Abstract
Enceladus and Europa are compelling targets for astrobiology investigations due to their potentially habitable subsurface oceans connected to the icy surface by geological processes. Both moons emit ice grains either ejected via micrometeoroid surface impacts or erupted from their interiors through plume activity. These grains can be sampled by spacecraft flybys, and their composition can be analyzed by impact ionization mass spectrometers, such as the SUrface Dust Analyzer (SUDA) onboard Europa Clipper, or similar instruments proposed for future Enceladus missions. These instruments can identify potential biomolecules, such as amino acids, down to nanomolar concentrations, as demonstrated through previous laboratory experiments. However, the identical masses of isomeric compounds could hinder the mass spectrometric identification and assignment of molecular biosignatures. Here, we investigate the general capability of impact ionization mass spectrometry to distinguish between isomeric compounds, validated with a test case of eight amino acid isomers with an identical molecular mass of 131.173 u and formula C6H13NO2, using quantum chemical calculations. We show that the amino acid isomers (including diastereoisomers) can be uniquely identified due to their distinct mass spectral features and fragmentation patterns, explained through intramolecular hydrogen bonding and other structural specificities of the individual isomers. Importantly, α-amino acids can be clearly differentiated from non-α-amino acids owing to several major mass spectral features. We show that SUDA-type instruments have sufficient capabilities to differentiate certain isomers and to identify biosignatures from ocean worlds with high confidence.
Introduction
Icy ocean worlds with liquid water oceans beneath their ice crusts represent potentially habitable environments within the solar system. The moons Enceladus and Europa, which orbit Saturn and Jupiter, respectively, have emerged as prime targets in astrobiology investigations, as they may fulfill the minimum conditions for sustaining life as we know it, that is, a sustained presence of liquid water, energy sources, and available essential elements (Des Marais et al., 2008). Both moons are targets of major upcoming missions, including ESA’s JUICE (Grasset et al., 2013), NASA’s Europa Clipper (Pappalardo et al., 2024), and ESA’s next large class mission (Martins et al., 2024). Enceladus, and potentially Europa, possesses a plume that ejects gas and ice grains sourced from the subsurface ocean into space (Hansen et al., 2006; Roth et al., 2014, 2025). The in situ analysis of emitted material by spacecraft investigations can offer valuable insights into the habitability of these moons and potential detection of biosignatures indicative of extraterrestrial life (Cable et al., 2021; Klenner et al., 2024).
The Cassini–Huygens mission provided strong evidence that Enceladus’ subsurface ocean is in direct contact with a hydrothermally active rocky seafloor (Hsu et al., 2015; Waite et al., 2017) and contains both inorganic (e.g., salts; Postberg et al., 2009, 2011, 2023) and organic components. In particular, Cassini’s Cosmic Dust Analyzer (CDA; Srama et al., 2004), an impact ionization mass spectrometer, detected and analyzed ice grains emitted by Enceladus, both in the moon’s plume and in Saturn’s E-ring. CDA facilitated the in situ compositional analysis of individual ice grains, which enabled sampling of ocean material during high-speed flybys. The CDA instrument, alongside the Ion and Neutral Mass Spectrometer (INMS; Waite et al., 2004), revealed that at least five of the six basic elements for life—CHNOPS—are present on Enceladus, although sulfur has only been tentatively identified (Khawaja et al., 2019, 2025; Postberg et al., 2018a, 2023; Waite et al., 2006, 2009, 2017). Detailed analysis revealed refractory complex macromolecules and low-mass N- and O-bearing volatiles alongside semipolar aromatics (Postberg et al., 2018a; Khawaja et al., 2019, 2025). These low mass volatile organics are key compounds for prebiotic chemistry, as they could potentially act as amino acid precursors (Khawaja et al., 2019, 2025; Barge et al., 2019).
These findings emphasize the significance of impact ionization mass spectrometry (MS) in characterizing extraterrestrial ocean worlds. A new generation of such instruments, based on the heritage of CDA, includes the SUrface Dust Analyzer (SUDA; Kempf et al., 2025) onboard Europa Clipper and the High Ice Flux Instrument (HIFI; Mousis et al., 2022) proposed for a future Enceladus mission. These instruments have enhanced capabilities as compared with CDA, including dual ion modes (while CDA only could detect cations) as well as significantly improved mass resolution, precision, and sensitivity, to enable more precise analyses of chemical compositions and related geochemical processes.
CDA’s remarkable discoveries about the composition of Enceladus’ ocean were underpinned by analog laboratory data, which allowed the comparison and interpretation of CDA mass spectra. To this end, analog experiments that used laser-induced liquid beam ion desorption (LILBID) MS were performed, as this technique well reproduces impact ionization mass spectra of water ice grains across a wide range of impact speeds (Klenner et al., 2019). Importantly, this laboratory technique is used not only to interpret data from past or ongoing space missions but also to predict the spectral fingerprints of a range of material relevant for ocean worlds and evaluate the capabilities of future spaceborne mass spectrometers. LILBID experiments have been performed with a large range of organic materials, including hydrothermally processed compounds, molecular biosignatures, and bacterial cells, alongside inorganic materials including salts and other compounds relevant for icy ocean moons (Dannenmann et al., 2023; Khawaja et al., 2019, 2022, 2023, 2024, 2025; Klenner et al., 2020a, 2020b, 2024; Napoleoni et al., 2023a, 2023b, 2024). Thousands of analog mass spectra recorded in these experiments are stored in a comprehensive database (Klenner et al., 2022), which will be key to the interpretation of data from upcoming missions.
LILBID analog experiments with organic compounds are especially relevant for astrobiology investigations on icy ocean moons. Amino acids carry particular significance as a major class of biomolecules: they are fundamental building blocks present in all terrestrial living systems, and they fulfill diverse metabolic roles (e.g., protein building blocks, metabolic intermediates, nutrients, and even energy sources; Lea and Miflin, 1977). Their unique molecular structures confer distinct properties, which enable them to partake in essential biochemical processes (Madigan et al., 2022). Some amino acids can also be generated abiotically via Strecker and Friedel–Crafts reactions during water–rock interaction (Miller, 1957; Ménez et al., 2018; Singh et al., 2022; Bada, 2023) and are thus considered ubiquitous byproducts of abiotic reactions. Amino acids formed abiotically can have hundreds of thousands of possible configurations (Cleaves et al., 2014), and more than 80 (α-, β-, γ-, and δ-type) have been identified in meteorites and samples from asteroids and comets (Altwegg et al., 2016; Elsila et al., 2016; Naraoka et al., 2023; Parker et al., 2023, 2025; Glavin et al., 2025), whereas terrestrial life typically uses a set of only 20 α-amino acids. Another key difference between abiotic and biotic amino acids is that those formed abiotically consist of a racemic mixture of amino acids, while amino acids typically used in peptide biosynthesis have the same chirality. Thus, the observed chemical distribution (type, abundance, and chirality) of amino acids can be used to identify distinctive patterns characteristic of biotic processes (e.g., Creamer et al., 2017; Davila and McKay, 2014). A noteworthy shift from a thermodynamically driven profile of amino acids (i.e., α-, β-, γ-, and δ-type racemic mixtures) to a more specialized, selective, or tailored amino acid profile (e.g., homochiral α-amino acids) is indicative of biotic processes and thus provides a distinction between abiotic and biotic formation processes (Koga and Naraoka, 2017; Davila and McKay, 2014; Cobb and Pudritz, 2014). While abiotic synthesis of amino acids results in an excess of the simplest amino acids such as glycine (due to thermodynamics; Dorn et al., 2011), biotic processes result in a profile with more complex amino acids—driven by structural requirements rather than thermodynamics. For carbon-based life, the ratio of more complex amino acids to glycine can thus be considered a biosignature characteristic of the biogeochemical processes of a system (e.g., Davila and McKay, 2014). The identification of specific amino acids, including their stereochemistry and their precise quantification, therefore represents a fundamental cornerstone in the search for extraterrestrial life (e.g., Georgiou, 2018; Neveu et al., 2018; Glavin et al., 2020).
