The Large-scale Interstellar Medium of SS 433/W50 Revisited

The radio nebula W50, also known as supernova remnant (SNR) G39.7−2.0 (e.g., Green 2017), has a large angular extent of ∼120' × 60', which surrounds its central bright compact point source of SS 433. SS 433, a close massive binary system, is a Galactic microquasar consisting of a compact object and a massive donor star (e.g., see reviews of Margon 1984; Fabrika 2004).

The well-known and unique object SS 433 has been widely studied in multiwavelength observations, as well as theoretical analyses and numerical simulations. Most of these studies focused on the properties of the energetic microquasar itself. On the other hand, many works were also concentrated on the large-scale environment of the unusual system (e.g., radio continuum studies in Geldzahler et al. 1980; Downes et al. 1981, 1986; Elston & Baum 1987; Dubner et al. 1998; Gao et al. 2011; Broderick et al. 2018; IR studies in Band 1987; Wang et al. 1990; Mirabel et al. 1996; optical studies in van den Bergh 1980; Zealey et al. 1980; Mazeh et al. 1983; Boumis et al. 2007; Abolmasov et al. 2010; X-ray studies in Watson et al. 1983; Yamauchi et al. 1994; Brinkmann et al. 1996; Safi-Harb & Ögelman 1997; Safi-Harb & Petre 1999; Brinkmann et al. 2007; the multiwavelength studies in Moldowan et al. 2005; H i and CO gas studies in Lockman et al. 2007; Yamamoto et al. 2008; and the very recent magnetic fields and ionized gas studies in Farnes et al. 2017).

Among the ample scope on the SS 433/W50 system, the interaction between SS 433/W50 and its surrounding interstellar medium (ISM) is a worthy topic for further studies. The prominent phenomenon of the system is that the elongation of W50 is exactly along the axis of the precession cone of the SS 433 jets. The extension of the relativistic jets of SS 433 (and/or jet counterparts in multiwavelength, e.g., X-ray, optical, and radio emission) has an orientation coincident with that of W50 nebula, suggesting the physical connection between them. Recently, Panferov (2017) suggested that a model of 10⁵ yr old SNR and <2.7 × 10⁴ yr old jets can explain the configuration of SS 433/W50. But for all this, the origin of them is still uncertain and debated (e.g., more details are given in Farnes et al. 2017).

In this paper, we use the Milky Way Imaging Scroll Painting (MWISP⁴ ) CO data and the complementary H i data from the Green Bank Telescope (GBT) 100 m and Arecibo Observatory 305 m telescopes to investigate the molecular and atomic gas properties of W50. Thanks to the Galactic CO and H i surveys with the high-dynamical range, we can gain much insight into the ISM environment of SS 433/W50 from the combined high-quality molecular and atomic line observations. In particular, the high-resolution (in spatial and velocity) CO and H i data allow us to investigate the gas properties and the kinematic features of the surrounding ISM of SS 433/W50 from the large-scale structures of several degrees to the small-scale features of ≲1 arcmin.

Throughout this paper, we use Galactic coordinates to identify directions on the sky, which is convenient to analyze the structure of the gas on a large scale, e.g., several degrees. In particular, the well-known western and eastern jets/lobes of SS 433/W50 in equatorial coordinates are described as the northwestern jet/lobe (the receding one toward the Galactic plane) and the southeastern jet/lobe (the approaching one away from the plane) in Galactic coordinates (see the radio morphology of W50 in Figure 1 and the red and blue lines in Figure 4), respectively.

Zoom In Zoom Out Reset image size

2.1. CO Data

2.2. H i Data and Radio Continuum Emission

Several works had been done to investigate the gas environment toward the SS 433/W50 system (e.g., Huang et al. 1983; Dubner et al. 1998; Durouchoux et al. 2000; Lockman et al. 2007; Yamamoto et al. 2008). However, conclusions from these studies, especially on the kinematic distance to SS 433/W50 (e.g., 2.2–5.5 kpc), are not consistent with each other. In these papers, the local standard of rest (LSR) velocity of the gas was used to determine the distance of the system based on H i absorption and/or CO+H i associations with W50 nebula. As an association between SS 433/W50 and its surrounding gas was established, the kinematic distance of them at a certain LSR velocity can be determined from the Galactic rotation curve model. Therefore, the key point to the argument is which of the gas components along the line of sight (LOS) is actually associated with SS 433/W50.

A detailed analysis of the combined molecular and atomic gas toward the direction could shed light on the connection between the SS 433/W50 system and its surrounding ISM. The new high-resolution and high-sensitivity CO observations, together with the complementary H i data, may also provide us a good opportunity to investigate interactions of the system with its ambient gas on a large scale.

Generally, a large amount of molecular gas traced by CO emission, as well as the complicated atomic gas traced by H i data, is seen in the region of SS 433/W50. For the convenience of discussion, the gas toward SS 433/W50 was divided into four parts according to different velocities of V_LSR ≲ 0 km s⁻¹, V_LSR = 0–20 km s⁻¹, V_LSR = 20–70 km s⁻¹, and V_LSR ≳ 70 km s⁻¹. Position−velocity (PV) diagrams across the radio nebula W50 (e.g., see the yellow arrows that are roughly perpendicular to the precession axis of SS 433 in Figure 1) were made in Figure 2 for subsequent analysis.

Zoom In Zoom Out Reset image size

3.1. Gas at ≲0 km s⁻¹

3.2. Gas at 0–20 km s⁻¹

3.3. Gas at 20–70 km s⁻¹

3.3.1. The 170 pc Long MC Filament G40.82−1.41 at 29–34 km s⁻¹

According to the CO intensity map in the interval of 20–40 km s⁻¹ (Figure 1), a giant molecular filament (GMF), which is named as G40.82−1.41 from its geometric center of the CO emission, is revealed to extend about 5° from (l = 429, b = −01) to (l = 387, b = −27). The distance to the GMF is estimated to be 1.6–2.0 kpc (e.g., the 3D extinction map from Green et al. 2015) or ∼1.8–2.1 kpc from the CO peak velocities of V_LSR ∼ 29–34 km s⁻¹ (the near kinematic distance from the A5 model of Reid et al. 2014). The length of the GMF is thus about 170 pc at a distance of ∼2.0 kpc.

