Extensive Analysis of gamma-Ray Periodicity in Jetted AGN from the 4FGL Catalog Using Fermi-LAT Observations

TL;DR

This study performs a comprehensive search for gamma-ray periodicity in 1492 jetted AGN using 12 years of Fermi-LAT data and a robust, multi-method pipeline. It replaces bootstrap-based significance with red-noise simulations and incorporates ARFIMA/ARIMA modeling, autocorrelation, and complementary analyses to discriminate genuine periodic signals from red-noise variability. No statistically significant periodic emissions are found after proper corrections, though ~2σ hints appear for a subset of sources that fail global significance tests; a small set of candidates remains for future monitoring. The results underscore the dominant impact of red-noise in blazar variability, advocate for multi-wavelength follow-up and theoretical modeling to interpret potential QPOs, and provide forward-p-looking predictions to guide subsequent observations.

Abstract

The quest to uncover periodic patterns within the $γ$-ray emissions of jetted active galactic nuclei (AGN) has recently emerged as a focal point in astrophysics. One of the primary challenges has been the necessity for prolonged exposures in the $γ$-ray energy band. In our investigation, we leverage 12 years' worth of observations from the \textit{Fermi}-LAT to systematically explore periodicity across 1492 jetted AGN cataloged in 4FGL, representing the largest sample analyzed to date. Our analysis involves a robust pipeline employing nine distinct techniques designed to detect potential periodic emissions within their $γ$ rays. We note that 24 objects with previous hints of periodicity are deliberately excluded in the present work since they were reanalyzed in a dedicated paper using a similar methodology. Using this thorough approach, we do not find any evidence for periodic signals in the 1492 jetted AGN $γ$-ray light curves analyzed here.

Figures (10)

• Figure 1: Left: Light curve of the blazar PKS 0405$-$385. Right: Light curve for the analysis with the information of the upper limits by using the likelihood profile of the flux distribution for extracting the 95% upper limit, represented by the red points
• Figure 2: The new Periodicity-search pipeline summarized activity diagram of Unified Modeling Language. It is organized in four stages. In this initial stage, we identify variable jetted AGN with $<$50% of upper limits present in their LCs. The second stage involves conducting a rapid periodicity analysis of the selected sources. In the subsequent stage, selected sources undergo a more comprehensive analysis using methods that require more computational time. In the final stage, the most significant sources exhibiting evidence of periodicity are selected.
• Figure 3: The distribution of upper limits across the analyzed LCs. The blue bars represent the normalized distribution of upper limits in the initial sample, with normalization based on the bin containing the highest number of sources. This distribution shows a peak concentration at 0% (i.e., detections in all time bins) and a median of 22.4%, indicating that most sources have moderate upper limits. In contrast, the orange bars illustrate the subset of sources exhibiting trends, also normalized based on the bin with the highest number of sources. This subset shows a lower overall percentage of upper limits, with a peak at 0% (detections in all time bins) and a median of 0.6%.
• Figure 4: Left: Examples of PSD fits for PKS 0405$-$385 using both the PL and BPL approaches, utilizing the power-spectral indices included in Table \ref{['tab:slopes']}. The vertical line denotes the peak associated with the period of 3.1 years. Right: The residuals of the PSD fits for PKS 0405$-$385 are compared between the PL and BPL approaches. From the analysis, it is evident that the BPL approach yields smaller residuals compared to the PL approach. This suggests that the BPL model provides a better fit to the observed data for PKS 0405$-$385. The reduced residuals in the BPL approach indicate that this model captures the underlying features of the data more effectively.
• Figure 5: Trial-factor correction applied to the experimental relationship between local-global significance to estimate P in the periodicity analysis. Eq. 1 represents the results obtained by applying Equation \ref{['eq:trial']} for a particular number of independent frequencies, reporting 24. The $P$ is chosen to adjust the test statistics of $\approx2.5\sigma$, thereby improving the correction of the most significant blazars of Table \ref{['tab:candidates_list']}.