© ESO 2021

1 Introduction

Chemical studies towards protostars are important because they provide a useful tool for investigating the physical processes involved in the formation of massive and low-mass stars (see Gerner et al. 2014 and Jørgensen et al. 2004, respectively). In particular, the evolution of high-mass star-forming regions is accompanied by an increase in the chemical complexity of the molecular environment (e.g. Tan et al. 2014; van Dishoeck & Blake 1998). Thus high-mass star-forming regions are a very suitable laboratory to study astrochemistry and particularly the formation of complex molecules (Coletta et al. 2020).

The cyano radical (CN), one of the first detected interstellar molecular species (Adams 1941; McKellar 1940), is a key molecule in many astrochemical chains. For instance, it is related to the chemistry of HCN and HNC, which are ubiquitous species in different interstellar environments (Loison et al. 2014). Given that CN is very reactive with molecules possessing C=C double and C ≡C triple bonds, it is involved in the formation of cyanopolyynes (Gans et al. 2017), molecules observed in high-mass star-forming regions and hot cores, and proposed as ‘chemical clocks’ in determining the age of this kind of source (Taniguchi et al. 2018; Chapman et al. 2009).

From interferometric observations, CN emission has been used to study protoplaneatry discs (Artur de la Villarmois et al. 2019; van Terwisga et al. 2019; Öberg et al. 2011) and some hot molecular cores (Qiu et al. 2012; Zapata et al. 2008). Beuther et al. (2004), based on the investigation of two massive star-forming regions, pointed out that CN does not trace the central molecular condensations but mainly gas in their near vicinity. The authors suggested that CN does not appear well suited for disk studies in massive protostars as in the case of low-mass stars, and thus they recommended the use of different molecules. Additionally, it was proposed that CN probes material in the boundary between the bulk protostellar envelope and its outflow cavity (Jørgensen et al. 2004). Besides some of the mentioned works, in the literature there are few works regarding the detection of CN towards massive protostars or high-mass star-forming regions using interferometric observations. Han et al. (2015) presented a study of CN, HCN, and HNC towards 38 high-mass star-forming regions, which included high-mass starless cores (HMSC), high-mass protostellar objects (HMPO), and HII regions, but using single-dish observations. Hence, we point out that it is necessary to study CN emission with high-angular resolution towards a sample of massive protostars with the aim of unveiling the spatial distribution at the small scale of the emission of this molecular species in relation to the star-formingprocesses. The Atacama Large Millimeter Array (ALMA) database offers the opportunity to perform such a study.

According to the National Institute of Standards and Technology (NIST) database¹, interstellar CN has a strong emission line at the rest frequency 226 874.764 MHz, corresponding to the N = 2–1 J = 5/2–3/2 F = 7/2–5/2 transition, which can be blended with the F = 5/2–3/2 and F = 3/2–1/2 lines (at 226 874.183 and 226 875.896 MHz, respectively). Therefore, we searched for observing projects in the ALMA database regarding high-mass star-forming regions observed at Band 6. In the following section we describe the data used in this work and the source selection.

2 Data and source selection

Data were obtained from the ALMA Science Archive². We used data from the project entitled ‘Tracing the evolution of massive protostars’ (Code: 2015.1.01312.S, PI: G. Fuller), which was observed in the ALMA Cycle 3 in configurations C36-2 and C36-3 in the 12 m array, at Band 6. We used the calibrated data that passed the second level of Quality Assurance (QA2). The theoretical maximum recoverable scale is about 6 arcsec. The angular resolution of this data set is about 0.′′ 7, with an almost circular beam of 0.′′ 6 × 0.′′8 (PA = 82. °9) and a spectral resolution of 1.13 MHz. The Common Astronomy Software Applications (CASA) was used to handle and analyse the data. The task imcontsub was used to subtract the continuum from the spectral lines and a first order polynomial was used. The 1σ rms noise level is about 1.5 mJy beam⁻¹ for both the continuum and the line emission.

We note that even though the data passed the QA2 quality level, which assures a reliable calibration for ‘science ready’ data, the automatic pipeline imaging process may give rise to a clean image with some problems or artefacts. For example, an inappropriate setting of the parameters of the clean task in CASA could generate artificial dips in the spectra. Therefore we reprocessed the raw data set of sources whose spectra exhibit conspicuous dips in or close to the main component of the CN emission to check whether the imaging process is responsible for such spectral features. Particular care was taken with the clean task. The images obtained from our data reprocessing are very similar to those obtained from the archive, and hence we conclude that such absorption features are not due to the clean.

