Asunción Fuente Centro de Astrobiología (CAB), CSIC-INTA, Ctra. de Torrejón a Ajalvir, km 4, 28850 Torrejón de Ardoz, Spain Maria N. Drozdovskaya Physikalisch-Meteorologisches Observatorium Davos und Weltstrahlungszentrum (PMOD/WRC), Dorfstrasse 33, CH-7260, Davos Dorf, Switzerland Silvia Spezzano Max-Planck-Institut für Extraterrestrische Physik, Giessenbachstrasse 1, 85748 Garching, Germany Gisela Esplugues Observatorio Astronómico Nacional (OAN), 28014 Madrid, Spain David Navarro-Almaida Centro de Astrobiología (CAB), CSIC-INTA, Ctra. de Torrejón a Ajalvir, km 4, 28850 Torrejón de Ardoz, Spain Marina Rodríguez Baras Observatorio Astronómico Nacional (OAN), 28014 Madrid, Spain Angèle Taillard Centro de Astrobiología (CAB), CSIC-INTA, Ctra. de Torrejón a Ajalvir, km 4, 28850 Torrejón de Ardoz, Spain Karin Öberg Center for Astrophysics — Harvard & Smithsonian, Cambridge, MA 02138, USA

Abstract

The isotopic ratios of deuterated methanol derived around protostars are commonly used to infer the physical conditions under which they formed in the earlier prestellar stage. However, there is a discrepancy in the ratio of the singly deuterated methanol isotopologues, CH2DOH/CH3OD, between low- and high-mass protostars, which puts into question whether prestellar isotopic ratios are generally preserved during the star- and planet-forming process. Resolving this puzzle is only made harder by the complete lack of data on this ratio in the prestellar stage. This work presents observations with the IRAM 30m telescope that securely detect CH3OD in the prestellar core L1448 in Perseus and tentatively in B213-C6 in Taurus. This work constrains the ratio of CH2DOH/CH3OD and the D/H ratios for both singly deuterated methanol isotopologues for the first time at the prestellar stage. Column densities calculated under the assumption of local thermal equilibrium lead to a CH2DOH/CH3OD ratio of 2.8–8.5 in L1448 and ≤\leq 5.7 in B213-C6. The values are marginally consistent with the statistically expected ratio of 3, but most assumptions put the values in an elevated range in line with values found around low-mass protostars. The D/H ratio in CH2DOH is between 3.6% and 6.8% in L1448 and in the range of 2.4–5.8% in B213-C6. The D/H ratio derived for CH3OD is lower, namely 1.4–4.4% in L1448 and ≤\leq 3.8% in B213-C6.

Astrochemistry, ISM: clouds, ISM: molecules

††facilities: IRAM 30m††software: cassis (Vastel et al. 2015), class(http://www.iram.fr/IRAMFR/GILDAS), gildas-class pipeline (Megías et al. 2023), matplotlib (Hunter 2007), numpy (van der Walt et al. 2011), scipy (Virtanen et al. 2020)