The GMF G40.82−1.41 is a clustering of tens of small filaments with somewhat different LSR velocities. These small filaments consisting of multiple components exhibit complicated structures. Many of them are oriented parallel to the long axis of the GMF and appear to be twisted between each other. The southwestern part of the GMF across the radio nebula W50 was suggested to have a connection with the SNR W50 (see Figure 3 and the discussions in Huang et al. 1983). Based on PV diagrams (e.g., LPV2 and LPV3 in Figure 2), however, we cannot find significant kinematic features of shock–MC interactions near and within the boundary of W50 nebula. Any asymmetric CO-line profiles, which can be seen everywhere both within and outside the W50 nebula, are obviously attributed to overlapping components at somewhat different velocities.

3.3.2. The MCs along the Precession Cone of SS 433 and the Corresponding H i Gas at ∼40–60 km s⁻¹

Several interesting MCs seem roughly aligned along the precession cone of SS 433, which was suggested to be the evidence of SS 433/W50–ISM interactions (see Figures 1–4 and the discussions in Yamamoto et al. 2008). Combined with the radio continuum and H i studies by Dubner et al. (1998), Yamamoto et al. (2008) then placed these MCs (the northern MCs of N1–N4 at V_LSR ∼ 53 km s⁻¹ and the southern MCs of S1–S6 at V_LSR ∼ 43 km s⁻¹; see Table 1 in their paper) and the associated SS 433 at a near kinematic distance of ∼3 kpc.

Figure 3 shows the WISE 22 μm emission overlaid with MWISP CO contours (blue, corresponding to MCs S1–S6 in Yamamoto et al. 2008) in the interval of 39–51 km s⁻¹ and GALFA H i contours (red) in the interval of 35–48 km s⁻¹. Obviously, the bright diffuse IR emission, which displays multiple shell-like or bubble-like features, is coincident with the gas emission at V_LSR ∼ 43 km s⁻¹, suggestive of an association between the IR features and the gas. The near kinematic distance of the gas is about 2.7 kpc (Reid et al. 2014), which is in good agreement with the 2.0–2.5 kpc from the 3D extinction map of Green et al. (2015). The atomic and molecular gas, together with the bright thermal dust association centered at (l ∼ 404, b ∼ −43) with a radius of ∼1°, is probably related to the star-forming regions at a distance of ∼2.5 kpc (or 190 pc below the Galactic plane). The total mass of the atomic+molecular gas within the region of ∼90 × 90 pc² is ≳1 × 10⁵ M_⊙. It is hard to believe that such amounts of gas can originate from the jet–ISM process of SS 433.

Zoom In Zoom Out Reset image size

On the other hand, MCs at 50–55 km s⁻¹ were suggested to be related to the W50 lobe toward the Galactic plane due to the spatial coincidence between them (Durouchoux et al. 2000; Chaty et al. 2001; Yamamoto et al. 2008). Most of the MCs, which are near the H ii region Sh 2-74 and close to the Galactic plane, also have corresponding bright IR emission. We argue that these MCs with prominent ¹³CO emission are probably associated with H ii region Sh 2-74 and/or belong to the nearby MC complex with active star formation (see Section 3.3.3), which agrees with suggestions from previous studies (e.g., Band & Gordon 1989; Moldowan et al. 2005).

At 38–44 km s⁻¹, a large H i void-like feature roughly shows the morphological correspondence with the radio emission of W50 (e.g., see Figure 14 in Lockman et al. 2007). However, the large-scale atomic cavity seen in the integrated map seems to be comprised of several separated substructures of H i emission. We cannot find fine morphological agreement between the radio continuum emission of W50 and its nearby atomic gas after analyzing the Arecibo H i data channel by channel. We agree with the conclusion of Lockman et al. (2007) that the gas in the velocity range is not physically associated with SS 433/W50.

3.3.3. Other Interesting MCs

The H ii region Sh 2-74 (l ∼ 399, b ∼ −13), which has a large radio size of ∼1° × 1° at a distance of about 3 kpc (e.g., Elston & Baum 1987; Paladini et al. 2003), lies near the northeastern edge of the W50 nebula (Figure 1). Using MWISP CO data, we find that the molecular gas at 40–55 km s⁻¹ (V_peak ∼ 48 km s⁻¹, which agrees well with the RRL central velocity of ∼47.7 km s⁻¹ for Sh 2-74; Alves et al. 2012), is physically associated with the H ii region according to the morphological agreement and the corresponding kinematic features between the ionized gas of the H ii region and the surrounding MCs (see LPV1 and LPV2 in Figure 2).

An interesting MC at (l ∼ 4066, b ∼ −242, V_peak ∼ 57 km s⁻¹) has broad CO wings in the velocity range of 45–72 km s⁻¹ (see LPV3 in Figure 2), which is very likely associated with the H ii region Du 22 centered at (l = 406567, b = −24658) (Dubout-Crillon 1976). The near kinematic distance of the object and the associated MCs is estimated to be ∼3.5 kpc. The far one of ∼9.0 kpc may be excluded due to the unreasonable scale-height of ∼400 pc.

