A search for ortho-benzyne (o-C6H4) in CRL 618
Susanna L. Widicus Weaver, Anthony J. Remijan, Robert J. McMahon, Benjamin J. McCall
aa r X i v : . [ a s t r o - ph ] N ov A search for ortho -benzyne ( o -C H ) in CRL 618 Susanna L. Widicus Weaver
Departments of Chemistry and Astronomy, University of Illinois at Urbana-Champaign,Urbana, IL 61801; [email protected]
Anthony J. Remijan
National Radio Astronomy Observatory, Charlottesville, VA 22903; [email protected]
Robert J. McMahon
Department of Chemistry, University of Wisconsin-Madison, Madison, WI 53706;[email protected] andBenjamin J. McCall
Departments of Chemistry and Astronomy, University of Illinois at Urbana-Champaign,Urbana, IL 61801; [email protected]
ABSTRACT
Polycyclic aromatic hydrocarbons (PAHs) have been proposed as potentialcarriers of the unidentified infrared bands (UIRs) and the diffuse interstellarbands (DIBs). PAHs are not likely to form by gas-phase or solid-state interstellarchemistry, but rather might be produced in the outflows of carbon-rich evolvedstars. PAHs could form from acetylene addition to the phenyl radical (C H ),which is closely chemically related to benzene (C H ) and ortho -benzyne ( o -C H ). To date, circumstellar chemical models have been limited to only a partialtreatment of benzene-related chemistry, and so the expected abundances of thesespecies are unclear. A detection of benzene has been reported in the envelopeof the proto-planetary nebula (PPN) CRL 618, but no other benzene-relatedspecies has been detected in this or any other source. The spectrum of o -C H issignificantly simpler and stronger than that of C H , and so we conducted deepKu-, K- and Q-band searches for o -C H with the Green Bank Telescope. Notransitions were detected, but an upper limit on the column density of 8.4 × cm − has been determined. This limit can be used to constrain chemical modelsof PPNe, and this study illustrates the need for complete revision of these modelsto include the full set of benzene-related chemistry. 2 – Subject headings: astrochemistry circumstellar matter - stars: individual (CRL618) radio lines: stars
1. Introduction
Polycyclic aromatic hydrocarbons (PAHs) are large molecules with carbon atoms ar-ranged in five- or six-membered rings and are thought to form from smaller carbon ringmolecules such as benzene (C H ) and its derivatives. PAHs are very stable against pho-todissociation, and so could be present in interstellar and/or circumstellar environments. Theunidentified infrared bands (UIRs) occur at frequencies characteristic of aromatic molecules,and so PAHs have been suggested as potential UIR carriers. Likewise, PAHs have beenproposed as carriers of the diffuse interstellar bands (DIBs). We refer the reader to a recentreview of the DIB problem by Sarre (2006) and references therein for a discussion of PAHsand their relation to the UIRs and DIBs. Significant laboratory spectroscopic work has beendedicated to PAHs and other benzene-related species to support astronomical observations,yet none of these species has been unambiguously detected in space. Only benzene has beententatively detected toward the proto-planetary nebula (PPN) CRL 618 (Cernicharo et al.2001a). One vibrational band of benzene was observed, and a column density of 5 × cm − and a kinetic temperature of 200 K were determined.CRL 618 is an evolved PPN with a central B0 star and an ultra-compact H II region sur-rounded by a carbon-rich molecular envelope (Cernicharo et al. 2001b; S´anchez Contreras & Sahai2004). This source has optical high velocity bipolar outflows (Trammell 2000) in additionto a low velocity expanding torus of molecular emission (S´anchez Contreras & Sahai 2004).CRL 618 has a rich molecular inventory including a variety of hydrocarbons [see Pardo et al.(2007) and references therein], but searches for biologically important molecules have pro-vided only upper limits (Remijan et al. 2005).Very few observational constraints have been placed on the chemical and physical prop-erties of PPNe, especially regarding benzene-related chemistry. Observational studies of thesimplest benzene derivatives are the essential first steps to understanding the formation ofPAHs in PPNe. The carbon-rich nature of CRL 618, coupled with the benzene detection,makes it an excellent target for initial benzene derivative searches. We have therefore con-ducted a search for the benzene derivative ortho -benzyne ( o -C H ) in CRL 618 with theGreen Bank Telescope (GBT) Ku-, K-, and Q-band receivers. An overview of circumstel-lar benzene-related chemistry, details of the observations, the results of our search, and adiscussion of these results are presented below. 3 –
2. Benzene-Related Circumstellar Chemistry
