Free Access
Issue
A&A
Volume 569, September 2014
Article Number A27
Number of page(s) 8
Section Interstellar and circumstellar matter
DOI https://doi.org/10.1051/0004-6361/201423858
Published online 12 September 2014

© ESO, 2014

1. Introduction

Deuterium is known to be ~105 times less abundant than hydrogen in the Universe (Linsky 2003); nonetheless, high abundances of D-containing isotopologues are common findings in many interstellar environments (e.g., Ceccarelli 2002; Roueff & Gerin 2003). During the past couple of decades, observations have revealed large molecular deuteration in low-mass pre-stellar cores and Class 0 protostars, where singly, doubly, and – in some instances – triply deuterated molecules have been detected (see, e.g., Ceccarelli et al. 2007).

Methanol (CH3OH) is one of the species that show the highest D-enhancements: together with ammonia (Lis et al. 2002; van der Tak et al. 2002), it is one of the two molecules for which a triply deuterated form has been observed. CD3OH has been revealed in the low-mass protostar IRAS 16293–2422 (Parise et al. 2004), a source showing extreme deuterium enhancements. It is, however, not a pathological case, because Parise et al. (2006) also found very high abundances of CH2DOH and CHD2OH in an extended sample of low-mass Class 0 protostars.

Theoretical and experimental studies predict that methanol is formed on the grain surface by subsequent additions of hydrogen to iced CO (Tielens & Hagen 1982; Watanabe & Kouchi 2002). This process is thought to take place during the cold and dense pre-collapse phase (e.g., Öberg et al. 2011). Later on, the molecule is released in the gas when the heating of newly formed protostar sublimates the ice mantles (Ceccarelli et al. 2001). Methanol deuteration is thus likely to be produced completely by active grain-surface chemistry, controlled by the atomic D content of the accreting gas. The high atomic D/H ratio required to account for the observed fractionation (0.1–0.2, Parise et al. 2002) has been explained by invoking an efficient transfer of atomic deuterium from the HD main reservoir via the intermediate H2D+/D2H+ ions (Roberts et al. 2003; Parise et al. 2006), which are very abundant in the CO-depleted pre-stellar gas (e.g., Caselli et al. 2003; Phillips & Vastel 2003; Vastel et al. 2004; Parise et al. 2011). Presently, all the measured methanol deuterations have been satisfactorily reproduced by the most recent coupled gas-grain models (Taquet et al. 2012; Aikawa et al. 2012), thus supporting the hypothesis that D-fractionation in methanol is a distinctive relic of the protostars’ past history.

In this context, it is very interesting to study methanol deuteration in the pre-stellar gas. Starless cores represent the early stage of low-mass protostar evolution and offer the opportunity to probe the initial conditions in the process of star formation. These objects have a simple structure with no central heating source and little (thermal) turbulence, thus providing a favourable environment to study molecular deuteration. In particular, measuring deuterated methanol in pre-stellar cores yields a further test of the process responsible for the build-up of D-bearing isotopologue reservoir onto grain mantles.

A few studies of methanol in the pre-stellar gas have been reported so far. Earlier detections of the parent species were accomplished in TMC 1, TMC 1C, L134N (=L183), and B335 (Friberg et al. 1988; Takakuwa et al. 1998, 2000), and also in some translucent clouds by Turner (1998). Later on, CH3OH was also observed in L1498 and L1517B by Tafalla et al. (2006) and, together with its deuterated variant CH2DOH, in the shocked gas of the Class 0 L1157 source (Codella et al. 2012). Previous detections of deuterated methanol in a sample of pre-stellar cores were also reported in a summarised form by Bacmann et al. (2007), but a detailed analysis of these observations is actually missing.

In this paper we report on the observation of CH2DOH towards the starless cloud L1544, thus providing an accurate assessment of the methanol deuteration in the cold pre-stellar gas. Multi-frequency analysis of the CH3OH emission (including non-LTE modelling) is performed in order to derive a reliable value for the column density of the main isotopologue. We finally compare the obtained fractionation ratio with the results derived in Class 0 protostars and discuss the implications suggested by the predictions based on gas-grain chemical models.

