Revealing an Oscillating and Contracting Compact Corona near the Event Horizon of the Supermassive Black Hole in 1ES 1927+654

TL;DR

This study presents the first observational evidence for an oscillating and contracting compact corona near a supermassive black hole, inferred from a phase-resolved analysis of a unique mHz QPO in 1ES 1927+654. By combining phase-resolved RMS/lag analyses with time-dependent Comptonization modeling and Monte Carlo radiative transfer simulations, the authors show that coronal temperature and/or optical depth oscillations inside a corona of a few $R_g$ reproduce the observed QPO properties, including a distinctive U-shaped lag-energy spectrum and energy-dependent variability. The QPO frequency increases rapidly over ~1 year and then plateaus, with the corona contracting as the frequency rises, consistent with magneto-acoustic resonance in a compact corona near the event horizon. These results bridge AGN QPO phenomenology with accretion-physics analogies from X-ray binaries and have implications for mapping inner accretion geometry, jet formation, and potential multi-messenger signatures in the mHz band. $

Abstract

Dynamic processes in the accretion flow near black holes produce X-ray flux variability, sometimes quasi-periodic. Determining its physical origin is key to mapping accretion geometry but remains unresolved. We perform a novel phase-resolved analysis on a newly discovered quasi-periodic oscillation (QPO) in the active galactic nucleus 1ES 1927+654. For the first time in a supermassive black hole (SMBH), we detect a unique `U'-shaped QPO lag-energy spectrum and observe coronal spectral variability over the QPO phase. We find that the QPO is adequately explained by plasma resonant oscillations within a corona. Modeling of QPO spectral properties and reverberation mapping reveal that the corona is contracting and confined to only a few gravitational radii regions near the SMBH, consistent with theoretical predictions for a decreasing QPO period of near 10 minutes. These results present the first observational evidence for an oscillating and contracting compact corona around an SMBH.

Figures (11)

• Figure 1: Overview of the X-ray QPO detected in 1ES 1927+654. (A) Long-term NICER light curves in energy ranges of 0.3--2 keV (blue circles) and 2--10 keV (red circles) beginning 2 months after the optical outburst 2025Natur.638..370M. The epoch of the QPO detected by XMM-Newton is shaded in gray. The timing of the mHz QPO detections suggests a possible association with the formation of the jet, as inferred from radio observations 2025ApJ...979L...2M2025ApJ...981..125L. (B) Evolution of the QPO frequency over time. Purple stars indicate significant detections of the QPO ($>6\sigma$), while gray circles denote marginal detections ($<6\sigma$). The presented QPO frequency and corresponding $1\sigma$ error bars are obtained from the PSD fitting with an additional Lorentzian. (C) Normalized phase-folded light curves (QPO waveforms) of the $\sim1.67$ mHz QPO (obsID 0915390701) across multiple energy bands, referenced to the 2--10 keV QPO phase. Data points with error bars are unbinned waveforms, while solid lines are best-fit sinusoidal functions. Data are plotted over two QPO cycles by repeating the pattern. Vertical dashed lines indicate the peak phases of the QPO waveforms obtained from sinusoidal fittings.
• Figure 2: RMS- and lag-energy spectra of the X-ray QPO in 1ES 1927+654 and corresponding fitting models. Error bars represent $1\sigma$ confidence intervals. The black dashed lines show the best-fit results from the joint fitting of time-averaged spectra, QPO RMS- and lag-energy spectra with the time-dependent Comptonization model. (A) QPO RMS-energy spectra. Black stars represent values from conventional PSD modeling using a Lorentzian function; colored circles show results from the phase-resolved analysis. (B) QPO lag-energy spectra with the 0.3--0.5 keV energy band as the reference (gray shaded region), defined as zero lag. The uncertainty in the reference band is based on the arithmetic mean of the $1\sigma$ errors in other energy bands. The blue solid lines represent a model with a soft-blackbody component and a hard-logarithmic power-law component, and the blue dotted lines show the corresponding hard-logarithmic component.
• Figure 3: Spectral properties of the QPO in 1ES 1927+654. (A) Comparison of QPO spectra with time-averaged energy spectra. Individual model components are indicated by different line styles and colors. QPO spectra and their corresponding fit models are scaled by a factor of 10 for better comparison. (B) Variability of the spectral index and hardness ratio over the QPO cycle. The hardness ratio is defined as the flux ratio between the 2--10 keV and 0.3--2 keV energy bands. The phase-folded QPO light curve in 2--10 keV is plotted in each panel as a gray dashed line for reference.
• Figure B1: Relationship between PSD FWHM of the extracted QPO mode and the VME $\alpha$ parameter. The horizontal dashed lines represent the FWHM values obtained from the PSD modeling of the mHz QPO with a Lorentzian function. In the VME analysis, the FWHM decreases monotonically with increasing $\alpha$, allowing us to estimate reasonable $\alpha$ values for two observations as 9078 and 6895, respectively.
• Figure B2: Posterior probability distributions for parameters derived from Bayesian modeling to estimate the VME parameter $\alpha$ for two observations. (A) Distributions corresponding to the observation with the 1.67 mHz QPO. (B) Distributions corresponding to the observation with the 2.21 mHz QPO.