PAHs have long been thought to be produced in carbon-star outflows, as it seemsunlikely that such large molecules could be produced by gas-phase or grain-assisted chemistrywithin the interstellar medium. Although the PPN phase of stellar evolution is typicallyvery short, lasting only about ten thousand years, the chemistry that takes place in theenvelope of the post-asymptotic giant branch (AGB) star leads to the formation of complexhydrocarbons. PPN chemistry is steeply dependent on density because circumstellar materialundergoes energetic processing, and formation mechanisms for large molecules are likelya combination of radical-molecule and ion-molecule reactions. This is in contrast to lowdensity interstellar chemistry that is almost exclusively driven by ion-molecule reactions.Both radical-molecule and ion-molecule reactions can lead to the formation of benzene andrelated species, and such benzene formation mechanisms have been included in PPN models(Redman et al. 2003; Woods et al. 2003). However, no PPN model includes both theradical- and ion-based benzene reaction networks. This striking deficiency in PPN modelsmakes interpretation of observational results quite difficult, and the radical- and ion-basedchemical mechanisms must be considered separately until more complete PPN models aredeveloped.The torus of a PPN is a high energy, high density environment, and the chemistryis expected to be similar to that observed in combustion (Frenklach & Feigelson 1989).Radical-based mechanisms based on combustion chemistry have been proposed for benzeneand PAH formation in the circumstellar shells of AGB stars (Cherchneff et al. 1992), and asubset of these reactions was also used to model the chemistry of clumps during PN evolution(Redman et al. 2003). A summary of radical-driven benzene chemistry is shown in Figure1. Here, hydrocarbon radicals react with acetylene (C H ), and ring-closure forms o -C H .Additional reactions lead to the phenyl radical (C H ) and benzene. Other routes to benzeneinclude either reaction of two C H radicals or, alternatively, the formation of straight-chainenergetic C H (n-C H ∗ ), which can then undergo ring-closure.As a PPN evolves and its surrounding gas expands, the density decreases and thematerial in the circumstellar envelope is subjected to photoprocessing from the central star,leading to ion-molecule chemistry. An ion-molecule benzene formation mechanism has beenproposed for PPNe (Woods et al. 2003), and a summary of this mechanism is shown inFigure 2. In this network, hydrocarbon ions react with acetylene, and ring closure formsC H +5 . Additional hydrogenation and/or electron recombination leads to o -C H , c-C H +7 ,and ultimately C H . It should be noted that the original PPN model investigated onlythe primary route to benzene shown in Figure 2 and did not include the o -C H formationreaction (Woods et al. 2003). 4 –One of the benzene derivatives formed in these reactions, C H , is a suggested precur-sor to PAHs in circumstellar environments (Frenklach & Feigelson 1989; Cherchneff et al.1992). As is illustrated in Figure 3, C H can undergo a series of radical-molecule reactionsinvolving acetylene to produce a naphthalene-like species (C H ) via ring closure. Subse-quent reactions of this nature can lead to larger PAHs.Studies of other high energy environments indicate that C H and o -C H may wellcoexist with benzene in CRL 618 if circumstellar chemistry is similar to combustion orplasma chemistry. The reaction network of Frenklach & Feigelson (1989) is based on sootproduction mechanisms in hydrocarbon flames, and theoretical studies show that benzeneunimolecular decomposition leads to C H and o -C H during combustion (Mebel et al.2001). Benzene electrical discharges also produce C H and o -C H (McMahon et al. 2003).There are no reliable predictions of C H , C H , and o -C H abundances in PPNe becauseof the partial treatment of their chemistry in models, but observations of C H and o -C H would provide important limits for future modelling studies. We therefore began a searchfor these species in CRL 618.C H has no permanent dipole moment and can therefore only be studied in the in-frared, but both C H and o -C H can be probed by radioastronomical techniques. Therotational spectra of o -C H and C H have been obtained in the laboratory (Brown et al.1986; Robertson et al. 2003; Kukolich et al. 2003; McMahon et al. 2003). Both speciesare carbon ring, planar, near-oblate asymmetric rotors with C v symmetry along their b inertial axes. The calculated dipole moment of o -C H is 1.38 D (Kraka & Cremer 1993)and that of C H is 0.9 D (McMahon et al. 2003). The unpaired electron in C H leads tohyperfine splitting of the lower rotational states (McMahon et al. 2003), and so there aremany rotational transitions to sample observationally. However, the larger dipole momentcoupled with the lack of hyperfine splitting yields much stronger lines for o -C H and makesthis species a more likely candidate for detection if the abundances are similar.
