DISENTANGLING THE CIRCUMNUCLEAR ENVIRONS OF CENTAURUS A: GASEOUS SPIRAL ARMS IN A GIANT ELLIPTICAL GALAXY

Some ellipticals show rotating disk-like features of gas and dust, which are expected to form as a result of gas accretion from the intergalactic medium, cannibalization of other galaxies, or re-accretion of material originally expelled during the major merger event that resulted in the formation of the elliptical itself (e.g., Henkel & Wiklind 1997; Young 2002; Davis et al. 2011; Young et al. 2011). Sage & Galletta (1993) proposed an evolutionary sequence of disk formation in dust lane ellipticals via accretion, from irregularly distributed gas to more evolved and settled disks. However, our understanding of the properties and evolution of such relatively settled disks is still poorly understood, partly because most observations of the gas in elliptical galaxies are hampered by poor angular resolution and/or sensitivity. In particular, it is not known if these newly created disks in elliptical galaxies are uniform once settled, or whether deviations from axisymmetry such as prominent spiral structures can develop, affecting the subsequent star formation (SF) patterns, and the overall evolution of the disk itself.

At a distance of D ≃ 3.8 Mpc (where 1'' roughly corresponds to 18 pc) (Harris et al. 2010), NGC 5128 (hereafter Cen A) is the closest giant elliptical (van den Bergh 1976; Israel 1998; Harris et al. 2012) with a prominent dust lane along its minor axis (Israel 1998; Harris et al. 2012; see Figure 1, left). This dust lane contains large amounts of gas and dust, as traced by Hα (Nicholson et al. 1992), mid-IR (Quillen et al. 2006), H i (van Gorkom et al. 1990; Struve et al. 2010), and CO (Quillen et al. 1992; Rydbeck et al. 1993; Espada et al. 2009).

Zoom In Zoom Out Reset image size

The H i emission still shows unsettled gas at r > 6 kpc, including tail/arm-like structures, and it is estimated that the merger event occurred within a timescale of only 0.3 Gyr (Struve et al. 2010). The gas in molecular phase and the dusty component inside the inner few kiloparsecs are well settled within a warped disk of about 7' with an inclination of i ∼ 70° (Quillen et al. 2006), although its detailed distribution at scales of a few hundreds of parsec has hitherto been poorly understood.

The most popular model to reproduce the large-scale disk structure is that of a homogenous warped and thin disk (e.g., Nicholson et al. 1992; Quillen et al. 1992, 2006, 2010; Kainulainen et al. 2009; Struve et al. 2010). According to these models, the disk crosses the line of sight at two radii. One of the most remarkable features in favor of this model is that in projection it can reproduce the observed parallelogram structure elongated along ∼3' (3 kpc) and a P.A. = 120° (Quillen et al. 2006).

However, the inner 1 kpc of Cen A (up to a few 100 pc from the nucleus) is very complex. Espada et al. (2009, hereafter Paper I) report ¹²CO(2–1) Submillimeter Array (SMA) observations of its nuclear region in a single pointing, with a resolution of 60 × 24 (100 × 40 pc), revealing the detailed distribution and kinematics of the molecular gas in the central inner kiloparsec, including a compact circumnuclear disk of molecular gas surrounding the Active Galactic Nucleus (AGN), as well as outer molecular gas partly associated with the parallelogram structure. While the warped disk models reproduce the observed features (distribution, kinematics, dust lane appearance) at a large scale, Paper I shows that the molecular gas distribution in the inner kiloparsec is not sufficiently in agreement with these models. A weak bisymmetric potential was proposed in Paper I to explain the deviations with respect to this model.

We present new observations using the SMA¹⁰ (Ho et al. 2004), which allow us to shed more light into the complex few inner kiloparsecs. The resolution of the mosaic we present here is a factor of 45 higher than existing ¹²CO(2–1) maps covering the dust lane, with a resolution of 23'' (or ∼380 pc; Rydbeck et al. 1993). Our mosaic covers an area three times larger than previously published interferometric CO maps (Paper I) to fully cover the entire parallelogram structure.

