Detection of Interstellar HC₅O in TMC-1 with the Green Bank Telescope

Observations of complex chemistry occurring outside the typical hot core environments are critical for understanding the underlying reaction mechanisms and chemical evolutionary processes at work in the larger interstellar medium (ISM). One of the prototypical sources for such investigations is the cold core TMC-1, which contains a rich chemical inventory distinct from star-forming regions. Indeed, while most hot core sources display a wealth of saturated organic molecules,⁸ such as methanol (CH₃OH), ethanol (CH₃CH₂OH), dimethyl ether (CH₃OCH₃), and ethyl cyanide (CH₃CH₂CN) (Neill et al. 2014), the inventory in TMC-1 is heavily weighted toward unsaturated species such as HC_nN (n = 1–9; Loomis et al. 2015), C_nH (n = 3–6), C₂S, C₃S, and C₃O (Kaifu et al. 2004). Since hydrogenation reactions are much more efficient on grains (e.g., CO → CH₃OH), saturated species will tend to be predominantly found on grain surfaces at the low temperatures within cold cores (≤20 K; Charnley et al. 1995; Garrod 2013). Therefore, gas-phase reactions (such as carbon insertion processes) that proceed rapidly at low temperature and density will tend to dominate the chemistry in these regions, resulting in the efficient production of unsaturated hydrocarbons and other carbon-chain molecules (Herbst & Millar 2008).

Because cold cores such as TMC-1 are at an early stage of stellar evolution, their relatively simple physical history and well-defined conditions allow one to model and test chemical pathways, often in great detail. For species that are observed in both cold and hot cores, it is then possible to study the effects of the vastly differing physical conditions (temperature, density, radiation field, etc.) on the chemical evolution. Despite the relative dearth of complex organic molecules, another important characteristic of cold cores is the ease with which new molecular species can be unambiguously identified, especially at centimeter-wavelengths, due to their cold excitation conditions, narrow linewidths, and uncrowded (∼1 line per 200 km s⁻¹) spectrum. For these reasons, it is not surprising that TMC-1 is one of the most well-studied astrochemical sources outside of hot cores and the carbon-star IRC+10216; several dozen new molecular detections have been reported there over the past several decades (see Kaifu et al. 2004 for an extensive review).

We have recently conducted high-sensitivity observations of TMC-1 in search of a number of new molecules. Here, we report the first of several new detections from this study: the carbon-chain HC₅O radical via the observation of four hyperfine components in the $J=17/2\to 15/2$ rotational transition. The observations are presented in Section 2, a review of the laboratory spectroscopy of HC₅O in Section 3, the results and analysis in Section 4, and a discussion of the astrochemical implications and future study in Section 5.

The observations were conducted over eight observing sessions from 2017 February to 2017 June using the 100 m Robert C. Byrd Green Bank Telescope in Green Bank, WV. The observations of TMC-1 were centered on α(J2000) = 04^h41^m42.5^s, δ(J2000) = 25°41'270. Pointing observations were conducted every hour; the pointing accuracy is estimated to be within $2^{\prime\prime}$. The K-band Focal Plane Array was used with the VEGAS spectrometer backend configured to provide 187.5 MHz total bandwidth in each of 10 spectrometer setups at 1.4 kHz (0.02 km s⁻¹) spectral resolution. These extremely high-resolution observations were necessary to resolve the ∼0.3 km s⁻¹ FWHM spectral features typical of TMC-1 (Kaifu et al. 2004). The total spectral coverage was 1875 MHz in 10 discontinuous 187.5 MHz windows within the range of 18 to 24 GHz.

Observations were conducted in position-switching mode, using a 1° offset throw, with 120 s of integration at each position and between ∼7.5 and 15 hr of total on-source integration depending on the frequency window. The resulting spectra were placed on the atmosphere-corrected T_A* scale (Ulich & Haas 1976). Data reduction was performed using the GBTIDL software package. The spectra were averaged using a weighting scheme that corrects for the measured value of T_sys during each 240 s ON–OFF cycle. The spectra were smoothed to a resolution of 5.7 kHz (0.08 km s⁻¹), sufficient to provide ≥3 points across each 0.3 km s⁻¹ FWHM. A polynomial fit was used to correct for baseline fluctuations. The final rms noise in the HC₅O window examined here was 2.4 mK.

