DISCOVERY OF MOLECULAR SHELLS ASSOCIATED WITH SUPERNOVA REMNANTS. I. KESTEVEN 69

Figure 1 shows three CO spectra toward the center (18^h32^m500, −10°06'42'') of the mapping area. There are several velocity components in the velocity range 0–120 km s⁻¹ in the ¹²CO (J = 1–0) spectrum, and no ¹²CO emission is detected out of this range in the whole mapping area. The ¹²CO emission peaks appear in the intervals ∼2–12, 40–60, 60–76, 76–100, and 107–120 km s⁻¹. We have found no morphological correspondence between the CO emission in the velocity intervals 2–12 km s⁻¹ and 40–60 km s⁻¹ and the emission from Kes 69 in any other waveband. The molecular component at around 110 km s⁻¹ also shows no morphological correspondence, which is near the tangent point in this direction. Applying the rotation curve of Clemens (1985) together with R₀ = 8.0 kpc (Reid 1993) and V₀ = 220 km s⁻¹, the tangent point in this direction is at 7.4 kpc at 113.7 km s⁻¹. For the ¹²CO emission, we focused our analysis on the velocity intervals 60–76 km s⁻¹ and 76–100 km s⁻¹, which respectively cover 69.3 km s⁻¹ and 85 km s⁻¹ at which the OH masers were detected. The ¹³CO (J = 1–0) emission is prominent at around 54 km s⁻¹ and 82 km s⁻¹, at which the ¹²CO emission is strong. The intensity of the C¹⁸O (J = 1–0) emission is too weak to examine its spatial distribution.

Zoom In Zoom Out Reset image size

We have produced ¹²CO intensity maps with interval 1 km s⁻¹ between 60 and 76 km s⁻¹ (Figure 2) to examine the molecular gas around 69.3  km s⁻¹, at which the compact OH maser is seen. In the northeastern compact maser region, some faint diffuse ¹²CO emission are present at 60–70 km s⁻¹. A distinct cloud at 69–76 km s⁻¹ is seen in the west.

Zoom In Zoom Out Reset image size

We have also produced ¹²CO intensity maps over the velocity range of 76–92  km s⁻¹ with interval 1 km s⁻¹ (Figure 3). By comparison with the radio continuum emission, an arc of the molecular gas at 77–86 km s⁻¹ is strikingly coincident along the southeastern rim. This arc is clearly seen in the close-up intensity images at 80–81 km s⁻¹ (Figure 4) and 79–82 km s⁻¹ (Figure 5). In the northwest (see Figures 4 and 5), there is another section of molecular arc, which can even be discerned in the wider range 79–86 km s⁻¹ (Figure 3). The two sections of molecular arcs are seen in similar velocity range and can be threaded with a circle of angular radius 87. Some bright, complicated ¹²CO features are present in the field of view at 80–88 km s⁻¹. There may be a contribution from the H ii region G21.902−0.368 in the northwest, where the recombination line at 79.5 km s⁻¹ has been detected (Lockman 1989). There is also a molecular cloudlet at 87–91 km s⁻¹, coincident with the strongest southern radio peak, at the western end of the incomplete radio shell. If this small cloud is associated with the SNR, the radio peak can be accounted for by the impact of the remnant shock on it, because of the drastic shock deceleration and hence the magnetic field compression and amplification.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The CO spectra from the northeastern compact OH maser region and the southeastern shell region are shown in Figure 6 (the two regions for CO spectrum extraction are shown in Figure 4). For the former region, the ¹²CO and ¹³CO emissions peak at both ∼65 and 85 km s⁻¹ (both with a signal-to-noise ratio (S/N)>3), essentially in agreement with the velocities at which the single maser arises (Green et al. 1997; Hewitt et al. 2008). The ¹²CO and ¹³CO lines at ∼65 km s⁻¹ have complicated profiles, each of which can be phenomenologically fitted with a combination of at least three Gaussians. The lines at 85 km s⁻¹ are slightly broadened in the blue wings. For the latter region, there are ¹²CO and ¹³CO lines at 85 km s⁻¹, with blue wings broadened (although the wing may be contaminated by the 72 km s⁻¹ peaks). The profile of each line is divided into a Gaussian at 85 km s⁻¹ and a residual part in the blue wing, with the fitted and derived parameters summarized in Table 2.

