DISCOVERY OF THE METHOXY RADICAL, CH₃O, TOWARD B1: DUST GRAIN AND GAS-PHASE CHEMISTRY IN COLD DARK CLOUDS^*

Complex organic molecules are detected toward several astrophysical environments as a product of gas and dust grain chemistry (Herbst & van Dishoeck 2009). In prestellar cores where the temperature is low and the gas well shielded against UV photons, it has been assumed that chemistry is dominated by ion–neutral reactions predicting rather low abundances for these complex species. Previous observations toward cold dense clouds have provided upper limits to most of these species typically detected in hot cores. However, the high sensitivity that can be reached with receivers currently installed in large radio telescopes has not yet been used to provide significant upper limits or detections for these species. Despite the large bandwidth provided by these receivers, it can be significantly reduced when high spectral resolution is needed, e.g., for quiescent clouds showing very narrow line profiles. Therefore, a systematic line survey would be extremely time consuming. Marcelino et al. (2005, 2007, 2009) and Marcelino (2007) started such an exploration and found that molecules such as propylene (CH₂CHCH₃), although exhibiting weak lines due to its low dipole moment, were easily detected (Marcelino et al. 2007). This discovery was followed by the detection of fulminic acid, HCNO (Marcelino et al. 2009) in B1-b and TMC1. B1-b contains two submillimeter point sources, B1-bN and B1-bS, separated by 20''(Hirano et al. 1999), and a protostar detected by Spitzer (see, e.g., Öberg et al. 2010; Hiramatsu et al. 2010). Although several outflows have been detected in the region, it is not clear if they are related to those sources (Jørgensen et al. 2006). Nevertheless, the kinetic temperature of B1-b remains relatively cold, 12–15 K (Marcelino et al. 2005, 2009). Observations of this cloud by Öberg et al. (2010) have shown the presence of methyl formate (CH₃OCOH), acetaldehyde (CH₃CHO), and tentatively of dimethyl ether (CH₃OCH₃). Some of these species were also reported by Marcelino (2007). Recent observations by Bacmann et al. (2012) also show the presence of complex molecules in the prestellar core L1689B. In this Letter, we report on the discovery in space of the methoxy radical, CH₃O, and on the detection of a large number of complex organic molecules toward the cold cloud B1-b. The derived abundances are discussed in the context of ice-laboratory experiments and gas-phase chemistry.

The observations presented in this Letter are part of a complete spectral line survey at 3 mm (Marcelino et al. 2005, 2007, 2009; N. Marcelino et al., in preparation), performed at the IRAM 30 m telescope at Pico Veleta (Spain). Using the 3 mm Eight MIxer Receivers (16 GHz of total instantaneous bandwidth per polarization) and the fast Fourier Transform Spectrometers with a spectral resolution of 50 kHz (which allows the observation of the inner 1.82 GHz of each band), only six receiver setups were needed to cover frequencies between 82.5 and 117.5 GHz. We observed the whole 3 mm band between 2012 January and March, toward B1-b ($\alpha _{{\rm J}2000}=03^{\rm h} 33^{\rm m} 20\mbox{$.\!\!^{\mathrm s}$}8$, δ_J2000 = 31°07'34'') and TMC1 ($\alpha _{{\rm J}2000}=04^{\rm h} 41^{\rm m} 41\mbox{$.\!\!^{\mathrm s}$}88$, δ_J2000 = 25°41'27''). Weather conditions were mostly average to moderate winter, with 3–10 mm of precipitable water vapor. System temperatures were between 80 and 120 K, which resulted in an average rms of ∼4–6 mK for both sources, except at the highest frequencies close to the end of the band for which increases up to ∼10–20 mK (T_sys ∼ 300–600 K). In 2012 May, we had more time available for the project, which was used to confirm a series of lines that initially we assigned to CH₃O and to observe additional lines of this species at 2 mm. All the observations were performed using Frequency Switching mode with a frequency throw of 7.14 MHz, which removes standing waves between the secondary and the receivers. Each frequency setup was observed for ∼2 hr, with pointing checkings in between on strong and nearby quasars. Pointing errors were always within 3''. The 30 m beam sizes at 3 mm are between 30'' and 21'', while at 137 GHz the half-power beam width is 18''. The spectra were calibrated in antenna temperature corrected for atmospheric attenuation using the ATM package (Cernicharo 1985; Pardo et al. 2001).

