Tuneable Microsecond-Pulsed Glow Discharge Design for the

Anal. Chem. 2009, 81, 2591–2599

Tuneable Microsecond-Pulsed Glow Discharge Design for the Simultaneous Acquisition of Elemental and Molecular Chemical Information Using a Time-of-Flight Mass Spectrometer Auristela Sola ` -Va´zquez, Antonio Martıń, Jose´ M. Costa-Ferna ´ ndez, Rosario Pereiro, and Alfredo Sanz-Medel* Department of Physical and Analytical Chemistry, Faculty of Chemistry, University of Oviedo, c/Juliań Claverıá, 8, 33006 Oviedo, Spain A microsecond-pulsed direct current glow discharge (GD) was interfaced and synchronized to a time-of-flight mass spectrometer MS(TOF) for time-gated generation and detection of elemental, structural, and molecular ions. In this way, sequential collection of the mass spectra at different temporal regimes occurring during the GD pulse cycle is allowed. The capabilities of this setup were explored using bromochloromethane as model analyte. A simple GD chamber, developed in our laboratory and characterized by a low plasma volume minimizing dilution of the sample but showing great robustness to the entrance of organic compounds in the microsecond-pulsed plasma, has been used. An exhaustive analytical characterization of the GD-MS(TOF) prototype has been performed. Calibration curves for bromochloromethane observed at the different time regimes of the GD pulse cycle (that is, for elemental, fragment, and molecular ions from the analyte) showed very good linearity for the measurement of the different involved ions, with precisions in the range of 7-13% (relative standard deviation). Actual detection limits obtained for bromochloromethane were in the range of 1-3 µg/L for elements monitoring in the GD pulse “prepeak”, in the range of 11-13 µg/L when monitoring analyte fragments in the plateau, and about 238 µg/L when measuring the molecular peak in the afterpeak regime. There is an increasing awareness in many scientific fields of the importance of knowing the chemical form in which an element is present in a given sample.1,2 State-of-the-art speciation analysis is today commonly accomplished by resorting to hyphenated analytical techniques, particularly to the use of a powerful chromatographic separation technique coupled to a highly sensitive atomic spectrometric detector. However, an urgent issue is to access to structural information for the identification of unknown or unexpected-to-be-found compounds in chromato* To whom correspondence should be addressed. Fax: 34-985103474. E-mail: [email protected]. (1) Sanz-Medel, A. Spectrochim. Acta, Part B 1998, 53, 197–211. (2) Szpunar, J.; Lobinski, R.; Prange, A. Appl. Spectrosc. 2003, 57, 102A–122A. 10.1021/ac802520q CCC: $40.75  2009 American Chemical Society Published on Web 03/03/2009

grams obtained during speciation of complex biological molecules.3,4 Soft ionization techniques (e.g., matrix-assisted laser desorption ionization or electrospray) coupled to tandem mass spectrometry ((MS)n) detection are essential tools for structure determination and characterization of biomolecules. Therefore, in addition to the element-selective information given by the atomic spectrometric detectors (e.g., inductively coupled plasma mass spectrometry, ICPMS), evaluation of the molecular structure requires complementary molecular detection techniques (e.g., electrospray ionization tandem mass spectrometry, ESI-(MS)n), with the resulting analysis time and cost increases. Thus, a reliable, selective, and sensitive system capable of providing, simultaneously or in a rapid sequence, both types of information (atomic and molecular) should be desirable.5 In this vein, one possible approach is the use of a “tandem source”, consisting of two complementary on-line ion sources possible, one providing the elemental mass spectrum and the other the molecular mass one6 or a “dual-source” approach offering such information by a parallel use of two sources (ICP and ESI) in a single time-of-flight mass spectrometer (MS(TOF)) instrument.7 In principle, a simpler approach would be the use of a single ionization source in different modes of operation to produce elemental and molecular ions. Modulated (or switched) ion sources, operated under two or more sets of conditions, each corresponding to a specific (atomic or molecular) ionization mode have also been described for the purpose.8,9 In such source, switching between the atomic and molecular modes of operation can be achieved by altering the discharge gas composition, the operating pressure, and the current. Unfortunately, quasisimultaneous atomic and molecular analysis using these approaches is hindered by the requirement of a continuous tuning of the operating parameters to obtain elemental or molecular ions. (3) Szpunar, J.; Lobinski, R. Hyphenated Techniques in Speciation Analysis; Royal Society of Chemistry: Cambridge, U.K., 2003. (4) Sanz-Medel, A. Anal. Bioanal. Chem. 2008, 391, 885–894. (5) Marcus, R. K.; Evans, E. H.; Caruso, J. A. J. Anal. At. Spectrom. 2000, 15, 1–5. (6) Ray, S. J.; Hieftje, G. M. Anal. Chim. Acta 2001, 445, 35–45. (7) Ray, S. J.; Andrade, F.; Gamez, G.; McClenathan, D.; Rogers, D.; Schilling, G.; Wetzel, W.; Hieftje, G. M. J. Chromatogr., A 2004, 1050, 3–34. (8) Guzowski, J. P.; Broekaert, J. A. C.; Ray, S. J.; Hieftje, G. M. J. Anal. At. Spectrom. 1999, 14, 1121–1127. (9) Guzowski, J. P.; Hieftje, G. M. J. Anal. At. Spectrom. 2000, 15, 27–36.

