A RADIO JET DRIVES A MOLECULAR AND ATOMIC GAS OUTFLOW IN MULTIPLE REGIONS WITHIN ONE SQUARE KILOPARSEC OF THE NUCLEUS OF THE NEARBY GALAXY IC5063

The list of emission lines that we detected in the NIR spectrum of IC5063 comprises $\mathrm{Pa}\alpha$, $\mathrm{Br}\gamma$, He i 2.0587 μm, [Fe ii] 1.6440 μm, [Si vi] 1.9634 μm, [Si xi] 1.9196 μm, [Ca viii] 2.321 μm, ${{\rm{H}}}_{2}$ (1-0) S(1), and ${{\rm{H}}}_{2}$ (1-0) S(3). A kinematic analysis was feasible for these lines, which were bright enough to be little affected by neighboring sky/telluric lines. The ${{\rm{H}}}_{2}$ (1-0) S(3) line could be contaminated by He i and/or [Fe ii] 1.954 μm, −550 km s⁻¹ away (Vanzi et al. 2008). Because we can neither test this assumption nor evaluate the strength of these lines, we opt to show the ${{\rm{H}}}_{2}$ (1-0) S(3) data, but to not use them in our kinematic analysis for velocities below −500 km s⁻¹. We detected numerous other emission lines in the spectra, but we did not create kinematic maps for them, either because their flux was low or because their profiles were substantially contaminated by sky/telluric lines. These lines comprise $\mathrm{Br}\delta$, Br10, Br11, [Fe ii] 1.5330 μm, ${{\rm{H}}}_{2}$ (1-0) S(0), S(2), S(5), S(7), and ${{\rm{H}}}_{2}$ (2-1) S(1), S(3). The possibility that potential Br12 1.6412 μm emission significantly affects the [Fe ii] 1.6440 μm emission (even though the Br13 and Br14 lines are undetected) is evaluated and rejected in Section 3.2. The ${{\rm{H}}}_{2}$ (1-0) S(4) line was so contaminated by sky lines that its flux could not be recovered. Finally, we detected ${{\rm{H}}}_{2}$ (1-0) Q(1), Q(2), and Q(3) emission, but the corresponding lines were discarded from our kinematic analysis because their restframe frequencies are only ∼1000 km s⁻¹ away and their fluxes are comparable: emission of ambient gas in one line could be mistaken for emission of outflowing gas in another line. In the rest of this work, when referring to ${{\rm{H}}}_{2}$ lines, we will refer only to lines probing transitions from the first to the fundamental vibrational state unless otherwise noted.

The emission line maxima at the location of the nucleus indicate that IC5063 is at a redshift of 0.0116 ± 0.0006. Collapsed images of the H-band continuum and of lines within ±500 km s⁻¹ from the systemic velocity are shown in Figure 1. The contours of these images, as well as the contours of a 17.8 GHz image (Morganti et al. 2007), are plotted over an optical (547 M filter) image taken with WFPC2 onboard the Hubble Space Telescope (Schmitt et al. 2003). The emission is spatially unresolved in [Si xi] and [Ca viii], which probe the galaxy nucleus. The stellar CO(3-1) absorption seen near [Ca viii] may be affecting the latter's distribution. In all other lines, the emission is resolved and follows the radio jet trail, like the [O iii] 5007 Å emission in the WFPC2 image. [Fe ii] shows the overall most diffuse emission. Four distinct regions can be identified in the NIR images: the nucleus (at 20:52:02.40−57:04:07.68), the south radio lobe (18 or 450 pc southeast of nucleus, at 20:52:02.60−57:04:08.68), the north radio lobe (22 or 520 pc northwest of the nucleus, at 20:52:02.17−57:04:06.55), and the extended clump near the northwest edge of the NLR ionization cone (38 or 900 pc away from the nucleus, at 20:52:02.06−57:04:05.05) that was previously seen by Kulkarni et al. (1998).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Spectra of these four regions are shown in Figure 2, integrated over a radius of 05. The spectrum of the nucleus has a rising (from 2.4 to 5.6 mJy) NIR continuum, characteristic of dust heated by the AGN. The highest number of detected lines is found in the nuclear region. $\mathrm{Pa}\alpha$ is the brightest of all, despite the strong atmospheric absorption window it is in. [Ca viii], [Si vi], and [Fe ii] 1.6440 μm are among other bright lines. Contrarily, the region with the lowest number of detected lines is the NLR clump 3.8'' from the nucleus. Its spectrum has a much flatter and fainter continuum, of ∼0.4 mJy, as a result of the lower AGN and stellar emission. The slope and the intensity of the continuum at the north radio lobe and at the south radio lobe are intermediate of those at the nucleus and at the NLR clump. However, the ${{\rm{H}}}_{2}$ and [Fe ii] emission is enhanced at the radio lobes—in particular at the north radio lobe, where these lines are brighter than at any other position in the central 2 × 2 kpc² of IC5063.