For some of the spectra presented here, we have also used data from the previous 3 mm survey (Marcelino et al. 2005, 2007, 2009). Errors in the Doppler tracked velocities could be important due to the large bandwidth covered in each observation, however the GILDAS software corrects for this issue (http://www.iram.fr/IRAMFR/GILDAS/).

The electronic ground state of CH₃O is ²E. Its microwave and submillimeter spectrum has been observed by Endo et al. (1984) and Momose et al. (1988) from which accurate rotational constants have been determined. Liu et al. (2009) have discussed the reflection parity assignments for CH₃O based on the analysis of all pure rotational lines and those of the òA₁–$\tilde{X}$²E electronic spectra. They conclude that the parity assignments have to be reversed and they derived a more accurate spin-orbit splitting of the electronic ground state. However, this change does not affect the frequencies of the transitions and energies of the levels of CH₃O associated to the present detection. The quantum numbers indicated in Figure 1, and the rest frequencies correspond to those reported by Endo et al. (1984). They quoted an experimental uncertainty of 30 kHz. The spin-orbit constant of CH₃O is −61.2719 cm⁻¹ (Liu et al. 2009). Consequently, the ²E_1/2 ladder is ∼90 K above the ²E_3/2 one. Taking into account the kinetic temperature of B1-b, T_K ≃ 12–15 K (Marcelino et al. 2005, 2009), only lines from ²E_3/2 could be expected in this source.

Zoom In Zoom Out Reset image size

Figure 1 shows the two components of the N = 1 − 0, K = 0, J = 3/2 − 1/2, F = 2 − 1 transition (top panels, rest frequencies of 82458.252 and 82471.825 MHz, respectively). There are three additional components for this transition (F = 1 − 0, Λ = ±1, and F = 1 − 1) with lower line strength. We have co-added these lines to improve the signal to noise ratio. The resulting spectrum is shown in the left bottom panel of Figure 1 which clearly shows a positive detection. The middle panels show the N = 2 − 1, J = 5/2 − 3/2 strongest hyperfine components. All of them have been detected. The weakest hyperfine components of this transition have also been co-added and the resulting spectrum is shown in the right bottom panel of Figure 1. Taking into account the low line density of B1-b, the number of detected lines of methoxy, and the fact that no lines are missing, we conclude that the discovery of this molecular species is fully secure.

In order to compute the column density of methoxy we have used the energy levels and line strengths from the JPL catalog (Pickett et al. 1998) and included the species in the MADEX code (Cernicharo 2013). We have assumed that the source fills the beam and used a beam efficiency of 0.81 and 0.74 at 3 and 2 mm, respectively. The rotational temperature has been derived from the relative intensities of the different hyperfine components at 3 and 2 mm to be ≃10 ± 3 K. The measured line width is 0.5 km s⁻¹. With these parameters the column density of methoxy is 7 × 10¹¹ cm⁻². For TMC1 we derive a 3σ upper limit to the column density of CH₃O of 10¹² cm⁻².

