Abstract
The formation and identification of prebiotic compounds in the organically rich atmospheres of Titan and Pluto are of great interest due to the potential implications such discoveries may have on theories of the origins of life on the early Earth. In past work, hindrances in detecting prebiotic molecules in lab-generated aerosol analogs have been the large number of products formed, often compounded by limited sample amounts. In this work, we detail a GC/MS/MS protocol that is highly selective (>30 simultaneously detectable compounds) and highly sensitive (limits of detection ∼1 picomole). Using this method to analyze aerosol analogs (tholins) generated by either cold plasma or photochemical irradiation of 1:1 mixtures of methane and carbon monoxide in nitrogen, this work has expanded the number of identifiable compounds in Titan/Pluto analog aerosols to include the nonbiological nucleobases xanthine and hypoxanthine in plasma aerosols and the first identification of glycine as a product in photochemical aerosols produced under reducing atmospheric conditions. Several species (glycine, guanidine, urea, and glycolic acid) were found to be present in both plasma and photochemical aerosols. Such parallel product pathways bring new understanding to the nature of plasma and photochemical aerosols and allow for new insights into the prebiotic chemistry of organically rich atmospheres including Pluto, Titan, and the early Earth.
Export citation and abstract BibTeX RIS
1. Introduction
The Oparin–Haldane paradigm posits that early life was formed abiotically from organic molecules that initially formed in the early atmosphere of Earth before settling out into primordial pools where aqueous reactions started the first metabolic processes. These early organisms were heterotrophic and fed upon the same abiotically produced organic compounds from which they arose (Oparin & Morgulis 1938). Early experimentation by Miller and Urey supported this hypothesis (Miller & Urey 1959). Using atmospheres of CH4/NH3/H2O that cycled continuously through a chamber with induced electrical sparks, the residue collected during these experiments was found to have amino acids, carboxylic acids, and hydroxy acids: molecules that could have supported early metabolisms. Since Miller and Urey, new experimentation and data accumulation has redefined the composition of the atmospheres under investigation (Kharecha et al. 2005; Domagal-Goldman et al. 2008; Shaw 2008), but the formation and detection of prebiotic molecules in primordial atmospheres still remains of great interest (Trainer et al. 2013).
The exact makeup of the Hadean and Archean atmospheres is currently unknown and continues to be a subject of debate. Models for methane vary from 100 ppm (Kiehl & Dickinson 1987) to over 1000 ppm (Kharecha et al. 2005). Other species in the atmosphere include carbon dioxide, sulfur dioxide, hydrogen, ammonia, and water (Oparin & Morgulis 1938; Kiehl & Dickinson 1987; Domagal-Goldman et al. 2008; Trainer 2013) in a predominantly nitrogen environment. As these ancient atmospheres no longer exist, planetary scientists use the atmospheres of Titan, Saturn's largest moon, and Pluto as analog primordial environments. (Sagan et al. 1992; Coustenis 1994; Clarke & Ferris 1997; McKay et al. 1999; Trainer et al. 2006; Neish et al. 2008; Lunine 2009; Trainer 2013).
With the abundance of data gathered by NASA's Cassini–Huygens mission, many comparisons of Titan's atmosphere to that of the early Earth have been made. Titan's atmosphere has methane concentrations ranging from ∼5% near the surface to ∼1.5% in the lower atmosphere (Niemann et al. 2005, 2010) with the bulk of the atmosphere being molecular nitrogen. A few other trace species, including carbon monoxide (∼50 ppm), make up the rest of the gas phase species. (Lutz et al. 1983; de Kok et al. 2007; Bellucci et al. 2009; Courtin et al. 2011). Titan is shrouded in an orange haze layer composed of organic aerosols (Sagan et al. 1992; Cable et al. 2012; Hörst 2017). Pluto has only recently risen to prominence in the area of photochemical hazes because of the recent flyby by the New Horizons spacecraft. Within the data sent back by the spacecraft is clear evidence of extended haze layers within the dwarf planet's atmosphere (Gladstone et al. 2016). Pluto's atmosphere is predominantly methane (0.2%–0.6%; Lellouch et al. 2009, 2015; Stern et al. 2015), carbon monoxide (∼0.05%; Greaves et al. 2011; Lellouch et al. 2011), and nitrogen. With the combination of CH4, CO, and N2 both Titan and Pluto have the potential to create biologically important compounds such as amino acids and nucleobases in their atmospheres.