The ALMA programme 2015.1.01312.S observed several young, high-mass, embedded sources based on previous surveys of the Galactic plane that have identified large populations of candidate young massive protostars. From these sources, we selected a sample that have been observed in the spectral window in which the above mentioned CN line lies. Additionally, we selected sources whose spectra present conspicuous lines that allow us to properly identify the CN emission. In Table 1 we present the sources, with their velocities, distances, and the spatial resolution of the data in each case. The velocity and distance were obtained from Wienen et al. (2012), who observed ammonia towards cold, high-mass clumps in the inner Galactic disk. Given that G23.389 and G34.821 are not included in Wienen et al.’s survey, the corresponding velocity and distance were obtained from Maud et al. (2015) and Navarete et al. (2019), respectively. Additionally, we used near-IR data at the Ks-band obtained from the UKIRT InfraRed Deep Sky Survey (UKIDSS)³, release DR11, and IRAC-Spitzer data obtained from the Galactic Legacy Infrared Mid-Plane Survey Extraordinaire⁴.

Fig. 1 Average spectra obtained from regions of about 5′′ in radius towards each source with the aim of identifying CN lines. The spectra are presented in rest frequency. We only show the frequency range in which transitions of CN appear. The vertical lines indicate such transitions: N = 2–1 J = 5/2–3/2 F = 7/2–5/2 (A), N = 2–1 J = 5/2–3/2, F = 3/2–3/2 and F = 1/2–1/2, (B) and (C), respectively.

3 Results

To identify CN lines, from the data cubes we extracted spectra of the averaged emission obtained from regions of about 5′′ in radius centred on each source. They are presented in Fig. 1. The frequency was converted to rest frequency using the velocities listed in Table 1, and we show the frequency range 226 860–226 900 MHz, in which, according to the NIST database, only lines of CN appear, except for one weak line of CH₃OCHO at 226 862.239 MHz. The strongest emission in all cases corresponds to the CN N = 2–1 J = 5/2–3/2 F = 7/2–5/2 transition, which, as mentioned above, may be blended with the weaker F = 5/2–3/2 and F = 3/2–1/2 lines, but the spectral resolution of the data does not allow us to resolve the lines properly. Additionally, the CN lines N = 2–1 J = 5/2–3/2, F = 3/2–3/2, and F = 1/2–1/2 at 226 887.399 and 226 892.151 MHz, respectively, appear clearly in all spectra. In some cases (see spectra of sources G13.656, G20.747, G24.464, and G33.133), the main component probably presents absorption or self-absorption features. This will be discussed below.

To appreciate the spatial distribution of the CN emission, in Fig. 2 we present maps of the CN emission averaged over the frequency interval of the main component. Additionally, the continuum emission at 1.3 mm is displayed in contours. Source G18.825 does not present continuum emission above the noise level. Given that we are presenting the spectrum of the millimeter emission of source G29.862 here, we point out that in a previous work, in which we performed adetailed multiwavelength study of this source (see Areal et al. 2020), the CN line was misinterpreted as CH₃OCHO and an erratum was recently published (Areal et al. 2021).

Table 2 presents the parameters of the cores observed at 1.3 mm, obtained from a two-dimensional Gaussian fitting. The peak intensity and the integrated flux density corrected for primary beam response are presented in Cols. 2 and 3, the source size in arcsec deconvolved from beam is presented in Col. 4, and in Col. 5, a rough estimate of the core size in astronomical units considering the average between the size of the semi-axes presented in Col. 4 and the distances listed in Table 1. The source of the errors in the size is mainly the uncertainties in the distances, which is about 20%.

