Article pubs.acs.org/est
Temporal Variations of Cyclic and Linear Volatile Methylsiloxanes in the Atmosphere Using Passive Samplers and High-Volume Air Samplers Lutz Ahrens,*,†,‡ Tom Harner,*,† and Mahiba Shoeib† †
Environment Canada, Air Quality Processes Research Section, Toronto, Ontario Canada, M3H 5T4 Swedish University of Agricultural Sciences (SLU), Department of Aquatic Sciences and Assessment, SE-750 07 Uppsala, Sweden
‡
S Supporting Information *
ABSTRACT: Cyclic and linear volatile methylsiloxanes (cVMSs and lVMSs, respectively) were measured in ambient air over a period of over one year in Toronto, Canada. Air samples were collected using passive air samplers (PAS) consisting of sorbent-impregnated polyurethane foam (SIP) disks in parallel with high volume active air samplers (HVAAS). The average difference between the SIP-PAS derived concentrations in air for the individual VMSs and those measured using HV-AAS was within a factor of 2. The air concentrations (HV-AAS) ranged 22−351 ng m−3 and 1.3−15 ng m−3 for ΣcVMSs (D3, D4, D5, D6) and ΣlVMSs (L3, L4, L5), respectively, with decamethylcyclopentasiloxane (D5) as the dominant compound (∼75% of the ΣVMSs). Air masses arriving from north to northwest (i.e., less populated areas) were significantly less contaminated with VMSs compared to air arriving from the south that are impacted by major urban and industrial areas in Canada and the U.S. (p < 0.05). In addition, air concentrations of ΣcVMSs were lower during major snowfall events (on average, 73 ng m−3) in comparison to the other sampling periods (121 ng m−3). Ambient temperature had a small influence on the seasonal trend of VMS concentrations in air, except for dodecamethylcyclohexasiloxane (D6), which was positively correlated with the ambient temperature (p < 0.001).
■
INTRODUCTION Cyclic and linear volatile methylsiloxanes (VMSs) are widely used in personal care products and in industrial applications such as solvents and coatings.1 Cyclic VMSs (cVMSs) are high production volume (HPV) chemicals and have received increasing public attention due to their persistence, bioaccumulation potential, and possible adverse effects on mammals.2−9 VMSs have a low water solubility10 and a large fraction of VMSs are released into the atmosphere during the production, use and recycling/disposal processes.11 Landfills and sewage treatment plant (STP) emissions have been identified as an important source for cVMSs and linear VMSs (lVMSs) in the atmosphere.12 The atmospheric lifetime of cVMSs range between 10 and 30 days depending on the OH radical concentration.13 VMSs have been detected globally in the atmosphere even in remote regions such as the Arctic which demonstrates their potential for long-range transport.12,14−21 There are few studies that have investigated the occurrence and fate of VMSs in the atmosphere.14−16,18−23 Analytically, the analysis of VMSs in the atmosphere is challenging due to potential losses of target analytes during sampling, sample preparation, and storage as well as the potential for contamination of the sample during trace analysis.24 A variety of sampling techniques have been used for measuring VMSs in © 2014 American Chemical Society