In general, the organic chemistry in atmospheres is initiated by either energetic particles (electrons, protons, and cosmic rays) or solar photons (Waite et al. 2007; Krasnopolsky 2009; Lavvas et al. 2011). While both processes can occur simultaneously, experimentally the two cases are studied separately using plasmas as a stand in for the electrons, protons, and cosmic rays and moderate to high intensity UV light sources (λ > 400 nm) to mimic aspects of solar radiation (Cable et al. 2012; Sebree et al. 2018). The aerosol analogs, often referred to as tholins, that result from these setups are composed of a complex mixture of thousands of different products with variable yields and composition. While plasma generated aerosols made from CH4 and nitrogen can yield on the order of 10 to 20 mg hr−1 or more depending on pressure and gas composition (Sciamma-O'Brien et al. 2010; Hörst et al. 2012; Hörst & Tolbert 2014), photochemical yields are 100 to 1000 times less (Trainer et al. 2006). This inherently lower yield of photochemical aerosols imposes minimal time limits for producing an analyzable quantity of material. In some cases, yields are so low as to make some analytical studies impossible (Tran et al. 2003; Coll et al. 2013). While higher photochemical yields have been obtained with the addition of other gas species (e.g., aromatics (Trainer 2013; Sebree et al. 2014; Yoon et al. 2014; Gautier et al. 2017a, 2017b), carbon monoxide (Hörst & Tolbert 2014), carbon dioxide (Trainer et al. 2006), and others (Tran et al. 2003, 2005, 2008)) to traditional methane/nitrogen gas mixtures, experimental production times for single samples can still run for hundreds of hours to generate sufficient material for ex situ analysis (Sebree et al. 2014, 2016; Gautier et al. 2017a, 2017b). In order to better understand the chemistry taking place in both plasma and photochemical aerosols, selective trace detection methods are needed that can detect small amounts of material in the complex matrix of aerosol analogs.
In previous work on plasma aerosols, GC/MS was used in addition to other techniques including ultra-high-resolution mass spectrometry (Orbitrap; Hörst et al. 2012; Gautier et al. 2014) and nuclear magnetic resonance (NMR) spectroscopy (He & Smith 2013). Starting from gas mixtures of methane and nitrogen, He & Smith (2013) isolated several small (<130 Dalton) compounds including cyanamide, guanidine, and 1,2,4-triazole. By using a combination of direct injection or derivatization of their tholins to increase the volatile nature of the products prior to GC/MS injection, Gautier et al. (2014) identified ∼20 molecules in 50–150 Da mass range (compared to >15 000 with Orbitrap MS) from their CH4/N2 aerosols. The majority of the compounds were polymeric in nature. While alanine and urea were detected, there were no oxygen source gases, so their presence was ruled to come from either contamination or the post-formation handling process. Starting from gas mixtures of CH4/CO/N2 exposed to cold plasma, Hörst et al. (2012) identified five biological nucleobases (cytosine, adenine, thymine, guanine, and uracil), and 15 of the biological amino acids using Orbitrap MS. However, only the structures of the five nucleobases, glycine, and alanine could be confirmed using traditional GC/MS. As is often the case when working with small amounts, selectivity and sensitivity often must be weighed against each other.
Tandem mass spectroscopy (GC/MS/MS) techniques using GC-triplequad setups offer a middle ground between the selectivity of a GC/MS and sensitivity of an Orbitrap. In GC/MS/MS, a parent mass from a particular compound (Figure 1) is trapped in the first quadrupole (Q1), ignoring all other masses. After a sufficient time, the ions are shuttled from Q1 to Q2 where they are fragmented using collisionally induced dissociation (CID) before the resulting daughter masses are scanned by the third quadrupole (Q3). This process has a two-fold benefit of preconcentrating a particular analyte of interest, and suppressing the background noise of other species that may coelute from the GC at the same time (Figure 1, bottom). Further noise suppression can be obtained by using Q3 as a mass filter and only allowing one or two masses (qualifier and quantifier) associated with the fragmentation to pass through to the detector. The combination of elution time, parent mass, qualifier, and quantifier have a combined effect of improving detection limits by up to 100x over GC/MS while maintaining the possibility of species identification.