3. Observations
Observations of o -C H were conducted with the NRAO
100 m Robert C. Byrd GBTbetween 2006 September 4 - 2007 January 28 using the Ku-band (12 - 15.4 GHz), K-band(18 - 22.5 and 22 - 26.5 GHz,) and Q-band (40 - 48 GHz) receivers. The eight intermediate-frequency (IF), 200 MHz, three-level GBT spectrometer configuration mode was used, pro- The National Radio Astronomy Observatory is a facility of the National Science Foundation, operatedunder cooperative agreement by Associated Universities, Inc. α =04 h m s .7, δ = +36 o ′ ′′ .0 and -27.5 km s − , respectively. Data were acquired in theOFF-ON position-switching mode. A scan included two-minute integrations for each posi-tion beginning with the OFF-source position, which was located 60 ′ East in azimuth of theON-source position. Antenna temperatures with estimated 20% uncertainties were recordedon the T ∗ A scale (Ulich & Haas 1976). The GBT half-power beamwidths are given by θ b =740 ′′ / ν (GHz). Dynamic pointing and focusing corrections were applied and observations ofthe quasar 0359+509 were used to adjust the zero points every two hours or less. Observa-tions from multiple nights and both polarizations were averaged for each frequency window,and the data were Hanning smoothed over three channels with the GBDish data reductionprogram. The resultant Q-band spectrum is shown in Figure 4.Table 1 lists the o -C H rotational transitions in the observed frequency windows. Thetransition quantum numbers, transition rest frequency ( ν ), the Einstein A coefficient of thetransition times the upper state degeneracy ( Ag u ), the energy of the upper level ( E u ), theobserved 1- σ RMS level ( T MB ), and the beam efficiency ( η ) are listed in the first six columns.The transition frequencies and intensities are from the Cologne Database for MolecularSpectroscopy (M¨uller et al. 2005). The d.rms routine in GBDish was used to calculate theRMS level in the T ∗ A scale for each spectral window. The RMS level is equal to the standarddeviation of the noise in a line-free region and was calculated after the Hanning smoothingwas applied. These values were then converted to the T MB scale by the relationship T MB = T ∗ A / η . The value of η was derived from a fit to the Ruze (1966) formulation, as is outlinedin Equation (2) of Hollis et al. (2007).
4. Results and Discussion
No spectral features associated with o -C H were detected during this search, and thereare no unidentified spectral features in any passband. Figure 4 shows a plot of the observedCRL 618 Q-band spectrum overlaid by a predicted o -C H spectrum at a column density of10 cm − , a rotational temperature of 200 K, and a source size of 10 ′′ . If these parametersare representative of o -C H , the blended 6 , − , and 6 , − , transitions at 40828.1686and 40829.9929 MHz, respectively, should have been easily detected at the 30 mK level.The RMS levels reached during the observations allow calculation of the o -C H columndensity upper limit in CRL 618, and the 3- σ upper limits determined from each observedtransition are presented in Table 1. These column density upper limits were calculated usingthe following expression, adapted from equation (1) of Nummelin et al. (1998): 6 – N T = Z ∞−∞ T b dv πkν hc Q ( T rot ) Ag u e E u /kT rot (1)where N T is the beam averaged total column density, R ∞−∞ T b dv is the transition integratedintensity, Q ( T rot ) is the rotational partition function, and T rot is the molecular rotationaltemperature. Since no lines were observed, the integrated intensity was approximated asthat of a line with a peak intensity at the 3- σ RMS level, T RMS , and an assumed linewidth∆ v . The value of T RMS was determined by T RMS =3 T MB / √ n , where n is the number ofchannels across the linewidth ∆ v . The 1/ √ n factor does not include a complete statisticaltreatment of the assumed Gaussian lineshape, but it does