Gas Engaged in Noncircular Motions in LITTLE THINGS Dwarf Irregular Galaxies

3.1.1. DDO 133

In optical images, DDO 133 appears to have a stellar bar. This is visible as a rotation in the position angle of isophotes between the inner and outer galaxy. In the V-band image shown in Figure 12 one can see the bright, boxy inner structure that is the stellar bar and the more diffuse oval structure that is the stellar disk. The bar is outlined on other images in Figure 13. Here we have included a velocity map of the weak noncircular gas map (wnonc1) where we have identified two wnonc regions that are coincident with the northern end of the bar (wnonc region 1) and the upper part of the body of the bar (wnonc region 2). Properties of the two wnonc regions and the bar are given in Table 3. The wnonc regions in DDO 133 are located on a ridge of H i in the mom0 map. Furthermore, there is a clump that is bright in Hα and FUV associated with wnonc region 1 at the end of the stellar bar.

Table 3.  DDO 133 Weak Noncircular Regions and Bar^a

Region	R.A.	Decl.	${l}_{\mathrm{maj}}$	${l}_{\min }$	P.A.	Area	ΔV
	(hh:mm:ss.s)	(dd:mm:ss)	(arcs)	(arcs)	(deg)	(arcs2)	(km s−1)
wnonc 1b	12:32:52.8	+31:33:38	⋯	⋯	⋯	1818	7
wnonc 2b	12:32:54.8	+31:32:36	⋯	⋯	⋯	1177	17
Bar	12:32:54.3	+31:32:15	158.5	75.5	165	11972	⋯

Notes.

^aParameters are defined in Table 2. ^bThe wnonc regions are outlined with polygons and so ${l}_{\mathrm{maj}}$, ${l}_{\min }$, and P.A. have no values.

Download table as:  ASCIITypeset image

In the upper left panel of Figure 13, velocity contours (in rainbow colors) have been placed on the V-band image of DDO 133. Note that along the outside edges of the bar (outlined in black), there is a crinkle pattern in the velocity contours. The same crinkle pattern along the edges of the bar are also seen by Lin et al. (2013) in the velocity field of the spiral galaxy NGC 1097. Their hydrodynamical simulation reproduces the velocity field pattern through streaming motions of the gas along the bar. Furthermore, it seems likely that the star formation at the end of the stellar bar is a result of gas piling up at the ends from streaming motions around the bar, and that this is also the cause of the wnonc motions there. The buildup of gas at the end of the bar could also have been influenced by the formation of an H i hole, which might once have been an expanding bubble, to the east of the northern end of the bar. This hole is evident as the circular purple region at the northern end of the mom0 image in the bottom left panel of Figure 13.

3.1.2. DDO 47

Simulations suggest that cold accretion of gas from the cosmic web onto galaxies is an ongoing process even today and a driver for ongoing star formation in spirals (e.g., Finlator & Davé 2008; Forbes et al. 2014; Sánchez Almeida et al. 2014). For example, Schmidt et al. (2016) found radial mass fluxes in five spiral galaxies from The H i Nearby Galaxy Survey (THINGS; Walter et al. 2008) that were comparable to their star formation rates (SFRs). Sánchez Almeida et al. (2015) find that large regions of star formation in tadpole galaxies are significantly lower in metallicity than the rest of the galaxy. They argue that these are possibly regions where accreting gas has been compressed as it enters the disk and a large star-forming event was triggered. Since the low metallicity gas rotates through the galaxy and dissipates in less than an orbital period, the accretion events must have happened recently for unusually low metallicity regions to be found.

Motivated by these observations, Ceverino et al. (2016) have run cosmological zoom-in simulations of the growth of a dIrr galaxy, tracking gas accretion and the distribution of star formation and metallicity in order to make connections between gas inflows and low metallicity regions in star-forming galaxies. They found that a number of their simulated galaxies have clumpy Hα regions with lower metallicity than the rest of the galaxy, such as observed in the tadpole galaxies. In the simulations, the velocities of the incoming streams of gas differ significantly from the bulk motion of the galaxy, and terminate at the clumpy Hα regions. This implies that the incoming gas streams are piling up and driving the high SFR, low metallicity pockets. To look for accretion events, therefore, we might expect to find incoming streams of gas that are more dense and moving at a different velocity than the bulk rotation of the existing H i disk. Furthermore, the impact of gas clouds on dIrrs has been observed to result in the formation of super star clusters (for example in NGC 1569 and NGC 5253; Johnson 2013; Turner et al. 2015). Thus, we might expect to find significant star-forming events to be associated with these streams.