Figure 1. Traditional GC/MS (MS1) of trisTBDMS-Xanthine. The parent peak at m/z = 437 is then trapped and fragmented using GC/MS/MS methods, resulting in an increased signal-to-noise ratio in the MS/MS results.
Download figure:
Standard image High-resolution imageGC/MS/MS methods have previously been used to detect metabolites (amino acids) in urine matrixes (Kvitvang et al. 2011), saccharides in urban (Earth) aerosols (Pashynska et al. 2002), and nucleobases in carbonaceous meteorites (Callahan et al. 2011), but have not been applied to primordial aerosol analogs. In this work, we use GC/MS/MS methodology in the detection of molecules in primordial aerosol analogs. This work both confirms the detection of many of the species in previous works (Hörst et al. 2012; He & Smith 2013; Gautier et al. 2014), and expands the number of species identified in both plasma aerosols and photochemical aerosols.
2. Experimental
2.1. Aerosol Production
Plasma aerosol were prepared using the Planetary HAZE Research (PHAZER) experimental setup at Johns Hopkins University (He et al. 2017). The gas mixture of CO, CH4, and N2 (5:5:90) was prepared in a stainless-steel cylinder and allowed to mix for a minimum of 12 hours before running the experiment. The gases continuously flow through a 15-meter stainless steel cooling coil immersed in liquid nitrogen (77 K), which cools the gases down to about 100 K and removes trace impurities in the gases. The temperature of the gases in the reaction chamber was determined based on Gay–Lussac's gas law (P1/T1 = P2/T2) before starting the experiment. After 1 hr flowing of the cold gases, valves on both sides of the chamber were closed and the pressure (P1) of the cold gases in the chamber was measured. Pressure (P2) was subsequently measured in the chamber after the gases have warmed up to 294 K (T2, room temperature).
The cold gases then flow through a stainless-steel reaction chamber where they are exposed to a glow discharge produced by AC power supply (6 kV/10 mA) initiating chemical processes that lead to the formation of new gas phase products and particles. Note that the AC glow discharge is a cold plasma source in which only a small fraction (for instance: a few percent) of the gas molecules are ionized with energies in the 5–15 eV range and the neutral gas heating is low, and therefore the total gas temperature is not significantly altered by the plasma. The produced plasma particles are not directly comparable to the energetic particles in Titan's atmosphere (Cable et al. 2012), but they provide similar range of energy to initiate chemical reactions in the chamber. The gas flow rate is maintained at 10 standard cubic centimeters per minute (sccm) by a mass flow controller (MKS Instrument, GM50A), so that the pressure in the reaction chamber is held at 2 Torr. The glow discharge operates for 72 hr. The continuous flow of the reactant gases through the glow discharge allows an average exposure time to plasma of about 3 seconds. After 72 hr of discharge flow, a red/brown film is deposited on the wall of the reaction chamber. The chamber is under vacuum for 48 hr to remove the volatile components, and then transferred to a glove box (Inert Technology Inc., I-lab 2GB) where the solids are collected under a dry N2 atmosphere (<1 ppm O2, H2O). The solids are stored in sealed glass vials, wrapped in aluminum foil, in the glove box until further analysis. To prevent contamination of the aerosols, prior to production, the chamber was first cleaned with ALCONOX powder cleaner, followed by ultrasonic cleaning for 30 minutes, rinsed with HPLC water for three times and with methanol for three times, and dried and baked in a 120 °C oven overnight. The chamber was then pumped for 24 hr under vacuum (10−2 Torr).
Photochemical aerosols were made using a UV-photolysis, continuous-flow chamber similar to those used previously (Trainer 2013; Sebree et al. 2014, 2018). The photochemical gas mixtures (Table 1) were created using benzene (Sigma Aldrich, ≥99.9%), pyridine (Sigma Aldrich, ≥99.9%), or methane (Airgas, 99.99%) and carbon monoxide (Sigma Aldrich, ≥99.0%) in nitrogen (Airgas, 99.999%). The mixtures of 50 ppmv benzene or pyridine were selected as medium yield (∼0.1–0.2 mg hr−1) photochemical aerosols (Trainer 2013; Sebree et al. 2014) to validate the GC/MS/MS method using an organic matrix that would not contain amino acids. For comparison to previous work on methane/carbon monoxide aerosols (Hörst et al. 2012; Hörst & Tolbert 2014), a mixture composed of 0.1% methane and 0.1% carbon monoxide in nitrogen was used.
