Accretion disk reverberation with Hubble Space Telescope observations of NGC 4593: evidence for diffuse continuum lags

TL;DR

This study uses Hubble Space Telescope data to analyze accretion disk reverberation in NGC 4593, revealing larger-than-expected disk sizes and evidence of diffuse continuum lags from the broad-line region, challenging standard models.

Contribution

First high-resolution UV/optical lag spectrum of NGC 4593 showing both disk and broad-line region contributions, highlighting the need to revise standard accretion disk models.

Findings

Lag spectrum follows τ ∝ λ^{4/3} relation.

Disk sizes are about three times larger than predicted.

Diffuse emission from the broad-line region significantly affects lag measurements.

Abstract

The Seyfert 1 galaxy NGC 4593 was monitored spectroscopically with the Hubble Space Telescope as part of a reverberation mapping campaign that also included Swift, Kepler and ground-based photometric monitoring. During 2016 July 12 - August 6, we obtained 26 spectra across a nearly continuous wavelength range of ~1150 - 10,000A. These were combined with Swift data to produce a UV/optical "lag spectrum", which shows the interband lag relative to the Swift UVW2 band as a function of wavelength. The broad shape of the lag spectrum appears to follow the $\tau \propto \lambda^{4/3}$ relation seen previously in photometric interband lag measurements of other active galactic nuclei (AGN). This shape is consistent with the standard thin disk model but the magnitude of the lags implies a disk that is a factor of ~3 larger than predicted, again consistent with what has been previously seen in other AGN. In all cases these large disk sizes, which are also implied by independent gravitational microlensing of higher-mass AGN, cannot be simply reconciled with the standard model. However the most striking feature in this higher resolution lag spectrum is a clear excess around the 3646A Balmer jump. This strongly suggests that diffuse emission from gas in the much larger broad-line region (BLR) must also contribute significantly to the interband lags. While the relative contributions of the disk and BLR cannot be uniquely determined in these initial measurements, it is clear that both will need to be considered in comprehensively modeling and understanding AGN lag spectra.