2. Observations

The methanol data presented here have been collected using the IRAM 30 m antenna, located at Pico Veleta (Spain) during three observing sessions in 2012–2013. Single-pointing observations towards the L1544 dust emission peak, located at coordinates α(J2000) = 05h04m17.21s and δ(J2000) = +25°1042.8′′ (Caselli et al. 2002a), were carried out in October 2012, April 2013, and October 2013. The 3 mm lines of CH3OH were observed using several tunings of the EMIR E090 receiver while surveying various organic molecules in L1544 (Spezzano et al. 2013, and in prep.). We used the FTS backend in the “fine” configuration, resulting in 7.2 GHz of instantaneous bandwidth divided in four sub-bands with a final unsmoothed resolution of 50 kHz. The CH2DOH transitions were only observed during the April 2013 observing block. The E090 and E150 receivers were tuned at 89.780 and 134.07 GHz, respectively, and the line signals were collected in the lower-inner (LI) sideband. As for CH3OH, we used the FTS backend in the “fine” mode. Frequency-switching was adopted as observing mode using frequency throws of 3.9 MHz at 3 mm and 7.8 MHz at 2 mm. The telescope pointing was checked every two hours on nearby bright radio quasars and was found accurate to 3–4′′. Typical system temperatures were 85–130 K at 97 GHz and 170–220 K at 108 and 134 GHz. The observed spectra were then converted from the to the Tmb temperature scale adopting Beff and Feff values taken from the IRAM documentation. The rest frequencies and other spectroscopic parameters of the observed methanol lines are reported in Table 1.

L1544 was mapped during the Autumn 2013 session. We performed a 3′ × 3′ on-the-fly (OTF) map centred on the source dust emission peak (see above). The reference position was set at (−180″,180″) offset with respect to the map centre. Three methanol lines at 96.7 GHz were observed using two different E090 setups in the LI sideband and FTS in “fine” mode. The spectral axis was thus sampled with a 50 kHz channel spacing. The map area was swept during 4.5 h of telescope time by moving the antenna along an orthogonal pattern of linear paths separated by 8′′ intervals, corresponding to roughly one third of the beam FWHM (25.4′′ at 96.7 GHz). The mapping was carried out in good weather conditions (τ ~ 0.03) and a typical system temperature of Tsys ≈ 90 K. The data processing was done using the GILDAS1 software (Pety 2005).

Table 1

Spectroscopic parameters of the observed methanol lines.

3. Results

thumbnail Fig. 1

Intensity maps (units of ) of the 20,2 − 10,1 (A+), 21,2 − 11,1 (E2), and 20,2 − 10,1 (E1) transitions of CH3OH integrated over 0.5 km s-1 velocity interval. The L1544 dust peak position located at α(J2000) = 05h04m17.21s, δ(J2000) = +25°1042.8′′ is indicated by the blue cross marker. The first contour is at 5σ and the increment is 5σ for all three maps; note however that the colour scale is different for the weak 20,2 − 10,1 (E1) line (right panel). The images were smoothed to a 30′′ angular resolution to increase the signal-to-noise ratio (1σ ≈ 10-2 K km s-1).

Open with DEXTER

thumbnail Fig. 2

Left panel: grey-scale map of the summed integrated intensity (units of ) of the 20,2 − 10,1 (A+) and 21,2 − 11,1 (E2) CH3OH lines (30′′ angular resolution). Five equally spaced contours from 0.23 to 1.5 K km s-1 are plotted. The blue dashed contours plot the 1.3 mm continuum intensity map of Ward-Thompson et al. (1999) smoothed at 22′′ to improve the signal-to-noise ratio. Contours are at 100, 140, 180, and 220 mJy flux density levels. The red crosses represent the offset positions at which the spectra have been extracted. Right panel: map spectra of the CH3OH transitions towards the nine red crosses shown in the left panel. The velocity axis is centred on the 20,2 − 10,1 (A+) line at 7.2 km s-1. The vertical axis of each spectrum represents the scale in K, as shown in the lower leftmost panel.

Open with DEXTER

thumbnail Fig. 3

CH3OH lines observed towards the (0′′, 0′′) position of L1544. Left: group of 21,2 − 11,1 and 20,2 − 20,1 transitions. Right: 00,0 − 11,1 (E1E2) line. The spectral rms is ~3 mK. The red solid line plots the Gaussian fit obtained using CLASS.

