Millimeter wave spectroscopy of cyanoketene (NC-CH=C=O) and an observational search in the ISM

Context. Ketene was detected in the interstellar medium (ISM) in 1977. Until now, only one derivative, the ketenyl radical, has been observed in this medium. Due to its large dipole moment value, cynaoketene is one of the best candidates for possible ketene derivative detection. Aims. To date, the measurements of the rotational spectra have been limited to 60GHz. The extrapolation of the prediction in the millimeter wave domain is inaccurate and does not permit an unambiguous detection. Methods. The rotational spectra were re-investigated up to 330GHz. Using the new prediction cyanoketene was sought after in a variety of astronomical sources: NGS63341, SgrB2(N), and ASAI sources. Results. A total of 1594 transitions were newly assigned and ﬁtted together with those from previous studies, reaching quantum numbers up to J = 82 and K a = 24. Watson’s asymmetric top Hamiltonian in the I r representation was used for the analysis; both reductions A and S were tested. Logically, the S reduction gave the best results conﬁrming that the molecule is very close to the prolate limit. Cynaoketene was not found in ISM; upper limits to the column density were derived in each source.


Introduction
Ketene (ethenone, H 2 C=C=O) was detected in the interstellar medium (ISM) more than 43 years ago (Turner 1977), and the only derivative, the ketenyl radical (HC=C=O), was observed 38 years later (Agúndez et al. 2015). In contrast, formaldehyde, a compound containing one less carbon and discovered 8 years before ketene (Snyder et al. 1969), has many derivatives detected in the ISM such as acetaldehyde, propanal, propenal, propynal, cyanoformaldehyde, methyl and ethyl formate, formamide, and acetamide 1 (McGuire 2018). However, it should be noted that many derivatives of ketene are very reactive compounds and have never been synthesized. Consequently, the microwave and millimetric spectra of simple substituted ketenes, such as hydroxy-, amino-, formate-, or ethynylketene, are missing. Ketene and methylketene, a simple substituted derivative, are easily isolated on a preparative scale by flash vacuum thermolysis of the corresponding carboxylic anhydride but are unstable compounds in the condensed phase at a temperature higher than −80 • C. Only very recently has the millimetric spectrum of methylketene been reported (Bermúdez et al. 2018) but this compound has not yet been detected. It is therefore too early to conclude that there are no substituted ketenes in the ISM before searching for several derivatives in Global fit, fitting files, and prediction catalog are only available at the CDS via anonymous ftp to cdsarc.u-strasbg.fr (130.79.128.5) or via http://cdsarc.u-strasbg.fr/viz-bin/ cat/J/A+A/638/A3 1 http://www.astrochymist.org/astrochymist_ism.html this medium. Cyanides represent a ubiquitous group of large astronomical molecules in the ISM, and are some of the most abundant and easily detectable in a wide range of environments. Simple cyanides, such as HCN, are found in nearly all astronomical environments ranging from comets to external galaxies. The cyanoketene, 3-oxo-2-propenenitrile (NCCH=C=O), a six-atom molecule containing only one hydrogen atom, appears to be a good candidate for the ISM. Using lab conditions to reproduce the chemistry on the interstellar grains, cyanoketene was found in the photolysis of cyanoacetylene with an oxygen atom generated from ozone, the photolysis of water with cyanoacetylene (Guennoun et al. 2005) or dicyanoacetylene (Guennoun et al. 2004). Cyanoketene is more thermodynamically stable than isomers such as cyanoethynol (NC-C≡C-OH), cyanooxirene (NC(c-C=CH-O), or 3-imino-1,2-propadien-1one (HN=C=C=C=O) (Maier et al. 1998). Combined with its physical properties and the presence of potential precursors in the ISM push us to record and analyze its millimetric spectrum. Using these species as both physical and chemical diagnostics can provide a better understanding and calibration of the physical environments where these species are found and of the formation of large astronomical molecules in general.

Experimental design
The flash vacuum thermolysis of 2-cyanoacetyl chloride at 750 • C produced cyanoketene with one equivalent of hydrogen chloride (Bock et al. 1981). The gaseous flow leaving the oven was directly introduced into the cell of the spectrometer.
Measurements in the frequency range under investigation (150-330 GHz) were performed using the Lille spectrometer (Zakharenko et al. 2015). The absorption cell was a stainlesssteel tube (6 cm in diameter, 220 cm in length). The sample pressure and temperature during measurements were about 10 Pa and room temperature, and the line width was limited by Doppler broadening. The frequency range 150-330 GHz was covered with various active and passive frequency multipliers from VDI Inc. and an Agilent synthesizer (12.5-18.25 GHz) was used as the source of radiation. Estimated uncertainties for measured line frequencies are 30 kHz and 50 kHz depending on the observed S/N and the frequency range.