Previous LILBID experiments by Klenner et al. (2020a, 2020b) provided insight into the spectral fingerprints of amino acids in both water and salt-rich matrices and showed that individual amino acids can be identified by their (de)protonated molecular peaks and characteristic fragments. The experiments showed that amino acids are typically detectable down to micromolar or nanomolar concentrations, although the sensitivity of LILBID varies between individual amino acids depending on their respective tendency to form ions that can be detected by MS. Amino acids with basic side chains, like lysine and arginine, have the lowest detection limits (1 nmol/L) in water matrices. These studies also showed that for individual amino acids at a given concentration, the amplitudes of molecular peaks vary significantly due to their molecular structure (e.g., position of functional groups, side chains, and chain lengths). Optimal ice grain encounter speeds (4–6 km/s) were suggested to optimally detect positively or negatively charged molecular peaks of amino acids to assist in mission planning of spacecraft to icy moons (Klenner et al., 2020a, 2020b; Jaramillo-Botero et al., 2021).
Though previous LILBID experiments have investigated amino acids with different molecular masses (Klenner et al., 2020a, 2020b), the ability of SUDA-type instruments to distinguish between isomeric amino acids (or any other isomeric species) of the same molecular mass or empirical formula remains unclear. The mass spectra of isomeric compounds typically produce a base peak at their (de)protonated molecular mass—which then allows the identification of the chemical formula of the compound; however, this is not always sufficient to decipher the structure of potential isomer(s). Analyzing a pattern of several peaks including fragment ions provides a significant advantage over investigating molecular peaks only. Due to their unique molecular structures, isomeric compounds can produce characteristic fragment ions (and different abundances thereof), which can be used to determine the isomeric species responsible for the spectrum. Ionization methods that induce extensive but mild fragmentation into large building blocks, such as LILBID and impact ionization at intermediate speeds, are preferred over those that either preserve the analyte’s structure or methods that break the structures up into mostly small fragments, such as electron-induced ionization (Charvat and Abel, 2007; Gross, 2017).
To test the diagnostic capabilities of impact ionization to distinguish organic isomers, we investigated the mass spectra of amino acid isomers by providing a comprehensive analysis of analog mass spectra. Our work showcases the possibility for biosignature detection with instruments onboard ongoing missions, such as SUDA on Europa Clipper (Kempf et al., 2025) or future missions to Enceladus (e.g., Mousis et al., 2022). In our work, we measured the LILBID mass spectra of eight isomeric amino acids that have an identical molecular mass of 131.173 u and formula C6H12NO2 but exhibit distinct differences in their molecular structures. Cation mass spectra were recorded, as this polarity shows better sensitivity to most amino acids than the anionic mode (Klenner et al., 2020a). The recorded mass spectra were investigated for spectral features that enable differentiation of the different isomeric amino acids, with the aid of quantum chemistry calculations. The data is stored in a comprehensive database (Klenner et al., 2022) used to interpret current and future spaceborne mass spectra of ice grains from ocean worlds.
Methods
Isomeric amino acid samples
Amino acid isomers were solubilized in pure Milli-Q water, and these solutions were then measured with LILBID-MS. We investigated 8 (of the 31 possible) isomeric amino acids with the formula C6H13NO2 and an identical molecular mass of 131.173 u (Table 1): 6-aminohexanoic acid (Thermo Scientific, CAS: 60-32-2, purity 99%), 3-amino-2-ethylbutanoic acid (BLD Pharmatec, CAS: 824424-73-9, purity 97%, racemic mixture of four diastereoisomers), 3-amino-2,2-dimethylbutanoic acid (from 3-amino-2,2-dimethylbutanoic acid hydrochloride, Ambeed, CAS: 180181-84-4, purity 95%, racemic mixture of its two diastereoisomers),
The Eight Isomeric Amino Acids (C6H12NO2) Studied in This Work, Their Respective Molecular Structures, and Their Respective Classification Based on the Structural Arrangement of the Amino and Carboxyl Groups
The Eight Isomeric Amino Acids (C6H12NO2) Studied in This Work, Their Respective Molecular Structures, and Their Respective Classification Based on the Structural Arrangement of the Amino and Carboxyl Groups
The amino acid samples were measured with LILBID time-of-flight (TOF) MS (Supplementary Fig. S1), a technique that accurately recreates impact ionization mass spectra obtained by CDA, SUDA, HIFI, or similar instruments (Klenner et al., 2019). The setup combines a vacuum chamber (5 × 10−5 mbar), within which the sample is ionized, and a reflectron-type TOF mass spectrometer, which analyzes ions accelerated toward its detector. The samples were injected (∼0.5 mL) individually into the vacuum chamber via a quartz nozzle (diameter 15 μm) to form a micrometer-sized liquid water beam with a constant flow rate of 0.32 mL/min. In the vacuum chamber, the water beam is intersected by a pulsed infrared laser (frequency: 20 Hz; pulse length: 7 ns; wavelength: 2840 nm; energy: 5.40 J): the water beam absorbs the laser energy and explosively disperses into charged atomic, molecular, and macroscopic fragments (Charvat and Abel, 2007; Wiederschein et al., 2015). Cations are accelerated toward and analyzed in the TOF mass spectrometer, supported by the principle of delayed extraction (Klenner et al., 2019), which allows the selection of ions based upon their initial velocities via a repelling electrode in front of the mass analyzer. The signals detected at the microchannel plate detector are then amplified and digitized, and this data is passed to a LabVIEW-controlled computer. To ensure reproducibility, the experimental setup was calibrated using a 10−6 M NaCl solution at the beginning of each measurement day and at three different delay times and laser intensity settings.
While the LILBID technique can be used to reproduce both cation and anion mode mass spectra of SUDA-type instruments, we focused on the former here because it permits capture of a greater range of features specific to the individual amino acids (Klenner et al., 2020a). Each recorded LILBID spectrum was an average of 300 individual coadded spectra, representing a total number of ions similar to the number of ions reaching the detector of SUDA during a single ice grain impact (Kempf et al., 2025; Klenner et al., 2024).
Sensitivity to different amino acids with impact ionization mass spectrometers varies with the impact speeds of ice grains onto the detectors; in LILBID experiments, these variations can be reproduced by using different combinations of delay times (i.e., delayed extraction) and laser intensities (Klenner et al., 2019). To ensure comparability of our experimental data, the laser energy and delay time were uniformly maintained at 90.4% (of 5.40 J) and 5.3–5.4 µs, respectively, during all measurements. That combination of laser energy and delay time corresponds to an ice grain impact speed onto a spaceborne detector of 5–7 km/s (Klenner et al., 2019), relevant to the Europa Clipper planned flyby velocities of Europa (4–5 km/s) and Enceladus plume flythroughs by Cassini (7 km/s) and future mission concepts (5–9 km/s). The concentrations of the amino acids were set to 1 mM, a concentration where every isomer shows both a clear protonated molecular peak and prominent fragment peaks. These experimental conditions do not necessarily represent the most sensitive instrument setting for each individual isomer, but they are nonetheless a good approximation for the average of all isomers and thus allow a meaningful comparison of the different isomers.
