Microplasma Ionization of Volatile Organics for Improving Air/Water Monitoring Systems On-Board the International Space Station
Low molecular weight polar organics are commonly observed in spacecraft environments. Increasing concentrations of one or more of these contaminants can negatively impact Environmental Control and Life Support (ECLS) systems and/or the health of crew members, posing potential risks to the success of manned space missions. Ambient plasma ionization mass spectrometry (MS) is finding effective use as part of the analytical methodologies being tested for next-generation space module environmental analysis. However, ambient ionization methods employing atmospheric plasmas typically require relatively high operation voltages and power, thus limiting their applicability in combination with fieldable mass spectrometers. In this work, we investigate the use of a low power microplasma device in the microhollow cathode discharge (MHCD) configuration for the analysis of polar organics encountered in space missions. A metal-insulator-metal (MIM) structure with molybdenum foil disc electrodes and a mica insulator was used to form a 300 μm diameter plasma discharge cavity. We demonstrate the application of these MIM microplasmas as part of a versatile miniature ion source for the analysis of typical volatile contaminants found in the International Space Station (ISS) environment, highlighting their advantages as low cost and simple analytical devices.
KeywordsPlasma ionization Microhollow cathode discharge Direct analysis in real-time Ion source miniaturization Air quality monitoring
Current in-flight air monitoring systems provide the necessary information to accurately assess air quality in the International Space Station (ISS), but the information available regarding water quality is still limited. Cabin air monitors in the ISS have typically used gas chromatography, ion and differential mobility spectrometry, and mass spectrometry (MS) to monitor part-per-million (ppm) levels of air contaminants [1, 2, 3, 4]. For water samples, electrochemistry and colorimetric solid-phase extraction/reflectance spectroscopy are used to determine total organic carbon and biocide concentrations, respectively [5, 6]. These on-board technologies, however, lack specificity or only focus on a narrow class of analytes, and are not implemented in a fashion that enables simultaneous air and water quality assessment. Therefore, the current environmental monitoring strategy still relies on ground analysis of returned water samples for obtaining more comprehensive chemical information. This limitation, if not addressed, could seriously limit the scope of future space exploration missions beyond low-Earth orbit. To address these shortcomings, more advanced in-flight air and water chemical analyzers are needed to reduce crew time, provide real-time data to manage nominal and off-nominal scenarios, and help assure mission success.
Recent work by our team demonstrated a new electrothermal vaporization (ETV) sample introduction device coupled to a direct analysis in real-time (DART) ambient plasma ion source and a high-resolution time-of-flight mass spectrometer for identifying and quantifying a variety of volatile organic compounds of importance to the ISS environment . The next logical step in the design of a combined air/water analyzer is the miniaturization of the plasma ion source itself by using a microplasma system, with the goal of reducing overall instrument size, required gas flow rates, and electrical power consumption.
Microplasmas are sub-millimeter scale non-equilibrium electrical discharges  that operate with unique geometries , and serve as effective sources for emission spectroscopy [10, 11], chemical ionization MS , and vacuum ultraviolet (VUV) photoionization MS . Steady growth in the development of microplasmas and other low power plasma for analytical chemistry applications has been seen in the recent literature [14, 15], the driving force being the potential for developing small, low power devices with reduced operational cost and increased fieldability [16, 17], with many focused on coupling to portable low resource MS instrumentation [18, 19, 20]. Microplasma-based miniature ion sources operated under ambient sampling/ionization conditions have the additional advantages of high sample throughput and simplicity, enabling direct analysis with little or no sample preparation [21, 22].
Here, we report on the performance of a microhollow cathode discharge (MHCD) microplasma-based ion source for the mass spectrometric analysis of a series of organic analytes of relevance to the ISS environment. The performance of the MHCD is compared with a commercial DART ion source. To the best of our knowledge, this is the first attempt to use MHCD microplasmas for the ionization of analytes of environmental importance in space exploration.