The pure rotational spectrum of HC₅O was precisely measured between 6 and 26 GHz by Mohamed et al. (2005). This radical was produced in an electrical discharge of HC₄H + CO, and its spectrum measured using a Balle-Flygare cavity Fourier-Transform microwave spectrometer (Balle & Flygare 1981). Rest frequencies were determined to better than 1 ppm (2 kHz; 0.03 km s⁻¹ at 20 GHz). The ground electronic state of HC₅O is ${}^{2}{{\rm{\Pi }}}_{r}$ with the ${}^{2}{{\rm{\Pi }}}_{1/2}$ fine structure level lying lowest in energy, many tens of K below the ${}^{2}{{\rm{\Pi }}}_{3/2}$ level. At high spectral resolution, its rotational spectrum displays well-resolved Λ-doubling but more closely spaced hydrogen hyperfine splitting. At the low rotational temperature characteristic of TMC-1, only transitions from the lower ${}^{2}{{\rm{\Pi }}}_{1/2}$ ladder are significantly populated; those that fall within the frequency coverage of the observations are given in Table 1. Of the 10 spectrometer windows used in these observations, only one, covering the range of 21766–21953 MHz, contained HC₅O transitions.

Table 1.  Measured and Observed Frequencies of HC₅O Transitions Covered in This Work as Well as Pertinent Line Parameters

${J}^{\prime }\to {J}^{\prime\prime }$	${F}^{\prime }\to {F}^{\prime\prime }$	e/f	Frequency (meas.)a	Frequency (obs.)b	Diff.	${\rm{\Delta }}{T}_{A}$ c	${\rm{\Delta }}V$ d	${S}_{{ij}}{\mu }^{2}$	Eu
			(MHz)	(MHz)	(kHz)	(mK)	(km s−1)	(Debye2)	(K)
17/2 → 15/2	9 → 8	e	21941.846	21941.848	−2	13.0	0.30	41.7	5.017
	8 → 7	e	21941.977	21941.980	−3	13.6	0.27	37.0	5.018
	9 → 8	f	21945.232	21945.231	1	15.1	0.26	41.7	5.019
	8 → 7	f	21945.370	21945.370	0	11.0	0.33	37.0	5.018

Notes.

^aMohamed et al. (2005); 1σ experimental uncertainty is ∼2 kHz. ^bGaussian fit to line at v_lsr = 5.64 km s⁻¹. 1σ uncertainty from Gaussian fit is ∼0.5 kHz. Given the SNR of the detected lines (∼5.4) and the linewidth, we estimate the uncertainty in the observed line centers to be ∼3.7 kHz. ^cWe estimate a conservative uncertainty of 30% in the overall flux calibration. ^dUncertainty in Gaussian fit is ∼0.01 km s⁻¹.

Download table as:  ASCIITypeset image

We observe emission from four hyperfine components of HC₅O in the $J=17/2\to 15/2$ rotational transition at ${v}_{\mathrm{lsr}}=5.64$ km s⁻¹, typical of molecules in this source (Kaifu et al. 2004); the parameters of the observed lines are given in Table 1 and the spectra toward TMC-1 are shown in Figure 1. Although these lines originate from a single J rotational level, taken together, the Λ-doubling and hydrogen hyperfine splitting provide a unique spectroscopic signature. Given the very low line density of the spectra and the coincidence of the line centers to less than the combined observational and experimental uncertainties, a mis-identification of these lines is extremely unlikely from these data alone.

Zoom In Zoom Out Reset image size

Molecular emission in TMC-1 has been shown to be well-modeled by a single excitation temperature (Remijan et al. 2006), typically ${T}_{\mathrm{ex}}=7\mbox{--}10$ K. Because of the negligible spread (∼0.1%) in upper state energies probed by the observed transitions, we assume ${T}_{\mathrm{ex}}=7$ K for our abundance determination. The column density of HC₅O was determined using the formalism of Hollis et al. (2004), given in Equation (1),

$$\begin{eqnarray}&&{N}_{T}=\displaystyle \frac{{{Qe}}^{{E}_{u}/{T}_{\mathrm{ex}}}}{\tfrac{8{\pi }^{3}}{3k}\nu S{\mu }^{2}}\times \displaystyle \frac{\tfrac{1}{2}\sqrt{\tfrac{\pi }{\mathrm{ln}(2)}}\tfrac{{\rm{\Delta }}{T}_{A}{\rm{\Delta }}V}{{\eta }_{B}}}{1-\tfrac{{e}^{h\nu /{{kT}}_{\mathrm{ex}}}-1}{{e}^{h\nu /{{kT}}_{\mathrm{bg}}}-1}},\end{eqnarray} \tag{ 1 }$$

where N_T is the column density (cm⁻²), E_u is the upper state energy (K), ${\rm{\Delta }}{T}_{A}{\rm{\Delta }}V$ is integrated line intensity (K cm s⁻¹), T_ex is the excitation temperature (K), T_bg is the background continuum temperature (2.7 K), ν is the transition frequency (Hz), S is the intrinsic line strength, ${\mu }^{2}$ is the transition dipole moment squared (Debye²),⁹ and ${\eta }_{B}$ is the beam efficiency (∼0.7 for the GBT at 20 GHz). The rotational partition function, Q, is calculated explicitly by direct summation of states (Q[7 K] = 491). We assume that the source fills the beam (see Loomis et al. 2015 for a detailed discussion).