Zoom In Zoom Out Reset image size

The flux ratio of ${{\rm{H}}}_{2}$ (1-0) S(1) over $\mathrm{Pa}\alpha$ is 2–8 times lower at the nucleus than within 100 pc from the north radio lobe, and $\lesssim 15$ times lower further south of this region (Figure 3). Likewise, this ratio is up to 2 times lower at the nucleus than within 100 pc from the south radio lobe, and $\lesssim 10$ times lower further north of this region. The regions with emission excess from molecular gas, which could be star-forming or associated with shocked media (Ogle et al. 2010; Guillard et al. 2012), do not necessarily show emission excess from iron ions. The [Fe ii]/ $\mathrm{Pa}\alpha$ flux ratio peaks mid-way between the radio lobes and the regions with ${{\rm{H}}}_{2}$ emission excess, whereas the [Fe ii]/ ${{\rm{H}}}_{2}$ (1-0) S(1) and [Fe ii]/ ${{\rm{H}}}_{2}$ (1-0) S(3) flux ratios are maximum in regions where the ${{\rm{H}}}_{2}$ (1-0) S(1)/ $\mathrm{Pa}\alpha$ flux ratio is minimum. High fractions of ionized to molecular gas emission are seen in regions elongated perpendicularly to the jet propagation axis. Such regions are likely to contain shocked gas, because Fe requires the passage of shock fronts to be liberated from dust grains and massively deposited into a diffuse gas reservoir. The borders of the NLR are prominently outlined by ridges of high [Fe ii]/ $\mathrm{Pa}\alpha$ flux ratio, either due to shocks or AGN illumination. The clump near the NLR northwest edge is bright in [Si vi]. The logarithm of the [Si vi]/ $\mathrm{Br}\gamma$ flux ratio in that region is in the range 0.1–0.3, which renders it compatible with shock and shock precursor ionization models, but also with AGN photoionization (Allen et al. 2008; Groves & Allen 2010; Bicknell et al. 2013).

Zoom In Zoom Out Reset image size

The effects of obscuration are evaluated by means of comparison of the $\mathrm{Pa}\alpha$/ $\mathrm{Br}\gamma$ ratio with the theoretically predicted value of 12 for case B recombination at 10⁴ K. In most regions, the ratio takes values between 12 and 9 (Figure 3). These numbers respectively translate to optical depths of 0.0–2.2 in the H band, and 0.0–1.6 in the K band, assuming that the absorption follows the Milky Way ${R}_{{\rm{V}}}$ = 3.1 model of Draine (2003). Data reduction artefacts and differences in the gas excitation conditions can explain values outside this range. We find that there is little or no obscuration near the north radio lobe, near the nucleus, near the NLR most illuminated edge, and in front of the dust lane (north of the jet trail). Dust could be affecting the observed gas properties ∼300 pc west of the nucleus and near the south radio lobe, but any such effects are not quantifiable. Regions with high obscuration and ${{\rm{H}}}_{2}$/ $\mathrm{Pa}\alpha$ excess coincide in projection with large-scale structures, rich in molecular gas (Morganti et al. 2015).

