EDP Sciences logo
Subscriber Authentication Point
Search
Menu
Home All issues Volume 706 (February 2026) A&A, 706 (2026) A316 Full HTML
Open Access
Issue
A&A
Volume 706, February 2026
Article Number A316
Number of page(s) 18
Section Interstellar and circumstellar matter
DOI https://doi.org/10.1051/0004-6361/202558316
Published online 23 February 2026

© The Authors 2026

Licence Creative CommonsOpen Access article, published by EDP Sciences, under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

This article is published in open access under the Subscribe to Open model. This email address is being protected from spambots. You need JavaScript enabled to view it. to support open access publication.

1 Introduction

Structural isomerism, in which molecules share an identical molecular formula but differ in the connectivity of their constituent atoms, represents a fundamental source of chemical diversity in the interstellar medium (ISM). From a chemical point of view, this diversity translates into distinct isomers often exhibiting markedly different electronic structures, thermodynamic stabilities, and spectroscopic signatures, making the identification of individual isomers a valuable tool for distinguishing among their plausible formation pathways (Shingledecker et al. 2019; Neill et al. 2012; Marcelino et al. 2021; Rivilla et al. 2023; García de la Concepción et al. 2023; San Andrés et al. 2024; Remijan et al. 2025; Sanz-Novo et al. 2025).

Within this context, the C3H6O2 isomeric family has garnered particular attention in recent years from the astrochemical community (Savarese et al. 2025). It encompasses molecules with diverse functional groups, ranging from families that are widespread in nature such as aldehydes (i.e., 2-hydroxypropanal or lactaldehyde, CH3CH(OH)C(O)H, 3-hydroxypropanal, HO(CH2)2C(O)H, and methoxyacetaldehyde, CH3OCH2C(O)H), esters (i.e., methyl acetate, CH3C(O)OCH3, and ethyl formate, CH3CH2OC(O)H), ketones (hydroxyacetone, CH3C(O)CH2OH), and carboxylic acids (propionic acid, CH3CH2C(O)OH), to more structurally exotic compounds such as epoxy alcohols (glycidol, c-CH2OCHCH2OH), diols, and enediols (i.e., prop-1-ene-1,2-diol, CH3C(OH)CHOH, prop-2-ene-1,2-diol, CH2C(OH)CH2OH, and prop-1-ene-1,3-diol, HOCH2CHCHOH).

Identifying which of these isomers are present in the ISM requires the synergy of accurate high-resolution laboratory rotational data, sensitive astronomical observations, and astrochemical modeling. These combined efforts have thus far yielded the detection of both the anti and gauche conformers of ethyl formate (Belloche et al. 2009; Tercero et al. 2013; Rivilla et al. 2017; Peng et al. 2019), methyl acetate (Tercero et al. 2013), and hydroxyacetone (Zhou et al. 2020) toward various starforming regions, as well as the recent tentative detection of 3-hydroxypropanal toward the Galactic center (GC) molecular cloud G+0.693-0.027 (hereafter G+0.693, Fried et al. 2025). Meanwhile, fruitless searches have been reported for propionic acid, the global minimum in energy, and for higher energy isomers such as methoxyacetaldehyde and lactaldehyde in Orion KL and Sagittarius B2(N) (Ilyushin et al. 2021; Belloche et al. 2025; Alonso et al. 2019; Kolesniková et al. 2018).

In the laboratory, experiments have shown that multiple C3H6O2 isomers can form under typical astrophysical conditions. Wang et al. (2023) reported the production of hydroxyacetone, methyl acetate, and 3-hydroxypropanal -all of which have already been identified in the ISM- along with their enol tautomers in UV-irradiated methanol-acetaldehyde ices, offering insights into their possible formation pathways. In a subsequent study, Wang et al. (2024) demonstrated that the irradiation of CO-ethanol ices yields, in addition to the known interstellar species ethyl formate and 3-hydroxypropanal, other isomers still uncharted in the ISM, such as 1,3-propenediol and lactaldehyde.

In light of these findings, a systematic survey of all the spectroscopically characterized C3H6O2 isomers (see Figure 1) will be essential for expanding the known interstellar chemical inventory of complex organic molecules (COMs, defined as carbon-based molecules comprised of six or more atoms; Herbst et al. 2020) and for unveiling whether a particularly complex, yet unknown underlying chemistry involving relatively large O-bearing species occurs in space.

Over the last few years, the search for COMs has increasingly turned toward unusual environments in the GC such as the GC molecular cloud G+0.693. This cloud appears as one of the most compelling sources for the potential detection of the remaining C3H6O2 isomers, as supported by the growing inventory of COMs discovered in it (see, e.g., Rivilla et al. 2023; Sanz-Novo et al. 2025). It includes new complex O-bearing species with one C atom such as cis-carbonic acid (HOCOOH; Sanz-Novo et al. 2023), two C atoms such as (Z)-1,2-ethenediol ((CHOH)2, Rivilla et al. 2022a) and trans-methyl formate (CH3OCHO; Sanz-Novo et al. 2025), and three C atoms such as propanol (C3H7OH; Jiménez-Serra et al. 2022) and 3-hydroxypropanal (Fried et al. 2025). The latter, which was tentatively detected earlier this year, motivated us to carry out the present follow-up study.

Our analysis has confirmed the presence of six additional C3H6O2 isomers in G+0.693, including the first interstellar detection of lactaldehyde and methoxyacetaldehyde, together with the second interstellar detection (confirmation) of methyl acetate and hydroxyacetone and the fourth detection of ethyl formate in the ISM. Propionic acid and glycidol remain undetected, but we have also reported the upper limits of their molecular column densities, setting new constraints on their abundances in the ISM.

thumbnail Fig. 1

Relative energies (including zero-point energy, ZPE, corrections) plotted as a function of the total dipole moment for the targeted C3H6O2 isomers, computed at the B2PLYPD3∕aug-cc-pVTZ level of theory (Grimme et al. 2011) using the Gaussian 16 program package (Frisch et al. 2016). Optimized 3D structures at the same level of theory are also shown (carbon atoms in gray, oxygen atoms in red, and hydrogen atoms in white). The status of detections in G+0.693 is indicated using the following color-code in the molecule name: new first detections in the ISM are shown in red, tentative detections in orange, non-detections in purple, and previous detections toward different sources in black. Previous detections toward other sources of methyl acetate (Tercero et al. 2013), anti and gauche ethyl formate (Belloche et al. 2009; Tercero et al. 2013; Rivilla et al. 2017; Peng et al. 2019), hydroxyacetone (Zhou et al. 2020), and 3-hydroxypropanal are also included. Also, new detections based in this work are highlighted in boldface. The 3D representations of all molecules are visualized with IQmol (https://www.iqmol.org/).
Table 1

Derived physical parameters for C3H6O2 isomers that we searched for toward G+0.693-0.027.

2 Observations

The astronomical search for the C3H6O2 isomers was performed toward the GC molecular cloud G+0.693, located approximately 55” north-east of the well-known high-mass star-forming region Sgr B2(N). This source exhibits intermediate H2 densities of 104−105 cm−3 that lead to the subthermal excitation of the molecular emission, with excitation temperatures between ~5-20 K (well below the kinetic temperature of the gas of 70-150 K; Requena-Torres et al. 2006; Zeng et al. 2018; Jiménez-Serra et al. 2020; Colzi et al. 2024). The spectra toward this source are, therefore, significantly less crowded than those observed toward hot cores, where some of the C3H6O2 isomers have previously been detected and where vibrationally excited states are also efficiently populated. This considerably reduces the line blending and facilitates the identification of new interstellar species through a smaller number of rotational transitions.

