Development of a sensitive methodology for the analysis of chlorobenzenes in air by combination of solid-phase extraction and headspace solid-phase microextraction

1 1 Development of a sensitive methodology for the 2 Analysis of Chlorobenzenes in Air by 3 Combination of Solid-Phase Extraction and 4 Headspace Solid-Phase Microextraction 5 6 Ruth Barro, Sergio Ares, Carmen Garcia-Jares*, Maria Llompart and Rafael Cela 7 8 Departamento de Química Analítica, Nutrición y Bromatología, Facultad de Química, 9 Instituto de Investigación y Análisis Alimentario, Universidad de Santiago de 10 Compostela, E-15782 Santiago de Compostela, Spain. 11 12 13 14 15 Corresponding author: 16 * Carmen Garcia-Jares 17 E-mail: qncg[email protected] 18 19 2 1 Abstract 2 In this study, a combination of solid-phase extraction and solid-phase microextraction 3 has been used to determine chlorobenzenes in air. Analytes were sampled by pumping a 4 known volume of air through a porous polymer (Tenax TA). Then, the adsorbent was 5 transferred into a glass vial and SPME was performed. The quantification was carried 6 out using GC-ECD or GC-MS. Several SPME coatings (100m PDMS, 75m 7 CAR/PDMS, 65m PDMS/DVB, 65m CW/DVB and 85m PA) were evaluated, 8 obtaining the highest responses with CAR/PDMS for the most volatile chlorobenzenes, 9 and with PDMS/DVB or CW/DVB fibers for the semivolatile compounds. To optimize 10 some other factors that could affect the SPME step, a factorial design was used. Kinetic 11 studies of the SPME process were also performed. Concerning the SPE step, 12 breakthrough was studied, showing that 2.5 m3of air could be processed without losses 13 of the most volatile compounds. The performance of the method was evaluated. 14 External calibration, which does not require the complete sampling process, 15 demonstrated to be suitable, obtaining good linearity (R2>0.99) for all chlorobenzenes. 16 Recovery studies were performed at two concentration levels (4 ng/m3and 40 ng/m3), 17 obtaining quantitative recoveries (>80%). Limits of detection at the sub ng/m3were 18 achieved for all the target compounds. 19 20 Key words 21 Air analysis, chlorobenzenes, gas chromatography-mass spectrometry, Solid-phase 22 microextraction, solid-phase extraction, factorial design, volatile compounds. 23 3 Introduction 1 Chlorobenzenes are a family of environmental pollutants that are produced in huge 2 quantities in industrial processes to be used as intermediates in the synthesis of other 3 organic chemicals and in the production of a wide range of consumer and commercial 4 products. The less chlorinated benzenes are widely used in cleaning and degreasing of 5 metal, leather, wool and paper, in dry cleaning and textile dyeing operations, as wood- 6 preserving compounds, in organic synthesis of pesticides and herbicides, as deodorizing 7 agents for garbage and sewage, as air fresheners, as heat transfer mediums in 8 maintenance equipment, and as magnetic coil coolants for the electrical and electronics 9 industries. They are also used in application or removal of surface coatings as solvents 10 for organic materials, waxes, resins, rubbers, oils and asphalts. Pentachlorobenzene is 11 used to make pentachloronitrobenzene, a fungicide and it is currently used as fire 12 retardant. Hexachlorobenzene was used, among other applications, as fungicide, in the 13 production of pyrotechnic compositions for the military, as a plasticizer agent for 14 polyvinyl chloride (PVC), although in many countries its production and use have 15 ceased[1]. 16 Release of chlobenzenes to the environment occurs primary during manufacture, and 17 incineration of chlorobenzenes may lead to the emission of polychlorinated dibenzo-p- 18 dioxins and dibenzofurans. Volatile chlorobenzenes are extensively used as solvents, so 19 large quantities are released to the air. However, atmospheric concentrations are usually 20 very low, often much less than a few g/m3. Nevertheless, indoor air concentrations 21 may be from one to three orders of magnitude higher where they are used [2]. Risks of 22 human exposure arising from contaminated indoor air are linked to the use of these 23 compounds as moth repellents and air fresheners [1]. 