The near-IR emission is usually useful to identify the central point-like sources of young stellar objects (YSOs) and traces diffuse emission arising mainly from warm dust and scattered light in the material surrounding the protostar(s), particularly in cavities carved out by jets on an infalling envelope of material (e.g. Muzerolle et al. 2013; Bik et al. 2006, 2005). Even though near-IR observations can sometimes miss deeply embedded protostars, the lack of near-IR emission or a point-like source at this wavelength in a molecular core can indicate that it is an starless core, which represents the transition between a diffuse molecular cloud and the next generation of stars to form therein (Schnee et al. 2012). Thus, it is interesting to analyse the Ks-band emission in comparison with the ALMA data with the aim of finding any connection relating the extended CN and near-IR emissions, and near-IR point-like sources with the 1.3 mm continuum cores. This comparison is presented in Fig. 3 for each source. Additionally, in order to make a similar comparison, but on a larger spatial scale, in Fig. 4 we present similar panels to those presented in Fig. 3 but using the 3.6, 4.5, and 8 μm emission observed with Spitzer-IRAC. While in some cases the emission is saturated, this is still useful to analyse the sources and the surrounding medium.

Fig. 2 Maps of the CN emission averaged over the frequency interval of the main component. The unit of the colour bar is Jy beam−1. The blue contours show the continuum emission at 1.3 mm with levels of 0.005, 0.010, and 0.020 Jy beam−1, and the names of the millimetre source used in Table 2 are indicated in each case. The 1σ rms noise level is about 0.0015 Jy beam−1 for both the continuum and the averaged line emission. The horizontal lines at the bottom left corner in each panel represent a size of 0.05 pc for each case.

4 Discussion

The first important result is that by analysing the spatial morphology of the CN emission at each source (see Fig. 2), we can conclude that CN traces both molecular condensations and the diffuse and extended gas surrounding them. It is worth noting that in general these condensations traced by the maximums of the CN emission do not spatially coincide with the peaks of the continuum emission at 1.3 mm, which may have contributions from both ionized gas emission and from the dust (Hernández-Hernández et al. 2014) usually associated with the smallest structures related to star formation. This phenomenon was also observed towards IRAS 20293+3952 by Beuther et al. (2004) using interferometric data of the CN N = 1–0 line with a similar angular resolution to the data presented in this work. The authors found that CN emission avoids the core peaks traced by the millimetre continuum emission and based on the fact that optical depth effects appear to be negligible, they suggest two likely main causes for the lack of CN emission towards the cores: (1) the source is too deeply embedded to excite the surface layers of the potential disc, and (2) the source is at such an early evolutionary stage that it does not generate enough UV photons to produce CN emission. Taking into account that some spectra presented in Fig. 1 probably show absorption or self-absorption features, and considering the works of Qiu et al. (2012) and Zapata et al. (2008) in which CN emission presents kinematical indications of infalling envelopes of gas, we point out that the non-coincidence between CN bulk emission and the cores traced by the continuum deserves deeper analysis (see Sect. 4.1).

From Fig. 3, we observe that in five sources (G20.762, G23.389, G29.862, G30.198, and G34.821) the continuum at 1.3 mm coincides with a point-like source (or almost point-like) at the near-IR, while in the others sources there is no near-IR emission associated with the 1.3 mm continuum. This suggests that our core sample is likely composed of regions at different stages of evolution in star formation: YSOs already emitting at the near-IR, and the ‘IR quiet’ sources representing YSOs at an earlier stage of evolution, or even starless cores (e.g. Motte et al. 2018; Schnee et al. 2012).

The difference in the evolutionary stages of the sources can also be seen in Fig. 4. The IR emission obtained from Spitzer-IRAC shows that the above-mentioned more evolved sources present point-like morphologies in these bands. Moreover, the excess in 8 μm suggests photodissociated gas due to ultraviolet photons from young stars. In the case of G13.656, it is important to note that, as also shown in Fig. 3, there is a point-like source in the field not related to the CN emission and the 1.3 mm continuum source.

It is important to remark that extended CN emission appears in all cases. In the cases in which we observe a point-like source at near-IR related to the 1.3 mm continuum, the related CN should be tracing gas in the near vicinity of the protostar as proposed byBeuther et al. (2004) and Jørgensen et al. (2004), and in the case of the ‘IR quiet’ sources, the CN emission may arise from a molecular core at a pre-stellar phase. Indeed, it can be observed that in regions with near-IR point-like sources associated with the 1.3 mm continuum, in general the morphology of the CN emission appears more localized than in the ‘IR quite’ sources. In any case, we conclude that the presence of CN may be ubiquitous during the different star formation stages.

Fig. 3 Near-IR emission at Ks-band obtained from the UKIDSS database. Yellow contours are the CN averaged emission and doted contours are the continuum emission at 1.3 mm, both as presented in Fig. 2. In all cases the first contour of the averaged CN emission is between 3 and 4σ rms noise level, and the increase is about 0.01 Jy beam−1. The horizontal white lines in each panel represent a size of 0.05 pc for each case.