the atmosphere including (i) active air samplers (AAS) using ENV+ resin (hydroxylated polystyrene−divinylbenzene copolymer) as the sorbent,16,17 (ii) passive air sampler (PAS) using sorbent impregnated polystyrene−divinylbenzene copolymeric resin (XAD) polyurethane foam (PUF) disks (SIP-PAS)12 and XAD-PAS,18 and (iii) direct analysis using atmospheric pressure chemical ionization-tandem mass spectrometry (APCI-MS/ MS).19 AAS and PAS techniques have been applied for monitoring of VMSs in outdoor environments,14−16,18−23 whereas the direct analysis of VMSs using APCI-MS/MS is limited due to detection limits of 4−6 μg m−3 which are higher than typical ambient air concentrations.19 Nevertheless, an evaluation of the different sampling techniques (AAS vs PAS) for assessing VMSs in the atmosphere is lacking. A reliable and robust method for the routine determination of these potentially toxic chemicals is needed for long-term atmospheric monitoring programs. The aim of this study was to evaluate the suitability of AAS, PUF−PAS, and SIP-PAS for assessing VMSs in the Received: Revised: Accepted: Published: 9374
April 28, 2014 July 23, 2014 July 29, 2014 July 29, 2014 dx.doi.org/10.1021/es502081j | Environ. Sci. Technol. 2014, 48, 9374−9381
Environmental Science & Technology
Article
series for the HV-AAS and analyzing the first and second set separately (n = 3). All samples were stored at −20 °C in the dark until extraction within 4 weeks. It has been shown that the VMS concentrations can change during the storage time.16,22 For example, the loss of 13C5-D5 ranged between 0.5% d−1 and 1% d−1 for spiked ENV+ cartridges stored at −18 °C.16 Details of the sampling, dates, air volume, and meteorological data are presented in Tables S2 and S3 in the SI. Sample Extraction and Instrumental Analysis. The extraction and instrumental analysis is based on the methods described elsewhere.12,25 Briefly, PUF/XAD-2 sandwiches, GFFs, SIPs, and PUFs were spiked with 5 μg absolute of an IS mixture containing of each of the 13C4-D4, 13C5-D5, and 13C6D6. The PUF/XAD-2 sandwiches were Soxhlet extracted with petroleum ether/acetone (85/15, v/v) for ∼6 h. The GFFs were extracted three times with dichloromethane by sonication. The SIPs and PUFs were extracted by using a pressurized liquid extraction (PLE) system (ASE 350, Accelerated Solvent Extraction System from Dionex Corporation, Sunnyvale, CA, U.S.). The extraction was carried out using petroleum ether/ acetone (83/17, v/v; 2 cycles) using the PLE conditions as follows: 100 °C, 5 min static cycle with a 100% flush and 240 s purge. All extracts were concentrated by rotary evaporation followed by gentle nitrogen blow-down to 0.5 mL using isooctane as a keeper for the extracts. Prior to injection, 0.1 μg absolute (10 μL of 10 ng μL−1) of mirex was added to the final extract as InjS. The separation and detection of the VMSs was performed using gas chromatography−mass spectrometry (Agilent 5975C; Agilent Technologies, Palo Alto, CA, U.S.) (GC/MS) in selective ion monitoring (SIM) mode using electron ioniziation (EI). Aliquots of 0.5 μL were injected on a DB-5 column (60 m, 0.25 mm inner diameter, 0.25 μm film, J&W Scientific, Folsom, CA, U.S.). The isotope dilution method was used for quantification, which is based on the ratio of the peak-areas of the target analyte to the IS (for details see Tables S4 and text in the SI).
atmosphere. The specific aims of this study include (i) to characterize SIP-PAS and PUF−PAS for selected cVMSs and lVMSs based on field calibration against high volume active air samplers (HV-AAS), (ii) to compare SIP-PAS vs HV-AAS for measuring VMSs in the atmosphere, and (iii) to evaluate seasonal trends and the influence of meteorological conditions on air concentrations of VMSs over one year. The results of this study will provide guidance on the use of HV-AAS and SIPPAS for monitoring of VMSs in air.