24 4 Hexachlorobenzene and 1,4-dichlorobenzene were the first compounds included in the 1 Third and Fifth (repectively) Annual Report on Carcinogens in the U. S. Department of 2 Health and Human Services as reasonably anticipated to be a human carcinogens based 3 on sufficient evidence of carcinogenicity in experimental animals [2]. The Clean Air 4 Act Amendments of 1990 list some chlorobenzenes as Hazardous Air Pollutants 5 (HAPs), so federal agencies and groups may develop recommendations to assist in 6 controlling exposure [3]. 7 Chlorobenzenes are frequently found in air at very low concentrations, so a 8 preconcentration step before the analysis is necessary. Most volatile chlorobenzenes are 9 usually analyzed following general procedures developed for the analysis of volatile 10 organic compounds (VOCs). In most procedures, once the analytes are extracted from 11 air to an appropriate sorbent by solid-phase extraction, a thermal or solvent desorption 12 step is carried out to transfer the target compounds into a standard gas chromatograph. 13 For trapping VOCs, a wide variety of sorbents have been applied, such as carbon-based 14 material [4], Tenax [5] or mixtures of Tenax with other sorbents [6, 7]. Some other 15 trapping materials can also be used [8]. To retain the less volatile chlorobenzenes, the 16 use of more adsorbent materials such as expanded polyurethane foam (PUF), both used 17 as monosorbent or mixtured with other polymeric sorbents in multibed cartridges, were 18 studied, and the retained compounds were extracted using Soxhlet solvent desorption 19 during 12 hours [9, 10]. However, there is currently an increasing demand for simple 20 and cost-effective sampling and analytical methods capable of achieving very low 21 detection limits in real or almost real-time [11]. Benefits of solid-phase microextraction 22 (SPME) in the analysis of air samples include the use of simple instrumentation like 23 lightweight and compact devices, so expensive cryotraps or thermal desorbers are not 24 required, dangerous and toxic organic solvents or reagents are not used, and short 25 5 extraction times are usually employed [12]. SPME is an equilibrium technique and in 1 consequence, analytes are not quantitatively extracted. Therefore, the main problem in 2 SPME analysis, especially in the case of air, is calibration. Up to now, different 3 strategies have been studied to overcome this drawback of the technique [13-19]. In 4 addition, the working concentration levels for air analysis by SPME are usually in the 5 range of g/m3and mg/m3[20, 21]. These levels are suitable for the determination of 6 volatile compounds that can be found at relatively high concentrations in air. 7 Nevertheless, some hazardous air pollutants need to be monitored at very low 8 concentrations and the levels achieved by SPME might be insufficient. 9 In the present paper, a method based in the combination of SPE and SPME techniques 10 is proposed to determine chlorobenzenes (including the less volatile in the family) in air 11 samples. The optimization of the method was performed using an experimental design 12 approach. External calibration, which does not require the complete sampling process, 13 demonstrates to be suitable. Limits of detection at the sub ng/m3were achieved for all 14 the target compounds. 15 Up to now, the combination of both techniques has only been applied to the analysis of 16 two high volatile organic compounds (toluene and benzene) in air samples [22, 23]. 17 18 Experimental 19 Reagents 20 1,3-dichlorobenzene (1,3-DCB), 1,4-dichlorobenzene (1,4-DCB) and 1,2,4- 21 trichlorobenzene (1,2,4-TCB) were supplied by Fluka (CH, Switzerland), 1,2- 22 dichlorobenzene (1,2-DCB) and 1,2,3-trichlorobenzene (1,2,3-TCB) were purchased 23 from Aldrich Chemie (Steinheim, Germany), 1,2,3,4-tetrachlorobenzene (1,2,3,4- 24 TeCB), 1,2,3,5-tetrachlorobenzene (1,2,3,5-TeCB) and 1,2,4,5-tetrachlorobenzene 25 6 (1,2,4,5-TeCB) were obtained from Riedel-de Haën (Seelze, Germany), 1 pentachlorobenzene (PeCB) and hexachlorobenzene (HCB) were supplied by Supelco 2 (Bellefonte, PA, USA). All organic solvents (isooctane, acetone, methanol and hexane) 3 were of pesticide grade and were obtained from Merck (Mollet del Vallés, Barcelona, 4 Spain). 