Open with DEXTER

3.1. Mapping of CH3OH

Figure 1 shows the maps obtained for the CH3OH 20,2 − 10,1 (A+), 21,2 − 11,1 (E2), and 20,2 − 10,1 (E1) transitions. The last is appreciably weaker because it comes from a level of higher energy (see Table 1). The reference centre of the map is indicated and coincides with the maximum of the continuum 1.3 mm emission (Ward-Thompson et al. 1999).

In the right-hand panel of Fig. 2, we plot a grid of spectra of the 20,2 −10,1 (A+) and 21,2 − 11,1 (E2) CH3OH lines. They were taken in nine positions separated by 30′′, as shown in the left-hand panel of the figure. Here, the grey scale represents the summed integrated intensity of the two lines, whereas the blue contours plot the 1.3 mm continuum emission map (Ward-Thompson et al. 1999) smoothed at 22′′. The lines peak strongly at the position 6, (0′′, 30′′), whereas the corresponding signals are reduced by one-half at the position 5, which coincides with the map reference centre and is also very close to the maximum of the dust emission.

The CH3OH emission thus differs from the dust continuum and presents a single peak offset to the north-east. A weak intensity enhancement located to the south-west is also discernible, thus suggesting the presence of a highly asymmetrical broken ring distribution and reflecting the presence of a central CO depletion hole. This feature is apparent in the C17O integrated intensity map shown in Caselli et al. (1999, see their Fig. 1). Also, the observed azimuthal asymmetry of the methanol emission is likely to be linked to slight inhomogeneities of the CO depletion, owing to the non-spherical and cometary-shaped morphology (e.g., Tafalla et al. 2004; Crapsi et al. 2007).

3.2. CH3OH single-pointing observations

Besides mapping, single-pointing, sensitive observations were performed towards the L1544 centre, corresponding to position 5 of Fig. 2 (dust emission peak). Four CH3OH emission lines were observed: three falling in a small frequency interval at 96.7 GHz (the same mapped, see Fig. 1), plus one line at 108.9 GHz. The observations are shown in Fig. 3. Line profiles were analysed using the GAUSS task of CLASS and the resulting data are reported in Table 2. The fit of the three closely spaced lines 20,2 − 10,1 (A+), 21,2 − 11,1 (E2), and 20,2 − 10,1 (E1) was carried out by adjusting only one “common” line width. Given the relatively small number of channels used in the least-squares fit, the statistical error on the optimised parameters yielded by the GAUSS procedure could be optimistic. Thus, to be on the safe side, we conservatively quote 3σ uncertainties in Table 2.

The emitting levels have energies ranging from 7 to 20 K, thus one may try to derive the CH3OH column density, N, and the average excitation temperature, Tex, through the population diagram method (Goldsmith & Langer 1999). We used here the modified method described by Nummelin et al. (2000), which also includes the cosmic background emission and the peak optical depth in addition to Tex and N. In this approach, the optimised parameters were sought by minimising the squared sum of the error weighted differences between observed and modelled line intensities. These intensities are derived through the radiative transfer equality (1)where ηbf is the source-beam filling factor, and Jν(T) is the equivalent Rayleigh-Jeans temperature. Assuming a Gaussian line profile, the peak opacity of each transition is obtained as (2)Here, ν is the emission frequency, Δv the FWHM line width in units of velocity, Eu the energy of the upper level (as listed in Table 1), gu the rotational degeneracy, A the Einstein’s coefficient for spontaneous emission, and Q(T) is the rotation partition function at temperature T. This is computed by summing over all (A + E) rotational levels (Xu & Lovas 1997), whose energies are available at the CDMS2 (Müller et al. 2005). Throughout the calculations, the average FWHM line width of 0.37 km s-1 was used and the beam filling factor ηbf was set to unity.

Table 2

Results of the CLASS GAUSS fit on the observed spectral profile of the methanol lines observed towards L1544.

thumbnail Fig. 4

Rotational diagram produced by the four CH3OH lines observed towards L1544. A considerable scatter is apparent. The red dashed line represent the “best” linear fit of all data and yields Tex = 6 ± 3 K (see text).