Large filamentary structures were identified in two of our galaxies: DDO 47 and DDO 133. The filament in DDO 133 is region 18 in Figure 7, and region 17 just above region 18, although somewhat detached, could also be part of this filament. There are 2.7 × 10⁶ M${}_{\odot }$ of gas in this filament, which is 3% of the total H i mass of the galaxy. The width of the filament is about 590 pc, the average column density of the H i is 1.8 × 10²⁰ atoms cm⁻², and the distance of the filament from the center of the galaxy in the face-on plane of the galaxy is about 3 kpc. As an example, Figure 16 shows cuts through the H i cube at three positions in the filament: the middle of region 17, the middle of the north half of region 18, and the middle of the south half of region 18. The intensity peaks due to the bulk and snonc motion gas are marked, and there is also a wnonc component clearly visible in two of the cuts. The offset in velocity between the filament and the bulk motion gas is of order 10 km s⁻¹. There is no connection between young regions seen in FUV or Hα and the filament.

The filament in DDO 47 is less coherent and is delineated by regions 9–12 (and possibly region 8) in Figure 2. The gas in the filament is 5.3 × 10⁶ M${}_{\odot }$, which is 1% of the total H i mass of the galaxy, and has median velocities of 4–8 km s⁻¹ higher than in the nearby bulk rotation gas. An example of a cut through the H i cube at a point in the middle of region 11 is shown in Figure 17. The gas in bulk motion at that location is moving around 240 km s⁻¹ along the line of sight, while there is even more gas there moving in noncircular motion at about 253 km s⁻¹. The width of the filament is about 760 pc, the average column density of the H i is 4.7 × 10²⁰ atoms cm⁻², and the distance of the filament from the center of the galaxy in the plane of the galaxy is about 4 kpc. As for DDO 133, there is no connection between young regions seen in FUV or Hα and the filament.

We examined the timescales for destruction of the filaments in DDO 133 and DDO 47. We first look at the rate of growth of shearing perturbations, given by Hunter et al. (1998) as $\pi G{{\rm{\Sigma }}}_{g}/c$, where G is the gravitational constant, Σ_g is the gas surface density, and c is the gas velocity dispersion. We multiply the H i surface density by 1.34 to include Helium. For the filament in DDO 133 we find that the growth rate is 5 × 10⁻¹⁸ s⁻¹, which gives a timescale of 6 × 10⁹ yr. Thus, it would take a very long time to destroy the filament by the growth of a shearing perturbation, if the filament is in the plane of the galaxy disk.

Next we consider the time for the filament to disperse, given the width and the velocity dispersion of the gas. The timescale to disperse (width divided by velocity dispersion) is 8 × 10⁷ yr, which is 19% of an orbital time for the galaxy. This implies that dissipation may be the primary destruction mechanism for the filament in DDO 133. This is the predominant mechanism for destruction found in the simulations of Ceverino et al. (2016).

For the filament in DDO 47 we find the timescale for the growth of shearing perturbations is 8 × 10⁷ yr and the timescale for dispersion is similar at 1 × 10⁸ yr. The latter is 31% of an orbital time. Thus the filament in DDO 47 could be destroyed by either shearing perturbations or dissipation.

Finally, we consider whether the filament in DDO 133 is unstable to gravitational instabilities that might enable it to collapse into star-forming clouds. The H i plus He gas density is 2.4 × 10²⁰ atoms cm⁻². This gas density is 20% of the Toomre (1964) critical gas density for instability at that radius in DDO 133 (Elmegreen & Hunter 2015). A typical region in the filament in DDO 47 has an H i+He column density of 6.3 × 10²⁰ atoms cm⁻², and this is 47% of the Toomre critical gas density. Therefore, the filaments are stable against internal collapse into self-gravitating clouds.