Fig. 4 Three-colour image presenting the IR emission at 8 (red), 4.5 (green), and 3.6 μm (blue) obtained from Spitzer-IRAC. The contours represent the averaged CN emission as presented in Fig. 3. The fields of view are about 1.5 arcmin for all cases except for G23.389 which is about 3 arcmin.

Fig. 5 Spectrum obtained towards core G13.656-mm1. Vertical dashed lines named A, B, and C indicate the rest frequencies of the CN transitions (see Fig. 1). The emission of CH3OCHO and CH2CHCN is indicated.

4.1 Analysis of sources with spectral absorption features

The spectra obtained towards sources G13.656, G20.747, G24.464, and G33.133 present some absorption features (see Fig. 2) and we decided to analyse them in more detail. Firstly, we note that these sources are those whose continuum cores are not associated with near-IR sources (see Fig. 3) suggesting that, as mention above, they may be cores at a pre-stellar phase. In that sense, the absorption features could be due to high optical depth effects. The cores with IR sources may have lower densities than the ‘IR quiet’ cores because outflow activity could have generated cavities in the molecular envelopes (Wheelwright et al. 2012), and the high optical depth effects affecting the CN emission, if there are any, are not so evident.

In addition, the spectra may present absorption features like regular or inverse P Cygni profiles, indicating expanding gas or that a gaseous envelope is falling into an accreting object as observed in CN transitions by Qiu et al. (2012) and Zapata et al. (2008). If the absorption features appear repeatedly along the observed frequency or velocity range, such absorption could also be occurring in the diffuse gas along the line of sight against a continuum source. Godard et al. (2010) studied this phenomenon towards star-forming regions using lines of HCO⁺, HCN, HNC, and CN from single dish observations.

As mentioned in Sect. 2, we discarded the possibility that such absorption features are due to the imaging process. However, it is important to take into account that they could also be due to instrumental effects. Interferometric emission not combined with single dish data can be affected by the missing flux coming from more extended spatial scales that are filtered out by the interferometer and that cannot be recovered even with the most accurate imaging process. This issue can produce not only missing flux but also such absorption features (e.g. Rodón et al. 2012; Beuther et al. 2004). Therefore, it is necessary to analyse this possibility or at least to take it into account before concluding anything about the nature of the observed spectral absorptions. Thus, for each of the mentioned sources, we analyse the spectra at the positions of the continuum cores.

4.1.1 At core G13.656-mm1

Figure 5 shows the spectrum obtained from a circle of 1.′′ 5 in radius at the position of G13.656-mm1. We observed a very complex spectrum in which a fairly deep and narrow absorption feature related to CN appears at 226 877 MHz. Additionally, we observed emission lines of CH₃OCHO (transitions 20(1,19)–19(1,18)E and A), and according to the NIST database, CH₂CHCN (transition ¹ν₁₁ 24(2,23)–23(2,22)), which are complex molecules usually detected towards hot molecular cores and pre-stellar cores (Jiménez-Serra et al. 2016; Herbst & van Dishoeck 2009; Friedel & Snyder 2008). CH₂CHCN is a complexcyanide whose observation and analysis, among other cyanides, can be used as a chemical clock of hot cores (Allen et al. 2018). Particularly, CN is important for the formation of this molecular species because it results from the reaction between CNand ethylene (Agúndez et al. 2008; Herbst & Leung 1990).

Even though the absorption feature at 226 877 MHZ is narrow, it could be interpreted as a P Cygni profile suggesting expanding gas in this molecular core. However, we cannot discard the possibility that it could be due to the instrumental effect discussed above.

Fig. 6 Spectra obtained towards core G20.747-mm1 and -mm2. Vertical dashed lines named A, B, and C indicate the rest frequencies of the CN transitions (see Fig. 1).

4.1.2 At cores G20.747-mm1/2

Figure 6 displays the spectra obtained towards cores G20.747-mm1 and -mm2 from a circle of 1.′′ 2 in radius. In both spectra, an absorption feature related to the CN main component appears at 226 877 MHz suggesting a P Cygni profile, and in the case of core G20.742-mm2, the CN main component has a dip at about 226 873 MHz, which could be explained as self-absorption suggesting that the emission is optically thick. As in the above case, we cannot discard the possibility that the P Cygni-like feature is due to the mentioned instrumental effect. However, the dip at the CN main component coincides exactly with the systemic velocity of the source, suggesting that it could be due to a high optical depth.

