- Email: [email protected]
Two-dimensional gas chromatography with electron capture detection for the analysis of atmospheric ozone depleting halocarbons
Accepted Manuscript Title: Two-dimensional Gas Chromatography with Electron Capture Detection for the Analysis of Atmospheric Ozone Depleting Halocarbons Authors: Chang-Feng Ou-Yang, Hsi-Che Hua, Yu-Chieh Chou, Ming-Kai Teng, Wen-Tzu Liu, Jia-Lin Wang PII: DOI: Reference:
S0021-9673(17)30522-8 http://dx.doi.org/doi:10.1016/j.chroma.2017.04.003 CHROMA 358436
To appear in:
Journal of Chromatography A
Received date: Revised date: Accepted date:
31-10-2016 27-3-2017 2-4-2017
Please cite this article as: Chang-Feng Ou-Yang, Hsi-Che Hua, Yu-Chieh Chou, Ming-Kai Teng, Wen-Tzu Liu, Jia-Lin Wang, Two-dimensional Gas Chromatography with Electron Capture Detection for the Analysis of Atmospheric Ozone Depleting Halocarbons, Journal of Chromatography Ahttp://dx.doi.org/10.1016/j.chroma.2017.04.003 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Two-Dimensional Gas Chromatography with Electron Capture Detection for the Analysis of Atmospheric Ozone Depleting Halocarbons Chang-Feng Ou-Yang1, Hsi-Che Hua2, Yu-Chieh Chou2, Ming-Kai Teng2, Wen-Tzu Liu2, JiaLin Wang2* 1
Department of Atmospheric Sciences, National Central University, Taoyuan 32001, Taiwan
2
Department of Chemistry, National Central University, Taoyuan 32001, Taiwan
*
Corresponding to: Jia-Lin Wang
Department of Chemistry, National Central University, Taoyuan 32001, Taiwan Tel: +886-3-4227151 ext.65906 Fax: +886-3-4277972 E-mail: [email protected] Highlights
» ODS and halocarbons at ambient level were analyzed by the TD-GC×GC-ECD method.
» The TD technique was coupled to GC×GC to in-line enrich atmospheric halocarbons.
» TD-GC×GC-ECD produces simplified results due to the high selectivity of ECD.
Abstract This study is to develop a GC×GC method with electron capture detection (ECD) to analyze atmospheric halocarbons in the concentration range of parts per trillion by volume (pptv). To enrich atmospheric halocarbons a home-built thermal desorption (TD) device was 1
coupled to the GC×GC-ECD. The technique of flow modulation was adopted using a Deans switch for GC×GC. Several column combinations of first and second dimensions were tested and the column set of DB-5×TG-1301 was found to show the best orthogonality for halocarbons. A series of modulation parameters were tested for their optimal settings. The modulation period (PM) was found to have minimal wrap-around when set at 3 s. The modulation ratio (MR) was determined to be 7.82 to ensure reproducible results and maximum sensitivity. The modulation duty cycle (DC) was calculated to be approximately 0.17. Nine halocarbons were separated successfully and seven were calibrated with the use of a certified standard gas mixture. The correlation coefficients (R2) were greater than 0.9972. The reproducibility was better than 1.90% as expressed in relative standard deviation (RSD; N=30) and the detection limits were in the range of pptv for the target halocarbons. A field test by continuous analyzing ambient air with hourly resolution was performed to show the stability of the method as suggested by the homogeneity of certain halocarbons, while also reflecting concentration variation for others when emissions did arise. Keywords: Deans switch; CFCs; ozone depletion substance; thermal desorption; twodimensional gas chromatography; electron capture detector.
1. Introduction Man-made halogenated trace gases are blamed for the stratospheric ozone depletion [1,2]. Due to the extremely long lifetimes of some selected halocarbons in the atmosphere, these compounds can drift to the stratosphere threatening the ozone layer by releasing chlorine or bromine atoms through photolysis [3]. The Montreal Protocol and its subsequent amendments were thus enacted to cease the production and consumption of anthropogenic chlorofluorocarbons (CFCs) and selected halocarbons [4,5]. Tropospheric abundances and emissions of most ozone depleting substances (ODS) hence started to decrease since mid-90s, resulting in the declines in the atmospheric burden of chlorine and bromine [2,6-8]. Although 2
the threat of the CFCs on ozone depletion has been greatly alleviated, the net radiative forcing of the halocarbons as greenhouse gases is still considerable since they still contribute up to 11.5% of the net radiative forcing with the inclusion of the CFC replacements of hydrochlorofluorocarbons and hydrofluorocarbons [9-13]. GC-ECD or GC-MS equipped with an enrichment device has been a common method to analyze ambient halocarbons [14-18]. The enrichment can be made either by using cryogens (e.g., liquid nitrogen) [19] or chemical sorbents with thermal desorption (TD) [16]. As more volatile compounds are invented and emitted into the atmosphere, the complexity of air composition increases over time and; thus, the analytical techniques with higher separation resolution become ever more in demand. The conventional GC techniques are often faced with limited separation efficiency that may not be adequate in the modern era when multidimensional separation becomes more assessable and common. Though GC-MS is considered as a multi-dimensional technique in concept with the retention time as the first dimension (1D) and the mass spectrum as the second dimension (2-D), its analytical power may still be hindered by insufficient GC separation when encountering complex samples. Given the full recognition of the achievements made by the conventional GC-ECD or GC-MS in analyzing trace-level atmospheric halocarbons in the past, the comprehensive GC×GC techniques can take the conventional halocarbon analysis to the next level by providing enhanced GC separation efficiency. Previous work by Wang et al. has demonstrated a cryogenic GC×GC-MS method for 30 atmospheric halocarbons [20]. Rather than using thermal modulation with a cryogen, flow modulation using a Dean switch was adopted for hardware simplicity and low cost in construction and operation. However, compared to the thermal modulation counterpart, flow modulation sacrifices some sensitivity and resolution, since minimal or no sample focusing is resulted from modulation [21]. Other than the cost and complexity issues, thermal modulation 3
is not without drawbacks. For instance, it has potential issues with the break-through of volatile analytes [22]. This study aims at developing a comprehensive TD-GC×GC-ECD method to analyze halogenated compounds at ambient level. ECD was employed as the detection method which is known to be extremely sensitive and rugged for highly halogenated compounds. Moreover, the high selectivity of ECD towards halocarbons can largely eliminate interferences from other air-borne volatile compounds bearing no halogen atoms to achieve rather simplified chromatography. Connecting TD-GC×GC to a more elaborate detection method such as timeof-flight although could greatly extend the analyte coverage; the simplicity in chromatography would be lost. The sampling rate of ECD of 50 Hz makes it sufficiently fast for the flow modulated TD-GC×GC applications. To our best knowledge, no or limited GC×GC work has been given to the ODS analysis and only few studies have shown the potential of the applications of GC×GC-ECD for the persistent organic pollutants such as pesticides residues, dioxins, chlorobenzenes, polychlorinated biphenyls, and fluorinated polycyclic aromatic hydrocarbons [23-27]. As a proof of concept, an analytical system of flow-modulated TDGC×GC-ECD was built in-house to demonstrate the feasibility of such an application. Tracelevel halocarbons of significant atmospheric importance, which have been monitored by one dimensional GC with a long history [14-19,28,29], can be readily analyzed via TD-GC×GCECD and viewed from a two-dimensional perspective.
2. Methodology 2.1. Halocarbon identification and quantification An aluminum AL-150 type cylinder (Spectra Gases, Branchburg, NJ, USA) filled with compressed ambient air was employed as the working standard and air sample for testing the analytical method. In this study, 9 halocarbons, CCl2Cl2 (CFC-12), CCl3F (CFC-11), methylene chloride (CH2Cl2), CCl2FCClF2 (CFC-113), chloroform (CHCl3), methyl chloroform 4
(CH3CCl3),
carbon
tetrachloride
(CCl4),
trichloroethylene
(CClH=CCl2),
and
tetrachloroethylene (CCl2=CCl2) were targeted due to their well-characterized properties, such as homogeneity, ubiquity, and the high sensitivity to ECD in detecting them [16], make them ideal targets to test the method. The concentrations of these compounds ranged from few pptv to sub-ppbv, which were well above the detection limits of ECD with the use of the TD unit in the system for in-line enrichment [16]. However, ECD is less sensitive for some hydrogencontaining halocarbons at ambient level ranging from few ppt to sub-ppbv even with the use of a pre-concentrator. The oxygen-doping technique by adding a small flow of zero air in the make-up gas of ultra-pure nitrogen can increase the sensitivity of ECD for the hydrogencontaining halocarbons (e.g., CH2Cl2 and CHCl3 in this study) [29,30]. The mixing was conducted by exploiting the electronic pressure controllers built in the GC to maintain very constant flows of nitrogen and air to attain the desired mixing ratio of 0.2% for oxygen in nitrogen. Compound identification was made by spiking known compounds in both the 1-D and GC×GC analysis, and the retention times can correlate rather well in the x - coordinate of both methods. In addition, since peaks with GC-ECD are usually very characteristic for the target halocarbons in ambient air, hence identification with spiking, assisted with characteristic fingerprinting, made identification relatively straightforward. Concentrations of the 6 target halocarbons (CFC-12, CFC-11, CH2Cl2, CFC-113, CH3CCl3, and CCl4) in ambient air were calibrated with a pressurized cylinder to serve as the working standard (or secondary standard). This working standard was in term calibrated with a standard cylinder purchased from National Oceanic and Atmospheric Administration/Global Monitoring Division in the US which is traceable to National Institute of Standards and Technology to serve as our primary standard. Calibration of CHCl3 was made by an US EPA type standard, i.e. TO-15. The concentrations of CFC-12, CFC-11, CH2Cl2, CFC-113, CHCl3, CH3CCl3, and CCl4 in the working standard 5
were calibrated to be 566.1, 247.9, 687.7, 75.2, 22.6, 5.7, and 94.6 pptv, respectively. Because no quality standards were available for tri- and tetrachloroethylene at the time of experiment, the concentration calibration was not possible and only compound identification was performed.