Cen A was observed at 1.3 mm using the SMA with seven antennas on 2008 April 22 and 26. The field of view is characterized by a half-power beam width of the primary beam of an SMA antenna of 52'' (0.9 kpc). We observe a five pointing mosaic to cover 156 × 52'' (or 2.6 × 0.9 kpc). The digital correlator was configured with 6144 channels (2 GHz bandwidth), resulting in a velocity resolution of about 0.5 km s⁻¹. The receivers were tuned to the redshifted ¹²CO(2–1) (ν_rest = 230.538 GHz) emission line in the upper sideband, using V_LSR = 550 km s⁻¹. Note that velocities are expressed throughout this Letter with respect to the LSR using the radio convention. We used R.A. = 13^h25^m276 and decl. = −43°01'088 (J2000) as our phase center (AGN position R.A. = 13^h25^m27615 and decl. = −43°01'08805; Ma et al. 1998). The maximum elevation of the source at the SMA site is ≃ 27°, forcing us to observe only under very stable atmospheric conditions, with zenith opacities of typically τ₂₂₅ ∼ 0.10. In order to have a beam shape as close to circular as possible we used a compact configuration with longer N–S baselines. Minimum and maximum projected baselines were 6 m and 108 m. The maximum angular scale observable with the shortest baselines is 25'', which considerably limits the amount of missing flux to emission with angular scales larger than this.

The two tracks were reduced independently. The editing and calibration of the data were done with the SMA-adapted MIR software.¹¹ 3C273 and 3C279 were used for passband calibration. An initial gain calibration was performed using J1337-129, which is at an angular distance of 30° from the target. We confirm that the continuum emission toward Cen A was found to be unresolved (Paper I). The gain calibration for the five pointings was then refined using the averaged line-free channels in Cen A itself (central pointing). Callisto and Ganymede were used as absolute flux calibrators. Overall, we estimate the absolute flux uncertainties on the order of 10%, from the comparison of both data sets. Finally, we combined both tracks.

An image of the ¹²CO(2–1) emission was produced using MIRIAD (Sault et al. 1995). A careful subtraction of the continuum was done using line-free channels with the task UVLIN. The data were cleaned using MOSMEM with uniform weighting. The synthesized beam is 44 × 19 (80 × 40 pc) with a major axis P.A. = 253. The task MOMENT was used to calculate the ¹²CO(2–1) integrated flux density distribution (2σ clipping) and the intensity-weighted velocity field distribution (clipped at 3σ). The rms noise level of the integrated intensity map is 6 Jy beam⁻¹ km s⁻¹. Figure 2 shows the SMA ¹²CO(2–1) integrated intensity map and velocity field along the inner 3' (∼3 kpc).

Zoom In Zoom Out Reset image size

From the ¹²CO(2–1) emission distribution in Figure 2, we confirm that the molecular gas is preferentially located along two filamentary structures resembling spiral-arm-like features to the SE and NW of the circumnuclear gas (galactocentric radius <200 pc and position angle (P.A.) = 155°; Paper I). The larger field of view of our observations demonstrates that these filamentary structures extend at least 10, and they are curved toward the NE and SW, respectively, as the projected distance from the nucleus increases. The approaching and receding sides of the disk are on the E and W, respectively, and the velocity range spans about 600 km s⁻¹, from 200 to 800 km s⁻¹.

The near side of the disk at r < 1.3 kpc is in the south (Quillen et al. 2010). From the morphology of the bisymmetric spiral-arm-like features and the velocity field in Figure 2, we infer that they are trailing (convex side advances). The projected angular widths of the spiral-arm-like features are ∼10''. Adopting an inclination of 70° (Quillen et al. 2010), their widths are 500 ± 200 pc.

The shape of galactic spiral arms is usually logarithmic in nature and independent of scale (e.g., Seigar & James 1998). One of the best geometric measures to represent the spiral arms is the pitch angle, ϕ, defined as the angle between the line tangent to a circle and the line tangent to a logarithmic spiral at a specified radius. We present in Figure 3 the deprojected CO(2–1) distribution, as well as a best-fit representation of a logarithmic spiral pattern with pitch angle ϕ = 20°. We used a set of models of bisymmetric spiral arms with an inclination of 70°. We varied the pitch angles in steps of 10° from 10° to 60°, and found that ϕ = 20° fitted best. This pitch angle is within the expected range for late-type spirals (Kennicutt 1981) for a maximum rotation velocity for Cen A of V ≃ 300 km s⁻¹.

Zoom In Zoom Out Reset image size

In Figure 4 we show a comparison between the CO(2–1) and mid-IR 24 μm emission obtained with Spitzer (Quillen et al. 2006), with resolutions of about 6'', which complements Figure 1 with mid-IR 8 μm emission. In general the molecular gas distribution is similar to that of dust emission.

