Discovery of Extra-Planar H i Clouds and a H i Tail in the M101 Galaxy Group with FAST

To show the H i (1420.4058 MHz) gas distribution surrounding M101, we mapped a 2${}^{\circ}\times$ 2^∘ region centered at position of R.A.= 14:03:12.6 and Dec.= 54:08:34.4 using the Five-hundred-meter Aperture Spherical radio Telescope (FAST), during January and April 2021. FAST is located in Guizhou, China. The aperture of the telescope is 500 m, and its effective aperture is about 300 m. The 19-beam array receiver system in dual polarization mode is used as front end. It formally works in the frequency range from 1050 MHz to 1450 MHz. For spectral-line observation, FAST is equipped with two digital backend systems, Spec(F) and Spec(W+N). We select Spec(F) backend, which has a bandwidth of 500 MHz and 1048576 channels, resulting a frequency resolution of 476 Hz and corresponding to a velocity resolution 0.1 km s^-1 at 1.4 GHz.

Mapping observation uses the Multibeam on-the-fly (OTF) mode. This mode is proposed to map the sky with 19 beams simultaneously, and has a similar scanning trajectory. The half-power beam width (HPBW) is about 2.9^′ at 1.4 GHz for each beam. The pointing accuracy of the telescope was better than 8^′′. To satisfy Nyquist sampling, the parameter of rotation angle is available in this mode. If the scan is along the latitude lines, the receiver platform needs to turn 23.4^∘. While along the longitude lines, it needs to turn 53.4^∘. After turning the angle, an interval between two adjacent parallel scans is about 1.14^′. Besides, we set the scan velocity of 15^′′ s^-1 and an integration time of 1 s per spectrum. During observations, system temperature was around 25 K. For intensity calibration, noises signal with amplitude of 10 K was injected under a period of two seconds. A total of 67 minutes is needed to acquire a 2${}^{\circ}\times$ 2^∘ map. In order to improve sensitivity, we observed two times for M101 with the Multibeam OTF mode. Jiang et al. (2019, 2020) gave more details of the FAST instrument.

For each spectrum with a bandwidth of 500 MHz, we intercept 2.2 MHz data for processing, which can cover the whole velocity range of M101. The data reduction was performed with the PYTHON package and GILDAS¹¹1http://www.iram.fr/IRAMFR/GILDAS package. First of all, we will calibrate the unit of the spectrum into kelvin for each polarization signal through the following transformation,

where $T_{A}$ is calibrated antenna temperature, while $T_{\rm cal}$ is noise diode temperature. $P_{\rm on}$ and $P_{\rm off}$ are power values when noise diode is on and off, respectively. The correction for the line intensity to brightness temperature ($T_{B}$) was made using the formula $T_{\rm B}=T_{\rm A}/\eta_{{}_{\rm MB}}$. Here $\eta_{{}_{\rm MB}}$ is the main beam efficiency, which can be determined by $\eta_{{}_{\rm MB}}(\lambda)=0.8899[\theta_{{}_{\rm MB}}/(\lambda/D)]^{2}\eta_{{}_{\rm A}}$ (Downes et al., 1989), where $\theta_{{}_{\rm MB}}$ is beamwidth in radians, $D$ is the effective aperture, $\lambda$ is the observing wavelength, and $\eta_{{}_{\rm A}}$ is the aperture efficiency. The $\theta_{{}_{\rm MB}}$ and $\eta_{{}_{\rm A}}$ values for each beam can be derived from Jiang et al. (2020). Here we only give the main beam efficiency of the central beam (Beam 1), $\eta_{{}_{\rm MB}}$ is $\sim$ 0.75 at 1.4 GHz.

Using the ArPLS algorithm, we mitigate radio frequency interference (RFI) by a fitting procedure of the data in the time-frequency domain. After combining the two polarizations, a first-order polynomial was used to fit baseline structure and subtract continuum sources to a line-free range of each spectrum. The L-band 19-beam receiver of FAST has a low level ($<$5%) of sidelobes near the frequency of 1420 MHz (Jiang et al., 2020; Han et al., 2021), hence we don’t make the sidelobe correction. Once the spectra have been fully calibrated, we apply them to a grid in the image plane. Each grid is adopted as 1.5^′. This grid can be made in a TAN projection projection in the World Coordinate System as described by Calabretta & Greisen (2002).