The target compounds were selected to include those representative of low-molecular-weight polar organics that are routinely detected in the ISS atmosphere and water. Analytes tested included methanol, ethanol, isopropanol, acetone, methyl ethyl ketone, and ethyl acetate. These were purchased from Fisher Scientific Inc. (Pittsburgh, PA, USA) and Sigma Aldrich (St. Louis, MO, USA) and used as received (>98% purity) without further purification. Gases used with both the MHCD and DART ion sources included industrial grade nitrogen (99.995%), ultrahigh purity helium (99.999%), and ultrahigh purity argon (99.999%) obtained from Airgas (Atlanta, GA, USA).
Direct Analysis in Real-Time
A detailed description of the commercial DART-SVP ion source (IonSense Inc., Saugus, MA, USA) is available elsewhere . Briefly, this ion source consists of a point electrode spaced ≤1 mm from a grounded disc electrode. A negative DC potential bias of several kV is applied to the point electrode. The effective discharge potentials vary depending on source gas type, and were measured using a high voltage probe (Fluke 179 True RMS multimeter) to be approximately –400 V for helium and argon, and –1850 V for nitrogen. Undesired background ions were filtered from the post-discharge plasma gas stream by a molybdenum grid electrode (+200 V) positioned at the exit nozzle, ensuring that metastable species expelled into the ambient ionization region were the primary reactants. The DART source ceramic heater was left powered off for an accurate performance comparison with the MHCD microplasma source, based purely on plasma characteristics and operating at ambient temperature. The DART exit nozzle was optimized at a position 5–8 mm away and centered on axis with the mass spectrometer inlet. DART gas flow rates were adjusted from 0.15 L min–1 to 2 L min–1 to evaluate ionization efficiency under typical DART settings and settings comparable to those used for the MHCD microplasma ion source.
TOF MS Instrumentation
Both the DART and MHCD ion sources were coupled to an orthogonal time-of-flight (TOF) mass spectrometer (JEOL AccuTOF, Tokyo, Japan). The mass spectrometer was operated in positive ion mode. Typical ion optics conditions were as follows: orifice lens 1 set to 10 V, ring lens at 6 V, and orifice lens 2 at 2 V with the desolvation temperature set to 100 °C. The ion guide reference voltage was set at –29 V and the pusher bias voltage was –0.28 V. Data was acquired using the JEOL Mass Center software in the 10–300 m/z range. The AccuTOF mass spectrometer provided a mass resolution ranging between 1700 and 4100 (FWHM) across the 18–121 Da range, corresponding to the target analytes.
Results and Discussion
Optimization and Comparison to DART
Qualitative analysis of each of the four volatiles tested (Figure 2b–e) was conducted by coating the end of a borosilicate glass melting point capillary with neat sample solutions of each and holding the capillary probe approximately 2 mm away from the ion source gas outlet. Experimental variability originating from sample capillary positioning and the amount of solution deposited on the probe was mitigated by averaging peak areas from repeated measurements. All target compounds were observed mainly as their protonated monomer [M + H]+ and dimer [2 M + H]+ ions, with the dimer ions in a majority of cases dominating the spectra. Some additional [M+ NH4]+ analyte species were also observed, but in lower abundances than protonated species. The detection of mostly clustered ions is attributed to the low inlet orifice potentials used, the neat solution concentrations sampled, and the unheated nitrogen ion source gas. Overall, the total ion signal intensity was lower for all analytes using the MHCD (approximately 35% of the DART signal on average), but the identities and relative ratios of ionic species for methanol, ethanol, or isopropanol did not change significantly between spectra for either ion source. For methanol, [M + H]+ ions at m/z = 33, [M + NH4]+ ions at m/z = 50, [2 M + H]+ ions at m/z = 65, and [3 M + H]+ ions at m/z = 97 were observed for both DART (Figure 2b left) and MHCD (Figure 2b right). Likewise, each ethanol