In this study, we employed an ultra-deep spectral survey of this source that combines centimeter- and millimeterwavelength observations acquired with two facilities. The Q-band range (31.075-50.424 GHz) was surveyed using the Yebes 40-m telescope in Guadalajara, Spain (projects 20A008 and 21A014), while the IRAM 30-m telescope in Granada, Spain, provided complementary coverage in two broadband millimeter-wavelength frequency windows (71.8-116.7 GHz, 124.8-175.5 GHz; see Rivilla et al. 2023; Sanz-Novo et al. 2023 and references therein; projects 172-18, 018-19, 133-19, 12322). Data were collected in position-switching mode, centered at α = 17h47m22s, δ = −28°21′27″ (J2000), with an off-source reference located at ∆α = −885″ and ∆δ = 290″. The Yebes telescope’s half-power beam width (HPBW) spans ~55″-35″ from 31 to 50 GHz (Tercero et al. 2021), whereas the IRAM 30-m HPBW ranges from ~32″ to ~10″ across the targeted frequencies. Given that molecular emission toward G+0.693 is extended compared to the beam size of the telescope (Jones et al. 2012; Li et al. 2020; Zheng et al. 2024), we report all spectra in antenna temperature units (T*A). Overall, we achieved root-mean-square (rms) noise levels lying between 0.25 and 0.9 mK across the whole Q band, 0.5-2.5 mK at 3 mm, and 1.0-1.6 mK at 2 mm. Further details on these observations and their reduction are available in Rivilla et al. (2023) and Sanz-Novo et al. (2023).

3 Results

3.1 LTE analysis

We carried out a local thermodynamic equilibrium (LTE) analysis of all C3H6O2 isomers for which rotational spectroscopic data are available using the Spectral Line Identification and Modeling (SLIM) tool (version from 15 June, 2024) within the MADCUBA software package (Martín et al. 2019). This tool allowed us to produce a synthetic spectrum of each target species and compare it directly with the observed spectra, as well as to evaluate the emission of all the molecules previously identified in the current molecular line survey (i.e., over 140 species).

To obtain the best-fit LTE model for each molecule, we used the AUTOFIT tool within SLIM (Martín et al. 2019) and performed a nonlinear least-squares LTE fit of a selected subset of clean transitions to the observed data using the LevenbergMarquardt algorithm. We note that heavily blended lines, contaminated by the emission of previously identified species, were excluded from the analysis. For the fit, we can leave the following as free parameters: the molecular column density (N), the excitation temperature (Tex), the radial velocity (vLSR), and the line width (FWHM). As described in the following subsections, we fixed certain parameters for some of the studied species to that derived for the most abundant C3H6O2 isomer, methyl acetate, which also presents the brightest transitions, to ensure convergence of the AUTOFIT.

In addition, to derive the fractional abundances with respect to molecular hydrogen for all targeted molecules, we adopted N(H2) = 1.35×1023 cm−2 from Martín et al. (2008), assuming an uncertainty of 15% of its value. For comparison and completeness, in Table 1 we list the physical parameters derived for all C3H6O2 isomers, including those reported for 3-hydroxypropanal in Fried et al. (2025).

thumbnail Fig. 2

Selected unblended or slightly blended transitions of methyl acetate, CH3C(O)OCH3, identified toward G+0.693, which were used to derive the LTE physical parameters of the molecule (see text; listed in Table A.1). The red line shows the best LTE fit of CH3C(O)OCH3, while the blue line represents the combined emission of all molecules identified in the survey (observed spectra are shown as gray histograms). The symmetry state of the detected transitions is shown at the bottom of each panel (full list of quantum numbers listed in Table A.1). The * symbol indicates that there are various transitions in the same panel, each comprising the same E3/E4/E1 symmetry states. The transitions are sorted by decreasing intensity. The molecular structure of CH3C(O)OCH3 is also shown (C atoms in gray, O atoms in red, and H atoms in white).

3.2 Detection of methyl acetate

Methyl acetate, CH3 C(O)OCH3, is the second most stable species within the C3H6O2 isomeric family and was first detected in Orion KL by Tercero et al. (2013). However, this molecule has not yet been identified in any other astronomical environment. Here, we searched for CH3C(O)OCH3 toward G+0.693, which resulted in an unequivocal detection (see Figure 2), thus pointing to a broader presence of methyl acetate in the ISM. For the astronomical search, we employed the spectroscopic catalog derived from the original analysis using the BELGI-Cs−2Tops code (Tercero et al. 2013), which we converted into the common SPCAT/SPFIT format (Pickett 1991) (see Appendix A). This molecule exhibits two non-equivalent CH3 internal rotors, which split the rotational energy levels into five substates (i.e., AA, AE, EA, and two EE species, E3 and E4, with statistical weights of 16, 16, 16, 8, and 8, respectively). In the transformed catalog, these symmetry species are labeled using the so-called vibrational quantum number as follows: 0 = AA, 1 = E1 (EA), 2 = E2 (AE), 3 = E3 (EE), 4 = E4 (EE) (see Table A.1).

In Figure 2, we show a selection of the brightest and least contaminated lines of CH3C(O)OCH3 that have been detected toward G+0.693, which were subsequently used to derive the physical parameters of the targeted molecular emission. Moreover, the fine structure corresponding to transitions belonging to the five symmetry states is, in some cases, partially resolved, giving rise to easily recognizable patterns, while in other cases it is fully merged into a single, broader spectral feature due to the typically broad line width of the molecular line emission measured toward G+0.693 (i.e., FWHM ~ 15-20 km s−1; Requena-Torres et al. 2006, 2008; Zeng et al. 2018; Rivilla et al. 2022b). However, this “autoblending” likewise facilitates the detection because it strengthens the intensity of the resulting line cluster (e.g., 161,16−151,15 transition at ~101.2512 GHz).

To yield the best LTE modeling for CH3C(O)OCH3, we used the AUTOFIT tool for the clean subset of transitions shown in Figure 2 (spectroscopic information reported in Table A.1). For the fit, we fixed the FWHM to a value of 21 km s−1, which reproduces the brightest and cleanest transitions well, while the N, Tex and vLSR were left as free parameters. We obtained a N = (2.2 ± 0.2) × 1014 cm−2, which corresponds to a fractional abundance with respect to molecular hydrogen of (1.6± 0.3) × 10−9. Furthermore, we derived a Tex = 15 ± 2 K and a vLSR = 68 ± 1 km s−1, consistent with those employed for the analysis of its isomer 3-hydroxypropanal (Fried et al. 2025) (see Table 1). We note that a change of 1-2 K in the Tex only slightly affects the derived N, and in all cases this variation is smaller than the uncertainty assumed in the derivation of the fractional abundance with respect to H2 (i.e., 15%).

3.3 Detection of anti-ethyl formate and tentative detection of gauche-ethyl formate

Next in energy, we find ethyl formate (CH3CH2OC(O)H). Its two lowest-in-energy conformers1, anti and gauche, are separated by 65 ± 21 cm−1 (or 94 K; Riveros 1967) with the anti configuration being the global minimum in energy. The anti conformer (hereafter a-CH3CH2OC(O)H) has been detected in three hot molecular cores so far: Sgr B2(N), Orion KL, and W51 e2 (Belloche et al. 2009; Tercero et al. 2013; Rivilla et al. 2017; Peng et al. 2019), with the latter two sources also harboring gauche CH3CH2OC(O)H (hereafter g-CH3CH2OC(O)H) in an approximately 1:1 abundance ratio.

In this work, we report a new detection of a-CH3CH2OC(O)H toward G+0.693, while g-CH3CH2OC(O)H is tentatively detected due to non-negligible line blending for the majority of lines (see Figures 3 and A.1). To perform the LTE analysis corresponding to the a-CH3CH2OC(O)H emission, we used the spectroscopic catalog from the Cologne Database for Molecular Spectroscopy (CDMS, entry 074514; Endres et al. 2016), which is based on the laboratory data by Riveros (1967); Demaison et al. (1984); Medvedev et al. (2009). For the g-conformer, we employed a new spectroscopic entry using the aforementioned data, since the CDMS entry includes the rotational lines of g-CH3CH2OC(O)H shifted by the corresponding ∆E(a-g); thus, they are not suitable for the analysis in regions whose molecular emission is well described by low excitation temperatures. Hence, we considered each conformer as distinct molecular species, as previously done for other molecules in G+0.693 (see, e.g., Jiménez-Serra et al. 2022; Sanz-Novo et al. 2023, 2025).