1 Introduction

Starless and dynamically evolved prestellar cores represent the birthplaces of low-mass stars like our Sun (M∗ << a few M⊙). They are characterized by cold temperatures (T ∼\sim 10 K) and gas densities ≥\geq 104 cm-3 (Benson & Myers 1989; Bergin & Tafalla 2007; André et al. 2014). Investigating their chemical composition allows us to learn about chemical processes at the initial stages of star formation, and lets us understand the initial molecular inventory that is inherited by subsequent stages of star and planet formation. Molecules are ubiquitous in space, and complex organic molecules (COMs; defined as carbon-bearing molecules with 6 or more atoms; Herbst & van Dishoeck 2009) have already been detected in the gas phase in prestellar cores (e.g., Bacmann et al. 2012; Jiménez-Serra et al. 2016). A key molecule to kickstart chemical complexity in space is methanol, CH3OH (Öberg et al. 2009; Chuang et al. 2016), it is the simplest and one of the most widespread COMs and is detected across all stages of star formation, and even in Solar System comets (Parise et al. 2006; Bizzocchi et al. 2014; Booth et al. 2021; Drozdovskaya et al. 2021). Methanol predominately forms in CO-rich ices via the hydrogenation of CO, or reactions between the intermediates of this hydrogenation sequence at conditions found in prestellar cores (T << 20 K, nH2{}_{\rm H_{2}} ≥\geq 105 cm-3; Watanabe & Kouchi 2002; Fuchs et al. 2009; Santos et al. 2022). Due to its efficient formation in cold ice and lack of gas-phase formation routes, methanol is uniquely suited to infer the past conditions under which ices formed (Geppert et al. 2006). A powerful tool to trace the historical physicochemical conditions during the star formation process is deuteration, a process that swaps out hydrogen for its heavier isotope, deuterium, and enriches molecules beyond the average interstellar D/H value of ∼\sim 10-5 (Linsky et al. 2006). D/H ratios on the order of 10-1–10-4 are routinely obtained towards prestellar cores and protostars (Nomura et al. 2022, Esplugues et al. 2022, Rodríguez-Baras et al. 2023). At temperatures below 20 K and gas densities >> 104 cm-3, enrichment in deuterium occurs via the reaction H+3{}_{3}^{{+} + HD ⇌\rightleftharpoons H2D+ + H2 + 232 K (Watson 1974). Once the temperature drops below 20 K, the backward reaction is quenched, and the ratio of H2D+/H+3{}_{3}}{+} increases compared to the elemental D/H ratio. The atomic reservoir that is now enriched in deuterium can participate in grain-surface reactions resulting in isotopic enrichments in molecules like methanol once neutral species such as CO freeze out. Thus, the D/H ratio of methanol directly depends on the elevated D/H ratio in the gas phase (Millar et al. 1989). Deuterated molecules are generally hard to detect in the solid state, only recently the first detection of a deuterated molecule in the ice, HDO, was claimed (Slavicinska et al. 2024). Instead, we rely on gas-phase spectroscopy following ice sublimation to trace the level of deuterium fractionation in ice chemistry products (Caselli & Ceccarelli 2012). Both singly deuterated methanol isotopologues, CH2DOH and CH3OD, have been readily detected around protostars of all masses (e.g., Parise et al. 2006; Peng et al. 2012; Neill et al. 2013; Belloche et al. 2016; Bøgelund et al. 2018; Jørgensen et al. 2018). At the prestellar stage, only CH2DOH has been detected so far (Bizzocchi et al. 2014; Ambrose et al. 2021; Lin et al. 2023). The D/H ratio of CH2DOH in low-mass prestellar cores is consistent with ratios obtained in low-mass Class 0/I protostars and comet 67P/Churyumov-Gerasimenko (Drozdovskaya et al. 2021), which suggests that their isotopic record is set at and subsequently inherited from the prestellar stage. H-D surface substitution that elevates D/H ratios in the CH3-group of methanol (Nagaoka et al. 2005) can explain the isotopic ratios of CH2DOH, as shown in astrochemical models (e.g., Taquet et al. 2012; Kulterer et al. 2022; Riedel et al. 2023). The D/H ratio in CH2DOH around high-mass protostars is lower compared to their low-mass counterparts, which could be due to slightly higher temperatures during their time of formation (van Gelder et al. 2022). This is also consistent with the D/H ratio of CH2DOH in high-mass prestellar cores (Fontani et al. 2015). There are some indicators; however, that the D/H ratio set at formation can be altered through ice chemical processes, veiling our extraction of ice formation environments from D/H ratios. The most pressing evidence is a claimed difference in D/H enhancement patterns in the OH-group of methanol between low- and high-mass protostars (Taquet et al. 2019). While the low-mass pre- and protostellar fractionation patterns generally conform to the expectations from theoretical models, the fractionation ratios towards high-mass protostars do not. In addition, the ratio of CH2DOH/CH3OD deviates from its statistically expected value of 3: for low-mass protostars, this value can exceed 10, while it drops towards unity and below for high-mass protostars (e.g., Parise et al. 2006; Bøgelund et al. 2018). It is not known which chemical process leads to the observed discrepancy in this ratio, and at which stage during the star formation process it takes place. In order to constrain whether this ratio is set at and inherited from the prestellar stage, or if it is affected by reprocessing of prestellar ices in the protostellar stage requires observations of CH3OD at the prestellar stage. A similar ratio of CH2DOH/CH3OD at pre- and protostellar stages would suggest that it is already set in prestellar cores, while a deviation would indicate that chemical processes during the warm-up stage alter the ratio, either through selective formation or selective destruction of one of the isotopologues. We present the first detection of CH3OD in the low-mass prestellar core L1448 in the Perseus star-forming region, and an upper limit in B213-C6 in the Taurus star-forming region, which allows us to put a first constraint on the ratio of CH2DOH/CH3OD in prestellar cores, and assess if the protostellar fractionation pattern of methanol is reflected by hydrogen fractionation chemistry in prestellar cores in the low-mass case. The paper is structured as follows: we detail the observations in Section 2, and describe the analysis process in Section 3. The detected singly deuterated methanol isotopologues and their isotopic ratios are described in Section 4, and discussed comparatively with detections across the star-forming sequence in 5 to shed light on potential formation pathways resulting in the observed ratios. We raise our conclusions in Section 6.