The profile of G20.747-mm1 presents a small absorption at 226 838 MHz, which does not appear in others regions, showing that it is not due to an instrumental effect and it should be real. The frequency of this absorption is close to that of the CH₂CHCN transition mentioned in the case of core G13.656-mm1, and hence we wonder if it may be due to absorption in a molecular envelope containing this molecular species against the bright continuum.

4.1.3 At core G24.464-mm1

The spectrum obtained from a circle of 1.′′3 in radius at the position of core G24.464-mm1 is presented in Fig. 7. It presents a conspicuous absorption dip at 226 874 MHz, very close to the systemic velocity of the source, suggesting that it could be due to high optical depths but missing flux coming from more extended spatial-scales filtered out by the interferometer cannot be discarded.

Fig. 7 Spectrum obtained towards core G24.464-mm1. Vertical dashed lines named A, B, and C indicate the rest frequencies of the CN transitions (see Fig. 1).

4.1.4 At cores G33.133-mm1/2/3

Spectra obtained from a circle of 1′′ in radius towards cores G33.133-mm1, -mm2, and -mm3 are displayed in Fig. 8. All the spectra present absorption at the main CN line. In the cases of G133.133-mm2 and -mm3, it could be due to absorption in a molecular envelope against the bright continuum. Again, missing flux produced in the interferometric observation cannot be discarded in any of these three cores. Emission of CH₃OCHO is evident inG33.133-mm2 and -mm3. From the spectra obtained towards these cores, we can say that it is clear that the non-coincidence between CN maximums and the continuum cores may be due to a combination of physical absorption with the described instrumental issue.

The spectrum of G33.133-mm2 presents a conspicuous absorption at 226 856 MHz, which also appears in the spectrum of G33.133-mm1. This absorption does not seem to be produced by instrumental effects. It is most likely real and we posit the same situation as the absorption found in G20.747-mm1.

4.2 Concluding about the absorption features

In conclusion, in most of the analysed absorption features it is possible that they are produced totally or in part by the missing flux coming from more extended spatial-scales that are filtered out by the interferometer. In the case of the other sources, which are those sources that do not present any absorption feature in the spectra shown in Fig. 1, we inspected the emission at the position of the associated 1.3 mm continuum cores and the sources do not present any important absorption features. It is possible that as these sources could be more evolved, given that they present near-IR point-like sources, the surrounding molecular envelopes at the more extended spatial-scales may be more diluted and/or evacuated due to the action of winds and outflows, and hence, the mentioned instrumental effect and/or the probable optical depth effects are not so significant.

Finally, we found that in cores with the largest integrated flux density at 1.3 mm continuum emission (cores G13.656-mm1, G24.464-mm1, and G33.133-mm2), the deepest absorption features in the whole field of view related to the CN emission positionally coincides with the peaks of the continuum emission. This suggests that high optical depth effects can be combined with the discussed instrumental issue.

Fig. 8 Spectra obtained towards cores G33.133-mm1, -mm2, and-mm3. Vertical dashed lines named A, B, and C indicate the rest frequencies of the CN transitions (see Fig. 1). The emission of CH3OCHO is indicated.

5 Summary and concluding remarks

Given that the CN radical is very reactive with molecules possessing C=C double and C ≡C triple bonds, it is involved in the formation of complex molecules in dark clouds, and in particular in those related to star-forming processes. Hence, the study of its emission at small spatial scales towards massive protostars is important. Using high-angular resolution data we investigated the CN emission in a sample of ten massive young stellar objects located at the first Galactic quadrant, and the main results found are summarized as below.

• (a)

  All analysed sources present emission of CN in the transitions N = 2–1 J = 5/2–3/2 F = 7/2–5/2, F = 3/2–3/2, and F = 1/2–1/2. We observe that CN traces both molecular condensations and the diffuse and extended gas surrounding them.

• (b)

  In general, the molecular condensations traced by the maximums of the CN emission do not spatially coincide with the peaks of the continuum emission at 1.3 mm as found in other works.