Analytical Chemistry, Vol. 81, No. 7, April 1, 2009

2591

Alternatively, a single “tuneable” micropulsed direct current (dc) glow discharge (GD) coupled to TOF mass spectrometry (MS) could be used.10 Different temporal ionization mechanisms occur within a short pulse in a GD.10,11 The pulsed cycle can be separated into three different temporal regions: prepeak, plateau, and afterpeak, during which, different ionization events occur. Initiation of the discharge (“prepeak”) generates a large number of electrons that can be used to produce electron impact-like spectra, resulting in the highest degree of fragmentation and in an efficient atomization and ionization of the analyte in their elemental constituents. In the “plateau” time regime, the highenergy electron population is diminished and a combination of electron impact, charge exchange, and Penning ionization processes occur. Therefore, in this time regime molecular fragmentation pattern could be obtained. Shortly after the plasma power is terminated (“afterpeak”), charged gas species quickly diminish and interactions between electrons and discharge gas ions yield metastable gas species. Due to the low energy on this latter species, molecular ions information is obtained.11,12 Thus, the coupling of such pulsed ion sources with a time-gated detection mass analyzer, such as a MS(TOF), would allow the quasisimultaneous collection of elemental, structural, and molecular ion information. Although “the proof of concept” was published11-14 actual development of analytical applications of such approach is very limited so far, mainly because the analytical performance characteristics of published systems are relatively poor. One possible cause for this final poor performance could be the design of the discharge chambers used in previous papers (not optimized for gaseous analysis). In this context, we present here the development, optimization, and analytical characterization of a new microsecond-pulsed dcGD, its coupling to a MS(TOF) analyzer, and its analytical performance for volatilized analytes examination. The GD chamber used was designed in our laboratory15,16 and is characterized by about 100 mm3 of internal volume and two opposite inert gas inlets. Additionally, the chamber was coupled to a homemade current stabilizer circuit (CSC). This design has proved to provide a very good electrical stability of the discharge, and this ensures a high robustness of the generated microsecondpulsed plasma to the entrance of volatile analytes (or to organic solvents) with minimal dilutions of the introduced sample in the plasma gas. EXPERIMENTAL SECTION Instrumentation. The GD chamber used and its coupling to a commercial time-of-flight mass analyzer (Renaissance ICPMS(TOF) from LECO, St. Joseph, MI) has been previously de(10) Klinger, J. A.; Barshick, C. M.; Harrison, W. W. Anal. Chem. 1991, 63, 2571–2576. (11) Jackson, G. P.; Lewis, C. L.; Doorn, S. K.; Majidi, V.; King, F. L. Spectrochim. Acta, Part B 2001, 56, 2449–2464. (12) Lewis, C. L.; Moser, M. A.; Hang, W.; Dale, D. E.; Hassel, D. C.; Majidi, V. J. Anal. At. Spectrom. 2003, 18, 629–636. (13) Fliegel, D.; Fuhrer, K.; Gonin, M.; Gu ¨ nther, D. Anal. Bioanal. Chem. 2006, 386, 169–179. (14) Lewis, C. L.; Moser, M. A.; Dale, D. E.; Hang, W.; Hassel, C.; King, F. L.; Majidi, V. Anal. Chem. 2003, 75, 1983–1996. (15) Pisonero-Castro, J.; Costa-Fernańdez, J. M.; Pereiro, R.; Bordel, N.; SanzMedel, A. J. Anal. At. Spectrom. 2002, 17, 786–789. (16) Pisonero, J.; Costa, J. M.; Pereiro, R.; Bordel, N.; Sanz-Medel, A. J. Anal. At. Spectrom. 2001, 16, 1253–1258.

