Morphology-aware periodicity searches in Fermi-LAT blazars

Abstract

Standard periodicity searches in active galactic nuclei assume sinusoidal modulation, whereas gamma-ray blazar emission often features asymmetric flares and localized bursts. We investigate how periodicity significance in Fermi-LAT light curves depends on intra-cycle morphology and data-quality cuts. We analyze 101 seven-day binned light curves from the Fermi-LAT Light Curve Repository (LCR) using the Fast Template Periodogram with sinusoidal, Gaussian burst, and empirical asymmetric templates. Significances are calibrated with red-noise simulations including a Poisson noise floor. A Random Forest classifier scores sources using periodogram, phase-coherence, and period-stability features. Unconstrained likelihood fits in a few bins flatten the periodogram and make the simulated null effectively white. After quality cuts, three sources exceed a per-source significance of 3 sigma, versus 0.14 expected by chance. One is the established candidate PG 1553+113, but none reaches 3 sigma after correcting for the 101 light curves searched. Template choice substantially affects significance: for 4FGL J0102.8+5824, it is 0.24 sigma with a sinusoid and 3.85 sigma with a Gaussian burst. Such a gain is expected for about one source in simulated noise, while about 40 percent of sources favor a sinusoid. The significance distribution has a heavier upper tail than the Gaussian null predicts, consistent with flares. We report no new periodicity detection. Non-sinusoidal templates complement sinusoidal searches, but a large significance gain alone does not establish periodicity. The classifier also responds to isolated flares, indicating that the null does not fully capture flare behavior. Our results agree with recent systematic searches. Applying the LCR convergence criterion, supplemented by a consistency check, is important for periodicity searches, especially with seven-day bins.

Substantial revision, results and conclusions changed. Submitted to A&A