Based on the above analysis, we argue that the MCs in the velocity range of 20–70 km s⁻¹ are accidental superpositions along the LOS and are not associated with the SS 433/W50 system. Both of the very long GMF (see the discussions in Huang et al. 1983) at V_LSR ∼ 29–34 km s⁻¹ and the gas features at V_LSR ∼ 40–60 km s⁻¹ (Dubner et al. 1998; Durouchoux et al. 2000; Yamamoto et al. 2008) appear to be just the foreground gas of SS 433/W50 in the FOV. We will not discuss these MCs any further because the above gas is not related to SS 433/W50 at a distance of 4.9 kpc (see Section 4).

3.4. Gas at ≳70 km s⁻¹

There are only a few CO clouds (V_LSR ≳ 70 km s⁻¹, T_peak ∼ 1–2 K, and size ∼1–2 arcmin²) within the radio boundary of the W50 nebula, e.g., (l = 39225, b = −1942, V_peak ∼ 73 km s⁻¹) and (l = 39692, b = −2450, V_peak ∼ 74 km s⁻¹), indicating that the CO emission at such velocities is very weak. Interestingly, two intriguing MCs, which are named as G39.315−1.155 and G40.331−4.302, are exactly aligned along the precession axis of SS 433 (see the two small boxes in Figure 4).

Zoom In Zoom Out Reset image size

Figure 5 shows a close-up view of the MCs G39.315−1.155 and G40.331−4.302. The northwestern MC G39.315−1.155 appears to display a coherent structure in the CO intensity map of 73–88 km s⁻¹, while the southeastern MC G40.331−4.302 consists of two parts, the faint CO emission in the eastern region (near PVSE lines in the right panel of Figure 5) and the main concentration with relatively strong emission in the western region (near PVS lines in the figure). Both of the two MC concentrations are along the direction from the northeast to the southwest, which are nearly perpendicular to the precession axis of SS 433 (see the red and blue lines in Figure 4).

Zoom In Zoom Out Reset image size

PV diagrams along selected lines (see the black arrows in Figure 5) are shown in Figures 6 and 7 for MCs G39.315−1.155 and G40.331−4.302, respectively. Typical spectra of some regions (see circles in Figure 5) are shown in Figures 8 and 9, respectively. It is interesting to note that MC G39.315−1.155 displays redshifted features while MC G40.331−4.302 displays somewhat blueshifted features in both of the PV diagrams and the typical spectra.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Table 1 lists properties of the atomic and molecular gas toward the main part of G39.315−1.155 and G40.331−4.302. The column density of atomic gas of G40.331−4.302 is calculated via the conversion factor of 1.823 × 10¹⁸ cm⁻² (K km s⁻¹)⁻¹ (Dickey & Lockman 1990). We cannot calculate the atomic-gas properties of G39.315−1.155 due to the strong background H i emission near the Galactic gas plane. The peak temperature of MC G40.331−4.302 is ∼3.4 K and ∼0.6 K for ¹²CO and ¹³CO, respectively. Note that if we use LTE assumption, an excitation temperature of 10 K, the beam filling factor of ∼0.5, and N(H₂)/N(¹³CO) ≈ 7 × 10⁵, the H₂ column density of G40.331−4.302 is about 1.1 × 10²¹ cm⁻², which is somewhat larger than the estimated value from the X-factor method (e.g., the mean CO-to-H₂ mass conversion factor of 2 × 10²⁰ cm⁻² (K km s⁻¹)⁻¹; Dame et al. 2001; Bolatto et al. 2013). It can be naturally explained because we adopt the peak emission of ¹²CO and ¹³CO to represent the property of the whole MC.

For the gas in G40.331−4.302, the velocity dispersion (${\sigma }_{v}=\displaystyle \frac{{\rm{\Delta }}{V}_{\mathrm{FWHM}}}{2.355}$) is estimated to be ∼6 km s⁻¹ for the atomic gas and ∼1.7 km s⁻¹ for the molecular gas, respectively. The mean density of the initial gas environment is on the order of ∼1 cm⁻³, which is a reasonable value estimated from the V_LSR ∼ 70–90 km s⁻¹ H i emission along the precession axis of SS 433. For MC G39.315−1.155, the velocity dispersion of the molecular gas is roughly 3–4 km s⁻¹, which is approximately two times larger than that of MC G40.331−4.302. The property indicates that MC G39.315−1.155 is much more turbulent than MC G40.331−4.302.

For G40.331−4.302, the molecular gas is mainly concentrated in a slab with a size of ∼32' × 8' (e.g., Figures 10 and 12). The total mass of G40.331−4.302 should be somewhat larger than the estimated value of ∼3 × 10³ M_⊙ in Table 1 because of the unaccounted gas in the northeastern region of the cloud (near PVSE in Figure 5) and the potential H₂ gas in CO-dark regions (see Section 6). For G39.315−1.155, the H₂ mass within a size of ∼17' × 8' is ∼1.1 × 10⁴ M_⊙.

Zoom In Zoom Out Reset image size

Obviously, the total mass of G39.315−1.155 is larger than that of G40.331−4.302. The two MCs also have different angular distances from SS 433 (e.g., ∼69' for G39.315−1.155 versus ∼129' for G40.331−4.302) and different extensions along the northeast–southwest direction (e.g., ∼17' for G39.315−1.155 versus ∼32' for G40.331−4.302). Actually, the northwestern radio lobe of W50 is famously shorter and brighter than the southeastern one, which is widely attributed to the denser ambient medium close to the Galactic plane (e.g., Dubner et al. 1998; Lockman et al. 2007; Goodall et al. 2011a). The two MCs features presented here can also be explained by the same reason of the denser environment near the Galactic plane, if these MCs are governed by the SS 433 jets/outflows (see Sections 4–6).