Finally, we make the calibrated data into the standard fits-cube format. For further processing of fits-cube data, we use GILDAS package, such as channel merging and regriding with a pixel of $1.5^{\prime}\times 1.5^{\prime}$. The velocity resolution is smoothed to 2.1 $\rm km~{}s^{-1}$, then the noise RMS is about 56 mk (2.6 mJy beam^-1) for the M101 region. To facilitate comparison with the previous observations of M101 by the HI4PI data (Collaboration et al., 2016), the FAST cube data are smoothed and resampled to the same angular resolution and pixel scale with the HI4PI data. Figure 1 shows comparison of the H i spectrum of NGC 5474 in the M101 galaxy group from the FAST map (black line) and HI4PI map (red line). Both spectra are integrated over whole the NGC 5474’s region. It is clear that the two spectra agree within the noise.

Compared with the previous H i observations toward the M101 galaxy group (Huchtmeier & Witzel, 1979; Walter et al., 2008; van der Hulst & Sancisi, 1988), we detected a more extended H i disk around M101 using FAST with a high sensitivity H i observation. On the southern edge of the extended H i disk, we identify three new H i clouds (C1, C2, and C3). The velocity field of the H i cloud C2 very similar to the dwarf companion NGC 5477 shows an irregular velocity gradient. From the position-velocity diagrams in panels c-e of Figure 4, we also see that the H i clouds C1, C2, and C3 similar to NGC 5477 rotate with the H i gas disk of M101. Unlike the companion NGC 5477, however, they have no optical counterparts. This situation is also very similar to nearby galaxy NGC 2403. In the NGC 2403, a newly detected cloud rotates with the main H i disk, see Figure 6 of de Blok et al. (2014).

From the observation of GBT, Mihos et al. (2012) also found two new H i clouds on the southwest of the M101’s H i disk. To distinguish the previously identified H i clouds, we made channel maps overlaid on the DSS $B$-band optical image, see Figure 5. From the panel (a) of Figure 5, we see that except for the H i gas associated with the main disk, there are two H i clouds located on the southwest, indicated by two black arrows. By comparing with velocity range and position, the two H i clouds in panel a should be those identified by Mihos et al. (2012). Hence, here the H i clouds C1, C2, and C3 in panels (b) and (c) of Figure 5 are identified for the first time from the M101’s H i disk.

The total mass of the three H i clouds in M101 can be calculated by $M=2.72mD^{2}\int N(x,y)dxdy$, where the factor 2.72 is the mean atomic weight, $m$ is the mass of atomic hydrogen, $D$ is the adopted distance to the clouds, $dx$ and $dy$ are the pixel sizes. The column density $N(x,y)$ in each pixel can be estimated as $N(x,y)=1.82\times 10^{18}\int T_{\rm B}dv$, where $T_{\rm B}$ is the line brightness temperature, and $dv$ is the channel width. We adopt the distance (6.9 Mpc) of M101 as those of the three H i clouds, yielding that the masses of C1, C2, and C3 are 1.3$\times$10⁷ $M_{\odot}$, 2.4$\times$10⁷ $M_{\odot}$, and 2.0$\times$10⁷ $M_{\odot}$, respectively. The obtained masses are a lower limit because we can only identify cloud emission when they are not projected against the main disk. Meanwhile, we also estimated that the sizes of C1, C2, and C3 are about 18 kpc, 19 kpc, and 15 kpc, respectively. From the Figure 5, we see that the H i clouds C1 and C3 have connected in projection with the main disk, while the H i cloud C2 seem not to overlap with the main disk in panel (c). These extra-planar gas in M101 are similar to the filaments found in NGC 891 (Oosterloo et al., 2007), NGC 2403 (Fraternali et al., 2002; de Blok et al., 2014), and M33 (Sancisi et al., 2008). Sancisi et al. (2008) suggested that such extra-planar gases are the most remarkable found in the halo regions of these galaxies because they may be direct evidence of cold gas accretion from outside.

For the origin of the extra-planar gas, it seems to has two mechanisms (Sancisi et al., 2008), fountain-driven accretion and accretion from intergalactic space. In fountain-driven accretion, the hot gas rises into the halo, condenses into cold clouds and returns to the disk. Here the extra-planar gas may be driven by star formation. Figure 6 shows the far-ultraviolet (FUV) emission image of M101 overlaid with the H i column-density contours. The FUV data, obtained from Dale et al. (2009), are used to trace emission from hot stars (de Blok et al., 2014). However, from the Figure 2 and 6, we did not see the evidences of violent star formation in the projection directions of these clouds and the HVC G1 identified by us. Thus, the fountain-driven accretion can not cause these IVC H i clouds and the HVC G1 detected in M101.