2.2. Sample pre-concentration Fig. 1 illustrates the apparatus of the analytical system used in this study. A TD unit was built in-house with multiport switching valves (A26UWE, Valco Instruments Co. Inc., USA) to guide sample flows during sample trapping and injection. A glass tube (10 cm × 2 mm I.D. × 1/8” O.D.) packed with carbon sorbent beds of Carboxen 1000, Carboxen 1003, and Carbotrap formed the sorbent trap. Each bed was about 1 cm in length isolated with glass wools. The air sample was drawn by an oil-free pump through the sorbent trap thermostated at 30°C for 5 min or a desired time interval. The sample flow was controlled by a mass flow controller at a constant flow rate of 50 mL min-1. Injection was then performed by flash heating from 30°C to 250°C within seconds, followed by back-flushing the sorbent trap kept at 300°C for 30 min for conditioning. All the tubing used in the TD unit is made of Silcosteel (1/16” O.D., 0.8 mm I.D., Restek, USA). Further detailed information of the TD unit can be referred to our previous works [16,31].
2.3. Heart-cut gas chromatography A GC (6890N, Agilent, USA) equipped with a micro-ECD thermostated at 300°C was used. A Deans switch device comprising three micro tee unions (MT1CS6, Valco Instruments Co. Inc., USA) connected with short sections of Silcosteel tubes (3 cm × 1/16” O.D. × 0.1 mm I.D., Restek, USA) served as the valve-based flow modulator in this work. A three-way solenoid valve (091-0094-900, Parker, USA) was connected to the two ends of the micro tee unions to control the flow direction for modulation.
6
A series of column combinations for the target halocarbons were tested. A non-polar column (either DB-1 or DB-5) with a typical length of 60 m was chosen to be the 1-D column throughout the 2D optimization process. Several columns of large difference in polarity and absorptivity from the 1-D column were systematically tested for the most desired separation efficiency and orthogonality for the halocarbons. Columns of various properties such as polarity, phase types (WCOT vs. PLOT) and dimensions were tested as the 2-D column. The pressures of the auxiliary and carrier gas were setup according to the conditions of different column combinations. The oven temperature program was initially set at 35°C for 10 min, then ramped at 5°C min-1 to 80°C, and finally ramped at 20°C to 200°C for 10 min. The detector signal were further processed and illustrated in 2-D contour plots using a commercially available software Surfer 8 (Golden Software, CO, USA).
3. Results and Discussion 3.1. Selection of GC×GC columns The selection of the 1-D and 2-D columns plays the central role in GC×GC. As illustrated in Fig. 2, a series of column combinations of various phases were tested for the target halocarbons. The separation of the 1-D non-polar column was generally based on the volatilities of the analytes. Several columns of large difference in polarity and absorptivity from the 1-D column were tested for the most desirable separation efficiency and orthogonality for the halocarbons. While keeping the non-polar 1-D column at a fixed length of 60 m throughout the test for the optimal 1-D separation, the 2-D columns of different phases varied in length from 0.3 m to 2 m and inner diameter varied from 0.18 mm to 0.32 mm were tested to yield the best separation, orthogonality and, at the same time, minimal wrap-around. Porous layer open tubular columns such as PLOT and Gaspro, which are generally used for separating compounds of high volatilities close to those of the target halocarbons, were first tested to exploit the high retention forces or absorptivity [32,33]. However, they both showed either severe peak tailing 7
or wrap-around problems working as the 2-D columns due to excessive retention (Figs. 2a and 2b). Three medium-polar wall-coated open tubular (WCOT) columns, i.e. Rti-200, DB-1701, and TG-1301, were also tested for suitability as the 2-D columns. These columns did not give rise to ideal orthogonality for the target compounds by coupling to DB-1 as the 2-D columns (Figs. 2c, 2d, and 2e), though the TG-1301 column showed a slightly better 2-D resolution among all the tested WCOT columns, since the peaks were more focused (Fig. 2e). Subsequently, as an alternative to DB-1 as the 1-D column, a DB-5 column of a smaller pore size was coupled to the TG-1301 column and resulted in a much improved result, possibly due to a better match in polarity and thus more apparent orthogonality (Fig. 2f). The chosen column set (DB-5×TG-1301) was then further optimized for the modulation properties of GC×GC as described below.
3.2. Optimization of GC and modulation parameters Because the Deans-switch was used as the flow modulator, the pressure of the auxiliary flow controls the rapid swings between the actions of sample loading and injection for modulation (Fig. 1). The pressure of the auxiliary flow ranging from 7-10 psi was first tested for the optimized setting. The elution of the analytes in the 2-D column needs to be rapid enough to avoid tailed or diffused peaks (Fig. S1a). However, if the pressure is set too high, the auxiliary gas would tend to impede the effluents from the 1-D column, causing less analytes to be sent into the 2-D column (Fig. S1c). Furthermore, serious co-elution could occur if the auxiliary flow pressure continues to increase (Fig. S1d). As a compromise, the auxiliary pressure was thus set at the optimal pressure of 7.5 psi (Fig. S1b). The sample loading time stands for the duration of the air sample band being sent to the sampling line for loading sample as indicated in Fig. 1 (3 cm in length), which only applies to the flow modulation GC×GC instead of the thermal modulation type, since no cryo-focusing is performed. The sample amount increases as the sample loading time increases. However, up to 8
a certain point, the filling will be complete and the excess sample will be vented. The optimal loading time is determined when the peak response is no longer increased. Various loading times ranging from 0.1 s to 0.7 s were tested (Fig. S2). As the loading time increased between 0 and 0.5 s (Figs. S2a-c), the peak intensity also increased accordingly. However, after 0.5 s the increase in intensity was no longer apparent (Figs. S2c and d), signaling an overflow of the sample. As a result, the optimized loading time was set at 0.5 s. There are three key common parameters that describe the modulation conditions of a GC×GC task for both the thermal and flow modulation types: the modulation ratio (MR), the modulation period (PM) and the modulation duty cycle (DC). MR is the ratio of the width of the 1-D peak to the PM, as defined by Khummueng et al. [34], which determines the reproducibility of the GC×GC results. For instance, MR = 1.5 can only suffice semi-quantitative analysis [34]. MR should be greater than 2.5 to permit satisfactory reproducibility of the transferred effluents (RSD ∼ 1%) [35], and MR = 3 should be aimed at if trace level analysis with high precision is desired [34]. As a result, for our ambient-level analysis, making MR greater than 3 so that each component produces 3-4 pulses is our minimal target to obtain precise results. However, this attempt is perplexed by choosing an appropriate PM value and, at the same time, avoiding wraparound. PM is the elapsed time for each transfer of a sample slice from the 1-D column to the 2-D column. Missing peaks and poor sensitivities can happen if the modulation period is too long. Conversely, if the modulation period is set too short, it is possible that wrap-around can occur. To inspect for wrap-around during the process of PM determination, a series of chromatograms were aligned back-to-back by repeated analyzing the standard air sample (Fig. S3). It was found that when the PM was set as 7 s, both criteria of MR > 3 (= 3.35) and the wrap-around avoidance was achieved. However, the halocarbons at trace ambient level such as CH2Cl2, CH3CCl3, and
9
CCl2=CCl2 became quantitatively undetectable in the GC×GC results if the PM was set too large. Therefore, PM was set as 3 s to achieve MR = 7.82 as a compromised condition. DC is defined as the fraction of PM during which the 1-D effluent is sampled by the modulation [35]. The DC value was determined to be approximately 0.3 in our final optimized GC×GC settings. We acknowledged that this value was rather low compared to that of the typical thermal modulation which has the DC value close to unity. In our case, DC was sacrificed largely for peak resolution. The high enrichment factor of an order of three with the TD device greatly compensated for the lost sample. Meanwhile, the inherited high sensitivity of ECD also tolerated greatly such a sample loss. After all the modulation parameters were determined and set, the GC×GC was then optimized for the GC temperature programing. Fig. 3 displays the typical chromatograms from analyzing 250 mL of ambient air resulted from both the conventional 1-D TD-GC-ECD method and the comprehensive TD-GC×GC-ECD method for comparison.