4.1. HC₅O

4.2. HC₃O, HC₄O, HC₆O, and HC₇O

We have also searched our observations, and those of Kaifu et al. (2004), for other members of the HC_nO (n = 3–7) family of oxygen-terminated hydrocarbon free radicals. At present, our observations only cover transitions of HC₇O, and no detection is seen in individual transitions (Figure 2(a)). If the four Λ-doubling components of the two J transitions covered in our observations are stacked in velocity space using a composite average approach (Kalenskii & Johansson 2010), some tentative indication of HC₇O is seen, suggesting the individual transitions remain at or just below our detection limit (Figure 2(b)).

Zoom In Zoom Out Reset image size

We do not find any definitive evidence in our survey, or that of Kaifu et al. (2004), for signal arising from any of the other HC_nO radicals. Assuming the same ${T}_{\mathrm{ex}}=7$ K as for HC₅O, we have calculated upper limits to the column densities using the strongest transition in the available frequency coverage (Table 2). Mohamed et al. (2005) lists the calculated dipole moments for the HC_nO radicals discussed here, all of which are approximately 2 Debye.

Table 2.  Derived Column Densities and Upper Limits for the HC_nO (n = 3–7) Family Compared to the Predicted Values from the HC_nN Family Abundance Trend (Section 5.1) and from the Chemical Model (Section 5.3)

Molecule	Column Density (1011 cm−2)			Transition	ν (MHz)	TA (mK)	References
	Observed	HCnN	Model
HC3O	$\lt 3$	200	0.5	${N}_{{k}_{a},{k}_{c}}={5}_{\mathrm{0,5}}\to {4}_{\mathrm{0,4}}$, $J=11/2\to 9/2$, $F=6\to 5$	45327.7	10a	(1), (2), (3)
HC4O	$\lt 6$	⋯	220	${N}_{J}={8}_{17/2}\to {7}_{15/2}$, $F=9\to 8$	36494.6	10a	(4)
HC5O	17	17b	16	See Table 1			(5)
HC6O	$\lt 10$	⋯	0.0003	$J=27/2\to 25/2$, $F=14\to 13$, f-parity	22165.2	10a	(5)
HC7O	≤5	10	5	$J=35/2\to 33/2$	19204c	2.4	(5)
				$J=37/2\to 35/2$	20302c	3.7	(5)

Notes.  Details of the transition used for the calculation, the assumed maximum brightness temperature, and the laboratory reference for the spectroscopy are also given.

^aFixed to the observed value. ^bKaifu et al. (2004). Accessible at http://www.cv.nrao.edu/SLiSE. ^cThese transitions display Λ-doubling and hyperfine splitting analogous to HC₅O. For simplicity, a central frequency for these splittings is given here; individual frequencies are available in Mohamed et al. (2005).

References. (1) Cooksy et al. (1992); (2) Chen et al. (1996); (3) Cooksy et al. (1995); (4) Kohguchi et al. (1994); (5) Mohamed et al. (2005).

Download table as:  ASCIITypeset image

The detection of HC₅O, combined with the non-detections of HC₃O and HC₇O, raises the question of whether the abundance and upper limits agree with the chemistry thought to be operative in the region and if these closely related species might be detectable in follow-up observations. To the former point, there are two logical avenues to explore: (1) a direct comparison of the abundance trends of the analogous carbon and nitrogen carbon-chain species in this source (HC_nN and C_nH) to explore if the formation and destruction pathways are perhaps similar and (2) state-of-the-art gas-grain chemical models and reaction networks. We explore each of these below and conclude by discussing the feasibility of future detections of other HC_nO radicals.