Analysis of the spectra
The first paper on the rotational spectra of cyanoketene was published by Hahn et al. (2004). The spectra of the parent molecule were measured in the centimeter-wave domain up to 60 GHz. These latter authors also measured the spectra of five isotopic species and calculated the structure of the molecule. The planarity of the molecule was confirmed and the nuclear quadrupole coupling of 14 N was also measured. Using Stark spectroscopy, Hahn et al. (2004) was able to determine accurate values of the dipole moment (µ a 2.844 D and µ b = 2.112 D), finding that it is usually large due to a CN group. This causes dense and intense rotational spectra in the millimeter and submillimeter wave range.
The analysis was relatively straightforward starting from predictions based on Hahn et al. (2004) parameters. First the most intense transitions, the a R ones, were analyzed and fitted up to 330 GHz. They were shifted by only a few megahertz from initial predictions. Subsequently, b R and b Q lines were sought and included in the fit up to 330 GHz. We used ASFIT (Kisiel 2001) for the fitting, and predictions were made with SPCAT (Pickett 1991). The global fits included the 70 microwave transitions from Hahn et al. (2004), and 1594 from this work. The maximal quantum numbers are J = 82 and K a = 24. We first attempted to use a Waston's A-reduced Hamiltonian for the analysis, however this reduction was unable to robustly treat K a values greater than 18. This is perhaps because the molecule is quite close to the prolate limit case (κ = 0.982), and thus would be better treated by a Waston's S-reduced Hamiltonian. The root mean square for A and S reduction considering the full dataset is respectively 210.5 and 24.5 kHz, using 26 parameters for A reduction and 23 for S reduction. Here we only provide the fitting results using the reduction S. In order to compare our parameters with the ones from Hahn et al. (2004), we refitted these data using reduction S. Results can be found in Table 1. Some of the new measurements are in Table 2. Figure 1 shows an example of an observed spectrum. Higher numbers of lines in the experimental spectrum compared to the predicted one come from excited vibrational states.
The complete version of the global fit Table (S1) is supplied at the CDS. The fitting files useable with SPFIT, namely .lin (S2), .par (S3), which contain the measured lines and the spectroscopic parameters respectively, and the prediction .cat (S4) are also available at the CDS.

Sources
We conducted a search for cyanoketene in a variety of astronomical sources. The observational details of these datasets have all been presented elsewhere (see Table 3); only a brief   description is provided here. Because these are nondetections, source-specific and molecule-specific parameters (e.g., source size, excitation temperature) must be assumed in order to calculate an upper limit. In each case, we used parameters representative of each source and complex molecules in that source from the literature references provided in the respective tables.  Fig. 1. Predicted (in red) and observed (in blue) rotational spectrum of NC-CH=C=O between 236.9 and 241.3 GHz. The low-frequency region of the spectrum shown here is dominated by a series of R-type transitions, J = 44-43, while the high-frequency region is dominated by a series of Q-type transitions, K a = 5-4. The slight inconsistency between predicted and observed spectrum intensities is due to source power and detector sensitivity variations. Notes. (a) Except where noted, the source is assumed to fill the beam. (b) For these interferometric observations, the intensity is given in mJy beam −1 rather than mK. ( †) Taken either as the 1σ rms noise level at the location of the target line, or for line confusion limited spectra, the reported rms noise of the observations. References.   Belloche et al. 2013). We used two publicly available datasets to cover both possible scenarios. The Prebiotic Interstellar Molecular Survey (PRIMOS) observations of the region cover the range from 1 to 50 GHz, and are most sensitive to cold, extended populations (Neill et al. 2012), whereas the IRAM 30 m survey of the region near 3 mm provides sensitivity to the warm, compact components (Belloche et al. 2013).
ASAI sources. The remaining sources are from the publicly available Astrochemical Surveys at IRAM (ASAI) Large Notes. (a) Calculated at the excitation temperature assumed for the source. See Table 3 Table 5. In each panel, the red trace shows the transition simulated using the derived upper limit column density and the physical parameters assumed for that source. The frequency of the transition is given in the upper right corner of each panel, and the quantum numbers for each transition in the upper left. The source name is given above each panel. Due to the large variations between observations, the intensity and velocity axes are not uniform across all panels.
Program. These spectra cover a range of source types from cold, dark clouds to Class 0/1 protostars and shocked outflows. Observational details are given in Lefloch et al. (2018).

