J0011+3443: a GPS compact symmetric object, gravitational lens, or dual AGN?

Abstract: We present new multi-frequency VLBA observations of \object{J0011+3443} (TXS\,0008+344, $z=0.89$) at 2.3, 4.9, 8.5, and, for the first time, 23.6\,GHz. The source consists of two compact components A and B at a projected separation of $314\pm2$\,pc, plus a third feature C detected at 23.6\,GHz at $0.6$\,mas from A. Archival low-resolution radio measurements confirm an integrated gigahertz-peaked spectrum, with an observed-frame peak frequency of $ν<em>\mathrm{peak}=0.73\pm0.08$\,GHz and a peak flux density of $S</em>\mathrm{peak}=879\pm125$\,mJy. Comparison with nearly frequency-matched low-resolution measurements shows that the VLBA recovers $0.85\pm0.11$ of the 4.85\,GHz flux density and $0.50\pm0.06$ of the 8.46\,GHz flux density. The lower recovered fraction at 8.5\,GHz suggests that low-surface-brightness emission is resolved out or falls below the VLBA surface-brightness sensitivity. We therefore interpret the VLBA component spectra as spectra of the compact recovered emission only. The 23.6\,GHz morphology, the absence of a detected flat-spectrum core, the similar compact spectra of A and B, and the steep integrated GHz spectrum favor an interpretation of \object{J0011+3443} as a GPS-class compact symmetric object, possibly in a short-lived or relic phase, although a dual-AGN origin cannot be excluded without multi-epoch astrometry.

• The paper presents multi-frequency VLBA observations showing stable 314 pc double morphology, symmetric steep-spectrum components, sub-equipartition brightness temperatures, and no detected flat-spectrum core, favoring a GPS compact symmetric object.
• The integrated spectrum peaks at 1.38 ± 0.15 GHz in the rest frame, matching the turnover–size relation, while missing VLBA flux—especially only 50% recovery at 8.5 GHz—limits interpretation of compact-component spectra and absorption physics.
• The study disfavors, but does not exclude, gravitational lensing and dual AGN scenarios, and proposes 23.6 GHz monitoring of component C for proper motion plus deeper optical/infrared imaging to search for a foreground deflector.

Introduction and observational context

The paper presents new multi-frequency VLBA observations of J0011+3443 (TXS 0008+344), a compact double radio source at $z = 0.89$, observed at 2.3, 4.9, and 8.5 GHz on 2022 March 8 and at 23.6 GHz — for the first time for this source — on 2022 February 8 (2608.19928). The source was flagged as a GPS candidate by its inverted low-frequency spectrum, and later as both a dual-AGN and gravitational milli-lens candidate by the SMILE program, which searches the Astrogeo VLBI database for compact multi-component sources. The central question addressed is which of three scenarios explains the two parsec-scale components A and B separated by $39.3 \pm 0.3$ mas ($314 \pm 2$ pc projected): a gigahertz-peaked-spectrum (GPS) compact symmetric object (CSO), a gravitationally milli-lensed background AGN, or a pair of independently active supermassive black holes.

The observations were conducted in phase-referencing mode against J0015+3216 with all ten VLBA stations, reduced with standard AIPS/Difmap procedures including ionospheric corrections, opacity-corrected amplitude calibration, self-calibration, and circular-Gaussian model fitting. Flux-density errors combine local rms noise with a 10% amplitude-calibration term.

Morphology and component properties

The double morphology is consistent across all four frequencies. Component A is brighter at every frequency, declining from $128.0 \pm 12.8$ mJy at 2.3 GHz to $1.9 \pm 0.2$ mJy at 23.6 GHz; B follows the same trend ($67.4 \pm 6.7$ to $1.0 \pm 0.1$ mJy). The flux ratio $S_\mathrm{A}/S_\mathrm{B} = 1.89 \pm 0.13$ is approximately constant over a factor of ten in frequency, indicating strong spectral symmetry between the components. At 23.6 GHz a third feature C is detected at $\sim 10\sigma$, $0.64 \pm 0.06$ mas from A at PA $39.3 \pm 0.3$0, offset by about $39.3 \pm 0.3$1 from the direction exactly opposite to B, and interpreted as sub-structure associated with A's hotspot region. No counterpart near B is seen.

Component A's fitted FWHM decreases monotonically from 3.42 mas at 2.3 GHz to 0.36 mas at 23.6 GHz. The authors attribute this not to core opacity stratification but to progressive resolution of low-surface-brightness hotspot/lobe emission. The peak brightness temperature of A is $39.3 \pm 0.3$2 K at 4.9 GHz — roughly an order of magnitude below the equipartition value of $39.3 \pm 0.3$3 K — while B has a lower limit $39.3 \pm 0.3$4 K. These sub-equipartition values are characteristic of lobe or hotspot plasma rather than Doppler-boosted cores, a point that carries significant weight in the later classification argument.

The A–B separation shows no statistically significant frequency dependence: the largest inter-frequency difference is $39.3 \pm 0.3$5 mas across 2.3–23.6 GHz. Combined with EVN separations of $39.3 \pm 0.3$6 and $39.3 \pm 0.3$7 mas measured in 2016 and 2020, the geometry is stable over six years.

