Sampling and analysis of polychlorinated biphenyls in indoor air by sorbent enrichment followed by headspace solid-phase microextraction and gas chromatography-tandem mass spectrometry

Sampling and analysis of polychlorinated biphenyls in indoor air by sorbent 1 enrichment followed by headspace solid-phase microextraction and gas 2 chromatographytandem mass spectrometry 3 4 Ruth Barro, Sergio Ares, Carmen Garcia-Jares*, Maria Llompart and Rafael Cela 5 6 Departamento de Quimica Analitica, Nutricion y Bromatologia, Facultad de Quimica, 7 Instituto de Investigacion y Analisis Alimentario, Universidad de Santiago de 8 Compostela, E-15782 Santiago de Compostela, Spain. 9 10 11 12 13 Corresponding author: 14 * Carmen Garcia-Jares 15 E-mail: qncg[email protected] 16 Phone: +34-981563100, ext. 14394; FAX: +34-981595012 17 Abstract 1 In this study, a combination of solid-phase extraction (SPE) and solid-phase 2 microextraction (SPME) techniques has been used to determine polychlorinated 3 biphenyls in air. Using a vacuum pump, a known volume of air was forced to pass 4 through a porous polymer (Tenax TA) where the target analytes were retained and 5 headspace SPME was carried out. The quantification was performed using gas 6 chromatography coupled to mass spectrometry in tandem. Some parameters that could 7 affect SPE and SPME steps were studied and a screening factorial design was used to 8 evaluate the influence of the addition of solvents to the adsorbent and the type of fibre 9 coating. Performance of the method was evaluated demonstrating that external 10 calibration, which does not require performing the complete determination procedure, 11 was suitable. The correlation coefficients were calculated and a lack-of-fit test was run 12 within the calibration data. Repeatability of the method was found adequate 13 (RSD12%). Limits of quantification were found below 0.100 ng/m3when only 2.5 m3 14 air were sampled, well lower than recommended exposure levels given by different US 15 health organizations. In addition, more sensitivity could be attained by increasing the 16 volume of air sampled (decrease in retention efficiency was not detected for sample 17 volumes up to 25 m3), and/or extending the extraction time in the SPME step. 18 19 20 Key words 21 Air analysis, polychlorinated biphenyls, gas chromatography-mass spectrometry, solid22 phase microextraction, solid-phase extraction, factorial design, chlorinated compounds. 23 Introduction 1 Polychlorinated biphenyls (PCBs) are compounds with high lipid solubility and high 2 stability and persistence in the environment. They have been manufactured in 3 substantial amounts since the 1920s for being used in the electrical, paint, pigments, 4 paper, and cardboard industries; so, they were spread to the remotest areas of the world 5 before any control on use or disposals was implemented. The application of PCBs can 6 cause potential adverse health effects to humans by contaminating soil, water, air, plants 7 and animal life. Due to their stable, low volatility and lipophilic chemical nature, PCBs 8 exhibit bio-accumulative, chronic health effects; therefore, monitoring the presence of 9 these compounds in ambient air is of great importance [1-3]. 10 The International Agency for Research of Cancer (IARC) has determined that PCBs are 11 probably carcinogenic to humans. US Environmental Protection Agency (US EPA) has 12 classified PCBs as a Group B2, probable human carcinogen. The National Institute for 13 Occupational Safety and Health (NIOSH) recommends workers not breathe air with 14 more than 0.001 milligrams of PCBs per cubic meter of air (0.001 mg/m3) for a 10-hour 15 workday, 40-hour workweek. The Occupational Safety and Health Administration 16 (OSHA) requires workplace exposure limits of 0.5 mg/m3(54 percent chlorine) or 1 17 mg/m3(42 percent chlorine) for a 8-hour workday to protect workers from non cancer 18 harmful health effects [4-7]. 