To determine the kinematic properties of the gas in the center of IC5063, we fit a Gaussian function along the profile of each line in each spatial pixel. Only pixels with S/N > 10 were used. The results of this computation are presented in Figure 4. The disk previously known from H i observations (Morganti et al. 1998) is clearly seen in all of our tracers. It has a projected minor axis of 240(±30) pc, calculated from the He i map and from the hydrogen recombination line maps in regions displaying circular orbits. Its southeast side is moving away from us and its northwest side is coming toward us. On the southeast side, the average circular velocity V_c is 185 ± 21 km s⁻¹ at 750 pc from the nucleus, and 250 ± 12 km s⁻¹ at 900 pc from the nucleus. On the northwest side, the disk is rotating with −157 ± 42 km s⁻¹, −190 ± 44 km s⁻¹, and −247 ± 7 km s⁻¹at 800 pc, 1 kpc, and 1.25 kpc from the nucleus, respectively. A confirmation of these numbers comes from the radio lobe spectra, where the jet-related shocks enhance the emission of the gas rotating with the disk. There, the shocked gas emits so prominently that disk-related spectral line displacements of order 200 km s⁻¹ are visible in the scale of the entire NIR spectrum (Figure 2). The resolution-corrected velocity dispersion of the gas, as measured from the same gas tracers as those used for the rotational velocity measurement, is 115 ± 20 km s⁻¹ at the nucleus and 95 ± 9 km s⁻¹ one kiloparsec northwest of it.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The examination of data cube slices around lines of interest reveals further details on the kinematics of the different gas components. The disk rotation can be clearly seen in Figure 5, where each panel shows an image with a velocity bin equal to two spectral pixels of the initial cube: the bulk of the emission is shifted from the west of the disk to the east of the disk with increasing velocity. The emission from the northwest branch of the NLR peaks at about −320 km s⁻¹, and it can be seen for another ±300 km s⁻¹. Its counterpart on the other side of the galaxy, the southeast branch of the NLR, can be barely detected in the 157 and 315 km s⁻¹ $\mathrm{Pa}\alpha$ bins. Faint structures almost perpendicular to the disk major axis (at position angles between 25° and 50°) can also be seen at zero velocity in [Fe ii] and $\mathrm{Pa}\alpha$. One of them starts at the center of the galaxy and continues southwest, whereas the other starts mid-way between the nucleus and the south radio lobe and continues northeast. They are both linked to the large-scale spiral arms seen in CO by Morganti et al. (2015).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

A striking result in Figure 5 is the detection of high-velocity gas near the projected positions of the radio lobes, where the jet is thought to interact with dense ISM clouds. Near the north radio lobe, the high-velocity flow is seen for multiple species (H, He, Fe, Si, ${{\rm{H}}}_{2}$) and for all gas phases (i.e., atomic ionized/neutral and molecular). [Fe ii], for example, is detected over a continuous velocity range of 2000 km s⁻¹ (Figure 5). The [Fe ii]-emitting gas at ∼1000 km s⁻¹ (Figure 5) is moving against the disk rotation by ∼1200 km s⁻¹. In the direction of the disk rotation, about half of the [Fe ii] emission comes from gas that is moving faster than the circular velocity along the line of sight (Figure 2; bottom). Despite the potential existence of Br12 1.6412 μm emission −510 km s⁻¹ away, we confirmed that the detection of the highly blueshifted [Fe ii] wing is significant: under the assumption that the Br12 flux can be as high as the Br11 flux, we removed the stacked Br11 image at −300 ± 180 km s⁻¹ from the stacked [Fe ii] image at −810 ± 180 km s⁻¹, and verified that the detection remains significant (Figure 6). Gas that is moving faster than V_c by more than 600 km s⁻¹ is seen in several other lines, including ${{\rm{H}}}_{2}$ (1-0) S(1).

Zoom In Zoom Out Reset image size

Near the south radio lobe, the jet effects on the gas kinematics are only a little less pronounced than near the north radio lobe, even though the gas emission is lower by a factor of 2 (Figure 2; lower panels). With V_c rising quickly from ∼100 to ∼200 km s⁻¹ near the south lobe (Figure 4), we again detect gas moving with velocities 600 km s⁻¹ above the disk circular velocity and 900 km s⁻¹ against it (Figure 5). This result, found for two sites, unambiguously shows that the irregular gas kinematics are due to jet-cloud collisions—not due to some coincidence (e.g., debris of a past merger; Colina et al. 1991; in the line of sight behind one of the lobes).

