Transient quasi-periodic oscillations in the gamma-ray light curves of bright blazars

Abstract

Transient quasi-periodic oscillations (QPOs) are extremely interesting observational phenomena. However, the precise physical mechanisms leading to their generation are still hotly debated. We performed a systematic search for transient QPO signals using Weighted Wavelet Z-transforms on the gamma-ray light curves of 134 bright blazars with peak flux exceeding $1\times 10^{-6}$ ph cm${}^{-2}$ s${}^{-1}$ as monitored by Fermi-LAT. Artificial light curves were generated from the power spectral density and probability distribution functions of the original light curves to assess the significance level of transient QPO. We discuss several physical mechanisms that produce transient QPOs, with the helical jet model providing the best explanation. This study identified four new transient QPO events. Interestingly, repetitive transient QPOs are observed in PKS 0537-441, and nested transient QPOs are detected in PKS 1424-41. Additionally, we find that transient QPOs tend to occur in the flare state of the blazar. Finally, we estimate the incidence of transient QPO events to be only about 3%.

1 Introduction

There are usually one or more supermassive black holes in the center of active galactic nuclei (the mass is millions to billions of times the mass of the sun), and the supermassive black hole accretes surrounding matter into bright radiation. There are many subclasses of active galactic nuclei, the type in which the jet is directed toward the observer is called a blazar (Urry & Padovani, 1995; Blandford, Meier, & Readhead, 2019). Blazars are also among the brightest objects in the universe, with radiation spanning the entire electromagnetic spectrum and significant long-term light variability. The energy spectrum of blazars shows a double-peak structure in the electromagnetic band, the low-energy peak is dominated by optical to X-ray emission, and the high-energy peak is mainly dominated by gamma-ray emission. Current studies have shown that low-energy photons mainly come from synchrotron radiation, while high-energy photons mainly come from inverse Compton scattering, also originated from the relativistic proton-related hadron radiation (Mannheim, 1993; Aharonian, 2000; Atoyan & Dermer, 2003). According to the intensity of spectral emission lines, blazars are classified as BL Lacertae objects (BL Lacs) and Flat-spectrum Radio Quasars (FSRQs) (Giommi et al., 2012). According to the synchrotron peaked frequency, blazars are also divided into four categories; Low Synchrotron Peaked, Intermediate Synchrotron Peaked, High Synchrotron Peaked, and Extremely High Synchrotron Peaked blazars (Abdo et al., 2010; Akiyama et al., 2016).

The light variability of blazars is very violent and usually exhibits irregular noise-like behavior. In the entire blazar family, there are only a few sources whose light curves change regularly, among which a very special type is the quasi-periodic oscillation (QPO). Periodic light variability is of great significance for searching the structure of binary black holes or studying the special dynamical behavior in the inner region of black holes. OJ 287 is the first blazar reported to have a QPO with a period time scale of 12 years in the optical V-band. This long-timescale QPO signal can be well explained in the binary black hole scenario (Sillanpaa et al., 1988). Subsequently, Kidger, Takalo, & Sillanpaa (1992) analyzed the optical B-band light curve of OJ 287 for about 100 years and found a QPO with a time scale of ($11.6\pm 0.5$ years). Since then, some interesting QPO candidates have been reported in the optical, radio, X-ray, and gamma-ray bands: AO 0235+16, RE J1034+396, PKS 2155-304, 2XMM J123103.2+110648, J1359+4011, PG 1553+113, 1H 0707-495, PKS 0537-441, Mrk 766, PMN J0948+0022, PKS 0219-164, PKS 2155-304, PKS 0426-380, PKS 0301-243, OJ 287, PKS J2134-0153 , PKS J0805-0111, PKS 0521-36, and PKS 1510-089 (Raiteri et al., 2001; Gierliński et al., 2008; Lachowicz et al., 2009; Lin et al., 2013; King et al., 2013; Ackermann et al., 2015; Zhang et al., 2018; Sandrinelli, Covino, & Treves, 2016; Zhang et al., 2017b, a; Bhatta, 2017; Zhang et al., 2017, 2017c, 2017d; Kushwaha et al., 2020; Ren et al., 2021a, b; Zhang et al., 2021; Li et al., 2023). These QPO candidates are mainly distributed on annual timescales, and most of them only appear in a single frequency band, with only a few sources (for example, PG 1553+113) appearing in two or more frequency bands. Long-time-scale QPOs have been more widely discussed in the supermassive binary black hole scenario, and these sources are also considered as possible candidates for supermassive binary black holes (Valtonen et al., 2008; Villforth et al., 2010). In addition, persistent jet precession (Romero et al., 2000; Rieger, 2004) and Lense-Thirring precession of accretion disks (Stella & Vietri, 1998; Liska et al., 2018; You, Bursa, & Życki, 2018) are also used in some scenarios to explain long-term QPO.

Zhou et al. (2018) first reported a 34.5-day QPO in the gamma-ray light curve of the blazar PKS 2247-131, which is also the first monthly-time-scale transient QPO reported in an active galactic nucleus. This QPO appeared in the PKS 2247-131 gamma-ray flare state with six cycles and a time span of about 210 days. Its physical mechanism is attributed to a helical jet scenario with an enhanced radiation zone. Subsequently, many transient QPOs have also been reported one after another. The transient QPO of the high-redshift FSRQ B2 1520+31 in the gamma-ray band is 71 days, with seven cycles and a total duration of about 500 days (Gupta et al., 2019). The FSRQ CTA 102 exhibits a transient QPO (quasi-periodic oscillation) event lasting approximately 7.6 days, simultaneously observed in both the gamma-ray and optical bands. This QPO event persists for about 60 days and undergoes eight cycles (Sarkar et al., 2020). There is a 47-day transient QPO in the gamma-ray variability of 3C 454 with nine cycles lasting 450 days, and a similar lower-confidence signal in the optical band (Sarkar et al., 2021). The TeV blazar PKS 1510-089 exhibits a 3.6-day transient QPO with five cycles, lasting approximately 17 days. Additionally, it shows a 92-day transient QPO with seven cycles, lasting for 650 days (Roy et al., 2022). Chen et al. (2022) searche for monthly timescale QPO signals among 1525 highly variable sources given in the Fermi-LAT lightcurve library. They found a 31.3$\pm$1.8-day transient QPO in the gamma-ray light curve of TeV blazar S5 0716+714, with seven cycles lasting about 220 days. Ren, Cerruti, & Sahakyan (2023) analyzed the light curves of the 35 brightest Fermi-LAT AGNs, including data from the start of the Fermi mission (August 2008) to April 2021, and energies from 100 MeV to 300 GeV. Two time-bin, 7 days and 30 days, were investigated. A search for quasi-periodic features was then performed using continuous wavelet transform, identifying 24 transient QPO candidates. Among them, there are 16 transient QPOs whose cycle times are greater than or equal to five. As shown in Table 1, we list the information of the more credible transient QPOs that have been reported so far (including source name, 4FGL name, type, QPO time scale, QPO cycle number, and literature sources). The physical mechanism of these transient QPOs is could be attributed to hotspots rotating on the innermost stable circular orbit of the supermassive black hole (Zhang & Bao, 1991; Gupta, Srivastava, & Wiita, 2009), magnetic reconnection events of nearly equidistant magnetic islands within the jet (Huang et al., 2013; Shukla et al., 2018), and helical motion of the enhanced emission region (or blob) inside the jet (Mohan & Mangalam, 2015; Sobacchi, Sormani, & Stamerra, 2017; Zhou et al., 2018).