5.1. Comparison to HC_nN

5.2. Comparison to C_nH

5.3. Gas-grain Chemical Model

Instead, as an initial attempt to model the chemistry of the HC_nO series with $n=3\mbox{--}7$, we have used the NAUTILUS-1.1 3-phase astrochemical model developed in Bordeaux (Ruaud et al. 2016) with the KIDA 2014 network (Wakelam et al. 2015), which we have updated to include reactions related to these species. In a previous study, Adams et al. (1989) predicted observable abundances of the $4\leqslant n\leqslant 6$ HC_nO radicals, assuming formation via the dissociative recombination reactions

$$\begin{eqnarray}&&{{\rm{H}}}_{2}{{\rm{C}}}_{n}{{\rm{O}}}^{+}+{e}^{-}\to {\mathrm{HC}}_{n}{\rm{O}}+{\rm{H}},\end{eqnarray} \tag{ 2 }$$

$$\begin{eqnarray}&&{{\rm{H}}}_{3}{{\rm{C}}}_{n}{{\rm{O}}}^{+}+{e}^{-}\to {\mathrm{HC}}_{n}{\rm{O}}+{{\rm{H}}}_{2}.\end{eqnarray} \tag{ 3 }$$

Since only H₂C₃O⁺ and H₃C₃O⁺ were included in the KIDA 2014 network, following Adams et al. (1989), we have added radiative association reactions (4) and (5) as formation pathways using the rate coefficients tabulated in Adams et al. (1989) for the 4 ≤ n ≤ 6 HC_nO series and a calculated Langevin rate coefficient for reaction (4) leading to the HC₇O radical. We further updated our chemical network to include the destruction of the HC_nO radicals by photons and ions, with the ion-polar rate coefficients calculated using the Su-Chesnavich capture approach (Woon & Herbst 2009).

$$\begin{eqnarray}&&{{\rm{C}}}_{n-1}{{\rm{H}}}_{2}^{+}+\mathrm{CO}\to {{\rm{H}}}_{2}{{\rm{C}}}_{n}{{\rm{O}}}^{+}+h\nu ,\end{eqnarray} \tag{ 4 }$$

$$\begin{eqnarray}&&{{\rm{C}}}_{n-1}{{\rm{H}}}_{3}^{+}+\mathrm{CO}\to {{\rm{H}}}_{3}{{\rm{C}}}_{n}{{\rm{O}}}^{+}+h\nu .\end{eqnarray} \tag{ 5 }$$

With the updated network, simulations were run using standard TMC-1 conditions (Hincelin et al. 2011), the results of which are shown in Figure 3. We find the best agreement with the observational results at a time of $\sim 2\times {10}^{5}\,\mathrm{years}$, and previous studies have noted that TMC-1 models are in good agreement with observations at around this time (Hincelin et al. 2011; Majumdar et al. 2016). We note that at this cloud age, there is excellent agreement between the observed and theoretical column densities for HC₅O, and values for HC₃O, HC₆O, and HC₇O are also below the upper limits derived in this work. At $\sim 2\times {10}^{5}\,\mathrm{years}$, however, we find that HC₄O is overproduced by a factor of about a few compared to our present upper limit. These preliminary results suggest that the ion-neutral reactions noted in Adams et al. (1989) can lead to significant abundances of HC_nO radicals that are in good agreement with observations at reasonable timescales. The overproduction of HC₄O in our preliminary models illustrates the need to further explore the chemistry of these radicals, particularly the possible importance of radical–radical reactions and of unconsidered destruction pathways, which we do not consider in this work due to the current lack of theoretical studies for those pathways.

Zoom In Zoom Out Reset image size

5.4. Future Directions

We have presented the discovery of the HC₅O radical through the observation of four well-resolved hyperfine components in high-sensitivity observations of TMC-1 with the Green Bank Telescope. A search for other HC_nO radicals ($n=3\mbox{--}7$) resulted in non-detections. A first-look chemical model well reproduces the observed column density of HC₅O and agrees with the upper limits inferred for HC₃O, HC₆O, and HC₇O. HC₄O, however, is overproduced in the model, likely indicating that additional reaction pathways not considered here contribute significantly to that chemistry. A detailed modeling study is now underway to more thoroughly examine these possibilities.

The National Radio Astronomy Observatory is a facility of the National Science Foundation operated under cooperative agreement by Associated Universities, Inc. The Green Bank Observatory is a facility of the National Science Foundation operated under cooperative agreement by Associated Universities, Inc. S.V.K. acknowledges support from Basic Research Program P-7 of the Presidium of the Russian Academy of Sciences. E.H. thanks the National Science Foundation for support of his astrochemistry program. A.M.B. is a Grote Reber Fellow, and support for this work was provided by the NSF through the Grote Reber Fellowship Program administered by Associated Universities, Inc./National Radio Astronomy Observatory. The authors thank the anonymous referee for comments that improved the quality of this manuscript.