© ESO 2022

1. Introduction

Ion-neutral reactions, together with neutral-neutral reactions, are thought to be responsible for the chemical synthesis of molecules in cold dense clouds (Agúndez & Wakelam 2013). Ions are therefore key intermediates in the buildup of chemical complexity in these environments. However, among the nearly 130 molecules detected toward TMC-1, the vast majority are neutral species, with only 11% being positive ions and 4% negative ions. Detecting ions has turned out to be a more difficult task than observing neutral species, most likely due to their lower abundances compared to neutrals. However, since many chemical routes involve positive ions, detecting them and constraining their abundances is essential to validate the chemical networks used to model cold dense clouds.

Positive ions detected toward cold dense clouds comprise, in addition to the widespread HCO⁺ and N₂H⁺, HCS⁺ (Thaddeus et al. 1981), HCNH⁺ (Schilke et al. 1991), HC₃NH⁺ (Kawaguchi et al. 1994), HCO (Turner et al. 1999), NH₃D⁺ (Cernicharo et al. 2013), NCCNH⁺ (Agúndez et al. 2015), H₂COH⁺ (Bacmann et al. 2016), NS⁺ (Cernicharo et al. 2018), and H₂NCO⁺ (Marcelino et al. 2018). In the last two years, the list has increased considerably with the discovery of the cations HC₅NH⁺ (Marcelino et al. 2020), HC₃O⁺, HC₃S⁺, CH₃CO⁺, C₃H⁺, C₅H⁺ (Cernicharo et al. 2020a, 2021a,b; Cernicharo et al. 2022), HCCS⁺, and HC₇NH⁺ (Cabezas et al. 2022a,b). Most of the cations observed, in fact all except NS⁺, are the protonated form of abundant neutral molecules.

The chemistry of protonated molecules in cold dense clouds has been described by Agúndez et al. (2015). In a simplified chemical scheme, protonated molecules are mostly formed by proton transfer to the neutral counterpart from abundant proton donors, although ion-neutral reactions different to proton transfer may compete, and they are mainly destroyed through dissociative recombination with electrons. Protonated molecules have abundances in the range 0.01–10% of their corresponding neutral counterpart. In fact, there exists a trend in which the protonated-to-neutral abundance ratio increases with increasing proton affinity of the neutral (Agúndez et al. 2015). Although the chemical composition of cold dense clouds is known to be regulated by chemical kinetics, rather than thermochemical considerations, the proton affinity matters in establishing how abundant a protonated species can be, most probably because a high proton affinity increases the number of possible proton donors and thus the overall protonation rate. It is worth noting that chemical models of cold dense clouds tend to underestimate protonated-to-neutral abundance ratios, which indicates that chemical networks are probably not accurate enough yet.

Here we report the discovery of a new protonated molecule in the cold dense cloud TMC-1, HCCNCH⁺. This ion is the protonated form of HCCNC, a metastable isomer of HC₃N, which is relatively abundant in TMC-1. We also briefly review the situation concerning protonated molecules in cold dense clouds, and discuss the prospects for finding new ions of this type.

2. Observations

The observational data used in this article consist of spectra of TMC-1 taken with the Yebes 40 m telescope and the 30 m telescope of the Institut de Radioastronomie Millimétrique (IRAM). The observed position corresponds to the cyanopolyyne peak of TMC-1, α_J2000 = 4^h41^m41.9^s and δ_J2000 = +25° 41″27.0″.

The Yebes 40 m observations are part of the ongoing line survey of TMC-1 QUIJOTE¹ (Cernicharo et al. 2021c). Observations were carried out during different observing runs between November 2019 and May 2021. The total on-source telescope time is 238 h in each polarization (twice this value after averaging the two polarizations). The QUIJOTE line survey uses a 7 mm receiver covering the Q band (31.0–50.3 GHz) with horizontal and vertical polarizations. Receiver temperatures during 2019 and 2020 varied from 22 K at 32 GHz to 42 K at 50 GHz. In 2021, some power adaptation carried out in the down-conversion chains lowered the receiver temperatures to 16 K at 32 GHz and 25 K at 50 GHz. The backends are fast Fourier transform spectrometers which provide a bandwidth of 8 × 2.5 GHz in each polarization, thus practically covering the whole Q band, with a spectral resolution of 38.15 kHz. The system is described in detail by Tercero et al. (2021). The QUIJOTE observations were performed using the frequency-switching observing mode with frequency throws of 8 or 10 MHz, depending on the observing session. The half power beam width (HPBW) of the Yebes 40 m telescope in the Q band can be fitted as a function of frequency as HPBW(″) = 1764/ν(GHz).