• (c)

  Our sample of sources is divided into those that present near-IR emission associated with the continuum at 1.3 mm and those that are ‘IR quiet’ sources, suggesting that they are protostellar objects at different stages of evolution. The CN is present at both, suggesting that this radical may be ubiquitous over the different star formation stages, and hence it may be involved in different chemical reactions occurring over time in the formation of the stars. For instance, CN participates in the formation of vinyl cyanide (CH₂CHCN), a molecule detected in core G13.656-mm1.

Finally, it is worth noting that a molecular line emission analysis at such small spatial scales can be affected by the missing flux coming from more extended spatial scales that are filtered out by the interferometer. This is an important issue because it can introduce artificial absorption features into the spectra that can be misinterpreted. We found that ‘IR-quite’ sources are in general more affected by this issue, and we suggest that in the more evolved sources, the surrounding molecular envelopes at the more extended spatial scales can be more diluted and/or evacuated due to the action of winds and outflows, and therefore this effect is not so significant. We cannot discard high optical depth effects combined with missing flux. Interferometric observations of the analysed CN line in a more compact configuration would be useful to properly evaluate this issue.

Acknowledgements

We thank the anonymous referee for her/his very useful comments. A.M. and M.B.A. are doctoral fellows of CONICET, Argentina. S.P. and M.O. are members of the Carrera del Investigador Científico of CONICET, Argentina. This work is based on the following ALMA data: 2015.1.01312.S. ALMA is a partnership of ESO (representing its member states), NSF (USA) and NINS (Japan), together with NRC (Canada), MOST and ASIAA (Taiwan), and KASI (Republic of Korea), in cooperation with the Republic of Chile. The Joint ALMA Observatory is operated by ESO, AUI/NRAO and NAOJ.

References

All Tables

All Figures

	Fig. 1 Average spectra obtained from regions of about 5′′ in radius towards each source with the aim of identifying CN lines. The spectra are presented in rest frequency. We only show the frequency range in which transitions of CN appear. The vertical lines indicate such transitions: N = 2–1 J = 5/2–3/2 F = 7/2–5/2 (A), N = 2–1 J = 5/2–3/2, F = 3/2–3/2 and F = 1/2–1/2, (B) and (C), respectively.
In the text

Fig. 2 Maps of the CN emission averaged over the frequency interval of the main component. The unit of the colour bar is Jy beam−1. The blue contours show the continuum emission at 1.3 mm with levels of 0.005, 0.010, and 0.020 Jy beam−1, and the names of the millimetre source used in Table 2 are indicated in each case. The 1σ rms noise level is about 0.0015 Jy beam−1 for both the continuum and the averaged line emission. The horizontal lines at the bottom left corner in each panel represent a size of 0.05 pc for each case.

In the text

Fig. 3 Near-IR emission at Ks-band obtained from the UKIDSS database. Yellow contours are the CN averaged emission and doted contours are the continuum emission at 1.3 mm, both as presented in Fig. 2. In all cases the first contour of the averaged CN emission is between 3 and 4σ rms noise level, and the increase is about 0.01 Jy beam−1. The horizontal white lines in each panel represent a size of 0.05 pc for each case.

In the text

	Fig. 4 Three-colour image presenting the IR emission at 8 (red), 4.5 (green), and 3.6 μm (blue) obtained from Spitzer-IRAC. The contours represent the averaged CN emission as presented in Fig. 3. The fields of view are about 1.5 arcmin for all cases except for G23.389 which is about 3 arcmin.
In the text

	Fig. 5 Spectrum obtained towards core G13.656-mm1. Vertical dashed lines named A, B, and C indicate the rest frequencies of the CN transitions (see Fig. 1). The emission of CH3OCHO and CH2CHCN is indicated.
In the text

	Fig. 6 Spectra obtained towards core G20.747-mm1 and -mm2. Vertical dashed lines named A, B, and C indicate the rest frequencies of the CN transitions (see Fig. 1).
In the text

	Fig. 7 Spectrum obtained towards core G24.464-mm1. Vertical dashed lines named A, B, and C indicate the rest frequencies of the CN transitions (see Fig. 1).
In the text

	Fig. 8 Spectra obtained towards cores G33.133-mm1, -mm2, and-mm3. Vertical dashed lines named A, B, and C indicate the rest frequencies of the CN transitions (see Fig. 1). The emission of CH3OCHO is indicated.
In the text