After inspecting the astronomical data, various unblended or slightly blended a-type transitions ranging from J = 6 to J = 9, with Ka < 3 and belonging to a-CH3CH2OC(O)H were detected. A selection of the brightest and cleanest transitions, which were subsequently used in the LTE fit with MADCUBA, are shown in Figure 3 (spectroscopic information reported in Table A.2). For completeness, we also show other transitions that appear blended, but their predicted intensities are nevertheless required to reproduce the observed spectra once the contribution of all species previously identified toward G+0.693 is taken into account. In this case, we fixed the Tex, vLSR and FWHM to the values derived for methyl acetate (i.e., Tex = 15 K, vLSR = 68 km s−1 , FWHM = 21 km s−1 ), leaving the N as the only free parameter in the AUTOFIT. We derived a N = (1.82 ± 0.15) × 1013 cm−2, which yields a fractional abundance with respect to H2 of(1.3±0.2) × 10−10.

Meanwhile, for g-CH3CH2OC(O)H we detected two unblended lines along with seven additional lines showing slight blending. Therefore, the overall detection of this conformer remains tentative (see Figure A.1, spectroscopic information listed in Table A.3). Following the same fitting approach as for the anti conformer, we obtained a N = (1.9 ± 0.2) × 1013 cm−2, which corresponds to a fractional abundance with respect to H2 of (1.4 ± 0.3) × 10−10, and does not produce any overly bright feature for other transitions. Therefore, based on the derived abundance, we infer an anti/gauche ratio of 0.9±0.1, in excellent agreement with the ratio reported previously toward Orion KL and W51 e2 (of ~1; Tercero et al. 2013; Rivilla et al. 2017). Furthermore, the total abundance of CH3CH2OC(O)H (i.e., the sum of the abundance of both a- and g-conformers) is a factor of 5.9±0.6 lower than that of CH3C(O)OCH3 in G+0.693, yielding a CH3C(O)OCH3/CH3CH2OC(O)H abundance ratio, comparable to that observed previously toward Orion KL (of ~5; Tercero et al. 2013).

3.4 Detection of hydroxyacetone

To date, hydroxyacetone (CH3C(O)CH2OH; the fourth most stable structural isomer) has only been detected toward the protostar IRAS 16293-2422B (Zhou et al. 2020). We conducted a new search for the molecule toward G+0.693, yielding the second interstellar detection (first one outside a hot core) and confirming its presence in the ISM.

To carry out the search, we used the spectroscopic entry of the ground vibrational state of CH3C(O)CH2OH included in the Jet Propulsion Laboratory (JPL) database (entry 74003; Pickett et al. 1998) based on high-resolution microwave, millimeter- and submillimeter-wave laboratory data by Braakman et al. (2010). The molecule presents a low barrier CH3 internal rotor, which causes the splitting of the energy levels into the A and E symmetry states. In this case, we detected several nearly complete progressions of R-branch a-type transitions ranging from J = 5 to J = 8 (a total of six fully unblended lines), with Ka = 0,1 and belonging to CH3C(O)CH2OH, along with few additional higher Ka lines (see Figure 4, spectroscopic information reported in Table A.4). To conduct the LTE analysis, we again left the N as the only free parameter and fixed the rest of the parameters to those obtained for methyl acetate (listed in Table 1). We retrieved a N = (2.11 ± 0.13) × 1013 cm−2, which translates into a fractional abundance with respect to H2 of (1.6± 0.2) × 10−10. Based on the derived abundance, CH3C(O)CH2OH is a factor of 10±1 less abundant than CH3C(O)OCH3, and 1.8±0.1 times less abundant than CH3CH2OC(O)H.

thumbnail Fig. 3

Transitions of a-ethyl formate, a-CH3CH2OC(O)H, identified toward G+0.693, which were used to derive the LTE physical parameters of the molecule (see text; listed in Table A.2). The red line shows the best LTE fit of a-CH3CH2OC(O)H, while the blue line represents the combined emission of all molecules identified in the survey, including a-CH3CH2OC(O)H (observed spectra are shown as gray histograms). The quantum numbers for each transition are shown at the bottom of each panel. The transitions are sorted by decreasing intensity. The molecular structure of a-CH3CH2OC(O)H is also shown (C atoms in gray, O atoms in red, and H atoms in white).

3.5 Detection of lactaldehyde

For lactaldehyde (2-hydroxypropanal, CH3CH(OH)C(O)H; ranked as the fifth most stable structural isomer), we incorporated the high-resolution rotational data collected by Alonso et al. (2019) into MADCUBA, which corresponds to entry 074519 from CDMS (Endres et al. 2016). The exquisite sub-mK sensitivity of the present spectral survey toward G+0.693 allowed us to detect three clear and clean emission features that could be unambiguously ascribed to the 50,5−41,4, 60,6—51,5 and 61,6−50,5 transitions of CH3CH(OH)C(O)H (see Figure 5, spectroscopic information reported in Table A.5). These correspond to the brightest Kα = 0,1 R-branch b-type transitions that fall within the ultra-deep Q-band data, all of which appear free of blending and exhibit good integrated signal-to-noise (S/N) ratios (S/N = 12, 16, and 18, respectively; rms = 0.37 mK, 0.34 mK, and 0.43 mK)2. We also detected the adjacent 70,7−61,6 transition, as well as a few additional higher Ka lines that appear blended but reproduce the observed spectra well once the contribution from all other identified species is considered (solid blue line). We note that CH3CH(OH)C(O)H is a near-prolate asymmetric top, exhibiting a dominant b-type spectrum (μa = 1.1 D, μb = 2.2 D, μc = 0.93; computed at the B2PLYPD3-aug-cc-pVTZ level). The fact that A ≫ B, C implies that the strongest R-branch b-type progressions (e.g., J1,J′′→J0,J″″$J'_{1,J'} \rightarrow J''_{0,J''}$, which appear at frequencies close to A + C + 2CJ″) are shifted toward higher frequencies compared to the a-type R-branch lines detected for other isomers, which occur near (B + C)(J″ + 1). Consequently, fewer intense transitions of CH3CH(OH)C(O)H are expected within the Q band, where the sensitivity is the highest (see Rivilla et al. 2023; Sanz-Novo et al. 2023; Sanz-Novo et al. 2025). Accordingly, only a few unblended or slightly blended transitions are detected at higher frequencies (e.g., 3 mm window; see last row of Figure 5), and the rest appear either heavily blended or the derived S/N is significantly lower. Nevertheless, in all cases the LTE model is consistent with the observed spectra, and we note that no missing lines of CH3CH(OH)C(O)H are detected across the entire survey.

Following the same fitting approach as described above for other isomers, the AUTOFIT analysis yielded a N = (1.10 ± 0.13) × 1013 cm−2, which translates into an abundance of (8.1 ± 1.6) ×10−11. Accordingly, we find that lactaldehyde is 20±3, 3.4±0.4, and 1.9±0.2 times less abundant than methyl acetate, ethyl formate, and hydroxyacetone, respectively, and is only slightly more abundant (a factor of 1.3±0.2) than its positional isomer 3-hydroxypropanal in G+0.693 (Fried et al. 2025).

thumbnail Fig. 4

Transitions of hydroxyacetone, CH3C(O)CH2OH, detected toward G+0.693 (see text; listed in Table A.4). The red line shows the best LTE fit of CH3C(O)CH2OH, while the blue line represents the combined emission of all molecules identified in the survey, including CH3C(O)CH2OH (observed spectra are shown as gray histograms). The quantum numbers for each transition are shown at the bottom of each panel (A-state lines unless stated otherwise). The transitions are sorted by increasing frequency. The molecular structure of CH3C(O)CH2OH is also shown (C atoms in gray, O atoms in red, and H atoms in white).

3.6 Detection of methoxyacetaldehyde

Methoxyacetaldehyde (CH3OCH2C(O)H) has previously been searched for, without success, toward Orion KL and Sgr B2, as well as toward the dark cloud Barnard 1 (Kolesniková et al. 2018). To perform the astronomical search, we implemented the spectroscopic entry into MADCUBA (CDMS entry 074517 Endres et al. 2016, based on Kolesniková et al. 2018). Although CH3OCH2C(O)H also presents methyl internal rotation, its V3 barrier is higher than that of methyl acetate, and the A/E splittings were not resolved at 3mm with a conventional absorption spectrometer (Kolesniková et al. 2018) or with the resolution of the present survey (Rivilla et al. 2023; Sanz-Novo et al. 2023).