3.3. System validation System validation from the aspects of precision, detection limits and linearity was evaluated for the TD-GC×GC-ECD system. The linearity was assessed by varying the amount of the trapped sample from 100 to 400 mL and five volumes were injected to test for linearity. The 7 target halocarbons showed the correlation coefficients (R2) ranging from 0.9969 for CFC12 to 0.9991 for CFC-11 (Fig. S4). These volume-based curves were then converted into calibration curves by equating the injection volumes to concentrations. Since the regular trapping time is five minutes at the flow rate of 50 mL min-1, which is our standard trapping conditions for routine analysis, any varied trapping time at 50 mL min-1 can correspond to different injection volumes and thus be equivalent to different concentrations since analytes were trapped and air matrix escaped. For instance, trapping two minutes of the same air would
10
be equivalent to 100 mL of the standard sample and thus 40% of the concentration as of the standard condition. The RSD (relative standard deviation) of 30 repeated aliquots (N = 30) of the working standard air sample ranged from 0.81% for CH3CCl3 to 1.90% for CFC-113. The method detection limits based on the precision for the 100 mL aliquots were in the range of 8-24 pptv. Table 1 summarizes the all QA results for the target halocarbons.
3.4. Field test The TD-GC×GC-ECD system was tested by continuous hourly measurements of ambient air for 2 days with the air intake extruding to outside the laboratory. The standard volume of 250 mL was analyzed for each hourly measurement (Fig. 4). Most halocarbons, i.e. CFC-12, CCl4, CFC-113, CH3CCl3, and CH2Cl2, showed steady concentrations with small variability, suggesting negligible emissions and that their presence in the atmosphere is the result of prior emissions and long lifetimes. By contrast, the levels of CFC-11 and CHCl3 elevated on 29th December, which might be caused by the nearby sources in the surrounding buildings. The field test suggests that the method is stable to perform automated continuous measurements, as indicated by the fairly constant levels for halocarbons such as CFC-113 and CCl4 [36,37], while also reflecting concentration variation for compounds such as CFC-11 and CHCl3 when emissions did arise.
4. Conclusions In this study, we have presented a TD-GC×GC-ECD method to measure atmospheric halocarbons using components that are widely available to conventional 1-D GC users. It was found that the column combination of DB-5 as the 1-D column and TG-1301 as the 2-D column offered the best orthogonality among four column combinations of various polarity pairs. A 11
series of modulation parameters were tested for GC×GC. The PM was set at 3 s to maximize MR, which was determined to be 7.82 to ensure high reproducibility and sensitivity. The DC was approximately 0.17, which although is low compared with that of the thermal modulation counterpart, it was greatly compensated by the in-line TD enrichment and the high sensitivity of ECD. Quality assurance including linearity, precision, and detection limits were assessed for the proposed method of TD-GC×GC-ECD to ensure robust measurement of halocarbons at ambient level. Since very limited GC×GC work in analyzing atmospheric ODS and volatile halocarbons has been reported in the literature, the ever increased availability of the GC×GC technique offers the atmospheric halocarbon analysis, which used more conventional GC-ECD or GCMS technique, an entirely new perspective. Acknowledgements This study was finically supported by the Ministry of Science and Technology (formerly National Science Council), Taiwan, through the contract no. NSC 102-2113-M-008002-MY3. References
[1]
S.A. Montzka, S. Reimann, A. Engel, K. Krüger, S. O'Doherty, W.T. Sturges, D. Blake, M. Dorf, P. Fraser, L. Froidevaux, K. Jucks, K. Kreher, M.J. Kurylo, A. Mellouki, J. Miller, O.J. Nielsen, V.L. Orkin, R.G. Prinn, R. Rhew, M.L. Santee, A. Stohl, D. Verdonik, Chapter 1 Ozone-Depleting Substances (ODSs) and Related Chemicals, Scientific Assessment of Ozone Depletion: 2010, Global Ozone Research and Monitoring Project - Report No.52, World Meteorological Organization, Geneva, Switzerland, 2011.
[2]
L.J. Carpenter, S. Reimann, J.B. Burkholder, C. Clerbaux, B.D. Hall, R. Hossaini, J.C. Laube, S.A. Yvon-Lewis, Update on Ozone-Depleting Substances (ODSs) and Other Gases of Interest to the Montreal Protocol, Scientific Assessment of Ozone Depletion: 2014, Global Ozone Research and 12
Monitoring Project - Report No.55, World Meteorological Organization, Geneva, 2014. [3]
M.J. Molina, F.S. Rowland, Stratospheric sink for chlorofluoromethanes: chlorine atom-catalysed destruction of ozone, Nature 249 (1974) 810-812.
[4]
UNEP, Montreal Protocol on Substances that Deplete the Ozone Layer: Final Act, Montreal, 1987.
[5]
UNEP, The Copenhagen Amendment (1992): The amendment to the Montreal Protocol agreed by the Fourth Meeting of the Parties, Copenhagen, 1992.
[6]
R.G. Prinn, J. Huang, R.F. Weiss, D.M. Cunnold, P.J. Fraser, P.G. Simmonds, A. McCulloch, C. Harth, S. Reimann, P. Salameh, S. O'Doherty, R.H.J. Wang, L.W. Porter, B.R. Miller, P.B. Krummel, Evidence for variability of atmospheric hydroxyl radicals over the past quarter century, Geophys. Res. Lett. 32 (2005) doi:10.1029/2004gl022228.
[7]
M. Rigby, R.G. Prinn, S. O'Doherty, S.A. Montzka, A. McCulloch, C.M. Harth, J. Muhle, P.K. Salameh, R.F. Weiss, D. Young, P.G. Simmonds, B.D. Hall, G.S. Dutton, D. Nance, D.J. Mondeel, J.W. Elkins, P.B. Krummel, L.P. Steele, P.J. Fraser, Re-evaluation of the lifetimes of the major CFCs and CH3CCl3 using atmospheric trends, Atmos. Chem. Phys. 13 (2013) 26912702.
[8]
S.A. Montzka, J.H. Butler, B.D. Hall, D.J. Mondeel, J.W. Elkins, A decline in tropospheric organic bromine, Geophys. Res. Lett. 30 (2003) doi:10.1029/2003gl017745.
[9]
M. Maione, U. Giostra, J. Arduini, F. Furlani, F. Graziosi, E. Lo Vullo, P. Bonasoni, Ten years of continuous observations of stratospheric ozone depleting gases at Monte Cimone (Italy) - Comments on the effectiveness of the Montreal Protocol from a regional perspective, Sci. Total Environ. 445 (2013) 155-164.
[10]
M.K. Vollmer, J. Muhle, C.M. Trudinger, M. Rigby, S.A. Montzka, C.M. Harth, B.R. Miller, S. Henne, P.B. Krummel, B.D. Hall, D. Young, J. Kim, J. Arduini, A. Wenger, B. Yao, S. Reimann, S. O'Doherty, M. Maione, D.M. Etheridge, S.L. Li, D.P. Verdonik, S. Park, G. Dutton, L.P. Steele, C.R. Lunder, T.S. Rhee, O. Hermansen, N. Schmidbauer, R.H.J. Wang, M. Hill, P.K. Salameh, R.L. Langenfelds, L.X. Zhou, T. Blunier, J. Schwander, J.W. Elkins, J.H. Butler, P.G. Simmonds, R.F. Weiss, R.G. Prinn, P.J. Fraser, Atmospheric 13
histories and global emissions of halons H-1211 (CBrClF2), H-1301 (CBrF3), and H-2402 (CBrF2CBrF2), J. Geophys. Res. 121 (2016) 3663-3686. [11]
S.A. Montzka, B.D. Hall, J.W. Elkins, Accelerated increases observed for hydrochlorofluorocarbons since 2004 in the global atmosphere, Geophys. Res. Lett. 36 (2009) doi:10.1029/2008gl036475.
[12]
M. Rigby, R.G. Prinn, S. O'Doherty, B.R. Miller, D. Ivy, J. Muhle, C.M. Harth, P.K. Salameh, T. Arnold, R.F. Weiss, P.B. Krummel, L.P. Steele, P.J. Fraser, D. Young, P.G. Simmonds, Recent and future trends in synthetic greenhouse gas radiative forcing, Geophys. Res. Lett. 41 (2014) 2623-2630.
[13]
J.H. Butler, S.A. Montzka, in, http://www.esrl.noaa.gov/gmd/aggi/aggi.html, 2016.
[14]
S.A. Montzka, R.C. Myers, J.H. Butler, J.W. Elkins, S.O. Cummings, Global tropospheric distribution and calibration scale of HCFC-22, Geophys. Res. Lett. 20 (1993) 703-706.
