A SPECTRAL LINE SURVEY OF CRL 2688 IN THE RANGE OF 85–116 GHz

Figure 3 gives the line profiles of CO and ¹³CO from CRL 2688 and IRC +10216. The ¹²CO spectra of CRL 2688 and IRC +10216 show an optically thick emission with parabolic shape, while the ¹³CO spectrum of CRL 2688 shows a partially optically thin emission with rectangular shape and that of IRC +10216 shows an optically thin and resolved emission with a double peak. The integrated intensity ratios ∫ T*_A(CO) dv/∫T*_A(¹³CO) dv are 4.76 and 10.0 for CRL 2688 and IRC +10216, respectively. The observational results of Kahne et al. (1992) with the IRAM 30 m telescope showed that these values are 7.5 for CRL 2688 and 8.65 for IRC +10216, respectively. The use of these integrated intensity ratios of an optically thick to thin line may lead to lower limits of their abundance ratios. Kahne et al. (1992) reported that molecular abundance ratios CO/¹³CO are 8 for CRL 2688 and 16 for IRC +10216 as lower limits. According to Wannier & Sahai (1987), abundance ratios of CO/¹³CO were estimated to be 20 for CRL 2688 and 48 for IRC +10216 based on CO J = 2–1 observations. Because of this lower CO/¹³CO ratio for CRL 2688 than for carbon stars (40–80), Jaminet et al. (1992) placed CRL 2688 among the J-type carbon stars. The line-center velocity and the full width at zero power of CO emission are −36.6 km s⁻¹ and 39.4 km s⁻¹, respectively. So the half width is 19.7 km s⁻¹ (in good agreement with Herpin et al. 2002), which corresponds to the expansion velocity of CRL 2688. The expansion velocity of IRC +10216 measured by the CO line is 15.6 km s⁻¹. For a young PPN CRL 2688, Cox et al. (2000) have detected a high-velocity gas emerging in two directions of the central star in the CO J = 2–1 line which is closely related to H₂ emission at 2 μm.

Zoom In Zoom Out Reset image size

The spectra of CS J = 2–1 and J = 3–2 toward CRL 2688 (top panel) and IRC +10216 (bottom panel) are shown in Figure 2, respectively. The mapping observations of CS for CRL 2688 were performed by Kasuga et al. (1997) using the Nobeyama Millimeter Array with angular resolutions of 3''×3''. They reported that the emission is extended by about 15''. This value is smaller than the emitting size of other molecules such as HCN and HC₃N (larger than 20''). Their modeling of the CRL 2688 envelope indicates that the density distribution is nearly spherically symmetric but the abundance of CS is enhanced near the polar regions. According to their results, the high-velocity flow may be limited in the inner region of the envelope. Generally the CS line traces a dense region (⩾ 10⁴ cm⁻³). In the case of IRC +10216, the line profile of 250 kHz resolution shows a flat-topped feature, which indicates a partially optically thin molecular shell. The line intensity of CS J = 3–2 from CRL 2688 is about a tenth of that from IRC +10216.

This observation is the first to detect a CCS line in CRL 2688; although this molecule has been detected in IRC +10216 (Cernicharo et al. 1987). CCS is 40 times less abundant than CS in the case of IRC +10216 (Cernicharo et al. 1987). Cernicharo et al. (1987) reported that CCS can be formed at chemical equilibrium in the outer stellar atmosphere and can also be formed further in the envelope from CS, by ion–molecule reactions. Similar formation processes may be also applied to CRL 2688.

The line profiles of HCN and H¹³CN from CRL 2688 and IRC +10216 are given in Figure 2. The integrated antenna temperature ratios of HCN to H¹³CN ∫T*_A(HCN) dv/∫T*_A(H¹³CN) dv are 2.24 and 2.44 for CRL 2688 and IRC +10216, respectively. The latter value, 2.44 for IRC +10216 is larger than the values 2.0, observed with the Onsala 20 m telescope (Olofsson et al. 1982), and 1.65, observed with the IRAM 30 m telescope (Kahne et al. 1988) probably due to different beam dilution factors. Nguyen-Q-Rieu & Bieging (1990) showed that the isotope ratio HCN/H¹³CN in CRL 2688 appeared to be 5 as a lower limit for optically thick lines and that its determination is difficult because of line saturation and hyperfine line blending. Dinh-V-Trung & Nguyen-Q-Rieu (2000) reproduced the observed spectrum at the center position of IRC +10216 with a radiative transfer model of [HCN]/[H₂] = 2.5 × 10⁻⁵, given an isotopic ratio ¹²C/¹³C = 55. The carbon isotopic ratio ¹²C/¹³C is an important value for understanding nucleosynthesis and the evolution stage (Iben & Renzini 1983; Balser et al. 2002), but it is difficult to determine the ratio by using optically thick lines of CO and HCN (Kwan & Hill 1977). The asymmetric features of the HCN and H¹³CN spectra which may be produced by self absorption and/or blending of hyperfine components (Nguyen-Q-Rieu et al. 1988; Nguyen-Q-Rieu & Bieging 1990; Bieging et al. 1988) were not observed in CRL 2688, unlike in IRC +10216.