Open with DEXTER

Once the best-fit N and Tex are determined through Eqs. (1)and (2), the match between observed and modelled data can be presented in a population diagram fashion, ln(Nu/gu) vs. Eu, where the Nu value for each transition is derived from the corresponding peak optical opacity through (3)The result of this comparison is illustrated in Fig. 4. The observed points exhibit a considerable scatter, largely exceeding the estimated error bars (~15%, including calibration and pointing uncertainties). The largest deviations are shown by the 21,2 − 11,1 (E2) line, which appears substantially brighter than expected. The best-fit modelled points instead lie on a straight line, not far from the linear fit obtained under the assumptions of optically thin emission and negligible background radiation (Goldsmith & Langer 1999). Indeed, the derived peak optical depths are moderate (τ< 0.4) and the cosmic background (Tbg = 2.7 K) emission merely acts as an offset. Figure 4 clearly indicates that excitation anomalies are present, thus preventing determination of a unique excitation temperature for all the rotational levels involved in the observed emissions. Indeed, the analysis yielded poorly constrained results: Tex = 6 ± 3 K and a column density value N = (1.9 ± 1.9) × 1013 cm-2.

3.3. Non-LTE modelling

To achieve a better constraint for the CH3OH column density in L1544, we performed a non-local thermodynamic equilibrium (LTE) modelling using the radiative transfer code MOLLIE (Keto & Rybicki 2010) and the L1544 physical model with central density of 1 × 107 cm-3 described in Keto et al. (2014).

thumbnail Fig. 5

Left panel: radial profiles for the E-CH3OH abundance (black) and density (red) corresponding to the best-fit input abundance of 9.2 × 10-10. The radial trends were produced by the simple chemistry model implemented in the radiative transfer code MOLLIE (see text).

Open with DEXTER

The model is computed with a spherical Lagrangian hydrodynamic code with the gas temperature set by radiative equilibrium between heating by external starlight and cosmic rays and cooling by molecular line and dust radiation (see Keto & Caselli 2008, for a comprehensive discussion of the theory). Even if the L1544 emission maps show that the cloud is not spherical but instead has an elongated shape, the adopted model is simple, physically motivated and certainly adequate to model observation data averaged over a single dish beam profile. For the statistical equilibrium calculation, we used de-excitation rates for p-H2/A-CH3OH and p-H2/E-CH3OH collisional systems (Rabli & Flower 2010a) available at the LAMDA database (Schöier et al. 2005). Collisional data for o-H2 are lacking, but this does not represent a problem for our modelling because the H2 ortho-to-para ratio (OPR) is expected to be very low in pre-stellar cores (see, e.g., Walmsley et al. 2004; Sipilä et al. 2013). Simulations were run separately for A and E symmetry species of methanol and considering only rotational levels below 36 K.

MOLLIE implements a simple CO chemistry (Keto & Caselli 2008) to take into account both the molecule freeze-out towards the inner cloud core and the photo-dissociation due to the UV stellar field at the edges. Since CO is considered to be the main precursors of methanol, we used the same model to describe the molecular abundance trend across the core. The “nominal” CH3OH abundance, provided as the input parameter to MOLLIE, is thus internally translated in a radial abundance profile. As an example, the E-CH3OH radial abundance and density profile corresponding to the best-fit input abundance (see below) are illustrated in Fig. 5. The resulting beam-averaged value of the CH3OH column density (including both A and E species) is 2.66 × 1013 cm-2.

At the end of the computation, the code outputs a data cube representing the spectral distribution of the emerging radiation field. After convolution with the appropriate telescope beam (25.5′′ at 97.6 GHz, 22.9′′ at 108.9 GHz), the modelled spectra were extracted from the central pixel and compared to the observations.

thumbnail Fig. 6

Observed vs. modelled spectrum of CH3OH in L1544. The black histogram shows the observations. Coloured lines indicate the best-fit modelled spectra. The velocity axis is centred on the 20,2 − 10,1 (A+) line at 7.2 km s-1.