[15]
P.G. Simmonds, S. Odoherty, G. Nickless, G.A. Sturrock, R. Swaby, P. Knight, J. Ricketts, G. Woffendin, R. Smith, Automated gas-chromatograph massspectrometer for routine atmospheric field-measurements of the CFC replacement compounds, the hydrofluorocarbons and hydrochlorofluorocarbons, Anal. Chem. 67 (1995) 717-723.
[16]
J.L. Wang, C.H. Wu, Construction and validation of a cryogen free gas chromatography-electron-capture detection system for the measurement of ambient halocarbons, Anal. Chim. Acta 461 (2002) 85-95.
[17]
B.R. Miller, R.F. Weiss, P.K. Salameh, T. Tanhua, B.R. Greally, J. Muhle, P.G. Simmonds, Medusa: A sample preconcentration and GC/MS detector system for in situ measurements of atmospheric trace halocarbons, hydrocarbons, and sulfur compounds, Anal. Chem. 80 (2008) 1536-1545.
[18]
C.H. Wang, S.W. Chiang, J.L. Wang, Simultaneous analysis of atmospheric halocarbons and non-methane hydrocarbons using two-dimensional gas chromatography, J. Chromatogr. A 1217 (2010) 353-358.
[19]
J.L. Wang, C.J. Chang, W.D. Chang, C. Chew, S.W. Chen, Construction and evaluation of automated gas chromatography for the measurement of anthropogenic halocarbons in the atmosphere, J. Chromatogr. A 844 (1999) 259-269. 14
[20]
Y. Wang, X.B. Xu, L.Y. Yin, H.B. Cheng, T. Mao, K.P. Zhang, W.L. Lin, Z.Y. Meng, J.A. Palasota, Coupling of comprehensive two-dimensional gas chromatography with quadrupole mass spectrometry: Application to the identification of atmospheric volatile organic compounds., J. Chromatogr. A 1361 (2014) 229-239.
[21]
J.V. Seeley, Recent advances in flow-controlled multidimensional gas chromatography, J. Chromatogr. A 1255 (2012) 24-37.
[22]
R.B. Gaines, G.S. Frysinger, Temperature requirements for thermal modulation in comprehensive two-dimensional gas chromatography, J. Sep. Sci. 27 (2004) 380-388.
[23]
E.M. Kristenson, P. Korytar, C. Danielsson, M. Kallio, M. Brandt, J. Makela, Evaluation of modulators and electron-capture detectors for comprehensive two-dimensional GC of halogenated organic compounds, J. Chromatogr. A 1019 (2003) 65-77.
[24]
A.M. Muscalu, E.J. Reiner, S.N. Liss, T. Chen, G. Ladwig, D. Morse, A routine accredited method for the analysis of polychlorinated biphenyls, organochlorine pesticides, chlorobenzenes and screening of other halogenated organics in soil, sediment and sludge by GCxGC-μECD, Anal. Bioanal. Chem. 401 (2011) 2403-2413.
[25]
J.M. da Silva, C.A. Zini, E.B. Caramao, Evaluation of comprehensive twodimensional gas chromatography with micro-electron capture detection for the analysis of seven pesticides in sediment samples, J. Chromatogr. A 1218 (2011) 3166-3172.
[26]
P. Haglund, P. Korytar, C. Danielsson, J. Diaz, K. Wiberg, P. Leonards, U.A.T. Brinkman, J. de Boer, GC×GC-ECD: a promising method for the determination of dioxins and dioxin-like PCBs in food and feed, Anal. Bioanal. Chem. 390 (2008) 1815-1827.
[27]
A.M. Muscalu, E.J. Reiner, S.N. Liss, T. Chen, Determination of polychlorinated biphenyls, organochlorine pesticides, chlorobenzenes in sludge and sediment samples by GC × GC-μECD, Int. J. Environ. Anal. Chem. 90 (2010) 1-13.
[28]
J.W. Elkins, D.W. Fahey, J.M. Gilligan, G.S. Dutton, T.J. Baring, C.M. Volk, R.E. Dunn, R.C. Myers, S.A. Montzka, P.R. Wamsley, A.H. Hayden, J.H. Butler, T.M. Thompson, T.H. Swanson, E.J. Dlugokencky, P.C. Novelli, D.F. 15
Hurst, J.M. Lobert, S.J. Ciciora, R.J. McLaughlin, T.L. Thompson, R.H. Winkler, P.J. Fraser, L.P. Steele, M.P. Lucarelli, Airborne gas chromatograph for in situ measurements of long-lived species in the upper troposphere and lower stratosphere, Geophys. Res. Lett. 23 (1996) 347-350. [29]
F.L. Moore, J.W. Elkins, E.A. Ray, G.S. Dutton, R.E. Dunn, D.W. Fahey, R.J. McLaughlin, T.L. Thompson, P.A. Romashkin, D.F. Hurst, P.R. Wamsley, Balloonborne in situ gas chromatograph for measurements in the troposphere and stratosphere, J. Geophys. Res. 108 (2003) doi:10.1029/2001jd000891.
[30]
E.P. Grimsrud, D.A. Miller, Oxygen doping of carrier gas in measurement of halogenated methanes by gas chromatography with electron capture detection, Anal. Chem. 50 (1978) 1141-1145.
[31]
C.F. Ou-Yang, W.C. Liao, P.C. Wang, G.J. Fan, C.C. Hsiao, M.T. Chuang, C.C. Chang, N.H. Lin, J.L. Wang, Construction of a cryogen-free thermal desorption gas chromatographic system with off-the-shelf components for monitoring ambient volatile organic compounds, J. Sep. Sci. 39 (2016) 14891499.
[32]
J.L. Wang, G.Z. Din, C.C. Chan, Validation of a laboratory-constructed automated gas chromatograph for the measurement of ozone precursors through comparison with a commercial analogy, J. Chromatogr. A 1027 (2004) 11-18.
[33]
W.C. Liao, C.F. Ou-Yang, C.H. Wang, C.C. Chang, J.L. Wang, Twodimensional gas chromatographic analysis of ambient light hydrocarbons, J. Chromatogr. A 1294 (2013) 122-129.
[34]
W. Khummueng, J. Harynuk, P.J. Marriott, Modulation ratio in comprehensive two-dimensional gas chromatography, Anal. Chem. 78 (2006) 4578-4587.
[35]
J.V. Seeley, N.J. Micyus, S.V. Bandurski, S.K. Seeley, J.D. McCurry, Microfluidic Deans switch for comprehensive two-dimensional gas chromatography, Anal. Chem. 79 (2007) 1840-1847.
[36]
C.C. Chang, G.G. Lo, C.H. Tsai, J.L. Wang, Concentration variability of halocarbons over an electronics industrial park and its implication in compliance with the Montreal Protocol, Environ. Sci. Technol. 35 (2001) 3273-3279.
[37]
C.F. Ou-Yang, C.C. Chang, S.P. Chen, C. Chew, B.R. Lee, C.Y. Chang, S.A. Montzka, G.S. Dutton, J.H. Butler, J.W. Elkins, J.L. Wang, Changes in the 16
levels and variability of halocarbons and the compliance with the Montreal Protocol from an urban view, Chemosphere 138 (2015) 438-446. Figure Captions Fig. 1. Schematics of the TD-GC×GC-ECD system using Deans switch as the modulator. Effluents from 1-D column are directed to (a) 2-D column for further separation (b) load sample through the loading section and discard the excess. Fig. 2. GC×GC separation of ambient halocarbons from six column combinations. (a) DB-1 (60 m× 0.32 mm × 0.25 μm) × PLOT (0.4 m × 0.32 mm × 0.8 μm), (b) DB-1 (60 m× 0.32 mm × 0.25 μm) × Gaspro (0.3 m × 0.32 mm), (c) DB-1 (60 m× 0.32 mm × 0.25 μm) × Rxi-200 (1 m × 0.18 mm × 0.2 μm), (d) DB-1 (60 m× 0.32 mm × 0.25 μm)× DB-1701 (0.35 m × 0.25 mm × 0.5 μm), (e) DB-1 (60 m× 0.32 mm × 0.25 μm)× TG-1301 (2 m × 0.25 mm × 0.5 μm), and (f) DB-5 (60 m × 0.32 mm × 0.25 μm) × TG-1301 (2 m × 0.25 mm × 0.5 μm). Fig. 3. A typical 1-D GC-ECD and comprehensive TD-GC×GC-ECD chromatograms from analyzing 250 mL of ambient air. The 1-D chromatogram in red illustrates the chromatogram of zero air spiked with trichloroethylene for compound identification. Fig. 4. Time-series hourly data of selected halocarbons from the campus measurements. ECD response rather than concentration is used as Y-coordinate to show better contrast in variability. The plot of concentration vs. time is shown in Fig. S5.
17
Figr-1
Fig. 1.
18
19
Fig. 2.
Fig. 3.
20
Fig. 4.
21
Table 1 Quality assurance results
Compound CFC-12 CFC-11 CH2Cl2 CFC-113 CHCl3 CH3CCl3 CCl4 a
R2 (5 points) 0.9969 0.9991 0.9987 0.9979 0.9982 0.9972 0.9988
RSD (N=30) 1.22% 1.39% 1.88% 1.90% 1.44% 0.81% 1.49%
Method Detection Limit (ppt)a
The method detection limits are calculated based on 3σ of the 100mL aliquots.