An inspection of the data of G+0.693 revealed a clear pattern of R-branch a-type transitions belonging to CH3OCH2C(O)H (see Figure 6). Among these, the 81,8−71,7, 91,9−81,8 and 90,9−80,8 transitions are fully unblended and show a good integrated S/N ratio (S/N = 12, 14, and 14, respectively; rms = 0.44 mK, 0.48 mK, and 0.48 mK). We also identified the 80,8−70,7, 92,7−82,6, and 101,10−91,9 transitions, which are slightly contaminated with 13CH2CHCN and two U-lines, respectively, as well as a few additional Ka = 0, 1, and 2 transitions that appear more blended. Nevertheless, all these transitions help to reproduce the observations once the emission from all other molecules previously identified toward G+0.693 are taken into account, further reinforcing the detection. The spectroscopic information of all these transitions is reported in Table A.6. We then used the aforementioned unblended and slightly blended transitions of CH3OCH2C(O)H to carry out the LTE analysis. We retrieved a N = (3.3 ± 0.3) × 1012 cm−2 with the AUTOFIT tool. This corresponds to a fractional abundance with respect to H2 of (2.4 ± 0.4) ×10−11. Consequently, CH3OCH2C(O)H shows the lowest abundance among all C3H6O2 isomers, that is, a factor of 67±9, 11±1, 6.4±0.7, 3.3±0.4, and 2.6±0.5 less abundant than methyl acetate, ethyl formate, hydroxyacetone, lactaldehyde, and 3-hydroxypropanal, respectively, in G+0.693.

thumbnail Fig. 5

Transitions of lactaldehyde, CH3CH(OH)C(O)H, identified toward G+0.693 (see text; listed in Table A.5). The red line shows the best LTE fit of CH3CH(OH)C(O)H, while the blue line represents the combined emission of all molecules identified in the survey, including CH3CH(OH)C(O)H (observed spectra are shown as gray histograms). The quantum numbers for each transition are shown at the bottom of each panel. For panels with multiple lines, we show the quantum number of the lowest-frequency transition. The transitions are sorted by increasing frequency. The molecular structure of CH3CH(OH)C(O)H is also shown (C atoms in gray, O atoms in red, and H atoms in white).

3.7 Non-detection of propionic acid and glycidol

Propionic acid, CH3CH2C(O)OH, besides being the most thermodynamically favored C3H6O2 isomer, also exhibits the lowest total dipole moment within this molecular family (i.e., μα = 0.2 D, μb = 1.6 D; computed at the B2PLYPD3-aug-cc-pVTZ level). It has previously been searched for toward Orion KL and Sgr B2(N) (Ilyushin et al. 2021); however, the molecule remains undetected. Here, we searched for CH3CH2C(O)OH toward G+0.693 using a catalog based on the spectroscopy reported in Jaman et al. (2015), but we did not achieve a detection. Although many transitions still leave room for the possible presence of the molecule, the majority of the brightest transitions are affected by significant line blending (see Figure A.2). To derive the upper limits for its column density, we fixed the Tex, vLSR, and FWHM to that of methyl acetate (listed in Table 1) and increased N until the LTE model (including also the contribution from other species) visually matched the observed spectra for the brightest Ka = 1 transition, 81,8−70,7 (at 49.3424 GHz). We estimated a N ≤ 2 × 1013 cm−2, which yields a fractional abundance with respect to H2 of ≤ 1.5 × 10−10.

We also targeted the cyclic species glycidol (c-CH2OCHCH2OH), which, to our knowledge, has not been searched for in the ISM to date. We used a newly generated catalog in SPFIT/SPCAT format, based on the spectroscopic parameters reported by Demaison et al. (2012), and implemented it into MADCUBA. After analyzing the data, we did not detect any bright, unblended transition of c-CH2OCHCH2OH (see Figure A.3). We note that this species presents a relatively low total dipole moment (i.e., the second lowest after CH3CH2C(O)OH; μa = 0.7 D, μb = 1.3 D, μc = 0.8 D, computed at the B2PLYPD3-aug-cc-pVTZ level), which will hamper its detection. To derive the line-integrated 3σ upper limit to its column density (σ being the rms noise of the spectra), we used the 55,1 −44,0 and 55,0−44,1 transitions (fully coalesced and located at ~97.0723 GHz), which is the brightest feature predicted for c-CH2OCHCH2OH, only negligibly blended with CH3(O)OCH3. We obtained a N ≤ 0.5 × 1013 cm−2, which translates into an abundance upper limit of ≤3.7×10−11.

4 Astrochemical implications and conclusions

The detection of six distinct C3H6O2 structural isomers toward G+0.693 (i.e., methyl acetate, hydroxyacetone, ethyl formate, 3-hydroxypropanal, lactaldehyde, and methoxyacetaldehyde), together with the non-detection of propionic acid and glycidol, allows us to delve further into the formation of these species in the ISM.

Here, we focus on possible grain-surface chemical formation pathways, which have previously been demonstrated to produce multiple C3H6O2 isomers upon UV irradiation of methanol-acetaldehyde and CO-ethanol ices (Wang et al. 2023; Wang et al. 2024). As shown in Figure 7, all the proposed routes proceed through a final radical-radical recombination reaction starting from the parent molecule CO, which can either be hydrogenated to form methanol (CH3OH; Jiménez-Serra et al. 2025) or oxygenated to yield HOCO and, eventually, CO2 (Garrod & Pauly 2011; Minissale et al. 2013; Molpeceres et al. 2025). In the former case, CH3 OH can undergo H-abstraction to produce the CH2OH and CH3O radicals, which may subsequently react with CH3 CO and CH2C(O)H to form hydroxyacetone, 3-hydroxypropanal, methyl acetate, and methoxyacetaldehyde, although to our knowledge the latter has not been detected in previous experiments (Wang et al. 2023; Wang et al. 2024). Alternatively, CH3OH can evolve into CH3CH2OH, which, upon H-abstraction, yields CH3CH2O, CH2CH2OH, or CH3CHOH, depending on the abstraction site. These radicals can then react with the H(O)C radical, thought to form efficiently via hydrogenation of CO (Brown et al. 1988) or H-abstraction from formaldehyde (H2 (O)C), leading to the formation of ethyl formate, 3-hydroxypropanal, and lactaldehyde, respectively. Remarkably, lactaldehyde has been proposed as a key building block for relevant prebiotic molecules, such as pyruvic acid, methylglyoxal, and lactic acid, which participate in the methylglyoxal cycle and are considered as precursors of amino acids, sugars, and sugar acids (Wang et al. 2024).

While the formation routes described above could indeed account for the formation of the targeted C3H6O2 isomers in the ISM, their efficiency depends critically on two ingredients: (i) a high cosmic-ray ionization rate (CRIR) to promote radical-radical recombination reactions, and (ii) access to the chemistry that occurs on the surface of dust grains, enabled by the desorption of ice-mantle species into the gas phase. This aligns perfectly with two of the key factors driving the exceptionally rich chemistry observed in G+0.693. First, large-scale, low-velocity shocks, likely resulting from a cloud-cloud collision scenario (Requena-Torres et al. 2006; Zeng et al. 2020; Colzi et al. 2024), enhance the sputtering of icy grain mantles, thereby exposing surface chemistry. Second, the high CRIR that is thought to affect the cloud (10−14−10−15 s−1 ), inferred from chemical modeling of the PO+ and HOCS+ cations (Rivilla et al. 2022b; Sanz-Novo et al. 2024), has been suggested to facilitate the formation of radicals on these icy mantles. Once generated, these radicals can efficiently interact and recombine, ultimately driving the synthesis of increasingly complex molecules (Rivilla et al. 2023).