As shown in Figure 2, spectra from 90.55 GHz to 90.8 GHz, the SiS (J = 5–4, v = 0) molecular line was detected in IRC +10216 while it was not detected in CRL 2688 in our observations. We estimated an upper limit on the SiS abundance for T*_A ∼ .06 K (rms = 0.02 K) in CRL 2688. We assumed the source size of 35'' for the SiS J = 5–4 line (Fukasaku et al. 1994). The beam dilution factor was corrected. As a result, we got a column density of 2.6 × 10¹³ cm⁻² and an abundance (relative to H₂) of 1.2 × 10⁻⁸, whose orders are comparable to the results of 4.5 × 10¹³ cm⁻² (Fukasaku et al. 1994) and 4 × 10⁻⁸ (Nguyen-Q-Rieu & Bieging 1990). In the case of IRC +10216, its column density and abundance were estimated to be 4.7 × 10¹⁵ cm⁻² (Kawaguchi et al. 1995) and 6.5 × 10⁻⁷ (Nguyen-Q-Rieu & Bieging 1990), respectively. These values show a lower abundance of the SiS molecule in CRL 2688 than in IRC +10216. SiS is known to be produced in regions of chemical equilibrium. Therefore, SiS abundance decreases as a star evolves beyond the AGB, through the PPN and PN phases (Bachiller et al. 1997a). But the HNC molecule is mainly formed by ion–molecule reactions. Glassgold et al. (1987) indicated that HNC is produced from a dissociative recombination of HCNH⁺. The higher HNC abundances in the more evolved object PPN CRL 2688 may be ascribed to a higher level of ionization in the gas originated from radiation of the central star (Fukasaku et al. 1994).

Six rotational transitions in the ground vibrational state of HC₃N were detected in our line survey toward CRL 2688. Many millimeter wave spectra produced by rotational transitions of ground and vibrationally excited states of HC₃N have been also detected from the C-rich proto-planetary nebula CRL 618 and have allowed quite a precise determination of physical conditions (Wyrowski et al. 2003; Pardo et al. 2004, 2007b). We also adopted the HC₃N lines for investigating physical conditions such as excitation temperature and column density in CRL 2688.

Rotation temperatures (T_rot) and column densities (N) of HC₃N in CRL 2688 and IRC +10216 were derived using standard LTE rotational diagram analysis (Turner 1991).

The equation for the analysis is given by

$$\begin{equation} {\rm log}\,L = {\rm log} {{3k \int\!T_{\rm A}^*\, dv}\over {8 \eta _B \pi ^3 \nu S\mu _i^2 g_I g_k}} = {\rm log}{N\over Q_{\rm rot}} -{{E_u {\rm log}\,e} \over {k T_{\rm rot}}}, \end{equation} \tag{ 1 }$$

where k is the Boltzmann constant, ∫T*_Adv is the integrated intensity, η_B is the main beam efficiency of the telescope, ν is the rest frequency of the line, S is the line strength, μ_i is the relevant dipole moment, g_I is the reduced nuclear spin weight, g_k is the K-level degeneracy, E_u is the upper state energy of the transition, Q_rot is the rotational partition function, and N is the column density.

The above equation is based upon the assumptions that lines are optically thin, level populations are characterized by a single T_rot (LTE), T_rot≫T_bg (background temperature), and the Rayleigh–Jeans approximation is valid for all transitions (Turner 1991). The values for g_I, g_k, E_u/k, Sμ²_i, log L, and η_B for HC₃N are given in Table 2. The partition function Q_rot for this molecule is given by 3[kT_rot/hB] (Blake et al. 1986), where B is the rotational constant (4549.059 MHz from Lafferty & Lovas 1978). In total, six rotational transitions were detected in CRL 2688 and IRC +10216 (Table 1). Observed antenna temperatures were converted to source brightness temperature (T_R), where we considered the effect of main beam efficiency (η_B) and beam dilution (η_BD) as follows,

