Very Large Array Multiband Radio Imaging of the Triple AGN Candidate SDSS J0849+1114

Both nucleus A and nucleus C are clearly detected in all four bands. Liu et al. (2019) reported 9.0 GHz (X-band) detection of nuclei A and C, and here we have new detections at 3.0, 6.0, and 15.0 GHz. Moreover, interesting structures are revealed in both nuclei.

Figure 2 provides a close-up view of nucleus A at 15.0 GHz and nucleus C at 10.0 GHz. In Figure 2(a), nucleus A exhibits an elongated morphology, reminiscent of a radio jet pointing along southeast–northwest at a position angle of ∼307° (east from the north) with an apparent extent of ∼06 (∼0.9 kpc). This elongated feature (larger than the corresponding synthesized beam size) is also evident in the 10.0 GHz image of nucleus A (Figure 1(c)). The intensity peak of nucleus A in the HST Y-band (orange contours; Figure 1(e)) and U-band images (cyan contours; Figure 1(f)) is consistent with the 15.0 GHz intensity peak at [RA, DEC] = $[{08}^{{\rm{h}}}{49}^{{\rm{m}}}5\buildrel{\rm{s}}\over{.} 526,+11^\circ 14^{\prime} 47\buildrel{\prime\prime}\over{.} 57]$ (marked as the white cross in Figure 2(a)) to within the astrometric uncertainty of ∼015 between the optical and radio images, suggesting that the latter may be the position of the jet base. If this were the case, a discrete knot seen at ∼03 southeast of the jet base may be the trace of an otherwise insignificant counterjet, likely due to the relativistic Doppler dimming effect. Due to the extended nature of nucleus A, we report its integrated flux density at each frequency (Table 3), which is measured using the CASA task IMFIT with a two-dimensional Gaussian model. The resultant integrated flux density (S_ν ) is 17.75 ± 0.56, 10.28 ± 0.56, 6.26 ± 0.21, and 4.21 ± 0.17 mJy at 3, 6, 10, and 15 GHz, respectively. Using orthogonal distance regression, we derive a spectral index of α = −0.90 ± 0.04 (S_ν ∝ ν^α ) for nucleus A. Accounting for the coherence loss, the measurements of nucleus A from the VLA agree with the results from the European VLBI Network (EVN) at 1.7 GHz (Gabányi et al. 2019).

Zoom In Zoom Out Reset image size

Figure 2(b) shows that the 10 GHz emission from nucleus C is dominated by a compact core, which peaks at [R.A., decl.] = $[{08}^{{\rm{h}}}{49}^{{\rm{m}}}05\buildrel{\rm{s}}\over{.} 436,+11^\circ 14^{\prime} 50\buildrel{\prime\prime}\over{.} 76]$ and is coincident with the southern intensity peak in the Y-band image (orange contours; Figure 1(e)), to within the astrometric uncertainty of 015. On the other hand, the U-band intensity peak (also coincident with the northern peak in the Y-band) is significantly offset (∼ 05 to the northeast) from the radio peak. If the latter marked the true position of the putative MBH, the U-band peak may be understood as the site of circumnuclear star formation. Significant 10.0 GHz emission seen to the immediate northeast of the radio core may be associated with this star-forming activity. Strong extended emission is also present to the west of the radio core, exhibiting an edge-brightened morphology reminiscent of a radio bubble. This bubble-like feature may be connected to the radio core with weak emission. The integrated flux density of nucleus C is measured to be 1.62 ± 0.09, 0.83 ± 0.07, 0.45 ± 0.04, and 0.32 ± 0.02 mJy at 3.0, 6.0, 10.0, and 15.0 GHz, respectively. The best-fit spectral index is −1.03 ± 0.04.

Despite the unprecedented sensitivity of these images, nucleus B remains undetected in all four bands. Therefore, we estimate a 3σ upper limit for the peak flux density of nucleus B, which is 15, 15, 15, and 18 μJy beam⁻¹ at 3.0, 6.0, 10.0, and 15.0 GHz, respectively. Here the 10 GHz upper limit is from the B-configuration concatenated image, which has an rms 1.4 times lower than that of the A-configuration image. These limits are much tighter than reported by Gabányi et al. (2019) based on the EVN observations, which is <450 μJy beam⁻¹ at 1.7 GHz. Figure 3 displays the observed radio spectral energy distributions of the triple nucleus (upper limits in the case of nucleus B), along with the best-fit power law.

