MeerKAT Reveals Evidence of a Radio Megahalo at GHz Frequency

The paper presents MeerKAT L-band (1.28 GHz) observations of the massive galaxy cluster RXC J0528.9−3927 (J0528) at z = 0.284, revealing faint diffuse radio emission surrounding its central radio halo. The known ∼1.14 Mpc halo is embedded within previously unreported low-surface-brightness emission extending to ∼2 Mpc. The surface-brightness profile of the emission shows a distinct flattening beyond ∼0.55R500 (where R500 = 1.26 Mpc), suggesting that the outer emission forms an additional component. These properties make the cluster a candidate megahalo system and potentially the first detected at GHz frequencies, although deeper multi-frequency observations are required for confirmation.

The observations used archival MeerKAT data with 12.9 hours on-source time across three epochs (7.5, 1.7, and 3.7 hours), reduced with the CARACAL pipeline. Compact sources were subtracted using a uv-range > 1 kλ, and the final diffuse emission map was produced with a 26′′ restoring beam (rms ∼ 7.5 µJy beam−1). At 1.28 GHz, the connected 6σrms contour encloses diffuse emission with a largest linear size of ∼1.14 Mpc, consistent with the previously reported halo. The map also reveals previously unreported low-surface-brightness emission traced by the connected 3σrms contour, extending the total diffuse emission to ∼1.96 Mpc.

The surface-brightness profile was derived using elliptical annuli centred on the halo peak, following the fitting procedure of Cuciti et al. (2022), with a beam-convolved single-exponential model. Beyond r ≃ 162.5′′ (∼697 kpc), the profile systematically departs from the model, with this transition occurring at ∼0.55R500, comparable to radial departures reported for other megahalos. An elongated feature morphologically consistent with a peripheral radio relic was detected west–northwest of the cluster and masked during profile derivation.

The central halo enclosed by the connected 6σrms contour has an integrated flux density of Shalo = 6.53 ± 0.33 mJy. After excluding this region, the directly detected candidate-megahalo emission within the connected 3σrms contour has Sext = 2.37 ± 0.13 mJy and a mean surface brightness of 34.08 µJy beam−1. Assuming the candidate megahalo extends through the region occupied in projection by the central halo with the same mean surface brightness, the total flux density of the candidate megahalo is estimated to be Smega = 4.48 ± 0.25 mJy. Assuming a typical radio-halo spectral index of α = −1.3, the 1.4-GHz radio powers are Phalo,1.4 = (1.59 ± 0.08) × 10 24 W Hz−1 and Pmega,1.4 = (1.09 ± 0.06) × 10 24 W Hz−1. The volume-averaged emissivities are jhalo,1.4 = 9.4 × 10−43 and jmega,1.4 = 1.68 × 10−43 erg s−1 Hz−1 cm−3, making the candidate megahalo ∼5.6 times less emissive than the central halo, with Vmega/Vhalo = 3.83.

The discussion notes that J0528 has been classified as dynamically mixed, hosts a cold front, and shows possible north–south accretion or merger activity, supporting the scenario of an off-axis or minor merger capable of generating turbulence without fully disrupting the core. This is consistent with recent results suggesting a connection between extended diffuse radio emission and sloshing/cold-front features in cool-core clusters. The authors stress that the detection remains tentative owing to the low surface brightness and lack of spectral constraints, and deeper multi-frequency observations with matched angular resolution and uv coverage are required to confirm the nature of the emission. The detection is significant as it represents a unique example of megahalo-like diffuse emission detected at 1.28 GHz, a relatively high frequency where steep-spectrum emission is typically difficult to recover, demonstrating the capability of deep MeerKAT observations to probe low-surface-brightness non-thermal components of the ICM at GHz frequencies.

Improvements to AI Systems:

1. Adaptive Surface-Brightness Profile Fitting with Automatic Component Separation

• Train a model to fit multi-component exponential profiles to radio surface-brightness data, automatically detecting radial breaks (e.g., at 0.55R500) and flagging statistically significant departures from single-component models.

• The improved system can identify candidate megahalos or other diffuse emission components in low-S/N radio maps without manual intervention, and quantify the confidence of each component detection.

2. Multi-Epoch Radio Data Fusion and Artifact-Aware Source Subtraction

• Develop an AI pipeline that optimally combines archival observations from multiple epochs (e.g., 7.5, 1.7, 3.7 hours) with varying uv-coverage, using learned weighting to minimize artifacts and maximize sensitivity to extended low-surface-brightness emission.

• The system can automatically subtract compact sources using uv-range cuts (e.g., >1 kλ) while preserving diffuse structure, and output a robust diffuse-emission map with per-pixel uncertainty estimates.

3. Low-Surface-Brightness Feature Detection at High Frequencies

• Implement a deep-learning segmentation model trained on synthetic and real MeerKAT-like data to detect connected emission at low significance levels (e.g., 3σrms) and distinguish it from noise or residual artifacts.

• The improved system can reliably trace faint diffuse halos, relics, and megahalo candidates at GHz frequencies, even when the emission is steep-spectrum and faint, enabling new discoveries in existing archival data.

4. Physical Parameter Estimation from Radio Morphology

• Build a regression model that, given a detected diffuse emission map and cluster properties (e.g., R500, redshift), predicts integrated flux, volume-averaged emissivity, and total radio power, with uncertainty propagation from surface-brightness noise.

• The system can automatically compute quantities like Shalo, Sext, Smega, and j1.4, and compare them across clusters to classify halo types (e.g., megahalo vs. standard halo) based on emissivity ratios and size.

5. Automated Merger-Dynamics Inference from Radio and X-ray Joint Analysis

• Create a multi-modal AI that combines radio diffuse emission maps with X-ray-derived features (e.g., cold fronts, sloshing) to infer merger type (off-axis, minor, major) and turbulence generation efficiency.

• The improved system can predict whether a cluster is likely to host megahalo-like emission, and prioritize targets for deeper multi-frequency follow-up observations.

6. Spectral Index Prediction and Steep-Spectrum Recovery

• Train a generative model to predict the spectral index distribution across diffuse emission regions from a single-frequency map and cluster dynamical state, enabling extrapolation to other frequencies (e.g., 1.4 GHz) and assessing detectability.

• The system can estimate the likelihood of detecting a candidate megahalo at higher frequencies, guiding observational strategies for confirmation.

What the Improved AI System Can Do:

• Automatically process deep radio observations to detect and characterize ultra-faint diffuse emission, including megahalo candidates, with minimal human bias.

• Provide probabilistic classifications of radio halos and relics, including whether an outer component is physically distinct or an extension of the central halo.

• Predict the detectability of steep-spectrum emission at GHz frequencies, optimizing telescope time for follow-up.

• Integrate multi-wavelength data (radio, X-ray, SZ) to infer cluster merger histories and link them to non-thermal particle acceleration processes.

• Deliver ready-to-use catalogs of diffuse emission properties (flux, size, emissivity, spectral index) for large cluster surveys, enabling statistical studies of megahalo formation.