The IRAM 30 m data correspond to observations in the 3 mm band and include data from our 3 mm line survey of TMC-1 (Marcelino et al. 2007; Cernicharo et al. 2012) and data taken in September 2021, which are described in Agúndez et al. (2022) and Cabezas et al. (2022a). Briefly, in the observations carried out in 2021, we used the Eight MIxer Receiver EMIR E090 connected to a fast Fourier transform spectrometer, providing a spectral resolution of 48.84 kHz. We used two frequency setups at slightly different central frequencies to check for spurious signals, image band contamination, and other technical artifacts. The frequency-switching observing mode was used with a frequency throw of 18 MHz. After averaging all IRAM 30 m data, the total on-source integration time was 35.4 h for each polarization. For the IRAM 30m telescope, HPBW(″) = 2460/ν(GHz).

The intensity scale in the Yebes 40 m and IRAM 30 m telescopes is antenna temperature, , which has an estimated uncertainty of 10% and can be converted to main beam brightness temperature, T_mb, by dividing by B_eff/F_eff. For the Yebes 40 m telescope², B_eff can be fitted as a function of frequency as B_eff = 0.738 exp[−(ν(GHz)/72.2)²] and F_eff = 0.97. For the IRAM 30 m telescope³, B_eff = 0.871 exp[−(ν(GHz)/359)²], and F_eff = 0.95 in the 3 mm band. All data were analyzed using the GILDAS software⁴.

3. Results

Among the unidentified lines present in the 7 mm and 3 mm data of TMC-1, there are four lines with peak antenna temperatures between 2 and 4 mK, which have a nearly perfect harmonic relation 4/5/9/10. Two lines lie in the 7 mm band at 37 315 MHz and 46 644 MHz, while the other two lie in the 3 mm band at 83 958 MHz and 93 287 MHz. The lines are shown in Fig. 1 and their accurate measured frequencies are given in Table 1. The line frequencies can be fitted to a rotational constant B = 4664.431891 ± 0.000692 MHz and a centrifugal distortion constant D = 519.14 ± 4.14 Hz, with an rms of 7 kHz. Given the nearly perfect harmonic relation, the probability that the four lines arise from different carriers is negligible. The fit predicts three additional lines in the 3 mm band at 74 630 MHz, 102 615 MHz, and 111 943 MHz. However, the sensitivity of our data at these frequencies is 4.0 mK, 3.5 mK, and 4.8 mK in , that is to say much poorer than for the other two lines in the 3 mm domain.

Fig. 1. Lines of HCCNCH+ observed toward TMC-1. The J = 4–3 and J = 5–4 lines were observed with the Yebes 40 m telescope and the J = 9–8 and J = 10–9 lines with the IRAM 30 m telescope. Unidentified lines are labeled as “U”. The red lines correspond to the line profiles calculated with the LVG method adopting a column density for HCCNCH+ of 3.0 × 1010 cm−2, a kinetic temperature of 10 K, a volume density of H2 of 4 × 104 cm3, a linewidth of 0.60 km s−1, and an emission size of 40″ in radius (see text).