To study the outflow extent between the radio lobes, we created maps of the outflowing-to-ambient gas emission (in [Fe ii] and ${{\rm{H}}}_{2}$). We did so for two velocity ranges of the outflowing gas. With the fast outflow used in Figure 6, we associated gas moving faster than 600 km s⁻¹ (i.e., with a velocity exceeding V_c by more than three times the disk velocity dispersion). With the intermediate-velocity outflow used in Figure 7, we associated gas moving slower than 600 km s⁻¹, but against the disk rotation. To be conservative, we used the maximum measured circular velocity, 250 km s⁻¹, and velocity dispersion, 115 km s⁻¹ in our definitions. By ambient, we refer to gas moving slower than V_c (in either direction). Figure 6 shows that the fast outflow extends along the radio jet trail from the vicinity of the nucleus to the north radio lobe. The intermediate-velocity outflow extends further away (Figure 7; see also Morganti et al. 2015 for a 200 km s⁻¹ outflow of cold gas probed by CO).

Zoom In Zoom Out Reset image size

Figure 6 further reveals that the highly blueshifted and the highly redshifted emission near the north radio lobe are not cospatial. The emission at 810(±)180 km s⁻¹ peaks 240 pc southeast of the emission at −810(±)180 km s⁻¹ (Figure 6).

To identify discrete outflow starting points, we used Hubble Space Telescope FOC optical images of high spatial resolution (∼10 pc). The FOC image shown in Figure 8 is the average of three individual exposures probing polarized light at 0, 60, and 120°, respectively. Taken with the F502M filter, these images comprise ${\rm{H}}\beta$ and [O iii] emission. Regions of interest are marked on Figure 8. With R1 and R2, we mark the beginning and the end of a 120 pc-long filamentary structure, which is perpendicular to the jet trail. The peak of the highly redshifted [Fe ii] emission is near R2. The dark region west of R2 is where the H-band optical depth is ∼2, where the ${{\rm{H}}}_{2}$/ $\mathrm{Pa}\alpha$ ratio is high (Figure 3), and where a large-scale CO structure reaches the inner gas disk (Morganti et al. 2015). With R3, we mark the NLR base, i.e., the region where the ionization cone becomes visible. With R4, we mark the peak of the highly blueshifted [Fe ii] and ${{\rm{H}}}_{2}$ emission. There, the emitting region in the FOC data is oriented north-south (i.e., with an inclination with respect to the jet trail). R1, R3, and R4, are respectively located 240, 320, and 400 pc away from the nucleus.

Zoom In Zoom Out Reset image size

Upstream the NLR cone and northwest of R4, more outflowing gas is detected by its [Fe ii] emission at 315 ± 90 km s⁻¹ (or ∼550 km s⁻¹ against V_c; Figure 7). A discrete starting point within this flow can be seen by the NLR clump identified by Kulkarni et al. (1998). We mark it with R5 in Figure 8.

The flow of accelerated gas is not just parallel to the jet trail. Figure 6 shows a diffuse outflow component that is oriented perpendicularly to the jet trail, and that spreads southwest of R1. Its extent in either [Fe ii] or ${{\rm{H}}}_{2}$ exceeds 300 pc at high velocities and reaches ∼500 pc at intermediate velocities (Figures 6, 7). The intermediate-velocity maps further reveal a similar diffuse outflow component spreading ∼400 pc northeast of R1 (Figures 7, 8). The overall outflow has, thus, a biconical shape. Both its cones extend beyond the projected disk thickness (as derived from regions with circular orbits in the He i and hydrogen recombination line maps). Conical shapes of gas moving against the disk rotation can also be seen in the [Fe ii], ${{\rm{H}}}_{2}$ (1-0) S(1), S(3), and $\mathrm{Br}\gamma$ kinematic maps of Figure 4.

