Probing the chemical complexity of amines in the ISM: detection of vinylamine (C₂H₃NH₂) and tentative detection of ethylamine (C₂H₅NH₂)

The term prebiotic molecules refers to species that are considered to be involved in the processes leading to the origin of life. From the prospective of understanding the prebiotic chemistry in the interstellar medium (ISM), primary amines have caught much attention because they contain the same NH₂ group as those considered to be the fundamental building blocks of life (e.g. amino acids, nucleobases, nucleotides and other biochemical compounds). Although the attempts to observe glycine (NH₂CH₂COOH), the simplest amino acid, in the ISM have not succeeded (e.g. Ceccarelli et al., 2000; Belloche et al., 2013), an increasing number of NH₂-bearing species have been reported in the last few years, which increases the chance to discover pre-biotic complex organics in space. This is in fact well-demonstrated towards G+0.693-0.027 (hereafter G+0.693), a molecular cloud located within the Sgr B2 complex in the Galactic Centre. As revealed by the census of N-bearing species towards G+0.693 (Zeng et al., 2018), several species containing NH₂ such as cyanamide (NH₂CN), formamide (NH₂CHO), and methylamine (CH₃NH₂) have been detected with abundance $\geq$10^-9. Following the search, two key precursors in the synthesis of prebiotic nucleotides, hydroxylamine (NH₂OH) and urea (NH₂CONH₂) have also been identified towards G+0.693 (Rivilla et al., 2020; Jiménez-Serra et al., 2020). The very recent discovery of by far the most complex amine, ethanolamine (NH₂CH₂CH₂OH), the simplest head group of phospholipids in cell membrances, towards G+0.693 attests the potential of this source for finding more complex molecules of prebiotic relevance (Rivilla et al., 2021). This has thus prompted us to keep hunting for more amines in order to understand the chemical processes yielding related ingredients for life in space.

Structurally analogous to amino acids, CH₃NH₂ is the only primary amine that has been unambiguously observed towards different astronomical objects (e.g. Kaifu et al., 1974; Zeng et al., 2018; Bøgelund et al., 2019; Ohishi et al., 2019). Vinylamine (also known as ethenamine, C₂H₃NH₂) and ethylamine (C₂H₅NH₂) have not yet been reported in the ISM although the latter has been identified in multiple meteorites (Aponte et al., 2020, and references therein) and materials returned by the Stardust mission from comet 81P/Wild 2 (e.g. Glavin et al., 2008). Due to some concerns on the possible contamination of the stardust sample, their cometary origin could not be confirmed until the robust detection in the coma of comet 67P/Churyumov-Gerasimenko (Altwegg et al., 2016). The detection of CH₃NH₂ and C₂H₅NH₂ together with glycine in these solar system objects enhanced the probability of amino acids being formed in space. Indeed, the retrosynthesis of amino acids revealed that molecules containing the -NH₂ functional group are likely precursors of amino acids (Förstel et al., 2017). For example, CH₃NH₂ and C₂H₅NH₂ may be the major constituent of glycine and alanine (NH₂CH₃CHCOOH) respecitvely. According to the detailed quantum chemical calculation, C₂H₃NH₂ is the next energetically stable isomer in C₂H₅N group after E- and Z- conformer of ethanimine (CH₃CHNH) (Sil et al., 2018), both of which have been detected in Sgr B2 (Loomis et al., 2013) as well as in G+0.693 (Rivilla et al. in prep.). C₂H₅NH₂ exists in two forms: anti- and gauche-conformer. The former is known to be more stable and has the higher expected intensity ratio than the latter conformer (see Sil et al., 2018, for details). Therefore anti-C₂H₅NH₂ should be the most viable candidate for the astronomical detection in the C₂H₇N group. In this letter, we present the first detection of C₂H₃NH₂ and tentative detection of C₂H₅NH₂ in the ISM towards G+0.693 through the identification of several rotational transitions of its millimeter spectrum.

