The Discovery of the Largest Gas Filament in Our Galaxy, or a New Spiral Arm?

It has been realized that filaments are one of the basic structures in the interstellar medium (Myers 2009; André et al. 2014). With high spatial resolutions and sensitivities in the submillimeter regime, Herschel observations reveal the ubiquitous presence of filaments in the dense parts of molecular clouds, which are also most closely related to star formation (André et al. 2010; Men'shchikov et al. 2010; Schneider et al. 2012, 2013; Gong et al. 2018; Zhang et al. 2019). The largest elongated molecular cloud structures are called giant (molecular) filaments with lengths greater than 10 pc. Jackson et al. (2010) have identified an archetypal giant filament "Nessie" with an extreme aspect ratio (80 × 0.5 pc) and claimed that the dense cores within it are the birthplaces of massive stars. Goodman et al. (2014) claimed that Nessie has a much larger length of 430 pc and referred it as the "bone" of the Scutum–Centaurus arm. After Nessie, a large number of observational studies aimed at identifying and characterizing giant filaments have been performed in various tracers, ranging from extinction maps at mid-infrared (Ragan et al. 2014; Zucker et al. 2015) to far-infrared/submillimeter dust emission (Wang et al. 2015; Abreu-Vicente et al. 2016) and CO line emissions (Su et al. 2015; Xiong et al. 2017; Li et al. 2018a, 2020).

Zucker et al. (2018) have performed a comprehensive analysis of the physical properties of large-scale filaments in the literature. Possibly related to the mechanisms shaping the structure and dynamics of the Milky Way (Smith et al. 2014a), together with the spatial resolution and sensitivity limitations of the observations, no giant filament is found in the Extreme Outer Galaxy (EOG) where the Galactocentric distance (R_gc) is greater than 15 kpc. 85% of giant filaments lie mostly parallel to <45°, and in close proximity to <30 pc, the Galactic plane. There are 30%–45% of giant filaments that are associated with spiral arms. The lengths of the giant filaments range from 11–269 pc and the furthest filament is located at R_gc ∼ 12 kpc, well within the most remote spiral arms (R_gc ∼ 20 kpc; Dame & Thaddeus 2011; Sun et al. 2015; Reid et al. 2016) and the furthest molecular clouds (R_gc ∼ 30 kpc; Digel et al. 1994; Matsuo et al. 2017).

Compared to giant molecular filaments, H i filaments are not well studied. Kalberla et al. (2016) and Soler et al. (2020) systematically searched for H i filaments in the Galaxy using the large-scale H i surveys Galactic All-Sky Survey (GASS; McClure-Griffiths et al. 2009) and the HI/OH Recombination line survey of the inner Milky Way (THOR; Beuther et al. 2016; Wang et al. 2020), respectively. Most of the H i filaments are aligned with the Galactic plane, which is similar to the situation of giant molecular filaments. Combined with the Planck all-sky map of the linearly polarized dust emission at 353 GHz, Kalberla et al. (2016) found that H i filaments tend to be associated with dust ridges and aligned with magnetic fields. They claimed that the cold neutral medium is mostly organized in sheets and that they are observed as filaments due to projection effects. The H i filaments are normally cold with a typical excitation temperature T_ex ∼ 50 K and are often associated with CO dark molecular gas (Kalberla et al. 2020). However, detailed physical properties of H i filaments, as well as their distribution in the Galaxy, are not well characterized. A specific case is presented in Soler et al. (2020): using the 40'' resolution observations in the THOR survey, they identified a very long H i filament "Magdalena" in the inner Galaxy (R_gc ∼ 12 kpc) with a length exceeding 1 kpc (see Figure 9 in their paper).

In this work we present Five-hundred-meter Aperture Spherical radio Telescope (FAST) observations of a newly detected giant filamentary H i structure named Cattail, which is possibly the furthest (R_gc ∼ 22 kpc) and largest (∼1.1 kpc) filament to date. Together with the archival HI4PI data, we find that Cattail could be even much longer (∼5 kpc). A new extension of the Outer Scutum–Centaurus (OSC) arm between 70° < l < 100° is also identified. The observations are described in Section 2 and the results and discussion are presented in Section 3.

