The Distance to High-velocity Cloud Complex M

The overall morphology of the area described in this paper is shown in Figure 1, which presents maps of galactic longitude (l) versus latitude (b) of the H i brightness at velocities that highlight some of the well-known features of Complex M. The data are from the λ-21 cm galactic neutral atomic hydrogen HI4PI survey (Ben Bekhti et al. 2016), which combines the northern hemisphere data from the Effelsberg–Bonn Hi Survey (EBHIS; Winkel et al. 2016) and the southern hemisphere data from the third revision of the Galactic All-Sky Survey (McClure-Griffiths et al. 2009). HI4PI has an angular resolution of Θ_FWHM = 162, a sensitivity of σ = 43 mK, and full spatial sampling, 5' in both galactic l and b.

Zoom In Zoom Out Reset image size

The contour intervals in Figure 1 and subsequent figures are chosen to highlight the morphological details. The naming of various H i structures is traditional but can be confusing as data quality improves. What appeared to be isolated "clouds" in maps from old surveys might now be seen as parts of filaments as sensitivity, resolution, and dynamic range improve. This may be the case with the well-known HV feature labeled as MI at (l, b) = (165°, 65°), which looks cloudlike in Figure 1(a) at a velocity of −120 km s⁻¹. Parts of the HV Complex C appear in the lower right corner of this panel.

Maps at lower velocities, such as Figure 1(b) at −100 km s⁻¹, show the connectedness between MI and the filamentary structure that is part of Complex M and includes a loop-the-loop shaped segment, which extends from l = 130° to 160 In some early papers on HV clouds (e.g., Hulsbosch 1968), both the cloud and the loop-the-loop segment were grouped together. MII is located to the left of MI. Different peaks have been given labels, but those are not always consistent in the literature. We designated two of these, MIIa and MIIb, which appear to extend the complex to greater longitudes. It may be that these features, which appeared as clouds in earlier studies are, rather, local density concentrations or line-of-sight enhancements in a long, twisted filament.

Figure 1(c) at −75 km s⁻¹ clearly shows emission bridging the gap between MI and MIIa & MIIb. Because of the connectedness revealed by the HI4PI data, we are able to follow the arching shape from about (l, b) = (105°, 53°) to (l, b) = (196°, 55°). We refer to the overall structure as Complex M throughout this paper. In addition, a distinctly different feature, the well-known IV Arch, appears below Complex M. This brings up a challenge related to the modern study of interstellar H i gas. Where clouds had clear boundaries and were isolated in the sky, filaments connect features in both position and velocity space. Defining the boundaries is not necessarily easy. Kalberla & Haud (2006) studied the connection between HV and IV gas in Complex M. Since this paper presents evidence for a radically different distance, we decided to focus on Complex M where the connections to MI were the strongest and exclude the IV Arch where the connections to MI were weaker/different/nebulous. It may very well be true that the distance to the IV Arch is much closer than the 0.8–1.8 kpc values quoted by Kalberla & Haud (2006), but we have no new evidence (yet) to support this claim. For the purposes of this paper, neither the IV Arch nor Complex C will be considered in the analysis.