We have carried out high-sensitivity spectral surveys at 7, 3, and 2 mm towards G+0.693 molecular cloud using the IRAM 30m¹¹1IRAM is supported by INSU/CNRS (France), MPG (Germany), and IGN (Spain) and Yebes 40m²²2http://rt40m.oan.es/rt40m en.php telescopes. The observations were centred at $\alpha$(J2000) = 17^h47^m22^s, $\delta$(J2000) = -28^∘21^′27^′′. The position switching mode was used in all observations with the reference position located at $\Delta\alpha$, $\Delta\delta$ = -885${}^{{}^{\prime\prime}}$,290^′′ with respect to the source position. The half-power beam width (HPBW) of the IRAM 30m and Yebes 40m telescope are in a range of 14^′′-36^′′ at observed frequencies between 30 GHz and 175 GHz. The intensity of the spectra was measured in unit of antenna temperature, T${}_{\rm A}^{*}$ as the molecular emission toward G+0.693 is extended over the beam (Zeng et al., 2020). The IRAM 30 m observations were performed in three observing runs during 2019: April 10-16, August 13-19 and December 11-15, from projects numbers 172-18 (PI Martín-Pintado), 018-19 (PI Rivilla) and 133-19 (PI Rivilla). It covered spectral ranges of 71.76-116.72 GHz and 124.77-175.5 GHz. We refer to Rivilla et al. (2020) for a full description of the IRAM 30 m observations. The Yebes 40 m observations were carried out during 6 observing sessions in February 2020, as part of the project 20A008 (PI Jiménez-Serra). The new Q band (7mm) HEMT receiver was used to allow broad-band observations in two linear polarisations. The spectral coverage ranges from 31.075 GHz to 50.424 GHz. We refer to Zeng et al. (2020) and Rivilla et al. (2020) for more detailed information on the Yebes 40 m observations.

The line identification and analysis were carried out using the SLIM (Spectral Line Identification and Modelling) tool implemented within the madcuba package³³3Madrid Data Cube Analysis on ImageJ is a software developed at the Center of Astrobiology (CAB) in Madrid; http://cab.inta-csic.es/madcuba/Portada.html.(version 21/12/2020, Martín et al., 2019). The spectroscopic information of C₂H₃NH₂, within 0${}^{+}\&$ 0^- (CDMS⁴⁴4Cologne Database for Molecular Spectroscopy (Endres et al., 2016). https://cdms.astro.uni-koeln.de/classic/ entry 43504) was obtained from Brown et al. (1990) and Mcnaughton & Robertson (1994). And the spectroscopic information of C₂H₅NH₂, anti-conformer (CDMS entry 45515) was obtained from Fischer & Botskor (1982) and Apponi et al. (2008). Table 1 summarises the unblended or only partially blended transitions of C₂H₃NH₂ and C₂H₅NH₂ detected towards G+0.693. Note that all the C₂H₅NH₂ lines in Table 1 are a-type transitions with selection rules of $\Delta$K_a=0 and $\Delta$K_c=$\pm$1. The rotational spectrum of C₂H₃NH₂ and C₂H₅NH₂ are characterised by inversion doubling due to the large amplitude inversion motion of the NH₂ group. Each rotational energy level, specified by the rotational quantum numbers J and K, is thus split into an inversion doublet. The 0⁺ or 0^- in Table 1 indicate the inversion doublet from which the transition arises. For C₂H₅NH₂, the variation of upper state degeneracy and integrated intensity of the same transition is due to the spin statistical weight of 3:1 between the symmetric and anti-symmetric sub-levels.

The analysis was performed under the assumption of local thermodynamic equilibrium (LTE) conditions due to the lack of collisional coefficients of C₂H₃NH₂ and C₂H₅NH₂. Due to the low density of G+0.693 ($\sim$10⁴-10⁵ cm^-3; Zeng et al., 2020), molecules are sub-thermally excited in the source and hence their excitation temperatures (in a range of 5 to 20 K; e.g. Requena-Torres et al., 2008; Rivilla et al., 2018; Zeng et al., 2018) are significantly lower than the kinetic temperature of the source ($\sim$150 K; e.g. Zeng et al., 2018). However, note that the transitions of C₂H₃NH₂ and C₂H₅NH₂ are well fitted using one excitation temperature for each species. Considering the effect of line opacity, MADCUBA-SLIM generated synthetic spectra that can be compared to the observed spectra. The MADCUBA-AUTOFIT tool was then used to provide the best non-linear least-squares LTE fit to the data using the Levenberg-Marquardt algorithm. It is important to note that not a single transition of C₂H₃NH₂ and C₂H₅NH₂ predicted by the LTE spectrum is missing in the data. To properly evaluate the line contamination by other molecules, over 300 species have been searched for in our dataset. This included not only all the molecules detected towards G+0.693 in previous studies (Requena-Torres et al., 2008; Zeng et al., 2018; Rivilla et al., 2018, 2019, 2020, 2021; Jiménez-Serra et al., 2020; Rodríguez-Almeida et al., 2021), but also those reported in the ISM⁵⁵5See https://cdms.astro.uni-koeln.de/classic/molecules. The molecule blended with the transition of C₂H₃NH₂ and C₂H₅NH₂ is listed in the last column of Table 1. We note that the line identification and fitting file used in this analysis is the same as the one used in previous works (e.g. Zeng et al., 2018; Rodríguez-Almeida et al., 2021) as well as on-going works. Therefore, the excitation temperatures and column densities of blending species are consistent with the ones reported in all these studies.