5 Standard stock solutions of 2000-4000 mg/mL of individuals were prepared in acetone 6 or isooctane, and working solutions were obtained by appropriate dilution. All solutions 7 were stored in amber colored vials and stored at –20ºC. 8 9 Air sampling and extraction of chlorobenzenes 10 Using a vacuum pump working at 100 L/min, a known volume of air was pumped 11 through a glass tube containing 25 mg of Tenax TA adsorbent (mesh size 60/80) 12 retained by glass wool (Aldrich, Madrid, Spain). For recovery experiments, the pump 13 was placed in a clean room provided of a laminar flow system, and a V-shaped tube was 14 inserted before the collecting Tenax tube. A solution of the target analytes in hexane 15 was then carefully placed in the V-shaped tube, and a selected volume of air was 16 pumped throughout the system (Figure 1). Thus, the air was enriched in the analytes 17 before reaching the Tenax tube. Only Teflon (PTFE) tubing was used for connections. 18 The adsorbent was then poured into a glass vial sealed with an aluminum cap furnished 19 with a PTFE-faced septum. The vial was placed into a water bath at 50 or 100 °C. 20 Compounds retained by the adsorbent were analyzed by exposing a SPME fiber to the 21 headspace of the vial (HS-SPME). Once finished the SPME process, the fiber was 22 immediately inserted into the injection port of the chromatograph and chlorobenzenes 23 were desorbed to the GC for 4 min. If necessary, vials containing chlorobenzenes 24 adsorbed on Tenax can be stored at –20ºC during a few days to further analysis. SPME 25 7 manual holders and fibers were obtained from Supelco. Fibers used in this work were: 1 100 m polydimethylsiloxane (PDMS), 65 m polydimethylsiloxane-divinylbenzene 2 (PDMS/DVB), 75 m carboxen-polydimethylsiloxane (CAR/PDMS), 65 m carbowax- 3 divinylbenzene (CW/DVB) and 85 m polyacrylate (PA). 4 5 Gas Chromatographic Analysis 6 Gas chromatography-mass spectrometry (GC-MS) analysis was performed in a Varian 7 3400 GC system, equipped with a Saturn 3 ion trap mass detector, operated by Saturn 8 version 5.4 software. A Varian VA-5MS or CP-Sil8 CB Lowbleed/MS (25 m x 0.25 9 mm i.d. x 0.25 m) column was used for the separation of chlorobenzenes. Working 10 GC-MS parameters are summarized in Table 1. Mass acquisition ranges were 11 programmed by time segments and centered on the ions characteristic of each group of 12 compounds (Table 2). 13 Gas chromatography with ECD analysis was performed in a Hewlett Packard 5890 14 Series II-Plus GC system, equipped with an electron capture detector and a 15 split/splitless injector, operated by HP Chemstation software. A SE-54 (30 m x 0.25 mm 16 i.d. x 0.25 m) column (Alltech, Deerfield, IL) was used to separate the target analytes. 17 The experimental GC-ECD parameters are shown in Table 1. 18 19 Results and discussion 20 One of the most relevant steps in the sample preparation method is the transfer of the 21 chlorobenzenes from the adsorbent to the SPME fiber. This clearly affects the amount 22 of compound adsorbed by the fiber and hence, the limits of detection and quantification 23 of the method. Therefore, the SPME process was studied before optimization of the 24 sampling step. 25 8 Different SPME fiber coatings (85 m PA, 65 m CW/DVB, 65 m PDMS/DVB, 100 1 m PDMS and 75 m CAR/PDMS) were evaluated using the same experimental 2 procedure: 100 mg of clean Tenax were placed in a glass vial and spiked with a standard 3 solution of chlorobenzenes in hexane to obtain 10 ng of each target compound per mg 4 of adsorbent. Tenax TA was the adsorbent selected for the SPE step. Due to its fast 5 desorption kinetics it is suitable to combine with SPME [8]. Solvent is left to evaporate 6 at room temperature and then, vials are closed and immersed in a water bath at high 7 temperature (100ºC) to favor desorption of the analytes. A SPME fiber is exposed to the 8 headspace over the spiked adsorbent for 15 min. In Figure 2, the chromatographic 9 responses obtained for some representative compounds with each fiber are shown. With 10 PA and PDMS fibers, the extraction efficiency was very poor for all compounds. 