Relationship between ices and gas phase organic compounds in simulated extraterrestrial environments

Abstract

Ices are significant sources of organic matter in various extraterrestrial environments, including the interstellar medium, protoplanetary disks, comets, or icy satellites. This study tested the impact of ice compositions on the formation of volatile organic compounds (VOCs) in the gas phase in simulating the warming up of photoprocessed icy grains at the edge of a protoplanetary disk. 39 VOCs (alkanes, alkenes, esters, ketones, ethers, alcohols, aldehydes, amines, nitriles, amides) were quantified during the warming of different H₂O:CH₃OH:NH₃ ices using gas chromatography coupled to high-resolution mass linked to a vacuum chamber. Different ice compositions showed direct relationships to VOC formation processes and unveiled the active role of NH₃ reactivity. These results indicate that quantifying VOCs in the gas phase could provide insights into composition and surface reactivity of icy extraterrestrial objects. This experimental work further strengths the role that ices play in the enrichment of gas phase of various extraterrestrial environments.

Introduction

Organic matter is found in various astrophysical environments, and a wide diversity is observed, depending on specific conditions and the analytical instruments used for the observation. Gas phase observations have revealed a high molecular diversity in different environments using submillimeter telescopes, such as ALMA^1,2. To probe the surfaces of objects, infrared telescopes are often employed, such as ISO or the most recent JWST^3,4. They provide information on chemical functions, but are still limited in obtaining full information on the large range of potential compounds present in such environments.

Space missions have developed various analytical systems to perform in situ analyses, offering complementary information to telescope observations. For instance, the Rosetta space mission has provided significant insights to understand the gaseous environment of the 67 P/Churyumov-Gerasimenko comet. Mass spectrometry⁵ analyses provided evidence for various organic compounds, including the simplest amino acid, glycine⁶. Using the VIRTIS infrared spectrometer, the first details about the surface of this comet were obtained, suggesting the presence of ammonium salts⁷.

However, information about the molecular diversity remains sparse, even with in situ analyses. Consequently, other space missions have developed an alternative approach aiming at returning samples from the targeted objects. The Stardust mission, for example, returned samples from the coma of the 81 P/Wild comet^8,9,10. Recent Hayabusa 2 mission returned samples from a carbonaceous asteroid, Ryugu¹¹. The most powerful ground-based analytical systems were used to analyze the organic content of these samples, revealing the high molecular diversity present on the surface of such interplanetary objects^12,13. Interestingly, these analyses confirmed previous findings of organics in meteorites without the risk of contamination¹⁴.

To deepen our understanding of the chemistry occurring in such astrophysical objects, laboratory experiments are crucial. Constrained by observations, these experiments simulate specific extraterrestrial environments. Among them, experiments on icy environments have demonstrated the potential to generate significant molecular diversity when exposed to UV photons, electrons, or ions^15,16,17,18. For instance, ices composed of H₂O, CO, CO₂, CH₃OH, and NH₃ formed at 20 K, considered as interstellar ice analogs in dense molecular cloud ices, once altered, exhibit a chemical network containing thousands of different compounds¹⁹. At room temperature, in the remaining refractory residue, tens of thousands of distinct organic molecules have been observed in the resulting solid phases, with masses up to 4000 Da²⁰.

The chemistry occurring is mainly driven at low temperature by radical chemistry²¹, while an additional thermal reactivity occurs during the warming processes²². Studies on the variation of ice composition have demonstrated its effect on the final solid organic samples²³. Competition between carbon and nitrogen sources has been observed, leading to various chemical families observed in the resulting organic solid phases. Furthermore, H₂O exhibits a significant effect on reactivity by maintaining molecules at temperatures higher than their desorption temperatures, thereby facilitating their reactivity during the ice warming²⁴.

In addition to analyzing the solid organic phases, other experiments have been developed to characterize volatile organic compounds (VOCs) potentially released in the gas phase during the warming of photoprocessed ice. Initial experiments on pure methanol ices irradiated with UV photons at Lyman α have demonstrated the release of a wide range of organic molecules^25,26,27. Interestingly, the abundance of molecules in the gaseous phase aligns with the radical chemistry proposed to occur during molecular formation²⁷. Furthermore, the large range of compounds observed in the gas phase depends on photon dose and temperature²⁸.

Regarding the gas phase, the influence of the ice composition on the nature and the abundance of organic molecules released²³ has been investigated. Focusing only on molecules with oxygen, carbon, and hydrogen (CHO molecules), studies have shown how doping methanol ice with H₂O and NH₃ impacts the abundance of six volatile organic compounds (VOCs): one ether, four aldehydes, and one alcohol initially produced from pure methanol ice. The effects of adding H₂O or NH₃ significantly vary, depending on the chemical family. For example, adding water to methanol ice increases the abundance of aldehydes detected in the gas phase, while it inhibits ether formation²³.

In this study, we explore how the ice composition of dense molecular clouds can affect the type and the abundance of VOCs released into the gas phase after warming up photoprocessed ices. Specifically, we investigated ternary ices consisting of varying molecular ratios of H₂O, CH₃OH, and NH₃. VOCs were analyzed using Gas Chromatography coupled with High-Resolution Fourier Transform Mass Spectrometry (GC-FT-Orbitrap-MS). This analytical system allowed to identify 24 VOCs, including alkanes, alcohols, aldehydes, alkenes, amines, ethers, esters, ketones, and nitriles. In addition, due to the high-resolution capabilities of the mass spectrometry, we detected 15 other compounds based on their stoichiometric formulas without direct identification. We obtained insights into how ice compositions influence 39 different VOCs detected in the gas phase. This provides new insights into the role of astrophysical ices in the enrichment of the gas phase with VOCs.