Table 1. Starting Gas Mixtures and Energy Sources of Aerosols Produced
| Energy Source | CH4 Conc. (%) | CO Conc. (%) | Aromatic Conc. | Yield (mg hr−1) |
|---|---|---|---|---|
| Plasma | 5.0 | 5.0 | ∼10 | |
| Photochemical | 0.1 | 0.1 | ⋯a | |
| Photochemical | 50 ppm Benzene | ∼0.1 | ||
| Photochemical | 50 ppm Pyridine | ∼0.2 | ||
Note.
aMass yield was below detectable limits.Download table as: ASCIITypeset image
Gas mixtures were allowed to mix for a minimum of eight hours to ensure homogeneity before flowing through the reaction vessel at 20 sccm at a total pressure of 600 torr. An air-cooled deuterium lamp (Hamamatsu, L11798) with MgF2 windows, emitting from 115 to 400 nm (3.1 to 10.8 eV) with the maximum emission features between 121 nm (Lyα) and 160 nm was inserted directly into the reaction cell for UV photolysis. The average VUV flux for the lamp was previously determined to be ∼7.2 × 1015 photons s−1 (Sebree et al. 2018). The entrained aerosols were then collected on a PTFE filter in a collector out of direct line of sight from the lamp. The collector could then be decoupled under vacuum and transported to a nitrogen purged gloved box for derivatization.
2.2. GC/MS/MS
GC/MS/MS method validation was carried out using a combination of standards (Table 2). The seven nucleobases (Sigma Aldrich, >99.0%) were dissolved in 0.1M NaOH at a concentration of ∼2 mg ml−1 for each base. A physiological amino acid (AA) standard (Sigma Aldrich) with 27 detectable components at 0.5 μmol mL−1 was used in tandem with a separate standalone standard (Sigma Aldrich, all ≥99.0%) of norvaline, norleucine, guanidine, glycolic acid, and methionine dissolved in 0.1 M HCl at ∼2 mg ml−1. 10 μL of each of a standard solution was then evaporated in a GC/MS vial in an oven (∼75–80 °C) under dried nitrogen. Several drops of dichloromethane were added to evaporate any remaining solvent. After all liquid was evaporated, 30 μL of dimethylformamide (DMF) and 30 μL of N-tert-butyldimethylsilyl-N-methyltrifluoroacetamide (MTBSTFA) were added, and the solution was left in the oven to derivatize for 30 minutes under dried nitrogen flow. After derivatization was complete, 1 μL of the solution was injected onto an Agilent 7000c Triple Quad. Separation of compounds was achieved with a Restek capillary column (RTX-5MS) with the He flow held constant at 1.3 ml minutes−1. The operating conditions were as follows: initial temperature, 100 °C; ramp rate, 10 °C minutes−1 to final temperature 270 °C; hold time 11.5 minutes. Figure 2 presents the GC/MS chromatogram for the nucleobase (a) and a combined amino acid (b) standards. CID energies of 15 keV were used during MS/MS operations. Each standard was run in a traditional (single) MS mode (MS1) and in tandem mode (MS/MS) to serve as a reference library (Figure 1).
Figure 2. GC/MS chromatograms for (a) the nucleobase standard and (b) the amino acid standard. For figure clarity, the peaks labeled with asterisks (*) denote the compounds in the physiological standard in the order of Table 2. Unlabeled peaks in both traces are expected background peaks from the derivatization process. The derivatized citrate is a preservative in the purchased standard.
