High-energy variability of the gravitationally lensed blazar PKS 1830-211

TL;DR

The paper uses gravitational lensing as a natural telescope to study the high-energy variability of the lensed blazar PKS 1830-211. By combining Swift/XRT, NuSTAR, and Fermi-LAT data from 2016 and 2019, the authors implement an enhanced autocorrelation analysis on the gamma-ray light curve and an unbinned, time-tagged photon approach on the X-ray data to derive a consistent lens-induced delay of $t_0 = 21.1 \pm 0.1$ days and a magnification factor $a = 0.13 \pm 0.01$, with these observables remaining time-invariant. X-ray spectra are remarkably stable across states, while gamma-ray emission varies dramatically; broad-band SED modeling favors a single ERC component (likely infrared from the dusty torus) dominating the high-energy hump, with rapid electron cooling providing the observed gamma-ray variability. The constancy of the delay across 16 years argues for a persistent emission region in the jet and supports a lensing interpretation, while the SED results imply substantial jet power, proton-dominated energetics, and potential implications for jet composition and acceleration physics in lensed blazars.

Abstract

The production site and process responsible for the highly variable high-energy emission observed from blazar jets are still debated. Gravitational lenses can be used as microscopes to investigate the nature of such sources. We study the broad-band spectral properties and the high-energy variability of the gravitationally-lensed blazar PKS 1830-211, for which radio observations have revealed two images, to put constraints on the jet physics and the existence of a gravitationally-induced time delay and magnification ratio between the images. We utilize Swift/XRT, Nustar, and Fermi-LAT observations from 2016 and 2019 to compare periods of low activity and high activity in PKS 1830-211. Short-timescale variability is elucidated with an unbinned power spectrum analysis of time-tagged NuSTAR photon data. To study the gravitationally-induced time delay in the gamma-ray light curve observed with Fermi-LAT, we improve existing autocorrelation function based methods. Our modified auto-correlation method yields a delay of t_0=21.1 +/- 0.1 d and magnification factor a=0.13 +/- 0.01. These parameters remain time-invariant. In data from 2016 and 2019, the X-ray spectra remain remarkably stable, contrasting with extreme changes in gamma-rays. Both states can be fitted with a single component from Comptonisation of infrared emission from the dusty torus, with different gamma-ray states arising solely from a shift in the break of the electron energy distribution. The detection of a consistent lag throughout the whole light curve suggests that they originate from a persistent location in the jet.

Figures (12)

• Figure 1: Literature values for the delay between the two major images of PKS 1830$-$211. Triangle markers are based on radio data while the rest is based on Fermi-LAT. The result of vanOmmen_1995 at $44 \pm 9$ for 1990 Jun - 1991 Jul and the additional delay of $76^{+25}_{-15}$ found in Neronov_2015 are not shown.
• Figure 2: Light curve of PKS 1830$-$211 in daily binning with an orange and a green vertical line indicating the two available NuSTAR observations. For the Fermi-LAT data (top) we neglect bins with TS $< 0$, bins with uncertainties ten times larger than the flux, and bins affected by solar activity (see \ref{['sec:data_fermi']}). The best piece wise constant step function is shown in the middle panel for a false alarm probability of 5% (p0 = 0.05) and blue vertical dotted lines indicate the division into sub-intervals (1-5). For the Swift/XRT data (bottom) we also omit upper limits
• Figure 3: Spectral fit of NuSTAR FPMA and FPMB data (red and black) and the previous and subsequent (blue and green) Swift/XRT observation from 2016 (top) and 2019 (bottom). The spectrum is fitted with constant$\,\times\,$ztbabs$\,\times\,$tbabs$\,\times\,$ cflux(powerlaw) which is accounting for calibration differences between the instruments, assuming a power law for the source intrinsic spectrum, and absorption in the lensing galaxy as well as the Milky Way. Corresponding fit parameters are shown in Tab. \ref{['tab:joint_2_spec']}.
• Figure 4: Power spectra for one good-time interval (GTI) of the 2019 NuSTAR pointing. The power spectra are from top to bottom: FP A, FP B, the corresponding co-spectrum, and the average of all three.
• Figure 5: Top: Autocorrelation of daily binned, filtered PKS 1830$-$211 Fermi-LAT data (black solid line) and its lower envelope (LE, gray dotted line). Bottom: the lower envelope excess (LEE, black solid line). Visible in both panels, the largest peak at 21 days is presumed to be due to strong lensing. The other peaks are probably signatures of correlated variability of PKS 1830$-$211 itself, see Sec. \ref{['sec:discuss']}. Significance lines were derived with percentiles, see text.