State-Dependent X-ray Variability in Cygnus X-1: A 12-Year NuSTAR Timing Study of Accretion Flow Geometry

TL;DR

This 12-year NuSTAR timing study of Cygnus X-1 delivers a cohesive, multi-faceted view of accretion geometry evolution across spectral states. By combining energy-resolved light curves, rms–flux scaling, color–based diagnostics, and frequency-domain timing, the work reveals a robust link between hard X-ray bimodality, disk truncation radii, and coronal evolution, with a precise mapping of break frequencies to $R_g$ via a truncated disk model. The discovery of a failed state transition (Obs 30302019006) challenges conventional transition scenarios and suggests wind-fed systems can follow distinct evolutionary pathways. Collectively, the results provide stringent observational constraints on disk–corona coupling and argue for universal scaling of timing properties across mass scales, with immediate implications for state classification in current and future X-ray timing missions.

Abstract

We present a comprehensive timing analysis of the black hole X-ray binary Cygnus X-1 using 26 NuSTAR observations spanning 2012-2024, providing the most detailed characterization to date of its accretion flow variability across spectral states. Our analysis reveals fundamental insights into the physics governing state transitions in stellar-mass black holes. We discover distinct bimodal flux distributions in the 8-79 keV band with well-separated peaks, contrasting with overlapping distributions in the 3-8 keV band. This energy-dependent bimodality establishes hard X-rays as the optimal diagnostic for state classification, directly tracing the geometric transformation between corona-dominated and disk-dominated configurations. Power spectral analysis uncovers state-dependent characteristic frequencies shifting from 0.050 Hz (hard) to 0.074 Hz (intermediate), with featureless red noise in soft states. These frequencies correspond to disk truncation radii evolving from $\sim$5.5 $R_g$ to $\sim$2 $R_g$, providing direct observational evidence for the inward progression of the accretion disk during state transitions. Frequency-dependent time lags evolve systematically from $\sim$50 ms hard lags at 0.1 Hz in hard states to near-zero in soft states, quantifying the collapse of the Comptonizing corona. Linear rms-flux relations persist across all states with parameters that precisely track the relative contributions of thermal versus non-thermal emission components. Most remarkably, we identify a failed state transition (observation 30302019006) exhibiting anticorrelated band behavior, suppressed variability ($F_{var}$ < 1.38\%), and apparent sub-ISCO truncation. This discovery challenges standard transition models and suggests new pathways for accretion flow evolution in wind-fed systems.

Figures (9)

• Figure 1: The long-term evolution of Cygnus X-1's intensity and spectral hardness. Each point represents a 100s time bin, with observation segments separated for visual clarity. The y-axis shows the total intensity in the 3-79 keV band, and the color of each point corresponds to the hardness ratio (HR), defined as the 8-79 keV rate divided by the 3-8 keV rate. Yellow/orange colors indicate a harder spectrum, while purple/blue colors indicate a softer spectrum, clearly showing the source transitioning between hard and soft states.
• Figure 2: Orbital phase coverage of NuSTAR observations separated by spectral state. Gray points show individual 100s time bins phase-folded using the 5.6-day orbital period, while colored lines represent smoothed profiles normalized to the mean count rate (horizontal dashed line). Data are shown over two cycles for clarity. Soft state (top): nearly complete phase coverage. Intermediate state (middle): approximately 50% coverage concentrated at phases 0.0-0.4. Hard state (bottom): comprehensive coverage across most of the orbit.
• Figure 3: Count rate distributions for Cygnus X-1 across various NuSTAR energy bands. (a) Count rate distributions for the 3-8 keV band. (b) Count rate distributions for the 8-79 keV band. (c) Count rate distributions for the 20-79 keV band. Each panel shows a histogram of the light curve count rates for the specified energy band. A clear bimodal distribution is seen in energy bands starting at 8 keV and the bimodality remains constant as the lower end of the energy band increases, motivating our choice of the hard and soft bands.
• Figure 4: Distributions of the count rates for (a) the full band, (b) the soft band, and (c) the hard band. The histograms are shown on a logarithmic rate axis. The dashed line in each panel represents the best-fit model as determined by the BIC.
• Figure 5: Log of observations showing the start and end orbital phases and the corresponding fractional variability and modulation factors. ObsIDs marked with an asterisk (*) cross the phase 0.0 boundary. For observations where the excess variance was negative, we report the 95% confidence upper limit on the fractional variability.