For C₂H₃NH₂, we detect five clean transitions and four slightly blended transitions with a blending contribution of $<$10$\%$ (see Figure 1). For C₂H₅NH₂, only four clean transitions are reported and three are slightly blended (Figure 2). Since the clean transitions of C₂H₅NH₂ are weak (S/N=4 in integrated intensity; Table 1), we conclude that this species is tentatively detected. The free parameters that can be fitted are: molecular column density ($\textit{N}_{\rm tot}$), excitation temperature ($\textit{T}_{\rm ex}$), radial velocity ($\textit{V}_{\rm LSR}$), full width half maximum (FWHM, $\Delta\nu$), and source size ($\theta$). For G+0.693, we assumed that the source size is extended in madcuba. As the algorithm did not converge when fitting C₂H₃NH₂ with all parameters left free, we fixed the FWHM and $\textit{V}_{\rm LSR}$ to 18 km s^-1 and 67  km s^-1 respectively by visual inspection of the most unblended lines. These values are consistent with those from many other molecules previously analysed in G+0.693 (FWHM$\sim$20 km s^-1, $\textit{V}_{\rm LSR}\sim$68 km s^-1; Requena-Torres et al., 2008; Zeng et al., 2018; Rivilla et al., 2018, 2019, 2020; Jiménez-Serra et al., 2020; Rodríguez-Almeida et al., 2021). The resulting LTE fit gives $\textit{T}_{\rm ex}$=(18$\pm$3) K and $\textit{N}_{\rm tot}$=(4.5$\pm$0.6)$\times$10¹³ cm^-2. In the case of C₂H₅NH₂, the $\textit{V}_{\rm LSR}$ was fixed to 67 km s^-1 and the LTE fit gives $\textit{T}_{\rm ex}$=(12$\pm$5) K, FWHM=(18$\pm$5) km s^-1 and $\textit{N}_{\rm tot}$=(2.5$\pm$0.7)$\times$10¹³ cm^-2.

The resulting best LTE fit to the C₂H₃NH₂ and C₂H₅NH₂ lines is shown in red line in Fig.1 and Fig.2 whilst blue line indicates the best fit considering also the total contribution of the LTE emission from all the other identified molecules. Adapting the H₂ column density inferred from observations of C¹⁸O ($\textit{N}_{\rm H_{2}}$ = 1.35$\times$10²³ cm^-2; Martín et al., 2008), the resulting abundance is (3.3$\pm$0.4)$\times$10^-10 and (1.9$\pm$0.5)$\times$10^-10 for C₂H₃NH₂ and C₂H₅NH₂ respectively.

Figure 3 presents the abundance with respect to H₂, in decreasing order, of NH₂-bearing species detected towards G+0.693. The results are compared to those derived towards three high-mass and low-mass star-forming regions that are chemically rich, i.e. Sgr B2(N), Orion KL, and IRAS 16293-2422 B. The low abundance or non-detection of -NH₂ species may indicate that their formation is less efficient towards Orion KL and IRAS 16293-2422 B. On the other hand, the detection with abundance $>$10^-11 suggests G+0.693 is a prominent -NH₂ molecule repository which allows us to study their origin as well as their chemical relation to other pre-biotic molecules.

In contrast to the three compared sources, G+0.693 lacks an internal heating source responsible for the rich chemistry. The high level of molecular complexity is attributed to dust grain sputtering by low-velocity shocks ($\leq$20 km s^-1), which is driven by the possible cloud-cloud collision occurring in the Sgr B2 complex (Zeng et al., 2020). This is indicated by the high abundances of shock tracers such as HNCO and SiO (Martín et al., 2008; Rivilla et al., 2018) and the presence of molecules that are known to be formed on grain surfaces (Requena-Torres et al., 2008; Zeng et al., 2018). In addition, the abundance ratio of HC₃N/HC₅N and C₂H₃CN/C₂H₅CN derived in Zeng et al. (2018) suggested that an enhanced cosmic-ray ionisation rate may also play a role in the chemistry of N-bearing species towards G+0.693. In the following Section, we evaluate the possible formation routes for the amines detected towards G+0.693 and the discussed formation routes are summarised in Figure 4.