The center of this biconical outflow is located 240 pc northwest of the galaxy nucleus. Its axis is perpendicular to the jet trail and parallel to the R1–R2 line in the FOC image, revealing an association between the outflow and the bright ${\rm{H}}\beta$ and [O iii] emission. For [Fe ii], the fraction of accelerated-to-ambient gas emission exceeds one near this line (Figures 6, 7). It is worth noting that this is not necessarily the highest intrinsic fraction, but the highest observed fraction due to ions scattered in our line of sight. The elongated region with the highest observed accelerated-to-ambient gas emission continues to the north as a ridge of high [Fe ii] velocity dispersion (Figure 4). Along that ridge, the gas dispersion is boosted from 140(±10) km s⁻¹ to 290(±20) km s⁻¹. On its west, the gas line of sight velocity goes down to −350(±20) km s⁻¹ in [Fe ii], $\mathrm{Pa}\alpha$, and [Si vi]. On its east, the iron ions are typically moving with 180–300 km s⁻¹.

The data cubes provide enough information as to how the spatial distribution of the diffuse gas emission changes with velocity (Figure 9). In this Figure, the [Fe ii], ${{\rm{H}}}_{2}$ (1-0) S(1) and S(3) cubes are plotted at their finest velocity step, and the nuclear emission has been subtracted for visualization purposes. The diffuse emission from gas moving against the disk rotation is also visible prior to the nuclear flux removal (Figure 5). The inspection of the individual cube channels reveals that the gas emission spreads from the nucleus to the north lobe with increasing velocity. The region that the diffuse gas occupies unfolds almost symmetrically with respect to the jet trail. At velocities ≳240 km s⁻¹, however, the symmetry breaks as the emission drops sharply north of the trail, near the ridge of high [Fe ii] velocity dispersion. South of the trail, the emission continues, forming an arc that connects (in projection) the biconical outflow base with the north radio lobe. The arc is most pronounced for the molecular gas. An explanation for this progression in the outflowing gas spatial distribution as a function of velocity is given in the following section.

Zoom In Zoom Out Reset image size

Position–velocity diagrams along position angles pertinent to the biconical outflow are shown in Figures 10 and 11. In Figure 10, we examine a slice that is nearly perpendicular to the jet axis and that passes from the starting point of the biconical outflow. All along 800 pc, gas is continuously detected with a velocity peak of ∼200 km s⁻¹ for all of [Fe ii], ${{\rm{H}}}_{2}$ (1-0) S(1) and S(3). For [Fe ii], we see that the emission spreads out to 1.2 kpc, and that the gas attains a sharp terminal velocity of 400 km s⁻¹. When reading this position-angle diagram from the jet propagation axis (i.e., from an offset of 3'') to outer regions (i.e., to offsets of 0'' or 6''), we find that the diffuse gas velocity increases with increasing distance. The velocity gradient is smooth, and close to symmetric in the two sides of the jet.

Zoom In Zoom Out Reset image size

In Figure 11, we examine slices nearly parallel to the jet axis. Along the northern border of the biconical outflow, which runs parallel to the jet propagation axis at a distance of ∼350 pc, there is intermittent [Fe ii] emission at 900 km s⁻¹ along both sides of the disk (Figure 11). This indicates that the ionized gas wind generated northwest of the nucleus is strong enough to dominate the [Fe ii] emission even northeast of the nucleus (see also Figure 4). This is not true for the warm molecular gas, even though ${{\rm{H}}}_{2}$ (1-0) S(3) emission is also seen along 1.3 kpc in both sides of the disk (at about −100 km s⁻¹). West of the nucleus, part of this emission is coming from the disk. East of the nucleus, the emission is blueshifted by 260 km s⁻¹ with respect to the disk rotational velocity. It is, thus, likely due to shocks generated east of the nucleus (Figures 4, 11). The region with negative velocities northeast of the nucleus can also be seen in the ${{\rm{H}}}_{2}$ (1-0) S(1) and S(3) velocity fields (Figure 4). This is another region where gas flows in both line of sight directions, as [Fe ii] emission is also detected there at ∼900 km s⁻¹. Diffuse outflowing gas is, thus, also present on the east side of the disk. Its emission is, however, too weak to reveal changes in the gas spatial distribution as a function of velocity.

Zoom In Zoom Out Reset image size

Notably, the (more tenuous) ionized gas attains higher velocities than the (denser) molecular gas at fixed positions (Figures 5, 9–11). This result is in agreement with multi-phase outflow acceleration models (Hopkins & Elvis 2010; Wagner & Bicknell 2011; Zubovas & King 2014).