2.1. FAST Data

2.2. HI4PI Data

The sky region of 20^h30^m48^s < R.A. < 20^h44^m04^s and 409 < decl. < 434 covers the main part of the Cygnus-X North molecular cloud, which has a velocity range of −30 to 20 km s⁻¹ (Schneider et al. 2010) and is located 1.4 kpc away from the Sun (Rygl et al. 2012; Xu et al. 2013). In Figure 1, however, there remains abundant atomic gas with velocities <−30 km s⁻¹, implying the presence of H i gas far behind Cygnus-X. In the velocity range −170 to −130 km s⁻¹, there appears to be H i emission that is not connected to the main velocity components of the Outer arm. The central velocity of the outermost gas is −151.1 km s⁻¹. According to the rotation model of Reid et al. (2014; 2016), the heliocentric distance of such an H i cloud is ∼21 kpc, corresponding to a distance from the Galactic center of ∼22 kpc. The outermost cloud exhibits a filamentary structure and lies approximately parallel to the Galactic midplane (b = 0, right panel of Figure 1). Its length is about 1.1 kpc, much larger than the giant molecular filament Nessie, and comparable with the large H i filament Magdalena. If the structure, which we name Cattail, lies at the kinematic distance, it is located far beyond all known giant gas filaments (R_gc < 12 kpc), molecular and atomic, in previous research (Zucker et al. 2018). Figure 2 presents the spectra of six positions located within (A1, B1, C1) and immediately outside of Cattail (A2, B2, C2). Compared to A2, B2, C2, the spectra of Cattail (A1, B1, C1) have an additional emission from −170 to −130 km s⁻¹. There is no self-absorption feature in the spectra.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Figure 3 shows the velocity channel map in the H i 21 cm emission of Cattail. It can be seen that the filament exhibits an overall velocity gradient along its major axis. The southwest end of Cattail has a less negative velocity compared to that of the northeast end. Most of the atomic gas of the filament has velocities in the range from −160 to −140 km s⁻¹, with a mean velocity of approximately −150 km s⁻¹. The intensity distribution and the width of the filament are almost uniform from northeast to southwest. The velocity gradient along the filament is estimated to be 0.02 km s⁻¹ pc⁻¹.

Zoom In Zoom Out Reset image size

Similar to Suri et al. (2019), we extract the radial profile of Cattail from the intensity map. First, we derive the direction of the filament spine, which is indicated by a red arrow in the right panel of Figure 1. Then, we extract the perpendicular intensity profile for each pixel along the filament spine. Finally, we average the profiles of all pixels along the filament and apply Gaussian fitting to the mean radial profile. In addition to the emission of Cattail, there is residual emission in the sky region which comes from the velocity distribution of the closer arms (Figure 2). In this case, the calculated width would increase with the fitting range (Smith et al. 2014b; Panopoulou et al. 2017). To be less influenced by this bias, we apply Gaussian fitting to the inner part of the radial profile. The intensity radial profile of Cattail is almost symmetrical (Figure 4). The width (FWHM) of the filament is 207 pc, which means that the aspect ratio of Cattail is about 5:1. Given that the width is comparable to the scale height of gas in a spiral arm (Kalberla & Kerp 2009), Cattail could also be a density enhancement along a spiral arm; we will come back to this point later.

Zoom In Zoom Out Reset image size

Considering that the peak antenna temperature is about 1 K (Figure 2) and there is no H i narrow self-absorption feature in the spectra (see Figure 2), the emission should be optically thin and the column density can be calculated according to the following formula (Wilson et al. 2009):

$$\begin{eqnarray}&&{N}_{{\rm{H}}\,{\rm\small{I}}}\ [{\mathrm{cm}}^{-2}]=1.823\times {10}^{18}\int {T}_{{\rm{B}}}(v){dv}\ [{\rm{K}}\ \mathrm{km}\ {{\rm{s}}}^{-1}]\end{eqnarray} \tag{ 1 }$$

We calculated the mass of Cattail to be 6.5 × 10⁴ M_⊙. The linear mass density is 60 M_⊙ pc⁻¹. The peak column density of Cattail is 1.7 × 10¹⁹ cm⁻² (0.1 M_⊙ pc⁻²), which is close to the median value of H i filaments with latitudes ∣b∣ > 20° (Kalberla et al. 2016) and lower than the threshold (N ≥ 1×10²² cm⁻²) for high-column-density gas adopted by Zucker et al. (2018). Assuming the depth to be the same as the width (207 pc), the peak volume density of the filament is calculated to be 0.03 cm⁻³, which is close to the average Galactic midplane volume density at R_gc = 22 kpc (0.01 cm⁻³) (Kalberla & Kerp 2009). The physical parameters of Cattail are listed in Table 1.