There are various possible carriers for the series of lines observed in TMC-1. The derived rotational constant of 4664.4 MHz is on the order of that of HC₃O⁺, 4460.5 MHz, which has been recently detected toward TMC-1 (Cernicharo et al. 2020a). In that study, calculations at the CCSD(T)-F12/cc-pCVTZ-F12 level of theory were carried out for various species with rotational constants on the order of that of HC₃O⁺. Among these species, HCCNCH⁺ has the closest rotational constant to the one derived here. We note that there is a typographical error in the calculated value of the rotational constant of HCCNCH⁺ given in Table 2 of Cernicharo et al. (2020a). The correct calculated value of B is 4664.74 MHz and not 4646.4 MHz, as confirmed by calculations carried out in this study at the same level of theory. The agreement with the observational value is excellent, with a difference of just 0.007%. The calculated value of the centrifugal distortion constant is 458.3 Hz, which is also in good agreement with the observational value. To provide more refined values of B and D, we used the experimental/calculated ratios obtained for HCCNC, which is adopted as a reference molecule, as scaling factors. The experimental values of B and D for HCCNC are 4967.838144 ± 0.000236 MHz and 626.945 ± 0.191 Hz, respectively (Vigouroux et al. 2000). Calculations at the same level of theory mentioned above yield B = 4967.79 MHz and D = 575.6 Hz for HCCNC. The experimental/calculated ratios are thus 1.00001 for B and 1.08920 for D. Applying these scaling factors to the theoretical constants of HCCNCH⁺ yield B = 4664.79 MHz and D = 499.2 Hz. The rotational constant of HCCNCH⁺ is essentially unchanged after the scaling (the difference is now 0.008%) because calculations match the experimental rotational constant of the reference molecule HCCNC very well. The very good agreement between the theoretical rotational constant of HCCNCH⁺ and the astronomical one provides strong support in favor of the assignment of the series of unidentified lines observed in TMC-1 to HCCNCH⁺. Other plausible candidates are HCNCN⁺, HNCNC⁺, HCNNC⁺, HNCCCH⁺, and HCCCO⁻, but their rotational constants, which were calculated at the same level of theory mentioned above by Cernicharo et al. (2020a), differ from the one observed in TMC-1 by 3–11%. These candidates can be ruled out because the level of calculation used provides errors smaller than 1%. Other plausible candidates containing four heavy atoms (C, N, and/or O) have been discarded as well (see Cernicharo et al. 2020a).

Using the line parameters derived in Table 1 and a dipole moment of 3.45 Debye, as calculated for HCCNCH⁺ by Cernicharo et al. (2020a), we have constructed a rotation diagram, which is shown in Fig. 2. We derived a rotational temperature of 7.2 ± 0.8 K, which is in the range of the rotational temperatures derived in TMC-1, between 5 and 10 K. The column density derived from the rotation diagram is (4.0 ± 1.0) × 10¹⁰ cm⁻². We also carried out excitation calculations using the large velocity gradient (LVG) method with a code written by M. Agúndez (see Agúndez 2009), which is similar to other codes such as MADEX (Cernicharo 2012) and RADEX (Van der Tak et al. 2007). We considered inelastic collisions with H₂ and He, adopting the rate coefficients calculated for collisions between NCCNH⁺ and He (Bop et al. 2018). We adopted an abundance of He of 0.17 relative to H₂ and scaled the rate coefficients calculated for He, when applied to H₂, by multiplying them by the squared ratio of the reduced masses of the H₂ and He colliding systems. We adopted a gas kinetic temperature of 10 K (Fehér et al. 2016) and a line width of 0.60 km s⁻¹, the arithmetic mean of the values derived for the four lines of HCCNCH⁺ in TMC-1 (see Table 1). We assumed that the emission is distributed in the sky as a circle with a radius of 40″, as observed for various hydrocarbons in TMC-1 (Fossé et al. 2001). We found that the best agreement with the observed line intensities is achieved for an H₂ volume density of 4 × 10⁴ cm⁻³ and a HCCNCH⁺ column density of (3.0 ± 0.5) × 10¹⁰ cm⁻². This latter value, which is adopted hereafter, is similar to the column density derived from the rotation diagram. The calculated line profiles using these parameters are shown in Fig. 1. The LVG calculations indicate that the J = 4–3 and J = 5–4 lines have excitation temperatures around 10 K, although the higher J lines, J = 9–8 and J = 10–9, have lower excitation temperatures in the range 6–7 K. These excitation temperatures are in line with the global rotational temperature derived from the rotation diagram, 7.2 K

Fig. 2. Rotation diagram of HCCNCH+ in TMC-1.