A question pertinent to galaxy evolution studies that NIR data help us to address is whether the jet heats up the molecular gas it accelerates. To evaluate this possibility, we compared the gas excitation temperature ${T}_{{\rm{ex}}}$ in the ambient medium and in the intermediate-velocity diffuse outflow.⁴ For this purpose, we created temperature maps of the warm ${{\rm{H}}}_{2}$ using the (1-0) S(1) and S(3) lines, assuming that the gas can be described by a single temperature component in local thermodynamic equilibrium (LTE). The results are shown in Figure 12. The ambient gas excitation temperature at the north radio lobe is 1500 ± 200 K (see also Tadhunter et al. 2014). In the rest of the disk, it is 2000 ± 200 K. At the edges of the disk, near the regions with excess of molecular-to-atomic-gas emission (Figure 3), it goes up to 2700 ± 500 K. The low temperature of the ambient gas at the north radio lobe could be indicative of a region that is denser and more self-shielded than its surroundings, which can be described by several components of subthermally excited gas.

Zoom In Zoom Out Reset image size

For the diffuse accelerated gas at 310 ± 70 km s⁻¹, ${T}_{{\rm{ex}}}$ varies significantly from one region of the map to another. In the south of the jet trail, ${T}_{{\rm{ex}}}$ is within the range of values found for the ambient medium. However, no ${T}_{{\rm{ex}}}$ solution can be found for many pixels north of the jet trail: north of the regions R3, R4, and the north radio lobe (where diffuse 17.8 GHz emission exists; Figure 8) the (1-0) S(3)/S(1) flux ratio is above the saturation limit (2.04) for a population that is described by a Boltzmann distribution and that decays spontaneously. This result implies that the molecular gas is in part non-thermally excited and that radiative decay in the form of fluorescence is plausible. The non-thermal source of excitation could be X-rays or UV light coming, for example, from the jet-related high-velocity shock wave or from the AGN radiation (Lepp & McCray 1983; Black & van Dishoeck 1987; Draine & Woods 1990; Mouri 1994). This result does not rule out a temporary temperature increase of the gas that got exposed to the non-thermal source of excitation: radiative cascade and heating are often coupled processes (e.g., Mouri 1994). Evidence of heating comes from the work of Young et al. (2007), who found that the dust is warmer at the north radio lobe than at south radio lobe.

The ${{\rm{H}}}_{2}$ mass distribution that we can create when taking the ${T}_{{\rm{ex}}}$ map at face value and assuming an ortho-to-para ratio of 3 is shown in Figure 13. The total ambient gas mass is 3500 M_⊙ (of which 990 M_⊙, 460 M_⊙, and 150 M_⊙ are found within a radius of 05 from the north lobe, the nucleus, and the south lobe, respectively). For the gas at 310 ± 70 km s⁻¹, the total mass is 170 M_⊙ (of which 30 M_⊙ and 80 M_⊙ are within 05 from the north lobe and the nucleus, respectively). The remaining 60 M_⊙ come from the diffuse medium. Figure 13 demonstrates that the maximum mass of the diffuse molecular gas (that can be detected from our line of sight) follows a filamentary distribution near the biconical outflow axis. The filamentary structure carries 35 M_⊙. For comparison, the mass of the molecular gas in the fast outflow at −810(±)210 km s⁻¹ would be 25 M_⊙, if all of the ${{\rm{H}}}_{2}$ (1-0) S(1) emission in it was due to thermalized gas at ${T}_{{\rm{ex}}}$ = 2000 K. The total mass in the outflow at all velocities and regions is, thus, unlikely to exceed some hundreds of solar masses (see also Tadhunter et al. 2014). We, thus, conclude that the contrast between the mass carried in the warm phase and the mass carried in the cold phase (2.3 × 10⁷–1.3 × 10⁸ M_⊙; Morganti et al. 2013) remains stark. If the acceleration of the molecular gas happens through its partial dissipation and reformation as proposed by Hopkins & Elvis (2010), then this result is a proof that microturbulence (Nesvadba et al. 2011) ceases rapidly enough for cooling and potentially for star formation to occur during the outflow propagation timescale.