Computational chemistry
We used quantum chemistry computations to obtain the relative energies of isomeric amino acids and their optimized molecular structures, with the goal of better interpreting the spectral appearance of the different isomers measured with LILBID-MS. To this effect, we performed computations using the ORCA 6.0.0 theoretical chemistry package (Bykov et al., 2015; Helmich-Paris et al., 2021; Izsák et al., 2012, 2013; Izsák and Neese, 2011; Neese et al., 2020; Neese, 2003, 2009, 2012, 2022, 2023) on all of the isomeric amino acids investigated with LILBID-MS in this work. We utilized the Curta high-performance computing cluster at the Freie Universität Berlin for these calculations (Bennett et al., 2020).
To establish true global energy minima for each molecular isomer, comprehensive exploration of the conformational space was performed for each case. LILBID measurements are performed with molecules initially in solution, albeit with a postionization vapor/plasma phase depending upon the laser energy density. First, initial zwitterionic geometries were optimized using second-order Resolution of Identity Møller–Plesset perturbation theory (RIMP2; Weigend et al., 1998), with the correlation-consistent polarized valence triple zeta (cc-PVTZ; Dunning, 1989) basis set and cc-PVTZ/C auxiliary basis set appropriate for RIMP2 computations. Harmonic vibrational frequencies were calculated to ensure that these structures represented true minima on the potential energy surface (PES). The zero-point corrected energies, vibrational frequencies, and optimized geometries for each structure are shown in the Supplementary Material.
Further geometric optimizations under simulated aqueous conditions were performed using the conductor-like polarizable continuum model (CPCM; Barone and Cossi, 1998). CPCM employs a cavity surface approach wherein the solute’s electrostatic potential generates surface charges at the solute–solvent dielectric boundary. These polarization charges self-consistently interact with the molecule’s electronic structure and thereby incorporate solvation effects into the molecular Hamiltonian. Many of the initial zwitterionic structures were optimized to the neutral form (i.e., a –COOH and a –NH2) in the gas phase, so in these cases, a proton was manually transferred from the carboxyl to the amine group prior to further calculations with the solvation model. Again, numerical frequencies were calculated to ensure true PES minima. Optimizations performed with the CPCM retained the zwitterionic form for all amino acid isomers.
From these solvated zwitterionic structures, the Global Optimizer Algorithm (GOAT; de Souza, 2025) feature within ORCA was performed to find the true global minimum energy structure, regardless of it being a zwitterion or having the neutral structure, and the ensemble of low-lying conformers on the PES. These runs were performed using the semiempirical extended tight-binding model (GFN2-xTB; Bannwarth et al., 2019), which, importantly, due to computational resource constraints, offers a faster sampling of the PES than density functional theory (DFT) or post-Hartree–Fock methods. GOAT calculations were performed using the extended CPCM (Stahn et al., 2023). Hydrogen atoms were also fixed in place during this run to sample only the zwitterionic conformer space preventing proton transfer during optimization.
The global minima and final ensembles were extracted from the constrained GOAT runs, with the former selected for further analysis at the xTB level of theory. These structures were relaxed (i.e., H atoms were no longer frozen), and stochastic explorations of the potential energy landscape were conducted again using ORCA’s GOAT.
Once these final GOAT runs were complete, final geometric optimizations on the entire conformational ensemble for each isomer, extracted from every prior GOAT and optimization calculation, were performed using DFT with the LibXC variant (Lehtola et al., 2018) of the
We also calculated proton affinities in solvation for all of the amino acid isomers investigated in this work. Protonated structures were again optimized at the
Results
LILBID mass spectra
The recorded LILBID cation mass spectra of isomeric amino acids (Figs. 1–8) exhibit (i) prominent protonated molecular peaks (labeled “M+H+” at m/z 132); (ii) a range of characteristic fragments typically formed by loss of one or several functional group(s) (labeled “M-X” or by their respective formulas, some of which are also observed by Klenner et al., 2020a); and (iii) sodium, potassium, and pure water clusters (of respective formulas (H2O) n Na+; (H2O) n K+; and (H2O) n H3O+) from the liquid matrix and sample impurities. The latter are common spectral artifacts in this system.
LILBID cationic mass spectrum of 6-aminohexanoic acid simulating an impact velocity of 5–7 km/s. The protonated molecular peak of 6-aminohexanoic acid is labeled “M+H+.” Fragment ions are identified in the mass spectrum and represented in the molecular structure as well. Ions from the matrix (pure water clusters and sodium and potassium water clusters) are labeled with blue triangles and yellow and green squares, and ammonium ions with black dots. The peak at m/z 114 is identified as protonated caprolactam (see Section 3.1).
LILBID cationic mass spectrum of 3-amino-2-ethylbutanoic acid simulating an impact velocity of 5–7 km/s. The protonated molecular peak of 3-amino-2-ethylbutanoic acid is labeled “M+H+.” Fragment ions are identified in the mass spectrum and represented in the molecular structure as well. Ions from the matrix (pure water clusters and sodium and potassium water clusters) are labeled with blue triangles and yellow and green squares, and ammonium ions with black dots. The fragment C2H6N+, labeled with green, is formed by a McLafferty rearrangement (see Section 3.2).
LILBID cationic mass spectrum of 3-amino-2,2-dimethylbutanoic acid simulating an impact velocity of 5–7 km/s. The protonated molecular peak of 3-amino-2,2-dimethylbutanoic acid is labeled “M+H+.” Fragment ions are identified in the mass spectrum and represented in the molecular structure as well. Ions from the matrix (pure water clusters and sodium and potassium water clusters) are labeled with blue triangles and yellow and green squares, and ammonium ions with black dots.
LILBID cationic mass spectrum of
LILBID cationic mass spectrum of
LILBID cationic mass spectrum of α-methyl-
LILBID cationic mass spectrum of
LILBID cationic mass spectrum of
Despite the identical amino acid molar concentration in all samples, the intensities of protonated molecular peaks and fragment peaks varied between isomers (Figs. 1–8; Table 2). The mass spectra of three of the isomers (6-aminohexanoic acid, 3-amino-2-ethylbutanoic acid, and 3-amino-2,2-dimethylbutanoic acid; Figs. 1–3) are dominated by the protonated molecular peaks and fragment peaks, with water and salt clusters present with much lower intensities. Conversely, for the remaining five isomers (Figs. 4–8), the peaks from water and salt clusters are predominant over the molecular and fragment peaks from the isomers.
Summary of the Cations Related to the Amino Acid Isomers, with Their Respective Molecular Formulas and Mass-to-Charge Ratio (m/z), Detected in the LILBID Mass Spectra for Each Isomeric Amino Acid in Figures 1–8
The amplitudes of the peaks were normalized to the total integrated ions over the spectra’s range (m/z 0–200). Peak amplitude values higher than 1% are in bold. Empty cells indicate a lack of detection of the corresponding cations. “UI” refers to unidentified cations.
continued
In the cationic LILBID mass spectrum of 6-aminohexanoic acid (Fig. 1), the protonated molecular peak (m/z 132) is the base peak (i.e., most prominent peak). Two water cluster species of the protonated molecular peak are also observed. The fragment peaks [M-COOH-NH3]+ (m/z 69), M-H2O-NH3+H+ (m/z 97), M-H2O+H+ (m/z 114), and [M-NH2]+ (m/z 115) are detected with relatively high intensities. Less prominent fragment peaks include NH4+ (m/z 18), CH2NH2+ (m/z 30), and [M-COOH]+ (m/z 86).