Most of the proposed species that motivated the present line survey have been detected toward B1-b. Figure 2 shows selected transitions of oxirane (CH₂OCH₂, laboratory data from Pan et al. 1998), methyl mercaptan (CH₃SH; Bettens et al. 1999), dimethyl ether (CH₃OCH₃; Lovas et al. 1979; Neustock et al. 1990), ketene (H₂CCO; Johns et al. 1992 and references therein), formic acid (HCOOH; Cazzoli et al. 2010), propynal (HCCCHO; Jones 1980), acetaldehyde (CH₃CHO; Kleiner et al. 1996), and methyl formate (CH₃OCOH; Carvajal et al., 2007). Strong lines of CH₃OH and other species such as HCO, H₂CO, and H₂CS have been already reported (Marcelino et al. 2005; Marcelino 2007). Except for formic acid and propynal, all the other species reported in this paper are detected for the first time in cold dense cores (see also Marcelino et al. 2007; Öberg et al. 2010; Bacmann et al. 2012). All these species are included in the MADEX code (Cernicharo 2013) which has been used to derive column densities. Assuming that the emission fills the main beam of the telescope and that the excitation temperature is similar to that of methoxy, i.e., T_rot ≃ 10 K we derive (in units of cm⁻²) N(CH₂OCH₂) = 2.0 × 10¹¹, N(CH₃SH A+E) = 1.6 × 10¹², N(CH₃OCH₃) = 3.0 × 10¹², N(H₂CCO) = 2.0 × 10¹², N(HCOOH) = 1.5 × 10¹², N(HCCCOH) = 5.5 × 10¹¹, N(CH₃CHO A+E) = 1.5 × 10¹², N(CH₃OCOH A+E) = 3 × 10¹², and N(HCO) = 2.7 × 10¹². The assumption of a rotational temperature of 10 K is consistent with previous works in which rotational diagrams provide T_rot ≃ 7–10 K (see, e.g., Marcelino et al. 2005, 2009; Lis et al. 2010). Hence, the use of a common rotational temperature for optically thin lines will introduce a maximum uncertainty of 30%–40%. Other potentially interesting species in our initial list such as glycolaldehyde (CH₂OHCHO) and ethanol (CH₃CH₂OH) have not been detected with a 3σ upper limit to their column density of 5 × 10¹¹ and 10¹² cm⁻², respectively. Methyl mercaptan was previously detected only toward clouds in the galactic center (Linke et al. 1979; Nummelin et al. 2000) and toward the hot core G327.3–0.6 (Gibb et al. 2000). Hence, our observation of this species in B1-b represents its first detection in cold molecular clouds. H₂CS has also a large abundance in this source and was studied by Marcelino et al. (2005). Taking into account the large H₂ column density toward B1-b, N(H₂) ≃ 1.5 × 10²³ cm⁻² (see, e.g., Hirano et al. 1999; Lis et al. 2002; Marcelino et al. 2009), all these species have abundances of a few 10⁻¹¹. Among these species only HCO, H₂CCO, CH₃CHO, and HCCCOH have been observed in TMC1 (N. Marcelino et al., in preparation).

Zoom In Zoom Out Reset image size

Laboratory experiments that simulate ice mantle processing by UV or ion irradiation can be used to evaluate the role played by dust grains in the formation of the species detected in the gas phase toward B1-b. A UV irradiation experiment of CO:H₂S = 100:17 ice under ultra-high vacuum conditions (A. Jiménez-Escobar et al. 2012, in preparation), using transmittance infrared spectroscopy to monitor the ice and quadrupole mass spectrometry (QMS) to detect the desorbing species in the gas phase, led to formation of several of the species detected in our observations. CH₃SH (m/z = 48) was detected by QMS during warm-up of the irradiated ice above 130 K, although no infrared absorption of these species was observed in the ice, which could be due to the lower sensitivity of the infrared spectrometer compared to the QMS and the fact that the infrared bands of CH₃SH overlap with those of either CH₃OH or H₂S. A rough estimate based on the QMS signal intensities gives a CH₃SH/CH₃OH production ratio of about 1% in this experiment (A. Jiménez-Escobar et al. 2013, in preparation). HCO and HCOOH, with infrared bands, respectively, at 1853 and 1685 cm⁻¹, were observed in the same experiment during UV irradiation at 8 K, the kinetics of product formation indicate that HCOOH is formed from HCO, a possible formation is via HCO + OH → HCOOH or more likely through CO₂ + 2 H → HCOOH, since CO₂ is one of the most abundant photoproducts in the experiment. Both HCO and HCOOH were previously reported in irradiation experiments of H₂O:CO ice with UV or protons (Allamandola et al. 1988; Hudson & Moore 1999). The presence of H₂CCO and HCCCHO could not be inferred from experimental data, but the mass fragment m/z = 26 (H₂CC or acetyline) was found to co-desorb with CO during warm-up (A. Jiménez-Escobar et al. 2012, in preparation).

Irradiation of CH₃OH ice, either pure or in CH₃OH:H₂S ice mixtures, also leads to the formation of HCO and HCOOH. The co-desorption of m/z = 26 and m/z = 42 supports the formation of H₂CCO desorbing from the UV-irradiated CH₃OH:H₂S ice above 120 K (A. Jiménez-Escobar et al. 2012, in preparation). In addition, CH₃OCH₃ and probably CH₃CHO were also formed in these CH₃OH-containing ices during UV irradiation (Öberg et al. 2009; A. Jiménez-Escobar et al. 2012, in preparation).