Zoom In Zoom Out Reset image size

In the derivation, we used two methods to estimate the H₂ column density and molecular mass. In the first method, the H₂ column density is estimated by the use of the conversion factor N(H₂)/W(¹²CO) ≈ 1.8 × 10²⁰ cm⁻² K⁻¹ km⁻¹s (Dame et al. 2001). In the second method, we assume local thermodynamical equilibrium (LTE) for the gas and optically thick condition for the ¹²CO (J = 1–0) line and use the relation N(H₂) ≈ 7 × 10⁵N(¹³CO) (Frerking et al. 1982).

There seems to be a weak feature at 82 km s⁻¹ in the ¹³CO line profile of the southeastern region, which could be either a broadened part of the 85 km s⁻¹ component or a chance coincidence of an irrelevant, unperturbed cloud in the line of sight. We performed a multiple Gaussian fit incorporating this feature and derived a molecular gas mass for this feature, which is about two orders lower than the virial mass, indicating that the cloud is perturbed. In view of this and the clear shell-like appearance at 80–82 km s⁻¹, we treat it simply as a small part in the broadened wing of the 85 km s⁻¹ gas.

In the long-time pointing observation toward the southern strongest radio peak or the cloudlet mentioned above, which targets at (18^h32^m500−10°12'42''), no HCO⁺ signal was detected. However, HCO⁺ emission was detected in the pointing observation toward another radio peak at (18^h33^m100, −10°12'42'') on the southeastern shell. The ¹²CO, ¹³CO, and HCO⁺ spectra of this point are shown in Figure 7. The HCO⁺ emission appears to be prominent only at ∼85 km s⁻¹, at which both the ¹²CO and ¹³CO emission peak, too. We calculated the column density of HCO⁺  assuming that the HCO⁺ emission is in LTE with the same excitation temperature as the ¹²CO emission (∼7.8 K) and is optically thin. For comparison with the CO column density, we smoothed the CO (J = 1–0) data to the same angular resolution as the HCO⁺ observation, and the results are listed in Table 3. Considering the ¹²CO/¹³CO ratio (30–70) of the general interstellar medium (Langer 1992 and references therein), the HCO⁺/CO abundance ratio is ∼ 5 × 10⁻⁶. It is somewhat less than that of undisturbed cold MCs such as TMC-1 or L134N (1 × 10⁻⁴; Ohishi et al. 1992). The HCO⁺ abundance is expected to be reduced in a slow, nondissociating shock, unless some enhanced ionization is present (Iglesias & Silk 1978; Elitzur 1983). We note that the extended maser emission on the southern rim is found at the same LSR velocity (Hewitt et al. 2008), which implies that the dense molecular gas at this velocity is disturbed by C-type shock. For the origin of HCO⁺ emission, though, further observation is needed to make a decision.

Zoom In Zoom Out Reset image size

Combination of the multiwavelength observations toward Kes 69 (Figure 5) shows some interesting morphological correlation. First, the southeastern 77–86 km s⁻¹ molecular shell is not only coincident well with the 1.4 GHz radio shell (as pointed out above), but also (as seen in Figure 8) with the mid-IR "ridge" described in Reach et al. (2006). This mid-IR ridge is prominent in 4.5 μm, with a relatively high brightness in 5.8 μm. A part of the 24 μm floclike emission in the southeast is also aligned with the molecular shell (Figure 8). The presence of the molecular shell is consistent with the HCO⁺ and extended 1720 MHz OH maser emission at 85 km s⁻¹ along the southern rim. This correspondence in morphological features is most probably another signature of the SNR-MC interaction, in addition to the 1720 MHz OH masers detected in Kes 69.