On a different note, the non-detection of propionic acid and glycidol provides additional valuable information. These species exhibit the lowest dipole moments among all isomers (see Figure 1), which significantly hampers their astronomical detection. Interestingly, while most C3 H6O2 isomers can be formed through the hydrogenation of CO (Figure 7), propionic acid follows a distinct chemical route. Its formation begins with the oxygenation of CO via the reaction CO + OH, producing the HOCO radical, which is widely considered a key precursor in the formation of interstellar carboxylic acids (see, e.g., Oba et al. 2010; Ioppolo et al. 2021; Ishibashi et al. 2024; Molpeceres et al. 2025). This species can either be converted into CO or CO2 (Molpeceres et al. 2025) or, alternatively, it could lead to the formation of more complex acids such as propionic acid. However, this latter pathway might represent only a minor channel, given the limited reservoir of complex carboxylic acids currently known in the ISM (Sanz-Novo et al. 2023). Furthermore, the chemical distinction proposed here between propionic acid and the other C3H6O2 isomers is consistent with the spatial differentiation observed in ALMA interferometric observations toward Orion KL by Tercero et al. (2018), which is clearly differentiated for methyl acetate and ethyl formate compared to that of the carboxylic acids (HCOOH and CH3COOH).

At this point, we can compare the upper limit to the column density derived for propionic acid with the abundance of the related trans-formic acid (t-HCOOH; N = (2.0 ± 0.4) × 1014 cm−2) and acetic acid (CH3COOH; N = (4.5 ± 0.2) × 1013 cm−2), both of which are also detected toward G+0.693 (Sanz-Novo et al. 2023). As has been observed for other chemical families such as alcohols, thiols, aldehydes, and isocyanates (Rodríguez-Almeida et al. 2021a,b; Jiménez-Serra et al. 2022; Sanz-Novo et al. 2022; Fried et al. 2025), the abundances typically decrease by roughly one order of magnitude with increasing molecular complexity (i.e., the substitution of a -CH3 group). In this case, the observed ratios are t-HCOOH/CH3COOH = 4.4 ± 0.9 and CH3COOH/CH3CH2COOH ≥2.0, which suggests that we may still be below the sensitivity level required to detect CH3CH2COOH.

Regarding glycidol, while the proposed formation route also begins with CH3OH, it requires an additional precursor for the suggested radical-radical recombination reaction: c-CH2OCH. This radical could be produced via H-abstraction of ethylene oxide (c-CH2OCH2). However, the relatively low abundance of c-CH2OCH2 observed toward G+0.693 (N ~ 1014 cm−2; fractional abundance of ~7 × 10−10; Requena-Torres et al. 2008) compared with the other precursors such as CH3OH and CH3CH2OH, which are ~2 and ~1 orders of magnitude more abundant than c-CH2OCH2, respectively (i.e., 1.1 ±0.2 × 10−7 and 4.6±0.6 × 10−9; Rodríguez-Almeida et al. 2021a), may hamper the detection of glycidol in the present survey. Additionally, we highlight that neither glycidol nor propionic acid were found in the experiments by Wang et al. (2023); Wang et al. (2024).

Finally, the comparison of the abundances of the detected C3H6O2 isomers (Table 1) leads to the following ordering (see Figure 8):

N (methyl acetate) > N (ethyl formate) > N(hydroxyacetone) > N(lactaldehyde) > N(3-hydroxypropanal) > N(methoxyacetaldehyde).

Methyl acetate clearly emerges as the dominant isomer within the C3H6O2 family, followed by ethyl formate and hydroxyacetone. Surprisingly, the observed abundance trend qualitatively follows the relative thermodynamic stability of the detected isomers (Figure 8). However, a major exception is propionic acid, the most stable member of the family, which remains undetected. This discrepancy shows that thermodynamic stability alone cannot be used as a reliable tool to predict the molecular abundances within this isomeric family in the ISM, as reported for other families (e.g., C3H2O, C2H4O2, C2H2N2, and C2H5NO2; Loomis et al. 2015; Mininni et al. 2020; Shingledecker et al. 2020; Rivilla et al. 2023; San Andrés et al. 2024). Instead, kinetics are likely responsible for distinguishing among species arising from separated chemical formation pathways (i.e., those originating from CH3OH or from HOCO; see Figure 7).

On the other hand, the significantly higher abundance of methyl acetate may suggest that the formation of this species is more efficient, potentially due to additional formation pathways occurring either on dust grains or in the gas phase (e.g., Das et al. 2015). Alternatively, if the relative dipole principle (RDP; Shingledecker et al. 2020) applies to this isomeric family, then the lower dipole moment of methyl acetate (μtot = 1.9 D, see Table 1) would lead to a slower destruction in the gas phase compared to the other detected species. Nevertheless, further theoretical, laboratory, and modeling effort is needed to shed conclusive light on this question.

In summary, our systematic survey of all C3H6O2 isomers toward G+0.693 has significantly expanded the census of complex organic molecules in the ISM, particularly within this isomeric family. We report the interstellar discovery of lactaldehyde and methoxyacetaldehyde, along with the confirmation of methyl acetate and hydroxyacetone, and new detections in G+0.693 of both anti - and g auche-conformers of ethyl formate. These results not only enhance our understanding of the molecular inventory in the ISM, but also emphasize the importance of highly sensitive spectral line surveys for detecting the less-abundant species. Furthermore, they provide a rationale for the more than decade-long gap between the detection of methyl acetate and ethyl formate and the new interstellar discoveries reported in this work. These results also underscore the need for future laboratory and observational efforts targeting other uncharacterized C3H6O2 isomers, including various enol tautomers (e.g., prop-1-ene-1,2-diol, prop-2-ene-1,2-diol, 1-methoxyethen-1-ol, prop-1-ene-1,3-diol) proposed by Wang et al. (2023), which may be efficiently produced via keto-enol tautomerization. Based on our findings, these species also appear as promising candidates for future interstellar detection.

thumbnail Fig. 6

Transitions of methoxyacetaldehyde, CH3OCH2C(O)H, detected toward G+0.693 (collected in Table A.6). The red line shows the best LTE fit of CH3OCH2C(O)H, while the blue line represents the combined emission of all molecules identified in the survey, including CH3OCH2C(O)H (observed spectra are shown as gray histograms). The quantum numbers for each transition are shown at the bottom of each panel. The transitions are sorted by increasing frequency. The molecular structure of CH3OCH2C(O)H is also shown (C atoms in gray, O atoms in red, and H atoms in white).
thumbnail Fig. 7

Suggested grain-surface chemical routes for the formation of targeted C3H6O2 isomers in the ISM. Dashed lines indicated processes involving more than one step. We highlight in boldface molecules identified toward G+0.693. C3H6O2 species that were not detected and whose formation has not been tested experimentally by Wang et al. (2023) and Wang et al. (2024) are highlighted with a light purple box. For simplicity, stereoisomerism is not considered in this plot.
thumbnail Fig. 8

Fractional abundances with respect to H2 of different C3H6O2 isomers sorted by increasing relative energy. The total dipole moment for each molecule (computed at the B2PLYP-D3/aug-cc-pVTZ level) is included. We used the same color-code as that in Fig. 1: new first detections in the ISM are shown in red, tentative detections in orange, non-detections in purple (upper limits indicated with black arrows), and previous detections toward different sources in black.