approximate the overestimation ofthe RMS level from spectral oversampling. Examination of previously reported upper limitcalculations indicates that, in nearly all cases, this factor is either neglected entirely, or atthe very least not discussed explicitly. We find this to be a gross omission for instrumentssuch as the GBT spectrometer, where the channel width is often significantly smaller thanthe linewidth. From the T RMS values, correction for beam dilution gave T b through theexpression T b = T RMS /B , where the beam filling factor, B , was calculated from the sourcesize, θ s , and the beam size, θ b , by the relationship B = θ s / [ θ s + θ b ]. The integrated intensitywas then approximated as R ∞−∞ T b dv = 1 . T b ∆ v , where the 1.064 factor arises from theassumed Gaussian lineshape. We have included this factor for completeness, although it islikely minimal compared to the uncertainty in T b .The calculated 3- σ o -C H upper limits for CRL 618 are presented in Table 1. Thesecalculations required assumptions for the values of T rot , ∆ v , and θ s . The linewidth wasassumed to be that observed for other identified lines, 10 km s − . Determination of anappropriate temperature and source size was less straightforward, as there is much discrep-ancy in the literature regarding these values. The kinetic temperature derived for benzeneis 200 K (Cernicharo et al. 2001a), but an IRAM 30-meter millimeter line survey indicatestemperatures of 250 - 275 K for the torus and 30 K for the circumstellar shell (Pardo et al.2007). It is expected that o -C H would have very similar physical properties to those ofbenzene in this source, and so a temperature of 200 K was assumed for the upper limitcalculations.There is similarly contradicting information regarding the CRL 618 source size. Pardo et al.(2007) found torus and circumstellar shell sizes of 1.5 ′′ and 3.0 - 4.5 ′′ , respectively, while mil-limeter interferometric observations with a ∼ ′′ beam gave source sizes ≤ ′′ for moleculesexpected to trace the extended envelope (Remijan et al. 2005). A molecule such as o -C H is likely to be present in the torus where higher densities shield it from photodissociation,but such a species could also be present in more extended regions if readily formed by ion- 7 –molecule chemistry. The GBT beam is ≥ ′′ , and so our observations probed both the torusand the extended envelope of CRL 618. A source size of 10 ′′ was assumed for the upper limitcalculations, as this includes the entire molecular envelope and therefore all possible regionsof emission.As is shown in Table 1, the most strict o -C H column density upper limit determinedfrom these observations is 8.4 × cm − . The observed benzene column density is 5 × cm − (Cernicharo et al. 2001a), and McMahon et al. (2003) report a C H column densityupper limit in CRL 618 of 4 × cm − . The o -C H upper limit is therefore much lowerthan the limits for the related species. Given the incomplete nature of the chemical models,however, the significance of this limit is unclear.Further interpretation of this limit will require additional observational and modellingstudies, as this work highlights the nearly total lack of information regarding the physicaland chemical properties relevant to benzene-related chemistry in circumstellar environments.Interferometric observations would prove quite useful, as such studies would probe the spatialdistribution of molecules in the torus and extended envelope, eliminating the uncertaintyin beam dilution effects. Laboratory measurements of the millimeter and submillimeterspectra of o -C H would aid these observations. We also strongly encourage complete revisionof existing PPN models to investigate both the radical- and ion-based benzene chemicalnetworks for comparison to o -C H , C H , and C H observations. Only a combination offurther observations, modelling, and laboratory measurements of benzene derivatives willlead to full understanding of PAH formation mechanisms in PPNe.We would like to thank the NRAO and the GBT support staff, especially Carl Bignell.Support for SLWW and BJM was provided by the NSF CAREER award (NSF CHE-0449592)and the UIUC Critical Research Initiative program. Support for RJM was provided by NSF-0412707. We are grateful to Matthew Redman and Paul Woods for providing additionaldetails about their chemical models. We would also like to thank Michael Remijan forprogramming support and development. REFERENCES