Open with DEXTER

Optimal χ2-fit between observed and modelled spectral profiles were found for the 20,2 − 10,1 (A+), 21,2 − 11,1 (E2), and 20,2 − 10,1 (E1) transitions using input abundances of 9.5 × 10-10 and 9.2 × 10-10, for A and E species, respectively. The modelling results for these three emissions falling at 96.7 GHz are shown in Fig. 6.

thumbnail Fig. 7

Excitation temperature vs. cloud radius plot for the 21,2 − 11,1, 20,2 − 10,1, and 00,0 − 11,1 transitions of E-CH3OH as computed by the present best-fit radiative transfer modelling.

Open with DEXTER

Despite the success in reproducing these three lines, the fit underestimates the brightness of the observed 00,0 − 11,1 (E1E2) emission at 108.9 GHz by a factor of 4. Figure 7 illustrates the trend of the excitation temperatures for the modelled E-CH3OH transitions as a function of the cloud radius. It shows that the 00,0 −11,1 line is sub-thermally excited even at the high gas density of the cloud centre, where the other two E lines are instead thermalised. The reason for this behaviour is easily understood by inspecting Fig. 8, which shows the bottom part of the rotational level diagram of E-CH3OH, together with the radiative and collisional transitions considered in the present modelling. The 20,2 and 21,2 upper levels are connected to their corresponding lower state within the Ka = 0 or Ka = 1 manifolds by “strong” collisional transitions (i.e., upward rate > 10-11 cm3 s-1 at 10 K), and other weaker E1E2 connections also exist. On the other hand, the collisional transitions connecting 00,0 to the lower 11,1 and 21,2 states have rate coefficients that are set to zero in the present E-CH3OH/p-H2 data set (Rabli & Flower 2010a). As discussed in Rabli & Flower (2010b), this is an artefact of the coupled state (CS) approximation used in the production of the collision cross sections, and it is very likely that the collisional 00,0 − 11,1 and 00,0 − 21,2 transitions actually have small, but non-zero, rate coefficients.

Repeated tests using altered sets of collisional data showed that rate coefficients as small as 3 × 10-11 cm3 s-1 (ca. one third of the 21,2 − 11,1 upward rate) allow for a satisfactory modelling of the 00,0 − 11,1 (E1E2) emission without significantly altering the fit quality of the 21,2 − 11,1 (E2) and 20,2 − 10,1 (E1) transitions. This finding thus suggests that the difficulties encountered in modelling the 00,0 − 11,1 (E1E2) line are indeed caused by small inaccuracies in the collisional dataset used.

thumbnail Fig. 8

Rotational energy plot for the lowest levels of E-CH3OH. Active collisional channels are indicated by the upward grey arrows, with solid lines marking “strong” collisional transitions (upward rate > 10-11 cm3 s-1 at 10 K, Rabli & Flower 2010a). Red solid arrows indicate the observed radiative transitions.

Open with DEXTER

An estimate of the error bar associated to the CH3OH column density was obtained as in Bizzocchi et al. (2013). The Gaussian width of the χ2 profile (plotted vs. the free parameter) is used to get an estimate of the uncertainty involved in the optimisation process (~19%). A further 10% error is added in quadrature to take the telescope calibration into account, yielding a final estimate of a 22% relative uncertainty. We thus ended with N(CH3OH) = (2.7 ± 0.6) × 1013 cm-2. This value is not far from the uncertain results derived using an LTE approach in Sect. 3.2, and it provides a more accurate constraint for the CH3OH column density in L1544.

3.4. Singly deuterated methanol (CH2DOH)

Two lines of CH2DOH have been detected towards L1544, 20,2 − 10,1 and 30,3 − 20,2, at 98.4 and 134.1 GHz, respectively. Both are a-type transitions belonging to the lower e0 torsional state (see Table 1). The observations are shown in Fig. 9, and the results of the CLASS Gaussian fits are reported in the last two rows of Table 2. Interestingly, both CH2DOH lines peak at a slightly bluer velocity compared to the ones of the parent species, which in turn are consistent with the systemic velocity of L1544 (7.2 km s-1, Caselli et al. 2002a). The difference, ca. 0.3 km s-1, is small but significant given the fit statistical errors (< 0.003 km s-1 for most lines) and the spectral channel spacing of the observations (0.11−0.15 km s-1). It is also unlikely to be due to uncertainties in the rest frequencies because they are predicted to be as small as ~0.005 km s-1. The slight discrepancy is thus suggestive of a complex cloud dynamics, and it suggests that deuterated methanol traces a more confined (maybe inner) region with respect to its parent species. To confirm this, a map of CH2DOH is required.

thumbnail Fig. 9

CH2DOH lines observed towards the (0′′, 0′′) position of L1544. Left: 20,2 − 10,1 (e0) – right: 30,3 − 20,2 (e0). The spectral rms is ~5 mK. The red solid line plots the Gaussian fit obtained using CLASS.

