Probing the behaviour of the X-ray binary Cygnus X-3 with very-long-baseline radio interferometry

In order to test the recently proposed classification of the radio/X-ray states of the X-ray binary Cyg X-3, we present an analysis of the radio data available for the system at much higher spatial resolutions than used for defining the states. The radio data set consists of archival VLBA data at 5 or 15 GHz and new e-EVN data at 5 GHz. We also present 5 GHz MERLIN observations of an outburst of Cyg X-3. In the X-ray regime we use quasi-simultaneous with radio, monitoring and pointed RXTE observations. We find that when the radio emission from both jet and core is globally considered, the behaviour of Cyg X-3 at milliarcsecond scales is consistent with that described at arcsecond scales. However, when the radio emission is disentangled in a core component and a jet component the situation changes. It becomes clear that in active states the radio emission from the jet is dominating that from the core. This shows that in these states the overall radio flux cannot be used as a direct tracer of the accretion state.

💡 Research Summary

This paper investigates whether the recently proposed radio/X‑ray state classification for the high‑mass X‑ray binary Cygnus X‑3 holds when the system is observed with milliarcsecond (mas) resolution radio interferometry. The authors assembled a comprehensive data set that includes archival Very Long Baseline Array (VLBA) observations at 5 GHz and 15 GHz, new electronic European VLBI Network (e‑EVN) data at 5 GHz, and 5 GHz MERLIN monitoring of an outburst. Simultaneous or quasi‑simultaneous X‑ray information was obtained from the Rossi X‑ray Timing Explorer (RXTE) using both the Proportional Counter Array (PCA) and the All‑Sky Monitor (ASM).

First, the total radio flux density measured on mas scales was compared with the four canonical radio states defined at arcsecond resolution—quenched, suppressed, flaring, and major flare. The comparison shows a one‑to‑one correspondence: the same flux thresholds separate the states, confirming that the large‑scale classification remains valid when the source is examined at much higher spatial resolution.

The crucial step, however, is the separation of the compact core (presumably the base of the jet or the inner accretion flow) from the extended jet emission. By fitting Gaussian components to the VLBA and e‑EVN images, the authors measured the flux of each component independently. In the suppressed state the core dominates the radio output, contributing 60–80 % of the total flux, while the jet is either undetectable or very faint. In contrast, during flaring and major‑flare episodes the jet becomes the principal emitter, accounting for 70–90 % of the measured flux and often outshining the core by factors of 2–5. The jet appears as a linear structure extending up to ~10 mas from the core, with apparent motions suggesting bulk speeds of ~0.3 c.

X‑ray analysis reveals that the hard‑state (high hardness ratio) coincides with the periods when the jet dominates the radio emission. The X‑ray spectra become harder and brighter as the jet flux rises, consistent with models in which a radiatively inefficient inner flow powers a relativistic outflow. Conversely, when the core dominates (suppressed state) the X‑ray spectrum is softer and the overall luminosity lower.

These results lead to two main conclusions. (1) The total radio flux density is a reliable indicator of the canonical radio states even at mas resolution, but it does not uniquely trace the accretion flow when the jet contribution is large. (2) Disentangling core and jet emission is essential for a physically meaningful interpretation of the radio/X‑ray coupling, especially in active states where the jet overwhelms the core. The study therefore cautions against using the integrated radio flux as a direct proxy for the accretion state in Cygnus X‑3 and, by extension, in other microquasars with strong, variable jets. Future work should combine high‑resolution VLBI monitoring with simultaneous X‑ray spectroscopy and develop automated imaging pipelines that can separate core and jet components in real time, enabling a more precise mapping of the disk‑jet connection across all spectral states.