Dielectric Barrier Discharge Ionization of Perfluorinated Compounds

Oct 26, 2015 - For calibration a GC was coupled to the same MS by an in-house built ... capillary show a very tiny signal at 435 m/z whereas an intens...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/ac

Dielectric Barrier Discharge Ionization of Perfluorinated Compounds Alexander Schütz, Sebastian Brandt, Sascha Liedtke, Daniel Foest, Ulrich Marggraf, and Joachim Franzke* Leibniz-Institut für Analytische Wissenschaften − ISAS − e.V., Bunsen-Kirchhoff-Str. 11, 44139 Dortmund, Germany ABSTRACT: The soft ionization ability based on plasma-jet protonation of molecules initiated by a dielectric barrier discharge ionization source (DBDI) is certainly an interesting application for analytical chemistry. Since the change of an applied sinusoidal voltage may lead to different discharge modes the applied discharge was powered by a square wave generator in order to get a homogeneous plasma. It is known that besides the protonation [M+H]+ of unpolar as well as some polar molecules the homogeneous DBDI can be used to ionize molecules directly [M]+. Here we prove that the DBDI can be applied to exchange fluorine by oxygen of perfluorinated compounds (PFC). PFC are organofluorine compounds with carbon−fluorine and carbon−carbon bonds only but no carbon−hydrogen bonds. While the position of the introduction into the plasma-jet is essential, PFC can be measured in the negative mass spectrometer (MS) mode.

F

of the Stockholm Convention might be much higher. Most of these screening studies searched for highly persistent and bioaccumulative chemicals, but did not include the long-range transport potential, which is a key criterion under the Stockholm Convention. Scheringer et al. applied the POP screening criteria defined in Annex D of the Stockholm Convention to a huge set of industrial chemicals and identified the substances that exceed these criteria. They defined a group of substances the so-called very POPs group which definitely exceeds the Annex D criteria of the Stockholm Convention. One class of these chemicals are the perfluorinated alkanes like perfluoroheptane, -octane, and -nonane.3 Usually PFOA, PFOS, and POSF can be easily measured by electrospray ionization (ESI)-MS whereas compounds like perfluorooctane or perfluorononane could not be ionized by ESI and weakly ionized by atmospheric pressure chemical ionization (APCI). The ionization conditions in APCI are considered to be somewhat “harder” than those in ESI.4 While APCI is understood to be primarily based on gas-phase ion−molecule reactions between analyte molecules and a solvent-based reagent gas, generated by a series of ion−molecule reactions initiated by electrons from the corona discharge needle as alternative to corona discharge generated plasmas, a DBD can be used to generate low-temperature plasmas at atmospheric pressure.5,6 Marotta et al. show that the reactivity of perfluorohexane in forming positive ions is highly dependent on the temperature of the APCI source. At room temperature no significant ionization

or years PFC are supposed to be harmful substrates which can be found in drinking water, clearing sludge and finally in human blood. Even Stockholm Convention on Persistent Organic Pollutants (POPs) is listing PFC because of fulfilling toxicity criteria. Compounds like perfluorooctanesulfonic acid (PFOS), its salts, and perfluorooctanesulfonyl fluoride (POSF) are identified as being extremely persistent and having substantial bioaccumulating and biomagnifying properties. Whereas other POPs show a partitioning into fatty tissue, PFC has an affinity to bind to proteins in the blood and liver.1 Also the Umweltbundesamt (German Federal Environmental Agency) has published some written warnings concerning the vulnerability of PFC.2 Although Stockholm Convention on Persistent Organic Pollutants and the Umweltbundesamt are aware of the risk of PFC they just point to perfluorooctanoic acid (PFOA), PFOS, and POSF. These anthropogenic substances have been discovered in the environment that have not been detected previously. In general they are often referred to as “contaminants of emerging concern” (CECs) because the risk to human health and the environment associated with their presence, frequency of occurrence, or source may not be known. It is hypothesized that PFOA or PFOS, which are common industrial products, are released into the environment where they may lose their functional group and become perfluorinated alkanes. These compounds show a very low intensity only when they are detected with conventional methods. Here a new ionization method would fill the gap. Under the Stockholm Convention on Persistent Organic Pollutants, currently 22 chemicals or groups of chemicals are regulated as POPs.3 However, various screening exercises performed on large sets of chemicals indicate that the number of substances fulfilling the screening criteria defined in Annex D © 2015 American Chemical Society