$$$ {\it T}_{\rm R} = {\it T}_{\rm A}^*/\eta _{B}\eta _{BD}. $$$

The correction factor for beam dilution is calculated using

$$$ \eta _{BD} = [{1+(\theta _B/\theta _S)^2}]^{-1}, $$$

where θ_B is the antenna beam width (FWHM) and θ_S is the source diameter (Bell 1993). We assumed that the source size of 20'' is uniform with our observing frequencies as Fukasaku et al. (1994) adopted. The beam size was obtained from an interpolation using the values, 64'' at 84 GHz and 46'' at 150 GHz. In the case of CRL 2688, all HC₃N data except for J = 12–11 were found to be in a good linear correlation in the distribution of E_u/k versus log L, as shown in Figure 4. This good correlation, except for the J = 12–11 transition datum, confirms the validity of the above assumptions for LTE rotation diagram analysis. The datum of the J = 12–11 transition probably has a calibration problem. Therefore, excluding the J = 12–11 transition datum, we made a linear least-squares fit and found T_ex = 25.8^+3.5_−2.7 K and N = 3.5^+1.4_−1.0 × 10¹⁴ cm⁻². In the case of IRC +10216, all data in the distribution of E_u/k versus log L were found to be in a good linear correlation. From the least-squares fit, we have also derived T_ex = 16.6^+0.9_−0.8 K, and N = 1.1^+0.2_−0.2 × 10¹⁵ cm⁻². These results along with literature values are given in Table 3. The HC₃N column density derived by us is consistent with the value in Fukasaku et al. (1994) within a factor of ∼ 2. Bachiller et al. (1997) also showed that the HC₃N column density is estimated to be 4.4 × 10¹⁴cm⁻² in CRL 2688. The observed HC₃N line shapes in CRL 2688 have showed an optically thick profile (Fukasaku et al. 1994) or, at least a partially optically thick profile (Olofsson et al. 1982) conflicting with the optically thin assumption above. Therefore, in order to confirm a better accurate optical depth of HC₃N in CRL 2688, it might be necessary to observe the isotope line of HC₃N by using a high-sensitivity radio telescope. The HC₃N column density of CRL 2688 estimated by us is lower than that of IRC +10216, as shown in Table 3. Nguyen-Q-Rieu & Bieging (1990) and Bieging & Tafalla (1993) also showed that HC₃N abundance relative to H₂ is lower in CRL 2688 than in IRC +10216 although Bujarrabal et al. (1988) did not find a difference in HC₃N abundance between CRL 2688 and IRC +10216. These facts may be related to the chemical evolution of the circumstellar envelopes of AGB stars. Specifically, Bachiller et al. (1997a) suggested that, as a star evolves beyond the AGB, through the proto-planetary and planetary phases, the abundance of HC₃N as well as the abundance of SiO, SiC₂, and CS decreases and HC₃N is not detected in the planetary nebulae.

Zoom In Zoom Out Reset image size

The C₂H and CN radicals are known to be extended to 50'' and 75'' in IRC +10216, respectively (Truong-Bach et al. 1987). Three hyperfine components of the N = 1–0 J = 3/2–1/2 spectra of C₂H were detected in both CRL 2688 and IRC +10216, as shown in Figure 2. The ethynyl radical, C₂H, was first detected in interstellar clouds by Tucker et al. (1974). It is known that the penetrating interstellar UV field plays a major role in dissociating HCN and C₂H₂ in the inner layer and producing CN and C₂H in the outer layer. In addition, these exotic molecules will cause the syntheses of linear long carbon chain molecules HC_2n+1N (Cyanopolynes with n = 1, 2, 3,...). One of these, HC₃N, was detected in six rotational transitions in this survey.

The N = 1–0 transitions of CN molecule with nine hyperfine components forming two fine-structure groups (J = 1/2–1/2 and J = 3/2–1/2) are shown in Table 1 and in Figure 2. The N = 1–0 J = 3/2–1/2 spectra of CN show a blended feature in both IRC +10216 and CRL 2688. The weakest component of the four N = 1–0 J = 1/2–1/2 spectra of CN (arrow mark in the CN N = 1–0 J = 1/2–1/2 spectrum of Figure 2) is not seen in the case of CRL 2688 because of lower sensitivity of the TRAO 14 m telescope compared with the IRAM 30 m telescope although all the four spectra of CN were detected by Bachiller et al. (1997a, 1997b). Bachiller et al. (1997a, 1997b) reported that the CN/HCN abundance ratio is 2 in CRL 2688 compared to about 0.5 in C-rich AGB envelopes including IRC +10216. This large abundance of CN in CRL 2688 may be related to the effect of chemistry influenced by the UV radiation of the central star during the PPN phase (Kwok 2000).