The paper aims to systematically search and study for transient QPO events in the gamma-ray band for a flux-limited bright blazar sample. The frequency (period) we are interested in ranges from 0.0025-0.05 Hz (20-400 days). It is well known that periodic signals often hide within noise, making them difficult to detect and prone to false alarms. The confidence in their presence is also often overestimated. Therefore, in this paper, we use more stringent criteria to screen and confirm transient QPOs. This work reports four new transient QPOs, also confirms four transient QPOs reported by previous works, and reevaluates the confidence level of signal detection. Furthermore, based on the search results of the flux-limited sample, we provide the first estimate of the incidence of transient QPO events. The possible physical origins of these transient QPO events are discussed. The paper is organized as follows. Section 2 describes the sample sources and the Fermi data processing procedure. Section 3 describes the search process and methodology for transient QPOs. Section 4 presents the analysis results of all transient QPOs. Section 5 is a discussion, and Section 6 is a summary.

Table 1: Summary of the transient QPOs of individual blazars as reported in the literature.

Source name	4FGL name	Source class	QPO time scale (days)	QPO cycle number	Reference
PKS 2247-131	4FGL J2250.0-1250	BCU	34.5$\pm$1.5	6	a, b
B2 1520+31	4FGL J1522.1+3144	FSRQ	179$\pm$42	6	b
			71$\pm$15	14	c, b
			39$\pm$11	17	b
PKS 1510-089	4FGL J1512.8-0906	FSRQ	3.6$\pm$0.07	5	d
			92$\pm$1.2	7	d
3C 454.3	4FGL J2253.9+1609	FSRQ	$47^{+0.97}_{-0.51}$	9	e
3C 279	4FGL J1256.1-0547	FSRQ	39$\pm$1	-	f
			24$\pm$1	-	f
			101$\pm$27	6	b
4C +01.02	4FGL J0108.6+0134	FSRQ	122$\pm$26	5	b
PKS 0402-362	4FGL J0403.9-3605	FSRQ	122$\pm$42	5	b
PKS 0426-380	4FGL J0428.6-3756	BL Lac	85$\pm$26	8	b
PKS 0447-439	4FGL J0449.4-4350	BL Lac	111$\pm$42	7	b
S5 0716+714	4FGL J0721.9+7120	BL Lac	324$\pm$77	5	b
1H 1013+498	4FGL J1015.0+4926	BL Lac	52$\pm$12	12	b
4C +21.35	4FGL J1224.9+2122	FSRQ	66$\pm$17	6	b
PKS 1424-41	4FGL J1427.9-4206	FSRQ	90$\pm$22	5	b
Mrk 501	4FGL J1653.8+3945	BL Lac	326$\pm$76	7	b
CTA 102	4FGL J2232.6+1143	FSRQ	178$\pm$40	5	b
			$7.6^{+0.36}_{-0.25}$	8	g

Notes: Column (1) is the name of the object; Column (2) is the name of Fermi 4FGL; Column (3) is the type of source (BL Lac or FSRQ); Column (4) is the QPO time scale and error; Column (5) is the QPO cycle number; Column (6) is the reference: (a) Zhou et al. (2018), (b) Ren et al. (2023), (c) Gupta et al. (2019), (d) Roy et al. (2022), (e) Sarkar et al. (2021), (f) Sandrinelli et al. (2016), (g) Sarkar et al. (2020).

2 The sample of Fermi-LAT light curves

2.1 Source Selection

Fermi-LAT monitors a large number of extragalactic bright sources for a long time and adds them to the monitoring source list when the source flux exceeds the monitoring flux threshold of $1\times 10^{-6}$ ph cm ${}^{-2}$ s ${}^{-1}$ (Monitored Source List). As the mission progressed, the list continued to grow, and until now there are 197 sources. This catalog contains 173 Blazars of interest, including 19 Blazar Candidate of Uncertain type (BCU), 32 BL Lacs, and 122 FSRQs. To ensure the continuity and authenticity of the light curve, we add two conditions to further screen the samples: (1) Except for Fermi-LAT observations, these sources must be found by other telescopes to correspond to other wavelength bands. (2) The light curve for each source should have a sufficient number of data points. If more than 60% of the data points are missing from the light curve, that source will be excluded from the sample. Based on the above two criteria, we finally identified 134 blazars as the research objects of this paper, which are given in Table 3, consisting of 7 BCUs, 31 BL Lacs, and 95 FSRQs. For each source, the source name, 4FGL name, source type, coordinates, and redshift in J2000 are listed from left to right (the last three quantities available in the 4FGL catalog)

2.2 Fermi-LAT Data Analysis

The Fermi satellite was launched on June 11, 2008, and operates at an altitude of 565 km. The Fermi satellite carries two instruments, the Large Area Telescope (LAT) pair-conversion detector and the Gamma-ray Burst Monitor (GBM, Atwood et al. (2009)). Fermi-LAT is a high-energy gamma-ray telescope with an observation energy range from $\sim$20 MeV to 500 GeV, with a large field of view ($>$2 sr about 20% of the all-sky). Since routine scientific observations began on August 4, 2008, the entire sky has been scanned every 3 hours. It provides an angular resolution of 5${}^{\circ}$ per event below 100 MeV, downscaling to 2${}^{\circ}$ at 300 GeV. The all-sky monitoring capabilities of the LAT provide us with nearly continuous observations of gamma-ray sources on different time scales.