The infrared absorptions of the methoxy radical, CH₃O, fall near 1040 and 2820 cm⁻¹, and overlap with the methanol ice bands (Gerakines et al. 1996). Therefore, the CH₃O radical was, to our knowledge, not observed in ice irradiation experiments, but the detection of H₂CO and CH₃OH by UV/ion irradiation of CO:H₂O or CO:H₂S ices suggests that CH₃O and/or CH₂OH are the intermediate species following the reaction H₂CO + H → CH₃O/CH₂OH + H → CH₃OH. The estimated reaction barrier for CH₃O formation from hydrogenation of H₂CO is low, 21.2 kJ mol⁻¹, compared to that of CH₂OH formation via the same process, 48.2 kJ mol⁻¹ (Bennett et al. 2007 and references therein). The CH₃O radical isomerizes to CH₂OH, especially in environments like ice mantles where proton exchange is possible. CH₂OH is formed by irradiation of ice mixtures containing CH₃OH (see, e.g., Gerakines et al. 1996). In cold dark regions, a significant CH₃O/CH₂OH ratio in the gas phase would be indicative of either CH₃O photodesorption or direct formation in the gas phase. Unfortunately, no spectroscopic information is available on the rotational transitions of CH₂OH. The photodesorption of the CH₂OH or CH₃O isomers can be followed by a QMS signal at m/z = 31 in CH₃OH ice irradiation experiments, but a negative result was obtained, indicating no photodesorption (G. A. Cruz-Diaz et al. 2012, in preparation). We therefore conclude that CH₃O in ice mantles preferably isomerizes to CH₂OH or leads to methanol formation. UV photodesorption of CH₃O is negligible, it might be released to the gas phase by ion sputtering but this possibility needs to be confirmed experimentally. Alternatively, CH₃O could be formed in the gas phase by dissociation of CH₃OH if UV photons are available or from reaction of CH₃OH with OH. The last reaction has an activation barrier of 415 ± 100 K (Jiménez et al. 2003) and has been studied between 227 and 360 K. These experiments indicate that the reaction proceeds through H-abstraction from methanol and that the main product will be CH₃O. More recent studies (Shannon 2012) at T_K = 63 K indicate an increase of the rate of the reaction OH+CH₃OH by two orders of magnitude (≃4 × 10⁻¹¹ molecule⁻¹ cm³ s⁻¹) without any pressure dependence. Taking into account the large methanol abundance in B1-b (Öberg et al. 2010), the possibility to form the methoxy radical directly in gas phase has to be considered. Unfortunately, very little information is available on the reactivity at low temperatures of OH with complex organic molecules in the gas phase, in particular down to the 12–15 K of B1-b. Possible feedback from star formation could provide some local warmer environments at T_K ∼ 60 K where such reactions could take place. Observations of NH₃ by Lis et al. (2010) suggests the presence of a compact source of warm gas. Our beam covers all the sources associated to B1-b. High angular resolution observations are needed to confirm the contribution from the different sources. Nevertheless, the role of shocks or outflows is not likely to be important, since they will have a rather low velocity seeing the relatively narrow lines in B1-b. For some of the detected species in B1-b (see Figure 2) we are in a similar situation to that of propylene in TMC1 (Marcelino et al. 2007). Hence, laboratory experiments of abundant gas-phase species, in particular radicals such as OH, with complex organic molecules are needed. It is worth noting that Shannon et al. (2010) have studied the reaction of OH with acetone, methyl ethyl ketone and dimethyl ether at 86 K and found that the rates, with respect those at 295 K, increase by a factor 10–300. Ice mantle chemistry is probably the main production mechanism for methanol and other species. However, the gas-phase chemistry that follows the desorption of these molecules must be studied in detail to have a complete description of the formation paths of complex molecules in cold prestellar cores.

J. Cernicharo and G. Muñoz Caro thank the Spanish MICINN for funding support through grants AYA2009- 07304 and CSD2009-00038. M. Gerin and E. Roueff acknowledge support from the French National PCMI Program.