Zoom In Zoom Out Reset image size

The VLA images also reveal remarkable features on the galactic scales. Most prominent is a linear feature through nucleus A in the 3.0, 6.0, and 10.0 GHz images (Figure 1), which is reminiscent of a two-sided jet. To distinguish it from the jet-like feature within half an arcsecond of nucleus A (Figure 2(a)), we shall refer to this linear feature as the outer jet (and the former as the inner jet). Remarkably, the position angle of the outer jet is ∼327°, i.e., deviating from that of the inner jet by 20°. The outer jet is more extended on the southeast side, reaching an apparent extent of at least ∼55 (∼8.1 kpc) in the 3.0 GHz image, which is most sensitive to extended features. The S-band image shows that the far side is brightened and slightly bent eastward, perhaps due to deflection by some external pressure. On the northwestern side, the far end of the outer jet is defined by an arc-shaped feature located at ∼18 from nucleus A, which is clearly seen in the 10.0 and 15.0 GHz images and reminiscent of a jet-driven shock. That the "arc" and nucleus A are connected by weak emission is evident in the 3.0 and 6.0 GHz images. The markedly different radio morphology on the two sides of nucleus A might be due to an intrinsic difference in the galactic-scale environment or the relativistic Doppler dimming. We measure the integrated flux densities of the arc in individual bands adopting a two-dimensional Gaussian model, which are given in Table 3 and plotted in Figure 3. The best-fit spectral index of the arc is found to be −0.99 ± 0.05, consistent with synchrotron radiation from shock-accelerated high-energy particles. The arc has no clear counterpart in the HST images. A visual examination of the Chandra image of (Liu et al. 2019, Figure 6 therein; see also Pfeifle et al. 2019a) suggests that a few X-ray photons are detected at the position of the arc, but deeper Chandra exposures are necessary to confirm this marginal excess. We have also examined the optical spectra presented in Liu et al. (2019) for potential shock-induced spectral features at the position of the arc. No clear evidence of this kind is found.

A two-sided linear feature through nucleus C is also evident in all four bands but most clearly seen in the 3.0 and 6.0 GHz images (Figure 1). This linear feature has a length of ∼15 (∼2.2 kpc) on both sides and a width of ∼06 (∼0.9 kpc). On the western side, the linear feature ends at a prominent knot (hereafter called the "apex"), which may be the top of the bubble-like feature shown in Figure 2(b). Emission on the eastern side is substantially fainter compared to its western counterpart and shows no sign of a jet-inflated bubble. Again, this difference on the two sides of nucleus C might be due to an intrinsic difference in the environment. We find no significant optical or X-ray counterpart of this two-sided radio feature. We measure the integrated flux densities of the apex in individual bands adopting a two-dimensional Gaussian model (Table 3 and Figure 3). The best-fit spectral index is −1.07 ± 0.08, again suggestive of a synchrotron origin.

The inner jet in nucleus A points along southeast–northwest at a position angle of ∼307° extending to 06 (0.9 kpc). The inner counterjet can be resolved at 15.0 GHz, which follows the same direction. On the scale up to 1'' (1.5 kpc) away from the peak of nucleus A, the outer jets on both sides deviate from the position angle of the inner jet by ∼20°. This indicates that the jet and the counterjet have both turned an angle of 20° on their way forward. However, it is too coincidental to explain the same offset angle of the jets on both sides due primarily to an external pressure. We speculate that a more natural scenario to explain the offset could be that a spinning black hole is powering episodic jets. An older jet points at a certain position angle. When a newly born younger jet emerges, the jet axis has rotated by another certain angle, resulting in the younger jet not moving forward at the previous position angle of the older jet. While galaxy mergers are not a necessary condition for radio jet reorientation, it is plausible that the ongoing galaxy merger may have promoted more gas inflows that could affect both the accretion states of the active MBHs and the dynamical state of the circumnuclear medium. Indeed, theoretical studies demonstrate that the spin axis of a strongly accreting MBH can be significantly altered and ultimately aligned with the angular momentum of a gas inflow feeding the accretion disk over a timescale of ≲Myr (Fiacconi et al. 2018; Cenci et al. 2021).

We estimate the minimum time gap between the old and young jets of nucleus A. By assuming the jet has an average speed of 0.1c (c is the speed of light), it takes ∼1.5 × 10⁵ yr to travel 4.7 kpc. As for the young jet, a smaller distance of 0.9 kpc needs ∼2.9 × 10⁴ yr of travel. Therefore, the minimum time gap is around 1.2 × 10⁵ yr, compatible with the aforementioned timescale suggested by theoretical studies (Fiacconi et al. 2018; Cenci et al. 2021). We note that the relatively strong magnetic field could explain a negative spectral index. A rapid synchrotron cooling happens in a strong magnetic field.