Integrated spectrum and missing flux

Combining archival low-resolution flux densities spanning 0.144–8.46 GHz, the authors fit both a synchrotron-self-absorption parameterization with free optically thick index and an internal free-free absorption model. Both describe the data comparably well, so the data do not determine the physical absorption mechanism — a limitation stated explicitly. The SSA-like fit yields $39.3 \pm 0.3$8 GHz observed frame ($39.3 \pm 0.3$9 GHz rest frame) and $314 \pm 2$0 mJy; both models give $314 \pm 2$1. This confirms the GPS classification independent of any VLBI spectral analysis.

A key quantitative result concerns VLBI flux recovery. Comparing summed VLBA flux densities with nearly frequency-matched single-dish/survey measurements:

Frequency	VLBA sum (mJy)	Low-res (mJy)	Recovered fraction
4.9	$314 \pm 2$2	$314 \pm 2$3 (87GB)	$314 \pm 2$4
8.5	$314 \pm 2$5	$314 \pm 2$6 (CLASS)	$314 \pm 2$7

The substantially lower recovered fraction at 8.5 GHz indicates that low-surface-brightness emission is resolved out or falls below the VLBA surface-brightness sensitivity. The authors therefore interpret the VLBA component spectra strictly as spectra of the compact recovered emission only, and caution that they should not be used to infer the source-wide turnover or absorption mechanism. This missing-flux effect also plausibly explains why their VLBA component spectra are steeper than the EVN index $314 \pm 2$8 reported previously.

Discriminating among the three interpretations

Gravitational lensing is disfavored but not excluded. At 23.6 GHz, A resolves into a compact double (and into three sub-components A0–A2 in archival EVN imaging) while B remains unresolved; since lensing preserves surface brightness modulo magnification and parity, a simple lens would produce broadly similar sub-structure in both images. Additionally, no luminous lower-redshift deflector is identified along the line of sight, though the authors concede that milli-lensing by a dark compact foreground mass cannot be ruled out.

Dual AGN is judged less natural than the CSO interpretation. Both components have steep compact spectra, sub-equipartition brightness temperatures, similar spectra, and a constant flux ratio — properties expected of lobe/hotspot plasma, not of two unrelated active nuclei, one of which would be expected to show a flat or inverted core spectrum. No flat-spectrum core is detected at either component even at 23.6 GHz, and the source is unpolarized at 8.4 GHz ($314 \pm 2$9). Notably, the steep spectrum places the source outside the SMILE flat/inverted selection criterion ($128.0 \pm 12.8$0), which the authors flag as a caveat for future dual-AGN searches relying on such criteria.

CSO interpretation is favored by several mutually consistent results. The rest-frame turnover of $128.0 \pm 12.8$1 GHz agrees excellently with the $128.0 \pm 12.8$2 GHz predicted by the O'Dea–Baum turnover–size anticorrelation for a 0.31 kpc source. Equipartition estimates from the 4.9 GHz compact emission, assuming filling factor unity, equal proton/electron energy, and no beaming, give $128.0 \pm 12.8$3–12 mG per component, consistent with CSO/HFP hotspots. For these fields the CMB inverse-Compton field at $128.0 \pm 12.8$4 is negligible, and synchrotron cooling times for GHz-emitting electrons are only 205–660 yr — far shorter than the $128.0 \pm 12.8$5 yr source-crossing time — supporting the presence of radiatively evolved plasma. The non-detection of the central engine at all frequencies is consistent with a short-lived or relic CSO phase in which jet activity has ceased or declined; in that picture, component C could represent recently injected plasma at the A-side hotspot. The authors label this scenario speculative and propose a specific test: multi-epoch 23.6 GHz imaging should reveal proper motion if C is a traveling feature, versus monotonic fading if it is aging lobe plasma.

They also note honestly that an asymmetric alternative — an active nucleus in A with B as a single hotspot — cannot be entirely excluded on morphology alone, though the spectral symmetry argues against it.

Limitations and open questions

Several limitations bear directly on the strength of the conclusions. The integrated-spectrum fits rely on non-simultaneous, heterogeneous-resolution archival measurements, so the SSA-versus-FFA distinction remains open. The recovered flux fractions are indicative rather than precise for the same reason. The magnetic-field estimates derive only from compact VLBA-detected emission with an uncertainty of at least a factor of two, and the cooling times computed from them are not spectral ages, since the rest-frame break frequency is not robustly constrained by these data. The homogeneous-SSA field estimate scales as $128.0 \pm 12.8$6 and yields unphysically large values; the authors take this, together with the sub-equipartition brightness temperatures, as evidence that pure homogeneous SSA does not fully explain the turnover, plausibly implicating free-free absorption. Definitive exclusion of the dual-AGN hypothesis requires multi-epoch astrometry that this paper does not yet provide.

Conclusion

This work establishes J0011+3443 as a GPS-class compact symmetric object candidate through converging evidence: a confirmed integrated turnover at $128.0 \pm 12.8$7 GHz rest frame consistent with the turnover–size relation, stable two-sided morphology over six years, symmetric steep component spectra with constant flux ratio, sub-equipartition brightness temperatures, absence of any detected flat-spectrum core, and hotspot-like equipartition fields. The dual-AGN and milli-lens interpretations are disfavored but not definitively eliminated. The paper identifies a clear path forward: multi-epoch 23.6 GHz astrometry of component C to test for proper motion, and deeper optical/near-infrared imaging to constrain any foreground deflector.