2 Observations

The prestellar core L1448 (RA = 03h25m49s.00, DEC = 30∘42′24′′.6) is located in the Perseus star-forming region; the prestellar core B213-C6 (RA = 04h18m08s.40, DEC = 28∘05′12′′) is located in Taurus. B213-C6 lies in the northern part of the B213 filament. This part is surrounded by ∼\sim 40 protostars (Luhman et al. 2009; Rebull et al. 2010). While the Perseus star-forming region is associated with two clusters in which pre-main sequence stars are currently forming (Lada et al. 1996; Luhman et al. 2003). L1448 is located in a quiescent part of the region that is not associated with active star formation. This work analyses data from two observational programs carried out at the IRAM 30m telescope. The first data set is the large program Gas phase Elemental abundances in Molecular CloudS (GEMS) and the second is project 071-23 (PI of both: A. Fuente). The GEMS observations were conducted in frequency-switching mode with a frequency throw of 6 MHz. By using the Eight MIxer Receivers (EMIR) in combination with the Fast Fourier Transform Spectrometers (FTS) as the backend; a spectral resolution of 49 kHz was achieved (Fuente et al. 2019). This paper makes use of the CH2DOH transitions at 85.2967, 86.6688, and 134.0658 GHz in the GEMS data. The rms for those transitions is between 13 and 20 mK. The second data set is project 071-23 with data taken on the 2nd and 3rd of May 2024 in excellent weather (pwv = 0–4.6 mm, τ\tau << 0.1) and at Tsys of 60–130 K. As for the GEMS data, EMIR with the FTS backend and a spectral resolution of 49 kHz was used. The data were also taken in frequency-switching mode with a frequency throw of 3.9 MHz. The four spectral windows were centered on 90.1, 93.6, 106.1, and 109.6 GHz, the rms for the spectral windows relevant for this work is 2.1–2.9 mK. The half power beam width (HPBW) is related to the frequency of the observed transition via HPBW(′′) = 2460/ν\nu111https://publicwiki.iram.es/Iram30mEfficiencies, where ν\nu is the frequency in GHz. This leads to a beam size of ∼\sim 29′′ at 85 GHz and a beam size of ∼\sim 18′′ at 134 GHz. The intensity scale of the data is the main beam temperature, TMB, which is related to the antenna temperature, T∗,A via TMB = (Feff/Beff) ⋅\cdot T∗,A. The beam efficiencies are listed in Table B.1 in Fuente et al. (2019). We note that CH2DOH at 89.4078 and 90.7798 GHz is detected in both projects in both cores, but due to the higher noise in the GEMS data, combining data sets did not improve the signal-to-noise significantly, and we thus do not include the GEMS data of those two lines in our analysis.

3 Analysis

The initial line identification process was conducted with the class package of the gildas222http://www.iram.fr/IRAMFR/GILDAS software, and the data were reduced with the automated gildas-class Pipeline (GCP333https://github.com/andresmegias/gildas-class-python/; Megías et al. 2023). Line frequencies and the spectroscopic parameters of the transitions were taken from the Jet Propulsion Laboratory (JPL; Pickett et al. 1998) for CH2DOH and from the Cologne Database for Molecular Spectroscopy (CDMS; Müller et al. 2001, 2005; Endres et al. 2016) for CH3OD based on the spectroscopic entry of Ilyushin et al. (2024). The spectroscopic entry of CH2DOH in JPL based on Pearson et al. (2012) states that column densities calculated from b- and c-type transitions are not reliable. We therefore only used the two a-type transitions at 89.4078 and 90.7798 GHz for determining the column density of CH2DOH. Partition function values below 10 K for CH3OD are not publicly available, so we calculated the partition function values at 6 and 8 K from the energy levels of the vt=0v_{t}=0 torsion level (H. Müller, private communication). For CH2DOH, we used an updated catalogue to obtain the partition function values (L. Coudert, private communication). The respective partition function values are listed in Table A2. The peak temperature of each detected line, TMB (K), alongside the center velocity, v0 (km/s), and the line width, δ\deltav (km/s), of each detected line was determined by fitting a Gaussian line profile utilizing the curve_fit package of scipy (Virtanen et al. 2020). The resulting values are shown in Table 1. The calculation of the column densities of CH2DOH and CH3OD was carried out under the assumption of local thermal equilibrium (LTE). For cores L1448 and B213-C6, it has been shown by Spezzano et al. (2022) and Kulterer et al. (submitted) that CH3OH cannot be fit using an LTE approach, but because collisional data for the deuterated methanol isotopologues are not available, we are left with using the LTE approximation. In the case of LTE, the line flux ∫\intTMBdv (K/km/s) is related to the column density in the upper state (Nup (cm-2)) via