Results

Volatile organic compounds analyzed with GC-FT-Orbitrap-MS

Figure 1 depicts chromatograms of volatile organic compounds (VOCs) detected in the sublimated gas-phase mixture after warming the photoprocessed ice samples to 300 K. We studied four different compositions of ice containing H₂O:CH₃OH:NH₃ (Table 1): a benchmark (3:1:2), H₂O-rich (8:1:2), NH₃-rich (2:1:5), and CH₃OH-rich (2:5:3). As shown in Fig. 1, the overall intensities of detected VOCs are influenced by the initial ice composition. The highest intensities are observed with an excess of CH₃OH, while the lowest intensities are observed with an excess of NH₃. Therefore, the ice composition directly affects the branching ratios within the ice during both its photo-processing and subsequent warming phases.

Fig. 1: Chromatograms observed from GC-FT-Orbitrap-MS analyses.

Chromatograms corresponding to H₂O:CH₃OH:NH₃ ices in the proportions 3:1:2 (A); 8:1:2 (B); 2:1:5 (C), and 2:5:3 (D). The number refers to VOCs identified with a standard (Table S1). Tentative assignments are not notes in the chromatograms.

Table 1 Composition of ice samples and the corresponding apparent irradiation rate relative to CH₃OH used in this work

Using retention times and fragmentation patterns of each peak, tentative identifications of volatile organic compounds (VOCs) were performed by comparing these data to our database containing 150 VOCs. Among the signals observed, 24 compounds were unequivocally identified, including ethers, aldehydes, ketones, hydrocarbons, esters, amines, alkenes, alkanes, and nitriles (Fig. 2 and Table S1). Additionally, many signals remained unidentified. Since a high-resolution mass spectrometer (resolution 60,000 at m/z 200) was operated, stoichiometric formulas were assigned to each ion in a fragmentation pattern, providing valuable information about their chemical composition (CH, CHO, CHN, CHNO) and degree of unsaturation (double bond equivalents, DBE). This approach yielded additional insights into the chemical groups of 15 other VOCs (Fig. 2 and Table S1). Overall, our study encompassed 39 different VOCs, spanning all chemical groups and families analyzable with our GC-FT-Orbitrap-MS protocol.

Fig. 2: Volatile organic compounds detected during the different experiments with GC-FT-Obitrap analyses.

In blue are the molecules identified with a standard, while in dark are the molecules tentatively identified. Additional information is displayed in Figs. S2–S16.

Effect of ice composition on abundances of volatile organic compounds

Evolution of identified compounds

To assess the impact of the initial ice compositions on abundances of the VOCs produced (Fig. 1), we compared our benchmark experiment (3:1:2) with other compositions (H₂O-rich 8:1:2, NH₃-rich 2:1:5, and CH₃COH-rich 2:5:3) for all identified compounds categorized by their chemical functions (as shown in Fig. 2). In the text, only absolute abundances are discussed, since same evolution are observed if relative abundances to the quantity of deposited CH₃OH are used (Fig. S19). Figure 3 presents one representative compound per chemical family, while Figs. S1 and S15 display trends for all detected compounds.

Fig. 3: Evolution of absolute intensities as a function of the ice composition for selected VOCs representative of their chemical family evolution.

Ethanol #12: alcohols; dimethylether #2: ethers; acetaldehyde #5: aldehydes; methylacetate #21: esters; ethylamine #14: amines; acetonitrile #20: nitriles; benzene #31: aromatics; pentane #10: alkanes. Error bars were obtained from the average area measured on three repeated experiments.

For alkanes (pentane #9) or alkenes (1-pentene #10), similar trends are observed (Figs. 3, S1 and S14). Higher amounts of CH₃OH, H₂O, or NH₃ in the ice increase the abundance of hydrocarbons (Figs. 3, S1 and S14), with a stronger effect observed with CH₃OH. However, for aromatic compounds (Fig. 3, benzene #31, and S1), trends differ. The abundance of these compounds decreases in NH₃-rich ice, while CH₃OH or H₂O-rich ices show a non-significant effect compared to the benchmark.

For alcohols (ethanol #12) or ethers (dimethylether #2), the most significant changes are observed in CH₃OH-rich ices (Figs. 3, S1 and S14), showing a strong increase in compound abundance. NH₃ and H₂O do not exhibit significant or observable effects in this context (Fig. 3). However, there is a notable difference with diethyl ether #13 (Fig. S1), which is weakly affected by the amount of H₂O, and moderately affected by the amount of CH₃OH. Regarding aldehyde (acetaldehyde #5) or ketones (acetone #15) compounds when increasing the amount of CH₃OH in the ice, their abundances slightly or moderately increase, except for 2-butanone #25, which shows a negative effect. In contrast, H₂O appears to have a significant positive influence on aldehydes compared to ketones (Figs. 3 and S1). The presence of NH₃ has a negative effect, resulting in a significant depletion for all compounds in this category. Notably, 2-butanone #25 was not observed in H₂O-rich ice conditions.