22
24.9 4.5 5.4 9.9 5.6 2.9 2.0
S0021-9673(17)30522-8 http://dx.doi.org/doi:10.1016/j.chroma.2017.04.003 CHROMA 358436
To appear in:
Journal of Chromatography A
Received date: Revised date: Accepted date:
31-10-2016 27-3-2017 2-4-2017
Please cite this article as: Chang-Feng Ou-Yang, Hsi-Che Hua, Yu-Chieh Chou, Ming-Kai Teng, Wen-Tzu Liu, Jia-Lin Wang, Two-dimensional Gas Chromatography with Electron Capture Detection for the Analysis of Atmospheric Ozone Depleting Halocarbons, Journal of Chromatography Ahttp://dx.doi.org/10.1016/j.chroma.2017.04.003 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Two-Dimensional Gas Chromatography with Electron Capture Detection for the Analysis of Atmospheric Ozone Depleting Halocarbons Chang-Feng Ou-Yang1, Hsi-Che Hua2, Yu-Chieh Chou2, Ming-Kai Teng2, Wen-Tzu Liu2, JiaLin Wang2* 1
Department of Atmospheric Sciences, National Central University, Taoyuan 32001, Taiwan
2
Department of Chemistry, National Central University, Taoyuan 32001, Taiwan
*
Corresponding to: Jia-Lin Wang
Department of Chemistry, National Central University, Taoyuan 32001, Taiwan Tel: +886-3-4227151 ext.65906 Fax: +886-3-4277972 E-mail: [email protected] Highlights
» ODS and halocarbons at ambient level were analyzed by the TD-GC×GC-ECD method.
» The TD technique was coupled to GC×GC to in-line enrich atmospheric halocarbons.
» TD-GC×GC-ECD produces simplified results due to the high selectivity of ECD.
Abstract This study is to develop a GC×GC method with electron capture detection (ECD) to analyze atmospheric halocarbons in the concentration range of parts per trillion by volume (pptv). To enrich atmospheric halocarbons a home-built thermal desorption (TD) device was 1
coupled to the GC×GC-ECD. The technique of flow modulation was adopted using a Deans switch for GC×GC. Several column combinations of first and second dimensions were tested and the column set of DB-5×TG-1301 was found to show the best orthogonality for halocarbons. A series of modulation parameters were tested for their optimal settings. The modulation period (PM) was found to have minimal wrap-around when set at 3 s. The modulation ratio (MR) was determined to be 7.82 to ensure reproducible results and maximum sensitivity. The modulation duty cycle (DC) was calculated to be approximately 0.17. Nine halocarbons were separated successfully and seven were calibrated with the use of a certified standard gas mixture. The correlation coefficients (R2) were greater than 0.9972. The reproducibility was better than 1.90% as expressed in relative standard deviation (RSD; N=30) and the detection limits were in the range of pptv for the target halocarbons. A field test by continuous analyzing ambient air with hourly resolution was performed to show the stability of the method as suggested by the homogeneity of certain halocarbons, while also reflecting concentration variation for others when emissions did arise. Keywords: Deans switch; CFCs; ozone depletion substance; thermal desorption; twodimensional gas chromatography; electron capture detector.
1. Introduction Man-made halogenated trace gases are blamed for the stratospheric ozone depletion [1,2]. Due to the extremely long lifetimes of some selected halocarbons in the atmosphere, these compounds can drift to the stratosphere threatening the ozone layer by releasing chlorine or bromine atoms through photolysis [3]. The Montreal Protocol and its subsequent amendments were thus enacted to cease the production and consumption of anthropogenic chlorofluorocarbons (CFCs) and selected halocarbons [4,5]. Tropospheric abundances and emissions of most ozone depleting substances (ODS) hence started to decrease since mid-90s, resulting in the declines in the atmospheric burden of chlorine and bromine [2,6-8]. Although 2
the threat of the CFCs on ozone depletion has been greatly alleviated, the net radiative forcing of the halocarbons as greenhouse gases is still considerable since they still contribute up to 11.5% of the net radiative forcing with the inclusion of the CFC replacements of hydrochlorofluorocarbons and hydrofluorocarbons [9-13]. GC-ECD or GC-MS equipped with an enrichment device has been a common method to analyze ambient halocarbons [14-18]. The enrichment can be made either by using cryogens (e.g., liquid nitrogen) [19] or chemical sorbents with thermal desorption (TD) [16]. As more volatile compounds are invented and emitted into the atmosphere, the complexity of air composition increases over time and; thus, the analytical techniques with higher separation resolution become ever more in demand. The conventional GC techniques are often faced with limited separation efficiency that may not be adequate in the modern era when multidimensional separation becomes more assessable and common. Though GC-MS is considered as a multi-dimensional technique in concept with the retention time as the first dimension (1D) and the mass spectrum as the second dimension (2-D), its analytical power may still be hindered by insufficient GC separation when encountering complex samples. Given the full recognition of the achievements made by the conventional GC-ECD or GC-MS in analyzing trace-level atmospheric halocarbons in the past, the comprehensive GC×GC techniques can take the conventional halocarbon analysis to the next level by providing enhanced GC separation efficiency. Previous work by Wang et al. has demonstrated a cryogenic GC×GC-MS method for 30 atmospheric halocarbons [20]. Rather than using thermal modulation with a cryogen, flow modulation using a Dean switch was adopted for hardware simplicity and low cost in construction and operation. However, compared to the thermal modulation counterpart, flow modulation sacrifices some sensitivity and resolution, since minimal or no sample focusing is resulted from modulation [21]. Other than the cost and complexity issues, thermal modulation 3
is not without drawbacks. For instance, it has potential issues with the break-through of volatile analytes [22]. This study aims at developing a comprehensive TD-GC×GC-ECD method to analyze halogenated compounds at ambient level. ECD was employed as the detection method which is known to be extremely sensitive and rugged for highly halogenated compounds. Moreover, the high selectivity of ECD towards halocarbons can largely eliminate interferences from other air-borne volatile compounds bearing no halogen atoms to achieve rather simplified chromatography. Connecting TD-GC×GC to a more elaborate detection method such as timeof-flight although could greatly extend the analyte coverage; the simplicity in chromatography would be lost. The sampling rate of ECD of 50 Hz makes it sufficiently fast for the flow modulated TD-GC×GC applications. To our best knowledge, no or limited GC×GC work has been given to the ODS analysis and only few studies have shown the potential of the applications of GC×GC-ECD for the persistent organic pollutants such as pesticides residues, dioxins, chlorobenzenes, polychlorinated biphenyls, and fluorinated polycyclic aromatic hydrocarbons [23-27]. As a proof of concept, an analytical system of flow-modulated TDGC×GC-ECD was built in-house to demonstrate the feasibility of such an application. Tracelevel halocarbons of significant atmospheric importance, which have been monitored by one dimensional GC with a long history [14-19,28,29], can be readily analyzed via TD-GC×GCECD and viewed from a two-dimensional perspective.
2. Methodology 2.1. Halocarbon identification and quantification An aluminum AL-150 type cylinder (Spectra Gases, Branchburg, NJ, USA) filled with compressed ambient air was employed as the working standard and air sample for testing the analytical method. In this study, 9 halocarbons, CCl2Cl2 (CFC-12), CCl3F (CFC-11), methylene chloride (CH2Cl2), CCl2FCClF2 (CFC-113), chloroform (CHCl3), methyl chloroform 4
(CH3CCl3),
carbon
tetrachloride
(CCl4),
trichloroethylene
(CClH=CCl2),
and
tetrachloroethylene (CCl2=CCl2) were targeted due to their well-characterized properties, such as homogeneity, ubiquity, and the high sensitivity to ECD in detecting them [16], make them ideal targets to test the method. The concentrations of these compounds ranged from few pptv to sub-ppbv, which were well above the detection limits of ECD with the use of the TD unit in the system for in-line enrichment [16]. However, ECD is less sensitive for some hydrogencontaining halocarbons at ambient level ranging from few ppt to sub-ppbv even with the use of a pre-concentrator. The oxygen-doping technique by adding a small flow of zero air in the make-up gas of ultra-pure nitrogen can increase the sensitivity of ECD for the hydrogencontaining halocarbons (e.g., CH2Cl2 and CHCl3 in this study) [29,30]. The mixing was conducted by exploiting the electronic pressure controllers built in the GC to maintain very constant flows of nitrogen and air to attain the desired mixing ratio of 0.2% for oxygen in nitrogen. Compound identification was made by spiking known compounds in both the 1-D and GC×GC analysis, and the retention times can correlate rather well in the x - coordinate of both methods. In addition, since peaks with GC-ECD are usually very characteristic for the target halocarbons in ambient air, hence identification with spiking, assisted with characteristic fingerprinting, made identification relatively straightforward. Concentrations of the 6 target halocarbons (CFC-12, CFC-11, CH2Cl2, CFC-113, CH3CCl3, and CCl4) in ambient air were calibrated with a pressurized cylinder to serve as the working standard (or secondary standard). This working standard was in term calibrated with a standard cylinder purchased from National Oceanic and Atmospheric Administration/Global Monitoring Division in the US which is traceable to National Institute of Standards and Technology to serve as our primary standard. Calibration of CHCl3 was made by an US EPA type standard, i.e. TO-15. The concentrations of CFC-12, CFC-11, CH2Cl2, CFC-113, CHCl3, CH3CCl3, and CCl4 in the working standard 5
were calibrated to be 566.1, 247.9, 687.7, 75.2, 22.6, 5.7, and 94.6 pptv, respectively. Because no quality standards were available for tri- and tetrachloroethylene at the time of experiment, the concentration calibration was not possible and only compound identification was performed.