The cation M-H2O+H+ (m/z 114) originates from an intramolecular condensation reaction (i.e., loss of water) of 6-aminohexanoic acid, forming caprolactam (C6H11NO) and water, as shown in Figure 9. Caprolactam is detected in its protonated form, that is, C6H12NO+. Other peaks that could originate from caprolactam are (i) the fragment M-H2O-NH3+H+ (m/z 97), which could be formed by a loss of ammonia from protonated caprolactam, and (ii) the fragment CH2NH2+ (m/z 30) (Mitera and Kubelka, 1971).
Condensation reaction of 6-aminohexanoic acid, leading to the formation of caprolactam and water.
In the cationic LILBID mass spectrum of 3-amino-2-ethylbutanoic acid (Fig. 2), the most prominent peak is a fragment of the parent molecule generated by a loss of the C2H5 sidechain ([M-C2H5]+ at m/z 102). This fragment peak is not observed in the spectra of any other isomer and is thus unique among the measured isomers. A water cluster [M(H2O)-C2H5]+ is observed at m/z 120. Other prominent peaks in this spectrum include the fragments C2H6N+ (m/z 44), [M-COOH-NH3]+ (m/z 69), [M-NH2]+ (m/z 115), and the protonated molecular peak M+H+ (m/z 132). Less prominent fragment peaks include NH4+ (m/z 18), [M-COOH]+ (m/z 86), M-H2O-NH3+H+ (m/z 97), and M-H2O+H+ (m/z 114), as well as peaks at m/z 65 and 83 that may be identified as C5H5+ and C5H7O+, respectively (i.e., [M-NH2-CH2-(H2O)2]+ and [M-NH2-CH2-(H2O)2]+).
While the peak at m/z 114 is assigned to protonated caprolactam in the spectrum of 6-aminohexanoic acid (Figs. 1 and 9), the peak at the same mass in the spectrum of 3-amino-2-ethylbutanoic acid (Fig. 2) is assigned to a protonated β-lactam (i.e., four-membered cyclic amide). This β-lactam is formed by an intramolecular condensation reaction of the amino acid, which leads to the formation of a β-lactam and water (Lam et al., 2008).
The cation C2H6N+ (m/z 44) can be produced by direct dissociation or by the β-cleavage of 3-amino-2-ethylbutanoic acid (McLafferty rearrangement) shown in Figure 10 (where the fragment C2H5N would be protonated and detected as C2H6N+). The lack of a C4H8O2+ (88 u) fragment in the mass spectrum (Fig. 2) argues for a formation of C2H6N+ by direct dissociation (leaving behind a neutral C4H8O2) or for a secondary fragmentation or rearrangement of C4H8O2+. McLafferty rearrangements were previously observed by Khawaja et al. (2019) in both LILBID mass spectra and CDA mass spectra of carbonyl compounds in Enceladus ice grains.
McLafferty rearrangement of 3-amino-2-ethylbutanoic acid: a H+ of the methyl group in the γ position transfers to the carbonyl oxygen atom, while the molecule undergoes β-cleavage. This leads to the production of two fragments, namely, C4H8O2+ (mass 88 u) and C2H5N (mass 43 u) which is later protonated to observed C2H6N+ (m/z 44).
In the cation LILBID mass spectrum of 3-amino-2,2-dimethylbutanoic acid (Fig. 3), the most prominent peak is the fragment peak at m/z 44, which can be interpreted as C2H6N
McLafferty rearrangement of 3-amino-2, 2-dimethylbutanoic acid: a H+ of the γ methyl group transfers to the carbonyl oxygen atom. This leads to the production of two fragments, namely, C4H8O2+ (mass 88 u) and C2H5N (mass 43 u). Note that the C4H8O2+ has a different structure as compared with Figure 9 and is observed at m/z 88 in Figure 3.
Other prominent peaks in that spectrum include the fragments NH4+ at m/z 18, C3H7+ and/or C2H5N
The cation LILBID mass spectrum of
Formation pathway of the C3H7+ and C2H6N+ fragments observed at m/z 43 and 44 in the mass spectrum of
The cationic LILBID mass spectrum of
Mass spectrum of α-methyl-valine
The cationic LILBID mass spectrum of α-methyl-valine (Fig. 6) is also dominated by sodium and pure water clusters (i.e., (H2O)
n
Na+, (H2O)
n
H3O+), and the protonated molecular peak (m/z 132) is the most prominent peak originating from the isomer. The relative intensities between M+H+, [M-COOH]+, and [M-NH2-CH2]+ are very similar to those same species in the spectra of
Mass spectrum of L -allo-isoleucine
The cationic LILBID mass spectrum of
Mass spectrum of l -isoleucine
The cationic LILBID mass spectrum of
Comparison of the isomers’ cationic mass spectra
General spectral features
The mass spectra of 6-aminohexanoic acid, 3-amino-2-ethylbutanoic acid, and 3-amino-2,2-dimethylbutanoic acid (Figs. 1–3) are largely dominated by peaks originating from the respective amino acids (molecular peaks and/or fragments)—and not by salt and water cluster peaks. The highest intensity peak in the mass spectrum of 6-aminohexanoic acid (Fig. 1) is the protonated molecular peak M+H+ (m/z 132), which is an order of magnitude higher in intensity than the pure water and salt cluster peaks. The highest intensity peak in the mass spectrum of 3-amino-2-ethylbutanoic acid (Fig. 2) is the fragment [M-C2H5]+ (m/z 102), which is an order of magnitude higher in intensity than the pure water cluster and salt cluster peaks. In the mass spectrum of 3-amino-2,2-dimethylbutanoic acid (Fig. 4), the most intense peak is the fragment C2H4O+ (m/z 44), which is slightly higher than the pure water cluster peaks (at m/z 37 and 55); and the protonated molecular peak M+H+ (m/z 132) is also highly prominent.
In contrast, the mass spectra of
Distinctive fragmentation and molecular features
Fragment peaks considerably differ in both nature and intensities between the different isomers (Figs. 1–8; Table 2). A summary of the main peaks originating from the isomers, including the protonated molecular peak and fragment peaks, is shown in Table 2, with their respective intensities normalized to the total integrated ions in the spectrum. Both the protonated molecular peak and the fragment peak due to the loss of a carboxyl group ([M-COOH]+ at m/z 86) are observed for all isomers. 6-aminohexanoic acid is the only isomer for which the protonated molecular peak is considerably higher in intensity than all of the other peaks in the spectra (Table 2). Most isomers also produce the fragments NH4+ (m/z 18), [M-COOH-NH2]+ (m/z 69), and [M-CH2-NH2]+ (m/z 101). Some fragments are produced by only one or two isomeric species, such as CH2NH2+ (m/z 30), [M-C2H5N]+ (m/z 88), or [M-C2H5]+ (m/z 102). The peaks at m/z 160 and 190 are uniquely observed for
The two cations produced by the condensation reaction forming lactams, that is, both M-H2O+H+ (m/z 114) and M-H2O-NH3+H+ (m/z 97), are only observed in the mass spectra of 6-aminohexanoic acid, 3-amino-2-ethylbutanoic acid, and 3-amino-2,2-dimethylbutanoic acid, but not for the remaining isomers. The peak at m/z 114 is especially intense in the spectrum of 6-aminohexanoic acid, where it is identified as protonated caprolactam. For the butanoic acid derivates, this peak is identified as a protonated β-lactam and has much lower intensities than the protonated caprolactam.
