Characterizing Short-Timescale Optical Variability in Non-blazar Active Galactic Nucleus PKS~0521$-$36 Using TESS

Abstract

We present a systematic analysis of high-cadence optical light curves of the non-blazar AGN PKS~0521$-$36 obtained with \textit{TESS} across three sectors: Sectors~5 and~6 (Cycle~1, 30~min cadence) and Sector~32 (Cycle~3, 10~min cadence). The source exhibits moderate variability with $F_\mathrm{var} \approx 0.69$--$1.19\%$, consistent with a mildly beamed jet. The power spectral density (PSD) in all sectors is better described by a bending power-law than a simple power law, with high-frequency slopes $α_1 \approx 2.1$--2.9, indicating red-noise dominated variability. Flux distributions require two-component models, with a double log-normal providing the best description, suggesting the presence of two distinct optical flux states associated with quiescent jet emission and episodic flaring activity. A significant QPO at $P = 2.838 \pm 0.078$~d is detected in Sector~5 at $>99.99\%$ confidence in the Lomb--Scargle periodogram, independently confirmed by WWZ ($2.839 \pm 0.110$~d) and supported at the $3σ$ level by DRW analysis. The signal spans $\sim$9 cycles within the 26.1-day baseline in Sector~5 and is absent in Sectors~6 and~32, indicating a transient feature. The PSD bending frequency ($ν_b \approx 0.308$~d$^{-1}$; $\sim$3.2~d) is consistent with the QPO period, suggesting a common origin. We interpret the oscillation as magnetohydrodynamic kink instabilities in the relativistic jet, consistent with the observed helicoidal structure. A moderate Doppler factor ($δ\approx 5$--10) naturally explains the day-scale periodicity. Together with previously reported $γ$-ray QPOs on longer timescales, this suggests a hierarchical variability structure, and, to the best of our knowledge, provides the first evidence for an optical QPO in a non-blazar AGN with a directly imaged helical jet.

Figures (6)

• Figure 1: Systematics-corrected TESS optical light curves of PKS 0521$-$36 obtained using the Quaver pipeline for Sectors 5, 6, and 32. The red points correspond to the full hybrid reduction adopted for the timing analysis in this work, while the blue points show the simple PCA (SPO) light curve plotted for comparison. The time axis is given in BTJD ($\mathrm{BTJD} = \mathrm{BJD} - 2457000$), and the flux is in instrumental units of $\mathrm{e^{-}\,s^{-1}}$. The prominent gap visible near the midpoint of each sector arises from routine spacecraft data downlink operations. Both reductions recover consistent variability patterns, supporting the reliability of the full hybrid light curves used in the subsequent analysis.
• Figure 2: Generalized Lomb--Scargle periodograms (LSP; blue solid curves) of the Quaver full hybrid TESS light curves of PKS 0521$-$36 for Sectors 5, 6, and 32. The dashed curves show the Monte Carlo significance levels derived from 20 000 simulated light curves following 2013MNRAS.433..907E, with the specific confidence levels indicated in each panel legend. The spectral window function is shown as a red solid curve in Sectors 5 and 32, and as a purple solid curve in Sector 6, which traces the power introduced by the uneven sampling pattern of the TESS observations. Filled purple circles mark peaks exceeding the 99.99% confidence level where present. A significant QPO candidate is detected in Sector 5, while no statistically significant periodic feature is identified in Sectors 6 or 32. The results are discussed in Section \ref{['sec:results']}.
• Figure 3: Weighted wavelet Z-transform (WWZ) analysis of the 4 hr binned TESS full hybrid light curve of PKS 0521$-$36 for Sector 5. Left panel: Two-dimensional WWZ time--frequency map showing the evolution of wavelet power as a function of time (BTJD) and frequency (d$^{-1}$). A persistent, localized band of enhanced power is visible near $f \approx 0.35$ d$^{-1}$, extending across the majority of the observing baseline, indicative of a quasi-periodic modulation sustained throughout the sector. Right panel: Time-averaged WWZ power spectrum with the 99.7% ($\sim$ 3$\sigma$) confidence level derived from $15\,000$ Monte Carlo simulations following 2013MNRAS.433..907E, shown as the dashed magenta curve.
• Figure 4: DRW-based significance assessment of the generalized Lomb--Scargle periodogram for the Sector 5 light curve of PKS 0521$-$36. The black solid curve shows the observed LSP periodogram. The red and blue dashed curves represent the $3\sigma$ and $2\sigma$ confidence levels, respectively, derived from an ensemble of 12 000 mock light curves generated using the maximum-likelihood DRW parameters inferred from the data via the EzTao package. The blue shaded region indicates the area between the $2\sigma$ and $3\sigma$ thresholds. The green solid curve shows the spectral window function, which traces the power introduced by the uneven sampling pattern of the TESS observations. The vertical cyan line on the left marks the minimum frequency corresponding to the total temporal baseline of the sector ($f_{\rm min} = 1/T$), while the cyan shaded region on the right delineates the frequencies approaching the Nyquist limit ($f_{\rm max} = 1/2\Delta T$), beyond which the periodogram is unreliable. The inset shows the full frequency range on a logarithmic scale for both the periodogram (black) and the spectral window (green). A prominent peak near $f \approx 0.34$ d$^{-1}$ exceeds the $3\sigma$ confidence threshold, providing independent support for the QPO candidate identified in the LSP analysis of Sector 5. The corresponding period and significance are discussed in Section \ref{['sec:results']}.
• Figure 5: Power spectral density analysis of the TESS full hybrid light curves of PKS 0521$-$36 for Sectors 5, 6, and 32. In each panel, the cyan points with error bars show the logarithmically rebinned periodogram, and the light cyan curve shows the raw Lomb--Scargle periodogram. The green dashed, blue solid, and red solid curves correspond to the best-fitting $M_1$ (simple power law), $M_2$ (bending power law-A), and $M_3$ (bending power law-B) models, respectively. The preferred model in each sector, selected on the basis of the minimum BIC value, is listed in the subcaption. Best-fitting parameters for all models are given in Table \ref{['tab:psd']}.