4. Discussion

For HCCNC, Cernicharo et al. (2020b) derived a column density of (3.0 ± 0.3) × 10¹² cm⁻². Therefore, the protonated-to-neutral abundance ratio HCCNCH⁺/HCCNC is 0.010 ± 0.002 in TMC-1. To put this result in context, we revisit the observational and theoretical situation of protonated molecules in cold dense clouds, which is summarized in Fig. 3. In this figure, protonated-to-neutral abundance ratios, MH⁺/M, are shown as a function of the proton affinity of the neutral M. Proton affinities were taken from the references listed in the caption of Fig. 3. To estimate the proton affinity of HCCNC, we performed calculations at the CCSD/cc-pVTZ level. This method, which tends to overestimate proton affinities, yields values of 767.8 kJ mol⁻¹ for HC₃N (the experimental value is 751.2 kJ mol⁻¹; Hunter & Lias 1998) and 800.1 kJ mol⁻¹ for HCCNC. Our best estimate for the proton affinity of HCCNC is thus 783 kJ mol⁻¹, which was obtained by scaling the theoretical value by the experimental/calculated ratio found for HC₃N.

Fig. 3. Observed and calculated protonated-to-neutral abundance ratios, MH+/M, in cold dense clouds (mostly in TMC-1) as a function of the proton affinity of the neutral M. The figure is based on Agúndez et al. (2015), with updates for H2COH+ (Bacmann et al. 2016), H2NCO+ (Marcelino et al. 2018), HC5NH+ (Marcelino et al. 2020), HC3O+ (Cernicharo et al. 2020a), HC3S+ (Cernicharo et al. 2021a), CH3CO+ (Cernicharo et al. 2021b), HCCS+ (Cabezas et al. 2022a), HC7NH+ (Cabezas et al. 2022b), and HCCNCH+ (this study). We also include calculated values for C3H+/C3 and C5H+/C5 because these two ions were recently detected toward TMC-1 (Cernicharo et al. 2022).

We compiled observed MH⁺/M ratios from the literature (see references in the caption of Fig. 3). The calculated MH⁺/M ratios in Fig. 3 were obtained from a pseudo time-dependent gas-phase chemical model of a cold dense cloud, similar to those presented by Agúndez et al. (2015), Marcelino et al. (2020), and Cabezas et al. (2022b). In Fig. 3 we plotted the MH⁺/M ratios calculated at steady state. A comparison between observed and modeled abundances indicates that the time at which chemical models best reproduce the chemical composition of TMC-1 is in the range 10⁵–10⁶ yr (the so-called early-time; see e.g., Wakelam et al. 2010; Agúndez & Wakelam 2013). Although fractional abundances relative to H₂ may experience important variations between the early time and the steady state, which is usually reached at much longer times, protonated-to-neutral abundance ratios reach a steady state after some 10⁵ yr (e.g., Agúndez et al. 2015; Cernicharo et al. 2022; Cabezas et al. 2022b). Therefore, when dealing with protonated-to-neutral abundance ratios, as in Fig. 3, it is well justified to adopt calculated values at steady state. The chemical network adopted is largely based on the UMIST RATE12 reaction network (McElroy et al. 2013)⁵. Among the features implemented, we updated the chemistry of HC₇NH⁺ (see Cabezas et al. 2022b), removed the reaction between C₂S and O (see Cernicharo et al. 2021a; Cabezas et al. 2022a), and expanded the network to include the chemistry of HCCNC and its protonated form, HCCNCH⁺, where reaction rate coefficients were mostly taken from the Kinetic Database for Astrochemistry (KIDA) (Wakelam et al. 2015)⁶.