For saturated compounds bearing two oxygen chemical functions, two ethers (dimethoxy methane #16) or ether -alcohol (methoxymethanol #28), the same trend as described previously in the corresponding mono-functional groups is observed. For esters (methyl acetate #21), the situation is different. They all present the same evolution with an increase of their abundance in CH₃OH-rich ice, and depletions in the presence of higher amounts of H₂O and NH₃. One exception is observed with ethyl acetate #26, where CH₃OH inhibits its formation, H₂O has a weak positive effect and NH₃ still presents a negative influence.

Two amines were detected in the gas phase: ethylamine #14 and propylamine #34. Ethylamine is present in all samples. The overabundance of NH₃ in the ice has a strong positive effect on amine abundances compared to the benchmark ice. With H₂O and CH₃OH, the influence is still positive, but weaker. Propylamine is only visible in the ice enriched in NH₃, confirming the large increase observed with ethylamine. Two nitriles were identified: acetonitrile #20 and propionitrile #27. In this case, NH₃, unlike amines, inhibits their formation. CH₃OH and H₂O have only a weak influence.

Unidentified compounds

Among the 39 compounds detected in our experiments, 15 were not identified using a standard. Their mass spectra are shown in Fig. S16. Tentative identifications were investigated by taking advantage of the high resolution of the Orbitrap instrument, which provided stoichiometric formulas for each ion detected in the mass spectra, allowing us to assign them to a family group (CH, CHO, CHN or CHNO). Then, mass spectra were compared to a standard list and the NIST (National Institute of Standards and Technology) database to propose some identifications based on specific fragmentation patterns, providing structural information (for more details, see Supplementary Information). Corresponding assumptions are given in italics in Table S1. For 11 of these compounds, a specific compound is suggested, while for three of them, only a chemical group attribution was proposed.

Particularly, two compounds were attributed to CHNO compounds. Only monounsaturated CHNO compounds (Table S1) were detected, which does not correspond to the detection of alcoholamines that have already been identified in such ices^24,29. Since no standards were available for the identification, it was difficult to determine the structure. However, if their fragmentation spectra are compared to those of amide compounds from the NIST database (Fig. S15), correlations are observed. Consequently, the CHNO compounds are probably acetamide (#4) and propylamide (#19). Their evolution shows similarities with aldehydes, ketones, and nitriles (Fig. S15). Compared to the benchmark ice, NH₃ has a strong negative effect, while CH₃OH and H₂O have positive influences (Fig. S15).

For unidentified hydrocarbons, investigations using their fragmentation spectra allowed for putative attributions to propane (#1), 1-butene (#3), butane (#9), and heptane (#33). Compared to identified compounds, they exhibit rather similar evolutions, except for propane, which shows an evolution similar to the alkene family.

Additional compounds carrying two oxygen were also tentatively identified as dimethoxymethane (#25), methoxymethanol (#28), and ethoxyethanol (#30), and probably an alcohol-ether (#38). Similar trends to CHO or compounds carrying two oxygen with DBE = 0 were observed, with a significant increase in their abundance in CH₃OH-rich ice and a depletion in NH₃-rich ice. Finally, a compound C₆H₆O₂ at 25.79 min (#37) with a DBE = 4 displays an evolution similar to CHO compounds with DBE = 0. Abundant evolutions of C₆H₆O₂ differ from aromatic compounds, such as toluene or benzene, which show a depletion in its abundance in CH₃OH-rich ice. At this stage, it is difficult to determine a precise structure except that the compound likely bears a hydroxyl moiety.

Discussion

In this work, ices composed of CH₃OH:NH₃:H₂O are altered through VUV irradiation and a subsequent warming. These three molecules are among the main constituents ices detected in dense molecular clouds by infrared spectroscopy, as well as on interstellar grains³⁰. We thus simulate the physical and chemical evolution that interstellar or circumstellar ices could undergo during the formation and development of a planetary system. This includes temperature conditions ranging from 10 to 100 K and photon fluxes varying from ~0 to 10⁸ photons cm⁻² s⁻¹, depending on the grain’s location within the nebula³¹.

In our experimental setup, the ice temperature is maintained at 78 K—representative of a grain located at the outer edge of a protoplanetary disk—and is exposed to a UV flux of 2 × 10¹⁴ photons cm⁻² s⁻¹ (see ref. ³¹). After 44 h of simultaneous deposition and irradiation, the total UV dose reaches 3 × 10¹⁹ photons cm⁻², corresponding to the cumulative exposure of a grain would receive over 10⁶ years in a disk environment at a flux of 10⁶ photons cm⁻² s⁻¹ (see ref. ³¹). The refractory residue obtained at the end of the experiment (under laboratory conditions of 300 K and 10⁻⁸ mbar) can thus be considered as a laboratory analog of the refractory organic materials that may exist on interstellar grains and could eventually be incorporated into comets or asteroids. Thousands of organic compounds are indeed observed in these refractory organic materials remaining on the sample holder^15,16. These results highlight the importance of ice reactivity in the formation of organic matter in various astrophysical environments. Regarding the gas phase, several experiments have demonstrated that ice processing might also release volatile organic compounds into the gas phase^{17,25,27,32,33}. In general, two main experimental methods are available to study ice reactivity under laboratory conditions. The first approach involves forming an ice layer that is subsequently irradiated. In contrast, the second approach consists of irradiating the sample during the ice formation process. These two methods simulate different astrophysical environments. In the second case, interactions between photons and gas-phase molecules are likely, which can mimic conditions near snowlines. This may lead to the formation of radicals both in the gas phase and within the growing ice. However, the synthesis of organic compounds generally requires higher molecular densities, which are more readily achieved in the solid phase than in the gas phase. Importantly, our current results are fully consistent with those obtained using the post-deposition irradiation method²⁸, reinforcing the conclusion that the observed COVs arise from chemical processes occurring within the ice rather than in the gas phase.