The extended and fragmented CO gas of the two MC concentrations is well confined to a cone with a small angular extent of ${\left(\displaystyle \frac{17}{69},\times ,\displaystyle \frac{180}{\pi }\right)}_{{\rm{G}}39.315}\sim {\left(\displaystyle \frac{32}{129},\times ,\displaystyle \frac{180}{\pi }\right)}_{{\rm{G}}40.331}\approx 14^\circ$ with respect to SS 433. We note that as such the angular extent of the molecular gas is comparable to the result from the current X-ray jet of SS 433 (e.g., ∼18°, also seen from SS 433; Brinkmann et al. 2007).

To search for possible evidence of the W50–ISM association, we constructed four integrated H i maps in the intervals of 85–88, 82–85, 79–82, and 76–79 km s⁻¹ (Figure 11). Several H i features at the velocity range are indeed found to be positionally coincident with the bright radio shell of W50 in the FOV. First, the H i maps exhibit a cavity-like structure near the Galactic plane, which coincides with the radio morphology of the W50 nebula (also see the static ring in Lockman et al. 2007). Second, several features with enhanced H i emission match the bright radio shells of W50 very well (e.g., see Figures 11(c) and (d)). Third, these H i enhancements, as well as the bright radio shells, roughly extend along the northwest–southeast direction, which is consistent with the trend of the precession axis of SS 433 (note that the axis is roughly perpendicular to the LOS with an inclination angle of ≈80°; Eikenberry et al. 2001).

Zoom In Zoom Out Reset image size

Finally, the H i gas at V_LSR ∼ 70–90 km s⁻¹ is associated with MC G40.331−4.302 (Figures 10–12). The H i concentration near MC G40.331−4.302, which seems to be a part of an expanding H i shell near the southeastern radio lobe of W50 (or the outermost parts of the approaching gas from the SS 433 jet), was suggested to be related to SS 433/W50 (Lockman et al. 2007). Using the Arecibo H i data, we find that the atomic gas, as well as the corresponding molecular gas at the velocity (e.g., the main part of MC G40.331−4.302), is indeed approaching the observer (Figures 11 and 12), indicative of the association between the gas and the SS 433 jet.

Zoom In Zoom Out Reset image size

Figure 4 displays that two MC concentrations exactly lie projected on the precession axis of the SS 433 jets. Further analysis shows that the kinematic features of the two MC concentrations are consistent with the behavior of the SS 433 jets. That is, MC G39.315−1.155 displays the redshifted feature (e.g., in particular, PVNW3 and PVNW4 in Figure 6) corresponding to the receding jet of SS 433, and MC G40.331−4.302 displays the blueshifted feature (e.g., in particular, PVS1 and PVS2 in Figure 7) corresponding to the approaching one.

The blueshifted feature of MC G40.331−4.302 is clearly confirmed from the accompanying atomic gas using the Arecibo H i data (see Figure 12). We speculate that MC G39.315−1.155 also has its accompanying receding atomic gas. However, the very strong background H i emission near the Galactic plane prevents us from obtaining its detailed kinematic information. It is very difficult to discern the possible kinematic feature at levels of several K from the strong background H i emission at several tens of K. We also emphasize that we cannot identify the disturbed molecular gas in the region closer to the Galactic plane (e.g., b ≳ −10) due to the complicated CO emission (e.g., strong ¹³CO emission and multiple CO peaks in the velocity interval of 70–85 km s⁻¹).

However, despite all of this, a H i cavity, which was identified by Lockman et al. (2007) using the GBT observations, does appear toward the W50 radio lobe near the Galactic plane using the new Arecibo data (Figure 11). MC G39.315−1.155 is roughly touching the top wall of the H i cavity (e.g., red contours in Figure 11(d)). Actually, a patch of atomic gas with enhanced H i emission at ∼80–90 km s⁻¹ (see the H i gas near red contours in Figure 11) seems to be coincident with the MC, suggesting the possible atomic gas away from the observer.

Additionally, the H i emission at ∼70–90 km s⁻¹ displays compelling morphological evidence for an association between W50 and the atomic gas (e.g., Figure 11). Combined the above analysis and results from Section 3.4, all the evidence points to the gas at ∼70–90 km s⁻¹. We thus argue that the ∼77 km s⁻¹ gas is physically associated with SS 433/W50, which agrees well with the previous study by Lockman et al. (2007).

We note that the heliocentric systemic radial velocity of 56 ± 2 km s⁻¹ was suggested to be related to SS 433/W50 from deep optical observations toward the whole system (Boumis et al. 2007). The heliocentric radial velocity of ∼56 km s⁻¹ can be transformed to an LSR velocity of ∼75 km s⁻¹ (e.g., see Reid et al. 2009, 2014), which is in excellent agreement with our finding of V_LSR = 77 ± 5 km s⁻¹ for the whole system on a large scale.

Accordingly, the LSR velocity of 77 ± 5 km s⁻¹, where the ±5 km s⁻¹ is the velocity error for possible peculiar motions, corresponds to a near kinematic distance of 4.9 ± 0.4 kpc (e.g., the A5 model in Reid et al. 2014). The far kinematic distance is excluded due to lacking H i absorption at the velocity near the tangent point (e.g., V_LSR ∼ 85 km s⁻¹ or d_tangent ∼ 6.4 kpc; see Figures 3 and 4 in Lockman et al. 2007).

We find that our new kinematic distance of 4.9 ± 0.4 kpc is somewhat smaller than the value of 5.5 ± 0.2 kpc from Lockman et al. (2007). The discrepancy of the two estimates comes from the different parameters of the Galactic rotation curve model. In the work of Lockman et al. (2007), they used a flat rotation curve with R₀ = 8.5 kpc (distance of the Sun from the Galactic Center) and V₀ = 220 km s⁻¹ (rotation speed of the Galaxy at R₀). Accordingly, the LSR velocity of 75 ± 6 km s⁻¹ from the atomic gas yielded the kinematic distance of 5.5 ± 0.2 kpc. The velocity uncertainty of ±6 km s⁻¹ in their work is the typical random motion of cool H i clouds. In our case, we used the new Galactic rotation curve model, in which the values of R₀ = 8.34 kpc and V₀ = 240 km s⁻¹ are adopted (see Table 4 in Reid et al. 2014). Thus, the 77 ± 5 km s⁻¹ from the CO gas leads to a distance of 4.9 ± 0.4 kpc. The velocity uncertainty of ±5 km s⁻¹, which is roughly comparable to the value of ±10 km s⁻¹ for the typical peculiar motions of the high-mass star-forming regions (Reid et al. 2014), is from the LSR velocity difference of the two MCs G39.315−1.155 and G40.331−4.302 (see Figures 6–9).