Zoom In Zoom Out Reset image size

Second, as described above, both of the sections of molecular arcs at ∼77–86 km s⁻¹ in the southeast and the northwest seem to be aligned along a circle. A faint 1.4 GHz bar in the north (highlighted by a red contour in Figure 5) is also along this circle. Moreover, as noted by Yusef-Zadeh et al. (2003), there is a faint radio shell in the northwest in the low-resolution 330 MHz image of Kes 69 (Kassim 1992), which is confused with the H ii region G21.902−0.368. We note that the peak of this patch of radio emission roughly coincides with the northwestern section of the molecular arc.

These morphological correlations demonstrate that Kes 69 is associated with the giant MC at the systemic velocity of ∼85 km s⁻¹. This association is strengthened by the detection of the extended and compact 1720 MHz OH maser emission and the HCO⁺ emission at 85 km s⁻¹ from the SNR region.

The southeastern half of the circle is roughly X-ray bright, while the other half appears to be X-ray faint but covered by the 79–82 km s⁻¹ molecular gas. This brightness anticorrelation between the X-ray and CO emission, however, does not imply that the molecular gas in the northwestern half obscures the X-rays. We estimate the hydrogen column density of this gas (in the interval 79–82 km s⁻¹) to be N_H ∼ 1 × 10²¹ cm⁻³. If there were an X-ray emitting gas, with similar properties to that observed in the southeast (temperature kT_x ∼ 1.6 keV and intervening hydrogen column 2.4 × 10²² cm⁻²; Yusef-Zadeh et al. 2003), behind this northwestern molecular gas, then an extra extinction by the N_H ∼ 1 × 10²¹ cm⁻³ gas would only cause an ∼8% decrease in the 0.5–2 keV X-ray flux and this X-ray emission could have been observable. Thus we conclude that the X-ray faintness in the northwest is not caused by the absorption of the diffuse 79–82 km s⁻¹ molecular gas, whether this gas is connected to the ∼85 km s⁻¹ cloud or not.

The association between SNR Kes 69 and the MC at systemic velocity 85 km s⁻¹ facilitates a convincing determination of the kinematic distance to the remnant. The systemic velocity 85 km s⁻¹ is suggestive of two candidate kinematic distances to the SNR/MC association, 5.2 kpc and 9.6 kpc. The choice can be made with the aid of the H i absorption along the line of sight. Following the method used in Tian et al. (2007) and Tian & Leahy (2008a), we produced three H i spectra of regions along the southeastern shell of Kes 69, as shown in Figure 9. Distinct absorption features appear at 4, 6, 15, 40, 48, 67, 70, 82, and 88 km s⁻¹. The ¹²CO components along the shell around 5 and 50  km s⁻¹ (Figure 6), due to the corresponding absorption features in the H i spectra, can be related to the chance-coincident foreground gas. No H i absorption features are present around the tangent point velocity 113.7 km s⁻¹, which indicates that the SNR is in front of the tangent point (at 7.4 kpc). Hence, the distance to the SNR is d = 5.2 kpc. We note that a similar estimate has in the meantime been given by Tian & Leahy (2008b).

Zoom In Zoom Out Reset image size

The southeastern molecular arc is revealed above to be coincident with the radio and IR shell and move at a velocity v_m of order 10 km s⁻¹ (matching the blueshifted line broadening which reflects the velocity component in the line of sight). The broadened blue wing of the ∼85 km s⁻¹ line profile of the molecular arc implies that the giant molecular cloud may have suffered a perturbation from the rear side. Three scenarios regarding the dynamical relation between the molecular arc and the SNR are discussed below.