Most of the experiments on VOC generated from ices focused on the irradiation of pure methanol ices. Öberg et al. used temperature-programmed desorption coupled with a mass spectrometer to obtain information on VOCs released into the gas phase from a VUV-photoprocessed ice of methanol²⁵. Coupled with infrared spectroscopy measurements of the photoprocessed ice, a preliminary reaction network was proposed to explain the formation of the observed compounds.

The reactivity of methanol is mainly based on radical formation, which recombines by diffusion in the ice to form the observed compounds²¹. The radical reactivity was later confirmed by the direct detection of radicals CH₃, CH₂OH, and HCO³⁴. The effect of the VUV photon doses was also investigated using GC-MS. Quantification of VOCs in the gas phase provided additional information on the reactivity occurring during the ice processing²⁸. In particular, it was demonstrated that ketones and esters form rapidly from CO, while aldehydes form from HCO, except for acetaldehyde, which forms from the photo-processing of ethanol^28,34. Alcohols are species that form more slowly. Other experimental setups were designed to directly analyze VOCs in the gas phase using submillimeter spectroscopy²⁶.

If the ice composition is more diverse by adding water and/or ammonia, the number of chemical pathways and the molecular diversity increase in the solid phase^23,24,35. The ratio of NH₃ to CH₃OH is particularly crucial, with nitrogen (N) and carbon (C) competing for the formation of the molecular skeleton^23,36. Regarding the gas phase, initial analyses of the VOCs with GC-MS have shown that ice reactivity can be directly observed²³. However, this initial experiment focused only on CHO compounds due to a lack of sensitivity. In this study, using a more sensitive analytical system, we were able to obtain data on other chemical groups: CH, CHO (with one or two O), CHN, and CHNO. These results provide new insights into the influence of ice composition on gas-phase VOCs sublimated from energetically processed ices, and indirectly on the low-temperature reactivity occurring in the ice once formed.

Reactivity leading to alkanes and alkenes

Alkanes and alkenes are mainly formed by the recombination of CH₃ radicals generated from the photodissociation of CH₃OH³⁴. Even if ethane is not detectable with our analytical protocol, it can be assumed to form initially:

$${}^{\cdot }\mathrm{CH}_{3}+{}^{\cdot }\mathrm{CH}_{3}\to {{\rm{C}}}_{2}{{\rm{H}}}_{6}$$

(1)

From ethane, various alkanes, alkenes, and radicals can be formed, which explains the diversity of alkanes and alkenes observed^37,38,39. It has to be noted that branched hydrocarbons were not yet identified. This could be due to analytical issues. Linear hydrocarbons are near their detection limit. Although branched hydrocarbons are thermodynamically more stable, branching ratios leading to these compounds could be different from linear hydrocarbons and led to different abundances. Their abundances may be lower than their linear part, thus being outside our detection limits. Additional investigations have to be performed to elucidate this aspect.

Reactivity leading to alcohols and ethers

As for alkanes, CH₃OH is the main precursor of the detected alcohols and ethers, as shown by its high positive effect on their abundances (Figs. 3, S1 and S14). As already discussed by Oberg et al.²⁵, they are formed by direct recombination of methanol radicals. For example, dimethylether is formed according to Eq. (2) and ethanol according to Eq. (3).

$${}^{\cdot }\mathrm{CH}_{3}+{}^{\cdot }\mathrm{OCH}_{3}\to {\mathrm{CH}}_{3}\mathrm{OC}{{\rm{H}}}_{3}$$

(2)

$${}^{\cdot }\mathrm{CH}_{3}+{}^{\cdot }\mathrm{CH}_{2}\mathrm{OH}\to {\mathrm{CH}}_{3}{\rm{C}}{{\rm{H}}}_{2}\mathrm{OH}$$

(3)

Alcohols with longer carbon chains can be formed from chain elongation of smaller alcohol via secondary and tertiary reactions, from the recombination of alkane and hydroxyl radicals or hydrogenation of their corresponding aldehydes^28,40. Same mechanisms could lead to diethers, such as dimethoxymethane, which requires recombination (4) and (5)

$${\mathrm{CH}}_{3}{\mathrm{OCH}}_{3}+\mathrm{hv}\to {\mathrm{CH}}_{3}{\mathrm{OC}{{\rm{H}}}_{2}}^{\cdot }+{{\rm{H}}}^{\cdot }$$

(4)

$${\mathrm{CH}}_{3}{\mathrm{OC}{{\rm{H}}}_{2}}^{\cdot }+{}^{\cdot }\mathrm{OCH}_{3}\to {\mathrm{CH}}_{3}\mathrm{OC}{{\rm{H}}}_{2}\mathrm{OC}{{\rm{H}}}_{3}$$

(5)

The influence of H₂O on the quantity of alcohol and ether are in total accordance with observation made in ternary ices by Abou Mrad et al.²⁷. An increase in the amount of H₂O inhibits the formation of alcohols and ethers. The mechanism proposed to explain this phenomenon was the availability of H^• radicals favoring recombination with CH₃O^•, thereby reducing the abundance of this radical.