	Nup=∫TMB​dv×8​π⋅kB⋅ν2h⋅c3⋅Aul,\rm N_{up}=\int T_{MB}dv\times\frac{8\pi\cdot k_{B}\cdot\nu^{2}}{h\cdot c^{3}\cdot A_{ul}},		(1)

where kB corresponds to the Boltzmann constant, ν\nu is the frequency of the transition, h is the Plank constant, c the speed of light, and Aul the Einstein A coefficient of the upper energy level. The peak temperature TMB (K) can be calculated from

	TMB=∫TMB​dvFWHM×2​2​l​n​(2)2​π,\rm T_{MB}=\frac{\int T_{MB}dv}{FWHM}\times\frac{2\sqrt{2ln(2)}}{\sqrt{2\pi}},		(2)

with the FWHM being the full width half maximum (δ\deltav) of the line. Nup, the column density in the upper energy state, is determined via

	Nup=gup⋅NtotQ​(Tex)⋅exp​(Eup/kB​Tex).\rm N_{up}=\frac{g_{up}\cdot N_{tot}}{Q(T_{ex})\cdot exp(E_{up}/k_{B}T_{ex})}.		(3)

Q(Tex) corresponds to the partition function at an excitation temperature Tex, Eup and gup correspond to the energy and the degeneracy of the upper state, respectively. As a first step, we attempted to fit one column density and excitation temperature if multiple lines of one molecule were detected with the cassis software444http://cassis.irap.omp.eu/ (Vastel et al. 2015). We explored a grid of Tex from 10 to 15 K in steps of 0.1 K and Ntot values of (5 ×\times 1011)–(5 ×\times 10)13{}^{{13}) cm-2 with a step size of 0.01 in logarithmic space. The source size was set to 27′′, which equals the beam size at 90 GHz, and we assumed an average line width from the detected transitions for each molecule. However, it was not possible to determine a good fit: the lowest χ2\chi}{2} obtained was ∼\sim 10, which may be due to non-LTE effects that are not included in the calculations. On the other hand, the temperature range was set to mimic the dust temperature range derived for L1448 and B213-C6 by Rodríguez-Baras et al. (2021); and the gas could potentially be colder. As a second approach, we instead calculated the best-fit column density for each transition for fixed Tex of 6, 8, and 10 K from the integrated intensity ∫TMB\int T_{\rm MB} dv. The column density for a molecule is calculated from the average of the individual best-fit column densities for each line. The column densities for the assumed grid of excitation temperatures are listed in Table 2.

4 Results

4.1 Detected Molecules

For the first time, CH3OD has been detected in a prestellar core. Its transitions at 90.6699, 90.7058, and tentatively at 110.1889 GHz are detected in L1448 (Fig. 1). We consider the detection at 110.1889 GHz as tentative, because the peak intensity of the line equals only 3 ×\times rms. A line at 90.7058 GHz is detected towards B213-C6 (Fig. 2), which matches the position of a CH3OD transition and within 1 MHz of this line, there is no other molecule identified in the online tool Splatalogue555https://splatalogue.online with transitions at upper level energies below 30 K. However, we only find this one line in the spectra of B213-C6, and therefore consider this only a tentative indication of CH3OD in this core and treat the column densities and ratios of CH3OD in B213-C6 as limits. At the position of the 90.6699 GHz transition in B213-C6 a signal just slightly above the noise level is detected with an intensity of 2.4 mK. In addition, we report the detections of five CH2DOH lines towards L1448 and three lines towards B213-C6.

Figure 1: Detected transitions of CH2DOH (top two rows) and CH3OD (bottom row) towards L1448. The line at 110 GHz is a tentative detection. The observed spectra plotted in black are overlaid by the Gaussian fits (see Table 1) in red. The silver dashed lines represent the value of ±\pm1 ×\times rms, and the dotted, black line represents the value of 3 ×\times rms. The purple dash-dotted vertical line represents the line rest velocity of CH3OH (Kulterer et al. submitted). Figure 2: Detected transitions of CH2DOH (top row) and CH3OD (bottom left panel) towards B213-C6. The observed spectra plotted in black are overlaid by the Gaussian fits (see Table 1) in red. The silver dashed lines represent the value of ±\pm1 ×\times rms, and the dotted, black line represents the value of 3 ×\times rms. The purple dash-dotted vertical line represents the line rest velocity of CH3OH (Spezzano et al. 2022).