We downloaded Fermi-LAT data for the Pass8 source class during MET 239557417-647393397 ($\sim$ 12.9 years). Based on the Fermi Science Toolkit $v11r05p3$ analysis was performed using standard analysis programs provided by the fermi-LAT collaboration. Select events belonging to the SOURCE class (evclass = 128, evtype = 3) with energies ranging from 100 MeV to 300 GeV, and exclude photons with zenith angles greater than $90^{\circ}$ (to reduce diffuse reflection from the Earth’s atmosphere photon effects). We use the standard filter (DATA_QUAL$>$0) && (LAT_CONFIG$==$1) to select a good time interval (GTI) to filter the data, and set the range of 15${}^{\circ}$ centered on the point source as the region of interest (ROI). For the diffuse background model file, we choose the Galactic diffuse emission model $gll\_iem\_v07.fits$ (the model includes the sun, moon, Fermi bubble, and gamma-ray radiation generated by cosmic ray propagation in the Milky Way) and the isotropic model for point source $iso\_P8R3\_SOURCE\_V3\_v1.txt$ (the model used to describe the extragalactic gamma diffuse radiation background provided by the extragalactic sub detection source and some cosmic ray background that cannot be completely deducted by the instrument itself). Then, we use the script $make4FGLxml.py$ to create the source model XML file, which includes the location of the source and the best prediction of the spectral form. Perform a maximum likelihood analysis on the XML spectrum file using the $gtlike$ tool and the instrument response function (IRF) $P8R3\_SOURCE\_V3$ to obtain the spectrum of the source. In the 4FGL catalog, the final source spectrum is modeled using a log-parabola (Abdollahi et al., 2020). We use the $gttsmap$ tool to perform a statistical test (TS) on the area around the target source, add a test point source model for each location, and use the formula TS=-2$(lnL0-lnL)$ to calculate the significant (L0 represents the maximum likelihood value of background model fitting, and represents the maximum likelihood value of fitting by adding test point sources in the background model). Through the above analysis and data reduction, we generated the flux data of the source. Subsequently, the light curves for each source were plotted, excluding the few data points with TS less than 9.

3 Transient quasi-periodic oscillations identification

Since the first case of transient QPO in AGN was reported in the gamma-ray emission of PKS 2247-131 (Zhou et al., 2018), the study of transient QPO has attracted extensive attention and discussion. The identification of transient QPOs is more complex than that of long-standing QPOs. This paper utilized the WWZ technique, known for its sensitivity to transient QPO, to detect QPO signals. Subsequently, we verified and examined the signals by eye to account for possible false positives and harmonics. Confidence levels were assessed by Monte Carlo simulations. Here is a brief introduction to the specific analysis process, as shown in Figure 1.

(1) Perform WWZ analysis on the global light curves of all samples, and determine the possible transient QPO on the wavelet power map.

(2) Determine the light curve of the transient QPO given by the WWZ method with the naked eye, and eliminate false positives.

(3) We independently analyze the light curves where transient QPOs occur using the WWZ and LSP methods, respectively. The artificial light curve is then generated from the power spectral density (PSD) and probability distribution function (PDF) of the original light curve to assess the confidence level of the periodic component.

Figure 1: Flowchart for Searching and Evaluating Transient QPOs.

3.1 WWZ global search

Compared with Fourier analysis, Wavelet Transform (WT) is more impressive in identifying transient QPO signals (Foster, 1995). WT evaluates the signal over a certain time range, searching for frequency components in each time segment separately to identify local properties in the time and frequency domains. Also, the WT time range width can be scaled to generate multiple resolutions in time and frequency by traversing different time scales. Therefore, it is sensitive to the local attributes and global attributes of time series data and can detect possible local transient periodic fluctuations and parameter changes in the signal. The continuous wavelet transform of a time series $x(t)$ is defined as

$$W(\omega,\tau;x(t))=\omega^{1/2}\int x(t)f^{\ast}(\omega(t-\tau))dt=\omega^{-1/2}$$ (1)

$$\int x(\omega^{-1}z+\tau)f^{\ast}(z)dz,$$

where $f^{\ast}$ is the complex conjugate of $f$, the function $f(z)$ is the mother wavelet, $\omega$ is the scaling factor and $\tau$ is the time shift. For a discrete dataset of observations x{$t_{\alpha}$, $\alpha$=1, 2, …, N}, the wavelet transform can be written as the discrete wavelet transform (DWT)

$$W(\omega,\tau;x(t))=\sqrt{\omega}\sum\limits^{N}_{\alpha=1}x(t_{a})f^{\ast}(\omega(t_{a}-\tau)),$$ (2)

introduce the mother wavelet function the Morlet wavelet $f(z)$

$$f(z)=e^{iz-cz^{2}}=e^{i\omega(t-\tau)-c\omega^{2}(t-\tau)^{2}},$$ (3)

replace the number of data points N with the effective number

$$N_{eff}=\frac{(\sum\omega_{\alpha})^{2}}{\sum\omega_{\alpha}^{2}}=\frac{[\sum e^{-c\omega^{2}(t-\tau)^{2}}]^{2}}{\sum e^{-2c\omega^{2}(t-\tau)^{2}}}.$$ (4)

Then Foster defines the Weighted wavelet transform (WWT),

$$WWT=\frac{(N_{eff}-1)V_{y}}{2V_{x}}$$ (5)

$$V_{x}=\frac{\sum\limits_{\alpha}\omega_{\alpha}x^{2}t_{\alpha}}{\sum\limits_{\beta}\omega_{\beta}}-\left[\frac{\sum\limits_{\alpha}\omega_{\alpha}xt_{\alpha}}{\sum\limits_{\beta}\omega_{\beta}}\right]^{2}$$

$$V_{y}=\frac{\sum\limits_{\alpha}\omega_{\alpha}y^{2}t_{\alpha}}{\sum\limits_{\beta}\omega_{\beta}}-\left[\frac{\sum\limits_{\alpha}\omega_{\alpha}yt_{\alpha}}{\sum\limits_{\beta}\omega_{\beta}}\right]^{2},$$

where $V_{x}$ and $V_{y}$ is the weighted variation of the data and model function, respectively (Foster, 1996). For sine or cosine signals, WWT tends to peak at lower frequencies. At lower frequencies, we have a wider "window" and thus effectively sample more data points (with a larger effective number $N_{eff}$), which makes the WWT increase as $\omega$ decreases and the data fit worse. The Z-statistic is a test statistic that is less sensitive to the amount of valid data, which will give a better estimate of the frequency of important peaks. Also known as Weighted Wavelet Z-transform (Foster, 1996).

$$Z=\frac{(N_{eff}-3)V_{y}}{2(V_{x}-V_{y})},$$ (6)