2.2. Sample pre-concentration Fig. 1 illustrates the apparatus of the analytical system used in this study. A TD unit was built in-house with multiport switching valves (A26UWE, Valco Instruments Co. Inc., USA) to guide sample flows during sample trapping and injection. A glass tube (10 cm × 2 mm I.D. × 1/8” O.D.) packed with carbon sorbent beds of Carboxen 1000, Carboxen 1003, and Carbotrap formed the sorbent trap. Each bed was about 1 cm in length isolated with glass wools. The air sample was drawn by an oil-free pump through the sorbent trap thermostated at 30°C for 5 min or a desired time interval. The sample flow was controlled by a mass flow controller at a constant flow rate of 50 mL min-1. Injection was then performed by flash heating from 30°C to 250°C within seconds, followed by back-flushing the sorbent trap kept at 300°C for 30 min for conditioning. All the tubing used in the TD unit is made of Silcosteel (1/16” O.D., 0.8 mm I.D., Restek, USA). Further detailed information of the TD unit can be referred to our previous works [16,31].
2.3. Heart-cut gas chromatography A GC (6890N, Agilent, USA) equipped with a micro-ECD thermostated at 300°C was used. A Deans switch device comprising three micro tee unions (MT1CS6, Valco Instruments Co. Inc., USA) connected with short sections of Silcosteel tubes (3 cm × 1/16” O.D. × 0.1 mm I.D., Restek, USA) served as the valve-based flow modulator in this work. A three-way solenoid valve (091-0094-900, Parker, USA) was connected to the two ends of the micro tee unions to control the flow direction for modulation.
6
A series of column combinations for the target halocarbons were tested. A non-polar column (either DB-1 or DB-5) with a typical length of 60 m was chosen to be the 1-D column throughout the 2D optimization process. Several columns of large difference in polarity and absorptivity from the 1-D column were systematically tested for the most desired separation efficiency and orthogonality for the halocarbons. Columns of various properties such as polarity, phase types (WCOT vs. PLOT) and dimensions were tested as the 2-D column. The pressures of the auxiliary and carrier gas were setup according to the conditions of different column combinations. The oven temperature program was initially set at 35°C for 10 min, then ramped at 5°C min-1 to 80°C, and finally ramped at 20°C to 200°C for 10 min. The detector signal were further processed and illustrated in 2-D contour plots using a commercially available software Surfer 8 (Golden Software, CO, USA).
3. Results and Discussion 3.1. Selection of GC×GC columns The selection of the 1-D and 2-D columns plays the central role in GC×GC. As illustrated in Fig. 2, a series of column combinations of various phases were tested for the target halocarbons. The separation of the 1-D non-polar column was generally based on the volatilities of the analytes. Several columns of large difference in polarity and absorptivity from the 1-D column were tested for the most desirable separation efficiency and orthogonality for the halocarbons. While keeping the non-polar 1-D column at a fixed length of 60 m throughout the test for the optimal 1-D separation, the 2-D columns of different phases varied in length from 0.3 m to 2 m and inner diameter varied from 0.18 mm to 0.32 mm were tested to yield the best separation, orthogonality and, at the same time, minimal wrap-around. Porous layer open tubular columns such as PLOT and Gaspro, which are generally used for separating compounds of high volatilities close to those of the target halocarbons, were first tested to exploit the high retention forces or absorptivity [32,33]. However, they both showed either severe peak tailing 7
or wrap-around problems working as the 2-D columns due to excessive retention (Figs. 2a and 2b). Three medium-polar wall-coated open tubular (WCOT) columns, i.e. Rti-200, DB-1701, and TG-1301, were also tested for suitability as the 2-D columns. These columns did not give rise to ideal orthogonality for the target compounds by coupling to DB-1 as the 2-D columns (Figs. 2c, 2d, and 2e), though the TG-1301 column showed a slightly better 2-D resolution among all the tested WCOT columns, since the peaks were more focused (Fig. 2e). Subsequently, as an alternative to DB-1 as the 1-D column, a DB-5 column of a smaller pore size was coupled to the TG-1301 column and resulted in a much improved result, possibly due to a better match in polarity and thus more apparent orthogonality (Fig. 2f). The chosen column set (DB-5×TG-1301) was then further optimized for the modulation properties of GC×GC as described below.
3.2. Optimization of GC and modulation parameters Because the Deans-switch was used as the flow modulator, the pressure of the auxiliary flow controls the rapid swings between the actions of sample loading and injection for modulation (Fig. 1). The pressure of the auxiliary flow ranging from 7-10 psi was first tested for the optimized setting. The elution of the analytes in the 2-D column needs to be rapid enough to avoid tailed or diffused peaks (Fig. S1a). However, if the pressure is set too high, the auxiliary gas would tend to impede the effluents from the 1-D column, causing less analytes to be sent into the 2-D column (Fig. S1c). Furthermore, serious co-elution could occur if the auxiliary flow pressure continues to increase (Fig. S1d). As a compromise, the auxiliary pressure was thus set at the optimal pressure of 7.5 psi (Fig. S1b). The sample loading time stands for the duration of the air sample band being sent to the sampling line for loading sample as indicated in Fig. 1 (3 cm in length), which only applies to the flow modulation GC×GC instead of the thermal modulation type, since no cryo-focusing is performed. The sample amount increases as the sample loading time increases. However, up to 8
a certain point, the filling will be complete and the excess sample will be vented. The optimal loading time is determined when the peak response is no longer increased. Various loading times ranging from 0.1 s to 0.7 s were tested (Fig. S2). As the loading time increased between 0 and 0.5 s (Figs. S2a-c), the peak intensity also increased accordingly. However, after 0.5 s the increase in intensity was no longer apparent (Figs. S2c and d), signaling an overflow of the sample. As a result, the optimized loading time was set at 0.5 s. There are three key common parameters that describe the modulation conditions of a GC×GC task for both the thermal and flow modulation types: the modulation ratio (MR), the modulation period (PM) and the modulation duty cycle (DC). MR is the ratio of the width of the 1-D peak to the PM, as defined by Khummueng et al. [34], which determines the reproducibility of the GC×GC results. For instance, MR = 1.5 can only suffice semi-quantitative analysis [34]. MR should be greater than 2.5 to permit satisfactory reproducibility of the transferred effluents (RSD ∼ 1%) [35], and MR = 3 should be aimed at if trace level analysis with high precision is desired [34]. As a result, for our ambient-level analysis, making MR greater than 3 so that each component produces 3-4 pulses is our minimal target to obtain precise results. However, this attempt is perplexed by choosing an appropriate PM value and, at the same time, avoiding wraparound. PM is the elapsed time for each transfer of a sample slice from the 1-D column to the 2-D column. Missing peaks and poor sensitivities can happen if the modulation period is too long. Conversely, if the modulation period is set too short, it is possible that wrap-around can occur. To inspect for wrap-around during the process of PM determination, a series of chromatograms were aligned back-to-back by repeated analyzing the standard air sample (Fig. S3). It was found that when the PM was set as 7 s, both criteria of MR > 3 (= 3.35) and the wrap-around avoidance was achieved. However, the halocarbons at trace ambient level such as CH2Cl2, CH3CCl3, and
9
CCl2=CCl2 became quantitatively undetectable in the GC×GC results if the PM was set too large. Therefore, PM was set as 3 s to achieve MR = 7.82 as a compromised condition. DC is defined as the fraction of PM during which the 1-D effluent is sampled by the modulation [35]. The DC value was determined to be approximately 0.3 in our final optimized GC×GC settings. We acknowledged that this value was rather low compared to that of the typical thermal modulation which has the DC value close to unity. In our case, DC was sacrificed largely for peak resolution. The high enrichment factor of an order of three with the TD device greatly compensated for the lost sample. Meanwhile, the inherited high sensitivity of ECD also tolerated greatly such a sample loss. After all the modulation parameters were determined and set, the GC×GC was then optimized for the GC temperature programing. Fig. 3 displays the typical chromatograms from analyzing 250 mL of ambient air resulted from both the conventional 1-D TD-GC-ECD method and the comprehensive TD-GC×GC-ECD method for comparison.
