© ESO, 2015

1. Introduction

The search for the cyanomethyl radical has been triggered by the detection of many unidentified transitions in the spectrum of the star-forming region SgrB2 (Cummins et al. 1986; Turner et al. 1975). CH₂CN has then been unambiguously detected a few years later for the first time in space towards the Sagittarius B2 star-forming region and the TMC-1 dark cloud by observation of rotational transitions by Irvine et al. (1988) using the spectroscopic laboratory measurements from Saito et al. (1988). Following these detections, the ground-state of the CH₂CN species has been characterised spectroscopically with the determination of the fine and hyperfine coupling constants of the molecule (Saito & Yamamoto 1997). Based on these calculations, Kaifu et al. (2004) identified many weaker lines in the TMC-1 dark cloud that were not reported in Irvine et al. (1988). Recent measurements have been performed by Ozeki et al. (2004) with the microwave spectrum of the cyanomethyl radical (CH₂CN) in its ground electronic state (X²B₁) observed for the lowest three rotational transitions (N = 1–0 to 3–2) with a precise frequency accuracy of several kHz. As part of the IRAM-30 m Large Program ASAI¹ (Lefloch et al., in prep.), we have carried out a highly sensitive, unbiased spectral survey of the molecular emission of the L1544 prestellar core with the highest spectral resolution permitted by the IRAM 30 m FTS broad-band spectrometers.

In the present study we report on the detection of the CH₂CN molecule in the L1544 prestellar core, which was found to agree very well with our computations of the molecular fine and hyperfine structure components between 99 and 145 GHz. This is the first detection in a prototypical prestellar core and also the first detection of the hyperfine structure at this frequency.

2. Observations

The observations were performed at the IRAM-30 m toward L1544 (α₂₀₀₀ = 05^h04^m17.21^s,δ₂₀₀₀ = 25°10′42.8″) using the broad-band receiver EMIR at 3 mm, connected to an FTS spectrometer in its 50 kHz resolution mode (see Vastel et al. 2014, for more details). Figures 1 and 2 show the detected transitions of the ortho and para forms of CH₂CN using the computed frequencies described in Sect. 3. The sensitivity of the observations is high enough that most of the hyperfine components are detected. We also detected several transitions of the para form in the range 101 520−101 535 MHz, although they are much weaker.

3. Calculation of line strengths

The hyperfine-resolved transition frequencies of CH₂CN are partly reported for transitions below the 60 GHz region by Ozeki et al. (2004). However, they are not available at higher frequencies. Then, we calculated them by using the molecular constants reported by Ozeki and collaborators. Since CH₂CN has a pair of identical hydrogen nuclei by symmetry, the spin states are classified into ortho and para states. As in the case of H₂, the para state has the total hydrogen nuclear spin of 0, while the ortho state has the total hydrogen nuclear spin of 1. The symmetry of the ground electronic state of CH₂CN is ${}^{\mathrm{2}}\mathit{B}_{\mathrm{1}}$, which is antisymmetric with respect to 180 degree rotation about the a-axis (the CCN axis). Hence, the nuclear spin state of I = 1 couples to rotational states with K_a = 0 (K_a is a projected quantum number of the end-over-end rotation, N), whereas the nuclear spin state of I = 0 couples to rotational states with K_a = 1. This situation is opposite to the case of H₂CO, where the ortho and para states correspond to the K_a = 1 and K_a = 0 states, respectively.

To calculate line strengths of hyperfine-resolved rotational transitions, we used the following coupling scheme of angular momenta: – J = N + S (S denotes the electron spin (=1/2)) – F₁ = J + I(N) (I(N) is the nuclear spin of the nitrogen nucleus (=1)) – F = F₁ + I(H) (I(H) is the nuclear spin of the hydrogen nucleus (=1)) for the ortho state, and there are six quantum numbers specifying each hyperfine level, F, F₁, J, N, K_a, K_c, where K_c is projection of N on the c axes. On the other hand, the coupling scheme is – J = N + S – F = J + I(N) for the para state because I(H) = 0. Hence, there are five quantum numbers, F, J, N, K_a, and K_c. With these basis sets, we set up the Hamiltonian matrices and obtained energy levels by direct diagonalisation of the matrices. We calculated the transition frequencies and the line strengths for a few low-N transitions with astronomical interests, as shown in Table 1. In calculations of line intensity, we rigorously considered mixing of wave functions used for the basis sets. Because of the relatively strong hyperfine interactions and the relatively weak spin-rotation interaction, the coupling scheme obtained above is only a poor approximation. Hence, considering the mixing of wave functions is indispensable for evaluating the line strengths.