Figure 3 shows that there is a trend in which the protonated-to-neutral abundance ratio MH⁺/M increases for increasing proton affinity of the neutral M. This behavior was already noticed by Agúndez et al. (2015) based on a smaller number of protonated molecules detected in cold dense clouds. The rationale behind this behavior is that since proton transfer to M is one of the main formation routes to MH⁺, the larger the proton affinity of M, the higher the number of available proton donors (which are all protonated species XH⁺ such that X has a lower proton affinity than M), which results in a faster formation rate of MH⁺ and a larger MH⁺/M abundance ratio. However, this behavior is only noticed when looking over a broad range of proton affinities, down to that of N₂, and some caution must be taken because in the region of low proton affinities, most species M are nonpolar and only calculated MH⁺/M ratios are available. If we focus on the right side of Fig. 3, that is for proton affinities larger than 700 kJ mol⁻¹, then we see that there is no clear trend in either the observed MH⁺/M ratios or in the calculated ones. That is, for proton affinities larger than 700 kJ mol⁻¹, observed and calculated MH⁺/M ratios cluster around the region 10⁻³–10⁻¹, with no clear correlation with the proton affinity of M. The case of C₃ is an exception because its protonated form, the ion C₃H⁺, is the only one among the protonated molecules discovered so far in cold dense clouds that reacts rapidly with H₂. This fact makes C₃H⁺ reach a low abundance and thus a low C₃H⁺/C₃ ratio (Cernicharo et al. 2022). It thus seems that there are two main regimes, depending on whether the proton affinity is lower or higher than that of CO. If the proton affinity of a molecule M is lower than that of CO, then the main proton donor forming MH⁺ would be H, while if the proton affinity is higher than that of CO, then HCO⁺, which is the most abundant ion in cold dense clouds, can act as a proton donor, which enhances the formation rate of MH⁺ by orders of magnitude. As a result, molecules M with proton affinities below that of CO reach low MH⁺/M ratios, in the range 10⁻⁶–10⁻⁴, while molecules M with proton affinities higher than that of CO reach substantially higher MH⁺/M ratios, in the range 10⁻³–10⁻¹.

It is worth noting that the chemical model systematically underestimates the MH⁺/M abundance ratio for all species considered in Fig. 3. To get insight into this systematic deviation, we must have a look at the chemistry of protonated molecules in cold dense clouds. As discussed by Agúndez et al. (2015), in a simplified chemical scheme, a protonated molecule MH⁺ is mainly formed by proton transfer to M from the most abundant proton donors, while it is mainly destroyed through dissociative recombination with electrons. Some ions MH⁺ may be efficiently formed by ion-neutral reactions different to proton transfer, and this increases the MH⁺/M ratio over the value expected under the simplified chemical scheme in which MH⁺ formation is dominated by proton transfer. It is likely that the lack of efficient ion-neutral reactions forming MH⁺, other than proton transfer, in the chemical network causes the underestimation of MH⁺/M ratios. The difference between observed and calculated MH⁺/M ratios is most noticeable for HCS⁺ and HC₃O⁺, suggesting that there are important gaps in the chemistry of these ions in current chemical networks. The disagreement is less severe for other species such as H₂NCO⁺ and HC₇NH⁺.

The ion presented here, HCCNCH⁺, shows the smallest discrepancy between the calculated and observed MH⁺/M ratio. The chemistry of HCCNCH⁺ in the chemical network is relatively simple, but subject to important uncertainties. According to the chemical model, the main formation reaction of HCCNCH⁺ is the ion-neutral reaction between C⁺ and CH₃CN. This reaction has been measured to be rapid (the rate coefficient at 300 K is 5.6 × 10⁻⁹ cm³ s⁻¹ and it is likely larger at 10 K), although information of the products is not available (Anicich 2003). The KIDA database assumes that there are two channels that yield C₂H + CN and HCCNCH⁺ + H with equal branching ratios. Interestingly, in the chemical model, this reaction is far more efficient than the proton transfer to HCCNC from abundant proton donors such as HCO⁺ and H₃O⁺, and this is behind the relatively high MH⁺/M ratio calculated, which happens to coincide with the observed one well. Although it is uncertain whether HCCNCH⁺ is truly produced in the reaction C⁺ + CH₃CN, an interesting lesson is that for the protonated molecule with the smallest disagreement between the chemical model and observations, there exists an efficient ion-neutral route to the ion, which is different from proton transfer to the neutral counterpart. This is in line with the argument given above.