spectrum presented [2 M + H]+ ions at m/z = 93, [M + NH4]+ ions at m/z = 64, [M + H]+ ions at m/z = 47, and [3 M + H]+ ions at m/z = 139, in descending order of intensity. The only prominent exception to the observed matching ion distributions between DART and MHCD spectra was for acetone, where the [M + NH4]+ signal equated to ~90% of the [2 M + H]+ signal for DART, but only 5% of the [2 M + H]+ abundance for the MHCD. It is not at all surprising that among these four simple volatile organic compounds, acetone showed the greatest variability between the two ion sources, given the DART’s and MHCD’s drastically different flow rates and fluid dynamics. Of all analytes, acetone has the highest vapor pressure at 0.53 atm , whereas methanol, ethanol, and isopropanol possess vapor pressures of 0.07, 0.15, and 0.13 atm, respectively . Likewise, the nature of the functional groups in each case (alcohol versus ketone) may offer very distinct stabilization of NH4+ adducts and different protonation sites, which would clearly have an influence on the relative spectral abundances. Such behavior could be useful in deciphering volatiles based on their fingerprint spectra (i.e., [M + H]+ versus [M + NH4]+ versus [2 M + H]+) when low resolution mass analyzers are employed. The differences in sensitivity between the MHCD and DART ion sources seen in Figure 2 are a consequence of differences in parameters such as gas flow rate and plasma current. The higher predicted gas velocity through the reduced dimensions of the MHCD disc electrode (i.d. = 300 μm) could also be a factor affecting ion transport under these conditions. The plasma generation conditions for each ionization source may also play a role in terms of differences in metastable abundances. However, the DART plasma operates in the corona-to-glow transition mode [25, 26], which is relatively similar to the MHCD glow mode, so in the absence of direct spectroscopic measurements, a solid conclusion cannot be drawn. Nevertheless, it appears that the limited scale and lower current of the MHCD are the primary reasons for a lower ion signal compared to the DART source, at least when the comparison is performed under gas flow regimes where DART is typically operated [26, 29].
To fully implement a system utilizing MHCD into the ISS, the ionization source must be characterized for limits of detection and ability for quantitation. Approximate limits of detection for methanol, ethanol, isopropanol, acetone, as well as methyl ethyl ketone (MEK) and ethyl acetate were determined and are shown in Supplemental Information (Table S1). Using the same capillary introduction system and using nitrogen plasma gas at a flow rate of 0.100 L/min, estimated LODs were defined as the minimum amount of material on the borosilicate capillary placed between the MHCD source and MS inlet that could be detected with high confidence. The values were found to be 3000 pmol (100 ppm) for methanol, 2000 pmol (100 ppm) for ethanol, 170 pmol (10 ppm) for isopropanol, and 430 pmol (25 ppm) for acetone. For MEK and ethyl acetate, LODs of 350 and 110 pmol (25 and 10 ppm) were achieved, respectively. These two compounds compare very well to previous work using our ETV source, which achieved quantitative limits of detection of 194 pmol for MEK and 2000 pmol for ethyl acetate injected into the sampling system . For the three organic compounds with Spacecraft Water Exposure Guideline (SWEG) values, methanol, acetone, and MEK, 100-day allowable exposures were found to be 40, 150, and 54 ppm, respectively [27, 28]. While the methanol detection limit is 2.5 times higher with the current MHCD set-up than SWEG values, we believe much of the sample is being lost when placing the borosilicate capillary in an open pathway between the source and MS inlet. MEK and acetone detection limits, however, were lower than the recommended values stated in the exposure guidelines. Using a modified introduction method where sample loss is minimized through more efficient ion/neutral collection, focusing, and mixing of the plasma gas, LODs could be lowered dramatically enabling quantitation, as previously demonstrated using the ETV set-up coupled to DART .