The fragment cation C2H6N+ (m/z 44), detected in the mass spectra of both 3-amino-2-ethylbutanoic acid and 3-amino-2,2-dimethylbutanoic acid, is formed via McLafferty rearrangement for both species. This cation is also observed in the mass spectrum of
Quantum chemistry computations
The
The global minimum energy structure for each amino acid isomer investigated in this work, optimized at the
Aqueous-Phase Proton Affinities Calculated at the MP2/cc-PVTZ//
Spectral analysis
We recorded LILBID mass spectra of eight structurally distinct isomeric amino acids that share an identical formula (C6H13NO2) and molecular mass of 131.173 u to simulate their impact ionization mass spectra as would be recorded by SUDA-type instruments. All the spectra were recorded under uniform instrumental settings, to replicate a scenario where the amino acids are enclosed within ice grains ejected from an ocean world (e.g., Europa or Enceladus) and subsequently encountered by a SUDA-type detector during a spacecraft flyby at 5–7 km/s (Klenner et al., 2019). For all the isomers, we detected their protonated molecular peak, which allows identification of the molecular mass of the isomers. We establish distinctive spectral features for each isomer based on (i) the relative abundance of peaks originating from the amino acids as compared with matrix peaks (i.e., water and salt clusters), which is linked to prevailing intramolecular hydrogen bonding, and (ii) the nature of observed fragment cations and their relative abundances. While our work hitherto has only investigated spectral features observable in cation mode, the ability of SUDA (and other future dust analyzers) to also record anion mode spectra will undoubtedly provide other means to eliminate coincidental interferences and further improve spectral interpretations.
On intramolecular hydrogen bonding
A predominance of peaks originating from the parent molecule (Figs. 2–4) is observed for isomers where the amine group is located far from the carboxylic acid group (i.e., 6-aminohexanoic acid, 3-amino-2-ethylbutanoic acid, and 3-amino-2,2-dimethylbutanoic acid). On the contrary, α-amino acids, that is, amino acids in which amino and carboxyl groups are bound to the same “α” carbon atom (i.e., α-methyl-
Hydrogen bonding can occur in both inter- and intramolecular interactions and has been shown to have a stabilizing effect in amino acids and proteins (e.g., Vogt et al., 1997; Giubertoni et al., 2020; Shahamirian and Azami, 2021; Rios et al., 2020). Previous LILBID experiments (Khawaja et al., 2023) also highlighted the importance of hydrogen bonding when comparing the spectra of isomeric organic compounds. They studied the LILBID mass spectra and fragmentation pathways of two isomeric derivatives of benzoic acid, namely, 2,3-dihydroxybenzoic acid and 2,5-dihydroxybenzoic acid. Subtle differences in the mass spectra of the isomers were found in both ion modes, particularly in the nature and intensities of fragments but also in the appearance of a dimer-related cation due to intermolecular hydrogen bonding in one of the isomers. However, we note that intermolecular interactions (such as dimer formation) will be dependent on species concentration, and the abundances of amino acids in ocean world-derived ice grains may be lower than the 1 mM concentrations investigated here. Our results, and those of Khawaja et al. (2023), strongly suggest that intramolecular interactions—which depend on the proximity of certain functional groups within a given molecule—influence both the ionization efficiencies and the fragmentation patterns of the molecules.
Among the eight studied isomers, the two species
Distinctive spectral and fragmentation features
Here, we show that the patterns of observed fragment cations and their relative abundances can be used to trace back their origin to parent molecular isomers. A systematic approach based on a decision-tree pathway is proposed (Fig. 14) as an aid for the identification of the eight isomeric amino acids studied in this work. First, the intensity of peaks derived directly from the isomers themselves relative to those arising from the matrix can be used to indicate the classification of the amino acid (i.e., α-amino acids or non-α-amino acids). Following this, the presence of certain characteristic fragment peaks and their relative intensities can be used to discriminate between the different isomers. Such a decision tree is specific for the impact speeds simulated in this work and for the particular molecules we studied, and it may differ not only at alternative impact speeds and different chemical conditions (pH, salinity, etc.) but almost certainly for other collections of isomeric molecules. For samples of more complex compositions, interference with other organics (especially with small molecules like ammonia or methylamine that can be present themselves or form as fragments of several larger species) or salts could also make spectral interpretation much more challenging. Nonetheless, this work marks a significant development relevant for the science objectives of Europa Clipper’s SUDA, ESA’s next L-Class mission to Enceladus, as well as other potential future missions that would conduct hypervelocity flybys to sample ice grains containing organics.
Decision tree for a systematic approach to identify the eight investigated isomeric amino acids in cation mode, depending on their fragmentation pattern in their respective mass spectra. The classification (i.e., α-amino acids or non-α-amino acids) is first determined by investigating the predominance of organic or matrix peaks. Subsequent investigations to discriminate specific isomers are based on the presence of key fragment ions. The specific indicators presented in this decision tree are only valid for the eight isomers measured individually in pure water and do not take into account possible interfering ions that are expected in samples of more complex composition.
Our results, which show that there are qualitative and quantitative differences in fragment peaks between isomeric species, are a substantial enhancement of those found by Khawaja et al. (2023). They also found a correlation between the simulated impact speed and the degree and diversity of differences between the spectra of the two isomers: while they found that there was no difference between the fragmentation pattern exhibited by the isomers at low simulated impact speeds (4–6 km/s), intermediate impact speeds (6.5–8.5 km/s) provided greater differences, especially in the negative mode; and high simulated impact speeds (9–11 km/s) exhibited significant spectral differences (although in the negative mode only). Our experiments focused on intermediate impact speeds (∼5–7 km/s), as this range so far shows the largest differences between isomers. We note that the flyby velocities of the Europa Clipper mission (4–5 km/s) fall at the lower end of this range. Future work could explore if the correlations found by Khawaja et al. (2023) are true for all types of compounds or if they are limited to certain organic families or functional groups. Anion mode mass spectra of a variety of isomeric compounds could also be investigated to study their respective spectral features and determine which ion mode allows best discrimination between specific isomers.
We propose interpretations of several fragment cations that we identified in this work, which can be further extrapolated to other amino acids:
The C2H6N+ cation can be used to discriminate The C2H6N+ cation was also detected with a high intensity by Klenner et al. (2020a) in the LILBID cation mass spectrum of a mixture of nine amino acids and was interpreted as [alanine-COOH]+. We argue that this cation could also be formed via fragmentation and subsequent McLafferty rearrangement of threonine—an amino acid that was among the nine amino acids measured and that has both a methyl and a hydroxyl group in its β carbon atom. Decarboxylation of ε-, γ-, and δ-amino acids forms unsaturated amines. Decarboxylation of β-amino acids can form tautomers (i.e., isomeric species that differ only in the positioning of a labile proton, which moves from one position to another, interconverting in rapid equilibrium; Antonov, 2014), which are a respective enamine and a respective imine. The intensity of the observed [M-COOH]+ fragment peak depends on the stability and ionization potential of these tautomers. Stability can be influenced by the substitution of double bonds (Smith and March, 2007), which might be why the ramified [M-COOH]+ fragments (such as those formed from 3-amino-2,2-dimethylbutanoic acid) have high-intensity [M-COOH]+ peaks, while 6-aminohexanoic acid, which has a less substituted double bond, has a low M-COOH peak (Table 2, Fig. 15). Decarboxylation of the isomeric amino acids studied in this work and the resulting cations observed in the LILBID spectra at m/z 86. Differences in the stability of the different isomers after decarboxylation explain the difference in observed intensities of the peak at m/z 86. The α-amino acids (last five isomers, with names colored in light green) produce an imine (R1R2-C=N-R3) directly after the cleavage of the carboxylic (–COOH) group. The iminium ions (i.e., protonated form of the imine) derived from these imines are the cations observed in the corresponding LILBID spectra at m/z 86. The non-α-amino acids, on the contrary, do not necessarily form imines after cleavage of the carboxylic group. The isomer 6-aminohexanoic acid forms an amine, which is detected in its protonated form. The isomers 3-amino-2-ethylbutanoic acid and 3-amino-2,2-dimethylbutanoic acid can both form compounds that present a certain isomerism called tautomerism: a proton in their structure moves from one position to another, interconverting in rapid equilibrium (brown circles). For each isomer, the tautomers are a respective imine (C–C=N, i.e., the double bond includes the N atom) and a respective enamine (C=C–N, i.e., the double bond is adjacent to and does not include the N atom). The enamine derived from 3-amino-2,2-dimethylbutanoic acid is more stable than the enamine derived from 3-amino-2-ethylbutanoic acid, as its double bond is more substituted. Usually, the most stable form out of the two tautomers is imine, which is thus the more likely tautomer formed, protonated, and observed at m/z 86 in the LILBID mass spectra. Besides the imine and enamine tautomers, the 3-amino-2-ethylbutanoic acid, upon cleavage of the –COOH group, can form molecule b, which can also be detected in its protonated form (cation 2b).