Acknowledgements

We are grateful to the IRAM 30 m and Yebes 40 m telescopes staff for their help during the different observing runs, highlighting project 21A014 (PI: Rivilla), project 018-19 (PI: Rivilla) and projects 123-22 and 076-23 (PI: Jiménez-Serra). The 40 m radio telescope at Yebes Observatory is operated by the Spanish Geographic Institute (IGN, Ministerio de Transportes, Movilidad y Agenda Urbana). IRAM is supported by INSU/CNRS (France), MPG (Germany) and IGN (Spain). M.S.-N. acknowledges a Juan de la Cierva Postdoctoral Fellow proyect JDC2022-048934-I, funded by the Spanish Ministry of Science, Innovation and Universities/State Agency of Research MICIU/AEI/10.13039/501100011033 and by the European Union “NextGen-erationEU”/PRTR”. V.M.R. acknowledges support from the grant RYC2020-029387-I funded by MICIU/AEI/10.13039/501100011033 and by “ESF, Investing in your future”. V.M.R., A.L-G. and D.S.A from the Consejo Superior de Investigaciones Científicas (CSIC) and the Centro de Astrobiología (CAB) through the project 20225AT015 (Proyectos intramurales especiales del CSIC), and from the grant CNS2023-144464 funded by MICIU/AEI/10.13039/501100011033 and by “European Union NextGenerationEU/PRTR”. I.J.-S., V.M.R., M.S.-N, L.C., A.M., D.S.A, A.L-G. and A.M.-H. acknowledge funding from grant No. PID2022-136814NB-I00 from MICIU/AEI/10.13039/501100011033 and by “ERDF, UE”. I.J.-S. and A.M. also acknowledge funding from the ERC grant OPENS (project number 101125858) funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the European Research Council Executive Agency. Neither the European Union nor the granting authority can be held responsible for them. D.S.A. also extends his gratitude for the financial support provided by the Comunidad de Madrid through the Grant PIPF-2022/TEC-25475. A.M.-H. acknowledges funding from grant MDM-2017-0737 Unidad de Excelencia “María de Maeztu” Centro de Astrobiología (CAB, CSIC-INTA) funded by MICIU/AEI/10.13039/501100011033. B.A.M. and Z.T.P.F. acknowledge support from Schmidt Family Futures. L.K. acknowledges financial funding from the Czech Science Foundation (GACR, grant No. 24-12586S). B.T. acknowledges Spanish Ministry of Science support from grants PID2022-137980NB-100 and PID2023-147545NB-I00. J.-C.G. thanks the Centre National d’Etudes Spatiales (CNES) for a grant and the Programme National “Physique et Chimie du Milieu Interstellaire” (PCMI) of CNRS/INSU with INC/INP cofunded by CEA and CNES. E.J.C. acknowledges financial support from the Spanish Ministry of Science and Innovation (MCIN/AEI, Project PID2023-147698NB-I00) and the Basque Government (Project IT1491-22) I.K. would like to thank the Agence Nationale de la recherche Scientifique (ANR) project ANR-25-CE29-7779.

References

  1. Alonso, E. R., McGuire, B. A., Kolesniková, L., et al. 2019, ApJ, 883, 18 [NASA ADS] [CrossRef] [Google Scholar]
  2. Belloche, A., Garrod, R. T., Müller, H. S. P., et al. 2009, A&A, 499, 215 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  3. Belloche, A., Garrod, R. T., Müller, H. S. P., et al. 2025, A&A, 698, A143 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  4. Braakman, R., Drouin, B. J., Widicus Weaver, S. L., & Blake, G. A. 2010, J. Mol. Spectrosc., 264, 43 [Google Scholar]
  5. Brown, P. D., Charnley, S. B., & Millar, T. J. 1988, MNRAS, 231, 409 [Google Scholar]
  6. Carvajal, M., Favre, C., Kleiner, I., et al. 2019, A&A, 627, A65 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  7. Colzi, L., Martín-Pintado, J., Zeng, S., et al. 2024, A&A, 690, A121 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  8. Das, A., Majumdar, L., Sahu, D., et al. 2015, ApJ, 808, 21 [NASA ADS] [CrossRef] [Google Scholar]
  9. Demaison, J., Boucher, D., BURIE, J., & Dubrulle, A. 1984, Z. Naturf. A, 39, 560 [Google Scholar]
  10. Demaison, J., Craig, N. C., Conrad, A. R., Tubergen, M. J., & Rudolph, H. D. 2012, J. Phys. Chem. A, 116, 9116 [Google Scholar]
  11. Endres, C. P., Schlemmer, S., Schilke, P., Stutzki, J., & Müller, H. S. P. 2016, J. Mol. Spectrosc., 327, 95 [NASA ADS] [CrossRef] [Google Scholar]
  12. Fried, Z. T. P., Motiyenko, R. A., Sanz-Novo, M., et al. 2025, ApJ, 992, 187 [Google Scholar]
  13. Frisch, M. J., Trucks, G. W., Schlegel, H. B., et al. 2016, Gaussian~16 Revision C.01 (Wallingford, CT: Gaussian Inc.) [Google Scholar]
  14. García de la Concepción, J., Jiménez-Serra, I., Corchado, J. C., et al. 2023, A&A, 675, A109 [Google Scholar]
  15. Garrod, R. T., & Pauly, T. 2011, ApJ, 735, 15 [Google Scholar]
  16. Grimme, S., Ehrlich, S., & Goerigk, L. 2011, J. Computat. Chem., 32, 1456 [Google Scholar]
  17. Herbst, E., Gianfranco, V., & Ceccarelli, C. 2020, ACS Earth Space Chem., 4, 488 [Google Scholar]
  18. Ilyushin, V. V., Margulès, L., Tercero, B., et al. 2021, J. Mol. Spectrosc., 379, 111454 [Google Scholar]
  19. Ioppolo, S., Kanuchová, Z., James, R. L., et al. 2021, A&A, 646, A172 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  20. Ishibashi, A., Molpeceres, G., Hidaka, H., et al. 2024, ApJ, 976, 162 [Google Scholar]
  21. Jaman, A. I., Chakraborty, S., & Chakraborty, R. 2015, J. Mol. Struct., 1079, 402 [Google Scholar]
  22. Jiménez-Serra, I., Martín-Pintado, J., Rivilla, V. M., et al. 2020, Astrobiology, 20, 1048 [Google Scholar]
  23. Jiménez-Serra, I., Rodríguez-Almeida, L. F., Martín-Pintado, J., et al. 2022, A&A, 663, A181 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  24. Jiménez-Serra, I., Megías, A., Salaris, J., et al. 2025, A&A, 695, A247 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  25. Jones, P. A., Burton, M. G., Cunningham, M. R., et al. 2012, MNRAS, 419, 2961 [Google Scholar]
  26. Kolesniková, L., Peña, I., Alonso, E. R., et al. 2018, A&A, 619, A67 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  27. Li, J., Wang, J., Qiao, H., et al. 2020, MNRAS, 492, 556 [Google Scholar]
  28. Loomis, R. A., McGuire, B. A., Shingledecker, C., et al. 2015, ApJ, 799, 34 [Google Scholar]
  29. Marcelino, N., Tercero, B., Agúndez, M., & Cernicharo, J. 2021, A&A, 646, L9 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  30. Martín, S., Requena-Torres, M. A., Martin-Pintado, J., & Mauersberger, R. 2008, ApJ, 678, 245 [CrossRef] [Google Scholar]
  31. Martin, S., Martin-Pintado, J., Blanco-Sánchez, C., et al. 2019, A&A, 631, A159 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  32. Medvedev, I. R., De Lucia, F. C., & Herbst, E. 2009, ApJS, 181, 433 [Google Scholar]
  33. Mininni, C., Beltrán, M. T., Rivilla, V. M., et al. 2020, A&A, 644, A84 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  34. Minissale, M., Congiu, E., Manicò, G., Pirronello, V., & Dulieu, F. 2013, A&A, 559, A49 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  35. Molpeceres, G., Enrique-Romero, J., Ishibashi, A., et al. 2025, MNRAS, 538, 1565 [Google Scholar]
  36. Neill, J. L., Muckle, M. T., Zaleski, D. P., et al. 2012, ApJ, 755, 153 [Google Scholar]
  37. Nguyen, H. V. L., Kleiner, I., Shipman, S. T., et al. 2014, J. Mol. Spectrosc., 299, 17 [Google Scholar]
  38. Oba, Y., Watanabe, N., Kouchi, A., Hama, T., & Pirronello, V. 2010, ApJ, 722, 1598 [CrossRef] [Google Scholar]
  39. Peng, Y., Rivilla, V. M., Zhang, L., Ge, J. X., & Zhou, B. 2019, ApJ, 871, 251 [NASA ADS] [CrossRef] [Google Scholar]
  40. Pickett, H. M. 1991, J. Mol. Spectrosc., 148, 371 [Google Scholar]
  41. Pickett, H. M., Poynter, R. L., Cohen, E. A., et al. 1998, J. Quant. Spec. Radiat. Transf., 60, 883 [Google Scholar]
  42. Remijan, A. J., Changala, P. B., Xue, C., et al. 2025, ApJ, 982, 191 [Google Scholar]
  43. Requena-Torres, M. A., Martín-Pintado, J., Rodríguez-Franco, A., et al. 2006, A&A, 455, 971 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  44. Requena-Torres, M. A., Martín-Pintado, J., Martín, S., & Morris, M. R. 2008, ApJ, 672, 352 [Google Scholar]
  45. Riveros, J. M. 1967, J. Chem. Phys., 46, 4605 [Google Scholar]
  46. Rivilla, V. M., Beltrán, M. T., Martin-Pintado, J., et al. 2017, A&A, 599, A26 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  47. Rivilla, V. M., Colzi, L., Jiménez-Serra, I., et al. 2022a, ApJ, 929, L11 [NASA ADS] [CrossRef] [Google Scholar]
  48. Rivilla, V. M., Jiménez-Serra, I., Martin-Pintado, J., et al. 2022b, Front. Astron. Space Sci., 9, 829288 [NASA ADS] [CrossRef] [Google Scholar]
  49. Rivilla, V. M., Sanz-Novo, M., Jiménez-Serra, I., et al. 2023, ApJ, 953, L20 [NASA ADS] [CrossRef] [Google Scholar]
  50. Rodriguez-Almeida, L. F., Jiménez-Serra, I., Rivilla, V. M., et al. 2021a, ApJ, 912, L11 [CrossRef] [Google Scholar]
  51. Rodriguez-Almeida, L. F., Rivilla, V. M., Jiménez-Serra, I., et al. 2021b, A&A, 654, L1 [CrossRef] [EDP Sciences] [Google Scholar]
  52. San Andrés, D., Rivilla, V. M., Colzi, L., et al. 2024, ApJ, 967, 39 [CrossRef] [Google Scholar]
  53. Sanz-Novo, M., Belloche, A., Rivilla, V. M., et al. 2022, A&A, 666, A114 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  54. Sanz-Novo, M., Rivilla, V. M., Jiménez-Serra, I., et al. 2023, ApJ, 954, 3 [NASA ADS] [CrossRef] [Google Scholar]
  55. Sanz-Novo, M., Rivilla, V. M., Jiménez-Serra, I., et al. 2024, ApJ, 965, 149 [Google Scholar]
  56. Sanz-Novo, M., Molpeceres, G., Rivilla, V. M., & Jimenez-Serra, I. 2025, A&A, 698, A36 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  57. Sanz-Novo, M., Rivilla, V. M., Endres, C. P., et al. 2025, ApJ, 980, L37 [Google Scholar]
  58. Savarese, A., Alessandrini, S., Melosso, M., et al. 2025, ACS Earth Space Chem., 10, 198 [Google Scholar]
  59. Shingledecker, C. N., Álvarez-Barcia, S., Korn, V. H., & Kästner, J. 2019, ApJ, 878, 80 [Google Scholar]
  60. Shingledecker, C. N., Lamberts, T., Laas, J. C., et al. 2020, ApJ, 888, 52 [Google Scholar]
  61. Tercero, B., Cuadrado, S., López, A., et al. 2018, A&A, 620, L6 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  62. Tercero, B., Kleiner, I., Cernicharo, J., et al. 2013, ApJ, 770, L13 [NASA ADS] [CrossRef] [Google Scholar]
  63. Tercero, F., López-Pérez, J. A., Gallego, J. D., et al. 2021, A&A, 645, A37 [EDP Sciences] [Google Scholar]
  64. Wang, J., Marks, J. H., Turner, A. M., et al. 2023, Phys. Chem. Chem. Phys., 25, 936 [Google Scholar]
  65. Wang, J., Zhang, C., Marks, J. H., et al. 2024, Nat. Commun., 15, 10189 [Google Scholar]
  66. Zeng, S., Jiménez-Serra, I., Rivilla, V. M., et al. 2018, MNRAS, 478, 2962 [Google Scholar]
  67. Zeng, S., Zhang, Q., Jiménez-Serra, I., et al. 2020, MNRAS, 497, 4896 [NASA ADS] [CrossRef] [Google Scholar]
  68. Zheng, S., Li, J., Wang, J., et al. 2024, ApJ, 961, 58 [NASA ADS] [CrossRef] [Google Scholar]
  69. Zhou, Y., Quan, D.-H., Zhang, X., & Qin, S.-L. 2020, Res. Astron. Astrophys., 20, 125 [CrossRef] [Google Scholar]