4. Physical conditions and abundances

After computing the molecule frequencies, line intensities, and level energies, we implemented the results in the CASSIS² private database to perform a simple LTE modelling, considering both ortho and para states. L1544 is a prototypical prestellar core in the Taurus molecular cloud complex (d ~ 140 pc) on the verge of gravitational collapse (Caselli et al. 2012, and references within). The core is characterised by a low temperature (~7 K) and a high central density (higher than 10⁶ cm^-3) in its centre. We note that the other detection of CH₂CN in a starless core was performed in the TMC-1 quiet dark cloud, which is characterised by a constant temperature of 10 K, and density of 2 × 10⁴ cm^-3, and shows no signs of gravitational collapse. This stage represents an earlier phase of prestellar core formation. The high CO depletion observed in L1544 at the dust peak emission leads to a high deuterium fractionation (e.g. Vastel et al. 2006). Following the recent analysis by Vastel et al. (2014), we used a value of 10 K for the excitation temperature to estimate the column density for both forms. In Figs. 1 and 2, we present the results from the LTE radiative transfer modelling for the lowest energy levels transitions in which the lines have been detected. The spectra are perfectly reproduced using a column density of the ortho and para form of 3.5 × 10¹¹ cm^-2 for an excitation temperature of 10 K, using a full width at half maximum of 0.3 km s^-1 at a systemic velocity of 7.25 km s^-1. The hyperfine structure for the ortho form is clearly identified. However, the hyperfine structure of the para form is not clearly detected, therefore we conclude that the observations at 99.7 and 101.5 GHz are compatible with an ortho-to-para (OPR) ratio of unity. The line at 101.53 GHz in Fig. 2 is detected through the blending of three nearby hyperfine structure transitions at 101.532089, 101.532038, and 101.532017 GHz (see Table 1).

An uncertainty on the column density and therefore abundance has to be considered using excitation temperatures between 5 and 15 K. Table 2 shows the column density for both ortho and para forms from a best-fit analysis, as well as the OPR.

Fig. 1 Observation of the ortho-CH2CN transitions around 100.6 GHz (in black) superposed with the LTE modelling in red. The upper red lines correspond to the frequencies in Table 1 shifted by 7.25 km s-1.

Fig. 2 Observation of the para-CH2CN transitions around 101.5 GHz (in black) superposed with the LTE modelling in red assuming OPR = 1 with a best-fit column density from ortho-CH2CN. The upper red lines correspond to the frequencies in Table 1 shifted by 7.25 km s-1.

5. Chemical modelling

The dissociative recombination (DR) of the ion CH₃CNH⁺ is the main pathway to form CH₂CN: $\mathrm{C}{\mathrm{H}}^{\mathrm{+}}\stackrel{{\mathrm{H}}_{\mathrm{2}}}{\mathrm{-}0\mathrm{\to }}\mathrm{C}{\mathrm{H}}_{\mathrm{2}}^{\mathrm{+}}\stackrel{{\mathrm{H}}_{\mathrm{2}}}{\mathrm{-}0\mathrm{\to }}\mathrm{C}{\mathrm{H}}_{\mathrm{3}}^{\mathrm{+}}\stackrel{\mathrm{HCN}}{\mathrm{-}0\mathrm{\to }}\mathrm{C}{\mathrm{H}}_{\mathrm{3}}\mathrm{CN}{\mathrm{H}}^{\mathrm{+}}\stackrel{{\mathrm{e}}^{\mathrm{-}}}{\mathrm{-}0\mathrm{\to }}\mathrm{C}{\mathrm{H}}_{\mathrm{2}}\mathrm{CN}\mathit{,}$(1)and the photodissociation by interstellar UV photons is the main destruction process. We used the Nahoon³ gas-phase chemical model (Wakelam et al. 2015), which computes the chemical evolution of a species as a function of time for a fixed temperature and density. The chemical network kida.uva.2014 contains 6992 unique chemical reactions and in total 7506 rate coefficients, and these reactions involve 489 different species. We do not pretend here to describe the situation with a self-consistent model, but to understand the influence of the C/O ratio needed to reproduce the observed CH₂CN. We ran different models with different initial parameters:

• Model 1: for the first step we adopted an elemental abundance for carbon of 5 × 10^-5 and C/O = 0.5, a cosmic ionisation rate of 3 × 10^-17 s^-1, a density of 2 × 10⁴ cm^-3, and a temperature of 10 K and A_v = 10 and let the chemical composition evolve until steady state was reached. The abundances from this first step were then used as initial abundances for the second step, where the density is 4 × 10⁶ cm^-3 with an A_v of 30, as in Vastel et al. (2014) (Fig. 3, plain line).

