The X-ray/UV Connection in NGC 5548: A Rapidly Varying Corona

Recent intensive monitoring campaigns of active galactic nuclei (AGN) have provided simultaneous X-ray, UV, and optical data of unprecedented quality. The observations reveal a strong correlation between the UV and optical variability, but a weaker correlation between the X-ray and UV bands, challenging the standard X-ray reprocessing scenario. We revisit the X-ray/UV connection in NGC 5548 by fitting archival 2014 HST and Swift/XRT light curves assuming X-ray reverberation from a dynamically evolving X-ray corona. Our results show that, as long as the corona height, photon index and power vary over time, X-ray reverberation can explain the observed UV and optical variability within 2% and 5%, respectively (on average). The evolution of the best-fit parameters suggests that fast changes in coronal geometry and energetics on a time scale of days are required to explain the observed variability.

💡 Research Summary

This paper addresses the long‑standing puzzle of why the ultraviolet (UV) and optical variability of the Seyfert galaxy NGC 5548 are strongly correlated with each other yet only weakly correlated with the X‑ray band, a situation that challenges the classic picture of X‑ray reprocessing in a static lamp‑post geometry. Using the 2014 intensive monitoring data from the Hubble Space Telescope (HST) and Swift/XRT, the authors construct a time‑dependent reverberation model in which the X‑ray corona is allowed to change its height above the accretion disc, its intrinsic photon index, and its power relative to the disc on day‑scale intervals.

The UV light curves consist of three far‑UV continuum bands (1158 Å, 1479 Å, and 1746 Å) from the AGN STORM campaign, corrected for Galactic and host‑galaxy extinction. The X‑ray light curve is derived from Swift/XRT spectra fitted with an absorbed power‑law plus warm absorber model; the authors assume that 10 % of the observed 2–10 keV flux originates from disc reflection, leaving 90 % as the primary coronal emission. The 2–10 keV luminosity is then used as the driving term in the reverberation calculation.

The reverberation kernel Ψλ(t) is computed with the relativistic code KYNXiltr, which implements a lamp‑post geometry around a black hole of mass 7 × 10⁷ M⊙, inclination 40°, spin zero, colour‑correction factor 1.7, and an accretion rate of 0.06 Eddington. The disc’s intrinsic (non‑illuminated) flux Fλ,NT(t) is obtained from KYNSED using the same global parameters. The model allows the corona height h (5–50 rg), the intrinsic photon index Γint (1.3–2.1), and the ratio of transferred power to disc power Ltransf/Ldisc (0.3–0.9) to vary independently in each time segment. A uniform grid of 20 values per parameter is pre‑computed, and the appropriate response function is interpolated for each epoch.

Because the Swift sampling is not dense enough to provide a full three‑day X‑ray history for every UV point, the authors restrict the fit to the well‑sampled portion of the campaign (starting at MJD 56713). They simultaneously minimise the residuals of the three UV light curves (91 points each) and the X‑ray 2–10 keV flux. The best‑fit solution reproduces the UV variability with a mean fractional error of ≤2 % and the optical variability (derived from the same model) within ≤5 %. The inferred corona parameters show rapid day‑scale fluctuations: the height h typically ranges from ~10 to ~30 rg, Γint varies between ~1.7 and ~2.0, and Ltransf/Ldisc oscillates around 0.6. These variations are required to match the observed lag amplitudes, which are larger than those predicted by a static disc‑reprocessing model.

The analysis demonstrates several key points. First, a dynamic corona can naturally explain the weak X‑ray/UV cross‑correlation while preserving the strong UV/optical coupling, because changes in h and Γint reshape the disc response function and thus the timing and amplitude of the reprocessed signal. Second, the UV light curves provide an indirect but powerful constraint on the intrinsic photon index, which is difficult to measure directly from the noisy Swift spectra. Third, the model’s success suggests that the geometry and energetics of the corona in NGC 5548 evolve on timescales of days, implying a highly variable magnetic or accretion‑driven process near the black hole.

The authors acknowledge limitations: the sparse X‑ray sampling prevents a fully self‑consistent convolution over the required three‑day pre‑history; the reflection fraction is fixed at 10 % based on mean‑SED fits, which may overlook genuine variability in the reflected component; and the linear parameter grid may miss more complex, non‑linear behaviour. Nevertheless, the work provides a compelling framework for interpreting multi‑wavelength variability in AGN.

In conclusion, by allowing the X‑ray corona to vary dynamically in height, spectral slope, and power, the authors achieve an excellent fit to the UV and optical light curves of NGC 5548, reproducing both the amplitude and the longer-than‑expected lags. This study highlights the importance of rapid coronal evolution in shaping AGN variability and sets the stage for future high‑cadence, broadband monitoring campaigns that can directly track coronal changes on day‑scale timescales.