Download figure:
Standard image High-resolution imageTable 2. Standards Used for GC/MS/MS Method Calibration
| Nucleobase Standards | Uracil, Thymine, Cytosine, Hypoxanthine, Adenine, Xanthine, Guanine |
| Physiological Amino Acid Standard (0.5 μmol mL−1) | Ethanolamine, Alanine, Glycine, Sarcosine, Beta-Isobutyric Acid, Beta-Alanine, Urea, Valine, Leucine, Isoleucine, Proline, Taurinea, Methioninea, Serine, Threonine, Phenylalanine, Aspartic Acid, Hydroxy-L-Proline, Cysteinea, Glutamic Acid, Ornithine, Homocysteinea, Methyl-Histidine, Lysine, Histidine, Tyrosine, Tryptophan |
| Additional Standards | Norvaline, Norleucine, Guanidine, Glycolic Acid, Methioninea |
Note.
aSulfur containing species, monitored as derivatization reaction standards.Download table as: ASCIITypeset image
Aerosol analog samples were prepped for analysis similar to previous methods (Rodier et al. 2001; Buch et al. 2009; Callahan et al. 2011; Hörst et al. 2012; Gautier et al. 2014). Where samples that were sufficient in size to be measured directly (plasma-produced and aromatic aerosols), aerosols were dissolved in a 50:50 methanol/acetone solution (5 mg ml−1) and the soluble portion was used for analysis. For the lower yield CH4/CO photochemical aerosol, a direct massing of the aerosol was not possible due to the low yields, rather the PTFE filter was washed with the 50:50 solution and the wash was collected. A 10 μL spike of a 2 mg ml−1 standard of methionine (Sigma Aldrich, >99.0%) in 0.1 M HCl was added as an internal reaction standard to each of the CH4/CO aerosol solutions. A 10 μL spike of the AA standard and nucleobase standard was added to the benzene and pyridine aerosol solutions for MS/MS validation. The solutions were then evaporated and derivatized as outlined above. The 1 μL of the resulting solutions was then injected into the GC triple quad running either in MS1 or MS/MS mode. Empty vials and PTFE filters were treated under the same conditions and carried through as blanks and contamination checks.
3. Results
3.1. MS/MS Extraction Validation
Figures 3 and 4 illustrate the efficiency of MS/MS when performing targeted species analysis in the complex aerosol matrix. Photochemistry of benzene or pyridine is not expected to yield any of the 38 compounds targeted in this work. With no oxygen source gases used in the starting mixtures, all the amino acids and six of the seven nucleobases could not be created, except through contamination. As shown in Figure 3(a), both the benzene and pyridine aerosols have a significant number of background species in a traditional MS1. The pyridine (bottom trace), with the possible formation of amine groups, includes both derivatized and underivatized species in the chromatogram.
Figure 3. Comparison of (a) GC/MS chromatograms to (b) GC/MS/MS chromatograms for a mixed standard of amino acids and nucleobases free of aerosol matrix (top traces), in a benzene aerosol matrix (middle traces), or in a pyridine aerosol matrix (bottom trace). The peak marked with an asterisk (*) is a derivatization side product.
Download figure:
Standard image High-resolution imageThe use of targeted trapping and CID fragmentation (Figure 1) from MS/MS significantly improves signal-to-noise ratios and suppresses the matrix background for all three samples (Figure 3(b)). The MS/MS method has the added benefit of removing the side products that happen from derivatization, leaving only the selected compounds of interest in the final chromatogram.
As evidenced in Figures 3 and 4, when identical standards are derivatized both with and without an aerosol matrix a GC run in single MS mode (MS1) will show all the standards, in addition to the derivatization background, and the aerosol background. Depending on how complicated the matrix is, direct deconvolution of the data may not be possible (Figure 3(a), bottom trace). However, the use of MS/MS does present the opportunity to detect only the species of interest with little to no loss of signal. As shown by the red trace in Figure 4, a direct comparison of the MS/MS chromatograms of a standard to that of a standard with a matrix, shows only minor changes in peak shape.
Figure 4. Comparison of the spike recovery in MS/MS. Chromatograms run using a single MS (MS1) of an AA spiked benzene aerosol (top trace) compared to a pure AA standard (bottom trace) show distinct differences in the number of peaks and background. The MS/MS scans (black traces) are nearly identical as shown by the residual (red trace) created by subtracting the two black traces. Figure note: all green and black traces have been offset vertically and intensities of the bottom two traces have been inverted to make comparisons easier. Relative intensities have not been changed.