We implement the WWZ algorithm using the Python code provided by Aydin (2017) (the author declares that the code is derived from the Fortran code provided by Templeton (2004)). In this work, we perform a WWZ global analysis of all light curves in 134 samples. To more accurately search for potential transient QPO signals in long-term light curves, we determined that the frequency range of interest for WWZ analysis was 0.0025-0.05 $d^{-1}$ (20-400 days). In the contour power map given by the WWZ method, we identified 38 high-power transient QPOs and their appearance time ranges (MJD) in the 35 light curves. To rule out false positive signals, we visually traversed the local light curves of these 38 transient QPOs to confirm the authenticity of the signals. We identified three screening conditions: (1) The number of transient QPO cycles is greater than or equal to five (Vaughan et al., 2016). (2) The number of missing data points in a single cycle cannot exceed 60%. (3) The transient QPO signal can be recognized by the naked eyes. Only when the light curve of a transient QPO satisfies the above three conditions, we can pay attention to it and add it to the transient QPO catalog (Table 2) for further analysis. Eight transient QPO signals were preliminarily identified in the six sources (4C +01.02, PKS 0336-01, PKS 0402-362, PKS 0537-441, PKS 1424-41, and PKS 1510-089). It is worth noting that there are two transient QPO signals in the light curves of PKS 0537-441 and PKS 1424-41.

Quasi-periodic oscillations in blazar light curves are known to be very rare and thus difficult to detect and identify. Compared with previous work, we have given a more stringent screening condition, and there is no doubt that the results we present are very interesting. Next, we analyze and discuss these eight transient QPOs.

3.2 Transient QPO determination

For the eight transient QPOs signals identified above, we extracted the light curve during QPO from the whole light curve according to the QPO time range given by wavelet analysis for local research. For each light curve, the WWZ frequency range of interest is reset according to the period of each transient QPO, and finally, the power contour profile is drawn according to the power at each period position given by the WWZ method. Then, we introduce the LSP method to independently test the results given by the WWZ method (Lomb, 1976; Scargle, 1982; VanderPlas, 2018). The LSP method is a well-known algorithm for detecting and characterizing the periodicity of time series and is widely used in time-domain astronomy research. This method can efficiently calculate the Fourier power spectrum of non-uniform sampling data and obtain possible periodic oscillation components. The following is a brief introduction to this method, assuming a data set $g(t)$ consisting of N equally spaced data points, according to the power spectrum, its classic periodogram can be defined as

$$P_{s}(f)=\frac{1}{N}\left|\sum^{N}_{n=1}g_{n}e^{-2\pi i\beta_{n}}\right|^{2}.$$ (7)

When the classical periodogram is applied to uniformly sampled Gaussian white noise, the periodogram presents a chi-square distribution, and the periodogram can be used to distinguish periodic components from the non-periodic components. The uniformly sampled classical periodogram can be generalized to non-uniformly sampled datasets by a simple change

$$P(f)=\frac{1}{N}\left[\left(\sum^{N}_{n=1}g_{n}cos(2\pi ft_{n})\right)^{2}+\left(\sum^{N}_{n=1}g_{n}sin(2\pi ft_{n})\right)^{2}\right].$$ (8)

The statistical characteristics of this form of periodogram are more complex, and to a certain extent, it can identify potential periodic signals in non-uniform sampling data, but its periodogram cannot clearly distinguish periodic components from non-periodic components. Scargle addresses this problem by considering a generalized form of the periodogram (Scargle, 1982)

$$P(f)=\frac{A^{2}}{2}\left(\sum_{n}g_{n}cos(2\pi f(t_{n}-\tau))\right)^{2}+$$ (9)

$$\frac{B^{2}}{2}\left(\sum_{n}g_{n}sin(2\pi f(t_{n}-\tau))\right)^{2},$$

where $A$, $B$, and $\tau$ are arbitrary functions of frequency $f$ and observation time ${t_{i}}$. The Scargle-corrected periodogram fits the sinusoidal function model to the data of each frequency and constructs the periodogram according to the goodness of fit. Lomb also discusses this model in some depth (Lomb, 1976), so it ends up being called Lomb-Scargle periodogram, whose generalized form is

$$P_{LS}(f)=\frac{1}{2}\left(\frac{[\sum_{n}g_{n}cos(2\pi f(t_{n}-\tau))]^{2}}{\sum_{n}cos^{2}(2\pi f(t_{n}-\tau))}\right)$$ (10)

$$+\frac{1}{2}\left(\frac{[\sum_{n}g_{n}sin(2\pi f(t_{n}-\tau))]^{2}}{\sum_{n}sin^{2}(2\pi f(t_{n}-\tau))}\right),$$

where the time phase correction $\tau$ is

$$\tau=\frac{1}{4\pi f}tan^{-1}\frac{\sum_{n}sin(4\pi ft_{n})}{\sum_{n}cos(4\pi ft_{n})}.$$ (11)

Similar to the WWZ analysis, we analyzed the local light curves of the eight transient QPOs identified above using the LSP method to identify potential periodic components in the light curves and finally draw LSP power contours. Next, we performed Monte Carlo simulations based on red noise to determine the confidence level of the transient QPO determined by the WWZ and LSP methods, and to determine the importance of the periodic component in the light curve. We simulate the artificial light curve according to the method given by Emmanoulopoulos, McHardy, & Papadakis (2013), which uses a smooth bending power-law model plus Poisson noise to model the PSD of the original light curve, and at the same time, uses gamma and lognormal distributions to model PDF modeling of the original light curve. Then simulate the light curve of the same PSD and PDF distribution, here we use the Python program given by Connolly (2016) to implement the simulation algorithm. Finally, we separately simulated 10000 artificial light curves with the same PSD and PDF distributions as the original light curves for the local light curves of each transient QPO. In the confidence evaluation stage, first, calculate the WWZ power and LSP power of each artificial dimming curve, then calculate 99%, 99.7%, and 99.994% confidence power, and finally draw 99%, 99.7%, and 99.994% confidence contour lines in the frequency range of interest for each light curve in the WWZ power map and LSP power map, respectively.

4 Results

In this section, we present the analysis results of the WWZ and LSP methods for all transient QPOs. Figure 2, Figure 3, Figure 4, Figure 5, Figure 6, and Figure 7 respectively give 4C+01.02, PKS 0336-01, PKS 0402-362, PKS 0537-441, TXS 1219+285, PKS 1424-41 and PKS 1510-089 analytical results of transient QPO in gamma-ray. Except for QPOs of 253 days in 4C +01.02 and 103 days in PKS 0402-362, the confidence level of the other 6 QPOs exceeds 3$\sigma$. And 91.8 days in PKS 1510-089 and 55 days and 54.7 days in PKS 0537-441, the confidence levels of these three transient QPOs are all greater than or equal to 4$\sigma$. The time scale of QPOs is all within one year, the longest being the 341-day QPO in PKS 1424-41, and the shortest being the 54.7-day QPO in PKS 0537-441. Below we introduce the transient QPO in each source in detail.