We therefore conclude that all α-amino acids are likely to exhibit high-intensity [M-COOH]+ peaks that correspond to iminium ions, while β-amino acids do not necessarily have high-intensity M-COOH peaks as they do not necessarily form iminium ions. Amino acids with longer distance between the functional groups, including ε-, γ-, and δ-amino acids, form unsaturated amines upon decarboxylation. Our results agree with those of Gogichaeva et al. (2007), who studied the fragmentation of amino acids using MALDI tandem MS and concluded that [M-COOH]+ is a dominant fragmentation pattern for α-amino acids. The intensity of the [M-COOH]+ peak can thus be used as an indicator to discriminate α-amino acids from non-α-amino acids: a low-intensity [M-COOH]+ peak (as compared with the intensities of other fragments and molecular ions) implies that the measured amino acid is not an α-amino acid.
These interpretations for specific fragmentation patterns are applicable to a wide range of other amino acids (or compounds with similar functional groups) that bear structural similarities with those measured here. They are also possibly applicable to other MS methods, such as MALDI or EI MS. However, we did not investigate the other 23 amino acid isomers with the formula C6H13NO2 nor other non-amino acid isobaric species. Additional LILBID experiments to simulate higher and lower impact speeds of ice grains onto impact ionization mass spectrometers and in different chemical conditions (pH, salinity) would allow the investigation of other fragments and molecular adducts that could be used to discriminate other isomeric molecules spanning a range of masses and chemical properties. Investigations into the fragmentation of individual amino acids could also help interpret the fragmentation patterns of peptides and other amino acid derivatives (Armirotti et al., 2007; Gogichaeva et al., 2007). Deconvolution of impact ionization mass spectra from complex mixtures of many amino acids—where interferences may substantially hinder spectral interpretations—is another area of future exploration.
Our results highlight the importance of a better understanding of the reactivity and fragmentation of organic compounds during the processes of both laser desorption and impact ionization. We showed that condensation reactions and McLafferty rearrangements are key processes to understand the formation of fragments specific for each isomeric amino acid. Future work will further investigate the reactivity of organic compounds in laser desorption experiments (Hortal Sánchez et al., 2024). In the following section, we complement our interpretations of the mass spectra and fragmentation patterns with quantum chemical insights into structures, energies, and proton affinities.
Solvated proton affinities
The solvated proton affinity calculations (Table 3) provide insight into the observed ionization efficiencies of the amino acid isomers in the LILBID mass spectra. The non-α-amino acids exhibit significantly higher solvated proton affinities (1214.5–1400.5 kJ/mol) compared with the α-amino acids (1097.5–1164.5 kJ/mol), which correlates with their enhanced molecular ion intensities observed experimentally. Notably, 3-amino-2-ethylbutanoic acid possesses the highest solvated proton affinity (1400.5 kJ/mol), consistent with its prominent molecular ion peak. The α-amino acids, with proton affinities ∼100–300 kJ/mol lower, show correspondingly weak molecular ions; thus, their spectra are dominated by matrix-derived peaks. This substantial difference in protonation efficiency arises from the intramolecular hydrogen bonding in α-amino acids, where the amine group already interacts with the carboxyl group in a cyclic framework, which reduces its availability for protonation. An exception is norleucine, which, despite having the lowest proton affinity (1097.5 kJ/mol), maintains a relatively prominent molecular ion, possibly explained by its stable structure suggested by the shortest predicted hydrogen bond length (1.76 Å) among the α-amino acids.
Proton transfer generally occurs on the order of picoseconds (Tuckerman et al., 1995), but conformational rearrangement upon protonation may require longer timescales that are inaccessible during LILBID-TOF analysis (Patrick et al., 2017). The protonated molecules that reach the mass analyzer probably reflect the structures of the zwitterionic forms of the amino acids from solution; therefore, these are selected as the most suitable starting geometries. In addition, in the present calculations, we find that protonation only slightly alters the geometry of these structures.
Conformational effects and comparisons to spectral intensities
According to our calculations, all α-amino acids investigated in this work exhibit five-membered cyclic ring hydrogen bonding between the amine and carboxyl groups (Fig. 13), which enhances their structural robustness and resistance to fragmentation and possibly leads to the dominance of the matrix-derived peaks observed in the LILBID mass spectra. Notably, however, fragments related to the loss of the C4H9 chain (around m/z 57) that could cleave away from the cyclic part of the molecule are not observed. Similarly, no organic fragment peaks appear near m/z 75, which would derive from the cycle itself. This suggests that this cyclic hydrogen bonding does not have a significant impact on the dissociation channels. The five α-amino acids examined here are calculated to be the lowest energy isomers, followed by the isomers with amine and carboxylic groups in the terminals. The predicted hydrogen bond lengths (1.65–1.82 Å) of the neutrals calculated in this work are relatively short and can thus be considered rather significant interactions that will greatly influence the behavior of these amino acids, especially in solution (Herschlag and Pinney, 2018). Notably, due to the distance between the amine and carboxyl groups, the global minima energy structures for the butanoic acid derivatives with the implicit solvation model are zwitterions. This could explain the losses of NH3 from the protonated molecular ions of each isomer.
The very short predicted hydrogen bond length (1.65 Å) in 6-aminohexanoic acid (Fig. 13) approaches the realm of covalent bond distances. This hydrogen bond between the terminal carboxylic hydrogen and the nitrogen atom at the other terminal would enhance the structural robustness of this isomer, explaining why 6-aminohexanoic acid is energetically lower than the two branched structures. The lack of any side chain or branching means that the molecule experiences relatively weak torsional strain, and the carbon backbone is able to fold significantly to minimize gauche interactions. This robust cyclic structure explains the high intensity of the molecular peak for this species. The strong hydrogen bonding interaction between the NH2 and COOH groups also explains the prominence of the [M-COOH-NH3]+ peak; these groups are tightly bound to one another and are thus likely to be cleaved together. This does not always occur, however, because the hydrogen bond can be instantaneously disrupted by local water molecules, which destabilize it and perhaps permit higher-energy dissociations such as the losses of water and NH2 from 6-aminohexanoic acid. Similarly, the fact that 6-aminohexanoic acid naturally exhibits a coiled structure in aqueous media supports a favorable cyclization reaction into caprolactam as described above.
The mass spectrum of
The quantum chemistry calculations reveal that
Interestingly, the global minimum energy for norleucine (Fig. 13) prefers the zwitterionic structure over the neutral (i.e., nonzwitterionic) form, unlike the other α-amino acids investigated in our work. It also shows the shortest hydrogen bond length of all the α-amino acids and has the lowest energy of all isomers (Fig. 13). This is probably due to its lack of branching, which reduces the steric hindrance and enhances the conformational flexibility, and allows the molecule to adopt the optimal geometry for zwitterionic stabilization. This reduced steric hindrance from side chain substituents means that the carboxyl and amine groups can achieve ideal positioning for strong hydrogen bonding, while the unbranched carbon chain can readily adjust its conformation to minimize torsional strain.