Finally, the distance to SS 433 is estimated to be about 4.5–5.5 kpc (depending on the different authors with somewhat different considerations; e.g., Hjellming & Johnston 1981; Vermeulen et al. 1993; Stirling et al. 2002; Blundell & Bowler 2004; Marshall et al. 2013; Panferov 2014) based on the kinematic model of the proper motions of the SS 433 jets. Our near kinematic distance of 4.9 ± 0.4 kpc is consistent with the above estimates from the traditional kinematic model originally pioneered by Hjellming & Johnston (1981).

We construct a schematic diagram to elucidate the association between the observed CO+H i features and the SS 433/W50 system (Figure 13). The main features of the ISM surrounding SS 433/W50, such as the H i cavity near the Galactic plane, the H i wall toward the approaching gas from the SS 433 jet, and the corresponding kinematic characteristics of the gas and the jets, are all included in the map.

Zoom In Zoom Out Reset image size

In Figure 13, the opening angle from the CO data is about ±10°, which is comparable to the results from the radio and X-ray studies of W50 at large distances (e.g., jet–ISM encounters at ∼35–80 arcmin from SS 433; Brinkmann et al. 1996, 2007; Dubner et al. 1998), but much smaller than the ±20° of the current precession cone of the SS 433 jets at the inner region (e.g., ≲6 arcsec from the compact source; Hjellming & Johnston 1981; Stirling et al. 2002). The discrepancy of the two opening angles with respect to SS 433 is possibly due to the changing state of the jet precession with time (e.g., Kochanek & Hawley 1990; Zavala et al. 2008; Goodall et al. 2011a) and/or some hydrodynamical recollimated mechanisms for a precessing jet (e.g., Eichler 1983; Peter & Eichler 1993; Monceau-Baroux et al. 2015).

Whatever the exact scenarios, the violent jet–ISM interactions seem to be taking place around the precession axis of SS 433 with a small opening angle. The bulk kinetic energy and momentum of the high-velocity gas will produce the high-pressure environment, leading to the rapid H₂ formation in such interaction regions (see Section 6). As a result, the newly formed molecular gas is likely distributed around an area near the outermost parts of the jet cones with a limited opening angle. On the other hand, the SS 433's jets can shock the ISM and produce the overpressured environment with respect to the surrounding gas, which may inflate a bubble propagating away from the central source. The cavity-like structures of the atomic gas are probably such a case (e.g., the roughly "eight-shaped" morphology of the atomic gas in Figure 11(d)). The size of the upper cavity-like feature of the atomic gas is much smaller than that of the lower one, which is also probably due to the relatively higher density environment close to the Galactic plane.

SS 433 shows prominent jet activity, which deposits amounts of kinetic energy into its surrounding ISM from several tens of astronomical units to dozens of parsecs (e.g., see Figure 8 in Fabrika 2004). Now that the energetic jets of SS 433 may leave some mark on its ambient ISM, it is interesting to search for some corresponding counterparts of jet–ISM interactions. Several works have been done in earlier studies using radio (e.g., Hjellming & Johnston 1981; Downes et al. 1986; Dubner et al. 1998; Blundell & Bowler 2004; Lockman et al. 2007), X-ray (e.g., Watson et al. 1983; Brinkmann et al. 1996, 2007; Safi-Harb & Ögelman 1997), and optical (e.g., Zealey et al. 1980; Boumis et al. 2007) observations. In this section, we propose that the energetic microquasar SS 433 probably has a more profound effect on its ambient ISM.

We turn our focus on the unusual MCs G39.315−1.155 and G40.331−4.302, which are outside the current radio boundaries of W50. Briefly, the intensity of ¹²CO emission of the two MCs is about 1–3 K and shows non-Gaussian profiles, while the ¹³CO emission of them is only marginally detected in some regions with relatively strong ¹²CO emission (Figures 8 and 9). Both of the two MC concentrations exhibit fragmented structures with weak diffuse CO emission around the relatively strong CO peaks (see Figure 5). These fragmented clouds, which have somewhat different peak LSR velocities, display broad CO-line profiles (e.g., Figures 6–9).

We suggest that the two clouds are the fossil record of interactions between the SS 433 jets and the surrounding ISM. The evidence is summarized as follows:

• (1)  
  Spatial coincidence. The two MC concentrations are exactly aligned along the precession axis of SS 433 within a small opening angle of ∼±7° (Section 3.4). Both of the MCs, which appear to be elongated from the northeast to the southwest, seem to form a concave appearance or arc-shaped structure toward the direction of SS 433 (e.g., Figures 4, 5, and 12). The curvature of the two arc-like structures also points in the direction of SS 433. These features are similar to the laboratory experiments (e.g., see Figure 4 in Lebedev et al. 2002), magnetohydrodynamical simulations (e.g., Asahina et al. 2014), and other astrophysical systems with jet–ISM interactions (e.g., examples for the Herbig-Haro complex HH 1-2, Hester et al. 1998; the black hole X-ray binary GRS 1915+105, Tetarenko et al. 2018; and the Seyfert 2 galaxy IC 5063, Oosterloo et al. 2000; Tadhunter et al. 2014; Morganti et al. 2015).
• (2)  
  Kinematic features. The kinematics of the two clouds is consistent with the jet properties of SS 433. The receding jet of SS 433 may be responsible for the redshifted feature of G39.315−1.155, and the approaching jet is responsible for the blueshifted feature of G40.331−4.302 (Section 4). The CO emission of the MCs also exhibits broad-line profiles and multipeaks with slightly different LSR velocities (Figures 6–9). The angular distance of MC G39.315−1.155 from SS 433 is much nearer than MC G40.331−4.302, which agrees with the fact that MC G39.315−1.155 is becoming much more turbulent (Section 3.4) in the relatively high density gas environment close to the Galactic plane. The high turbulence and multiple gas components in the clouds may originate from the shock process of SS 433.