Upper-limit analysis and results
We find no definitive evidence for the presence of cyanoketene in any of the sources we have studied. We derive upper limits to the column density in each source using the formalisms outlined in Turner (1991), which assume the molecules are well described by a single excitation temperature and include corrections for optical depth. Frequencies, energy levels, degeneracies, and line strengths were obtained from the laboratory spectroscopy described in this work.
The rotational partition function was obtained and used without modification from SPCAT. In all but the coldest (T ∼ 10 K) sources, the vibrational contribution to the partition function must also be considered. To that end, we carried out quantum chemical calculations using Gaussian 09 (Frisch et al. 2009) at the wB97XD/6-311++G(d,p) level of theory and basis set to determine the harmonic frequencies of cyanoketene. The structure was first optimized, resulting in equilibrium rotational constants (A = 30 101 MHz, B = 2814 MHz, C = 2573 MHz) in good agreement with the experimental values. The harmonic frequencies are given in Table 4.
For each source, a simulated spectrum was generated using the physical conditions for the source and the line parameters measured in this work. The 1σ upper limit to the column density A3, page 4 of 5 was then derived using the line that gave the most rigorous constraint (i.e.s the line that would have the highest signal-to-noise ratio in the event of a detection). These lines (or a representative transition when multiple transitions contributed to a single spectral feature) and their parameters are provided in Table 5 along with the resulting upper limits. The spectra from each dataset, along with a simulated line profile of cyanoketene at the established upper limit, are displayed in Fig. 2.
From a chemical formation point of view, reactions resulting in addition of a CN group (-CN-addition reactions) involving either neutral or radical species are a viable pathway to the formation of larger cyanides (Bennett et al. 2010;Garrod et al. 2017). Remijan et al. (2008) presumed that a chemical formation route to cyanoformaldehyde (CNCHO) through -CN addition to H 2 CO might take place given the high abundances of H 2 CO and -CN in the gas phase. Current studies (Das et al. 2013) have followed up this supposition to measure not only the gas-phase and grain-surface reaction rates for this formation pathway but also possible destruction pathways of NCCHO. The results of these analyses showed that NCCHO could be formed efficiently on the grain surfaces, yet a viable desorption mechanism would need to be present to liberate the abundance of NCCHO into the gas phase -such a mechanism could exist in the form of low-velocity shocks. In this work, it is proposed that cyanoketene (NCCHCO) can be formed in similar reaction pathways to cyanoformaldehyde through the following chemical reactions: 1. H 2 CCO + CN → NCCHCO + H (radical-neutral), 2. HCCO + CN → NCCHCO (radical-radical), 3. C 3 N + H 2 O → NCCHCOH → followed by H-atom dissociation to form NCCHCO. In all cases, no experimental or theoretical values exist for the reaction rates of either of the three pathways. Reaction (3), if it is indeed barrier-less (or has a very low barrier), is more likely to form: 1. C 3 N + H 2 O → HC 3 N + OH (or NCCCOH + H).
As such, additional spectroscopy and astronomical searches are needed for the ethynol radical (HCCOH), a structural isomer of ketene, and its -CN substituted species cyanoethynol (NCC-COH). Of the two remaining reactions, (1) would suffer from the same limitations as the formation of NCCHO, and would most likely be most efficiently formed on grain surfaces. Furthermore, while the radical-radical reaction shown in reaction (2) would most certainly occur, the reaction products would also only survive if it occurs on a grain surface. However, unlike the formation and subsequent desorption of NCCHO into the gasphase, the nondetection of HCNCCO could indicate that: (1) the abundance of NCCHCO on the grain surface is too low to be detectable via current instrumentation when the species is liberated from the grain surface; (2) the desorption mechanisms to liberate the species into the gas phase are either too inefficient or result in the destruction of the molecule; (3) high barriers exist to its formation either in the gas phase or on grain surfaces, and/or (4) effective and efficient gas-phase destruction pathways are present that destroy the molecule quickly after liberation into the gas phase. If this is the case, NCCHCO may only be detectable on grain surfaces, and in any case, further investigation into the formation, destruction, and desorption pathways of NCCHCO is warranted.
The millimeter wave spectra of cyanoketene, a cyano derivative of ketene detected in the ISM, were measured and analyzed. This enabled the accurate prediction of the spectrum up to 400 GHz. It was searched for in major COM sources, but was not detected; upper limits on column density are provided. This work will benefit the astrophysics community for a possible detection in the future.