Possible quasi-periodic optical oscillations of ZTF blazars

TL;DR

This study analyzes five years of ZTF optical data (g- and r-bands) for 10 blazars to search for year-to-month scale QPOs within their variability. Using RMS–Flux relations, flux-distribution fitting, and multifaceted QPO tests (WWZ, LSP, ARIMA) with rigorous red-noise significance assessments, the authors find linear RMS–Flux relations and predominantly log-normal flux distributions, indicative of multiplicative nonlinear variability. Four blazars show candidate QPOs with strong local significances, but after trial corrections only J 1503.5+4759 and J 1652.7+4024 remain as tentative QPOs (global p-values ~0.008 and ~0.017–0.024, respectively), necessitating further monitoring. The results point to potential SMBH binary or jet-origin mechanisms for the detected QPOs and demonstrate the value of long-term, multi-band optical time-domain studies for probing the innermost regions of blazars.

Abstract

Based on the Zwicky Transient Facility (ZTF), we selected 10 blazars as our sample sources. Among these, we found four blazars (J 0923.5+4125, J 1221.3+3010, J 1503.5+4759, and J 1652.7+4024) showing possible indications of quasi periodic oscillations (QPOs) modulation. We conducted a detailed analysis of their optical light curves (g- and r-bands) over the past five years using the root mean square (RMS)-Flux relation, flux distribution, and QPO detection methods to investigate their variability characteristics. A linear RMS-Flux relation is present in both bands, and their flux distributions follow a log-normal form. This suggests that optical variability may arise from multiplicative, nonlinear processes across different timescales and flux states. Further QPO analysis using the weighted wavelet Z-transform (WWZ), Lomb-Scargle periodogram (LSP), and autoregressive integrated moving average (ARIMA) methods identified candidate periodic signals in four blazars. J 0923.5+4125 (period $\sim$ 205 days) and J 1221.3+3010 ($\sim$ 630 days) show local significances of $\sim 3 σ$, whereas J 1503.5+4759 ($\sim$ 38.5 days) and J 1652.7+4024 ($\sim$ 48 days) reach $\sim 4 σ$. After accounting for the look-elsewhere effect, the global significances for J 1503.5+4759 in the g- and r-bands are $\sim 2.7 σ$, while for J 1652.7+4024 they are approximately $\sim 2.5 σ$ in both bands. These two blazars warrant further monitoring and investigation.

Figures (8)

• Figure 1: G – R vs. G-band magnitudes for the four blazars, he blue solid line represents the result of linear regression analysis.
• Figure 2: Results of the CCF analysis. The yellow shaded region represents the 1$\sigma$ uncertainty, and the gray dashed line marks the 99% theoretical significance.
• Figure 3: Graphs of 10 data points per bin for the RMS–Flux relation. The red-shaded area denotes the 95% confidence interval. The four blazars exhibit a strong linear RMS–Flux relation in both the g- and r-band.
• Figure 4: Same as Figure \ref{['fig3']}, but the binning is 30 days.
• Figure 5: The best-fit normal (red solid curves) and log-normal (green solid curves) function to the flux distribution (hatched blue histograms) of the blazars. The black dotted curves represent the kernel density estimation (KDE) results.