Download figure:
Standard image High-resolution image3.2. Detection of Prebiotic Molecules
For the purpose of this discussion, it is important to note that "prebiotic" is used to refer to any abiologically formed, organic species that could have formed on Earth prior to the formation of life. The term "biological" is used to refer to those compounds used by life today. Of the 34 possible compounds targeted in this work, 16 were identified in the plasma and/or photochemically produced CH4/CO aerosols. The complete results are summarized in Table 3. Despite having an estimated 1000x less photochemically produced aerosol compared to the plasma-produced aerosol, it was possible to identify four compounds (m/z < 90 Daltons) using the tandem MS/MS method: guanidine, urea, glycine, and glycolic acid. Of the four, glycolic acid was present in sufficient amounts, as was observed in the MS1 scan (Figure 5, top plot). Comparison of the sample mass spectrum with that of the standard was used to verify the peak assignment as glycolic acid.
Figure 5. Top plot: MS1 chromatogram of photochemical aerosol (showing significant signal for glycolic acid (boxed). Bottom plot: GC/MS mass spectra of glycolic acid expanded to show the full mass range of the fragments produced by dissociative ionization of the derivatized molecules. Strong agreement between the standard (top trace) and the sample (bottom) verifies the identity of glycolic acid. Peaks marked with asterisk (*) are derivatization background peaks.
Download figure:
Standard image High-resolution imageTable 3. Summary of Identified Molecules from CH4/CO Aerosols in This Work
| Name | M/Z | Plasma | Photochemical |
|---|---|---|---|
| Nucleobase | |||
| Cytosine | 111 | MS1a | |
| Uracil | 112 | MS1a | |
| Thymine | 126 | MS1a | |
| Adenine | 135 | MS1a | |
| Hypoxanthine | 136 | MS1 | |
| Guanine | 151 | MS1a | |
| Xanthine | 152 | MS1 | |
| Amino Acid | |||
| Glycine | 75 | MS1a | MS/MS |
| Alanine | 89 | MS1a | |
| Sarcosine | 89 | MS/MS | |
| Beta-aminoisobutyric acid | 103 | MS/MS | |
| Leucine | 131 | MS/MSb,c | |
| Other Species | |||
| Guanidine | 59 | MS1d | MS/MS |
| Urea | 60 | MS1 | MS/MS |
| Ethanolamine | 61 | MS/MS | |
| Glycolic Acid | 76 | MS1 | MS1 |
Notes. All samples that were confirmed with GC/MS (MS1) were also observed in MS/MS runs.
aPreviously confirmed by GC/MS in Hörst et al. (2012). bPreviously identified by Orbitrap only in Hörst et al. (2012). cOnly the leucine isomer was observed. Isoleucine and norleucine were not detected. dPreviously confirmed by GC/MS in He & Smith (2013).Download table as: ASCIITypeset image
The plasma generated aerosol yielded a larger variety of detectable prebiotic molecules. This is expected as a more aerosol sample was used in the analysis. In addition to the five biological nucleobases identified in previous work (Hörst et al. 2012), two nonbiological nucleobases, hypoxanthine and xanthine, were also identified with sufficient signal to be confirmed in MS1 scans. The presence of these nonbiological nucleobases is important as all the compounds in this study should be from nonbiological sources. If the nucleobases detected were from a biological contaminant, hypoxanthine and xanthine would not have been detected.
Of the 14 amino acids previously identified (Hörst et al. 2012), only glycine and alanine had sufficient signals for MS1 analysis in this study. In the previous work, the Orbitrap was unable to distinguish between leucine and isoleucine. In this work, leucine has been positively identified using the GC/MS/MS results. It is important to note that this does not rule out the presence of isoleucine (or any of the other 10 amino acids) in the aerosol, but rather sets a relative upper limit on the amount that may be present in the aerosol. As Orbitrap MS is at least 10x more sensitive than GC/MS/MS, there could be as much as 10x less of any of the previously detected amino acids compared to leucine.
An additional six prebiotic compounds have also been identified in the plasma aerosol using the MS/MS methods. Among them, guanidine was previously identified in CH4/N2 plasma aerosols by NMR and GC/MS (He & Smith 2013). While alanine and urea were detected by Gautier et al. (2014), their presence was attributed to post-irradiation handling, not direct synthesis. While the presence of these additional molecules is not surprising, they do serve as an indicator as to the number of compounds that are still unidentified in the aerosol matrix.