Effect of Gas Type and Flow
Regardless of the plasma gas chosen, all microplasma spectra showed mainly [2 M + H]+ proton-bound dimers, as was the case in Figure 2. It is clear for all test volatiles that the most intense [M + H]+ and [2 M + H]+ signals were achieved with helium while sensitivity for argon and nitrogen remained essentially equal for each analyte. In the case of methanol (Figure 3a), ethanol (Figure 3b), and isopropanol (Figure 3c), the dimer intensity was about 3–5 times higher using helium than when using argon and nitrogen. For acetone, the signals for [2 M + H]+ with nitrogen and argon source gas were significantly greater, at approximately 50% and 75% of the dimer intensity detected using helium, respectively. Given the higher vapor pressure of acetone mentioned before, the higher acetone dimer ion abundances observed are reasonable. The much larger ion signal observed for helium was expected, since its higher thermal conductivity and metastable energy result in greater signal than is possible for either nitrogen or argon. Furthermore, it is of note that in order to achieve the highest intensity spectra for helium, an optimized flow rate of 1.0 L min–1 was used whereas for argon and nitrogen optimal signal was achieved at flow rates of 0.30 and 0.50 L min–1, respectively. For higher density plasma gases such as Ar and N2, higher flow rates appeared to destabilize the discharge, requiring the flow rates where maximum signal was achieved to be about 30%–50% of the optimized flow rates for helium.
In Supplementary Figure S2 it is seen that formation of [M + NH4]+ adducts for each of the volatile analytes of interest was much more pronounced with DART than with the MHCD ion source at all flow rates tested, and with all plasma gases. Typically, MHCD [M + NH4]+ ion signals averaged 0.1%–5.0% of those produced by helium DART. Given the lower ionization potential of ammonia (10.02 eV) versus that of water (12.62 eV), and the respective proton affinities of ~850 kJ mol–1 for ammonia and ~690 kJ mol–1 for water, the probability of ammonia protonation over water is greater, and ammonia adduction with an analyte may be preferred in instances where the analyte proton affinity is lower than that of ammonia [30, 31].
As seen in Figure 4, regardless of the plasma gas used or analyte measured, the DART signals decreased drastically if gas flow rates fell below 1.0 L min–1. Although decreasing the DART-inlet distance could improve ion transmission at lower flow rates, for gases such as helium the increased gas load on the mass spectrometer vacuum system could have short- and long-term detrimental effects. Additionally, it is suspected that for nitrogen plasma gas, the longer path length between the DART discharge chamber and sampling region may contribute to depletion (or relaxation) of metastable excited-state species in the plasma afterglow prior to interaction with the analyte, comparatively reducing ion signal intensity. Regardless, the nitrogen MHCD does perform well at a fraction of the flow rate needed for DART, as it also does with argon and helium. For this reason, MHCD microplasmas are an attractive alternative when considering the limited resources aboard semi-autonomous space systems such as the ISS.
Power and Lifetime Characterization
The device power consumption depended on the discharge gas, as different applied voltages were required in each case to sustain an equivalent current. The power curves for each gas (Figure 5) strongly indicate there is a clear difference in the power profile and the wattage used for nitrogen gas compared with argon and helium gases, as expected. In addition, it was observed that when using nitrogen and argon, the lifetime of the electrode at higher currents was significantly longer compared with helium. For this gas, lifetimes at powers above 0.120 mA were approximately 2–4 hours, with evident signs of electrode deterioration along the edges of the bored cavity, resulting in performance loss. Operation with nitrogen and argon was more robust and consistent even at these higher currents, with the ability to operate at over 24 h of constant use. Furthermore, operation at lower powers (<50 uA) increased lifetimes to over 100 h of operation for all plasma gas types examined. It is speculated that the rate of device degradation using helium is caused by the higher metastable energies  and higher plasma gas temperature because of the superior heat capacity, thermal conductivity, and excited state distributions of helium versus nitrogen and argon. Driving MHCD devices with AC waveforms rather than a DC bias could further reduce electrode sputtering and prolong lifetime . However, if operated with currents in the order of 0.050 mA, and equipped with molybdenum electrodes (E1 and E2, Figure 1), we expect MHCDs could provide reliable operation for months if powered intermittently or upwards of several days when operating the discharge continuously.