3.3. System validation System validation from the aspects of precision, detection limits and linearity was evaluated for the TD-GC×GC-ECD system. The linearity was assessed by varying the amount of the trapped sample from 100 to 400 mL and five volumes were injected to test for linearity. The 7 target halocarbons showed the correlation coefficients (R2) ranging from 0.9969 for CFC12 to 0.9991 for CFC-11 (Fig. S4). These volume-based curves were then converted into calibration curves by equating the injection volumes to concentrations. Since the regular trapping time is five minutes at the flow rate of 50 mL min-1, which is our standard trapping conditions for routine analysis, any varied trapping time at 50 mL min-1 can correspond to different injection volumes and thus be equivalent to different concentrations since analytes were trapped and air matrix escaped. For instance, trapping two minutes of the same air would
10
be equivalent to 100 mL of the standard sample and thus 40% of the concentration as of the standard condition. The RSD (relative standard deviation) of 30 repeated aliquots (N = 30) of the working standard air sample ranged from 0.81% for CH3CCl3 to 1.90% for CFC-113. The method detection limits based on the precision for the 100 mL aliquots were in the range of 8-24 pptv. Table 1 summarizes the all QA results for the target halocarbons.
3.4. Field test The TD-GC×GC-ECD system was tested by continuous hourly measurements of ambient air for 2 days with the air intake extruding to outside the laboratory. The standard volume of 250 mL was analyzed for each hourly measurement (Fig. 4). Most halocarbons, i.e. CFC-12, CCl4, CFC-113, CH3CCl3, and CH2Cl2, showed steady concentrations with small variability, suggesting negligible emissions and that their presence in the atmosphere is the result of prior emissions and long lifetimes. By contrast, the levels of CFC-11 and CHCl3 elevated on 29th December, which might be caused by the nearby sources in the surrounding buildings. The field test suggests that the method is stable to perform automated continuous measurements, as indicated by the fairly constant levels for halocarbons such as CFC-113 and CCl4 [36,37], while also reflecting concentration variation for compounds such as CFC-11 and CHCl3 when emissions did arise.
4. Conclusions In this study, we have presented a TD-GC×GC-ECD method to measure atmospheric halocarbons using components that are widely available to conventional 1-D GC users. It was found that the column combination of DB-5 as the 1-D column and TG-1301 as the 2-D column offered the best orthogonality among four column combinations of various polarity pairs. A 11
series of modulation parameters were tested for GC×GC. The PM was set at 3 s to maximize MR, which was determined to be 7.82 to ensure high reproducibility and sensitivity. The DC was approximately 0.17, which although is low compared with that of the thermal modulation counterpart, it was greatly compensated by the in-line TD enrichment and the high sensitivity of ECD. Quality assurance including linearity, precision, and detection limits were assessed for the proposed method of TD-GC×GC-ECD to ensure robust measurement of halocarbons at ambient level. Since very limited GC×GC work in analyzing atmospheric ODS and volatile halocarbons has been reported in the literature, the ever increased availability of the GC×GC technique offers the atmospheric halocarbon analysis, which used more conventional GC-ECD or GCMS technique, an entirely new perspective. Acknowledgements This study was finically supported by the Ministry of Science and Technology (formerly National Science Council), Taiwan, through the contract no. NSC 102-2113-M-008002-MY3. References
[1]
S.A. Montzka, S. Reimann, A. Engel, K. Krüger, S. O'Doherty, W.T. Sturges, D. Blake, M. Dorf, P. Fraser, L. Froidevaux, K. Jucks, K. Kreher, M.J. Kurylo, A. Mellouki, J. Miller, O.J. Nielsen, V.L. Orkin, R.G. Prinn, R. Rhew, M.L. Santee, A. Stohl, D. Verdonik, Chapter 1 Ozone-Depleting Substances (ODSs) and Related Chemicals, Scientific Assessment of Ozone Depletion: 2010, Global Ozone Research and Monitoring Project - Report No.52, World Meteorological Organization, Geneva, Switzerland, 2011.
[2]
L.J. Carpenter, S. Reimann, J.B. Burkholder, C. Clerbaux, B.D. Hall, R. Hossaini, J.C. Laube, S.A. Yvon-Lewis, Update on Ozone-Depleting Substances (ODSs) and Other Gases of Interest to the Montreal Protocol, Scientific Assessment of Ozone Depletion: 2014, Global Ozone Research and 12
Monitoring Project - Report No.55, World Meteorological Organization, Geneva, 2014. [3]
M.J. Molina, F.S. Rowland, Stratospheric sink for chlorofluoromethanes: chlorine atom-catalysed destruction of ozone, Nature 249 (1974) 810-812.
[4]
UNEP, Montreal Protocol on Substances that Deplete the Ozone Layer: Final Act, Montreal, 1987.
[5]
UNEP, The Copenhagen Amendment (1992): The amendment to the Montreal Protocol agreed by the Fourth Meeting of the Parties, Copenhagen, 1992.
[6]
R.G. Prinn, J. Huang, R.F. Weiss, D.M. Cunnold, P.J. Fraser, P.G. Simmonds, A. McCulloch, C. Harth, S. Reimann, P. Salameh, S. O'Doherty, R.H.J. Wang, L.W. Porter, B.R. Miller, P.B. Krummel, Evidence for variability of atmospheric hydroxyl radicals over the past quarter century, Geophys. Res. Lett. 32 (2005) doi:10.1029/2004gl022228.
[7]
M. Rigby, R.G. Prinn, S. O'Doherty, S.A. Montzka, A. McCulloch, C.M. Harth, J. Muhle, P.K. Salameh, R.F. Weiss, D. Young, P.G. Simmonds, B.D. Hall, G.S. Dutton, D. Nance, D.J. Mondeel, J.W. Elkins, P.B. Krummel, L.P. Steele, P.J. Fraser, Re-evaluation of the lifetimes of the major CFCs and CH3CCl3 using atmospheric trends, Atmos. Chem. Phys. 13 (2013) 26912702.
[8]
S.A. Montzka, J.H. Butler, B.D. Hall, D.J. Mondeel, J.W. Elkins, A decline in tropospheric organic bromine, Geophys. Res. Lett. 30 (2003) doi:10.1029/2003gl017745.
[9]
M. Maione, U. Giostra, J. Arduini, F. Furlani, F. Graziosi, E. Lo Vullo, P. Bonasoni, Ten years of continuous observations of stratospheric ozone depleting gases at Monte Cimone (Italy) - Comments on the effectiveness of the Montreal Protocol from a regional perspective, Sci. Total Environ. 445 (2013) 155-164.
[10]
M.K. Vollmer, J. Muhle, C.M. Trudinger, M. Rigby, S.A. Montzka, C.M. Harth, B.R. Miller, S. Henne, P.B. Krummel, B.D. Hall, D. Young, J. Kim, J. Arduini, A. Wenger, B. Yao, S. Reimann, S. O'Doherty, M. Maione, D.M. Etheridge, S.L. Li, D.P. Verdonik, S. Park, G. Dutton, L.P. Steele, C.R. Lunder, T.S. Rhee, O. Hermansen, N. Schmidbauer, R.H.J. Wang, M. Hill, P.K. Salameh, R.L. Langenfelds, L.X. Zhou, T. Blunier, J. Schwander, J.W. Elkins, J.H. Butler, P.G. Simmonds, R.F. Weiss, R.G. Prinn, P.J. Fraser, Atmospheric 13
histories and global emissions of halons H-1211 (CBrClF2), H-1301 (CBrF3), and H-2402 (CBrF2CBrF2), J. Geophys. Res. 121 (2016) 3663-3686. [11]
S.A. Montzka, B.D. Hall, J.W. Elkins, Accelerated increases observed for hydrochlorofluorocarbons since 2004 in the global atmosphere, Geophys. Res. Lett. 36 (2009) doi:10.1029/2008gl036475.
[12]
M. Rigby, R.G. Prinn, S. O'Doherty, B.R. Miller, D. Ivy, J. Muhle, C.M. Harth, P.K. Salameh, T. Arnold, R.F. Weiss, P.B. Krummel, L.P. Steele, P.J. Fraser, D. Young, P.G. Simmonds, Recent and future trends in synthetic greenhouse gas radiative forcing, Geophys. Res. Lett. 41 (2014) 2623-2630.
[13]
J.H. Butler, S.A. Montzka, in, http://www.esrl.noaa.gov/gmd/aggi/aggi.html, 2016.
[14]
S.A. Montzka, R.C. Myers, J.H. Butler, J.W. Elkins, S.O. Cummings, Global tropospheric distribution and calibration scale of HCFC-22, Geophys. Res. Lett. 20 (1993) 703-706.
[15]
P.G. Simmonds, S. Odoherty, G. Nickless, G.A. Sturrock, R. Swaby, P. Knight, J. Ricketts, G. Woffendin, R. Smith, Automated gas-chromatograph massspectrometer for routine atmospheric field-measurements of the CFC replacement compounds, the hydrofluorocarbons and hydrochlorofluorocarbons, Anal. Chem. 67 (1995) 717-723.