19 Due to their low concentration in air, polychlorinated biphenyls have been extensively 20 sampled by solid phase extraction (SPE), pumping air through a solid sorbent or 21 mixtures of solid sorbents, where the compounds are retained. Florisil [8], silica gel [9], 22 polyurethane foam (PUF) [10, 11], XAD-2 resin [12], Carbosphere activated carbon 23 [13], functionalized styrene-divinylbenzene [14], are sorbents used to retain PCBs from 24 air. Tenax, a 2,6-dipheyl-p-phenylene oxide porous polymer, presents hydrophobic 25 nature and low interference by moisture adsorption for sampling humid air. 1 Consequently, Tenax has been extensively used for the recovery of volatile organic 2 compounds from contaminated air [15-18]. Its hydrophobic nature is an advantage over 3 some common hydrophilic sorbents, such as charcoal and silica gel, because air 4 humidity may reduce the sorption efficiency. [19]. 5 Analytes retained by Tenax can be desorbed using an organic solvent or by thermal 6 desorption, prior to gas chromatographic analysis. Tenax is incompatible with many 7 solvent systems and then, it is rarely used for the retention of PCBs, which are mostly 8 extracted from the sorbents using Soxhlet extractors. On the other hand, thermal 9 desorption is more indicated for volatile analytes. Desorption of some low volatility 10 PCBs would require the application of high temperatures to the Tenax, and regardless of 11 its relatively inert nature, thermal desorption can cause drawbacks, as thermally, 12 mechanically and chemically degradation of the sorbent, as well as carryover problems 13 coming from poorly desorptions [20]. Some authors have noted changes in Tenax TA 14 from re-used thermal desorption tubes [21] and a few degradation products from Tenax 15 GC, such as benzaldehyde and acetophenone are well known [22]. 16 Saba and co-workers have proposed the use of SPME following a preconcentration step 17 on Tenax to determine benzene and toluene in air [23,24], and recently, the authors have 18 optimized the experimental conditions to analyze volatile and semivolatile 19 chlorobenzenes in indoor air [25]. 20 Solid-phase microextraction (SPME) provides some advantages over traditional 21 extraction methods. It offers solvent-free operation, and in spite of the limited amount 22 of analyte extracted, all is introduced into the GC injection port, allowing for good 23 sensitivity, with cost effectiveness and operational simplicity [26-28]. In addition, 24 SPME quantitative analysis is feasible in non-equilibrium situations once experimental 25 parameters are held constant, so a much shorter sampling time can be used for 1 quantitative analysis [29]. 2 The aim of the present paper is to demonstrate that the combination of SPE-SPME 3 using Tenax as adsorbent can be useful to develop a method for the analysis of 4 polychlorinated biphenyls in indoor air samples. Thus, a study of the parameters 5 influencing SPME was carried out with the help of an experimental design strategy, 6 which reduces the experimental work required and allows accounting for possible factor 7 interactions. The performance of the method was also studied, demonstrating that limits 8 of detection in the low-medium pg/m3can be achieved. In addition, the method was 9 applied to a real contaminated air sample. 10 11 Experimental 12 Reagents 13 2,4,4’-trichlorobiphenyl (PCB-28), 2,2’,5,5’-tetrachlorobiphenyl (PCB-52), 2,2’,4,5,5’- 14 pentachlorobiphenyl (PCB-101), 2,2’,4,4’,5,5’-hexachlorobiphenyl (PCB-138), 15 2,2’,3,4,4’,5’-hexachlorobiphenyl (PCB-153), 2,3,3’,4,4’,5’- hexachlorobiphenyl (PCB16 156), 2,2’,3,4,4’,5,5’-heptachlorobiphenyl (PCB-180) were supplied by Ultra Scientific 17 (North Kingston, RI). All organic solvents used (isooctane, acetone, methanol and n18 hexane) were of pesticide grade and were obtained from Merck (Mollet del Vallés, 19 Barcelona, Spain). 