Open with DEXTER

The CH2DOH column density can be calculated from each transition assuming LTE conditions and optically thin emission once a suitable constraint for the excitation temperature, Tex, is available. It holds that (4)Given the uncertainty of the excitation temperature determined for the parent species, the column density of the deuterated variant is derived through Eq. (4)assuming Tex ranging between 5 and 8 K. The results are collected in Table 3.

Table 3

CH2DOH column densities determined in L1544 assuming optically thin emission and excitation temperature in the 5–8 K interval.

The rotational partition function, Q(T) (whose values are also reported in Table 3) refers to the entire CH2DOH population (i.e., it includes e0, e1, and o1 levels) and was calculated by summing over rotational levels using the spectroscopic data of Pearson et al. (2012). The error bar of the CH2DOH column density is estimated adding in quadrature the calibration error (~15%) and the maximum dispersion due to the uncertainty in Tex. In this way we obtain N(CH2DOH) = (2.4 ± 0.9) × 1012 cm-2.

We also performed sensitive observations at the frequency of the 11,0 − 10,1 (AA+) line of the singly deuterated CH3OD isotopologue, but they resulted in a non-detection. The achieved 3σ sensitivity was 7 mK km s-1 (assuming a line width similar to that found for CH2DOH) thus, with Tex constrained in the 5–8 K interval, we derived a 3σ upper limit of 2.4 × 1011 cm-2 for the beam-averaged CH3OD column density (see next section for a possible explanation of this non-detection).

4. Discussion and conclusion

We have detected two lines of CH2DOH in L1544 and carried out an accurate evaluation of the methanol deuteration in a cold pre-stellar gas. From the beam-averaged column densities computed in Sects. 3.3 and 3.4, we obtain a fractionation ratio [CH2DOH]/[CH3OH] = 0.10 ± 0.03, a value significantly lower than the ones measured in low-mass Class 0 protostars (0.4–0.6, Parise et al. 2006), and lower than the deuterium fraction measured in molecules such as N2H+, NH3, and H2CO (Caselli et al. 2002b; Bacmann et al. 2003; Roueff et al. 2005; Crapsi et al. 2007). However, it should be noted that, unlike methanol, all these molecules can also form in the gas phase, so that only our results reflect the surface chemistry activity directly.

State-of-the-art gas–grain chemical models of deuterium chemistry have been recently published by Aikawa et al. (2012) and Taquet et al. (2012). These studies follow the molecular evolution and the D-fractionation as star formation proceeds from the pre-collapse phase to a proto-stellar core. Both models are able to reproduce the high [CH2DOH]/[CH3OH] ratio observed in Class 0 protostars if D and H abstraction and substitution are included in the chemical network. At densities of 5 × 106 cm-3 and T = 10 K, the observed deuterium enhancement is reached in the molecular ice after 104−105 yr (relatively faster, ~5 × 103 yr, following Taquet et al. calculations). On the other hand, a recent study of the gas-phase D-fractionation process (Kong et al. 2013) indicates that the gas within the inner core of L1544 has a deuteration age (estimated through N2H+) of five to eight times the free-fall time scale, which is of the order of 104 yr. This suggests that the deuterium reservoir frozen on grains should be fully developed, at least in the central region of the core.