2592


Figure 1. Scale drawing of the source and the finally selected interface between the GD chamber and the TOF(MS).

scribed.16 Different MS(TOF) instrumental variables were adjusted for maximum signal during the GD-MS(TOF) interface optimization. A detector voltage of -2200 V was selected during the optimization experiments (to reduce their aging). In quantitative experiments, the detector voltage was increased up to -2400 V. This instrumental design was initially proposed in our research group for continuous dc- or rf-GD-MS,17 where the voltage applied or the dc bias developed in the rf mode was lower than -1000 V. In our experiments, the dc-GD pulse voltages were between -1500 and -2200 V, and this forced us to cover the area of the surface not exposed to the plasma (see Figure 1) with a thin and highdielectric barrier sheet consisting of polymeric vinyl, laminated on a 120 g/m2 one-side silicon kraft liner, with 70 µm of thickness (KEMICA, Tecmark 5000). In this way, electrical arc formation between the anode and cathode can be prevented during pulsing, avoiding possible short circuits.18 The GD was generated by applying microsecond dc voltage pulses (IRCO, model M3kS-20N, Maryland) to a solid sample which acts as cathode electrode in the discharge. Initial experiments carried out in our laboratory19 demonstrated that a good cathode material is iron when using the GD source for analysis of gaseous analytes, mainly due to the iron low sputtering rate. Moreover, the generated iron ions do not interfere with the mass spectra of organohalogens used in this work as model analytes. Thus, an iron standard material ref M BS 50 D (MBH Analytical Ltd., Herts, U.K.) was selected for the cathode. Moreover, the cathode surface was polished and cleaned daily to avoid potential problems related to performance deterioration and/or shortcircuiting due to cathode erosion and redepositions related with longer times of sputtering in the GD. We observed that if the average current exceeds 8 mA the pulser shuts down and so the plasma cannot be ignited. In order to overcome this electrical inconvenience, a CSC was designed and connected between the (17) Pisonero, J.; Costa, J. M.; Pereiro, R.; Bordel, N.; Sanz-Medel, A. Anal. Bioanal. Chem. 2004, 379, 658–667. (18) Martıń, A.; Pereiro, R.; Bordel, N.; Sanz-Medel, A. J. Anal. At. Spectrom. 2007, 22, 1179–1183. (19) Centineo, G.; Fernańdez, M.; Pereiro, R.; Sanz-Medel, A. Anal. Chem. 1997, 69, 3702–3707.

cathode and the pulser.18 A digital oscilloscope (Tektronix, TDS3012, 100 MHz, 1.25 GS/s) synchronized with the pulsed power supply allows plasma stability monitoring. A 50 µs pulse width and 500 Hz were selected for the measurements. The discharge gas pressure inside the GD ion source was controlled using the nebulizer gas mass flow controller from the ICPMS(TOF). The pressure was measured with the first-stage gauge of the TOF mass spectrometer. High-purity argon (99.999% minimum purity) from Air Liquide (Oviedo, Spain) was employed as the discharge gas. Sample Introduction System. Samples containing the analyte (a highly volatile hydrocarbon) were introduced into the micropulsed GD-MS(TOF) using an exponential dilutor, consisting on a 500 mL glass vessel containing a magnetic stirrer to allow for gas homogeneity. Variable volumes of analyte/solvent mixtures were introduced in the exponential dilution chamber (through a septum, using a 1 µL syringe) to give the required analyte concentration in the 500 mL vessel. The exponential dilutor is provided with two three-way stopcocks that allow the Ar carrier gas to pass directly into the discharge chamber or to go through the vessel, dragging the sample to the plasma. Initial analyte concentration C0 in the vessel is diluted continuously according to the formula C(t) ) C0 exp(-Ft/V)