The quantum chemical structures and energies (Fig. 13) were obtained using an implicit solvation model. In LILBID, the desorption of ions from the liquid water beam involves a rapid phase transition into the gas phase. Many zwitterionic amino acids are, generally, thought to be unstable and undergo conversion into neutral molecules during phase transition. The stability of the zwitterionic form of amino acids, if present in an extraterrestrial subsurface ocean, would be strongly coupled to the solution pH and the concentrations of other ions. A basic pH, as is thought to be the case for Enceladus (Glein et al., 2015; Postberg et al., 2023; Glein and Truong, 2025), would indeed support zwitterionic equilibria for most amino acids. Europa’s subsurface ocean is thought to be more acidic than Enceladus’ (Pasek and Greenberg, 2012; Johnson et al., 2019), which suggests that some amino acids could exist as a protonated zwitterion (i.e., with –NH3+ and –COOH groups) in solution, perhaps enhancing the detectability of the protonated molecular ion in mass spectra. If incorporated into ice grains, freezing processes relevant for icy moons (Hamp et al., 2024; Klenner et al., 2025; Reynoso et al., 2025) could facilitate the retention of aqueous-phase nonequilibrium states.
Implicit CPCM(-X) solvation in this work is used to obtain the structure of the amino acids in the bulk aqueous phase and has been validated for use in both neutral and ionic systems (Ho and Ertem, 2016; Takano and Houk, 2005). Differences with proton affinities reported in the literature (e.g., Gronert et al., 2009) can be explained by our optimizations with the CPCM model generating alternative structural minima in the aqueous phase. We also note that the ωB97M-V functional that we utilize in the present work is better suited for modeling noncovalent interactions than the B3LYP (Santra and Martin, 2019) functional that others have used to investigate the proteinogenic amino acids (e.g., Bleiholder et al., 2006). In any case, our proton affinities for leucine and isoleucine lie within 2 kcal/mol of those reported by Gronert et al. (2009), reasonable for DFT calculations due to the challenges of approximating exchange-correlation terms.
However, explicit solvation (i.e., adding physical water molecules to the simulation) may be more relevant for calculating the structures of organic compounds in the solid phase, albeit at a much higher computational cost (Rai et al., 2011). This would introduce a discrete amount of material into the system, where the effects of water molecules on factors such as hydrogen bonding could be studied, and could lead to zwitterionic minima. Similarly, our calculations are performed at the nonphysical temperature value of absolute zero, whereas ice grains in the saturnian and jovian systems have temperatures that can exceed 100 K. Full thermochemical treatment would require extremely computationally intensive molecular dynamics calculations (e.g., Bera et al., 2025). The quantum chemical computations in this work are not intended to simulate precise fragmentation mechanisms but are instead used to explain the appearances and relative intensities of peaks in the mass spectra. Therefore, we only explore the global minima structures for each isomer (and their protonated forms) and not those of the fragment ions that we assign in this work. Similarly, we assume that only the ground state dissociates in this work, and we glean general structural insights from the ground state minima. Analysis of specific fragmentation pathways could enhance our findings by the calculation of transition and excited states and exploration of the PES between the parent molecule and its fragmentation products. Ongoing development of more comprehensive quantum chemistry methodologies for impact ionization and LILBID will assist in better extrapolating our results to icy moon environments (e.g., O’Sullivan et al., 2026).
Implications for ocean moon exploration with MS
Our results have significant implications for Europa Clipper’s SUDA instrument and the next generation of impact ionization mass spectrometers (e.g., HIFI; Mousis et al., 2022), as well as for other applications of laser desorption ionization–MS, given their substantial current progress for the development of in situ space exploration missions (e.g., Arevalo et al., 2023; Ligterink et al., 2020). Although our experiments were performed under controlled laboratory conditions (high concentrations of individual isomers and low salt levels), they open new perspectives on the ability of dust analyzers to distinguish between different isomeric organic compounds. Our results demonstrate that differentiation of isomers with dust analyzers is theoretically possible—at least for cases of ice grains that contain relatively high concentrations of a single isomer. While this scenario cannot be ruled out (especially at the plume of Enceladus, where organic compounds in ice grains appear to be compositionally segregated; Postberg et al., 2018a,b; Khawaja et al., 2019, 2025), more complex ice grain compositions are also possible at ocean moons, including mixtures of salts, oxidants, and organic material. In such cases, the deconvolution of an unknown mixture of several isomers could be difficult, especially if salts are present in the ice matrix, inducing suppression effects (Napoleoni et al., 2023a,b). Such mixtures are definitely expected for instruments that need larger sample sizes and cannot analyze individual micrometer-sized ice grains. There, a separation step (e.g., capillary electrophoresis, liquid or gas chromatography) of constituents prior to their mass spectrometric analysis is typically suggested in spaceborne techniques that aim at thorough organic analysis (e.g., Mora et al., 2022; Blase et al., 2024). However, for dust analyzers specifically, the formation of complexes with metal ions may also allow the differentiation of isomers, including enantiomers (e.g., Ito et al., 2005; Karthikraj et al., 2014), without the strict need for a hyphenated technique for compound separation. Such possibilities should be further investigated through dedicated laboratory experiments that explore chemical compositions likely to be encountered by missions to ocean moons.
Our investigation was performed in optimal laboratory conditions: relatively high concentrations (mmol) of amino acid isomers, low levels of salts, and no other organics were used. In reality, measurements by impact ionization instruments would be more challenging, as they would require relatively stable impact speeds and “ideal” ice grain compositions to achieve a detection of isomers as robust as in our study. Indeed, interferences with other compounds (both organic and inorganic) present in the ice grain matrix and salt-related suppression effects (Napoleoni et al., 2023a, 2023b) could complicate the detection of isomeric compounds. A mixture of several organic compounds (instead of individual compounds) could also complicate the spectral interpretation by possible peak interferences and challenges associated with reconciling certain fragment peaks with a parent molecule. Still, due to the many spectral differences between α- and non-α-amino acids, our work suggests it may be feasible to distinguish them even in the nonoptimal conditions that may be prevalent at Europa or Enceladus. The distinction between different stereoisomers may be more challenging than the distinction between isomers that have more substantial structural differences. Future experiments could further investigate the effect(s) of, for example, salts, pH effects, and variations in concentrations on the detection of isomers, and determine detection limits for isomeric compound distinctions—which must be higher than the detection limit of the molecular mass or the protonated peaks. Here, we measured the detection limits of 6-aminohexanoic acid to be of 1 µmol/L, which suggests that the differences in peak amplitudes described in this work should be observable even if the isomers have concentrations that are lower by one order of magnitude or more. The ability of an impact ionization instrument to analyze individual ice grains—which, at Enceladus, have segregated compositions (Postberg et al., 2018b)—suggests that interferences may not always be an issue. Besides, the use of Orbitrap MS in the next generation of spaceborne instruments (e.g., Sanderink et al., 2023; Selliez et al., 2020) would bring new possibilities of analysis and precise identification of fragment ions, including for more complex molecules (e.g., isomeric peptides or proteins; Xiao et al., 2016; Jiang et al., 2020) or complex samples.