[16]
J.L. Wang, C.H. Wu, Construction and validation of a cryogen free gas chromatography-electron-capture detection system for the measurement of ambient halocarbons, Anal. Chim. Acta 461 (2002) 85-95.
[17]
B.R. Miller, R.F. Weiss, P.K. Salameh, T. Tanhua, B.R. Greally, J. Muhle, P.G. Simmonds, Medusa: A sample preconcentration and GC/MS detector system for in situ measurements of atmospheric trace halocarbons, hydrocarbons, and sulfur compounds, Anal. Chem. 80 (2008) 1536-1545.
[18]
C.H. Wang, S.W. Chiang, J.L. Wang, Simultaneous analysis of atmospheric halocarbons and non-methane hydrocarbons using two-dimensional gas chromatography, J. Chromatogr. A 1217 (2010) 353-358.
[19]
J.L. Wang, C.J. Chang, W.D. Chang, C. Chew, S.W. Chen, Construction and evaluation of automated gas chromatography for the measurement of anthropogenic halocarbons in the atmosphere, J. Chromatogr. A 844 (1999) 259-269. 14
[20]
Y. Wang, X.B. Xu, L.Y. Yin, H.B. Cheng, T. Mao, K.P. Zhang, W.L. Lin, Z.Y. Meng, J.A. Palasota, Coupling of comprehensive two-dimensional gas chromatography with quadrupole mass spectrometry: Application to the identification of atmospheric volatile organic compounds., J. Chromatogr. A 1361 (2014) 229-239.
[21]
J.V. Seeley, Recent advances in flow-controlled multidimensional gas chromatography, J. Chromatogr. A 1255 (2012) 24-37.
[22]
R.B. Gaines, G.S. Frysinger, Temperature requirements for thermal modulation in comprehensive two-dimensional gas chromatography, J. Sep. Sci. 27 (2004) 380-388.
[23]
E.M. Kristenson, P. Korytar, C. Danielsson, M. Kallio, M. Brandt, J. Makela, Evaluation of modulators and electron-capture detectors for comprehensive two-dimensional GC of halogenated organic compounds, J. Chromatogr. A 1019 (2003) 65-77.
[24]
A.M. Muscalu, E.J. Reiner, S.N. Liss, T. Chen, G. Ladwig, D. Morse, A routine accredited method for the analysis of polychlorinated biphenyls, organochlorine pesticides, chlorobenzenes and screening of other halogenated organics in soil, sediment and sludge by GCxGC-μECD, Anal. Bioanal. Chem. 401 (2011) 2403-2413.
[25]
J.M. da Silva, C.A. Zini, E.B. Caramao, Evaluation of comprehensive twodimensional gas chromatography with micro-electron capture detection for the analysis of seven pesticides in sediment samples, J. Chromatogr. A 1218 (2011) 3166-3172.
[26]
P. Haglund, P. Korytar, C. Danielsson, J. Diaz, K. Wiberg, P. Leonards, U.A.T. Brinkman, J. de Boer, GC×GC-ECD: a promising method for the determination of dioxins and dioxin-like PCBs in food and feed, Anal. Bioanal. Chem. 390 (2008) 1815-1827.
[27]
A.M. Muscalu, E.J. Reiner, S.N. Liss, T. Chen, Determination of polychlorinated biphenyls, organochlorine pesticides, chlorobenzenes in sludge and sediment samples by GC × GC-μECD, Int. J. Environ. Anal. Chem. 90 (2010) 1-13.
[28]
J.W. Elkins, D.W. Fahey, J.M. Gilligan, G.S. Dutton, T.J. Baring, C.M. Volk, R.E. Dunn, R.C. Myers, S.A. Montzka, P.R. Wamsley, A.H. Hayden, J.H. Butler, T.M. Thompson, T.H. Swanson, E.J. Dlugokencky, P.C. Novelli, D.F. 15
Hurst, J.M. Lobert, S.J. Ciciora, R.J. McLaughlin, T.L. Thompson, R.H. Winkler, P.J. Fraser, L.P. Steele, M.P. Lucarelli, Airborne gas chromatograph for in situ measurements of long-lived species in the upper troposphere and lower stratosphere, Geophys. Res. Lett. 23 (1996) 347-350. [29]
F.L. Moore, J.W. Elkins, E.A. Ray, G.S. Dutton, R.E. Dunn, D.W. Fahey, R.J. McLaughlin, T.L. Thompson, P.A. Romashkin, D.F. Hurst, P.R. Wamsley, Balloonborne in situ gas chromatograph for measurements in the troposphere and stratosphere, J. Geophys. Res. 108 (2003) doi:10.1029/2001jd000891.
[30]
E.P. Grimsrud, D.A. Miller, Oxygen doping of carrier gas in measurement of halogenated methanes by gas chromatography with electron capture detection, Anal. Chem. 50 (1978) 1141-1145.
[31]
C.F. Ou-Yang, W.C. Liao, P.C. Wang, G.J. Fan, C.C. Hsiao, M.T. Chuang, C.C. Chang, N.H. Lin, J.L. Wang, Construction of a cryogen-free thermal desorption gas chromatographic system with off-the-shelf components for monitoring ambient volatile organic compounds, J. Sep. Sci. 39 (2016) 14891499.
[32]
J.L. Wang, G.Z. Din, C.C. Chan, Validation of a laboratory-constructed automated gas chromatograph for the measurement of ozone precursors through comparison with a commercial analogy, J. Chromatogr. A 1027 (2004) 11-18.
[33]
W.C. Liao, C.F. Ou-Yang, C.H. Wang, C.C. Chang, J.L. Wang, Twodimensional gas chromatographic analysis of ambient light hydrocarbons, J. Chromatogr. A 1294 (2013) 122-129.
[34]
W. Khummueng, J. Harynuk, P.J. Marriott, Modulation ratio in comprehensive two-dimensional gas chromatography, Anal. Chem. 78 (2006) 4578-4587.
[35]
J.V. Seeley, N.J. Micyus, S.V. Bandurski, S.K. Seeley, J.D. McCurry, Microfluidic Deans switch for comprehensive two-dimensional gas chromatography, Anal. Chem. 79 (2007) 1840-1847.
[36]
C.C. Chang, G.G. Lo, C.H. Tsai, J.L. Wang, Concentration variability of halocarbons over an electronics industrial park and its implication in compliance with the Montreal Protocol, Environ. Sci. Technol. 35 (2001) 3273-3279.
[37]
C.F. Ou-Yang, C.C. Chang, S.P. Chen, C. Chew, B.R. Lee, C.Y. Chang, S.A. Montzka, G.S. Dutton, J.H. Butler, J.W. Elkins, J.L. Wang, Changes in the 16
levels and variability of halocarbons and the compliance with the Montreal Protocol from an urban view, Chemosphere 138 (2015) 438-446. Figure Captions Fig. 1. Schematics of the TD-GC×GC-ECD system using Deans switch as the modulator. Effluents from 1-D column are directed to (a) 2-D column for further separation (b) load sample through the loading section and discard the excess. Fig. 2. GC×GC separation of ambient halocarbons from six column combinations. (a) DB-1 (60 m× 0.32 mm × 0.25 μm) × PLOT (0.4 m × 0.32 mm × 0.8 μm), (b) DB-1 (60 m× 0.32 mm × 0.25 μm) × Gaspro (0.3 m × 0.32 mm), (c) DB-1 (60 m× 0.32 mm × 0.25 μm) × Rxi-200 (1 m × 0.18 mm × 0.2 μm), (d) DB-1 (60 m× 0.32 mm × 0.25 μm)× DB-1701 (0.35 m × 0.25 mm × 0.5 μm), (e) DB-1 (60 m× 0.32 mm × 0.25 μm)× TG-1301 (2 m × 0.25 mm × 0.5 μm), and (f) DB-5 (60 m × 0.32 mm × 0.25 μm) × TG-1301 (2 m × 0.25 mm × 0.5 μm). Fig. 3. A typical 1-D GC-ECD and comprehensive TD-GC×GC-ECD chromatograms from analyzing 250 mL of ambient air. The 1-D chromatogram in red illustrates the chromatogram of zero air spiked with trichloroethylene for compound identification. Fig. 4. Time-series hourly data of selected halocarbons from the campus measurements. ECD response rather than concentration is used as Y-coordinate to show better contrast in variability. The plot of concentration vs. time is shown in Fig. S5.
17
Figr-1
Fig. 1.
18
19
Fig. 2.
Fig. 3.
20
Fig. 4.
21
Table 1 Quality assurance results
Compound CFC-12 CFC-11 CH2Cl2 CFC-113 CHCl3 CH3CCl3 CCl4 a
R2 (5 points) 0.9969 0.9991 0.9987 0.9979 0.9982 0.9972 0.9988
RSD (N=30) 1.22% 1.39% 1.88% 1.90% 1.44% 0.81% 1.49%
Method Detection Limit (ppt)a
The method detection limits are calculated based on 3σ of the 100mL aliquots.
22
24.9 4.5 5.4 9.9 5.6 2.9 2.0