While for the other mechanism, the extra-planar gas is either directly accreting from the intergalactic medium or is the result of a minor interaction with a neigboring dwarf galaxy, as the case in NGC 2403 (de Blok et al., 2014). According to Dekel & Birnboim (2006) and Mihos et al. (2013), if massive galaxies at low redshift undergone a cold flow accretion, they will host a massive hot halo. While the hot gas traced by the X-ray emission in M101 is diffuse and only correlated with the spiral arms (Kuntz et al., 2003). Furthermore, as shown in Figure 6, the FUV emisson in M101 is also similar to the X-ray emission. Thus, the M101 does not host a massive hot, and even warm halo, arguing that the massive galaxy M101 ($M_{\rm tot}\simeq 10^{12}M_{\sun}$, Huchtmeier & Witzel (1979)) would not be experiencing any significant cold flow accretion from the intergalactic medium at the present day (Mihos et al., 2013). Simultaneously, it also means that the main galaxy M101 may not have enough energy to eject such massive clouds found by us and Mihos et al. (2012). Hence, the extra-planar gas in M101 may be the result of a minor interaction with a neigboring dwarf galaxy, but there is still a lack of reliable evidence for the interaction.

From Figure 3, we see that the H i velocity field within the optical disk region of M101 is regular, indicating that the relatively strong disturbance occurs in the outer disk. While the asymmetric outer disk of M101 has also been believed to arise from an interaction with a companion (Beale & Davies, 1969; Rownd et al., 1994; Waller et al., 1997; Mihos et al., 2013, 2018). The two closest companions to M101 are NGC 5477 and NGC 5474. NGC 5477 has a low luminosity ($M_{V}$ =-15.3). Mihos et al. (2012) suggested that it can’t be massive enough to creat such a strong morphological response in a giant Sc spiral like M101. For the other companion NGC 5474, it is more luminous than NGC 5477. Huchtmeier & Witzel (1979) and Mihos et al. (2012) detected some diffuse H i gas between M101 and NGC 5474 in a H i column-density map and a PV diagram, respectively, strengthening the case for a recent interaction between the two galaxies.

While our high sensitivity H i observation with a relatively high resolution (2.9^′) did not detect the diffuse H i bridge connecting M101 with NGC 5474 in a H i column-density map (see Figure 2) in a PV diagram (see the panel b of Figure 4). From Figure 2, we obtain that the distance between M101 and NGC 5474 is about 15^′. While the angular resolutions of the previous obsevations from Huchtmeier & Witzel (1979) and Mihos et al. (2012) are about 9^′, which is nearly close to the distance between the two galaxies. Hence, above two observations may not be able to distinguish the gas between them well. We suggest that the previously detected bridge between the two galaxies is likely due to beam dilution created by the role of low spatial resolution. Generally, galaxy-galaxy interaction can give rise to star formation in tidal debris, while Garner et al. (2021) did not detect an abundance of star formation between M101 and NGC 5474, also suggesting that the gas between the two galaxies is very weak or there is no gas like bridge. We did not detect a H i bridge between the two galaxies, but identify a slightly extended H i disk for NGC 5474. It means that the two galaxies have undergone only a weak interaction, just like the judgement of Garner et al. (2021).

In addition, as shown in Figure 2, we detected a new H i tail on the southern edge of the extended H i disk of M101. Generally, the H i tail indicates ongoing minor merger and recent arrive of external gas, and then can be considered as direct evidence of cold gas accretion (Sancisi et al., 2008). Compared with the positions of the H i tail, the detected new H i clouds seems to can represent other H i tails. Here the H i tails detected in M101 gives a reliable evidence for M101 interaction with NGC 5474. Thus, we conclude that the extra-planar gas (three H i clouds) and the H i tail detected in the M101’s disk are likely to origin from a minor interaction with the companion NGC 5474. Furthermore, because FAST has high sensitivity and relatively high angular resolution, it can detect and separate the relatively weak and small gas structures around galaxies. This will play an important role in the study of galaxy evolution.