The respective frequencies, upper level energies, Einstein A coefficients of the lines, as well as their fitted peak intensities, integrated intensities, line widths, and center velocities are listed in Table 1. The center velocities of the lines deviate by less than 0.5 km/s from the main methanol isotopologue: v0 for CH3OH in L1448 is 4.45 km/s (Kulterer et al. submitted) and 6.85 km/s in B213-C6 (Spezzano et al. 2022). The line widths of the deuterated isotopologues are slightly different from that of CH3OH in L1448 (δ\deltavCH3​OH{}_{\rm CH_{3}OH} = 0.57 km/s; Kulterer et al. submitted). The majority of the CH2DOH lines are 5–20 % narrower, while the CH3OD lines are within 20% of the line width of CH3OH. Within errors, the lines of the isotopologues lie within 2-3 channels at most from each other, so they emit from nearly the same region. For B213-C6, the lines of the deuterated isotopologues are on average only about half the line width of the main isotopologue (δ\deltavCH3​OH{}_{\rm CH_{3}OH} = 0.53 km/s; Spezzano et al. 2022), which hints that CH3OH in B213-C6 detected in the data might stem from multiple gas layers, while deuterated methanol emits from only one region. It is to note that the peak of the suspected CH3OD transition at 90.7058 GHz in B213-C6 differs by four channels (∼\sim 0.38 km/s) from the velocity of CH3OH, while the CH2DOH lines are within two channels.

Table 1: Observed transitions of CH2DOH and CH3OD in L1448 and B213-C6. Tentative detections are marked with an asterisk in front of the transition frequency. The plots of the spectra and their fits are shown in Figs. 1 and 2.

Following the approach in Section 3, the column density of CH2DOH in L1448 is in the range of (0.9–1.5) ×\times 1013 cm-2 and for CH3OD in the range of (1.2–3.5) ×\times 1012 cm-2 for excitation temperatures of 6–10 K (Table 2). In B213-C6, the column density of CH2DOH is in the range of (2.5–5.9) ×\times 1012 cm-2 and ≤\leq 1.3 ×\times 1012 cm-2 for CH3OD.

4.1.1 Potential Caveats

The line shape and intensity of the CH3OD transition at 90.6699 GHz may be affected by the use of a frequency throw of 3.9 MHz during frequency switching. There is a nearby HNC line at 90.6636 GHz; and the artifact produced by the automated data reduction pipeline of the IRAM 30m telescope during the folding of the data is close to the CH3OD transition (Fig. A.1). It is not possible to say if potential wiggles at the wing of the feature either increase or decrease the intensity of the line. The peak intensity of the CH3OD transition is >> 6 ×\times rms in L1448, so we consider this line detected, but we doubled the errors in the integrated intensity that we used for the fitting of the column densities compared to the value listed in Table 1. The peak intensity at the position of this transition in B213-C6 is only at the noise level. This non-detection can be attributed to non-LTE effects, because in LTE, the column density derived from the detected 90.7058 GHz transition should have led to a detection of this line too (Fig. A.2). Another caveat that has to be considered is that the calculation of the column densities has been carried out based on the assumption that deuterated methanol is in LTE. This might not be the case. However, collisional data for CH2DOH and CH3OD are not available. A hint that deuterated methanol could be in non-LTE comes from the two a-type CH2DOH transitions at 89.4078 and 90.7798 GHz. In LTE, their line intensity ratio should be ∼\sim 2, but it is 4–5 in L1448 and B213-C6. In addition, given the sensitivity of the observations, the CH3OD transitions at 90.6699 and 110.1889 GHz in B213-C6 should have just been detected at 3σ\sigma for Tex of 6–10 K based on the upper limit on the column density derived from the transition at 90.7058 GHz if CH3OD would be in LTE.

4.2 Isotopic Ratios of Singly Deuterated Methanol

In general, the column densities of the deuterated methanol isotopologues are one to two orders of magnitude less abundant than the main isotopologue. The column density of CH3OH is (3.5 ±\pm 0.4) ×\times 1013 cm-2 in B213-C6 (Spezzano et al. 2022) and (8.2 ±\pm 0.7) ×\times 1013 cm-2 in L1448 (Kulterer et al. submitted). In order to get from column densities to the D/H ratios of the two functional groups in methanol, one has to take statistical weighting into account, as a deuterium atom is three times more likely to land in the CH3-group than the OH-group. The D/H ratio of singly deuterated methanol can therefore be calculated via

	NCH2​DOH/NCH3​OH=3​(D/H)CH3​OH\rm N_{CH_{2}DOH}/N_{CH_{3}OH}=3(D/H)_{CH_{3}OH}		(4)

and

	NCH3​OD/NCH3​OH=(D/H)CH3​OH.\rm N_{CH_{3}OD}/N_{CH_{3}OH}=(D/H)_{CH_{3}OH}.		(5)