MHCD operation using nitrogen provided an ideal balance of low power consumption and gas cost while affording adequate sensitivity and device longevity. As a possible alternative to nitrogen, argon microplasmas showed a lower onset discharge potential than helium (Figure 5), and provided a similar sensitivity to nitrogen for volatile analysis at low gas flow rates (Figure 3).
This study demonstrated the performance capabilities of a MHCD microplasma ion source for detection of organic volatiles important in environmental monitoring applications. The low power consumption and compact design make this type of plasma device a promising alternative ion source for miniature and field mass spectrometers. Both the DART and microplasma ion sources produced similar reactant ion populations and relative distributions of analyte ions, with higher-order ion clusters detected in abundance in both cases, given the relatively “soft” mass spectrometer transfer optics settings used. Analytes were observed as protonated neutrals and ammoniated adducts in most cases, with protonated dimers being the most common. Comparison experiments with various plasma gases provided significant evidence that the source of ammonia was ambient and that between nitrogen, argon, and helium at equivalent power use, there was a clear advantage for helium use in terms of sensitivity, but not power consumption or device durability.
In practice, nitrogen usage as the plasma gas would provide the most practical approach for implementing MHCD ionization aboard the ISS. It is available in large quantities and given the low flow rate required, should not make a significant impact on existing resources. It has been previously used for the volatile organic analyzer (VOA)  and despite the power and signal performance observed for nitrogen operation, it would be a promising avenue for further development in designing a better water and air analysis platform. However, additional research is needed to enable the detection and quantitation of lower levels of analytes in these environments. Future studies will be aimed at improving ion transfer from the AP ionization region into the mass spectrometer. Other pursuits include device operation in AC versus DC power modes to assess device lifetime and sampling efficiency, and investigation of more complex molecular systems using flow-through sample introduction schemes for practical use aboard a manned space vessel, such as the ISS.
M.C.B, P.D., J.M.S., and F.M.F. acknowledge support through NASA award number NNX13AF51G. J.D.K., T.M.O., and F.M.F. acknowledge additional support through the Center for Chemical Evolution, jointly sponsored by NSF and the NASA Astrobiology Program (NSF CHE-1004570). W.T.W and T.F.L. acknowledge funding under NASA contract no. NAS 9-02078.
- 4.Limero, T., Martin, M., Reese, E.: Validation of the volatile organic analyzer (VOA) for ISS operations. Int. J. Ion Mobil. Spectrom. 6, 5–10 (2003)Google Scholar
- 9.Bruggeman, P., Brandenburg, R.: Atmospheric pressure discharge filaments and microplasmas: physics, chemistry and diagnostics. J. Phys. D Appl. Phys. 46, (2013). doi:10.1088/0022-3727/46/46/464001
- 18.Kumano, S., Sugiyama, M., Yamada, M., Nishimura, K., Hasegawa, H., Morokuma, H., Inoue, H., Hashimoto, Y.: Development of a portable mass spectrometer characterized by discontinuous sample gas introduction, a low-pressure dielectric barrier discharge ionization source, and a vacuumed headspace technique. Anal. Chem. 85, 5033–5039 (2013)CrossRefGoogle Scholar
- 24.Haynes, W.M. (ed.): Vapor pressures of organic compounds. In: CRC Handbook of Chemistry and Physics, Internet Version 2016, 96th edn, section 6, p. 92. CRC Press/Taylor and Francis, Boca Raton, FL (2016)Google Scholar
- 27.Garcia, H.D.: Spacecraft water exposure guidelines for selected contaminants, vol 2, p. 29. National Research Council, National Academy Press, Washington, DC (2007)Google Scholar
- 28.Garcia, H.D.: Spacecraft water exposure guidelines for selected contaminants, vol 3, p. 139 and 160. National Research Council, National Academy Press, Washington, DC (2008)Google Scholar