20 Standard stock solutions of 800-1000 g/mL of individuals were prepared in isooctane, 21 and a stock solution containing a mixture of the seven target PCBs at 50 mg/L was 22 prepared in acetone. Working solutions were obtained by appropriate dilution in 23 isooctane, n-hexane or acetone. All solutions were stored in amber colored vials and 24 stored at –20ºC. 25 1 Air sampling and extraction of PCBs 2 Using a vacuum pump working at 100 L/min, a known volume of air was pumped 3 through a glass tube containing 25 mg Tenax TA adsorbent (mesh size 60/80) (Aldrich, 4 Madrid, Spain). A schematic view of the sampling device is shown in Figure 1. Only 5 Teflon (PTFE) tubing was used for connections. The adsorbent was then poured into a 6 10-mL glass vial and sealed with an aluminium cap furnished with a PTFE-faced 7 septum. As it will be discussed later, a known volume of organic solvent (n-hexane, 8 acetone or a mixture of both solvents) is added to the adsorbent. Then, solid phase 9 microextraction was carried out immersing the vial into a water bath maintained at 50 or 10 100ºC and exposing a SPME fiber to the headspace of the vial (HSSPME). Experiments 11 at 150ºC were carried out inside a conventional GC oven. The extraction time was fixed 12 at 30 or 60 minutes. To achieve good repeatability, vials should be immersed up to the 13 neck into the thermostated water bath. Once finished the SPME process, the fiber was 14 immediately inserted into the injection port of the gas chromatograph during 4 minutes 15 at 260ºC. 16 To study the retention of PCBs on Tenax, 100 L of standard mixtures of the target 17 PCBs in n-hexane were directly spiked on 25 mg of the adsorbent. The spike was left to 18 homogenize with the adsorbent for several hours. Then, the spiked Tenax was treated as 19 described above. In some experiments performed to detect the possible breakthrough of 20 the adsorbent, a second glass tube containing 25 mg of non-spiked Tenax was 21 connected in series with the first spiked one, and both portions of adsorbent were 22 individually extracted using the SPME procedure. 23 SPME manual holders and fibers were obtained from Supelco. Fibers used in this work 1 were: 100 µm polydimethylsiloxane (PDMS) or 65 µm polydimethylsiloxane2 divinylbenzene (PDMS/DVB). 3 4 Gas Chromatography-In tandem Mass Spectrometry 5 Analyses were performed in a Varian 3800 gas chromatograph (Varian Chromatography 6 systems, Walnut Creek, CA, USA) equipped with a 1079 split/splitless injector and an 7 ion trap mass detector Varian Saturn 2000 with a waveboard for MSnanalysis. The 8 system was operated by Saturn GC-MS WorkStation v5.4 software. The target 9 compounds were separated on a 25 µm length x 0.25 mm i.d., Varian CP-Sil8 CB Low 10 bleed/MS column coated with a 0.25 µm film. The GC oven temperature program was: 11 60ºC hold 3 min, rate 20ºC/min to 180ºC hold 7.5 min, rate 5ºC/min to 260ºC hold 2 12 min with a total acquisition program of 34.5 min. Helium was employed as carrier gas, 13 with a constant column flow of 1.2 mL/min. The injector was operated in the splitless 14 mode and programmed to return to the split mode after 2 min from the beginning of a 15 run. Split flow was set at 50 mL/min. Injector temperature was held constant at 270ºC. 16 The mass spectrometer was operated in electron ionization (EI) mode at 70 eV. The 17 mass range was scanned from 40 to 650 m/z at 1 s/scan for the full scan mode. For 18 MS/MS, all compounds were analyzed using a resonant waveform type. Each segment 19 included an ion preparation method (IPM) that defines MS-MS parameters and m/z scan 20 range [30]. Trap, manifold and transfer line temperatures were maintained at 250ºC, 21 50ºC and 280ºC, respectively. 22 1 Results and discussion 2 The SPME process was studied before optimization of the sampling step because the 3 transfer of the polychlorinated biphenyls from the adsorbent to the fiber might seriously 4 affect the sensitivity of the whole extraction method, and the amount of PCBs retained 5 by the fiber greatly depends on the experimental conditions used to carry out the 6 microextraction. 