The dynamical timescale of the jet process can be estimated as $\sim \displaystyle \frac{\mathrm{length}}{\mathrm{velocity}}\gtrsim 2\times {10}^{3}$ years when we consider the possible deceleration of the jets. Here, the angular distance of G40.331−4.302 from SS 433 is measured to be ∼215 (or ∼180 pc at a distance of 4.9 kpc), and the jet velocity is assumed to be at a constant of ≲0.26c.

On the other hand, the powerful jets of SS 433 may accumulate considerable material at the end of the jet shock. For G40.331−4.302, the velocity difference along the LOS (or the radial velocity component) is about 7 km s⁻¹ (Figures 7, 9, and 12). Assuming that the gas moves roughly along the precession axis of the SS 433 jets, the total velocity of the gas is $\displaystyle \frac{{\rm{\Delta }}{V}_{\mathrm{LSR}}}{\cos (i)}\,\sim$ 40 km s⁻¹, where i represents the inclination angle of the shock to our LOS (e.g., i ≈ 80°; Margon 1984; Vermeulen et al. 1993; Eikenberry et al. 2001).

The atomic gas surrounding MC G40.331−4.302 displays multilayers with several km s⁻¹ velocity difference (e.g., the red, green, and blue emission in Figure 12). The separation of the atomic-gas layers is measured to be ∼4' (or 5.7 pc at a distance of 4.9 kpc), which is approximately half of the thickness of the whole gas slab (Section 3.4). Using the above velocity of the shocked molecular gas of ∼40 km s⁻¹, we obtain the dynamical age of the moving gas of $\displaystyle \frac{3}{5}\times \displaystyle \frac{5.7\,\mathrm{pc}}{40\,\mathrm{km}\,{{\rm{s}}}^{-1}}\,\sim 0.8\times {10}^{5}$ years (e.g., Δl_gas ∝ t^3/5; Kaiser & Alexander 1997). The dynamical age of the approaching atomic gas is much larger than the timescale of the current jet process, supporting our hypothesis of previous dynamical interactions between the SS 433 jets and its surrounding ISM at ∼10⁵ years ago.

The kinetic energy of the disturbed gas around G40.331−4.302 is estimated to be ∼ 5 × 10⁴⁹ erg (e.g., gas mass of ∼3 × 10³ M_⊙ and velocity of ∼40 km s⁻¹). Adopting the kinetic luminosity of ≳10³⁹ erg s⁻¹ for the jets (e.g., Fabrika 2004; Begelman et al. 2006), the kinetic energy of SS 433 is sufficient to power the disturbed gas on the timescale of ≲0.8 × 10⁵ years for the energy transfer efficiency of 2%. On consideration of possible intermittent jet activities of SS 433, higher values of the kinetic energy input and/or the energy transfer efficiency are possible.

Recently, van den Heuvel et al. (2017) proposed that the SS 433 system, which consists of a compact object of 4.3 ± 0.8 M_⊙ and a supergiant donor star of 12.3 ± 3.3 M_⊙ (Hillwig & Gies 2008), can stably survive for the timescale of 10⁴–10⁵ years. Because the massive donor has a radiative envelope, the system can avoid going into a common-envelope phase and is able to gently spiral in with stable Roche-lobe overflow (King & Begelman 1999; King et al. 2000; van den Heuvel et al. 2017). The timescale of ∼10⁴–10⁵ years is controlled by the thermal timescale of the envelope of the 12.3 M_⊙ A-supergiant in the SS 433 system (van den Heuvel et al. 2017). We find that such a timescale is also consistent with the average time spent in the appearance of the binary systems as ultraluminous X-ray sources state of ∼10⁵ years (Pavlovskii et al. 2017).

For the moving atomic gas surrounding G40.331−4.302, the dynamical age of ∼0.8 × 10⁵ years estimated above is comparable to the evolutionary timescale of the unusual Be/X-ray binary with stable Roche-lobe overflow (e.g., for the case of the mass ratio <3.5; van den Heuvel et al. 2017), which is also consistent with the rapid H₂ formation timescale of ≲0.1 Myr for the gas far away from the Galactic plane at high pressure (see Section 6).

The previous jet episodes (or possible intermittent jet activities; e.g., Goodall et al. 2011a, 2011b), the winds from the system (e.g., Begelman et al. 1980, 2006; Konigl 1983; Fabrika 2004; Panferov 2017), and the loop magnetic field on a large scale (Farnes et al. 2017) may play significant roles in the formation of the SS 433/W50's current configuration.

Figures 10–12 display that a wall structure is clearly seen toward the locus of the extension of the precession axis of the SS 433 jets, in which the column density of H i increased precipitously. Enhanced CO emissions are also found to be nicely associated with the wall structure seen in H i emission. MC G40.331−4.302 has the H i counterpart, which displays more extended arc-like structure nearly perpendicular to the precession axis of the SS 433 jets (Figure 12). The positional coincidence between G40.331−4.302 and the precession axis of SS 433, as well as their coincident kinematic features, strongly suggests the physical connection between them (see Sections 4 and 5).

