Constraining quasar structure using high-frequency microlensing variations and continuum reverberation

Abstract

Gravitational microlensing is a powerful tool for probing the inner structure of strongly lensed quasars and for constraining parameters of the stellar mass function of lens galaxies. This is achieved by analysing microlensing light curves between the multiple images of strongly lensed quasars and accounting for the effects of three main variable components: (1) the continuum flux of the source, (2) microlensing by stars in the lens galaxy, and (3) reverberation of the continuum by the broad line region (BLR). The latter, ignored by state-of-the-art microlensing techniques, can introduce high-frequency variations which we show carry information on the BLR size. We present a new method that includes all these components simultaneously and fits the power spectrum of the data in the Fourier space rather than the observed light curve itself. In this new framework, we analyse COSMOGRAIL light curves of the two-image system QJ 0158-4325 known to display high-frequency variations. Using exclusively the low-frequency part of the power spectrum, our constraint on the accretion disk radius agrees with the thin-disk model estimate and the results of previous work where the microlensing light curves were fit in real space. However, if we also take into account the high-frequency variations, the data favour significantly smaller disk sizes than previous microlensing measurements. In this case, our results are only in agreement with the thin-disk model prediction only if we assume very low mean masses for the microlens population, i.e. ⟨M⟩ = 0.01 M⊙. At the same time, including the differentially microlensed continuum reverberation by the BLR successfully explains the high frequencies without requiring such low-mass microlenses. This allows us to measure, for the first time, the size of the BLR using single-band photometric monitoring; we obtain RBLR = 1.6−0.8+1.5 × 1017 cm, in good agreement with estimates using the BLR size–luminosity relation.

Evidence for a milliparsec-separation Supermassive Binary Black Hole with quasar microlensing

“…We show in Appendix A that this mechanism is not sufficient to reproduce the large amplitude of the extrinsic variations seen in the microlensing light curve. We still use this physically motivated model to generate 5000 simulated light curves from a Damped Random Walk (DRW) and compute the microlensing curve for each of them, using the same differential microlensing model as presented in Paic et al (2022). The simulated light curves have the same sampling and photometric noise as the real data (see Appendix A for the details of this test).…”

“…We adopt a fiducial macro lens model from Morgan et al (2008) for a stellar mass fraction, f M/L = 0.9 (κ = 0.72, γ = 1.03, κ /κ = 0.92 for image B). For the population of microlenses used in Section 4.3, we make similar assumptions to Paic et al (2022), that is, a Salpeter initial mass function (IMF) with mean stellar mass M = 0.3M and a mass ratio of 100 between the heaviest and the lightest microlenses. The mean Einstein radius R E of the microlenses, projected into the source plane, is defined as:…”

“…The fast variations have been tentatively attributed to microlensing by a population of planet mass microlenses (Schild 1996), variations of the accretion disk size over time (Blackburne & Kochanek 2010), inhomogeneities in the accretion disk (Gould & Miralda-Escudé 1997;Schechter et al 2003;Dexter & Agol 2011) or broad absorption clouds shadowing the quasar (Wyithe & Loeb 2002). Other works by Sluse & Tewes (2014) and Paic et al (2022) also propose that a differential magnification of the reverberated flux by the BLR could produce extrinsic variations on the same timescale as the intrinsic variations of the quasar.…”

“…The continuum emission region is smaller than a microlens Einstein radius, and is therefore most prone to microlensing, while the emission from the broad line is much less microlensed. As explained in Paic et al (2022), the spectra of Q J0158−4325 observed by Faure et al (2009) indicate that ∼ 40% of the R−band flux arises from the broad line region (Fig 2022), we have emulated the continuum signal F c using a Damped Random Walk model (Kelly et al 2009;MacLeod et al 2010), and added to it a reverberated BLR signal responding to the intrinsic variations with a lag of τ = 65 days through a tophat transfer function Ψ(t, τ). Following this procedure, the flux of image i (i.e.…”

Evidence for a milliparsec-separation supermassive binary black hole with quasar microlensing

“…The fast variations have been tentatively attributed to microlensing by a population of planetmass microlenses (Schild 1996), variations in the accretion disk size over time (Blackburne & Kochanek 2010), inhomogeneities in the accretion disk (Gould & Miralda-Escudé 1997;Schechter et al 2003;Dexter & Agol 2011), or broad absorption clouds shadowing the quasar (Wyithe & Loeb 2002). Works by Sluse & Tewes (2014) and Paic et al (2022) also propose that a differential magnification of the reverberated flux by the BLR could produce extrinsic variations on the same timescale as the intrinsic variations of the quasar.…”