■
EXPERIMENTAL SECTION Chemicals. The cVMSs included hexamethylcyclotrisiloxane (D3, C6H18O3Si3), octamethylcyclotetrasiloxane (D4, C8H24O4Si4), decamethylcyclopentasiloxane (D5, C 10 H 30 O 5 Si 5 ), and dodecamethylcyclohexasiloxane (D 6 , C12H36O6Si6) and the lVMSs included octamethyltrisiloxane (L 3 (MDM), C 8 H 24 O 2 Si 3 ), decamethyltetrasiloxane (L 4 (MD 2 M), C 10 H 30 O 3 Si 4 ), dodecamethylpentasiloxane (L5 (MD3M), C12H36O4Si5) (Gelest, Morrisville, PA, U.S. and Sigma-Aldrich, Oakville, ON, Canada). In addition, masslabeled 13C4-D4, 13C5-D5, 13C6-D6 were used as internal standards (IS) (Moravek Biochemicals, Brea, CA, U.S.) and mirex (C10Cl12) was used as an injection standard (InjS) (see Table S1 in the Supporting Information (SI)). Sampling. The calibration of the SIP-PAS and PUF-PAS were performed against HV-AAS at a semiurban meteorological station in Toronto (Environment Canada field site, 43°46′N, 79°28′W) from March 30 to October 13, 2010 as described in our study on per- and polyfluoroalkyl substances (PFASs) using the same sampling approach.25 Under the calibration study, the PAS were deployed for 7, 21, 28, 42, 56, 84, 112, 140, 168, and 197 days. Duplicate PAS were collected on days 28, 84, and 197 to verify reproducibility. After the completion of the calibration study, SIP-PAS, PUF-PAS, and HV-AAS sampling continued until the end of April 2011 for investigating seasonal trends. SIP and PUF disks were individually housed inside precleaned (rinsed with acetone and methanol) stainless steel chambers (“original chamber”, model TE-200-PAS, Tisch Environmental) and deployed ∼2 m above the ground during field deployment. The SIP and PUF disks were deployed for ∼28 days over 12 sampling periods. To compare different chamber configurations for SIP and PUF disks, four different chambers were used with different gaps between the two stainless steel bowl housings (i.e., original chamber’ with 1 cm overlap, “flush chamber”, “1 cm gap chamber”, “2 cm gap chamber”). For active sampling, high volume air samples (in average, ∼330 m3 over 24 h periods, one to two times a week) were collected from March, 2010 to April, 2011 using PS-1 type sampler (Tisch Environmental, Cleves, OH, U.S.) made of stainless steel (VMS-free sealings were used in the HV-AAS sampling head). The high volume air sampler uses glass-fiber filters (GFFs) (Type A/E Glass, 102 mm diameter, Pall Corporation) for collecting the particle-phase (n = 70) followed by a PUF/XAD-2 cartridge for trapping the gas-phase compounds (n = 70). Details on the preparation of SIP-PAS and PUF/XAD-2 cartridges are described in the SI and elsewhere.26,27 Field blanks were collected by exposing SIPs (n = 6), PUFs (n = 6), GFFs (n = 6), and PUF/XAD-2 cartridges (n = 6) for 1 min in the sampler at the sampling site and then treating them like real samples. To check the efficiency of the collection of VMSs in the gas-phase, breakthrough experiments were conducted by operating two sets of PUF/XAD-2 cartridges in
■
RESULTS AND DISCUSSION Quality Assurance/Quality Control. VMSs were not detected in PUF-PAS and GFFs indicating a low sorptive capacity of PUF-PAS for VMSs and a low sorption potential of VMSs to the particle-phase. To minimize contamination of the samples with VMSs, VMS-containing materials and personal care products were avoided and exposure to indoor air was minimized during sample preparation, extraction, and analysis (for details see SI). The blank concentrations, limits of detection (LODs), and limits of quantification (LOQs) are given in Tables S5 and S6 in the SI. All concentrations are corrected for blanks by subtracting the average of individual VMSs in the field blanks (ng absolute) from the amount (ng) in the samples. The reproducibility of the SIP-PAS samples ranged between 2.9 and 13% (n = 3). The average surrogate IS recoveries for 13C4-D4, 13C5-D5, and 13C6-D6 were 57 ± 20%, 66 ± 21%, and 70 ± 21% for HV-AAS (n = 76) and 49 ± 9%, 63 ± 11%, and 68 ± 11% for SIP-PAS (n = 42), respectively. The recoveries were in the same range as previously published.12,15,16 Breakthrough experiments were conducted to check the efficiency of the UF/XAD-2 sandwich for trapping the gas-phase compounds using HV-AAS (n = 3, air volume ∼330 m3). The breakthrough values for individual VMSs were determined by 9375
dx.doi.org/10.1021/es502081j | Environ. Sci. Technol. 2014, 48, 9374−9381
Environmental Science & Technology B=
[adsorbent]back [adsorbent]front + [adsorbent]back
Article