Table 2: List of transient QPO candidates and their associated confidence levels.

Source name	Source class	Time span (MJD)	WWZ period (days)	LSP period (days)	Significance	QPO cycle number
4C +01.02	FSRQ	56422-58000	252.2 $\pm$ 20.9	253.2 $\pm$ 20.0	$\sim$ 2.7 $\sigma$	6
PKS 0336-01	FSRQ	57020-57600	94.5 $\pm$ 7.2	94.6 $\pm$ 6.8	$\sim$ 3.4 $\sigma$	6
PKS 0402-362	FSRQ	56817-57363	103.5 $\pm$ 7.9	103.9 $\pm$ 6.7	$\sim$ 2.6 $\sigma$	5
PKS 0537-441	BL Lac	56803-57183	55.1 $\pm$ 3.1	55.0 $\pm$ 3.3	$\sim$ 4.1 $\sigma$	7
		57636-58036	54.6 $\pm$ 2.9	54.7 $\pm$ 3.3	$\sim$ 4.0 $\sigma$	7
PKS 1424-41	FSRQ	56998-57331	57.2 $\pm$ 4.1	57.1 $\pm$ 4.5	$\sim$ 3.2 $\sigma$	6
		56611-58644	336 $\pm$ 37.3	341 $\pm$ 25.8	$\sim$ 3.7 $\sigma$	6
PKS 1510-089	FSRQ	58197-58955	91.8 $\pm$ 5.3	92.3 $\pm$ 5.2	$\sim$ 4.3 $\sigma$	8

Notes: Column (1) is the name of the object; Column (2) is the type of source; Column (3) is the time span of the transient QPO; Column (4) shows the QPO and error given by the WWZ method; Column (5) shows the QPO and error given by the LSP method; Column (6) is the confidence level of the peak signal (take the maximum significance given by the previous two methods); and Column (7) is the QPO cycle number.

4.1 4C +01.02

4C +01.02 is a high-redshift FSRQ (z=2.099), which has GeV radiation observed by Fermi-LAT in 2013. Multi-wavelength studies have shown that the light variability of this source is very active, and the maximum flux of gamma-ray even reaches 61 times of the basic flux. The one zone leptonic model can well reproduce the broadband energy spectrum distribution, but its different flare states need to consider the contribution of BLR and IR torus seed photons to the EC process (Malik et al., 2022). Figure 2 shows the light curve of 4C +01.02 and the analysis results of the local transient period. It can be seen intuitively from the figure that there is a QPO MJD 56422-58000, with the period of 253.2 $\pm$ 20.0 days and six cycles (The uncertainty is estimated as the half width at half maximum of the Gaussian function fitting the power peak. As shown in Figure 2, the red dotted line gives the number of cycles). Similar QPOs have been reported in the related literature, and their analysis results showed that there were 286 and 123 days of QPO between MJD 56900-58000 and MJD 57300-57900, respectively, with four and five cycles, respectively (Ren, Cerruti, & Sahakyan, 2023). Among them, the 286-day QPO is relatively close to the 253-day result given in this paper. It is worth noting that the results given in the relevant literature are based on 30-day binning data, while our 253-day results are based on 7-day binning data. Overall, these results all point to an interesting fact that there is a transient QPO behavior in FSRQ 4C +01.02, and this periodic light variability may originate from some special activities of the black hole.

Figure 2: Top panel: The graph above shows the weekly binned light curve for 4C +01.02, with MJD on the horizontal axis, flux and error on the vertical axis, and the light curve where the transient QPO phase occurs is shaded in gray. The figure below shows a magnified view of the gray shaded area, where the red dotted line is used to describe the cycle number of QPO. Bottom panel: The left panel shows the LSP power spectrum and significance evaluation results, where the black solid line is the LSP power and the green dashed line is the 99% confidence level contour. The figure on the right shows the wavelet power spectrum and significance evaluation results, where the black solid line is the WWZ power, and the green dotted line is the 99% confidence level contour.

4.2 PKS 0336-01

PKS 0336-01 is an FSRQ with a redshift of 0.852, and its radio, optical, infrared, and gamma-ray band light variability all have obvious flares. As shown in Figure 3 for the PKS 0336-01 gamma-ray local light curve, we first report a 94.6 $\pm$ 6.8 day transient QPO with six cycles between MJD 57020-57600. Figure 3 also shows the analysis results of the local light curve WWZ and LSP methods and the confidence level contours. The confidence level of the periodic modulation signal in the light curve exceeds 3$\sigma$. Therefore, this is also a more plausible example of a transient QPO.

Figure 3: Same description as Figure 2 for PKS 0336-01 (The blue dotted lines in the bottom panel represent the 99.7% confidence level contours).

4.3 PKS 0402-362

PKS 0402-362 is an FSRQ with a redshift of 1.5. Das, Mondal, & Prince (2023) studied the long-term light curve of gamma-ray for about 13 years. They found that most of the gamma-ray flare peaks had an asymmetrical profile, suggesting slow particle cooling times or changes in the Doppler factor were the main cause of these flares. There is no time lag between optical-IR and gamma-ray, consistent with a simple one-zone emission model. Among other things, they found that the light emission at different flux phases is dominated by the hot disk, suggesting that it is a good source for examining disk-jet coupling. The light curve is shown in Figure 4, and we report a 103.5 $\pm$ 7.9-day transient QPO between MJD 56817-57363, which experienced five cycles. There are similar reports in the related literature (Ren, Cerruti, & Sahakyan, 2023). They detected 221-day and 112-day QPO in the 30-day binned data and 7-day binned data, respectively, and believed that the 112-day QPO may be a harmonic frequency of 221 days.

Figure 4: Same description as Figure 2 for PKS 0402-362.

4.4 PKS 0537-441

PKS 0537-441 is a BL Lac object with a redshift of 0.89, and its light curve is shown in Figure 5. Earlier studies showed that the source was active in infrared, optical, ultraviolet, and X-ray, and brightened by about a factor of 2 in all bands (Tanzi et al., 1986). PKS 0537-441 shows a strong flux density variation in the 1.42 GHz band and erupts at J.D.2449011, characterized by a volatility index of about 15%, a variation of about 45%, and a time scale of about $10^{4}$ s. This extremely violent behavior may be the result of strong scattering from dense ionized structures in the interstellar medium (Romero, Combi, & Colomb, 1994). Sandrinelli et al. (2016) studied the long-term light curve of PKS 0537-441 and reported the long-term QPO of 285 days for gamma rays for the first time based on the LSP method, and Ren, Cerruti, & Sahakyan (2023) gave similar results with the CWT method. Their results all prove that there is an annual QPO in the long-term light curve of PKS 0537-441. Different from the above studies, we are more interested in the transient QPO in the light curve. As shown in Figure 5, we report for the first time two transient QPOs in PKS 0537-441, between MJD 56803-57183 and MJD 57636-58036, with periods of 55.0 $\pm$ 3.3 and 54.7 $\pm$ 3.3 days, respectively, with seven cycles. These two periods are very close to about 55 days, and there happens to be a relatively obvious Flare state in the middle of them. These two transient QPOs may reveal a very interesting regular radiation process inside the PKS 0537-441.