The ability of spaceborne mass spectrometers to discriminate between isomeric compounds in ice grains ejected from icy ocean moons is significant in the context of astrobiology research. Knowing the exact molecular structures of the amino acids (or other key organic compounds) that could be detected on icy ocean moons can provide key information for interpreting—or ruling out—certain compounds and/or chemical distribution as biosignatures, particularly because the differentiation between a biotic and an abiotic origin in extraterrestrial environments can be difficult (Barge et al., 2022; Neveu et al., 2018). On Enceladus specifically, where a rich organic inventory is known to be present (Postberg et al., 2018a; Khawaja et al., 2019, 2025) and prebiotic chemistry may be ongoing (Kahana et al., 2019), it is of prime importance to fully understand the abiotic chemical possibilities before concluding on the presence (or absence) of biological systems.
Our results show that SUDA-type mass spectrometers could enable the identification of specific amino acid isomers and could thus provide information about the structures and functions of amino acids and possibly more complex peptides. The ability to identify structural and catalytic functionalities of extraterrestrial organic molecules is key to assessing the habitability of ocean worlds. Functional properties of amino acids are determined by their sidechain structure, the types of terminal groups, and their positioning relative to the side chain. The search for specific groups in the side chains of amino acids (or other organic molecules) that are characterized by their functional propensity for biochemical catalysis (e.g., hydrolysis reactions) is targeted in life-detection objectives (Georgiou, 2018). The differentiation of biogenic against abiogenic amino acids could also be achieved through consideration of the catalytic and/or structural features presented by certain amino acids. It is therefore crucial to accurately determine the exact structure—and thus the functional properties—of such amino acids. Protein adaptations in extremophilic organisms, including thermophilic, psychrophilic, and halophilic adaptations, can also be considered with patterns in the polarity, charge, and hydrophobicity of the amino acid specific to these adaptations (Georgiou, 2018).
A significant finding of our work is that α-amino acids can robustly be differentiated from non-α-amino acids with impact ionization MS. The relative abundance of α-amino acids to non-α-amino acids is often proposed as a possible approach to search for biogenic patterns in extraterrestrial samples (e.g., Burton et al., 2011; Davila and McKay, 2014). Terrestrial life only uses α-amino acids to build proteins by joining individual amino acids into polymers via covalent bonds connecting the carboxyl group of one amino acid to the amine group of another. A specific sequence of amino acids results in a specific suite of physicochemical characteristics of the protein, including protein structure and function. The resulting activity of the protein, including its interaction with other molecules, is thus determined by the succession of variable side chains of the amino acids that compose it. Proteins use only α-amino acids due to their structural advantages: they can form horizontal planes of α-imino (N-H)-α-carbonyl (C=O) peptide bond polymers that can easily rotate and generate hydrogen bonds for folding. While β- and γ-amino acids are not totally absent in terrestrial life, they are only used in small amounts, for functions other than protein synthesis. Abiogenic chemistry, however, produces a heterogenous mixture of α-, β-, and γ-amino acids. It has been proposed that α-amino acids may be a universal feature of protein peptide bonds and thus present in both terrestrial and extraterrestrial life (Georgiou, 2018), although some other work suggests that β-amino acid-based life could in theory be possible as well (Seebach et al., 1996). The distinction between α-amino acids and non-α-amino acids would also be useful in constraining possible terrestrial contamination.
Conclusions
We performed a laboratory and computational case study to investigate the ability of impact ionization MS to detect and differentiate amino acid isomers. We demonstrated the successful differentiation of amino acid isomers with LILBID-MS via analysis of the cation mass spectra of eight isomeric amino acids that share an identical molecular mass of 131.173 u and identical formula C6H12NO2 but exhibit distinct molecular structures. The recorded LILBID mass spectra simulate cation impact ionization mass spectra of SUDA-type instruments; thus, they reproduce the spectral features of isomeric amino acids as if enclosed in water-rich ice grains ejected from icy ocean moons and detected by SUDA-type mass spectrometers during a spacecraft flyby at 5–7 km/s. While we only investigate here a simplified scenario (high concentration of individual amino acid isomers in pure water), our results suggest that impact ionization MS has high potential to be an effective tool for the discrimination of organic isomers at icy moons. The unique advantage of single-grain analysis by dust analyzers may be particularly valuable at moons such as Enceladus, where organic species appear to be compositionally segregated from both salts and other organic compounds (Postberg et al., 2018a; Khawaja et al., 2019, 2025). Future studies will further investigate its analytical capabilities on more challenging conditions (e.g., samples of complex compositions, salty matrix, varying impact velocities), as well as the possibilities of deconvolution of a fully unknown mixture of isomers with dust analyzers.
We showed that the isomers can be discriminated owing to the different nature and intensities of molecular peaks and fragments and that several spectral features are linked to structural specificities of the individual isomers. Our results are applicable not only to the studied amino acids but also to other isomeric amino acids, peptides, and a wide range of organic compounds that present structural similarities and/or identical moieties. Importantly, α-amino acids can be differentiated from non-α-amino acids owing to several major spectral features, including (i) the relative abundance of peaks originating from the amino acids as compared with matrix peaks (due to the close proximity of the functional groups in α-amino acids), (ii) the relative intensities of the [M-COOH]+ and [M-COOH-NH3]+ fragment peaks, and (iii) the formation of lactams by condensation reactions.
We complemented our mass spectral results with quantum chemical calculations and found that many observed fragments and their intensities can be explained through intramolecular hydrogen bonding and other structural effects originating from the parent molecules. Notably, non-α-amino acids have higher proton affinities than α-amino acids, which result in higher ionization efficiencies for non-α-amino acids than for α-amino acids. Although stereoisomers are more challenging to differentiate than nonstereoisomers, we were able to distinguish between two diastereoisomers owing to (i) a higher intensity of the protonated molecular peak for
While Klenner et al. (2020a) showed that impact velocities of 4–6 km/s provide the highest intensity of molecular peaks for amino acids, our work as well as that of Khawaja et al. (2023) shows that slightly higher impact velocities of 5–9 km/s deliver better diagnostic capabilities of isomeric detection due to unique fragmentation patterns, while still preserving the molecular peaks. In the future, variations of the experimentally simulated impact velocities could also reveal significant differences in the isomers’ mass spectra and/or improved sensitivity for given isomers. Specifically, at lower simulated impact speeds, one could observe the formation of adducts of the amino acid with some of its fragment(s) (Napoleoni et al., 2023a), which, as demonstrated here, can strongly differ between isomers and thus further ease identification of the amino acids. In addition, anion mode mass spectra could also provide additional information that could be key in the discrimination between different isomers (we note that SUDA is capable of operating in both cation and anion mode). Further laboratory investigations with a wider range of isomeric compounds studied with LILBID in both ion modes and with a wider range of experimental parameters would add valuable datasets to inform the exploration of icy moons.
Authors’ Contributions
J.B.: Investigation, formal analysis, writing—original draft, and visualization. M.N.: Formal analysis, writing—original draft, writing—review and editing, and visualization. F.K.: Investigation, supervision, and writing—review and editing (ORCID: 0000-0002-5744-1718). T.R.O.’S.: Formal analysis, visualization, and writing—review and editing (ORCID: 0000-0003-2769-5140). L.H.S.: Formal analysis and writing—review and editing. N.K.: Writing—review and editing. P.P.B.: Writing—review and editing (ORCID: 0000-0003-0843-3209). M.J.M.: Conceptualization and writing—review and editing. M.L.C.: Conceptualization and writing—review and editing. F.P.: Conceptualization, resources, supervision, and writing—review and editing.
Footnotes
Author Disclosure Statement
No competing financial interests exist.
Funding Information
M.N., F.K., and F.P. were supported by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (ERC Consolidator Grant 724908-Habitat OASIS). N.K. and L.H.S. were supported by the European Research Council under the European Union’s Horizon 2020 research and innovation program (
Abbreviations
Associate Editor: Victor Parro
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