observed in this study between the Veq over the range of wind speeds, temperatures and other meteorological parameters for the different chamber housing designs (p > 0.05, Pearson Correlation, see Table S3 in the SI). In this study, the log KSIP−A values for the VMSs were in the range 5.5−6.2 and decreased slightly with increasing molecular weight (Table 1) These log KSIP−A values agree well with those reported by Cheng et al. (2011) (i.e., 5.7−6.4) for an indoor calibration of SIP for VMSs.12 However, the indoor study did not exhibit the trend of increasing log KSIP−A values with decreasing molecular weight. Part of this discrepancy may be due uncertainties in the derivation of log KSIP−A values due to variable ambient air concentrations of the VMSs, especially indoors. It is noteworthy that the average R-value for SIP reported by Cheng et al. (2011) was 2.6 m3 d−1 which is lower than the average in the current study of 4.2 m3 d−1.12 Lower Rvalues have been reported previously for indoor studies using PUF-PAS and attributed to more stagnant air masses indoors that results in reduced air-side controlled mass transfer rates.29,31 The uptake profile for SIP-PAS can be described by the uptake constant kU (d−1).29
(1)
where B is the vapor phase breakthrough for the HV-AAS and the [adsorbent] is the individual VMS concentration (ng m−3) in the back and front PUF/XAD-2 cartridge.27 The mean B value for VMSs ranged between 0.15 (L5) to 0.36 (D3). Values of B lower than 0.5 are desirable and indicative of small breakthrough (see Figure S1 in the SI).28 However, measured concentrations based on HV-AAS might be underestimated due to breakthrough which can result in an overestimation of sampling rates for SIP-PAS. Uptake Profile and Sampling Rates for SIP-PAS. The uptake of chemicals by the passive sampler medium (SIP) is compound-specific and based on the diffusivity in the air and SIP-air partition coefficient (KSIP−A).29 The volume of air sampled after a given exposure period is defined as the equivalent air volume (Veq, m3) for a passive air sampler. m Veq = SIP cA (2) where mSIP is the amount of chemical in the SIP (ng SIP−1) and cA is the total concentrations of the target analyte in ambient air using the HV-AAS (ng m−3). The uptake profiles of VMSs over the deployment time are shown in Figure 1. The VMSs had a short linear uptake of a few
kU =
ASIP kA × VSIP KSIP − A
(3)
where ASIP is the planar area of the passive sampler in m2 (i.e., 0.027 m2), VSIP is the volume of the SIP in m3 (i.e., 0.00021 m3), kA is the air-side mass-transfer coefficient (m d−1), and KSIP−A is the SIP-air partition coefficient. KSIP−A can be calculated using the volume of ambient air (VAIR) that contains an equivalent amount of chemical contained in a SIP disk having a volume (VSIP). The kU can be used to calculate the extent of the linear uptake phase which we define here as the time up to 25% of equilibrium (t25) (t25 = ln(0.75)/kU) and the time of 95% of equilibrium value (t95 = ln(0.05)/kU). The estimated t25 ranged between 9 and 24 days, while the estimated t95 ranged between 3 and 8 months (Table S7 in the SI). The t25- and t95-values for D6 and L5 were in this study 1 order of magnitude lower than estimates for XAD-PAS but comparable for D3, D4, D5, L3, and L4.18 The correlation of log KSIP−A against log octanol-air partition coefficient (KOA) for VMSs in SIP-PAS was investigated (see Table S8 and Figure S2 in the SI). The log KSIP−A values for the cVMSs were relatively constant and not well correlated with log KOA. The log KSIP−A values for the lVMS were lower than for cVMS and also not correlated with log KOA. These results indicate that octanol is probably not a suitable surrogate for sorption of VMSs to SIPs (XAD). VMSs are weak hydrogen bond acceptors. Consequently they will form bonds with octanol which is both a hydrogen bond donor and acceptor. However, XAD is aromatic and therefore acts only as a hydrogen bond acceptor and not expected to undergo hydrogen bonding with XAD, which is also a hydrogen bond acceptor. The sampling rate (R, m3 d−1) was derived from the linear uptake phase of the uptake profiles, by taking the slope of the plot of mSIP/cA versus time. For cVMSs, the R-values ranged between 4.1 and 5.7 m3 d−1, whereas the R-values for lVMSs were generally lower ranging from 3.3−3.5 m3 d−1. Differences of the R-values can be explained by varying air concentrations of VMSs, different physicochemical properties of VMSs, lack of correction for the breakthrough of cVMSs in HV-AAS and analytical or experimental variability. These R-values are very
Figure 1. Uptake profiles of individual lVMSs and cVMSs in SIP-PAS. The equivalent air volume (Veq, m3) for a given chemical is calculated by dividing the amount of chemical collected in the SIP (mSIP, ng SIP−1) by the estimated average air concentrations of the target analyte using the HV-AAS (cA, ng m−3), over this time period.