4. Implications
4.1. GC/MS/MS as a Screening Method
The use of tandem MS/MS screening methods presents itself as a powerful tool for better understanding the nature of prebiotic aerosols. The GC/MS/MS results above show some distinct advantages over that of the more sensitive Orbitrap. The sensitivity of an Orbitrap comes from its ability to trap a broad range of masses for extended periods of time, effectively boosting the signal of lower abundance ions to detectable levels, while sacrificing structural information beyond that of exact molecular formula. On samples with relatively few species, this is not a disadvantage, but prebiotic aerosols may contain between 8000 and 15,000 different molecular formulas (Hörst et al. 2012; Gautier et al. 2014), many of which will have more than one structural isomer, as shown by the previous assignment of m/z = 131 to either leucine or isoleucine. The addition of a GC to the front end of an Orbitrap can be used to filter products with different retention times. The effect of adding a GC can be seen in Figure 4, where two different sets of structural isomers (valine/norvaline and leucine/isoleucine/norleucine) can be separated in time for analysis.
A single MS by itself is still limited, regardless of Orbitrap or quadrupole, as all coeluting peaks are filtered at once. This has the effect of raising the background and can make deconvolution of individual species difficult. The use of tandem MS/MS for filtering individual species greatly enhances the data return where coeluting samples are of concern. As shown in Figure 3, by using the first quadrupole to trap a unique identifying peak, followed by fragmentation and analysis, it is possible to extract species of interest from a complicated mixture. This is best evidenced by the recovery of the added spike to the pyridine-based aerosol (Figure 3, blue traces).
A primary disadvantage of GC/MS/MS is that it requires foreknowledge of the sample so targeted method parameters can be programmed into the instrument. Species that are not looked for are suppressed, and thus not seen. GC/MS/MS is not a standalone technique but works best when used in conjunction with traditional GC/MS or an Orbitrap where sample amounts are too small for GC/MS analysis. In addition, as evidenced by the observation of fewer amino acids compared to previous Orbitrap work (Hörst et al. 2012), GC/MS/MS is a less sensitive technique than Orbitrap by at least 10x, and thus will miss the species with the lowest yields.
4.1.1. Aerosol Nature
The results presented in this work show a clear correlation between energy input and complexity of aerosol species. On the timescale of laboratory experiments, photochemical light sources are limited in the amount of energy they can provide to the reaction system. Not only is this seen in the traditionally low amounts of aerosols made (Tran et al. 2003; Coll et al. 2013), but is also evidenced by the simplistic (nonring structure, m/z < 100) nature of the identifiable products from the photochemically generated aerosol. If ringed structures and large polymers are part of the photochemical process, it would take either significantly longer dwell times of the aerosols in the irradiation zone of the lamp or more photons to push the initial photo-products toward the higher-order species. Plasma, with its higher energy maximum input (15 eV particles versus 10.8 eV photons from the UV-lamp), is able to generate larger amounts of material in a laboratory time frame with an apparent larger degree of molecular complexity, both in number and form. The molecular nature of plasma aerosol was consistent with previous studies with MS (Hörst et al. 2012; He & Smith 2013; Gautier et al. 2014; He & Smith 2014) in that aromatic species were identified along with other species with masses >100 daltons.