We obtain that the column density, N_H = N(H i) + 2 × N(H₂), is about 2.1 × 10²¹ cm⁻² for the cloud G40.331−4.302, which corresponds to A_V ≈ 1.1 mag assuming N_H ≈ 1.9 × 10²¹ cm⁻² mag⁻¹ × A_V. At the column density, the molecules are shielded from UV radiation, and self-shielding allows significant H₂ molecules to exist. The total mass of the main concentration of the cloud is ∼3 × 10³ M_⊙ within a radius of ∼66 (Table 1). The molecular fraction ${f}_{{{\rm{H}}}_{2}}=2N({{\rm{H}}}_{2})$/N_H is about 0.87, suggesting that the dominant of the gas in the cloud is in molecular form.

Table 1.  Properties of Gas toward G39.315−1.155 and G40.331−4.302

Name	Tracer	Areaa	Radiusb	Imeanc	Column Densityd	Masse	Volume Densitye,f
		(arcmin2)	(arcmin)	(K km s−1)	(×1020 cm−2)	(×103 M⊙)	(cm−3)
G39.315−1.155	12CO	136.5	6.6	8.8	N(H2) = 17.6	$M({{\rm{H}}}_{2})=10.7{d}_{4.9}^{2}$	$n({{\rm{H}}}_{2})=45.5{d}_{4.9}^{-1}$
G40.331−4.302	H i	135.0	6.6	154.2	N(H) = 2.8	$M({\rm{H}})=0.8{d}_{4.9}^{2}$	$n({\rm{H}})=7.3{d}_{4.9}^{-1}$
	12CO	53.3	4.1	4.6	N(H2) = 9.2	$M({{\rm{H}}}_{2})=2.2{d}_{4.9}^{2}$	$n({{\rm{H}}}_{2})=38.2{d}_{4.9}^{-1}$

Notes.

^aThe area of the emission with I(CO)>5 K km s⁻¹ for G39.315−1.155 and I(H i)>130 K km s⁻¹ and I(CO)>2.5 K km s⁻¹ for G40.331−4.302. ^bThe effective radius is calculated from (Area/3.14)^0.5. ^cThe mean intensity of the H i (70–90 km s⁻¹) and CO (73–88 km s⁻¹) emission. ^dThe column density of the atomic and molecular gas is calculated from 1.823 × 10¹⁸ × I_mean(H i) and 2 × 10²⁰ × I_mean(CO), respectively. See the text. ^eParameter d_4.9 is the distance to the cloud in units of 4.9 kpc (see Section 4). ^fThe volume density of the atomic and molecular gas is calculated from the mass and the effective radius of the emission.

Download table as:  ASCIITypeset image

The virial mass exceeds the total mass of the cloud by more than one order of magnitude, indicating that the cloud is in a gravitationally unbound state. The fragmented subclouds (e.g., with a size of ∼1 pc) in MC G40.331−4.302 may dissipate on a timescale of several Myr (sound crossing time) without other support. Alternatively, the MC may be confined by an external pressure of ≳4.4 × 10⁴ K cm⁻³ (e.g., ρ (H i) σ²(H i)) when the magnetic pressure and the self-gravity are negligible.

We emphasize that G40.331−4.302 is the only detected MC within the range of l = 349–451, −51 ≲ b ≲ −30, and V_LSR ≳ 70 km s⁻¹. It indicates that such an MC is located about 370 pc below the Galactic plane, where a kinematic distance of 4.9 kpc is adopted. However, the MCs within the solar circle are concentrated in the Galactic plane with a Gaussian FWHM of ∼90–100 pc (e.g., see Section 4.5 and Figure 6 in Heyer & Dame 2015). This suggests that G40.331−4.302 is an unusual MC at the locus far away from the Galactic plane. An interesting question is how to form the enormous molecular gas with a limited gas supply.

First of all, the formation timescale of the molecular gas, which was converted from the atomic gas, should be considered. The H₂ formation timescale, ${t}_{{{\rm{H}}}_{2}}\simeq \displaystyle \frac{{10}^{9}\ \mathrm{years}}{n\ ({\mathrm{cm}}^{-3})}$ (Hollenbach et al. 1971), is at least 10 Myr assuming the mean number density of ∼100 cm⁻³ (${n}_{{\rm{H}}}+2\times {n}_{{{\rm{H}}}_{2}}$; see Table 1). However, the above estimation does not take into account dynamical processes, which are ubiquitous in the universe. The processes such as shocks, turbulence, and instabilities have a great impact on the effective H₂ formation rate (e.g., Glover & Mac Low 2007).

Here, we propose that the previous jets/outflows of SS 433 may play crucial roles in the process of molecular gas formation of G40.331−4.302. The H₂ formation timescale can be simply written as ${t}_{{{\rm{H}}}_{2}}\simeq 7\times {10}^{5}{f}_{\mathrm{dust}}{\left(\displaystyle \frac{2\times {10}^{5}{\rm{K}}{\mathrm{cm}}^{-3}}{{p}_{\mathrm{th}}}\right)}^{0.95}\ \mathrm{years}$, connecting the H₂ formation timescale with the thermal pressure (here, f_dust is the dust mass fraction remaining in the gas; Guillard et al. 2009). In our case, the ram pressure, which is defined as p_ram = ρv², is probably dominant in the clouds of interest. For the ram pressure of ∼2 × 10⁶ K cm⁻³ (e.g., n ∼ 7.3 cm⁻³ and v ∼ 40 km s⁻¹), and f_dust = 1, the H₂ formation timescale is thus ≲0.1 Myr, indicating very rapid MC formation. Several numerical models were proposed that H₂ formation occurs rapidly with dynamical processes considered (see, e.g., Koyama & Inutsuka 2000; Bergin et al. 2004; Glover & Mac Low 2007; Glover et al. 2010). The high external pressure due to the prevailing shock was also suggested to accelerate the transition from atomic gas to molecular gas in a short time (e.g., Röhser et al. 2014; Kaneko et al. 2017).