However, it should be noted that the column densities have been derived from observations towards the core centre, where CH3OH suffers a considerable depletion (see Fig. 1). Thus, the present observations may mostly be sensitive to the outer parts of the core, where the deuterium fractionation is lower than in the region traced by N2H+, a molecular ion not significantly depleted at high gas densities (e.g., Bizzocchi et al. 2013). The observed methanol deuteration is intermediate between the N2D+/N2H+ (0.2) and DCO+/HCO+ (0.04) found by Caselli et al. (2002b) in L1544. This gives the hint that methanol deuteration is indeed tracing the region where CO is freezing out (at densities of a few 104 cm-3), where the D/H ratio is not high enough to reach D-fractions close to the ones found towards Class 0 sources or the centre of L1544. Inner regions are lost to view, since freeze-out is probably too efficient or because the product of deuteration remains on the grains, because centre grains are covered with N2 towards the core (see, e.g., Bertin et al. 2013). Also, the morphology of the CH2DOH emission in L1544 is not known, so the low measured [CH2DOH]/[CH3OH] ratio might be produced by chemical inhomogeneities present in this pre-stellar core. Further progress in this study requires sensitive, interferometric observations aimed at deriving a detailed view of the fractionation in the different regions of the source.

From our non-LTE modelling results, we may infer that gas-phase CH3OH in L1544 is composed of an almost equi-molar mixture of A and E species, and the resulting [ E ] / [ A ] ratio is 0.97 ± 0.26. Given the large associated uncertainty, we may conclude that this value is consistent with the one implied by the picture of methanol formation on thermalised dust grains, (e.g., [ E ] / [ A ] ≈ 0.7 at 10 K).

The singly-deuterated CH3OD isotopologue was not detected by the present observations, yielding an upper limit of 2.4 × 1011 cm-2 for its column density. The corresponding ratio between the singly-deuterated forms of methanol is [CH2DOH]/[CH3OD] ≥ 10, higher than the nominal value of 3 that is predicted on a statistical basis assuming that D atoms are randomly distributed in the methanol isotopologues (Rodgers & Charnley 2002). Our finding confirms the trend reported by Bacmann et al. (2007), and it agrees with the results of Class 0 protostars, where CH3OD appears to be under-abundant with respect to the other methanol isotopic species ([CH2DOH]/[CH3OD] ~ 14–20, Parise et al. 2006). Also, large [CH2DOH]/[CH3OD] ratios have been found by Ratajczak et al. (2011) in a sample of low- to high-mass protostars.

The explanation of these “anomalous” ratios is still a challenge for the gas-grain chemical models, because H and D abstraction and substitution reactions – whose rates on ices are not very well constrained – are crucial to reproducing the observed abundances of the various methanol deuterated forms (e.g., Taquet et al. 2012). However, a laboratory study of low-temperature formaldehyde hydrogenation (Hidaka et al. 2009) shows that the formation of CH2DOH in ice dominates CH3OD, owing to the higher velocity of the H–D substitution process compared to the D-atom addition.

The CH3OH emission has a highly asymmetric annular distribution surrounding the dust peak, where CO is mainly frozen onto dust grains. Methanol is expected to form via successive hydrogenation of CO on the surface of dust grains and then partially released in the gas phase upon formation (i.e., part of the formation energy is used to evaporate, in a process called reactive desorption; Garrod et al. 2006) and/or upon photo-desorption by UV photons produced by cosmic-ray impacts with H2 molecules (Prasad & Tarafdar 1983). Evaporated methanol will then freeze-out onto dust grains in a time scale inversely proportional to the gas number density (~ 109/nH yr; van Dishoeck et al. 1993). Thus, the gas phase abundance of methanol is expected to decrease towards the centre of the core, where the density (and the freeze-out rate) is higher and where the outer layers of ice mantles may be rich in N2, preventing hydrogenation of the underlying CO-rich layers (Bertin et al. 2013; Vasyunin & Herbst 2013).


1

See GILDAS home page at the URL: http://www.iram.fr/IRAMFR/GILDAS.

2

Cologne Databases for Molecular Spectroscopy, URL: http://www.astro.uni-koeln.de/cdms/

Acknowledgments

We are grateful to the IRAM 30 m staff for their support during the observations. L.B. and E.L. gratefully acknowledge support from the Science and Technology Foundation (FCT, Portugal) through the Fellowships SFRH/BPD/62966/2009 and SFRH/BPD/71278/2010. P.C. acknowledges the financial support of the European Research Council (ERC; project PALs 320620). L.B. also acknowledges travel support to Pico Veleta from TNA Radio Net project funded by the European Commission within the FP7 Programme.