Figure 2 shows detailed l − b maps of Complex M using the EBHIS λ-21 cm data, which has a spatial resolution of 108 (Winkel et al. 2016). The panels show how the structure of MI changes with velocity from (a) the familiar two-peaked distribution at −116 km s⁻¹ to (b) the flattened ring at −108 km s⁻¹ (Giovanelli et al. 1973). The remaining panels show how this ring slowly morphs from a relatively isolated HV cloud shape to the extended loop-the-loop segment of Complex M. It is not possible to say definitively where the cloud ends and the filament begins, and there is neither a spatial nor a velocity boundary separating the cloud from the filament. On the higher longitude side, however, the connection between MI and MIIa is much more tenuous. We will refer to this as the M&M gap. Panels (e)–(h) show that this gap is traversed by a narrow bridge of emission in the velocities from −82 to −70 km s⁻¹, implying that whatever created the gap at higher negative velocities did not succeed in removing all the H i at these lower velocities. We will return to the M&M gap in Section 3.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The HI4PI data cube allows us to examine Complex M in different ways. The area maps shown above give us one perspective, but the far less commonly used longitude–velocity maps give us another view. Figure 3 shows a series of l − v plots at different latitudes that cover the longitude range of interest. The figure shows clear H i ridges at several velocities. The low-velocity gas is generally thought to be relatively close to the Sun, certainly less than 100 pc distant at these latitudes. Normally, this gas is centered at zero velocity over the whole sky, and the offset from zero seen here may be related to the presence of the Local Bubble (Zucker et al. 2022). The second ridge of emission at about −60 km s⁻¹ forms part of the IV Arch. The third set of peaks around −85 km s⁻¹ marks the loop-the-loop segment of Complex M.

Zoom In Zoom Out Reset image size

The panels of Figure 3 were chosen to highlight the connectedness of MI and the loop-the-loop segment. In Figure 3(a) at b = +648, MI appears as an isolated cloud with two main peaks. In Figure 3(b) at b = +656, a string of peaks at around −90 km s⁻¹ begins to appear. They seem to be a series of clouds but are, rather, enhancements created where the l − v plot at this latitude intersects the twisted structures that define the loop-the-loop segment. This panel also shows the location of MIIb, which marks the continuation of Complex M to higher longitudes as seen in Figure 1(c). Figure 3(c) shows the strengthening of the connection between MI and the intensifying peaks of the loop-the-loop segment; the IV Arch is also dominant in this panel. Figure 3(d) shows the strongest connection between MI and the lower-longitude gas features. Figure 3(e) shows the M&M gap. In addition, H i in the low-velocity ridge at −5 to −10 km s⁻¹ is missing between l = 162° and 173°. These directions mark the boundaries of the low-velocity cavity created by the MI supernova event discussed by Schmelz & Verschuur (2022). MIIa is brightest in Figure 3(f).

Figure 3 shows how the connection between MI and the loop-the-loop segment evolves with increasing latitude. The connection starts to form in panel (b) and continues to strengthen throughout the image sequence until they appear to blend (morph) in panel (f). In addition, the data show the M&M gap on the high-longitude side. We see in Figure 1 that the arched filament appears to be moving coherently across the area, which indicates that its axis must be nearly in the plane of the sky. Why then is there an asymmetry as MI connects strongly with anomalous-velocity gas at lower longitudes but only tentatively to MIIa and MIIb at higher longitudes? The emission from the bridge discussed above is not plainly evident in these l − v plots because of its narrow angular width and relatively low brightness. It appears along the velocity edge of the unrelated emission that peaks around −60 km s⁻¹. This highlights our contention that position–position maps and position–velocity maps accentuate different facets of H i structure. We will return to this asymmetry in Section 3.

Additional evidence to support our hypothesis that MI is interacting with Complex M comes from Gaussian decomposition of H i emission profiles. The semiautomated Gaussian fitting method described by Verschuur (2004) uses the Microsoft Excel SOLVER ¹ algorithm. The initialization of SOLVER requires a first guess for the Gaussian parameters, including the approximate center velocity, line width, and amplitude of apparent components until the residuals resemble the noise. SOLVER is then run, adjusting the Gaussians to minimize the residuals and then calculating the reduced chi-squared value for the fit. We then examine the solution. If the value of reduced chi-squared is between about 0.9 and 1.1, the fit is judged to be satisfactory. If it is too high or too low, we can add or subtract components, respectively, and rerun SOLVER. This process determines the number of individual Gaussians that are required for a final good fit to the data. If a series of closely spaced profiles are to be decomposed, the best-fit solution for the first profile is used as the starting point for the next in the series. Only components with brightness temperature of ≥0.1 K and column densities ≥10¹⁸ cm⁻² were considered in the statistics. We selected two areas of the EBHIS data cube for comparison. The first involved a mapping of MI using profiles every 02 in longitude from 1670 to 1698 and every 01 in latitude over the range of the HV gas. This sample involved 330 profiles that produced 3182 Gaussians covering all velocities of which 659 pertain to HV H i. A second set of 50 profiles is located along the axis of the relatively straight and narrow section of the Complex M from longitude 1400–1468 and a latitude range that covered the filament. The Gaussian analysis produced 393 components at all velocities of which 68 pertain to HV gas.