Appendix A Complementary tables and figures

In Tables A.1, A.2, A.3, A.4, A.5 and A.6 we provide the spectroscopic information of the transitions detected for methyl acetate, a-ethyl formate, g-ethyl formate, hydroxyacetone, lactaldehyde and methoxyacetaldehyde, respectively, which are shown in Figures 2, 3, A.1, 4, 5 and 6.

Regarding the preparation of the line catalog of methyl acetate in SPCAT format from the BELGI output, we also provide in Table A.1 the values of the line strength (Sgμg2$S_g \mu_g^2$; Tercero et al. 2013), which were used to compute the log I(nm2 MHz) at 300 K via: I(T)=4.16231⋅10−5⋅ν⋅gI⋅Sg⋅μg2⋅exp⁡(−E″/kT)−exp⁡(−E′/kT)Qr(T)I(T) = 4.16231 \cdot 10^{-5} \cdot \nu \cdot g_I \cdot S_g \cdot \mu_g^2 \cdot \frac{\exp(-E''/kT) - \exp(-E'/kT)}{Q_{\text{r}}(T)}

with ν in MHz and μ in D.

We note that the original laboratory data covered up to J = 19 and Ka ≤ 7, although predictions were initially computed up to J = 30 (Tercero et al. 2013). In this work, we adopt the rotational partition function values computed from first principles, derived from the rotational constants reported in Nguyen et al. (2014). Specifically, we obtained a Qr (AA, 300 K) = 72174.2179 (without considering spin weights and all states), which is approximately 15% larger than the value used in Tercero et al. (2013) of Qr (AA, 300 K) = 61648.6 (I. Kleiner, priv. communication). The latter was obtained by applying a correction factor of 2.6 to the Qr (AA) computed through direct summation of the ground-state energy levels up to J = 30, which is needed to consider energy levels up to J = 65. Nevertheless, it may still underestimate the real partition function value if full convergence has not been achieved (Carvajal et al. 2019).

In contrast to the previous work, where each substate was treated as a separate species (Tercero et al. 2013), we have included all five substates in the same catalog. Consequently, the Eup (K) reported in Table A.1 accounts for the ZPE of the AA species (99.9450 cm−1 ; Tercero et al. 2013). To consider all of the states and their weights, Qr(Tot, 300 K) = Qr (AA, 300 K) × f, where f = (16+16+16+8+8) = 64.

In Figure A.1 we show the tentative detection of g-ethyl formate, g-CH3CH2OC(O)H, toward G+0.693, while in Figures A.2 and A.3 we present the LTE simulation of the propionic acid and glycidol emission, respectively, using the upper limits to the column density derived toward G+0.693.

thumbnail Fig. A.1

Tentative detection of g-ethyl formate, g-CH3CH2OC(O)H, toward G+0.693. The selected transitions are sorted by decreasing intensity and listed in Table A.3. The result of the LTE model of the g-CH3CH2OC(O)H emission is shown in orange, together with the expected molecular emission from all the molecular species identified to date in our survey, including g-CH3CH2OC(O)H (in blue), both overlaid on the observations (gray histogram). The molecular structure of g-CH3CH2OC(O)H is also shown (C atoms in gray, O atoms in red and H atoms in white).
thumbnail Fig. A.2

LTE simulation of the propionic acid, CH3CH2C(O)OH, emission using the upper limit column density derived toward G+0.693 (in purple) together with the expected molecular emission from all the molecular species identified to date in our survey (in blue), both overlaid on the observations (gray histogram). The quantum numbers for each transition are shown at the bottom of each panel.
thumbnail Fig. A.3

LTE simulation of the glycidol, (c-CH2OCHCH2OH, emission using the upper limit column density derived toward G+0.693 (in purple) together with the expected molecular emission from all the molecular species identified to date in our survey (in blue), both overlaid on the observations (gray histogram). The quantum numbers for each transition are shown at the bottom of each panel.
Table A.1

Spectroscopic information of the selected transitions of methyl acetate detected toward G+0.693, shown in Figure 2.