The 15 GHz (Ku-band) image well describes the outline of the western lobe that has a length of ∼15 (∼2.2 kpc) and a width of ∼06 (∼0.9 kpc). The apex seems to be the hot spot of the western lobe. Approximating the lobe as an ellipsoid with a circular cross section and assuming that the line of sight is close to edge-on, the volume of the lobe is ∼0.8 kpc³. The integrated flux density in such a region is 0.81 ± 0.03, 0.52 ± 0.04, and 0.39 ± 0.02 mJy at 6.0, 10.0, and 15.0 GHz, respectively, and at least 1.35 mJy at 3.0 GHz. The average radio spectral index of this region is −0.78 ± 0.02. Adopting the classic energy equipartition assumption, the minimum energy density (${u}_{\min }$; Equation (25) in Govoni & Feretti (2004)) in the lobe is estimated to be ∼2.3 × 10⁻⁹ erg cm⁻³, for an equipartition magnetic field strength of ${B}_{\mathrm{eq}}={(24\pi {u}_{\min }/7)}^{\tfrac{1}{2}}\approx 160\ \mu$ G. The total energy (i.e., a sum of magnetic and particle energy) of the lobe amounts to 5.0 × 10⁵⁵ erg, which is comparable to the work done during its expansion.

Because the steep spectrum extends to 3.0 GHz, the synchrotron cooling timescale is at least more than 4 × 10⁵ yr based on the equipartition magnetic field strength. If the jet had a speed of 0.1 c, it would need a travel time of 7 × 10⁴ yr. In fact, the lobe expansion could take more time by at least a factor of 10. Here, we choose a minimum timescale of 7 × 10⁵ yr and the jet power is estimated to be ∼4 × 10⁴² erg s⁻¹. Such power is plausibly available from nucleus C, which is a Seyfert galaxy with an estimated black hole mass of 5 × 10⁶ M_⊙ (Reines & Volonteri 2015; Liu et al. 2019). On the other hand, it is hard to determine any visible star-forming impact from the jet/lobe feedback in such a short timescale.

The lobe associated with nucleus C is an interesting analog to the famous case of the Circinus galaxy. Exhibiting two extended edge-brightened radio lobes each with a size of ∼1.5 kpc, the Circinus galaxy has an estimated jet power of ∼10⁴¹ erg s⁻¹ (Mingo et al. 2012). This is about 1 order of magnitude lower than the one we find above for the lobe associated with nucleus C, and so is the total energy involved, 2 × 10⁵⁵ erg (Mingo et al. 2012).

The nondetection of nucleus B gives an upper limit of 15, 15, 15, and 18 μJy beam⁻¹ at 3, 6, 10, and 15 GHz, respectively. Assuming S_ν ∝ ν^−0.59, which is appropriate for star-forming galaxies (Klein et al. 2018), we extrapolate our 3 GHz measurement to 1.45 GHz, which results in 28 μJy beam⁻¹. Using the L_{1.4 GHz}–SFR_UV+TIR correlation (Davies et al. 2016, 2017), the inferred 3σupper limit of star formation rate (SFR) is 0.4 M_⊙ yr⁻¹. This is slightly more stringent than but is broadly consistent with the 3σ SFR upper limit of <0.8 M_⊙ yr⁻¹ reported in Liu et al. (2019). A flatter spectral index would lead to a lower SFR; thus the inferred limit is a conservative value.

Liu et al. (2019) also found the dust attenuation corrected SFR is ∼0.2–7 M_⊙ yr⁻¹ derived from the HST U band and continuum index. Based on near-IR spectroscopy from the Large Binocular Telescope, Pfeifle et al. (2019a) obtained an SFR of ∼0.48 M_⊙ yr⁻¹ for nucleus B. Both estimates are consistent with our radio limit. In addition, based on the Kennicutt–Schmidt law (Kennicutt & Evans 2012), our limit is also in agreement with the nondetection of molecular gas in nucleus B at arcsecond resolution, which corresponds to an upper limit of 79 M_⊙ pc⁻² in molecular gas surface density (M. Hou et al. 2022, in preparation). Therefore, our VLA results reinforce the conclusion that star-forming activity alone is insufficient to account for the observed X-ray flux in nucleus B, and an additional heating source such as an AGN and/or shocks are required (Liu et al. 2019; Pfeifle et al. 2019a), although such additional power is not necessarily produced by the putative MBH in nucleus B.

In summary, we have presented new VLA observations of SDSS J0849+1114 at 3.0, 6.0, 10.0, and 15 GHz, which provide an unprecedented radio view for this triple AGN candidate. Two of the three nuclei, nuclei A and C, are detected for the first time at 3.0, 6.0, and 15 GHz. They both show a steep spectrum over 3–15 GHz, consistent with an origin of synchrotron radiation. The high-resolution images also reveal kiloparsec-scale extended features related to both nuclei, which can be attributed to AGN-driven jets/outflows. Nucleus B remains undetected at all four frequencies, and further studies are warranted to unambiguously determine the nature of this nucleus.