11 CAR/PDMS fiber was the most efficient in extracting the two and three chlorine 12 substituted compounds, while PDMS/DVB and CW/DVB fibers provided the most 13 efficient extraction for the semivolatile chlorobenzenes (tetra-, penta- and hexachloro- 14 benzenes). Therefore, CAR/PDMS and PDMS/DVB fibers were considered for further 15 optimization with other experimental parameters using an experimental design 16 approach. CW/DVB was discarded due to the low stability of its coating. 17 To simultaneously optimize the experimental factors and to evaluate the parameters that 18 can mainly affect the mass transfer to the SPME fiber, a factorial design was run [24]. 19 Four factors were studied at two levels (extraction temperature, fiber coating, addition 20 of water to the solid, and stirring of the sample) (see Table 3). The design selected was a 21 multifactor screening 3*2(3-1) mixed level fraction, which involves 12 experiments. 22 Extraction time was fixed at 15 min in all cases. 23 The analysis of the results produced the standardized Pareto charts show in Figure 3. 24 The length of each bar in the graphs is proportional to the absolute value of its 25 9 associated standardized effect. The standardized effect is obtained by dividing the 1 estimated effect of each factor or interaction by its standard error. The effects are 2 displayed in decreasing order of importance, which allows easy identification of the 3 most important factors. Vertical lines indicate the statistical significance of the effects at 4 a confidence level of 95%. A factor is not significant for a particular chlorobenzene 5 when its bar does not reach the critical line [24]. Figure 4 shows the main effect plots 6 for two selected chlorobenzenes. These plots show the main effects with a line drawn 7 between the low and the high level for the corresponding factors. The length of the lines 8 is proportional to the magnitude of the effect of each factor in the microextraction 9 process, and the sign of the slope indicates the level of the factor that produces the 10 highest response. 11 As can be seen in Figures 3 and 4, the temperature of extraction and the type of fiber 12 coating were the most relevant factors for the extraction of all chlorobenzenes. 13 Temperature greatly affects the kinetics of extraction. The Pareto charts in figure 3 14 show that temperature was significant in the extraction of all chlorobenzenes, but for 15 dichlorobenzenes the temperature bar length only slightly exceed the critical value, 16 while the bars length for more chlorinated benzenes exceeds to a great extent the critical 17 value. Figure 4 shows that in all cases the best extraction temperature was the maximum 18 (100ºC). 19 For highly volatile compounds, best results were achieved with CAR/PDMS coating 20 (labeled as low level in Table 3, and in the main effects graph shown in Figure 4), while 21 semivolatile compounds showed higher affinity for the PDMS/DVB fiber (the high 22 level for this factor in Table 3 and in Figure 4). Nevertheless, the responses for 23 semivolatile compounds were lower than the responses for the volatile ones, regardless 24 of the fiber used. As the aim of this work was to develop a method to determine both 25 16 [19] G. Xiong, Y. Chen, J. Pawliszyn, J. Chromatogr. A 999 (2003) 43. 1 [20] J. Namieśnik, A. Jastrzębska, B. Zygmunt, J. Chromatogr. A 1016 (2003) 1. 2 [21] J.A. Koziel, P.A. Martos, J. Pawliszyn, J. Chromatogr. A 1025 (2004) 3. 3 [22] A. Saba, A. Raffaelli, S. Pucci, P. Salvadori, Rapid Commun. Mass Spectrom. 13 4 (1999) 1899. 5 [23] A. Saba, A. Cuzzola, A. Raffaelli, S. Pucci, P. Salvadori, Rapid Commun. Mass 6 Spectrom. 15 (2001) 2404. 7 [24] Statgraphics-Plus, Experimental design, Appendix C, Manugistics Inc., Rockville, 8 Maryland, 1996. 9 [25] D.T. Burns, K. Danzer, A. Townshend, Recommendations for the Use of the Term 10 “Recovery” in Analytical Procedures, IUPAC Recommendations 2001 (see 11 http://www.iupac.org/reports/provisional/ abstract01/burns_prs.pdf). 12 [26] D.W.-M. Sin, Y.-C. Wong,W.-C. Sham, D. Wang, Analyst 126 (2001) 310. 