Figure 4 shows examples of the H i profiles for (a) MI and (b) the filament segment. Each profile has multiple peaks representing the complex structure seen in Figure 3. The rightmost spectral feature represents the local gas, and the peak at about −60 km s⁻¹ highlights the IV Arch. Although these portions of the profiles are quite different, it is the leftmost peak, with the highest negative velocity, that is of greatest interest to us here. The panels also show the Gaussian fits—one broad component (red) is seen in every direction with one or two weaker narrow components superimposed. The broader components have an average line width of 25.7 ± 3.4 km s⁻¹ for MI and 24.5 ± 1.5 km s⁻¹ for the filament segment, thus indicating that the underlying properties of both HV regions are similar.

Zoom In Zoom Out Reset image size

Gaussian line widths of order 25 km s⁻¹ are traditionally attributed to a warm neutral or ionized medium with a temperature of about 10,000 K, which might be in pressure balance with both colder and hotter interstellar components. If the gas is this warm, then we might expect to see its spectral signatures at wavelengths other than λ-21 cm. The Wisconsin Hα Mapper surveyed the distribution and kinematics of ionized gas in the galaxy above decl. −30° with an angular resolution of 1° (Haffner et al. 2003). The resulting profiles reveal ionized gas in nearly every direction in the sky including many anomalous-velocity features at high latitudes.

Figure 4(c) compares the H i and Hα structures at −80 km s⁻¹ with a 10 km s⁻¹ wide band. The higher H i contour levels are overlaid on the corresponding Hα emission in color, and the correlation between the two appears to be significant. A more directed study by Tufte et al. (1998) found clear Hα emission in six directions toward MI at around −100 km s⁻¹ and at two locations toward MIIb at about −80 km s⁻¹. Since MI is far from any known source of ionizing radiation, another mechanism is required to heat the gas and produce the observed Hα emission. That mechanism may be the interaction between MI and Complex M discussed here.

Herbstmeier et al. (1995) compared X-ray data from the ROSAT 1/4 keV survey with the structure of the Complex M Hi. They found an extended nebulous cloud of X-rays as well as enhanced X-rays toward both MI and MIIb. Blom et al. (1997) report γ rays detected in the range 0.75–3 MeV with COMPTEL from the area below the Complex M arched filament seen in Figure 1(c) with a peak at l, b = 165°, 57 This high-energy emission is often attributed to nonthermal electron bremsstrahlung arising from HV cloud interactions with matter at the halo−disk interface, but since this border is absent in the Local Chimney, a different explanation is required.

The Sun appears to lie in a region of low-density gas known as the Local Bubble, which was first detected using interstellar extinction (Fitzgerald 1968) and Lyα absorption (Bohlin 1975) measurements toward nearby stars. Pervasive, soft X-ray emission (Williamson et al. 1974; Sanders et al. 1977) and O vi absorption toward several solar-neighborhood stars (Jenkins 1978) suggested that the Local Bubble may be filled with a high-temperature, low-density plasma. Observations indicate that the Local Bubble appears to extend to 100–200 pc in all directions (Zucker et al. 2022) and is surrounded by an irregular, higher-density gas boundary near the galactic plane. These regions of low-density gas may extend, however, through the disk and into the halo in certain high-latitude directions in a feature known as the Local Chimney (Welsh et al. 1999). The bubble and chimney were mapped in 3D in a series of papers (see Lallement et al. 2003 and references therein) using absorption characteristics of the interstellar Na i D-line doublet at 5890 Å toward over 1000 target stars lying within 350 pc of the Sun as determined by parallax measurements made with the Hipparcos satellite. Figure 5 shows a cross section of the Local Chimney at longitude 165° from Lallement et al. (2003) highlighting the location of MI and 56 UMa (Schmelz & Verschuur 2022).