A moderate negative effect of NH₃ is observed for both families. This can be explained by the formation of the amino radical ^•NH₂ generated by the photodissociation of NH₃ (6)^41,42 that recombines with alkane precursors (^•C_xH_y) (7), decreasing their availability to lead to alcohol and ether formation.

$${\mathrm{NH}}_{3}+\mathrm{hv}\to {}^{\cdot }\mathrm{NH}_{2}$$

(6)

$${}^{\cdot }\mathrm{NH}_{2}+{}^{\cdot }{\rm{C}}_{{\rm{x}}}{{\rm{H}}}_{{\rm{y}}}\to {\mathrm{NH}}_{2}{{\rm{C}}}_{{\rm{x}}}{{\rm{H}}}_{{\rm{y}}}$$

(7)

Reactivity leading to aldehydes and ketones

In the case of aldehydes, two different mechanisms are involved in their formation. In the case of acetaldehyde, a dehydrogenation of the corresponding alcohol is suggested (7)^28,34, while for others aldehydes, a recombination between the corresponding alkyl radical R^• and HCO^• (8 and 9) can explain their formation²⁸.

$${\mathrm{CH}}_{3}{\mathrm{CH}}_{2}\mathrm{OH}+\mathrm{hv}\to {\mathrm{CH}}_{3}\mathrm{CHO}+{{\rm{H}}}_{2},$$

(7)

$${{\rm{H}}}_{2}\mathrm{CO}+\mathrm{hv}\to {\mathrm{HCO}}^{\cdot }+{{\rm{H}}}^{\cdot }$$

(8)

$${\mathrm{HCO}}^{\cdot }+{{\rm{R}}}^{\cdot }\to \mathrm{HCOR}$$

(9)

It was indeed shown that in pure CH₃OH ice, aldehydes are formed before their corresponding alcohol²⁸, at the difference of acetaldehyde. The recombination of HCO and ^•CH₃ that could lead to acetaldehyde, should form CH₄ and CO in pure CH₃OH^34,43,44. However, in the presence of H₂O and NH₃ in the ice, the existence of this second pathway of aldehyde formation seems to be favored. Indeed, while no effect of H₂O is observed on ethanol formation, an important increase of acetaldehyde production is observed, confirming the first experiment on ternary ices²⁷. If acetaldehyde was formed from ethanol, the same trend between both compounds should have been observed. Even if this observation does not rule out the first pathway, ethanol dehydrogenation alone cannot explain it. In fact, the presence of H₂O in the ice increases the abundance of HCO^• (10, 11, 12), which is consistent with the increase of the 1854 cm⁻¹ band attributed to HCO^• in the presence of a higher amount of water in the ice (Table S2). This supports previous observations recorded for other aldehydes^28,45,46.

$${\mathrm{OH}}^{\cdot }+{\mathrm{OH}}^{\cdot }\to {{\rm{H}}}_{2}{{\rm{O}}}_{2}$$

(10)

$${}^{\cdot }{\rm{CH}}_{2}{\rm{OH}}+{{\rm{H}}}_{2}{{\rm{O}}}_{2}\to {{\rm{H}}}_{2}{\rm{CO}}+{{\rm{H}}}_{2}{\rm{O}}+{}^{\cdot }{\rm{OH}}$$

(11)

$${{\rm{H}}}_{2}{\rm{C}}{\rm{O}}+{\rm{h}}{\rm{v}}\to {{\rm{H}}{\rm{C}}{\rm{O}}}^{\cdot }+{{\rm{H}}}^{\cdot }$$

(12)

The effect of the initial ice composition is different for ketones. Their formation occurs according to the mechanism (13) and (14), taking the example of the formation of acetone proposed by Tenelanda-Osorio et al.²⁸, consistent with the influence of H₂O and CH₃OH observed.

$$\mathrm{CO}+{}^{\cdot }\mathrm{CH}_{3}\to {\mathrm{CH}}_{3}{\mathrm{CO}}^{\cdot }$$

(13)

$${\mathrm{CH}}_{3}{\mathrm{CO}}^{\cdot }+{}^{\cdot }\mathrm{CH}_{3}\to {\mathrm{CH}}_{3}(\mathrm{CO}){\mathrm{CH}}_{3}$$

(14)

NH₃ has the strongest negative effect on aldehyde and ketone formation, corroborating the first observation on acetaldehyde²⁷. As for other families, NH₂ radical probably inhibited the availability of alkane chains (7), but mainly of CO or HCO^• by reacting with their radical precursor ^•CH₂OH (15) or directly on HCO^• (16), consistent with the amide detection²⁷. NH₃ could also react with H₂O₂⁴⁷, depleting formaldehyde formation (11) and indirectly HCO^• (12).

$${\mathrm{CH}}_{3}{\mathrm{CO}}^{\cdot }+{}^{\cdot }\mathrm{NH}_{2}\to {\mathrm{CH}}_{3}(\mathrm{CO}){\mathrm{NH}}_{2}$$

(15)

$${\rm{H}}{\mathrm{CO}}^{\cdot }+{}^{\cdot }\mathrm{NH}_{2}\to {\rm{H}}(\mathrm{CO}){\mathrm{NH}}_{2}$$

(16)

Reactivity leading to esters

Esters formate are suggested to form from the recombination of HCO^• with the corresponding alkoxy group (17)^25,34,48,49. For other esters whose acyl groups contain a longer carbon chain, a two-step mechanism can be considered: starting from the reaction between an alkyl radical and CO (13) to form the acyl group, followed by the recombination of this acyl group with an alkoxy radical (18)^{25,28,34,48,49}.