We calculated the D/H ratios of CH2DOH and CH3OD under the assumption that they emit from the same region as the main isotopologue, and present the values derived for excitation temperatures of 6, 8, and 10 K for deuterated methanol in Table 2. Depending on the assumed excitation temperature, the D/H ratio derived from CH2DOH in L1448 is between 3.6% and 6.8%, while in B213-C6, it is found to range from 2.4% to 5.8%. The highest excitation temperature leads to the highest D/H ratio; and the ratio decreases with decreasing excitation temperature. The D/H ratio in CH3OD is found to be between 1.4% and 4.4% in L1448, and ≤\leq 3.8% in B213-C6. Assuming that the excitation temperature of CH2DOH and CH3OD is the same, the D/H ratio in the CH3-group is always higher than in the OH-group. In the case of L1448, it is in the range of 40–50%, where the lower percentage corresponds to the lowest excitation temperature. For B213-C6, the D/H ratio in CH2DOH is ∼\sim 33% more elevated regardless of the excitation temperature. Furthermore, the deuterium fraction in CH2DOH is ∼\sim 30% higher in L1448 than in B213-C6, independent of the choice of excitation temperature. Again, we assume that CH2DOH and CH3OD emit from the same region in order to calculate the ratio of CH2DOH/CH3OD. Based on statistical weighting, the ratio is expected to be 3, but most assumptions put the obtained values above the expected value. Depending on Tex, the ratio is between 2.8 and 8.5 in L1448, ≤\leq 5.7 in B213-C6. At 6 K, the ratio is 4.9 ±\pm 2.1 in L1448 and ≤\leq 5.6 in B213-C6; at 8 K, it is 5.5 ±\pm 2.3 in L1448 and ≤\leq 5.6 in B213-C6; and the highest values are found for 10 K, namely 6.0 ±\pm 2.5 in L1448 and ≤\leq 5.7 in B213-C6. Only one line of CH3OD has been detected in B213-C6, hence the ratios are treated as limits for this core.

Table 2: Column densities of both singly deuterated methanol isotopologues for Tex of 6, 8, and 10 K, and D/H ratios of CH2DOH and CH3OD for Tex = 6, 8, and 10 K relative to CH3OH. We assume a column density for CH3OH of (8.18 ±\pm 0.70) ×\times 1013 cm-2 (Kulterer et al. submitted) for L1448 and a column density of (3.50 ±\pm 0.4) ×\times 1013 cm-2 for B213-C6 (Spezzano et al. 2022). In the last column, Ncorr corresponds to the D/H ratio of CH2DOH where statistical weighting has been taken into account. No statistical weighting is required for the D/H ratio of CH3OD.

5 Discussion

5.1 Origins of Prestellar Methanol Deuterium Fractionation Patterns

The average D/H ratio in the ISM is ∼\sim 10-5 (Linsky et al. 2006). However, the D/H ratios in CH2DOH and CH3OD are elevated beyond this value in prestellar cores, which shows that processes leading to their efficient deuteration must be present in prestellar cores. Methanol can form through successive hydrogenation of CO and from methane oxidation in ices (Watanabe & Kouchi 2002; Qasim et al. 2018; Santos et al. 2022) under conditions that are found in prestellar cores. In the case of methanol formation starting from CO, deuterium fractionation can occur through at least three mechanisms. One possibility to form CH2DOH and CH3OD in prestellar ices is by replacing one H atom with one D atom in the hydrogenation sequence

	CO→HCO→H2​CO→CH3​O,CH2​OH→CH3​OH,\rm CO\rightarrow HCO\rightarrow H_{2}CO\rightarrow CH_{3}O,CH_{2}OH\rightarrow CH_{3}OH,		(6)

when temperatures below 20 K and gas densities at ≥\geq 104 cm-3 (Watanabe & Kouchi 2002; Caselli & Ceccarelli 2012) lead to freeze-out of CO from the gas and enhanced availability of D atoms in the gas. This mechanism has been tested by models in Kulterer et al. (2022), and the models found that ratios of CH2DOH/CH3OD beyond the statistically expected value of 3 are only found for models with dust temperatures of 10 K, core ages >> 3 ×\times 105 yr, and gas densities ≥\geq 106 cm-3. While this mechanism surely contributes to the abundance of deuterated methanol, it is not sufficient to explain the ratios of CH2DOH/CH3OD in low-mass prestellar cores and protostars (e.g. Taquet et al. 2012; Kulterer et al. 2022; Riedel et al. 2023). In addition to the hydrogenation pathway, methanol can also form directly via

	CH3​O+H2​CO→CH3​OH+HCO.\rm CH_{3}O+H_{2}CO\rightarrow CH_{3}OH+HCO.		(7)