Furthermore, in the vicinity of G40.331−4.302 there may be considerable dark molecular gas (DMG; e.g., Grenier et al. 2005; Abdo et al. 2010; Langer et al. 2010, 2014; Planck Collaboration et al. 2011), which is nearly invisible using CO rotational emission but considerable H₂ molecules may exist in CO-dark or CO-faint regions. Simply speaking, the scale height (l_height) of the gas layer probably follows the relationship of l_height(H i) > l_height(DMG) > l_height(CO) when we see the edge-on Milky Way. Once there exists sufficient H₂ molecules, CO formation may occur rapidly in the turbulent DMG by considering the accumulation and compression of the material by ram pressure from the jet. The possible scenario can explain the unusual scale height of MC G40.331−4.302, which is nearly 370 pc below the Galactic plane. Observations such as OH 18 cm lines, CH 3.3 GHz lines, and 158 μm [C ii] lines toward the direction are advocated to investigate the nature of the gas for further study.

In combination with CO data from the MWISP project and H i data from GBT 100 m and Arecibo 305 m telescopes, we investigate the large-scale ISM environment toward the SS 433/W50 system. Our main findings and comments are summarized as follows:

• 1.  
  Two MC concentrations, MC G39.315−1.155 at V_LSR ∼ 73–77 km s⁻¹ and MC G40.331−4.302 at V_LSR ∼ 77–84 km s⁻¹, are found to be well aligned along the precession cone of SS 433 jets within a smaller opening angle of ∼±7°, indicating the possible connection between them. Further analysis shows that kinematic features of the molecular gas are consistent with the characteristics of the SS 433 jets. That is, MC G39.315−1.155 toward the Galactic plane displays redshifted profiles for the receding gas, while MC G40.331−4.302 away from the Galactic plane exhibits somewhat blueshifted features for the approaching material.
• 2.  
  The two MCs have corresponding atomic features traced by H i emission at V_LSR ∼ 70–90 km s⁻¹. For G39.315−1.155, the molecular gas is located at the top of the atomic gas cavity (see Figure 11(d)), although the strong H i emission near the Galactic plane prevents us from analyzing its kinematic feature of the associated atomic gas. For G40.331−4.302, the H i emission also exhibits the blueshifted velocity structure, as well as that of the embedded molecular gas in the apex of the atomic gas wall. Additionally, the H i emission at such velocities actually displays the morphological resemblance with the radio nebula W50.
• 3.  
  Based on the CO and H i data, the gas at V_LSR = 77 ± 5 km s⁻¹ is suggested to be physically associated with the SS 433/W50 system. The LSR velocity of the associated gas from radio observations is consistent with those from optical observations. Accordingly, the near kinematic distance of the system is 4.9 ± 0.4 kpc, which agrees well with the results from the kinematic model of the proper motions of the SS 433 jets within the error (e.g., see the views in Panferov 2014). The far distance of the system can be excluded due to lacking the H i absorption near the tangent point, e.g., V_LSR ∼ 85 km s⁻¹.
• 4.  
  The interesting MCs, which are exactly at the locus of the extension of the precession axis of the SS 433 jets, are outside the current boundaries of the extended nonthermal radio emission of W50. We argue that such gas is probably related to the very early process of the SS 433 jets and is the fossil record of jet–ISM interactions. Analyzing the moving atomic gas surrounding MC G40.331−4.302, we suggest that the dynamical process of the jet–ISM collisions probably began ∼0.8 × 10⁵ years ago.
• 5.  
  Moreover, G40.331−4.302 is the only MC detected in CO emission at such a low latitude (∼370 pc away from the Galactic plane). Connecting the blueshifted gas of G40.331−4.302 with the approaching jet of SS 433, we argue that the energetic jets of the microquasar have more profound effects on its ISM environment, leading to rapid MC formation of ≲0.1 Myr for the high ram pressure of ∼2 × 10⁶ K cm⁻³. Such a rapid H₂ formation timescale is also comparable to the dynamical age of the moving atomic gas around MC G40.331−4.302 and the evolutionary timescale of the unusual binary system (e.g., a low-mass black hole and a relatively moderate-mass donor star; van den Heuvel et al. 2017).
• 6.  
  There may be considerable H₂ gas with a little CO emission in the surroundings of the fragmented MC G40.331−4.302. We are advocating that further observations such as OH 18 cm lines, CH 3.3 GHz lines, and 158 μm [C ii] line are helpful in investigating the nature of the CO-faint MC and the nearby regions.

The authors acknowledge the staff members of the Qinghai Radio Observing Station at Delingha for their support of the observations. We would like to thank the anonymous referee for valuable comments and suggestions that helped to improve this paper. This work is supported by the National Key R&D Program of China through grants 2017YFA0402701, 2017YFA0402702, 2017YFA0402600, and 2015CB857100. Y.S. thanks F. J. Lockman for providing us with the GBT HI data in our preliminary study. Y.C. thanks the National Natural Science Foundation of China for the support through grants 11773014, 11633007, and 11851305. X.C. acknowledges support by the NSFC through grant 11473069. This publication utilizes data from the Galactic ALFA HI (GALFA HI) survey data set obtained with the Arecibo L-band Feed Array (ALFA) on the Arecibo 305 m telescope. Arecibo Observatory is part of the National Astronomy and Ionosphere Center, which is operated by Cornell University under Cooperative Agreement with the U.S. National Science Foundation. The GALFA HI surveys are funded by the NSF through grants to Columbia University, the University of Wisconsin, and the University of California.

Facility: PMO 13.7m. -

Software: GILDAS/CLASS (Pety 2005).