Table 1. Physical Parameters of Cattail

Parameter	Value
Galactic longitude l	8149
Galactic latitude b	065
R.A. αJ2000	20h37m54s
Decl. δJ2000	+42d14m40s
Center velocity	−151.1 km s−1
Kinematic distance D	21 kpc
Galactocentric radius Rgc	22 kpc
Length	1.1 kpc
Width (FWHM)	207 pc
Aspect ratio	5:1
Mass	6.5 × 104 M⊙
Linear mass	60 M⊙ pc−1
Line width ΔV	12 km s−1
Peak column density	1.7 × 1019 cm−2 (0.1 M⊙ pc−2)
Peak volume density	0.03 cm−3

Note. Rows 1–5 give the central positions in the position–position–velocity (PPV) space. The kinematic distance is derived according to the Galactic rotation model A5 of Reid et al. (2014). The volume density is defined as N_{H i} /depth, where depth is assumed to be the same as width.

Download table as:  ASCIITypeset image

Both simulations (Inutsuka & Miyama 1992, 1997) and observations (André et al. 2010; Men'shchikov et al. 2010; Schneider et al. 2012, 2013) have shown that filaments tend to be vulnerable to radial collapse if the line mass is greater than the critical value defined as ${M}_{\mathrm{vir}}=2{\sigma }_{\mathrm{tot}}^{2}/G$ (Fiege & Pudritz 2000). The total velocity dispersion σ_tot can be obtained by the FWHM line width (${\rm{\Delta }}v/\sqrt{8\mathrm{ln}2}$), which is 12 km s⁻¹ for Cattail. H i filaments are normally cold with a typical excitation temperature T_ex ∼ 50 K (Kalberla et al. 2020), corresponding to an isothermal sound speed of approximately 0.65 km s⁻¹. Therefore, the turbulence may be supersonic in Cattail. The critical line mass M_vir is calculated to be 1.1 × 10⁴ M_⊙ pc⁻¹, which is much larger than the linear mass density M_line (60 M_⊙ pc⁻¹). Thus, Cattail is gravitationally unbound and may be in an expanding state, unless confined by an external pressure (Fischera & Martin 2012).

A velocity range of ∣V_LSR∣ > 90 km s⁻¹ is generally used to separate high-velocity clouds (HVCs) from gas at low and intermediate velocities. However, this operational definition is insufficient considering that the velocities of some clouds can be well understood in terms of Galactic rotation but exceed 90 km s⁻¹ (Reid et al. 2016). Similarly with Wakker (1991), we adopt a "forbidden velocity" of ∣V_LSR − V_rot∣ > 50 km s⁻¹ to identify HVCs, where V_rot is the rotation curve of the Milky Way. In the direction of Cattail, the filament can be an HVC only when its heliocentric distance is within 12.5 kpc (R_gc < 14 kpc). However, the Galactic latitude of Cattail is 065, which would put it within the flaring FWHM of the disk when R_gc < 14 kpc (Kalberla & Kerp 2009). As HVCs travel through hot halo gas, they would be heated and photon ionized, which causes thermal instabilities between the skin and the inner region of the clouds (Putman et al. 2012). Moreover, the headwind pressures can induce dynamical instabilities, e.g., shear-driven disturbance in HVCs, which makes them fragment into head-tail clouds (Butler Burton et al. 2004) and increase their line widths (Putman et al. 2011). The H i 21 cm line width of Cattail is 12 km s⁻¹, significantly smaller than the typical value of 20–30 km s⁻¹ for HVCs (Lockman 2003; Tumlinson et al. 2017; Barger et al. 2020). The cloud does not seem to be fragmenting. Also, Cattail has a filamentary morphology parallel to the Galactic midplane with a mass less than 10⁵ M_⊙, and if it is an HVC, it would have been fully ionized while passing through the Galaxy halo (Bland-Hawthorn et al. 2007; Heitsch & Putman 2009; Kwak et al. 2011). Therefore, the properties of Cattail do not appear to be compatible with those of an HVC.

Assuming that the velocity of Cattail is attributed to the Galactic rotation, the structure is located beyond all the known spiral arms in the first quadrant. We then examine other possible origins of the extreme velocity of Cattail. We have discussed above that Cattail does not seem to be an HVC. Combining the results of Anderson et al. (2014) and Green (2014), there are no supernova remnants or H ii regions with regular sizes >25 near Cattail. There is no dwarf satellite galaxy near Cattail either (McConnachie 2012). Hence, there is no evidence for Cattail's velocity coming from the influences of stellar feedback or a tidal stream, and the most likely explanation is the Galactic rotation. Moreover, taking the column density and the depth of Cattail to be 1.7 × 10¹⁹ cm⁻² and ∼05, respectively, the volume density of Cattail would be significantly lower than the typical value of the Milky Way when it is located at R_gc < 15 kpc (Kalberla & Kerp 2009). Both a filament and a spiral arm should have a large density contrast compared to the surrounding medium. Therefore, we consider that Cattail is located at a heliocentric distance and Galactocentric distance of ∼20 kpc.