Zoom In Zoom Out Reset image size

These higher-energy emissions suggest that Complex M is doing more than just traveling aimlessly through the cold, remote depths of the galactic halo. In fact, these emissions are all easier to explain if Complex M is much closer to the Sun. Schmelz & Verschuur (2022) showed that the diffuse X-rays reported by Herbstmeier et al. (1995) are located in the lower left half of the MI cavity where the integrated H i emission is weakest. They also interpreted the enhanced X-rays associated with MI as evidence for a bow shock that forms as the 56 UMa binary system and its supernova cavity travel through interstellar space and interact with ambient gas. Interactions between MI and the arched filament could create the energy needed to ionize hydrogen, generate the Hα emission, and produce the X-rays associated with the denser, fastest-moving features. If the magnetic field is important here, the X-rays could be generated by reconnection or wave heating. The γ-ray emission may not be part of the Complex M arched filament that is the focus of this paper, but it could be yet another manifestation of the process that generated the HV gas and point the way to understanding the energetics of the phenomenon that created Complex M.

The supernova scenario described by Schmelz & Verschuur (2022) to explain the high anomalous velocity of MI was only possible because 56 UMa survived the explosion of its binary companion. They deliberately limited their investigation to explain the origin and energy of MI and resisted the temptation to apply this scenario to other anomalous-velocity features where the evidence may be less convincing. Many neutron stars might be hiding in plain sight over much of the sky, creating cavities in the surrounding matter, but without a bright companion, a pulsar signature, or mass transfer, they would be all but invisible to us.

Schmelz & Verschuur (2022) found the radius of the MI cavity to be about 45 or 13 pc, far too small to be responsible for the Complex M arched filament seen in Figure 1. Yet, the previous section presents compelling evidence that the MI cloud and the loop-the-loop segment are interacting. Figure 1 shows that the arched filament appears to be moving coherently at a velocity of about −85 km s⁻¹. It may be longer than it appears, but this length is what we can see easily when we look up the Local Chimney because there is little intervening gas to confuse the structure. We propose the following scenario—the arched filament is created and launched by an unknown mechanism (see Section 3.3). Then, about 100,000 yr ago, the invisible companion of 56 UMa exploded, blasting MI away at a velocity of 120 km s⁻¹. Recently, at least astronomically speaking, MI caught up with and began interacting with/crashing into/passing through the arched filament. If this scenario is correct, that would place Complex M at the same distance as MI—about 150 pc.

With this distance, we can estimate the mass and energy of Complex M. Based on a composite H i spectrum from l = 135° to 165° and b = 65° to 68°, it is possible to estimate the average column density for the loop-the-loop segment of Complex M to be about 20.8 × 10¹⁸ cm⁻². Making the standard assumption that the depth is about equal to the width of the filament (05), the volume density is 5.3 cm⁻³, the mass is about 40 M_⊙, and the energy is 2.8 × 10⁴⁸ erg. If the original source for the motion of Complex M were a supernova, this derived mass would be a combination of the matter from the original explosion as well as gas swept up via the snowplow effect (Spitzer 1978) as the blast wave moves through interstellar space. Without knowing the original radial velocity of this hypothetical supernova progenitor, we can only estimate the energy required to accelerate the matter from zero velocity. The full extent of Complex M may be 3 times greater than just the loop-the-loop segment, implying a mass of 120 M_⊙ and an energy of 8.4 × 10⁴⁸ erg. These values require the same assumptions as those described above. Integrating over 4π sr, the total energy for a spherically symmetrical explosion is estimated to be 1.9 × 10⁵⁰ erg, well within the energy budget of a typical supernova (10⁵¹ erg).