(1)

where C(t) is the concentration in the vessel at a time t, F is the carrier gas flow rate, t is the time spent after the stopcock was turned, and V is the volume of the vessel. Using eq 1 we could calculate the concentration of an injected sample along the elapsed time.19,20 A restriction was made with a needle valve (SS-4MG-MH Swagelok) on the Tygon sampling tubing, prior to the GD chamber, to match the rate of aspiration of the vacuum pump with the total flow rate of the carrier introduced into the chamber. A sidearm directed into two on-line coupled 50 mL Erlenmeyer flasks, the last one filled with sulfuric acid, was incorporated into the sampling line as a means of controlling the equilibration of the flow rate. By maintaining stable this sulfuric level in the flask a sampling efficiency of near 100% could be obtained. With the above-described system, the sample introduction manifold before to the restrictor is at atmospheric pressure. Changing the Ar carrier flow rate (controlled by using the nebulizer mass flow controller of the Renaissance ICPMS(TOF) system) different plasma pressures were achieved. An argon flow rate of 1.78 mL/s was used to achieve an optimal pressure of 0.39 Torr, measured in the first stage of the TOF mass analyzer. Reagents. Bromochloromethane, the selected model analyte, was obtained from Sigma-Aldrich (Steinheim, Germany). This relatively simple organic compound was selected as model because of the plethora of literature available on its fragmentation patterns and its low boiling point (to avoid problems associated to condensations in the dilutor or in the interface). Hexane and diglyme were obtained from Prolabo and SigmaAldrich, respectively. All the chemicals were of analytical reagent grade and used as received, without any further purification. (20) Lovelock, J. E. Anal. Chem. 1961, 33, 162–178.

RESULTS AND DISCUSSION GD-MS(TOF) Interface Design Considerations. The design of the GD-MS(TOF) interface, based on a conventional samplerskimmer system, included originally a thin stainless steel plate (∼1 mm thick) with a 2 mm diameter orifice, as sampler, and a skimmer of about 1.5 mm diameter central orifice commonly used in ICPMS. Ion transport processes were evaluated via observed sensitivity in relation to the presence-removal of the sampler and to discharge chamber-skimmer distance. In order to investigate the chamber-skimmer distance relationship a variable number of flat stainless steel spacer disks (1.3 or 2.5 mm thick) with an 8 mm diameter orifice were placed between the sampler (or the GD chamber) and the skimmer cone. An amount of 0.3 µL of bromochloromethane/hexane (1:1) was injected inside the exponential dilutor system, and the ion signals from the analyte were monitored in the MS(TOF) analyzer. The GD was operated at a dc voltage of 2000 V, pulsed at a frequency of 500 Hz, and with a pulse duration width of 50 µs. It was observed that element ion intensities (e.g., 79Br+ or 35 + Cl ) were more than 3 times higher when removing the sampler. On the other hand, analyte molecular ion and ions of molecular fragments were only detected without the sampler. These observations could be attributed both to the lower sensitivity expected with the sampler (as was observed for the elemental analyte ions) and also to the fact that the plasma region extracted with the sampler is more constrained (probably the more energetic central region of the plasma, i.e., the region with richer elemental information is preferentially sampled). Therefore, we decided to remove the sampler from the original GD-MS(TOF) interface for further experiments. Regarding the distance of the GD chamber-skimmer cone, experiment showed that the ion intensities from fragments and molecular species decreased sharply if less than two stainless steel separation disks (i.e., below 2.6 mm distance) were used. Significant differences in the ion intensities were not observed when two or three separation disks of 1.3 mm thickness were used. However, when a third spacer disk of 2.5 mm thickness was added, a stable vacuum in the GD chamber could not be secured and the plasma discharge became unstable. Therefore, working without sampler and using two stainless steel spacer disks of 1.3 mm thickness were chosen (see Figure 1). Solvent Selection. With the use of the optimized experimental conditions described in the previous GD-MS(TOF) interface design studies, different organic solvents were tested in order to check the robustness of the GD plasma formed (organic solvents could easily quench such low-pressure plasmas type). In a first approach (2-methoxyethyl)ether (diglyme) was tested, due to its high boiling point (162 °C). Injection of samples prepared in such solvent did not affect the electrical conditions of the discharge, but serious instabilities in the intensities of the ions from the analyte were found. Poor dissolution or reduced volatility of the dissolved analyte could be responsible for such instability of the MS signals. Thus, hexane was evaluated next. With the use of this latter highly volatile solvent, neither detectable plasma cooling nor analyte signals instabilities, were observed and so hexane was the solvent finally selected for further studies. When we represented the ion signal of the analyte along the elapsed time, we Analytical Chemistry, Vol. 81, No. 7, April 1, 2009