Zoom In Zoom Out Reset image size

As mentioned previously, a warped and thin disk model is usually used to reproduce the observed properties at large scales (>1 kpc), especially the parallelogram structure seen in dust emission (Quillen et al. 2006, 2010). However, the morphology of CO(2–1) emission resembles better spiral arms rather than a parallelogram structure. This rules out the single contribution of a homogeneous disk, otherwise we would also see filaments to the NE and SW of the circumnuclear gas, as the warped and thin disk model predicts that they are of comparable brightness as the observed NW and SE filaments (see, for example, Figure 4 in Quillen et al. 2006). This cannot be explained by flux loss due to missing spacings. From the model, these filaments should be of the order of less than 10'', compactness similar to the filaments to the NE and SW, and thus they would have been detected.

Therefore, we interpret the two CO(2–1) filamentary structures as spiral arms. Their main properties are similar to those found in disk galaxies: they are trailing, the linear widths are a few hundred parsecs, and the pitch angle is typical of spiral galaxies. This provides evidence that spiral arms can develop within a giant elliptical if enough cold gas exists. However, note that the molecular gas properties in the spiral arms might be considerably different from those in spiral galaxies, as it coexists within a triaxial potential and the stellar surface density changes more abruptly with radius in the case of a giant elliptical.

A consequence of the compression produced by these spiral arms is that it is expected to trigger SF on the leading edge. This is consistent with abundant SF traced by Paα (Marconi et al. 2000) for example, coincident at least with the southern molecular spiral arm (nearest part and thus likely the least obscured). The spiral features seen in the ¹²CO emission is also associated with poorer angular resolution maps, such as the SCUBA 450 μm emission map in Leeuw et al. (2002), as well as the pure rotational line of molecular hydrogen H₂ (J = 2–0) S(0) (28.22 μm) emission observed with Spitzer/IRS by Quillen et al. (2008). The molecular hydrogen transition H₂ (J = 2–0) S(0) indicates the presence of gas with T ∼ 200 K (Quillen et al. 2008), which is likely tracing photodissociation regions associated with abundant SF.

A small perturbation could have triggered this spirality, such as a non-axisymmetric weak potential (Paper I) or a minor merger after the disk was relatively well settled (e.g., Purcell et al. 2011). From the H i structure and kinematics we infer that the formation of spiral arms took place in less than 0.3 Gyr (Struve et al. 2010), which is likely the most accurate measure of time at our disposal since H i is one of the most sensitive components of interaction. Although evidence suggests that spiral arms are transient (Sellwood 2011; Foyle et al. 2011), recent simulations claim that spiral structures may be long lived (D'Onghia et al. 2012). Three-dimensional self-gravitating models for late-type spirals suggest that the pitch angle we find here would correspond to a long-lasting feature, because the density response tends to avoid larger pitch angles (Pérez-Villegas et al. 2012). Similar simulations in the deep potentials of giant ellipticals would be needed to investigate the response of the gas.

The study of the properties of recently accreted molecular gas in the deep potential wells of elliptical galaxies is a powerful tool in disentangling how galaxies form and evolve. Recent smoothed-particle hydrodynamic cosmological simulations in the Λ cold dark matter scenario (Komatsu et al. 2011) are able to reproduce most of the properties of disks (Agertz et al. 2011; Guedes et al. 2011; Doménech-Moral et al. 2012) at the current epoch, including their fine structure morphology, the observed angular momentum, the Tully–Fisher relation, and the star formation rate to gas surface density ratio (Kennicutt–Schmidt). However, to understand galaxy evolution it is essential to anchor these simulations with the actual properties of nascent disk galaxies. The properties of disks, in particular the existence of spiral arms, in these systems which are in the early stages following significant accretion can be compared with numerical simulations. Not only can we compare the observed gas distribution and kinematics of recently formed disks to models, but also the effect of SF and AGN feedback in the disk evolution. Moreover, with the advent of ALMA, it is now possible to validate theories of disk formation as they enter a deep gravitational potential, performing similar studies in objects along the evolutionary sequence of accreted gas in dust lane ellipticals (Sage & Galletta 1993), from systems whose gas is still irregularly distributed to more evolved and settled disks such as in Cen A.

We thank the referee for useful comments that improved the focus of this Letter. This research has made use of NASA's Astrophysics Data System Bibliographic Services, and has also made use of the NASA/IPAC Extragalactic Database (NED) which is operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration. This research was partially supported by a Marie Curie International Fellowship within the 6th European Community Framework Programme.

Facility: SMA - SubMillimeter Array