2024MNRAS.531.3310L

In this work we used a sample of 20 variable active galactic nuclei at 0.3 < z < 0.8 from the Ultra-VISTA survey to test whether their emission in the near-infrared (NIR) is consistent with emission from the dusty torus or the accretion disc, and whether the radius-luminosity relation found for local sources, is also seen when we consider sources from our sample. We carried out a spectral energy distribution analysis and conclude that 14 of the 20 sources have a variable component dominated by the dusty torus, and four have a variable component dominated by the accretion disc and two sources give uncertain results. Correcting for the disc contamination in the observed-frame Ks band, however, allowed us to determine dust lags for 13 sources after performing cross-correlation analysis. Special care was taken to consider lags that were not biased by seasonal gaps. Our determined lags are systematically found below the radius-luminosity relationship determined for local sources. Following previous works, we introduce an empirical relation that corrects lags by the rest-frame wavelength of the band that samples the dust emission, as shorter wavelengths arise from hotter regions of the torus. When the correction is introduced, our results are consistent with previous findings. We finally discuss whether other intrinsic parameters might also contribute to this effect and how these corrections will impact future large optical and NIR surveys.