Received: July 28, 2015 Accepted: October 26, 2015 Published: October 26, 2015 11415

DOI: 10.1021/acs.analchem.5b03538 Anal. Chem. 2015, 87, 11415−11419

Article

Analytical Chemistry could be observed whereas at a temperature of 300 °C many different ionic products are formed. The negative ion chemistry is likewise interesting. Here, the authors state that perfluorohexane gives stable molecular anions, [M]−, which at low temperature or, in humidified air, also at high temperature, are also detected as hydrates, [M+(H2O)]−. Complexes like [F +(H2O)n] are also observed in humidified air. Most interesting for our measurements was that they also detected [M+O−F]− getting a rather stable signal when Vcone is increased and additional signals appear which are assigned to products of reaction with [O]−.7 However, the DBD is typically formed between two electrodes, with at least one dielectric layer which separates the electrode from the plasma. The DBD plasmas are suitable for the atomization of volatile species.8−10 Furthermore, different geometrical arrangements have also served as an ionization source for ambient MS11−14 and ion mobility spectrometry.15−17 In this respect, DBD was proven to be an efficient ionization for LC/MS applications. Consequently, a microplasma ionization based on dielectric barrier discharge (DBDI) was implemented into commercial APCI interface for LC/MS applications. Therefore, a heterogeneous compound library was investigated by DBDI to illustrate the potential use of the miniaturized plasma as an alternative ionization technique to ESI, APCI and APPI. This study presents the use of a dielectric barrier discharge ionization for the measurement of perfluorinated alkanes by exchanging fluorine and oxygen by collisions with N2+ resulting in the measurement of [M+O−F] − or [M-3] − . The perfluorinated samples will be introduced directly by a GC capillary from headspace or gas chromatography without any thermospray (GC-DBDI-MS).

Figure 1. Schematic of DBDI-MS experiment: A DBDI was used for soft ionization while gas chromatography or headspace was used for sample apply. The capillary with nitrogen flow can be adjusted by a micro moving stage.

and a platinum electrode is plugged into the third connector to create a connection between solvent and high voltage. Mass Spectrometry. Detection by mass spectrometry was performed by an ion trap MS (Thermo LCQ Deca XP) in the negative ion mode. Instead of automatic gain control (AGC) a fixed injection time of 200 ms was set to prevent signal cutoff. Analyte supply was done by headspace (HS) because of the low vapor pressure of PFC. Therefore, an uncoated GC-capillary (id: 50 μm, od: 360 μm) was mounted in front of the plasmajet and MS inlet. A 10 mL/min nitrogen flow transported molecules from a flask through the capillary (length about 20 cm). To prevent memory effects the transport system was cleaned continuously by nitrogen flow and vaporized methanol and hexane. For calibration a GC was coupled to the same MS by an in-house built heated transfer line.

■

■

EXPERIMENTAL SECTION Chemicals and Solvents. PFC were purchased from Sigma-Aldrich Germany. Most of the chemicals were shipped as liquids with purity about 97%. Typically solutions were mixed by diluting the stock solution by adding solvents like methanol or hexane. Safety Considerations. As written in the beginning some PFC are already known as hazardous and risky chemicals. In the same way used solvents such as methanol or hexane are toxic. Gloves, protection goggles, and laboratory coat must be worn. An extractor hood was mounted above the experiment as well. During a high voltage is needed for plasma based ionization source it is important to take precautions like adequate isolation for energized cables, warning signs, and switch off high voltage if not necessary. Ambient Air Ionization Techniques. DBDI was used as ambient air soft ionization source for mass spectrometry analysis of PFC. A glass capillary (id: 450 μm, od: 900 μm) with two ring electrodes (distance 1 cm) has been implemented to ignite a miniaturized soft plasma-jet shown in Figure 1. Helium 5.0 (purity 99.999%) with a flow rate of 150 mL/min was chosen as gas because of its low ignition voltage and excellent ionization efficiency under homogeneous plasma conditions.18,19 A voltage of 3.5 kV was applied to the front electrode by an in-house build high voltage square wave generator at a frequency of 20 kHz. For nano ESI (nESI) an emitter tip (tip: 8 μm, id: 20 μm) from New Objective is attached to one side of a tee connector whereas the opposite side is connected to the sample syringe

