20 years of monitoring: PKS 2155-304 and PKS 1510-089 in the eyes of Swift and Fermi. I. The case of PKS 2155-304

Abstract

We present a comprehensive 20-year multiwavelength variability study of the blazar PKS 2155-304, one of the most luminous and extensively monitored high-frequency-peaked BL Lac objects in the southern hemisphere. Using Fermi-LAT $γ$-ray data together with Swift-XRT and UVOT observations spanning 2005-2024, we trace the long-term evolution of its flux, interband correlations, and spectral behaviour across the optical, X-ray, and $γ$-ray bands. All flux distributions are compatible with log-normality. Interestingly, the optical domain exhibited a notable baseline change around 2009, but this has no strong influence on the fit of the flux distribution. While interband flux-flux correlations are found, no stable temporal lags emerge. This implies varying correlation patterns between epochs. The X-ray emission displays a robust harder-when-brighter trend, however with epoch-dependent slopes, while the $γ$-ray spectra show only mild flux dependence. The fractional variability increases systematically with energy within a given radiation component. No direct correlation of the year-wise fractional variability with the corresponding average flux could be found. Interestingly, a pronounced X-ray spectral upturn, detected during a low state in 2012, points to an additional radiative component. As the connection from this upturn to the $γ$-ray spectrum is not smooth, it probably is not the onset of the inverse-Compton component, but more likely points either to a hadronic contribution or an additional spatially-separate emission zone. These findings reveal the complexity of variability patterns in PKS 2155-304 and the non-uniform nature of its particle acceleration and emission processes.

Figures (7)

• Figure 1: Time-evolution of PKS 2155$-$304 from 2005 to 2024. (a) HE $\gamma$-ray light curve from Fermi-LAT. The black dotted line marks the average flux of the data shown. (b) HE $\gamma$-ray spectral index with the black dashed line indicating its weighted average. The grey dotted lines marks a value of $2.0$. (c) Same as (a) but for the X-ray data from Swift-XRT. (d) Same as (b) but for the X-ray data from Swift-XRT. (e) B and V band light curves from Swift-UVOT with dashed lines indicating the respective average. (f) Year-wise $F_\text{var}$ values for the $\gamma$-ray, X-ray and B bands. The coloured dashed lines indicate the respective $F_\text{var}$ values considering all data of a given band.
• Figure 2: (Top row) Correlation plots of flux vs flux for X-rays vs HE $\gamma$ rays (left), B band vs HE $\gamma$ rays (middle), and B band vs X-rays (right). (2nd row) HE $\gamma$-ray index vs flux (left), and the same but only for soft spectra (right). The individual points of the latter are given in Tab. \ref{['tab:2155_HE_soft_spectra']}. (Bottom row) X-ray index vs flux (left), curvature vs flux (middle), and curvature vs index (right). In each panel, the Person R coefficient for the correlation is given, as well as the p-value of the probability that a similar or higher R coefficient could be obtained from an uncorrelated system.
• Figure 3: DCCF with a time resolution of $\Delta\tau=5\,$d of PKS 2155$-$304 between the various light curve bands as labelled. The grey bands mark the confidence intervals as indicated.
• Figure 4: X-ray spectral index versus integrated X-ray flux for PKS 2155$-$304. Dashed lines show year-wise (years indicated by colour) weighted linear fits to index-vs-flux data, while the shaded regions are constructed from the errors on the fit parameters. The gaps between dashes indicate the goodness-of-fit with large gaps representing a worse fit or higher rejection significance.
• Figure 5: Flux distributions of the HE $\gamma$-ray (left), X-ray (middle) and B band (right) data using logarithmic binning. The red lines mark Gaussian fits with its mean, standard deviation and the p-value (based on the KS test) of the fit indicated. The golden histogram for the B band is the data taken before 2009. The golden solid line fits the golden histogram (parameters "1"), while the golden dashed line fit the black histogram without the golden data (parameters "2").