weeks and equilibrated after about three months. The Veq for the cVMSs after equilibration (286−487 m3) was about three times higher compared to the lVMSs (101−138 m3) (Figure 1). The uptake profile can be influenced by the chamber housing design and meteorological factors such as wind speed and temperature.29,30 However, no significant correlation was 9376
dx.doi.org/10.1021/es502081j | Environ. Sci. Technol. 2014, 48, 9374−9381
Environmental Science & Technology
Article
Table 1. Calibration Results for SIP-PAS for Linear and Cyclic Volatile Methylsiloxanesa cA (ng m−3) D3 D4 D5 D6 L3 L4 L5 a
1.35 13.6 86.9 8.48 1.01 2.05 2.06
mSIP (ng disk−1) 658 4288 20 226 2422 140 257 207
Veq (m3)
mSIP (ng m−3 disk−1)
KSIP‑Ab
487 316 233 286 138 125 101
× × × × × × ×
× × × × × × ×
3.14 2.05 9.65 1.16 6.67 1.22 9.88
06
10 1007 1007 1007 1005 1006 1005
2.32 1.51 1.11 1.36 6.59 5.97 4.80
06
10 1006 1006 1006 1005 1005 1005
log KSIP‑Ab
kA (m d−1)
R (m3 d−1)
6.37 6.18 6.05 6.13 5.82 5.78 5.68
155 129 112 130 89 96 91
5.7 4.8 4.1 4.8 3.3 3.5 3.4
VSIP = 0.00021 m3, ASIP = 0.027 m2. bAverage air temperature was 18 °C between 20/07/2010 to 13/10/2010.
Figure 2. Comparison between total concentrations in air (ng m−3) using HV-AAS and SIP-PAS using linear regression for the lVMSs and cVMSs. The dotted 1:1 line represents a perfect agreement and the dashed line represents a difference by the factor of 2 between the two sampler types.
close to the suggested R-value of 4 m3 d−1 reported previously for persistent organic pollutants (POPs) and poly- and PFASs indicating similar uptake characteristics for POPs, PFASs, and VMSs using SIP-PAS.25,32 The R-values for VMSs using XADPAS were about 1 order of magnitude lower (0.42−0.50 m3 d−1)18 compared to SIP-PAS. For analytes which are still in the linear phase, the sample volume (VAIR, m3) can be simply calculated by multiplying the R-value of the analyte with the days of deployment (t). However, the linear phase for VMSs was lower (i.e., 0.05, Student’s t-test) (Figure S3 in the SI). This shows that different configurations of the SIP-PAS capture VMSs with similar efficiency. The air concentration of VMSs measured by the HV-AAS (representing 14−29% of the time for the monthly average) was compared with the air concentration derived by the SIPPAS (which sample 100% of the time) using linear regression (Figure 2). The average difference between the two sampling techniques was less than 50% for individual VMSs and no significant differences were found for the VMS concentrations measured by the SIP-PAS and HV-AAS (p > 0.05, Kruskal− Wallis test) (Figure 2, Figure S4 in the SI). The lVMS concentrations showed a higher scattering of the data (r2 = 0.47) which can be explained by low environmental concentrations in air of these compounds. Consequently, these lower concentrations may approach detection limits for lVMSs that have greater analytical uncertainty. In contrast, the cVMS concentrations showed a good linear regression r2 = 0.92). However, individual VMSs showed a different linear
⎛ ⎡ ⎛A kA ⎞ VAIR = KSIP − A × VSIP × ⎜⎜1 − exp⎢ −⎜ SIP × ⎟ ⎢⎣ ⎝ VSIP KSIP − A ⎠ ⎝ ⎤⎞ × t ⎥⎟⎟ ⎥⎦⎠
(4) 3
3
The VAIR-values ranged between 69 m (L5) and 105 m (D3) for a one month deployment period. The concentration of the analyte in air (cA, ng m−3) is calculated by dividing the amount of chemical in the SIP (mSIP, ng SIP−1) by the VAIR-value. cA = mSIP /VAIR
(5)
Overall, VMSs showed a relatively short linear uptake phase (i.e.,