RESULTS AND DISCUSSION ESI is used extensively in biomolecular analysis, including proteomics, metabolomics, and glycomics. The nano ESI (nESI) method uses flow rates smaller than 1000 nL/min and a smaller diameter spray tip as conventional ESI. Advantages of nESI over conventional ESI include much improved ionization efficiency and ion transmission, greatly reduced ion suppression and matrix effects, smaller sample volumes and lower absolute sample amounts, which improve compatibility with high-efficiency separation techniques.20 Usually, ESI as well as nESI can be used as ionization source for liquid samples by mixing the sample with methanol (MeOH) and water (H2O). In the following measurements the nESI solution (relative volume concentration: 49.5% MeOH, 49.5% H2O, 1% acetic acid) contains no analyte like PFOA and is just used to generate ions by coulomb explosion of the charged droplets. While most of the PFCs here perfluorohexane, perfluoroheptane, perfluorooctane, and perfluorononane exhibit an even higher vapor pressure than PFOA, they can easily be introduced by HS directly into the taylor conus by an auxiliary gas and do not have to be dissolved. Figure 2 shows that PFOA could be excellently measured by HS-ESI-MS while the PFC represented by perfluorooctane shows no signal at all. Therefore, PFC seems not to be ionized and another ionization technique should be applied in order to measure PFC. Similar to APCI the DBDI is known as a very soft protonation based ionization technique. Since electrons, 11416

DOI: 10.1021/acs.analchem.5b03538 Anal. Chem. 2015, 87, 11415−11419

Article

Analytical Chemistry

Figure 2. ESI spectra of perfluorooctane acid (5.74 μL/mL) and perfluorooctane (4.16 μL/mL). Perfluorooctane could definitely not be measured by electrospray.

metastable gas atoms, and nitrogen ion molecules are produced in the plasma-jet the question raises whether the DBDI can also be used as negative ionization source because the ions of the PFC are supposed to be measured in the negative MS mode. In contrast to direct ionization and protonation another ionization mechanism should be responsible. A homogeneous plasma was ignited in helium using a dielectric barrier discharge powered by an in-house made high voltage square wave generator. In order to prevent dissociation of the molecules the analyte should not be introduced into the plasma between the electrodes but into the plasma-jet. The MS was operated in the negative ion mode when perfluorooctane was introduced by HS into the plasmajet at different positions of the analyte supply. Mass spectra of perfluorooctane with the respective photos of the experimental arrangements are shown in the first three parts of Figure 3a. In all photos a helium plasma-jet is shown on the left and the inlet of the MS on the right side. The HS analyte was introduced at three different positions. In case of the upper spectrum the introduction part of the HS was in front of the inlet of the MS, in the middle spectrum between the MS inlet and the end of the DBDI capillary and in the lower spectrum the HS introduction part is positioned next to the end of the DBDI capillary. The spectra which were recorded when the analyte supply was not in the vicinity of the end of the DBDI capillary show a very tiny signal at 435 m/z whereas an intense signal could be detected when the analyte was introduced in the vicinity of the end of the plasma-jet. The 435 m/z represents a mass signal of [M-3]− which can be interpreted as an exchange of one fluorine atom against an oxygen radical. The chemical structures of perfluorohexane, perfluoroheptane, and perfluorononane are very similar to that of perfluorooctane and are distinguished by the amount of carbon atoms only. The spectra were also recorded when the analyte is introduced at different positions. With the analyte introduced near the end of the capillary, the best signal-to-noise ratio was obtained. At this optimum position an additional measurement by introducing perfluorohexane has been performed and is depicted in the fourth part of Figure 3a. Figure 3b contains mass spectra of perfluoroheptane, perfluorooctane, and perfluorononane from GC-DBDI-MS with a concentration of 50 nL/mL. Mass signals of [M-3]− with 385 m/z, 435 m/z, and 485 m/z were detected, respectively.

Figure 3. (a) Mass spectra of perfluorooctane (4.16 μL/mL) when the analyte is introduced by headspace at three different positions between the DBDI capillary end and the inlet of the MS (please note the different scale of y-axis). The last spectrum shows an additional measurement for perfluorohexane (3.95 μL/mL) at optimum position. (b) Mass spectra of perfluoroheptane, perfluorooctane, and perfluorononane from GC-DBDI-MS with a concentration of 50 nL/mL. (c) Calibration curve by GC-DBDI-MS for perfluorononane, perfluorooctane, and perfluoroheptane.