Figure 5: Same description as Figure 2 for PKS 0537-441 (The blue and red dashed lines in the bottom panel represent the 99.7% and 99.99% confidence level contours, respectively).

4.5 PKS 1424-41

PKS 1424-418 is an FSRQ with a redshift of 1.524. A study by Nesci et al. (2013) showed a clear correlation between NIR and gamma-ray emissions. Buson et al. (2014) analyzed in detail the four flares in the multiband light curve of PKS 1424-418 from 2008 to 2011, showing a high degree of correlation from optical to gamma-ray energy ranges. For all flares, the SED can be adequately represented by a Lepton model that includes inverse Compton emission from an external radiation field with similar parameters. Yang et al. (2021) reported a 355-day QPO between MJD 54963-58743 in the gamma-ray band at PKS 1424-418. Later, Ren, Cerruti, & Sahakyan (2023) also reported a 94-day transient QPO between MJD 56100-56500. The results of our systematic analysis showed that there were 57.2 $\pm$ 4.1 and 341 $\pm$ 25.8 days of transient QPO between MJD 56998-57331 and MJD 57636-58036, respectively, with six cycles. Among them, the 341-day cycle is very close to the 355-day cycle of Yang et al. (2021), the 57-day transient QPO is reported for the first time.

Figure 6: Same description as Figure 2 for PKS 1424-41 (The blue dotted lines in the bottom panel represent the 99.7% confidence level contours).

4.6 PKS 1510-089

PKS 1510-089 is a very bright FSRQ with a redshift of 0.36, which has attracted attention because of its high variability, and its multi-wavelength has been detected and widely discussed by many telescopes for a long time. In recent years, Sandrinelli et al. (2016); Castignani et al. (2017); Ahnen et al. (2017); Oramas et al. (2021); Roy et al. (2022); Li et al. (2023) have discussed the QPO of the PKS 1510-089 gamma-ray-radio band. These results are very interesting and have a certain significance for the study of multi-band light variability and the radiation mechanism of this source. In the gamma-ray that this paper focuses on, a QPO of 115 $\pm$ 5 days was reported by Sandrinelli et al. (2016). Roy et al. (2022) reported QPOs of 3.6 and 92 days in the PKS 1510-089 gamma-ray band of MJD 54906-54923 and MJD 58200-58850, with five and seven cycles, respectively, and they discussed several possible physical models. The result we give is 91.8 $\pm$ 1.8 days similar to that of Roy et al. (2022). The difference is that the time range can be extended to MJD 58197-58955, with eight cycles, and the confidence level is about 4.3 $\sigma$, as shown in Figure 7. This result undoubtedly confirms the QPO of about 92-day for the gamma-ray of PKS 1510-089.

Figure 7: Same description as Figure 2 for PKS 1510-089 (The blue and red dashed lines in the bottom panel represent the 99.7% and 99.99% confidence level contours, respectively).

5 Discussion

5.1 Compare with previous works

The research on the QPO phenomenon in the light curve of blazars has always been a very interesting subject. Zhou et al. (2018) reported the first case of a transient QPO on a monthly timescale (34.5 days) in blazar PKS 2247-131, which has attracted widespread attention from the astronomical community. Subsequently, transient QPO events on day-to-month time scales have been reported in different sources (see, e.g., Gupta et al., 2019; Sarkar et al., 2020, 2021; Roy et al., 2022; Chen et al., 2022). However, so far, high-confidence transient QPO cases are still rare, and the analysis methods used in different works varies, preventing systematic analysis of the sample. For this reason, an research effort to systematically search for transient QPOs, especially to limit the incidence of transient QPOs, would be of great importance.

Compared with long-term stable QPOs, searching for transient QPOs as this paper focuses on is more complicated, mainly because the time scale of the cycle for transient QPOs is shorter. The time scale of long-term QPO reported so far is generally more than one year or even more than ten years, while the time scale of transient QPO is generally within one year, mainly tens of days. This leads to higher requirements on the search method. Common period search methods such as power density spectroscopy (PDS), REDFIT, LSP, etc. are difficult to identify the transient QPO in a specific time period in the light curve. Different from these methods, thanks to the flexibility of wavelets, the WWZ method can identify the transient QPO signal in a specific time period in the light curve, but there are still cases of false positives, which generally require careful visual inspection and more detailed analysis to be more convincing.

In a very recent study, Ren, Cerruti, & Sahakyan (2023) analyzed the Fermi-LAT light curves of the 35 brightest blazars from the start of the Fermi mission (August 2008) to April 2021, energies from 100 MeV to 300 GeV. Two binning strategies, 7-day and 30-day binning light curves, were used in the analysis. Finally, 24 candidate quasars with transient QPO were identified in 35 active galactic nuclei, and their QPO periods were mainly concentrated within one year. Different from the work of Ren, Cerruti, & Sahakyan (2023), here we use a larger search sample in order to search based on a flux-limited complete sample as much as possible, and we use the WWZ method that is more sensitive to transient QPOs. In addition, we use more conservative (strict) criteria to determine the transient QPO: (1) Requiring the number of cycles of the QPO to be greater than or at least equal to five. (2) The data points in the light curve are required to be as complete as possible (those sources whose data points in the light curve are missing more than 60% are discarded). (3) The periodicity of those light curves that cannot be identified by the naked eye is discarded. Finally, we identified eight transient QPO candidates in six sources, and the confidence level of six of these transient QPO candidates quantitatively assessed by Monte Carlo simulation was not less than 3$\sigma$. The 92-day transient QPO in PKS 1510-089 was consistent with the report in Roy et al. (2022). The transient QPOs in 4C +01.02 (253-day), PKS 0402-362 (103-day) and PKS 1424-41 (341-day) are close to the reports in Yang et al. (2021) and Ren, Cerruti, & Sahakyan (2023). Moreover, we found for the first time a 94-day transient QPO in PKS 0336-01, 55-day and 54-day QPOs in PKS 0537-441, and a 57-day QPO in PKS 1424-41.