As to whether either of these filaments could be cosmic cold accretion, we cannot tell for sure. However, the filaments that we observe are fairly well behaved, being nearly engaged in the rotation pattern of the galaxy. By contrast, a cloud of gas falling onto the BCD VIIZw 403 has a very complex kinematic pattern relative to the body of the galaxy (see Figure 17 in Ashley et al. 2017). The simulations of cold accretion in dwarf galaxies by Ceverino et al. (2016; see their Figure 4) also show significant deviations in the velocities of the infalling gas relative to the galaxy. Furthermore, there does not seem to be consequences of these filaments to the galaxies in terms of star formation. While the observational signatures of accretion might depend on the details of the infall and on the density of the gas it is falling onto, it seems that these filaments, whatever their nature, are not having much of an impact on their host galaxies. By the same argument and the fact that these galaxies were chosen to be fairly isolated, it seems unlikely that these filaments are the consequence of a recent interaction with another galaxy. However, transient instability modes, as discussed for DDO 47 above, are possibilities.

We plot the SFR surface density of each galaxy against the ratio of mass of H i in that galaxy engaged in significant noncircular motion to the total H i mass of the galaxy. This is shown in Figure 18 with the SFR surface density determined from FUV emission on the left and that from Hα on the right. Galaxies without gas engaged in noncircular motions (at ${M}_{\mathrm{snonc}}/{M}_{{\rm{H}}{\rm{I}},\mathrm{tot}}=0$) cover the full range of SFRs, but for the rest of the galaxies there is a slight decrease in FUV SFR surface density with increasing ${M}_{\mathrm{snonc}}/{M}_{{\rm{H}}{\rm{I}},\mathrm{tot}}$, although this is not evident for the Hα SFR surface density. The FUV trend is opposite to the expectation that noncircular motions might facilitate star formation. On the other hand, if the presence of large amounts of snonc gas results in starbursts that are short lived, we might be less likely to catch those events.

On a local level, too, there is no pattern of correspondence between star-forming regions and strong noncircular motion gas. Generally speaking, in a given galaxy, a few nonciruclar motion regions might be coincident with FUV or Hα emission, but other regions are not and there can be lots of star formation in a galaxy not connected to gas moving differently from the bulk rotation.

Holes in the H i gas of the LITTLE THINGS dwarfs have been cataloged by Pokhrel (2016; see Bagetakos et al. 2011 for the methodology in identifying the holes). As an example, in Figure 19 we plot the location of the eight holes in DDO 133 and compare those to the location of snonc and wnonc features. We see that three of the gas holes overlap with snonc features. The most interesting hole is number 3, which overlaps with snonc region 9 and is adjacent to the northeast corner of the stellar bar and to the east of the northern part of a wnonc feature at the end of the bar (large red area in the bottom panel of Figure 19). Expansion of the shell around hole 3 might have contributed to the snonc and wnonc features, although currently the hole has blown out of the disk and is not expanding (Pokhrel 2016). Hole number 5 also appears to be connected with wnonc gas seen in the bottom panel of Figure 19 and snonc region number 13, and hole number 4 overlaps with snonc region 12.

In the bottom right panels of Figures 1–11 we plot the snonc features on the H i velocity dispersion (mom2) maps of the galaxies. Here we are looking for correlations between the snonc features with velocity dispersion that would indicate that these motions generate excess turbulence. Ashley et al. (2014), for example, found excess turbulence at the impact points of accreting gas in IC 10. In most galaxies there are some snonc features in regions at the upper range of turbulence values in the galaxy. For example, in CVnIdwA both regions are adjacent to the two regions of higher velocity dispersion (red in the mom1 map in Figure 1). Other snonc regions are located in the central parts of the galaxies where the velocity dispersion is overall higher, as in DDO 47 (Figure 2), and the association is not likely specific to the snonc nature of the regions; the higher velocity dispersion regions are located throughout the galaxy and the snonc regions appear to be randomly associated, as in DDO 50 (Figure 3); or a few snonc regions are associated with higher velocity dispersion but many others are not, as in DDO 70 (Figure 5). Thus, generally there is no systematic correlation between snonc regions and regions of high velocity dispersion.

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

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

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

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

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

However, there are two exceptions of note. In DDO 168 (Figure 8), all of the snonc features are located on the eastern edge of a large part of the galaxy where the velocity dispersion exceeds 14 km s⁻¹. There is nothing obviously unusual in that region of the galaxy at other wavelengths, so it is not clear why this large section of the galaxy has such a high velocity dispersion. In DDO 210 (Figure 9), just outside the snonc regions to the northeast and southeast are areas where the velocity dispersion of the gas is >11 km s⁻¹. These high velocity dispersion regions line the edges of the large-scale arc structures.