7 Initial experiments were performed using dry Tenax (no solvent addition). The amount 8 of adsorbent was selected according to previous results [25]. Portions of 25 mg 9 adsorbent were spiked with the target analytes, and SPME was conducted at different 10 extraction temperatures (50, 100 and 150ºC). Similar results (a factor 0.9-1.2 for all 11 compounds) were obtained working at 100ºC and 150ºC, while an important detriment 12 in response (a factor of 15-100) was observed when temperature was maintained at 13 50ºC. Then, an extraction temperature of 100ºC was chosen to carry out next studies. 14 To improve extraction of target analytes from Tenax and their transfer to the fiber, the 15 addition of a small volume (100 µL) of different solvents (hexane, acetone, water, 16 methanol) and solvent mixtures (hexane/acetone and methanol/acetone) previously to 17 SPME step was also investigated. A considerable improvement of the chromatographic 18 response was observed when hexane or acetone was used to wet the adsorbent. The 19 responses obtained were 20-70-fold higher than those obtained when no solvent was 20 added. Therefore, the addition of these solvents was considered for further study using 21 an experimental design approach. 22 A multifactor screening 3x2^2 mixed level factorial experimental design (type V 23 resolution), was carried out to study in 12 runs the influence of three main factors in the 24 SPME process [31]. The advantage of this design is that it allows the study of main 25 effects, as well as two-factor interactions. The factors considered in this design were: 1 percentage of acetone:hexane mixture and volume of solvent added to Tenax, both as 2 continuous factors; and type of fiber coating as discontinuous factor. The fiber included 3 in this study, PDMS and PDMS/DVB, were selected considering previous experience in 4 SPME of PCBS [32] In Table 1, the upper and lower levels given to each factor, as well 5 as the factor key, are presented. 6 The results of the experimental design indicated the statistical significance of some of 7 the main factors. Table 2 summarizes the analysis of variance for main factors. 8 Interactions were not included in this table since they were not significant with the 9 exception of BC (extraction volume and fiber coating) for PCB-52 and PCB-101. This 10 interaction will be discussed later. A factor is significant when its p-value is lower than 11 0.05 (95% confidence level). As can be seen in this table, solvent volume was 12 significant for the extraction of PCBs 52, 101 and 180. The percentage of hexane in the 13 solvent (acetone) was significant for PCBs 28, 52 and 138. In addition, fiber coating 14 was a significant factor for the extraction of the most chlorinated PCBs (PCB-156 and 15 PCB-180). 16 Figure 2 shows the main effects graphs for the target PCBs excluding PCB-153 since, 17 for this compound, none of the factors was statically significant (see Table 2). In these 18 plots, obtained by drawing a line betwen the low and the high levels of main factors, we 19 can see the magnitude of the effect of each factor on the microextraction process, as 20 well as the level of the factor that produces the highest response. The influence of each 21 factor is clearly appreciated in this figure. The fiber coating is an important factor in the 22 extraction of the most chlorinated PCBs. For these compounds, higher response is 23 achieved by PDMS coating. For the other compounds, responses obtained were 24 independent on the fiber used. So, PDMS can be selected as the most suitable coating 25 [13] A. L. Garcia, A.C. Den Boer, A.P.J.M. De Jong, Environ. Sci. Technol. 30 (1996) 1 1032. 2 [14] M. Ramil, I. Rodriguez, R. Cela, J. Chromatogr. A 963 (2002) 65. 3 [15] N. M. Bradshaw, J.A. Ballantine, Environ. Technol. 16 (1995) 433. 4 [16] M.P. Baya, P.A. Siskos, Analyst 121 (1996) 303. 5 [17] R.A Hallama, E. Rosenberg, M. Grasserbauer, J. Chromatogr. A 809 (1998) 47. 6 [18] S.F. Patil, S.T. Lonkar, J. Chromatogr. A 684 (1994) 133. 7 [19] V.G. Berezkin, Y.S. Drugov, Gas Chromatography in Air Pollution Analysis, 8 Journal of Chromatography Library, Volume 49, Ed. Elsevier, Amsterdam, 1991. 9 [20] M. Harper, J. Chromatogr. A 885 (2000) 129. 