2593

Figure 2. Optimization of GD pressure and voltage. (a) Effect of the discharge voltage on the net intensity of 35Cl+ (triangle points, gray dotted line) and 79Br+ (diamond points, black dotted line) in the prepeak time regime, m/z 49 (circle points, striped line) and m/z 93 (square points, solid line) in plateau time regime. (b) Influence of the discharge pressure on the net intensity of 35Cl+ (triangle points, gray dotted line) and 79Br+ (diamond points, black dotted line) in the prepeak time regime, m/z 49 (circle points, striped line) and m/z 93 (square points, solid line) in plateau time regime, m/z 130 (crossed points, dotted striped line) measured in the afterpeak. The error bars correspond to five replicates.

obtained the expected exponential curve from the “exponential dilutor” (used as sample introduction system). If we have had plasma cooling, a decrease in the expected intensity would have been observed in the first part of the graph (corresponding to higher masses of analyte and hexane entering the plasma). Optimization of the Microsecond-Pulsed GD Operating Conditions. The effect of pulsed dc voltage and plasma pressure parameters on the different ion signals (from elements, fragments, and molecular species) was then investigated. Bromochloromethane was diluted in hexane (1:1), and 0.3 µL of such sample solution was injected in the exponential dilutor system. Ion signals of the analyte were monitored running the pulsed GD ion source at a frequency of 500 Hz with a pulse duration of 50 µs width. Figure 2a shows that the intensity of 79Br+ ions increased with the voltage up to 2000 V (for this study a pressure of 0.3 Torr was selected). Similar signals were observed for 2100 V. However, as voltages higher than 2000 V (the upper stable voltage) should bring about plasma instabilities this value was selected for further investigations. Similar behavior was observed for 35Cl+ and fragment ions from the analyte. In Figure 2b the influence of variations in the discharge gas pressure (as measured at the first stage of the mass spectrometer) on the analyte intensities is shown. Previous experiments carried out in our laboratory15 working with this system without sampler demonstrated that the pressure in the third stage of the mass analyzer increases up to 10-5 Torr at pressure above 0.6 Torr in first stage. At such high pressure in the ion detector the TOF mass analyzer goes off (it is self-protected to avoid arc discharges between the different ionic lenses, which could result in a serious damage of the mass detector). As can be seen in Figure 2b, the highest elemental intensities, measured in the prepeak, were achieved when working at the 2594


Figure 3. Schematic representation of the setup for time-gated ion detection. The dotted line represents the dc power pulse temporal profile in the GD. The continuous line represents the time profile applied to the repeller pulse of the MS(TOF) for time-gated detection of the ions generated during each analytical temporal regime of the glow discharge pulse cycle (prepeak, plateau, and afterpeak).

assayed lower pressure (0.30 Torr), where highly energetic electronic impact processes predominate.12 On the other hand, the higher fragments (m/z 49 and m/z 93) observed intensities occurred at a pressure of 0.35 Torr. Molecular ions (m/z 130) were observed only at pressures higher than 0.39 Torr (at such pressures there is higher concentration of metastable gas species and so less energetic discharge is obtained).12 In order to have a tuneable ion source, then, a pressure of 0.39 Torr in the first stage was finally selected. Repeller Delay Selection. As pointed out in the introduction, pulsed GD operation affords distinct temporal regions within the pulse sequence, each offering substantially different ionization pathways.21 An appropriate synchronization between the pulse of the GD ion source (running at a frequency of 500 Hz with a 50 µs pulse width) with the MS(TOF) repeller delay (RD) should be ensured for successful acquisition and detection of the ions generated at the different temporal regions of the GD pulse cycle. Therefore, the desired tuneable ion source could be achieved by just changing appropriately the repeller delay, as can be seen schematically in Figure 3 for different temporal positions of the repeller delay. In our experiments the MS(TOF) repeller delay has been fixed to 2 µs width. Thus, the ion signals generated in the plasma discharge, after injecting 0.3 µL of the mixture bromochloromethane in hexane (1:1) through the exponential dilution system, were monitored at controlled and different MS(TOF) repeller delays. Intensity signals from elemental 79Br+, 35Cl+, two characteristically fragments of the analyte (m/z 49 and m/z 93), and the molecular ion (m/z 130) were monitored aiming at getting the best repeller delays providing the most intense ion signals in each part (time region) of the discharge pulse. Figure 4 shows that the highest intensities for the elemental ions (m/z 79) were obtained when selecting repeller delays of 10 µs, that is, in the prepeak of the discharge pulse. The same behavior was obtained for ion m/z 35. The highest signals for molecular fragments ions (m/z 93) were obtained when selecting a repeller delay of around 50 µs (plateau time regime). Similarly, highest intensities were obtained (21) Steiner, R. E.; Lewis, C. L.; Majidi, V. J. Anal. At. Spectrom. 1999, 14, 1537–1541.