13 14 17 Figure captions 1 2 Figure 1. Schematic diagram of the air-sampling device. 1: vacuum pump, 2: PTFE 3 connectors, 3: flow meter, 4: Tenax TA, 5: glass wool, 6: V-shaped glass tube to contain 4 the analytes in the recovery experiments. 5 6 Figure 2. Comparison of different fiber coatings on the extraction of some 7 representative chlorobenzenes. 8 9 Figure 3. Pareto charts for main effects. Vertical lines indicate the statistical 10 significance of the effects (95% confidence level). 11 12 Figure 4. Graphics showing the influence of main effects on the extraction of two 13 chlorobenzenes: low level (-), high level (+). 14 15 Figure 5. Extraction time profiles obtained in the optimal SPME conditions. 16 17 Figure 6. Variation of the response (expressed as area counts) with the volume of air 18 sampled for two selected chlorobenzenes. 19 20 Figure 7. Variation of the response (expressed as area counts) with the amount of 21 adsorbent for chlorobenzenes. 22 23 Figure 8. Mass chromatograms obtained by a) SPME extraction at the optimal 24 experimental conditions, and b) 2 L injection of a 1 mL n-hexane extraction of the 25 18 target chlorobenzenes adsorbed onto Tenax. The concentration of each compound in the 1 air samples was 40 ng/m3. 2 3 Table 1. Experimental GC-MS and GC-ECD conditions GC-MS Oven temperature program 60ºC (2 min) 100ºC (10ºC/min) 120ºC (3ºC/min) 210ºC (10ºC/min, 2 min)) Injection Splitless mode (2 min) 260-300ºC (depending on the fiber used) Transfer line 260ºC Carrier gas He, 8 psi at 60ºC Ionization mode EI (70eV) Manifold temperature 200ºC GC-ECD Oven temperature program 50ºC (1 min) 250ºC (10ºC/min, 5 min) Injection Splitless mode (1 min) 260ºC Detector temperature 250ºC Carrier gas Nitrogen, 1 mL/min Make-up gas Nitrogen, 40 mL/min Table 2. Quantification ions, mass acquisition ranges, and time segments selected for the determination of each group of chlorobenzenes using an ion trap mass detector. Compound Quantification ions Mass acquisition range (m/z) Time (min) Segment time length (min) DCBs 146 144-148 0.00 to 8.00 8.00 TCBs 180+182 178-184 8.00 to 13.00 5.00 TeCBs 216 214-218 13.00 to 16.50 3.50 PeCB 248+250 246-252 16.50 to 19.00 2.50 HCB 280-288 280-290 19.00 to 22.66 3.66 Table 3. Factors and levels considered in the experimental design. Factor Low level High level Continuous Water (L) 0 300 Yes Temperature (ºC) 50 100 Yes Fiber CAR/PDMS PDMS/DVB No Agitation No Yes No Table 4. Collection efficiency of chlorobenzenes at two concentration levels. 4 ng/m3 40 ng/m3 Apparent recovery (%) %RSD Apparent recovery (%) %RSD 1,3-DCB 86 6 84 5 1,4-DCB 133 4 97 4 1,2-DCB 99 7 77 5 1,2,4-TCB 100 9 118 11 1,2,3-TCB 95 11 97 11 1,2,3,5- TeCB 98 9 88 8 1,2,4,5- TeCB 100 7 89 6 1,2,3,4- TeCB 105 9 82 7 PeCB 126 10 87 9 HCB - - 116 - Table 5. Linearity, repeatability (%RSD), and limits of detection and quantification of the method (see text for more details). Compound Linearity Repeatability (%RSD) Detection limits (S/N=3, ng/m3) Quantification limits (S/N=10, ng/m3) Correlation coefficient (R2) F-test P-value SPME of spiked Tenax (n=6) SPE-SPME (n= 4) 1,3-DCB 0.9973 1.52 0.3256 12 10 0.022 0.062 1,4-DCB 0.9962 1.84 0.2582 13 10 0.012 0.033 1,2-DCB 0.9971 2.10 0.2187 13 12 0.011 0.029 1,2,4-TCB 0.9968 0.66 0.6434 11 9 0.007 0.018 1,2,3-TCB 0.9970 0.56 0.7025 11 9 0.004 0.012 1,2,3,5-TeCB 0.9971 0.34 0.8407 9 11 0.019 0.042 1,2,4,5-TeCB 0.9986 0.21 0.9249 6 10 0.005 0.011 1,2,3,4-TeCB 0.9983 0.64 0.6549 8 11 0.015 0.041 PeCB 0.9920 0.41 0.7989 8 8 0.035 0.075 HCB 0.9980 0.25 0.8987 9 8 0.108 0.238 FIG. 1 1 2 3 4 5 6 air FIG. 2 0 100000 0 200000 0 300000 0 400000 0 500000 0 1,2-DCB 0 50000 0 100000 0 150000 0 200000 0 250000 0 1,2,4,5-TCB 0 100000 0 200000 0 300000 0 400000 0 500000 0 1,2,4-TCB 0 100000 0 200000 0 300000 0 400000 0 500000 0 600000 0 700000 0 PeCB PA CW/DVB PDMS/DV B PDMS CAR/PDM S response (area cts) FIG. 8b HCB PeCB 1,2,3,4-TeCB 1,2,3,5-TeCB + 1,2,4,5-TeCB 1,2,3-TCB 1,2,4-TCB 1,2-DCB 1,4-DCB 1,3-DCB 5 10 15 20 t (min) kCounts 0 1 2 3 4 0 0.5 1 1.5 0 1 2 3 0 2 4 0 1 2 3 Ions: 146+148 Ions: 180+182 Ions: 214+216+218 Ions: 248+250+252 Ions: 284+286+288 Segment 1 Segment 2 Segment 3 Segmt 4 Segment 5