• Model 2: same as model 1 with C/O = 0.7 (Fig. 3, dot-dashed line).

• Model 3: same as model 1 with C/O = 1 (Fig. 3, three-dots-dashed line).

• Model 4: same as model 1 with C/O = 1.2 (Fig. 3, dashed line).

• Model 5: only one step with an elemental abundance for carbon of 5 × 10^-5 and C/O = 1, a cosmic ionisation rate of 3 × 10^-17 s^-1, a density of 4 × 10⁶ cm^-3, and a temperature of 10 K and A_v = 30 (Fig. 3, red line).

• Model 6: same as model 5 with C/O = 0.5 (Fig. 3, pink line).

The total column density of H₂ has been estimated in L1544 to be about 1.5 × 10²³ cm^-2. If we consider that CH₂CN is emitted in the whole region, its abundance varies between 1.9 × 10^-12 (at 15 K) and 1.6 × 10^-10 (at 5 K). The chemical modelling is clearly dependant on the C/O ratio. When steady state is reached, the abundance is consistent with the computed abundance for an excitation temperature of 10–15 K for models 3 and 5. From the chemical reactions leading to the production of CH₂CN, we can also assume that the emission comes from the border of the L1544 core. Therefore, we must take into account a much lower H₂ column density. A column density of 5 × 10²¹ cm^-2, probing a region where CH₃OH is emitted (and most likely other COMs and C₃O; see Vastel et al. 2014), would lead to a higher abundance of 6 × 10^-11 at 15 K. Even this higher value is significantly lower than the abundance found in the TMC-1 dark cloud (~10^-9: Irvine et al. 1988 using the dipole moment calculated by Ozeki et al. 2004). It is linked to the chemical evolution from a typical dark cloud where carbon chain molecules are much more abundant than in the more evolved prestellar cores. CH₂CN and CH₃CN, which are assumed to form through the same chemical reactions (Eq. (1)), might then be good indicators of the chemical evolutionary stage of dark cloud cores (Vastel et al., in prep.). More precise computation need to be performed with a non-LTE modelling when the collision coefficients are available for

temperatures below 20 K, and a study of the spatial extent of CH₂CN is needed to understand where it forms.

Fig. 3 Chemical modelling of the CH2CN species using the most recent chemical network kida.uva.2014 (see text for the variation of parameters).

6. Conclusions

We conducted an unbiased spectral survey in the 3 mm band at the IRAM 30 m telescope within the Large Programme ASAI, towards the L1544 prestellar core. We detected the CH₂CN species at ~101 GHz with a complex structure that we attributed to an hyperfine structure. Following this detection, we computed the spectroscopic parameters for the lower energy transitions of the ortho and para transitions. This led to the first detection of the fine and hyperfine structure of the CH₂CN species in a prototypical prestellar core. It is the first detection of the hyperfine structure of the cyanomethyl radical at 3 mm, and the first detection of the hyperfine structure in its para form. We estimated the column density and abundance for both ortho and para CH₂CN and compared the observations with a gas-phase chemical modelling. We found that abundances between 10^-12 and 10^-11 are compatible with the observations for excitation temperatures between 10 and 15 K, much lower than the abundance found in the dark cloud TMC-1.

References

All Tables

All Figures

	Fig. 1 Observation of the ortho-CH2CN transitions around 100.6 GHz (in black) superposed with the LTE modelling in red. The upper red lines correspond to the frequencies in Table 1 shifted by 7.25 km s-1.
In the text

	Fig. 2 Observation of the para-CH2CN transitions around 101.5 GHz (in black) superposed with the LTE modelling in red assuming OPR = 1 with a best-fit column density from ortho-CH2CN. The upper red lines correspond to the frequencies in Table 1 shifted by 7.25 km s-1.
In the text

	Fig. 3 Chemical modelling of the CH2CN species using the most recent chemical network kida.uva.2014 (see text for the variation of parameters).
In the text