$${\rm{H}}{\mathrm{CO}}^{\cdot }+{}^{\cdot }\mathrm{OCH}_{3}\to \mathrm{HCOO}{\mathrm{CH}}_{3}$$

(17)

$${\mathrm{CH}}_{3}{\mathrm{CO}}^{\cdot }+{}^{\cdot }\mathrm{OCH}_{3}\to {\mathrm{CH}}_{3}\mathrm{COO}{\mathrm{CH}}_{3}$$

(18)

These mechanisms clearly explain the difference observed in the various ices between the abundances of methyl esters and esters formate, and the abundances of ethyl acetate. Methyl esters and ester formate are formed from primary radicals directly originating from the methanol photolysis. Their formation is favored by the increase of the methanol quantity initially present in the ice, to the detriment of ethyl acetate, which requires precursors coming from a more diverse reactivity and implying alkyl and alkoxy radicals of longer carbon chain.

It is also interesting to compare the evolution of aldehyde and esters formate, since they are suggested to form both from HCO^•. However, regarding to the ice composition, they do not follow the same trend. Aldehydes present similarities with the evolution of alkanes, while esters formate follow the trend of alcohols. These observations are consistent as their formation is not depending on HCO^• (shared between both chemical families) but on other radicals. Therefore, their abundance differences rely on the relative availability of alkyl for aldehydes, and alkoxy radicals for esters formate. As observed for alkanes and alcohols, these radicals do not present same trends for a same ice composition. The same observation can be made for ketones and methyl esters.

Finally, a strong decrease of their abundances is observed with a high abundance of NH₃ in the ice. Same reactivity as discussed for aldehydes and ketones could explain this behavior, meaning a consumption of radical precursors by NH₂ radical (7, 15, 16).

Reactivity leading to amines and nitriles

One of the most surprising results of this study is the antagonistic effect of the initial composition of the ice on the abundance of amines and nitriles. If the increase in the quantity of CH₃OH and H₂O is insignificant for these two chemical families, the quantity of NH₃ significantly increases the abundance of amines, while strongly decreasing the formation of nitriles.

For the amine formation, several mechanisms are proposed⁵⁰. One way is using the hydrogenation of hydrogen cyanide (19)⁵¹, or directly by ion-molecule reaction (20) then (21).

$$\mathrm{HCN}+{4{\rm{H}}}^{\cdot }\to {\mathrm{CH}}_{3}{\mathrm{NH}}_{2}$$

(19)

$${\mathrm{NH}}_{3}+{\mathrm{CH}}_{3}^{+}\to {\mathrm{CH}}_{3}{\mathrm{NH}}_{3}^{+}$$

(20)

$${\mathrm{CH}}_{3}{\mathrm{NH}}_{3}^{+}+\mathrm{hv}\to {\mathrm{CH}}_{3}{\mathrm{NH}}_{2}+{{\rm{H}}}^{+}$$

(21)

However, regarding the evolution of amine as a function of the ice composition, it seems that the main route in our operating conditions is the direct recombination of ^•CxHy and ^•NH₂ radicals (22)⁵², confirmed by the great increase of their abundances in NH₃-rich ices.

$${}^{\cdot }{\rm{C}}_{2}{{\rm{H}}}_{5}+{}^{\cdot }\mathrm{NH}_{2}\to {{\rm{C}}}_{2}{{\rm{H}}}_{5}{\mathrm{NH}}_{2}$$

(22)

Nitriles are proposed to be formed directly from the photo-processing of their corresponding amines (23–25), with imines as intermediates⁵³.

$${}^{\cdot }\mathrm{NH}_{2}+{{\rm{C}}}_{2}{{{\rm{H}}}_{5}}^{\cdot }\to {{\rm{C}}}_{2}{{\rm{H}}}_{5}{\mathrm{NH}}_{2}$$

(23)

$${{\rm{C}}}_{2}{{\rm{H}}}_{5}{\mathrm{NH}}_{2}+{hv}\to {\mathrm{CH}}_{3}\mathrm{CH}=\mathrm{NH}+2{\rm{H}}$$

(24)

$${\mathrm{CH}}_{3}{\mathrm{CH}}={\mathrm{NH}}+{hv}\to {\mathrm{CH}}_{3}{\mathrm{CN}}+2{\rm{H}}$$

(25)

However, the antagonist evolution of amines and nitriles in NH₃-rich ice implies that a side reaction occurs consuming nitrile precursors, inhibiting nitrile formation. Indeed, if such, a correlation between amine and nitrile formations should occur. Here, a clear abundance decrease of nitriles is observed, in the same sense as observed with other oxidized chemical families (aldehydes, ketones, esters). In fact, ^·NH₂ radical consumes a large part of the first generation of methanol radicals (^·CH₃, ^·CH₂OH, ^·OH) precursors to form amine compounds, opening a competitive reactivity to the formation of CHO compounds (Fig. 4). This explains the high increase of amines with enriched NH₃ ice. Once amine, such as methylamine or ethylamine, are formed, irradiation probably leads to their corresponding imine (24), first step toward the formation of nitrile compounds (25). However, imines are highly reactive in our experimental conditions^54,55,56,57. They easily react with ^·NH₂, NH₃, HCOOH or ^·CN/^─CN to give refractory macromolecules that are part of the remaining organic residues formed after the photoprocessed ice warming. Consequently, a higher amount of nitrogen is incorporated in organic residue before nitrile formation, providing to a lower abundance of nitriles observed in the gas phase, since their imine precursors are trapped in refractory macromolecules.