Zoom In Zoom Out Reset image size

The bottom line of the results presented in Section 3 is that numerous gas components with irregular kinematics are detected in the central square kiloparsec of IC5063. Which mechanism drives the observed gas motions? In this Section, we argue that the propagation and the deflection of the radio jet plasma does so.

The diffuse gas geometric characteristics are not in line with an AGN-related or a starburst-related mechanism of acceleration. The biconical outflow in IC5063 cannot be associated with a bubble driven by the AGN radiation (as in NGC 4258; Jiménez-Vicente et al. 2010), because it is not centered at the nucleus. The gas kinematics are also incompatible with those of an expanding bubble: we observe no spherical symmetry in the velocity gradient of the accelerated Fe ions. No ultra-fast outflow (as in Wagner et al. 2013) has been detected in X-ray spectra of IC5063 (Tombesi et al. 2014), and no [Ca viii] or [Si xi] outflow is seen in the SINFONI data. A nuclear starburst is ruled out for the same symmetry reasons as those applicable to the AGN. An obscured starburst located near region R1 of Figure 8 is also unlikely. A biconical outflow similar to that in, e.g., M82 (Walter et al. 2002; Veilleux et al. 2009) would mainly show redshifted emission in the northern hemisphere of the galaxy and blueshifted emission in the southern hemisphere of the galaxy. The SINFONI data indicate a symmetry of the redshifted emission in the north and south hemispheres, and an asymmetry of the redshifted versus blueshifted emission to the west and east of region R3.

Contrarily to an AGN or starburst driven wind, numerous findings support the jet–ISM interaction: the gas emission is brighter northwest than southeast of the nucleus (like the plasma emission itself), the non-thermally excited gas lies parallel to the jet trail, outflows are initiated in at least four discrete regions along the jet trail, the highest gas line of sight velocities (<−600 or >600 km s⁻¹) are attained near the proposed jet–ISM interaction points, and the gas emission can be blueshifted/redshifted in the same hemisphere.

As the plasma propagates, bow shocks push the surrounding tenuous gas and form an expanding cocoon. On its turn, the cocoon applies ram pressure to clouds that it encounters as it expands. The emission from the latter clouds can be either blueshifted or redshifted on the same side/hemisphere of the galaxy, depending on the positioning of the jet and the gas with respect to us. If the jet travels between us and background clouds, then we see redshifted emission from the background clouds that are being pushed further away from us by the cocoon. If foreground clouds exist between the jet and us, then we see blueshifted emission from the foreground clouds that are being pushed closer to us by the cocoon (see also Garci´a-Burillo et al. 2014). In regions where the jet plasma directly impacts clouds, then both blueshifted and redshifted emission from the scattered material can be detected. If the jet propagates through a clumpy disk, the plasma can get deflected on several dense regions and the cocoon can push several clouds. Then, the cloud distribution is most important for determining how many outflows will be initiated and in which directions they will be headed (Wagner & Bicknell 2011).

One schematic representation of the jet–ISM geometry explaining all observed outflows in IC5063 is presented in Figure 14. It follows the cloud naming convention of Figure 8, under the assumption that the large-scale ${\rm{H}}\beta$ and [O iii] emission in the FOC image is associated with the jet trail, as advocated by Morganti et al. (2007). Figure 14 shows a jet propagating through a clumpy medium and its cocoon on the disk plane. Northwest of the nucleus, most clouds are located behind the jet, whereas southeast of the nucleus, most clouds are located in front of the jet (with respect to us). This is in line with several results: with the small inclination of the jet with respect to the disk plane (Figure 8, Morganti et al. 2007) making the northwest radio lobe visible below the dust lane in the southern galaxy hemisphere, with the NIR obscuration being higher at the southeast radio lobe than at the northwest radio lobe (Figure 3), and with the gas emission in the FOC image being diffuse southeast of the nucleus—possibly due to high ram pressure (Figure 8).