In the last years, there has been a sharp increase in the number of protonated molecules detected in cold dense clouds. Most of them are the protonated form of neutral species, which have an electronic closed shell and are the most stable isomer. However, recently, Cabezas et al. (2022a) detected a protonated radical, HCCS⁺, for the first time and here we report the first detection of a protonated metastable isomer, HCCNCH⁺. It is likely that in the near future, more protonated molecules will be discovered in cold dense clouds such as TMC-1. There are several neutral molecules which are known or expected to be abundant and whose protonated forms are good candidates for detection. Among them, we can mention the carbenes H₂C₃ and H₂C₄; various nonpolar or weakly polar hydrocarbons such as C₂H₂, CH₃CCH (and its isomer CH₂CCH₂), CH₂CHCH₃, CH₂CHCCH, and the cycles c-C₅H₆ and c-C₉H₈; nitriles such as CH₃CN, CH₂CHCN, and HC₉N; O-bearing molecules such as CH₃OH, CH₃CHO, and HC₅O; and S-bearing molecules such as H₂CS, OCS, SO, and SO₂. All of them are abundant in TMC-1 and their proton affinities, when available, are above that of CO, which suggests that MH⁺/M ratios could be in the range 10⁻³–10⁻¹. However, for many of them, the rotational spectrum has not been characterized in the laboratory.

5. Conclusions

We reported the first identification in space of the ion HCCNCH⁺ based on observations of TMC-1 and high level ab initio calculations. We derived a HCCNCH⁺/HCCNC abundance ratio of 0.010 ± 0.002, which is in the range of protonated-to-neutral abundance ratios found in cold dense clouds for neutral molecules with proton affinities above 700 kJ mol⁻¹. A state-of-the-art chemical model reproduces the observed HCCNCH⁺/HCCNC ratio well, although important discrepancies remain for other protonated molecules. Other neutral molecules known or expected to be abundant in cold dense clouds are likely to have relatively abundant protonated counterparts, which await detection through sensitive radioastronomical observations.

Acknowledgments

We acknowledge funding support from Spanish Ministerio de Ciencia e Innovación through grants PID2019-106110GB-I00, PID2019-107115GB-C21, and PID2019-106235GB-I00 and from the European Research Council (ERC Grant 610256: NANOCOSMOS). We thank the referee for a report that helped to clarify some points.

References

All Tables

All Figures

Fig. 1. Lines of HCCNCH+ observed toward TMC-1. The J = 4–3 and J = 5–4 lines were observed with the Yebes 40 m telescope and the J = 9–8 and J = 10–9 lines with the IRAM 30 m telescope. Unidentified lines are labeled as “U”. The red lines correspond to the line profiles calculated with the LVG method adopting a column density for HCCNCH+ of 3.0 × 1010 cm−2, a kinetic temperature of 10 K, a volume density of H2 of 4 × 104 cm3, a linewidth of 0.60 km s−1, and an emission size of 40″ in radius (see text).

In the text

	Fig. 2. Rotation diagram of HCCNCH+ in TMC-1.
In the text

Fig. 3. Observed and calculated protonated-to-neutral abundance ratios, MH+/M, in cold dense clouds (mostly in TMC-1) as a function of the proton affinity of the neutral M. The figure is based on Agúndez et al. (2015), with updates for H2COH+ (Bacmann et al. 2016), H2NCO+ (Marcelino et al. 2018), HC5NH+ (Marcelino et al. 2020), HC3O+ (Cernicharo et al. 2020a), HC3S+ (Cernicharo et al. 2021a), CH3CO+ (Cernicharo et al. 2021b), HCCS+ (Cabezas et al. 2022a), HC7NH+ (Cabezas et al. 2022b), and HCCNCH+ (this study). We also include calculated values for C3H+/C3 and C5H+/C5 because these two ions were recently detected toward TMC-1 (Cernicharo et al. 2022).

In the text