Marotta et al. used collision-induced dissociation (CID) to create a plenty of fragments of their APCI spectra.7 Here, CID was used to prove stability of the PFC species represented by perfluorooctane. Until a CID voltage of 30 V no respectable change of the spectra was found. Increasing the CID voltage from 30 V up to 70 V yields a decreasing signal intensity of the [M-3]− peak at 435 m/z whereas the fragments showed a 11417

DOI: 10.1021/acs.analchem.5b03538 Anal. Chem. 2015, 87, 11415−11419

Article

Analytical Chemistry proportional signal decrease. CID does not destroy [M-3]− but decreases the detection sensitivity. Therefore, [M-3]− is a stable molecule which can be applied for the detection of perfluoroalkanes excellently. To present DBDI for PFC as a suitable method for analytical sciences a calibration using gas chromatography (GC) was performed to separate a solution of different PFC (perfluoroheptane, perfluorooctane, and perfluorononane). Instead of ionization by electron impact (EI) like in commercial GC-MS the eluate was directly introduced at the end of the DBDI capillary into the plasma-jet at ambient air conditions. The calibration of perfluorononane, perfluorooctane, and perfluoroheptane is shown in Figure 3c in a double log scale. The 3σ detection limit (LOD) of perfluorooctane and perfluorononane was 5 nL/mL while perfluoroheptane had its LOD at 10 nL/ mL. As an example the content level of PFOA from Baltic Sea is in between 30 and 5900 pg/L (preconcentration factor 37200). 21 As shown, the detection sensitivity of the perfluorinated alkanes is comparable with the detection sensitivity of PFOA when it is ionized by ESI or DBDI. For the measurement of real samples such as perfluorononane with DBDI-MS a preconcentration method in the range of at least a factor of 3 × 105 might be necessary. Besides these compounds perfluoro-2-methyl-2-pentene (PFMP) and perfluoro(methylcyclohexane) (PFMCH) also belonging to the PFC could also be detected at [M-3]− with spectra shown in Figure 4. Even though the chemical structures

Figure 5. Perfluorooctanoic acid (5.74 μL/mL) detected by DBDIMS.

oxygen like in the case of the perfluorinated alkanes takes place. Therefore, the DBDI is not only able to ionize nonpolar but also polar molecules as ESI is mostly used for.22 In order to understand the possible mechanism the characteristics of the DBDI has to be explained. A helium capillary DBD was investigated by means of time-resolved optical emission spectroscopy with the aim of elucidating the process of the formation of the plasma-jet.17 It has been shown that two independent plasmas, separated in time, are formed: first, the plasma-jet and the early plasma in between the electrodes; second, the discharge in the capillary. It was found that the plasma-jet is formed only during the positive halfperiod of the applied voltage. In the early plasma as well as in the plasma-jet, the helium atom excitation propagation starts in the vicinity of the high voltage electrode and is departing in both directions with velocities in the range from 1 to 50 km/s. The excitation is initiated by electrons which will be attracted by the positive electrode. Shortly after the metastables are generated, N2+ are formed. Due to the electric fields starting from the high voltage electrode the N2+ will be accelerated in both directions against the electron movement. Therefore, collisions of electrons, He metastables (HeM) or N2+ with a PFC might be responsible for the loss of fluorine atoms of the perfluorinated molecules when it is treated by the plasma-jet. Since an electron gun was not successful, electrons might not be the collision partner and due to the fact that a plasma-jet only surrounded by He will not lead to any ionization of a PFC a direct collision with HeM can also not be the reason for the ionization. Therefore, collisions with N2+ might be the only explanation.

■

CONCLUSION The general usage of DBDI-MS for the detection of representatives for PFC has been demonstrated. Position and structure dependent effects have been studied. The exchange of fluorine by oxygen radicals might result from collision by N2+. Finally a calibration was performed to prove a future analytical application to extent common ESI-MS experiments.

Figure 4. Perfluoro(methylcyclohexane) (PFMCH, 4.07 μL/mL) and perfluoro-2-methyl-2-pentene (PFMP, 2.20 μL/mL).