5.2 Physical explanation

The generation mechanism of QPO in active galactic nuclei is still controversial, and some interesting physical models have also been proposed. The supermassive binary black hole system achieved great success in explaining the $\sim$ 12-year QPO of OJ 287, and subsequently, this model was often used to explain some stable long-duration QPOs (Valtonen et al., 2008; Villforth et al., 2010). In addition, the continuous jet precession and Lense-Thirring precession of the accretion disk can also explain the long-duration QPO. There are extensive discussions in the literature such as Romero et al. (2000); Rieger (2004); Stella & Vietri (1998); Liska et al. (2018); You, Bursa, & Życki (2018). For transient QPOs, it is clear that other physical models need to be considered. One possible mechanism is that hotspots rotating in the innermost stable circular orbit of a supermassive black hole dominate the periodic modulation process in the light curves (Zhang & Bao, 1991; Gupta, Srivastava, & Wiita, 2009). In this model, the rotation of the hot spot will participate in modulating the seed photon field for external inverse Compton scattering (EC) within the jet, thereby modulating the gamma-ray emission. The transient QPO that we observe lasts only a few days due to the Doppler enhancement effect (Roy et al., 2022). Second, magnetic reconnection events of nearly equally spaced magnetic islands inside the jet can modulate the QPO of BH systems at different scales (Huang et al., 2013). This model requires the formation of uniformly equidistant magnetic islands in a region with an Eddington strength magnetic field close to the black hole to periodically enhance the flux and thus generate the observed transient QPOs. Shukla et al. (2018) modeled the extremely fast variability ($\sim$ 5 minutes) of gamma-ray during the FSRQ CTA 102 outburst based on a magnetic reconnection process. This model can explain some short and fast transient QPOs, but explaining transient QPOs over weeks or even months is obviously more difficult and requires more theoretical studies and numerical simulations to explore.

In addition, a model for the helical motion of the enhanced emission region (or blob) inside the jet has been widely discussed in flare transient QPO studies (Mohan & Mangalam, 2015; Sobacchi, Sormani, & Stamerra, 2017; Zhou et al., 2018). In the one zone leptonic model, the blazar emission is dominated by the emission region (or blob) in the jet, and the seed photons in the emission region (or blob) produce high-energy gamma-ray emission through a synchrotron self-Compton process. The large-scale magnetic field or some dynamical reasons may cause the jet or the emission region (or blob) within the jet to produce a helical motion, which is also known as the helical jet structure (Mohan & Mangalam, 2015; Sobacchi, Sormani, & Stamerra, 2017). In this scenario, the viewing angle ($\theta$) changes with the motion of the emission region (or blob), which can be described as

$$cos\theta=cos\phi cos\psi+sin\phi sin\psi cos(2\pi t/P_{obs}),$$ (12)

where $\psi$ is the angle between the jet and the observer’s line of sight, $\phi$ is the pitch angle of the helical path of the emission region (or blob) with respect to the jet axis, and $P_{obs}$ is the periodicity in the observed light curve (Sobacchi, Sormani, & Stamerra, 2017; Zhou et al., 2018). For the observer, the gamma-ray emission is enhanced by relativistic clustering effects, so that there is a translation between the true physical driving period $P$ of the emission region (or blob) and the observed period, $P=P_{obs}/(1-\beta cos\phi cos\psi)$, where $\phi$ and $\psi$ we take typical values of $2^{\circ}$ and $5^{\circ}$ (Zhou et al., 2018). The bulk Lorentz factor ($\Gamma=1/(1-\beta^{2})^{1/2}$) of the emission region (or blob), where Lorentz factor we also take some typical values, for FSRQ type and BL lac type we take $\Gamma=20$ and $\Gamma=8.5$ respectively. Then, by substituting specific parameters on this basis, the real physical driving period $P$ of each transient QPO, the distance $D$ that the emitting region (or sphere) moves in one cycle ($D=c\beta Pcos\phi$), and the total projected distance of the entire transient QPO phase $D_{p}=NDsin\psi$ can be estimated (see Table 3). In terms of timescales, the helical jet structural scenario provides a more plausible explanation for most of the transient QPOs we discuss here.

Table 3: Helical Jet Parameters for Mentioned Features.

Source name	$P_{obs}$ (days)	P (years)	D (pc)	$D_{p}$ (pc)
4C +01.02	253.2 $\pm$ 20.0	$\sim$ 122.6	$\sim$ 37.5	$\sim$ 19.6
PKS 0336-01	94.6 $\pm$ 6.8	$\sim$ 19.6	$\sim$ 14.0	$\sim$ 7.3
PKS 0402-362	103.5 $\pm$ 7.9	$\sim$ 50.1	$\sim$ 15.3	$\sim$ 6.7
PKS 0537-441	55.0 $\pm$ 3.3	$\sim$ 13.3	$\sim$ 4.0	$\sim$ 2.5
	54.7 $\pm$ 3.3	$\sim$ 13.2	$\sim$ 4.0	$\sim$ 2.5
PKS 1424-41	57.2 $\pm$ 4.1	$\sim$ 27.7	$\sim$ 8.5	$\sim$ 4.4
	341 $\pm$ 25.8	$\sim$ 165.1	$\sim$ 50.5	$\sim$ 26.4
PKS 1510-089	92.3 $\pm$ 5.2	$\sim$ 44.7	$\sim$ 13.7	$\sim$ 9.5

In this work, we also note two special cases, PKS 0537-441 and PKS 1424-41. We find a transient QPO of about 55 days at MJD 56803-57183 of PKS 0537-441, and the same transient QPO appears again within the error range about 450 days later at MJD 57636-58036. The reappearance of the transient QPO in PKS 0537-441 in the helical jet scenario implies that the spiraling emitting region (or blob) within the jet is reproducible, in other words, after 450 days the helicalling emitting region (or blob) again modulates the PKS 0537-441 gamma ray emission in the same way. This is the first case of a repeated transient QPO reported to date.

In addition, the two cases of transient QPO in PKS 1424-41 are also very interesting. Transient QPOs of approximately 57 days and approximately 341 days occurred at MJD 56998-57331 and MJD 56611-58644, respectively, with the 57-day transient QPO occurring between the second and third cycles of the 341-day transient QPO. And the occurrence of the 57-day transient QPO causes the peak flux in the second period of the 341-day transient QPO to be lower than the average of the peak flux in the other five periods. Obviously, it is difficult to explain this special nested transient QPO appearing in PKS 1424-41 in a simple helical jet model. If the origin of the 341-day transient QPO is indeed the helical motion of the emission region (or blob) within the jet, then it is likely that the origin of the 57-day transient QPO involves a more intricate set of physical mechanisms.