Experiments by Santos et al. (2022) have shown that this is the dominant formation pathway for methanol at temperatures of 10–16 K. These experiments have also shown that deuteration in this formation pathway occurs preferentially in the CH3-group, if H2CO is replaced by HDCO in Eq. 7. Thus, this formation mechanism is another means to a) increase the D/H ratio in CH2DOH, and b) boost the ratio of CH2DOH/CH3OD at the prestellar stage, but it remains to be tested in astrochemical models. Another formation pathway that leads to efficient and successive deuteration of the methyl group of methanol is abstraction. Here, an impinging H or D atom removes an H atom from the methyl group, e.g., CH3OH + H →\rightarrow CH2OH + H (Nagaoka et al. 2005). Depending on whether CH2OH reacts with H or D afterward, it either forms CH3OH again or forms CH2DOH. Furthermore, CH2DOH can get successively deuterated, if abstraction occurs again, experiments have shown that once incorporated in the molecule, D atoms do not get abstracted anymore (Nagaoka et al. 2005; Hidaka et al. 2009). This mechanism can thus lead to a) a ratio of CH2DOH/CH3OD higher than the statistically expected value of 3, and b) high D/H ratios in CH2DOH compared to CH3OD. Indeed, models by Kulterer et al. (2022) have shown that implementing this formation pathway is crucial for elevating the ratio of CH2DOH/CH3OD beyond the statistically expected value of 3 at prestellar stages. Moreover, Riedel et al. (2023) have implemented this abstraction scheme in a source-tailored model for the prestellar core L1544 and are able to closely reproduce the observed column densities of CH3OH and CH2DOH in L1544. In addition, this mechanism should lead to D/H ratios in CHD2OH and CD3OH that should be higher than in CH2DOH. This has indeed been confirmed by observations (e.g., Drozdovskaya et al. 2022; Ilyushin et al. 2023; Lin et al. 2023; Scibelli et al. 2025). This mechanism can thus explain the high D/H ratios in CH2DOH. In line with experimental work (Nagaoka et al. 2005; Hidaka et al. 2009), the lower dust temperatures in low-mass star-forming regions favor this process, higher temperatures in high-mass star-forming regions quench its efficiency, leading to lower D/H ratios in CH2DOH for high-mass objects, and a similar abundance of CH2DOH and CH3OD. This can explain the trend that we see in CH2DOH in Fig. 3 in low-mass prestellar cores compared to high-mass prestellar cores. Methanol formation can also occur with methane as a starting point via two routes. Experiments by Qasim et al. (2018) have shown that CH3OH can form in H2O-rich ices before CO freezes out. In such a scenario, methanol forms from

	CH4+OH→CH3+H2​O,\rm CH_{4}+OH\rightarrow CH_{3}+H_{2}O,		(8)

followed by

	CH3+OH→CH3​OH.\rm CH_{3}+OH\rightarrow CH_{3}OH.		(9)

Deuterated methanol can form if deuterated methane or OD are present. However, it is not known, if the D atom in CH3D prefers to go into CH2D or into HDO in Eq. 8; and thus, if it even has the potential to form deuterated methanol. Lastly, methanol can also form upon insertion of excited O(1D) into CH4, where excited O(1D) forms after O2 gets photodissociated into O(3P) + O(1D) (Bergner et al. 2017). Similar to the formation pathway investigated by Qasim et al. (2018), this pathway could also occur starting from CH3D, but it is again unknown if this preferentially forms CH2DOH or CH3OD. Whether the formation of the singly deuterated methanol isotopologs behaves statistically, or if the difference in energies required for breaking a C-H vs. a C-D bond is high enough to make a difference has to be investigated by either further experiments or computational studies. Only then it will be possible to assess how this formation pathway influences the ratio of CH2DOH/CH3OD and the D/H ratio of the methanol isotopologues in prestellar ices. Our current knowledge from experiments and observations suggests that the abstraction scheme can explain the observed D/H patterns in CH2DOH best. The bulk of CH3OD is likely formed via atom addition to CO, whether the routes with methane as a starting point preferentially form CH3OD and thus, change the ratio of CH2DOH/CH3OD, is yet to be investigated. The D/H ratios from low-mass prestellar cores translate well to protostars (Fig. 4), which suggests that the majority of deuterated methanol indeed stems from the prestellar stage, at least in the low-mass case.