Figure 4. Effect of variations on the repeller delay on the net intensity of monitored ions from bromochloromethane at different discharge time regimes: (a) 79Br+ ions; (b) fragment ions at m/z 93; (c) the molecular ion at m/z 130. The error bars correspond to five replicates.

for RD 50 µs when monitoring m/z 49. Finally, the molecular ions (m/z 130) were only detected at repeller delays greater than the discharge pulse duration. A value of 90 µs was selected to acquire this type of information (afterpeak). The MS fragmentation patterns of the analyte observed with our system at the different discharge time regimes are collected in Figure 5 (the signals observed at m/z 54 and m/z 56 correspond to the sputtered iron from the cathode). The spectra obtained at each time regime with our system shows a rather similar behavior to those previously reported for other analytes12,22 (e.g., elemental information was mostly obtained during the prepeak time regime, plateau spectra collected preferably fragment ion information, and molecular weight information of analyte parent ion was achieved during the afterpeak). The standard NIST mass spectra of the model analyte, obtained via electron impact ionization with an electron energy of 70 eV, is also shown in Figure 5. A detailed comparison between the analytical regions of interest in the spectra obtained by both ionization modes is given in Figure 6. As Figure 6a shows, the prepeak mass spectra of bromochloromethane was dominated by elemental ions (see bromine isotopes at m/z 79 and 81), whereas the measured bromine isotopic ratios are very close to the theoretical ratios. Figure 6b illustrates the observed spectrum in the plateau region: as can be seen, structural information is rich compared to the (22) Fliegel, D.; Waddell, R.; Majidi, V.; Gu ¨ nther, D.; Lewis, C. L. Anal. Chem. 2005, 77, 1847–1852.

Figure 5. Mass spectra obtained with micropulsed dc-GD-MS(TOF) after injection of bromochloromethane/hexane: (a) standard NIST mass spectra obtained via electron impact ionization; (b) mass spectra obtained in the prepeak time regime; (c) mass spectra obtained in the plateau time regime; (d) mass spectra obtained in the afterpeak time regime.

spectra collected during the prepeak. Mass spectra detected with our instrument working in the plateau time regime provided a similar pattern to standard NIST electron impact (EI) mass spectra (see inset in Figure 6b). As an example, the mass spectrum of ClC fragment ions of bromochloromethane Analytical Chemistry, Vol. 81, No. 7, April 1, 2009

2595

Figure 6. Mass spectra obtained with micropulsed dc-GD-MS(TOF) after injection of bromochloromethane and comparison with NIST spectra collected (in a squared box on the right): (a) elemental bromine isotopes detected in the prepeak discharge time regime; (b) fragments obtained in the plateau regime for m/z from 46 to 52. The gray line in the graphs correspond to a GD-MS(TOF) spectra registered when injecting hexane (blank).

