leaves. For comparison, only chemicals found in all the matrices were reported (ACY, ACE, B[a]PYR, I[cd]PYR, DB[ah]ANTH, B[ghi]PER, and COR are not shown). Air and leaf concentrations can be found in Terzaghi et al.,27 while window film concentrations can be found in Table S10, Supporting Information. Data were grouped according to the periods (P) reported in Table S1, Supporting Information. Moreover, for display purposes, film concentrations were normalized on a weekly basis because exposure period varied from 6 d (P4) to 24 d (P8). The most abundant chemicals found in film samples were (average %: min−max %): PHE (22%: 12%−39%), PYR (22%:

(

1 δ

×

kA KF‐A

)×t )

1 δ

×

kA KP‐A

)×t )

exp

(6)

CP

(

KP‐A × 1 − exp −

(

exp

(7) 3

where CA is the gaseous PAH air concentration (ng/m ), CF is the PAH concentration in the window film (ng/g), KF‑A is the film/air partition coefficient (dimensionless), CP is the PAH plant concentration (ng/g dw), KP‑A is the plant/air partition coefficient (dimensionless), kA is the air-side mass transfer D

DOI: 10.1021/es5056929 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

Figure 2. Time required by the three samplers to reach 95% of equilibrium.

particles” still has to be evaluated in comparison to film, they are probably more susceptible to particle erosion caused by wind and rain. PER was found in Pinus pinea needles of 2 years collected near the sampling site of the present study in July 2007 (data not shown), probably due to the encapsulation within the cuticle of the captured particles and their associated chemicals that occur after a long period of exposure as demonstrated by Terzaghi et al.33 PHE was also the most abundant chemical in air (61%: 42%−69%), followed by FLUOT (12%: 7%−14%) and PYR (9%: 6%−13%). Also the air fingerprint did not seem to change over time, with the exception of P9. Figure 1b shows the total PAH (sum of FLUO, PHE, ANTH, FLUOT, PYR, b(a)ANTH, CHR, B(e)ACEP, and PER) trend of air, film, cornel, and maple leaves. PAH film concentrations showed a decreasing trend with time. PAH film concentrations showed a decreasing trend with time. This behavior is in line with that of air concentrations, also showing a decreasing trend with time (Figure S2, Supporting Information), due to the variability of meteorological (mainly temperature and PBL height) and ecological (specific leaf area, leaf area index, and leaf biomass) parameters as shown by the principal component analysis (PCA) reported in Terzaghi et al.27 In brief, both sets of factors together with the emission variations did intervene in determining the variability of air concentrations with time. In particular, a marked increase (about 50%) in the overall surface area of leaves (LAI) from March 15 to April 5 contributed to a sudden decrease in air concentrations during P3. Wind and rainfall were shown not to be important factors ruling air variability. In the present study, total PAH window concentrations ranged between 22 (P9) and 127 (P1) ng/m2 week while the total PAH amount for the whole sampling period (from March to June) was 550 ng/m2. These values are within the same order of magnitude of those reported for two urban areas in Stockholm (79−467 ng/m2)16 and Hong Kong (210−630 ng/ m2),17 one suburban area in Baltimore (78 ng/m2),14 and one rural area in Toronto (60−600 ng/m2).12 Cornel and maple leaves accumulated more PAHs than film when considering 1 week of exposure (P2 and P4 for cornel and maple, respectively). The higher amount in leaves (a factor of about 20 to 40) is consistent with their higher volume or thickness (see Section 3.2) and could depend on their sorbing capacity, e.g., composition of cuticle waxes.

16%−29%), PER (21%: 13%−40%), and FLUOT (16%: 8%− 22%). PHE, FLUOT, and PYR, but not PER, predominated also in films of other studies.12,14,17,18 As reported in Terzaghi et al.,27 the most abundant PAH found in cornel and maple leaves was (average %: min−max %) PHE (49%: 28%−65% in cornel; 35%: 16%−47% in maple), followed by PYR (20%: 13%−25% in cornel; 29%: 22%−35% in maple), FLUOT (17%: 12%−21% in cornel; 28%: 19%− 32% in maple), and CHRY (12%: 7%−17% in cornel; 7%: 7%− 7% in maple). The other PAHs were never or occasionally detected in leaves due to their lower air concentrations and/or their log KOA values (lower than 7 or higher than 11); in fact, the log KOA range of 7 to 11 was shown as the one in which high deposition to vegetation occurs (forest filter effect46). While leaf fingerprints did not seem to change over time, with the exception of P9, film fingerprints showed a different behavior. This may be attributed to the fact that (1) film can respond faster than leaves to a change in air emission due to its quick response rate (see Section 3.2); (2) sampled leaves were generally less exposed to sunlight and precipitation and therefore less susceptible to photodegradation and washoff of accumulated chemicals than film since they were protected by the upper canopy foliage.47 Moreover, film accumulated PER, which was not found in leaf samples, with the exception of P9. This chemical has a log KOA of 11.7, and as shown by Terzaghi et al.,27 it predominated in air particulate phase (representing the 13%: 5%−36% of the total particulate PAHs) rather than in the gaseous one (representing the 1.8%: 0.3%−3.6% of the total gaseous PAHs). This means that PER may have reached and accumulated on window organic film thanks to particle deposition. Liu et al.14 reported that the organic phase of the film develops rapidly (hours to days) after which particles accumulate, presumably due to the increased dry deposition flux caused by the sticky nature of the film. The sticky nature of the film reduces particle rebound and resuspension. During P8 and P9 (about 25 days of exposure), PER represented 44% and 33% of total PAHs, while during the previously weeks (about 7 days of exposure), it ranged between 13% and 21%. This is probably due to the fact that film represents both gas phase constituents that exchange between the film and air and particle-sorbed constituents that accumulate over time.12 This increase of PER in P8 and P9 may also occur because the sampling time was longer. Leaves can also trap particles and their associated PAHs,33 but although the leaf “stickiness to E

DOI: 10.1021/es5056929 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

Figure 3. Comparison between predicted and measured air concentrations for PHE, FLUOT, PYR, and CHR. Asterisk (∗) may mean: air not sampled, leaves not sampled because bud burst had not occurred yet, or contaminant not detected.

3.2. Time for Reaching Equilibrium in Film and Leaves. Steady State Environmental Condition Calculation. Figure 2 shows t95 values of PHE, FLUOT, PYR and CHR for leaves and film calculated employing “static” and “not compound specific” kA and ambient temperature KPS‑A. kU, t25, and texp are reported in Table S11, Supporting Information. For PHE, FLUOT, and PYR, texp is higher than t95; this means that all these compounds reach equilibrium in 1 week. On the contrary, CHR remains in the curvilinear phase (with the exception of cornel which approaches equilibrium at the end of the week). More specifically, film is generally faster than leaves in reaching equilibrium (due two its lower thickness) in the case of PHE (