5.2 Survival of Prestellar D/H Ratios in Methanol

The observed D/H levels in methanol across star-forming stages, source types, and isotopologues are not consistent. As seen in Fig. 3, the D/H fractionation in CH2DOH in low-mass prestellar cores and low-mass protostars is higher than in CH3OD. On the other hand, the average D/H fractionation is similar in CH2DOH and CH3OD for high-mass protostars. Consequently, the ratio of CH2DOH/CH3OD is higher for the low-mass case than the high-mass case (Fig. 4). No measurement of CH3OD exists in high-mass prestellar cores, so it is not possible to verify if this trend also extends to high-mass prestellar sources. However, independent of the formation stage or environment, all D/H ratios are higher by at least one order of magnitude compared to the ISM average of ∼\sim 10-5 (Linsky et al. 2006). The here-presented D/H ratios of methanol and the CH2DOH/CH3OD ratio in prestellar cores strongly suggest that the bulk of deuterated methanol is formed at the prestellar stage (Fig. 3, 4), but that there might be processes that alter the inherited isotopic fingerprint at subsequent stages. One such mechanism is the H-D exchange between the OH-group of methanol and water, which can alter the abundances of the main methanol isotopologue and CH3OD. This isotope exchange occurs at temperatures as low as 120 K (Ratajczak et al. 2009). H-D exchange on the methyl group of methanol has been found to be inefficient (Osamura et al. 2004), and thus, only the abundance of hydroxyl-deuterated methanol is going to be influenced by this process. For instance, this mechanism has been used to explain the column densities of CH3OD in Orion KL (Wilkins & Blake 2022). However, this thermal exchange of H and D between methanol and water has not only the potential to form CH3OD, it can also destroy it to form CH3OH. Whether formation or destruction of CH3OD is dominating depends on the reaction dynamics, and the starting ratios of D/H in water and methanol. In general, the likelihood of a CH3OD molecule meeting a H2O molecule in the ice is a lot higher than CH3OH encountering HDO or D2O. H2O is the most abundant ice species, while the abundance of the main methanol isotopologue is a few percent w.r.t. to water (Boogert et al. 2015). It is thus likely that this mechanism leads to changes in the D/H ratio of prestellar CH3OD, but the extent of this alteration is unknown. Once sublimated from the grains, the abundance of CH3OD can be altered via ion-molecule reactions in the gas phase (Osamura et al. 2004) via

	CH3​OD+X−H+→CH3​OHD++X,\rm CH_{3}OD+X-H^{+}\rightarrow CH_{3}OHD^{+}+X,		(10)

where X-H+ is a protonating ion, followed by the reaction

	CH3​OHD++e−→CH3​OH+D.\rm CH_{3}OHD^{+}+e^{-}\rightarrow CH_{3}OH+D.		(11)

To which extent this could change the inherited isotopic ratios of methanol has not been tested in models. The discussed mechanisms suggest that the D/H ratios in methanol could be mainly set at the prestellar stage and potentially marginally altered during the protostellar stage. To verify this hypothesis, we calculated D/H ratios of methanol and ratios of CH2DOH/CH3OD by compiling CH2DOH and CH3OD detections from the literature, which are presented in the Appendix in Table A1. The table has been adapted from Drozdovskaya et al. 2021, such that observations toward objects from this previous compilation are included if, both, singly deuterated isotopologues are detected or an upper limit for CH3OD has been derived. If only upper limits of CH2DOH are available for a source, it has been excluded. In addition, it has been updated with detections since the publication of that paper. The average D/H ratio in CH2DOH between low-mass prestellar cores and low-mass protostars differs by less than a factor of 2 (left panel Fig. 3). For low-mass prestellar cores, the values are in the range of ∼\sim 1–10%. The spread of values is larger for the protostellar evolutionary stage, with values in the range of ∼\sim 0.2–30% including the most extreme cases. It is to note that all prestellar core values have been determined from single-dish facilities, while some of the protostellar data have been observed with interferometers, which leads to the possibility to derive the abundances from emission at smaller spatial scales. Single-dish observations of protostars capture emission of the cold envelope, where material is less affected by the presence of the central heating source, and the D/H ratios are most likely still pristine. The temperature structure around protostars is less homogeneous than in prestellar cores, so the collected data stems from different emitting regions, which could explain the larger spread in ratios observed around protostars. The overall high D/H ratios favor a prestellar origin; and the current data strongly suggest that the isotopic fingerprints from the two stages in CH2DOH are connected. The average D/H ratio in CH2DOH does not change from high-mass prestellar cores to high-mass protostars, but, again, the spread in values is larger for the protostellar case. High-mass protostars display D/H ratios between 10-4 and 10-1; in high-mass prestellar cores, the values are in the range of ∼\sim 0.05–2%. In general, while the D/H ratios in CH2DOH are consistent for low-mass prestellar cores and low-mass protostars, and for high-mass prestellar cores and high-mass protostars, the ratios obtained towards high-mass sources are an order of magnitude lower than towards low-mass sources. Observations of the D/H ratio in CH2DOH around high-mass protostars that have been analyzed in van Gelder et al. (2022) and compared to the modeling work of, e.g., Bøgelund et al. (2018) and Taquet et al. (2019), show that dust temperatures at the prestellar phase of ≥\geq 20 K and core ages corresponding to the free-fall time, tff, can explain the observed D/H ratio of CH2DOH in high-mas