If MI and the loop-the-loop segment are physically interacting, the supernova that created the MI cavity would have ionized the ambient H i so that the M&M gap in the Complex M ridge would be centered on MI. But the emission expected from Complex M across this region is only missing between the longitude of MI, 165° and 175°, and not at longitudes less than 165 Why this asymmetry? This phenomenon, which we have referred to as the M&M gap, is clearly visible in Figures 1–3. Figure 1(c) shows a thin filament that bridges the gap, which has either survived the impact of the supernova shell or the shell has not yet reached it.

Why, then, does the supernova shell centered on MI produce this asymmetric pattern of missing gas in the overall morphology of Complex M? Figure 6 offers a possible explanation. This asymmetry can be accounted for if the axis of the loop-the-loop segment, which is otherwise fairly straight, is inclined along our line of sight. Overall, its orientation may reflect the modulating influence of the local interstellar magnetic field. Radio polarization vectors at 408 MHz (Brouw & Spoelstra 1976) and optical polarization data for stars (Mathewson & Ford 1970) show that the local interstellar magnetic field is normal to the line of sight at l = 140°.

Zoom In Zoom Out Reset image size

Figure 6 shows schematically and to scale the relationship between the MI cavity carved out by the 56 UMa-related supernova and its interaction with the arched filament. Its overall axis is likely to be curved in depth as well in the plane of the sky. The supernova shell interacts with the loop-the-loop segment at longitudes greater than that of MI at l = 165° to remove the ambient Hi gas that would otherwise be found at −85 km s⁻¹. Instead, the M&M gap is created. According to the geometry in Figure 6, the gas at longitudes less than 165° has not yet been affected by the expanding shell. Beyond longitude 175° the H i gas in Complex M is again relatively undisturbed.

Although the connection to MI allows us to bootstrap the distance to Complex M, its origin remains a mystery. There are several leading contenders for the responsible mechanism, including an old, local supernova, which is an obvious candidate. This model, which was rejected by Oort (1966) based on what was known about anomalous-velocity gas at the time, was resurrected by Schmelz & Verschuur (2022) to explain MI. The supernova that created the cavity centered on the 56 UMa binary system is too small to have produced the entire arched filament (Figure 1), but another larger, older explosion may be responsible. The energy calculated above makes this scenario possible.

Another mechanism that could be responsible for the origin of Complex M is the galactic fountain. Observations from early rocket-borne X-ray detectors revealed widespread diffuse high-energy emission (Williamson et al. 1974). These X-rays were thought to originate from hot gas produced by supernovae in the disk that could escape into the halo. Shapiro & Field (1976) suggested a convective-radiative galactic-fountain model to cool and condense hot halo gas as it fell back toward the galactic plane. Bregman (1980) then examined the kinematics and found good agreement with the observed velocities of HV features.

A third possible mechanism is the Perseus supershell. Verschuur (1993) modeled the velocities and spatial distribution of Complex M and other nearby anomalous-velocity features as part of a vast supershell with an elliptical cross section normal to the galactic disk. Its origin appears to be in the Perseus arm toward l = 131° ± 4° with an apparent radius of 1920 pc and an expansion velocity of 281 ± 10 km s⁻¹. The Perseus arm is home to dozens of OB associations, making multiple supernovae a possible origin for the subset of anomalous-velocity gas features in the second quadrant of the northern galactic hemisphere. According to that model, gas in the near side of the shell has impacted a local structure marked by the radio continuum spur known as Loop III. The derived distances of the anomalous-velocity H i driven by the supershell are dependent on location and velocity, ranging from 100 pc or less for gas in the near face to as much as a kpc.