(m/z 47, 48, 49, 50) obtained in the plateau time regime is shown. When compared with the inset (EI mass spectrum) the small intensity differences observed can be attributed to additional ionization mechanisms occurring in the pulsed GD (such additional mechanisms can result in changes in the observed relative abundances of ion fragments in comparison to those reported in the corresponding NIST spectra).12,22 Figure 7 illustrates that the molecular weight peaks of analyte parent ions could only be detected during the afterpeak, 40 µs after power termination (m/z 128, 130, 132). That is, the mass spectrum collected in this discharge time regime showed M+ information. It is worth noting that (in contrast to previous reported designs analyzing chlorinated hydrocarbons and aromatic compounds12) protonated parent ions, MH+ (originated from proton-transfer processes), were not detected in our setup, resulting in simpler mass spectra. 2596


These first experiments confirm the analytical potential of the proposed micropulsed GD as a tuneable ion source for real-time chemical speciation at the elemental, structural, and parent molecular ion level. The type of information obtained can be selected in our approach just by changing the repeller delay of the MS(TOF). Analytical Performance Characteristics. Once the ability of the system for qualitative speciation was demonstrated, the analytical performance characteristics of the developed prototype were evaluated at the optimized operating conditions, using bromochloromethane as model analyte. The analytical figures of merit obtained, after such analytical evaluation experiments, are collected in Table 1. Calibration graphs were constructed at each time regime of the GD by measuring the intensities observed for elemental, fragment, and molecular ions of bromochloromethane. Although

Figure 7. Mass spectra obtained with micropulsed dc-GD-MS(TOF) for molecular bromodichloromethane: (a) prepeak region; (b) plateau region; (c) afterpeak of the discharge pulse and comparison with NIST spectra (in a squared box on the right of the figure). Signals in gray in the GD-MS(TOF) spectra correspond to a hexane blank.

elemental and fragment ions were detectable in the prepeak and in the plateau regimes, the highest intensity for elemental ions was obtained in the prepeak, whereas the best signals for analyte fragments were observed in the plateau region. Therefore, only data from quantifications in such differential time regimes are shown in Table 1. The analyte molecular ions were only detected in the afterpeak of the discharge pulse, so the corresponding

calibration data refers to data observed in this time regime of the pulse. In all cases a good linearity was achieved (with regression coefficients >0.99), even for molecular ion calibration, where mass signals are significantly lower. Such good linearity is particularly noteworthy considering that no internal standardization was used for any of these measurements. For elemental and fragment determination, the dynamic ranges were always between 1 and 2 Analytical Chemistry, Vol. 81, No. 7, April 1, 2009

2597

Table 1. Analytical Performance Characteristics for the Determination of Bromochloromethane in Gas Media by Using the Developed Microsecond-Pulsed dc-GD-MS(TOF) System analytea 35

Cl Br m/z 49 m/z 93 m/z 130 79

slope

y-intercept

R2

linear range (µg/L)

DLanalog (µg/L)

DLanalog (pg)

precision (%)/(concn, µg/L)b

0.048 ± 0.004 0.0188 ± 0.0005 0.0042 ± 0.0004 0.0055 ± 0.0005 0.00043 ± 0.00004

0.04 ± 0.07 0.1 ± 0.1 0.02 ± 0.09 0.1 ± 0.1 0.01 ± 0.02

0.993 0.999 0.986 0.987 0.987

86 216 431 431 551

3 1 13 11 238

4.5 1.5 19.5 16.5 357

9 (255) 7 (283) 10 (207) 8 (186) 13 (350)

a Elements are measured in the prepeak region, fragments in the plateau, and the molecular peak (m/z 130) in the afterpeak. b Precision is expressed as the relative deviation for five replicates of a sample with analyte concentration showed in brackets (µg/L).

orders of magnitude. In the case of the parent ion, the dynamic range extended at least up to 550 ppb (maximum concentration assayed). Relative standard deviations observed for five replicate injections were in the average range of 7-13% for the three studied time regimes. Such precision figures achieved can be considered rather satisfactory, considering that samples were manually injected in the exponential dilution chamber. Precisions observed were similar or slightly better than those previously reported.13,22 The corresponding detection limits (DLs) were then calculated (using the elemental ion intensity signals of 79Br+ and 35Cl+ obtained in the prepeak, the intensity of fragment ions at m/z 49 and m/z 93 in the plateau time regime, and the intensity of the m/z 130 molecular ion in the afterpeak). The criteria approached for DLs calculation was the minimum concentration of bromochloromethane producing a net signal equal to 3 times the standard deviation of the background intensity (10 replicates, σB) in each case: DL )