Brown, R. D., Godfrey, P. D., & Rodler, M. 1986, J. Am. Chem. Soc., 108, 1296Cernicharo, J., Heras, A. M., Tielens, A. G. G. M., Pardo, J. R., Herpin, F., Gu´elin, M., &Waters, L. B. F. M. 2001a, ApJ, 546, L123 8 –Cernicharo, J., Heras, A. M., Pardo, J. R., Tielens, A. G. G. M., Gu´elin, M., Dartois, E.,Neri, R., & Waters, L. B. F. M. 2001b, ApJ, 546, L127Cherchneff, I., Barker, J. R., & Tielens, A. G. G. M. 1992, ApJ, 401, 269Frenklach, M. & Feigelson, E. D. 1989, ApJ, 341, 372Hollis, J. M., Jewell, P. R., Remijan, A. J., & Lovas, F. J. 2007, ApJ, 660, L125Kraka, E., & Cremer, D. 1993, Chem. Phys. Lett., 216, 333Kukolich, S. G., Tanjaroon, C., McCarthy, M. C., & Thaddeus, P. 2003, J. Chem. Phys.,119, 4353McMahon, R. J., McCarthy, M. C., Gottlieb, C. A., Dudek, J. B., Stanton, J. F., & Thad-deus, P. 2003, ApJ, 590, L61Mebel, A. M., Lin, M. C., Chakraborty, D., Park, J., Lin, S. H., & Lee, Y. T. 2001, J. Chem.Phys., 114, 8421M¨uller, H. S. P., Schl¨oder, F., Stutzki, J. & Winnewisser, G. 2005, J. Mol. Struct. 742, 215Nummelin, A., Dickens, J. E., Bergman, P., Hjalmarson, ˚A., Irvine, W. M., Ikeda, M., &Ohishi, M. 1998, A&A, 337, 275Pardo, J. R., Cernicharo, J., Goicoechea, J. R., Gu´elin, M., & Ramos, A. A. 2007, ApJ, 661,250Redman, M. P., Viti, S., Cau, P., & Williams, D. A. 2003 MNRAS, 345, 1291Remijan, A. J., Wyrowski, F., Friedel, D. N., Meier, D. S., & Snyder, L. E. 2005, ApJ, 626,233Robertson, E. G., Godfrey, P. D., & McNaughton, D. 2003, J. Mol. Spectrosc., 217, 123S´anchez Contreras, C. & Sahai, R. 2004, ApJ, 602, 960Sarre, P. J. 2006, J. Mol. Spectrosc., 238, 1Ruze, J. 1966, Proc. IEEE, 54, 633Trammell, S. R. 2000, in ASP Conf. Ser. 199, Asymmetrical Planetary Nebulae II: FromOrigins to Microstructures, ed. J. Kastner, N. Soker, & S. Rappaport (San Francisco:ASP), 147 9 –Ulich, B. L., & Haas, R. W. 1976, ApJS, 30, 247Woods, P. M., Millar, T. J., Zijlstra, A. A., & Herbst, E. 2002, ApJ, 574, L167Woods, P. M., Millar, T. J., Herbst, E., & Zijlstra, A. A. 2003, A&A, 402, 189
This preprint was prepared with the AAS L A TEX macros v5.2.
10 –Table 1: Summary of o -C H Observations Toward CRL 618 N T J ′ K ′ a ,K ′ c − J ′′ K ′′ a ,K ′′ c ν Ag u E u T MB η Upper Limit(MHz) (s − ) (K) (mK) (cm − )4 , − , × − × , − , × − × , − , × − × , − , × − × , − , × − × , − , × − × , − , × − × , − , × − × , − , × − × , − , × − × , − , × − × , − , × − × , − , × − × , − , × − × , − , × − × , − , × − × , − , × − × , − , × − × , − , × − × , − , × − × , − , × − × , − , × − ×
11 – C H -H C H -h ν +H-H C H +H CCC CHH H C H +H +Hn-C H *n-C H n-C H -H C H C H C H C H o-C H + .. Fig. 1.— Radical-based reaction scheme for PPNe benzene chemistry based onFrenklach & Feigelson (1989). 12 –
CCC C HH H + C H + C H + C H C H C H HCO + -CO-H -H C H -h ν H H-h ν -H-H + C H + + C H + +H +e - C H +e - o-C H Fig. 2.— Ion-based chemical scheme for PPNe benzene chemistry based on Woods et al.(2003). C H -H C C H C H +HC H -H+H-H C H C CH H -H C CH H CC HH C C H C C HC CH H-H C H PAHs . . .. . .
Fig. 3.— Potential PAH formation routes in circumstellar shells (Frenklach & Feigelson1989; Cherchneff et al. 1992). 13 – T M B ( K ) HC N , J = - H α Fig. 4.— Observed CRL 618 Q-band spectrum (grey) overlaid with a simulation of the o -C H spectrum (black) at a column density of 10 cm − , a rotational temperature of 200K, and a source size of 10 ′′′′