References

All Tables

Table 1

Spectroscopic parameters of the observed methanol lines.

Table 2

Results of the CLASS GAUSS fit on the observed spectral profile of the methanol lines observed towards L1544.

Table 3

CH2DOH column densities determined in L1544 assuming optically thin emission and excitation temperature in the 5–8 K interval.

All Figures

thumbnail Fig. 1

Intensity maps (units of ) of the 20,2 − 10,1 (A+), 21,2 − 11,1 (E2), and 20,2 − 10,1 (E1) transitions of CH3OH integrated over 0.5 km s-1 velocity interval. The L1544 dust peak position located at α(J2000) = 05h04m17.21s, δ(J2000) = +25°1042.8′′ is indicated by the blue cross marker. The first contour is at 5σ and the increment is 5σ for all three maps; note however that the colour scale is different for the weak 20,2 − 10,1 (E1) line (right panel). The images were smoothed to a 30′′ angular resolution to increase the signal-to-noise ratio (1σ ≈ 10-2 K km s-1).

Open with DEXTER
In the text
thumbnail Fig. 2

Left panel: grey-scale map of the summed integrated intensity (units of ) of the 20,2 − 10,1 (A+) and 21,2 − 11,1 (E2) CH3OH lines (30′′ angular resolution). Five equally spaced contours from 0.23 to 1.5 K km s-1 are plotted. The blue dashed contours plot the 1.3 mm continuum intensity map of Ward-Thompson et al. (1999) smoothed at 22′′ to improve the signal-to-noise ratio. Contours are at 100, 140, 180, and 220 mJy flux density levels. The red crosses represent the offset positions at which the spectra have been extracted. Right panel: map spectra of the CH3OH transitions towards the nine red crosses shown in the left panel. The velocity axis is centred on the 20,2 − 10,1 (A+) line at 7.2 km s-1. The vertical axis of each spectrum represents the scale in K, as shown in the lower leftmost panel.

Open with DEXTER
In the text
thumbnail Fig. 3

CH3OH lines observed towards the (0′′, 0′′) position of L1544. Left: group of 21,2 − 11,1 and 20,2 − 20,1 transitions. Right: 00,0 − 11,1 (E1E2) line. The spectral rms is ~3 mK. The red solid line plots the Gaussian fit obtained using CLASS.

Open with DEXTER
In the text
thumbnail Fig. 4

Rotational diagram produced by the four CH3OH lines observed towards L1544. A considerable scatter is apparent. The red dashed line represent the “best” linear fit of all data and yields Tex = 6 ± 3 K (see text).

Open with DEXTER
In the text
thumbnail Fig. 5

Left panel: radial profiles for the E-CH3OH abundance (black) and density (red) corresponding to the best-fit input abundance of 9.2 × 10-10. The radial trends were produced by the simple chemistry model implemented in the radiative transfer code MOLLIE (see text).

Open with DEXTER
In the text
thumbnail Fig. 6

Observed vs. modelled spectrum of CH3OH in L1544. The black histogram shows the observations. Coloured lines indicate the best-fit modelled spectra. The velocity axis is centred on the 20,2 − 10,1 (A+) line at 7.2 km s-1.

Open with DEXTER
In the text
thumbnail Fig. 7

Excitation temperature vs. cloud radius plot for the 21,2 − 11,1, 20,2 − 10,1, and 00,0 − 11,1 transitions of E-CH3OH as computed by the present best-fit radiative transfer modelling.

Open with DEXTER
In the text
thumbnail Fig. 8

Rotational energy plot for the lowest levels of E-CH3OH. Active collisional channels are indicated by the upward grey arrows, with solid lines marking “strong” collisional transitions (upward rate > 10-11 cm3 s-1 at 10 K, Rabli & Flower 2010a). Red solid arrows indicate the observed radiative transitions.

Open with DEXTER
In the text
thumbnail Fig. 9

CH2DOH lines observed towards the (0′′, 0′′) position of L1544. Left: 20,2 − 10,1 (e0) – right: 30,3 − 20,2 (e0). The spectral rms is ~5 mK. The red solid line plots the Gaussian fit obtained using CLASS.

Open with DEXTER
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.