Zoom In Zoom Out Reset image size

In this picture, the jet and its cocoon drive a mostly redshifted outflow northwest of the nucleus and a mostly blueshifted outflow southeast of the nucleus, clearing out some of the gas in the disk. Northwest of the nucleus, consecutive major jet–ISM interactions are occurring, starting at the base of the biconical outflow (near R1; Figure 14). There, the passage of the plasma by the near side of a dense ISM structure, as that seen in the ALMA CO(2-1) data (Morganti et al. 2015), causes the redshifted biconical emission. Part of the plasma is likely to be scattered to the south of the disk plane, as indicated by the weak 17.8 GHz emission, which extends along the southern biconical outflow axis and which ends where the fast outflow ends (Morganti et al. 2007). Then, near the north radio lobe, in region R4, an interaction between the jet and the far side of another gas clump generates the outflow coming toward us (Figure 14). This clump, oriented north south, could have been shaped by plasma that got deflected at the north radio lobe and that drilled a channel on its way out of the disk (Figure 8). Again, the 17.8 GHz image of Morganti et al. (2007) shows diffuse radio emission north of region R4 (starting north of R3 and ending near R5). Between regions R1 and R4, at the NLR base (R3), we see either the spatial overlap of the outflows at R1 and R4, or another jet–ISM interaction that scatters gas in many directions. Further up the NLR cone, in region R5, the jet again pushes away from us the near side of a cloud. On the other side of the nucleus, near the south radio lobe, scattered gas is accelerated in both lines of sight near the tip of the radio jet (Figures 8, 14).

The progression in the location of the diffuse gas emission with velocity can be explained by the jet propagation in more than one way. We could be observing tenuous ISM material that got accumulated and accelerated before impacting the dense structure at R1, material that got scattered from that dense structure, or a combination of the two. In the first case, we observe the motions of a cocoon-driven outflow that gets mass loaded as it moves from the nucleus to the outer parts of the galaxy. The velocity of this outflow will be higher near the jet cocoon (i.e., closer to the dense structure at the biconical outflow starting point). Once at the dense structure, the mass-loaded outflow will appear deflected in two directions perpendicular to the initial trajectory because of line of sight effects. In the second case, we observe an outflow that got mainly mass loaded after the impact of the jet upon the dense structure, when gas got bounced to different directions. The gas scattered close to our line of sight appears to be moving faster and to be located closer to the scattering surface, whereas the gas scattered nearly perpendicular to our line of sight appears to be moving slower and to be closer to the nucleus. In either way, scattering of the plasma links the gas entrainment along the jet trail and perpendicular to the jet trail, as well as in small and large scales. The high-velocity-dispersion regions near R1–R4 (e.g., the ridge northeast of R1 in the [Fe ii] kinematic map) are indicative of scattering. A spatial overlap of two accelerated/scattered gas components in projection or in volume (i.e., at the interface between two flows) can add to the observed dispersion.

The diffuse outflowing gas that we detected within the central square kiloparsec of IC5063 reaches hundreds of parsec away from visible jet breaking points. Its total northeast to southwest extent exceeds 700 pc. In the south of the jet trail, the outflowing gas emission fills the entire volume between the galaxy minor axis and the NLR southwest branch (Figure 8). Likewise, in the north of the jet trail, the regions with an intermediate-velocity outflow (Figures 4, 8) together with the region of oversaturated H₂ (1-0) S(3)/S(1) emission (Figure 12), again fill the volume between the galaxy minor axis and the NLR northwest branch. This emission, thus, spreads for 700–900 pc along its two southeast to northwest borders. The outflow detection in an area of order 1 kpc² emphasises the role of jets in affecting the reservoir for star formation. Notably, the simulations of Wagner & Bicknell (2011) mimicking the propagation of a jet in a clumpy disk, predicted that the turbulence would be so wide-spread.

The distance out to which we detect the diffuse gas depends on the flow travel time or on the presence of a physical obstacle such as star-forming regions (with excess ${{\rm{H}}}_{2}$/ $\mathrm{Pa}\alpha$ emission; Figure 3), in addition to the depth of the observations. The gas travel time gives us a limit on the timescales of the (intermittent or not) relativistic plasma ejection. If the Fe ions were moving at an average projected velocity of 900 km s⁻¹, then 4 × 10⁵ years would be required for them to travel from the disk to 350 pc north of it. For the molecular gas moving at −100 km s⁻¹, the timescale would be an order of magnitude higher, but it would be more uncertain: the fraction of molecular gas that forms in the flow from dense atomic gas and the pertinent timescales for its formation are unknown, rendering the use of a constant, low velocity unclear.