Table A.2

Spectroscopic information of the transitions of a-ethyl formate detected toward G+0.693-0.027, shown in Figure 3.

Table A.3

Spectroscopic information of the transitions of g-ethyl formate detected toward G+0.693-0.027, shown in Figure A.1.

Table A.4

Spectroscopic information of the transitions of hydroxyacetone detected toward G+0.693-0.027, shown in Figure 4.

Table A.5

Spectroscopic information of the transitions of lactaldehyde detected toward G+0.693-0.027, shown in Figure 5.

Table A.6

Spectroscopic information of the transitions of methoxyacetaldehyde detected toward G+0.693-0.027, shown in Figure 6.

All Tables

Table 1

Derived physical parameters for C3H6O2 isomers that we searched for toward G+0.693-0.027.

In the text
Table A.1

Spectroscopic information of the selected transitions of methyl acetate detected toward G+0.693, shown in Figure 2.

In the text
Table A.2

Spectroscopic information of the transitions of a-ethyl formate detected toward G+0.693-0.027, shown in Figure 3.

In the text
Table A.3

Spectroscopic information of the transitions of g-ethyl formate detected toward G+0.693-0.027, shown in Figure A.1.

In the text
Table A.4

Spectroscopic information of the transitions of hydroxyacetone detected toward G+0.693-0.027, shown in Figure 4.

In the text
Table A.5

Spectroscopic information of the transitions of lactaldehyde detected toward G+0.693-0.027, shown in Figure 5.

In the text
Table A.6

Spectroscopic information of the transitions of methoxyacetaldehyde detected toward G+0.693-0.027, shown in Figure 6.

In the text

All Figures

thumbnail Fig. 1

Relative energies (including zero-point energy, ZPE, corrections) plotted as a function of the total dipole moment for the targeted C3H6O2 isomers, computed at the B2PLYPD3∕aug-cc-pVTZ level of theory (Grimme et al. 2011) using the Gaussian 16 program package (Frisch et al. 2016). Optimized 3D structures at the same level of theory are also shown (carbon atoms in gray, oxygen atoms in red, and hydrogen atoms in white). The status of detections in G+0.693 is indicated using the following color-code in the molecule name: new first detections in the ISM are shown in red, tentative detections in orange, non-detections in purple, and previous detections toward different sources in black. Previous detections toward other sources of methyl acetate (Tercero et al. 2013), anti and gauche ethyl formate (Belloche et al. 2009; Tercero et al. 2013; Rivilla et al. 2017; Peng et al. 2019), hydroxyacetone (Zhou et al. 2020), and 3-hydroxypropanal are also included. Also, new detections based in this work are highlighted in boldface. The 3D representations of all molecules are visualized with IQmol (https://www.iqmol.org/).
In the text
thumbnail Fig. 2

Selected unblended or slightly blended transitions of methyl acetate, CH3C(O)OCH3, identified toward G+0.693, which were used to derive the LTE physical parameters of the molecule (see text; listed in Table A.1). The red line shows the best LTE fit of CH3C(O)OCH3, while the blue line represents the combined emission of all molecules identified in the survey (observed spectra are shown as gray histograms). The symmetry state of the detected transitions is shown at the bottom of each panel (full list of quantum numbers listed in Table A.1). The * symbol indicates that there are various transitions in the same panel, each comprising the same E3/E4/E1 symmetry states. The transitions are sorted by decreasing intensity. The molecular structure of CH3C(O)OCH3 is also shown (C atoms in gray, O atoms in red, and H atoms in white).
In the text
thumbnail Fig. 3

Transitions of a-ethyl formate, a-CH3CH2OC(O)H, identified toward G+0.693, which were used to derive the LTE physical parameters of the molecule (see text; listed in Table A.2). The red line shows the best LTE fit of a-CH3CH2OC(O)H, while the blue line represents the combined emission of all molecules identified in the survey, including a-CH3CH2OC(O)H (observed spectra are shown as gray histograms). The quantum numbers for each transition are shown at the bottom of each panel. The transitions are sorted by decreasing intensity. The molecular structure of a-CH3CH2OC(O)H is also shown (C atoms in gray, O atoms in red, and H atoms in white).
In the text
thumbnail Fig. 4

Transitions of hydroxyacetone, CH3C(O)CH2OH, detected toward G+0.693 (see text; listed in Table A.4). The red line shows the best LTE fit of CH3C(O)CH2OH, while the blue line represents the combined emission of all molecules identified in the survey, including CH3C(O)CH2OH (observed spectra are shown as gray histograms). The quantum numbers for each transition are shown at the bottom of each panel (A-state lines unless stated otherwise). The transitions are sorted by increasing frequency. The molecular structure of CH3C(O)CH2OH is also shown (C atoms in gray, O atoms in red, and H atoms in white).
In the text
thumbnail Fig. 5

Transitions of lactaldehyde, CH3CH(OH)C(O)H, identified toward G+0.693 (see text; listed in Table A.5). The red line shows the best LTE fit of CH3CH(OH)C(O)H, while the blue line represents the combined emission of all molecules identified in the survey, including CH3CH(OH)C(O)H (observed spectra are shown as gray histograms). The quantum numbers for each transition are shown at the bottom of each panel. For panels with multiple lines, we show the quantum number of the lowest-frequency transition. The transitions are sorted by increasing frequency. The molecular structure of CH3CH(OH)C(O)H is also shown (C atoms in gray, O atoms in red, and H atoms in white).
In the text
thumbnail Fig. 6

Transitions of methoxyacetaldehyde, CH3OCH2C(O)H, detected toward G+0.693 (collected in Table A.6). The red line shows the best LTE fit of CH3OCH2C(O)H, while the blue line represents the combined emission of all molecules identified in the survey, including CH3OCH2C(O)H (observed spectra are shown as gray histograms). The quantum numbers for each transition are shown at the bottom of each panel. The transitions are sorted by increasing frequency. The molecular structure of CH3OCH2C(O)H is also shown (C atoms in gray, O atoms in red, and H atoms in white).
In the text
thumbnail Fig. 7

Suggested grain-surface chemical routes for the formation of targeted C3H6O2 isomers in the ISM. Dashed lines indicated processes involving more than one step. We highlight in boldface molecules identified toward G+0.693. C3H6O2 species that were not detected and whose formation has not been tested experimentally by Wang et al. (2023) and Wang et al. (2024) are highlighted with a light purple box. For simplicity, stereoisomerism is not considered in this plot.
In the text
thumbnail Fig. 8

Fractional abundances with respect to H2 of different C3H6O2 isomers sorted by increasing relative energy. The total dipole moment for each molecule (computed at the B2PLYP-D3/aug-cc-pVTZ level) is included. We used the same color-code as that in Fig. 1: new first detections in the ISM are shown in red, tentative detections in orange, non-detections in purple (upper limits indicated with black arrows), and previous detections toward different sources in black.
In the text
thumbnail Fig. A.1

Tentative detection of g-ethyl formate, g-CH3CH2OC(O)H, toward G+0.693. The selected transitions are sorted by decreasing intensity and listed in Table A.3. The result of the LTE model of the g-CH3CH2OC(O)H emission is shown in orange, together with the expected molecular emission from all the molecular species identified to date in our survey, including g-CH3CH2OC(O)H (in blue), both overlaid on the observations (gray histogram). The molecular structure of g-CH3CH2OC(O)H is also shown (C atoms in gray, O atoms in red and H atoms in white).
In the text
thumbnail Fig. A.2

LTE simulation of the propionic acid, CH3CH2C(O)OH, emission using the upper limit column density derived toward G+0.693 (in purple) together with the expected molecular emission from all the molecular species identified to date in our survey (in blue), both overlaid on the observations (gray histogram). The quantum numbers for each transition are shown at the bottom of each panel.
In the text
thumbnail Fig. A.3

LTE simulation of the glycidol, (c-CH2OCHCH2OH, emission using the upper limit column density derived toward G+0.693 (in purple) together with the expected molecular emission from all the molecular species identified to date in our survey (in blue), both overlaid on the observations (gray histogram). The quantum numbers for each transition are shown at the bottom of each panel.
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.