First, the arc is likely to be a flake of molecular gas that is shocked by the slow transmitted cloud shock after the SNR blast wave hits the molecular cloud. In this case, the radio continuum emission may arise from the SNR shock that is blocked by dense cloud (Frail & Mitchell 1998) or the blast wave that propagates in the intercloud medium (Blandford & Cowie 1982), and the thermal X-ray emission may be ascribed to the hot gas behind the shocked molecular gas or just behind the blast wave. Thus there can be a crude pressure balance between the cloud shock and the X-ray emitting hot gas (Zel'dovich & Raizer 1967; McKee & Cowie 1975): n₀v²_b ∼ n_mv²_m, where n_m denotes the number density of the hydrogen atoms ahead of the cloud shock, n₀ the density of the undisturbed intercloud medium, and v_b the velocity of the blast wave. Velocity v_b is related to the postshock temperature kT_x as $v_b=[16kT_x/(3\bar{\mu }m_{\rm H})]^{1/2}$, where m_H is the hydrogen atom mass and $\bar{\mu }=0.61$ is the average atomic weight. Adopting kT_x ∼ 1.6 keV from the ROSAT X-ray observation (Yusef-Zadeh et al. 2003), we have v_b ∼ 1.2 × 10³ km s⁻¹ and n_m ∼ 1.3 × 10³(n₀/0.1 cm⁻³)(v_m/10 km s⁻¹)⁻² cm⁻³, implying that the blast wave hits a very dense matter in the molecular arc (here n₀ is assumed to be similar to the mean density of the X-ray emitting gas, ∼0.1 cm⁻³, as obtained from a reproduced ROSAT X-ray spectral analysis⁵). In this scenario, assuming an adiabatic expansion and adopting the SNR extent represented by the circle of radius ∼87 (Section 3.2) or r_s ∼ 13d_5.2 pc, the SNR's age is estimated as t = (2r_s)/(5v_b) ∼ 4.4d_5.2 kyr and the explosion energy is E = (25/4ξ)(1.4n₀m_H)v²_br³_s ∼ 6 × 10⁵⁰(n₀/0.1 cm⁻³)d³_5.2 erg (where ξ = 2.026).

The 4.5 μm and 5.8 μm mid-IR emissions are suggested to be likely dominated by lines of shocked gas (Reach et al. 2006). They note that the 5.8 μm emission is relatively strong, but it is not clear whether H₂ or [Fe ii] lines (as likely mechanisms) are responsible for this emission, unlike the case of IC 443, in which the [Fe ii] and H₂ lines are very clearly segregated to the northern and southern regions of the remnant, respectively (see Rho et al. 2001). The H₂ emission seems to be consistent with the slow molecular shock, while [Fe ii] emission could not be ruled out with the present observations if the SNR shocks propagate in a complicated multiphase medium. The 24 μm emission along the arc is likely to arise from the shocked molecular gas (e.g., OH and H₂O), shock-heated dust grains, and even probably ions. The OH maser, HCO⁺, and H₂O (if any) emissions are consistent with a C-type molecular shock.