Fig. 4: Ice reactivity explaining VOCs formation.

Scheme presenting the reactivity possibly occurring in the icy phase after its photo-processing and warming, highlighting by the analysis of volatile organic compounds in the gas phase.

This scenario is in total accordance with the previous experiment focused on the influences of ice composition on residue composition, showing that an increase of NH₃ in the ice can lead to an increase of its nitrogen content²³. It is also supported by the decrease of all carbonyl compounds observed in the gas phase in NH₃-rich ice, corroborated by the decrease of the higher generation of methanol photoproducts, as observed with infrared spectroscopy (CO₂, NCO⁻, CO, HCO, H₂CO, Table S2 and Fig. S7).

Reactivity leading to amides

The two detected amides, acetamide and propionamide, have roughly the same evolution as ketones. They present a higher positive effect with H₂O, a claim supported by previous results showing that H₂O increases nitrogen incorporation into compounds^17,18,23. A negative effect of NH₃ is also explained by the decrease of CO formation, since their formation is based on the same precursors as ketones (13, 14), strengthening their identification as amides. An alternative formation pathway would be the direct incorporation of H₂O in nitriles. However, no correlation was observed between evolutions of nitriles and amides, specifically in the presence of a higher amount of H₂O that implies a decrease of nitriles, while amides increase. If a correlation would exist, a same trend should be observed, or a stronger decrease of nitriles in correlation with the amide increase observed, which is not the case.

Figure 4 summarizes the reactivity in the solid phase that possibly leads to the VOCs identified in the gas phase with our GC-FT-Orbitrap-MS. These results demonstrate that the abundances of volatile organic compounds detected in the gas phase can be correlated with the chemistry occurring in the ice.

Even if further developments have to be performed to estimate the accuracy of using the gas phase as a probe of the surface composition, our current observations corroborate initial experiments focused on CHO compounds, confirming that H₂O plays a trapping role. Interestingly, by identifying nitrogen compounds, we gained insights into the role of NH₃ in the reactivity, and can directly connect gas phase abundance of VOCs to reactivity leading to refractory residue.

It appears that once the NH₃ proportion in the ice is high, an important competition occurs with carbonyl compounds (Table S2). It limits the precursor formation of aldehydes, ketones, and esters in the ice, reducing their abundances, as well as for nitriles. A CH₃OH increase gives an increase of VOC abundance. All these observations are consistent with the organic content of residues that are formed at room temperature after the warming up of the photoprocessed ices^15,23,58. It was indeed demonstrated that an increase of H₂O in the ice leads to an increase of nitrogen incorporation in organic residues in the same sense as it is observed in the gas phase. The same observation was made about the amines detected in the gas phase. Furthermore, it was shown that a higher NH₃ content in the ice leads to an increase of nitrogen incorporation, which is intuitive, but with a decrease in the average DBE²³. This is in agreement with our gas phase observation, where a higher amount of NH₃ leads to the detection of higher abundances of saturated compounds (ethylamine and propylamine), while unsaturated compounds in the gas phase present a decrease of their abundances due to the incorporation of their precursor (imines, ketenimines) in refractory residues.

Consequently, probing the gas phase may give information on the composition of the surface of icy environments, by taking into account the ratio between chemical families observed in the gas phase. This approach may be used in a near future to analyze observations in various astrophysical environments such as hot cores, protoplanetary disks or cometary environments.

Conclusion

Ice in extraterrestrial environments are known to be an important source of organic matter in various environments. In this contribution, we evaluate ices as a possible source of organics in the gas phase. Using gas chromatography coupled to a high-resolution mass spectrometry linked to a vacuum chamber, 24 volatile organic compounds were detected in the sublimated gas phase mixture, and tentatively identified 15 additional VOCs, including alkane, alkene, ester, ketone, ether, alcohol, aldehyde, and to the best of our knowledge, for the first time amine, nitrile and amide compounds, as well as some bifunctional compounds. Abundances of these VOCs were monitored depending on ice compositions to obtain a complete scheme of the reactivity leading to their formation. Furthermore, variation of molecular ratios in the initial ice provided information on VOCs formation. Particularly, in our operational conditions, NH₃ tends to compete with carbonyl compound formation and drive the formation nitrogen refractory compounds. Finally, all observations of VOCs evolution are directly explained by the reactivity occurring in the ice during its formation and photo-processing. It thus appears that ice can be an important reservoir of VOCs that can be released in the gas phase once in a warmer zone. Furthermore, the quantification of VOCs in the gas phase could give indirect information on the composition and reactivity occurring on the surface of an icy astrophysical object. Finally, this experimental work demonstrates the necessity to correlate ALMA observations giving information on the gas phase with the ones of JWST focusing on solid phase observations. By coupling these observations to the present work, information could be obtained to better understand the role of ices in VOCs observed in gas phases.