According to models of Titan's atmosphere (Krasnopolsky 2009, 2014), it is expected that the starting gas molecules of the atmosphere should form simple compounds that then go on to form more complex structures if given time and energy. This process is directly mirrored in the results from the CH4/CO aerosol in this work (Figure 6). All of the simple molecules that could be identified in the photochemical aerosol were also in the plasma aerosol. The plasma aerosol, however, showed an abundance of more complex species, including aromatics, that were not detected in the photochemical aerosol. This could be due either to the inherent differences in photochemical and plasma chemical pathways. While the exact chemical reactions involved are not fully understood, the energy sources (photons versus particles) interact differently with the reactant gases. For a photochemical reaction to proceed, the photon must first be absorbed by a reactant species. Nitrogen does not show any absorbance across the UV spectrum used in the study, requiring wavelengths shorter than 100 nm (Krasnopolsky 2009). Only UV photons corresponding to the absorption features of CO and CH4 can initiate photochemical reactions. This corresponds to 0.2% of the total photochemical gas mixture in this work. Recent work (Trainer et al. 2012; Sebree et al. 2016; Hörst et al. 2018) showed that CHN based compounds could be formed via photochemical processes on CH4/N2 gas mixtures with the exact mechanism of N2 incorporation still an ongoing study. Plasma reactions, however, are less selective, requiring the energetic particle (proton/electron) to have an energy similar to, or higher than the ionization energy of the reactant molecule (Krasnopolsky 2009). With ionization energies near or below 15.5 eV (Linstrom & Mallards 2011), all three reactant gases (CO, CH4, and N2) could potentially be ionized, or at least excited, by the plasma with a maximum particle energy of ∼15 eV. Even if only a few percent of the mixture is initially excited, this is still 10x more than is possible using a photochemical light source. The inclusion of excited/ionized nitrogen, also means the inclusion of an initial reactive species that is not present in far-UV photochemical experiments.
Figure 6. Hierarchical summary of the main products from plasma and photochemical initiation of reactions starting from CO, CH4, and N2 in this work. Aerosol products are consistent with those reported in previous works. (Hörst et al. 2012; He & Smith 2013; Gautier et al. 2014).
Download figure:
Standard image High-resolution image
It is also important to examine the detection limits imposed by the lower formation rates of photochemical versus plasma generated aerosols as a factor for the implied complexity of the plasma generated aerosols. It is likely that the use of larger amounts of photochemical generated samples would result in the detection of more species using the tandem MS methods. However, as current photochemical yields currently are ∼1000x less than plasma generated aerosol yields, such speculation of possible results is beyond the scope of this work but the use of an Orbitrap MS would be a viable option for studying these photochemically generated aerosols further for better comparisons.
5. Conclusions
In conclusion, GC/MS/MS methods open up the possibility to better understand the nature of prebiotic aerosols through targeted species identification. We have shown that the 100x increase in sensitivity over GC/MS can be used to great effect when looking for a set of molecules within an aerosol matrix. In addition, the use of targeted trapping followed by CID can act as a filter, both removing background peaks, and confirming the structural identity of the molecule through its fragmentation pattern. GC/MS/MS also has some limits that should be kept in mind when using on samples of unknown compositions. Without foreknowledge of the sample composition, GC/MS/MS is only sensitive to compounds that are directly scanned for, with the signal from other species being ignored/suppressed. In addition, the technique can be hampered by small sample size, and falls short of the sensitivity of Orbitrap MS techniques. Like all techniques, GC/MS/MS is best used in conjunction with others.
While the aerosol mixtures in this study had an enriched CO/CH4 ratio by 10x to 1000x compared to Pluto or Titan, many important conclusions can still be drawn from the work presented. First, the incorporation of nitrogen into the photochemical generated aerosols is consistent with previous far-UV studies (Trainer et al. 2012; Sebree et al. 2016; Hörst et al. 2018). In addition, we have extended those results to show that CHON molecules, including glycine and urea, can be formed via far-UV irradiation of a CH4/CO/N2 gas mixture. In plasma generated aerosols, the expansion of the previously reported (Hörst et al. 2012) nucleobases from the five biological bases to include the nonbiological analogs xanthine and hypoxanthine expands the inventory of prebiotic compounds that could have formed in the primordial atmosphere.
Finally, the ability to directly compare plasma generated and photochemical generated aerosols using the same analytical technique has allowed for an initial direct comparison of the two aerosol systems and some commonalities and differences between them. While GC/MS/MS may not reach the sensitivity of Orbitrap MS, it will prove a useful tool for our future work under more Titan-like (temperature, pressure, composition) conditions, where yields may be similar or lower.
This material is based on work supported by NASA under award No. NNX16AL88H. E.R.S. was supported by an undergraduate research fellowship administered by the Iowa Space Grant Consortium. C.H. was supported by the Morton K. and Jane Blaustein Foundation. The authors would also like to thank the anonymous reviewer for timely, useful and thorough feedback.