Second, let us discuss the scenario in which the arc represents the interstellar material swept up all the way by the SNR shock wave. The mass of the southeastern molecular arc is in the range of (0.5–1.1) × 10⁴d²_5.2 M_☉, with the gas mass observed in the whole line profile adopted as the upper limit and that of the broadened part as the lower limit (Table 2). In the swept-up case, the southeastern arc consists of the molecular gas that was originally of mean density n(H₂) ∼ 30–70d⁻¹_5.2 cm⁻³, distributed in an approximate 1/8th volume of a sphere of radius r_s ∼ 87 ∼ 13d_5.2 pc (considering the one-sided line broadening). In this scenario, the velocity of the SNR shock is represented by the arc's slow expansion (v_m ∼ 10 km s⁻¹), which would imply that the SNR is in the radiative phase. The explosion energy of the SNR, E ∼ 7 × 10⁵⁰(r_s/13 pc)^3.12[n(H₂)/50 cm⁻³]^1.12(v_m/10 km s⁻¹)^1.4 erg (Chevalier 1974), would have seemed to be normal, despite the seldom large age t ∼ 0.31r_s/v_m ∼ 4 × 10⁵(r_s/13 pc)(v_m/10 km s⁻¹)⁻¹ yr for X-ray-bright SNRs. However, the X-rays are bright along the southern shell and two X-ray emission peaks are essentially coincident with the two radio peaks (Figure 5; toward which the pointing molecular line observation was made). A radiative shock as slow as 10 km s⁻¹ could not be responsible for such X-rays along the SNR rim. On the other hand, to accelerate particles to relativistic energies for emitting radio synchrotron, the lower limit of the shock velocity is (Draine & McKee 1993): v_s > 64(x_i/10⁻⁵)^1/8[n(H₂)/50 cm⁻³]^1/8 (ϕ_cr/0.1)^−1/4(T/100 K)^0.1 km s⁻¹, where x_i is the ionization fraction, ϕ_cr the efficiency of the particle acceleration of the shock, and T the preshock molecular gas temperature. It is much larger than the observed expansion velocity of the molecular shell, and thus new particle acceleration would be difficult to take place at this shock. The existing relativistic electrons would rapidly stream freely away from the shock front for a very low ionization fraction (Blandford & Cowie 1982). Moreover, because of the ambipolar diffusion, the magnetic field will separate from the neutral gas in a timescale (Blandford & Cowie 1982): ∼10³(δr/1 pc)²(x_i/10⁻⁵) [n(H₂)/50 cm⁻³]²(B/10⁻⁶ G)⁻²(s/100)⁻² yr (where B is the magnetic field strength in the unshocked gas, s the density compression ratio, and δr the shell thickness), very likely to be smaller than the remnant age in this case (even for s ∼ 10). Therefore, such a slow expansion of molecular cloud is difficult to account for the SNR's radio emission. In view of the above points, this scenario does not apply to the molecular shell of Kes 69.

Third, the molecular shell may be the debris of the cooled, condensed material, which was swept up by the stellar wind of the supernova progenitor from the molecular gas [n(H₂) ∼ 50 cm⁻³] and is now hit by the SNR shock. This possibility could be compatible with the first scenario and naturally explain why there is pre-existing, very dense material that is now hit by the SNR shock. A molecular wind-bubble shell has recently been discovered coincident with the ring nebula G79.29+0.46 surrounding a luminous blue variable star (Rizzo et al. 2008), which provides an instance for this scenario. If there was a stellar-wind bubble, it was created (Castor et al. 1975; Weaver et al. 1977) ∼1.4 × 10⁶(r/13 pc)^5/3L^−1/3₃₆[n(H₂)/50 cm⁻³]^1/3 yr ago, where L₃₆ is the mechanical luminosity of the stellar wind in units of 10³⁶ erg s⁻¹, and the present velocity of the shell is ∼6(r/13 pc)^−2/3L^1/3₃₆[n(H₂)/50 cm⁻³]^−1/3 km s⁻¹. This velocity is comparable to the blueshift in the broadened line profile of the molecular gas along the shell.

There appears to be a blowout morphology outlined by the extension of the radio/CO shell to the northeast out of the circle (Figures 4 and 5), somewhat similar to the blowout morphology in SNR N132D (e.g., Dickel & Milne 1995; Xiao & Chen 2008). Actually, SNR N132D, which is in the vicinity of an MC, has been suggested to be shaped by the shock impacting on the stellar wind-bubble shell (Hughes 1987; Chen et al. 2003). The northeastern compact masers at both 69 and 85 km s⁻¹ are projected in the blowout region. In this case, as a possibility, the 69 km s⁻¹ maser could arise from a dense clump deviating from the systemic velocity by a strong perturbation, although it cannot be excluded from being nonassociated with the same SNR.