Methods

Ice photo-processing and volatile organic compounds recovery

The experimental setup for the synthesis of interstellar analogs is described in detail in refs. ^59,60,61. Briefly, it is composed of a sample holder holding a MgF₂ window in a vacuum chamber (2.10⁻⁷ to 8.10⁻⁸ mbar). The sample holder where the ice is formed is set at 77 K using liquid nitrogen (Linde). The gas deposition for the ice formation is performed at the same time as the irradiation using an H₂ discharge lamp with continuous flow at 5 mbar (Lyman α, 121 nm, (3.7 ± 0.6).10¹⁴ photon*cm⁻²*s⁻¹). A small contribution is also present at 160 nm. As previously described in Danger el al.¹⁵ These conditions simulate an environment near the CO snowline, where condensation on grain can occur, while the UV photon flux is still abundant. Furthermore, during this process, photodissociation of H₂O, CH₃OH, and NH₃ in the gas phase probably occurs as well as in the ice phase, since the photodissociation cross section is higher in the gas phase²¹. Consequently, radical precursors can have these two origins, while organic molecules observed are formed in the icy phase³². The photon flux was calibrated using replicated actinometry on a methanol ice irradiation, monitoring the CO band of methanol (1047–1006 cm⁻¹: (1.8 ± 0.4).10⁻¹⁷ cm*molecule⁻¹)⁶². Gas mixtures for ice formation were prepared in a reservoir with different proportions of H₂O (purified with a Millipore Direct Q5 system), CH₃OH (prepared from Fisher Chemicals, 99.99%) and NH₃ (purchased from AlphagazTM, 99.98%). The deposition rates and the molecular compositions of ice are estimated during the first hour of experiments and without the UV irradiation using a BRUKER Vector 22 FTIR Spectrometer in transmission mode (20 scans from 960 to 4000 cm⁻¹ with 2 cm⁻¹ of resolution and 400 scans of background). The following band strengths were used for each ice species H₂O (3612–3003 cm⁻¹: 2.10⁻¹⁶cm*molecule⁻¹), CH₃OH (1047–1006 cm⁻¹: (1.8 ± 0.4).10⁻¹⁷cm*molecule⁻, 3612–3003 cm⁻¹: (1.0 ± 0.2).10⁻¹⁶cm*molecule⁻¹) and NH₃ (¹1192 –1084 cm⁻¹: (1.6 ± 0.3).10⁻¹⁷cm*molecule⁻¹, 3612–3003 cm⁻¹: (2.3 ± 0.5).10⁻¹⁷cm*molecule⁻¹)⁶². The deposition rate was obtained by recording infrared spectra at different times during the first hour of ice formation. Using bands between 3612 and 3003 cm⁻¹, 1047 and 1006 cm⁻¹, and 1192 and 1084 cm⁻¹ for H₂O, CH₃OH, and NH₃, abundances were estimated at each time, and the slope for each molecule gave the deposition rate. The apparent irradiation rate was obtained by dividing the deposition rate by the photon flux of the UV lamp. Table 1 summarizes ice compositions, the apparent irradiation rate (T) and the deposition rate estimated previously. Once the photoprocessed ice is formed, the chamber is isolated from the turbo-molecular pumps and the cooling system is switched off, leading to an average heating rate of 2 K/min up to room temperature to obtain VOCs desorption. VOCs are then transferred to the GC-FT-Orbitrap-MS by adding ultrapure argon (Linde Ar 6.0) in the chamber to obtain a total pressure of 200 mbar. VOCs are preconcentrated in a specific interface in a 2.5-mL stainless steel tube cooled at 77 K with liquid nitrogen. Once recovered, the tube is warmed up and helium is added to a total pressure of 100 mbar. The resulting VOCs/helium gas mixture is then transferred to the injection unit of the GC that is formed by a stainless injection loop of 500 µL at 10⁻²mbar. Once filled out, the gas sample is driven with helium to GC injector. Additional information can be obtained in the following refs. ^60,63. Control experiments were performed to evaluate possible contamination coming from the experimental set-up: one consisting in depositing ice at 77 K without irradiation, and another one performing irradiation at 77 K without ice deposition (Fig. S17).

GC-FT-Orbitrap-MS configuration

The GC-MS analysis program is identical to the one used in ref. ⁶⁴. The chosen chromatographic column is Rxi-624-Sil MS from Restek (60 m × 0.25 mm × 1.4 μm). The helium flow rate (Linde-France Scientific 6.0) is set at 1 mL/min with a vacuum compensation, a purge flow rate of 5.0 mL/min and a split of 10.0 mL/min. The temperature in the oven was from 35 °C (3 min) up to 130 °C at 2 °C/min and finally reached 280 °C for 2 min at 30 °C/min. The transfer line and injectors are set at 280 °C while the ionization source is set at 250 °C with electrons accelerated to 70 eV. VOCs are first concentrated in a deactivated silicate column with a cryotrap from Frontier Lab⁶¹.

Data treatment

To obtain the detection of peaks in the chromatogram and their quantification, TraceFinder 4.1 software was used with a signal-to-noise greater than three. The integrating area (R) of the ion of each peak was corrected with the blank corresponding to the GC-FT-Orbitrap-MS analysis of an empty preconcentration loop. Control experiments have indeed shown that contamination came from this part of the experimental set-up. That is why blanks were performed between each experiment, as well as three replicates per ice composition. The compounds detected are then tentatively identified by comparing the fragmentation mass patterns to a high-resolution mass database containing 150 VOCs. When the database is not matching the experimental mass spectra, the tentatively identified VOCs are classified in groups (CH, CHO, CHO₂, CHN or CHNO) by taking into account the high resolution of the Orbitrap giving stoichiometric formula for each ion. Furthermore, mass patterns of tentatively identified VOCs are compared to our database and the NIST database to tentatively identify common patterns in order to propose a possible identification, noted in italic in Table S1. Additional information on compound identifications are given in Supplementary Information (Figs. S15 and S16).