Using the HI4PI data, we have investigated the complete morphology of Cattail (Figure 5). Cattail exhibits an even larger dimension while retaining a coherent velocity of −160 to −140 km s⁻¹. The FAST observations cover the southwest end of the full structure while the northeast end is located at about l = 93° and b = 55. We designate them as the FAST end and HI4PI end in this work, respectively. The full length of the complete Cattail reaches ∼15° (5 kpc).

Zoom In Zoom Out Reset image size

In the velocity range from −130 to −110 km s⁻¹, there is another new elongated structure between 70° < l < 100°. From its morphology (Figure 5) and velocity (Figure 6), it is apparently the extension of the OSC arm previously only seen at 0° < l < 70° (Dame & Thaddeus 2011). The full structure of the OSC arm in the first Galactic quadrant is then seen for the first time. In projection, Cattail is connected to the OSC arm at about l = 93°. However, Cattail does not appear to be physically connected to the OSC arm. First, the velocity of Cattail is different from that of the OSC arm by 30 km s⁻¹. Second, if Cattail is connected to the OSC arm at a Galactic longitude of about 93° and extends to about 80°, it means that Cattail joins the OSC arm at the downstream end with respect to the Galactic rotation and leaves the arm toward the upstream end, i.e., in a pattern opposite to the inertial direction of the Galactic rotation; such a structure is inconsistent with the picture of a spur originating from a spiral arm (e.g., Dobbs & Bonnell 2006). Thus, when combining the large-scale map of HI4PI, Cattail may have a much larger dimension but still appears to be located well above the OSC arm.

Zoom In Zoom Out Reset image size

Based on the above analysis, we suggest two possible explanations for Cattail: it is a giant filament with a length of ∼5 kpc, or part of a new arm in the EOG (Figure 7). Until recently, simulations of giant filaments appeared in the literature. As the numerical simulations show, the shear from Galactic rotation plays a critical role in the formation of large-scale filaments (Smith et al. 2014a; Duarte-Cabral & Dobbs 2016). Since the shearing motion tends to stretch out and align gas with spiral arms, giant filaments tend to form in spiral arms and interarm regions. Unlike most of the giant molecular filaments that are associated with spiral arms and lie within 30 pc from the physical Galactic midplane (Wang et al. 2016; Zucker et al. 2018), Cattail is far beyond the outermost OSC arm. Figure 7 shows that the Galactic disk is significantly warped in the first quadrant of the EOG, which is in line with the forecasts from Levine et al. (2006), Kalberla et al. (2007), and Kalberla & Kerp (2009). The Z-scale height of the HI4PI end is about 2 kpc, which approximates the warped scale of the physical Galactic midplane at that Galactocentric distance. Comparatively, the FAST end of Cattail is located at a Galactic latitude of 065, corresponding to about 200 pc upon the IAU-defined midplane (b = 0°), which is flat from the Galactic center to the edge of the Galaxy. Thus, the complete Cattail may originate from the warped physical Galactic midplane at a longitude of ∼93° and extend to a longitude of ∼80° with Z ∼ 200 pc. If Cattail is a gas filament located beyond the OSC arm, how is such a huge structure formed? Alternatively, if Cattail is part of a new spiral arm in the EOG, it is also puzzling that the new arm does not fully follow the Galactic warp, given that the FAST end is about 1.8 kpc away from the warped Galactic disk midplane. While these questions remain open with the existing data, the observations provide new insights into our understanding of the Galactic structure.

Zoom In Zoom Out Reset image size

This work is supported by the National Key R&D Program of China No. 2017YFA0402600 and the National Natural Science Foundation of China (NSFC) grant U1731237. K.Q. acknowledges the science research grant from the China Manned Space Project with No. CMS-CSST-2021-B06. C.L. acknowledges the supports from NSFC grant 12103025, China Postdoctoral Science Foundation No. 2021M691532, and Jiangsu Postdoctoral Research Funding Program No. 2021K179B. Y.C. is partially supported by the Scholarship No. 201906190105 of the China Scholarship Council and the Predoctoral Program of the Smithsonian Astrophysical Observatory (SAO). We thank the referee for the thoughtful comments which improved this paper. We would like to thank the FAST staff for their support during the observation and thank Junzhi Wang, Zhiyu Zhang, and Bing Liu for the helpful discussions on data reduction. This work makes use of publicly released data from the HI4PI survey which combines the EBHIS in the Northern Hemisphere with the GASS in the Southern Hemisphere. EBHIS is based on observations with the 100 m telescope of the MPIfR (Max-Planck-Institut fur Radioastronomie) at Effelsberg. The Parkes Radio Telescope is part of the Australia Telescope which is funded by the Commonwealth of Australia for operation as a National Facility managed by CSIRO.