3σB m

(2)

where m is the slope of the calibration curve. Table 1 collects all the calculated DLs, demonstrating that, working in the analog mode of the MS(TOF) system, DLs in the range between 1 and 200 µg/L are achieved (of course, obtained without using any preconcentration technique). The results show that DLs achieved for elemental and fragments detection were improved above 1 order of magnitude as compared with previous results13,22 measuring their analyte ions in the same pulse regimes (comparisons for the molecular ion cannot be made because such data were not reported). In any case, it is important to note that the studied analytes in previous reports are other than bromochloromethane selected in our work. The employed commercial MS(TOF) was originally designed to be operated with a continuous ionization source (e.g., an ICP), and so analog and digital measurement modes, at mass spectral generation frequencies typically of about 20 kHz, can be used as desired. A GD (in continuous mode) was previously coupled to this MS(TOF) for direct solid analysis, and an improvement in the observed DLs was apparent when digital instead of the analog detection mode18 was employed. Coupling a micropulsed ionization source to this MS(TOF) requires a modified electronic cardboard for the TOF to trigger a synchronized MS acquisition system with the external GD pulse generator. Unfortunately, using such cardboard, the spectral acquisition frequency of the MS(TOF) becomes limited by the low frequency rates at which the pulsed 2598


ion source is operated (500 Hz in our experiments). At such low frequencies the MS(TOF) digital acquisition mode did not work, and so the analog mode had to be employed. Operation of this MS(TOF) in digital acquisition mode with our micropulsed GD, however, would be feasible by replacing the used commercial modified cardboard by a new one designed to allow an appropriate triggering of the pulses in the GD, while maintaining the capabilities of the MS(TOF) at its common spectral acquisition frequency of 20 kHz. If such improved detection can be materialized, and considering an average factor of 25 times gain in the DLs attainable with the digital mode,18 DLs for bromochloroethane could be improved to the level of nanograms per liter (values in the range of 40-120 ng/L for elements in the prepeak, 400-600 ng/L for fragments in the plateau, and about 10 µg/L for the molecular ion can be estimated). Research in our laboratory to improve the electronic data acquisition and processing of this GD-MS(TOF) prototype are underway. CONCLUSIONS The capabilities of the developed microsecond-pulsed GDMS(TOF) system for the acquisition of molecular, fragments, and elemental information has been demonstrated. Thus, in a single GD pulse, elemental, structural, and molecular information can be obtained, at least for a volatile organohalogenated compound (used as model analyte). Interestingly, MS spectra acquired during the afterpeak temporal region provide information on just intact molecular ions. Of course, this fact could make molecular mass determinations relatively straightforward in many cases. The proposed system uses a simple design of a homemade GD chamber, that proved to provide a very stable pulsed discharge, which can be directly mounted onto the mass spectrometer interface without the need of a sampler cone. Therefore, the distance of skimmer-chamber can be minimized up to values as low as 3 mm (enhancing significantly the ion transmission to the mass analyzer). Additionally, the small internal volume of this chamber resulted in very low sample (analyte) dilutions allowing for increased sensitivities. The analytical figures of merit obtained compare very favorably with those reported so far in previous related approaches.13,22 In fact, achieved DLs here fulfill the sensitivity requirements of the threshold levels of trihalomethanes in European drinking waters legislation.23 Additionally, the GD chamber proposed has proved to be very resistant to the entrance of vapors of organic solvents (23) European Directive (3rd November, 1998) relative to the quality of waters intended for human consumption; Directive 98/83/EC, 5th December, 1998. Off. J. Eur. Communities: Legis. 1998, L-330, 0032-0054.

(i.e., plasma disturbance is virtually undetectable). High electrical stability, comparatively high sensitivity, and the proven compatibility with the entrance of vapors from organic solvents into this GD design guarantee its special suitability and potential for its future coupling as a chromatographic detector for more practical speciation studies. Particularly, if DLs can be eventually lowered 25 times or so, as discussed in the preceding section, qualitative and quantitative speciation of a wide variety of organic molecules in real samples with this pulsed dc-GD-MS(TOF) instrument can be envisaged by its coupling to classical GC separations.

ACKNOWLEDGMENT This work was supported by project CTQ2006-02309/BQU (Ministerio de Educacioń y Ciencia, Spain). A.S.-V. acknowledges the financial support from the project I3-2007-521-2 (Principado de Asturias). This work has been rewared in JAI 2008 by SEEM (Spain).

Received for review November 28, 2008. Accepted February 2, 2009. AC802520Q


2599

Tuneable Microsecond-Pulsed Glow Discharge Design for the

Recommend Documents