■

contain carbon and fluorine only the spatial pattern is different. A CC double bond inside PFMP is very characteristic to have a huge affinity to bind oxygen. Also PFMCH has a kind of open attached fluorine which might be exchanged by oxygen more easily. Although PFMP and PFMCH were detected at [M-3]− no dependence on position was observed. The difference in the chemical structure compared to the previous compounds could cause a faster and more efficient exchange. Therefore, the sensitivities measured for PFMP and PFMCH are not a function of the place where the analyte is introduced into the plasma-jet. Furthermore, PFOA can be ionized by DBDI too as it is shown in Figure 5. But here similar to the ionization by ESI a deprotonation due to a detachment of the hydrogen from the OH instead of a change of fluorine against

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: +49 231 1392-174. Fax: +49 231 1392-120. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The financial support by the “Ministerium für Innovation, Wissenschaft und Forschung des Landes Nordrhein-Westfalen”, by the “Bundesministerium für Bildung und Forschung”, and by the “Deutsche Forschungsgemeinschaft (DFG)” is gratefully acknowledged. 11418

DOI: 10.1021/acs.analchem.5b03538 Anal. Chem. 2015, 87, 11415−11419

Article

Analytical Chemistry

■

REFERENCES

(1) Stockholm Convention, http://chm.pops.int/TheConvention/ ThePOPs/ListingofPOPs/tabid/2509/Default.aspx, 2015. (2) Umweltbundesamt, http://www.umweltbundesamt.de/themen/ chemikalien/chemikalien-reach/stoffgruppen/per-polyfluoriertechemikalien-pfc, 2015. (3) Scheringer, M.; Strempel, S.; Hukari, S.; Ng, C. A.; Blepp, M.; Hungerbuhler, K. Atmos. Pollut. Res. 2012, 3, 383−391. (4) Niessen, W. M. A. J. Chromatogr. A 1999, 856, 179−197. (5) Miclea, M.; Kunze, K.; Musa, G.; Franzke, J.; Niemax, K. Spectrochim. Acta, Part B 2001, 56, 37−43. (6) Snyder, H. R.; Anderson, S. K. IEEE Trans. Plasma Sci. 1998, 26, 1695−1699. (7) Marotta, E.; Paradisi, C.; Scorrano, G. J. Mass Spectrom. 2004, 39, 791−801. (8) Zhu, Z. L.; Zhang, S. C.; Lv, Y.; Zhang, X. R. Anal. Chem. 2006, 78, 865−872. (9) Zhu, Z. L.; Liu, J. X.; Zhang, S. C.; Na, X.; Zhang, X. R. Spectrochim. Acta, Part B 2008, 63, 431−436. (10) Yu, Y. L.; Du, Z.; Chen, M. L.; Wang, J. H. J. Anal. At. Spectrom. 2008, 23, 493−499. (11) Na, N.; Zhao, M. X.; Zhang, S. C.; Yang, C. D.; Zhang, X. R. J. Am. Soc. Mass Spectrom. 2007, 18, 1859−1862. (12) Na, N.; Zhang, C.; Zhao, M. X.; Zhang, S. C.; Yang, C. D.; Fang, X.; Zhang, X. R. J. Mass Spectrom. 2007, 42, 1079−1085. (13) Ma, X. X.; Zhang, S. C.; Lin, Z. Q.; Liu, Y. Y.; Xing, Z.; Yang, C. D.; Zhang, X. R. Analyst 2009, 134, 1863−1867. (14) Zhang, Y.; Ma, X. X.; Zhang, S. C.; Yang, C. D.; Ouyang, Z.; Zhang, X. R. Analyst 2009, 134, 176−181. (15) Michels, A.; Tombrink, S.; Vautz, W.; Miclea, M.; Franzke, J. Spectrochim. Acta, Part B 2007, 62, 1208−1215. (16) Vautz, W.; Michels, A.; Franzke, J. Anal. Bioanal. Chem. 2008, 391, 2609−2615. (17) Olenici-Craciunescu, S. B.; Michels, A.; Meyer, C.; Heming, R.; Tombrink, S.; Vautz, W.; Franzke, J. Spectrochim. Acta, Part B 2009, 64, 1253−1258. (18) Meyer, C.; Muller, S.; Gilbert-Lopez, B.; Franzke, J. Anal. Bioanal. Chem. 2013, 405, 4729−4735. (19) Muller, S.; Krahling, T.; Veza, D.; Horvatic, V.; Vadla, C.; Franzke, J. Spectrochim. Acta, Part B 2013, 85, 104−111. (20) Karas, M.; Bahr, U.; Dulcks, T. Fresenius' J. Anal. Chem. 2000, 366, 669−676. (21) Theobald, N.; Gerwinski, W.; Caliebe, C.; Haarich, M. Umweltbundesamt, 41. (22) Hayen, H.; Michels, A.; Franzke, J. Anal. Chem. 2009, 81, 10239−10245.

11419

DOI: 10.1021/acs.analchem.5b03538 Anal. Chem. 2015, 87, 11415−11419