10 [21] H. Rothweiler, P.A. Wager, C. Schlatter, Atmos. Environ. 25B (1991) 679. 11 [22] E.D. Pellizzari, B. Demian, K.J. Krost, Anal. Chem. 56 (1984) 793. 12 [23] A. Saba, A. Cuzzola, A. Raffaelli, S. Pucci, P. Salvadori, Rapid Commun. Mass 13 Spectrom. 15 (2001) 2404. 14 [24] A. Saba, A. Raffaelli, S. Pucci, P. Salvadori, Rapid Commun. Mass Spectrom. 13 15 (1999) 1899. 16 [25] R. Barro, S. Ares, C. Garcia-Jares, M. Llompart, R. Cela, J. Chromatogr. A 1045 17 (2004) 189. 18 [26] J. Pawliszyn, Solid Phase Microextraction. Theory and Practice, Wiley-VCH, New 19 York, 1997. 20 [27] J.A. Koziel, M. Jia, J. Pawliszyn, Anal. Chim. Acta 400 (1999) 153. 21 [28] A. Khaled, J. Pawliszyn, J. Chromatogr. A 892 (2000) 455. 22 [29] J. Ai, Anal. Chem. 69 (1997) 1230. 23 [30] R. Barro, S. Ares, C. Garcia-Jares, M. Llompart, R. Cela, Anal. Bioanal. Chem. 24 2004, in press. 25 [31] Statgraphics-Plus, Manugistics Inc., Rockville, Maryland, 1996. 1 [32] M. Llompart, M. Pazos, P.Landin, R. Cela, Anal. Chem. 73 (2001) 5858. 2 Table 1. Ion preparation method (IPM) parameters for each of the six segments of the GC-MS/MS method Segment PCB Start time (min) Finish time (min) Parent ions (m/z) Scan range (m/z) Quantification ions (m/z) Excitation storage level (m/z) Excitation amplitude (V) Collision induced dissociation frequency offset (Hz) Solvent delay - 0.00 13.0 - - - - - - 2 28 13.00 16.5 258.0 [M+2] 100270 186+188 133.0 1.15 600 3 52 16.5 19.5 292.0 [M+2] 140300 220+222 157.0 1.15 600 4 101 19.5 23.5 325.9 [M+2] 180330 254+256+289+291 181.0 1.29 700 5 153, 138, 156 23.5 28.7 359.8 [M+2] 180375 288+290 206.0 1.59 700 6 180 28.7 30.0 395.8 [M+4] 220400 324+326 231.0 1.71 800 Table 2. Factors and levels considered in the experimental design. Factor Key Low level High level Continuous % hexane A 0 100 yes solvent volume (L) B 100 200 yes fiber coating C PDMS/DVB PDMS no Table 3. ANOVA results showing the significance of main effects Compound % hexane solvent volume fiber coating F-ratio p-value F-ratio p-value F-ratio p-value PCB-28 9.60 0.02 4.87 0.06 0.26 0.63 PCB-52 5.92 0.04 8.39 0.02 0.06 0.81 PCB-101 0.38 0.56 7.90 0.03 0.01 0.92 PCB-153 3.17 0.12 0.02 0.90 1.32 0.29 PCB-138 5.92 0.04 0.46 0.52 1.63 0.24 PCB-156 3.45 0.11 1.79 0.22 5.17 0.05 PCB-180 2.45 0.16 6.65 0.04 7.16 0.03 Table 4. Linearity, recovery, repeatability (%RSD), and limits of detection of the method. Compound Linearity Recovery (%) Repeatability (%RSD) Detection limits (S/N=3, ng/m3) Coefficient of determination (R2) F-test p-value 4ng/m3 40ng/m3 SPME SPE-SPME PCB-28 1.000 0.67 0.6402 100 92 14 11 0.011 PCB-52 0.999 4.76 0.0588 101 94 10 10 0.017 PCB-101 0.999 3.64 0.0946 90 90 13 7.8 0.016 PCB-153 0.997 0.64 0.6585 99 95 7.2 6.9 0.018 PCB-138 0.996 1.18 0.4187 108 97 8.8 9.9 0.037 PCB-156 0.993 0.37 0.8206 99 92 8.4 12 0.030 PCB-180 0.992 0.27 0.8859 101 92 8.8 11 0.096 Figure captions Figure 1. Schematic plot of the air sampling device. 1: vacuum pump, 2: PTFE connectors, 3: flow meter, 4: glass tube containing Tenax TA. Figure 2. Graphics showing the influence of main effects on the extraction of the target PCBs. Figure 3. Interaction plot for PCB-52. Figure 4. Extracted ion current chromatograms for an air sample containing 4 ng/m3of the target PCBs. Figure 5. Extracted ion current chromatograms for a real air sample. Figure 1 4 air 1 2 3 Figure 2. Figure ¿???. PCB52 area counts % hexane 0 100 solvent volume 100 200 fiber coating DVB PDMS 220 240 260 280 300 PCB156 area counts % hexane 0 100 solvent volume 100 200 fiber coating DVB PDMS 76 81 86 91 96 101 PCB28 area counts % hexane 0 100 solvent volume 100 200 fiber coating DVB PDMS 180 200 220 240 260 PCB101 area counts % hexane 0 100 solvent volume 100 200 fiber coating DVB PDMS 210 220 230 240 250 260 270 PCB138 area counts % hexane 0 100 solvent volume 100 200 fiber coating DVB PDMS 95 105 115 125 135 PCB180 area counts % hexane 0 100 solvent volume 100 200 fiber coating DVB PDMS 62 66 70 74 78 82 86 Figure 3 AB AC BC PCB-52 190 220 250 280 310 340 0 100 - - + + 0 100 - - + + 100 200 - - + + area counts