5.3 Incidence rate of Transient QPO Events

Based on the light variability properties of transient QPOs reported here and some transient QPOs reported in the literature, we noticed some interesting and important clues: (1) transient QPOs tend to appear in the flare states of the blazars. (2) Most of the reported transient QPOs in the literature appear to happen mainly in FSRQs, with comparatively fewer occurrences in BL Lacs. Our sample shows a similar pattern. However, whether this phenomenon is due to sample selection bias because of FSRQ’s brighter luminosity than BL Lacs in the Fermi-LAT gamma-ray band requires further investigation in the future. (3) Of the 134 bright sources examined in this paper, only 6 were found to have transient QPOs, and of these 6, only 4 exhibited transient QPOs with a confidence level greater than 3$\sigma$. The incidence of high-confidence transient QPO events is about 3%, which suggests that transient QPO phenomena are very rare in bright flares.

It seems plausible that the appearance of QPO indicates the formation of some ordered low-entropy dynamic structure within the black hole system. Physically, the formation of such an ordered low-entropy structure would require certain specific physical processes within the system, such as the helical motion of the enhanced emission region (or blob) inside the jet, hotspots rotating on the innermost stable circular orbit of the supermassive black hole, or magnetic reconnection between nearly equidistant magnetic islands within the jet, and the probability of its formation depends on these specific physical mechanisms. In the framework of interpreting QPOs as arising from the helical motion of an enhanced emission region (or blob) within a jet, a large-scale stabilizing magnetic field would be required to underlie the jet’s helical structure. And the probability of establishing such an ordered large-scale magnetic field configuration would determine the occurrence of transient QPO phenomena. Thus, the low rate of transient QPO events at least strongly implies that transient QPO may not be dominated by a universal physical process. Tighter empirical constraints on the occurrence of transient QPO events through future observations, alongside comparative analyses of the incidence of specific processes invoked by different physical models producing transient QPO phenomena, have the potential to provide strong discriminatory power between different theoretical frameworks.

6 summary

In this paper, a systematic search for transient QPO signals was conducted in 134 blazars with a peak flux exceeding $1\times 10^{-6}$ ph cm${}^{-2}$ s${}^{-1}$, based on the 0.1-300 GeV gamma-ray light curve from Fermi-LAT. The main results are summarized in the following:

1. This work reports four new transient QPO events, including an approximately 94-day event in PKS 0336-01, two events lasting approximately 55 days each in PKS 0537-441, and an approximately 57-day event in PKS 1424-41.

2. We discussed the physical origin of transient QPOs, and the presence of a helical jet structure in the enhanced emission region provides the best explanation.

3. Unique transient QPO events were identified in two sources. The occurrence of repeated transient QPOs in PKS 0537-441 suggests that the emission region (or blob) with helical motion modulates the gamma-ray emission again. The special nested transient QPO events in PKS 1424-41 may be related to more complex physical scenarios.

4. We found that transient QPO events tend to occur in the flaring state of blazars, and they are more likely to occur in the population of flat-spectrum radio quasars. Additionally, the occurrence rate of transient QPO events is estimated to be approximately 3%.

DATA AVAILABILITY

Monitored Source List Light Curves: https://fermi.gsfc.nasa.gov/ssc/data/access/lat/msl_lc/
The Fermi-LAT data used in this article are available in the LAT data server at: https://fermi.gsfc.nasa.gov/ssc/data/access/
The Fermi-LAT data analysis software is available at: https://fermi.gsfc.nasa.gov/ssc/data/analysis/software/
The weighted wavelet Z-transform method: https://github.com/eaydin/WWZ/
This light curve simulation method: https://github.com/samconnolly/DELightcurveSimulation/,

Appendix A Fermi-LAT bright source samples

Table 4: Fermi-LAT bright source samples.

Source name	4FGL name	Source class	R.A.(J2000)	Dec.(J2000)	Redshift
PMN J0017-0512	4FGL J0017.5-0514	FSRQ	4.3992	-5.2116	0.227
TXS 0059+581	4FGL J0102.8+5824	FSRQ	15.6907	58.4031	0.644
4C +01.02	4FGL J0108.6+0134	FSRQ	17.1615	1.5834	2.099
4C 31.03	4FGL J0112.8+3208	FSRQ	18.2097	32.1382	0.603
PKS 0130-17	4FGL J0132.7-1654	FSRQ	23.1812	-16.9135	1.02
PKS 0131-522	4FGL J0133.1-5201	FSRQ	23.274	-52.0011	0.02
PKS 0202-17	4FGL J0205.0-1700	FSRQ	31.2403	-17.0222	1.7395
0208-512	4FGL J0210.7-5101	FSRQ	32.6925	-51.0172	0.999
CGRaBS J0211+1051	4FGL J0211.2+1051	BL Lac	32.8049	10.8597	0.2
S3 0218+35	4FGL J0221.1+3556	FSRQ	35.2729	35.9372	0.944
3C 66A	4FGL J0222.6+4302	BL Lac	35.665	43.0355	0.37
PKS 0226-559	4FGL J0228.3-5547	FSRQ	37.09	-55.7676	2.464
4C +28.07	4FGL J0237.8+2848	FSRQ	39.4684	28.8025	1.213
0235+164	4FGL J0238.6+1637	BL Lac	39.6622	16.6165	0.94
PKS 0244-470	4FGL J0245.9-4650	FSRQ	41.5005	-46.8548	1.385
PKS 0250-225	4FGL J0252.8-2219	FSRQ	43.1998	-22.3237	1.427
PKS 0301-243	4FGL J0303.4-2407	BL Lac	45.8604	-24.1198	0.266
PKS 0336-01	4FGL J0339.5-0146	FSRQ	54.8789	-1.7766	0.852
PKS 0346-27	4FGL J0348.5-2749	FSRQ	57.1589	-27.8204	0.991
4C +50.11	4FGL J0359.6+5057	FSRQ	59.916	50.959	1.52
PKS 0402-362	4FGL J0403.9-3605	FSRQ	60.974	-36.0839	1.4228
PKS 0426-380	4FGL J0428.6-3756	BL Lac	67.1684	-37.9388	1.11
NRAO 190	4FGL J0442.6-0017	FSRQ	70.6611	-0.2954	0.844
PKS 0454-234	4FGL J0457.0-2324	FSRQ	74.2632	-23.4145	1.003
PKS 0458-02	4FGL J0501.2-0158	FSRQ	75.3034	-1.9873	2.286
PKS 0502+049	4FGL J0505.3+0459	FSRQ	76.3466	4.9952	0.954
PKS 0514-459	4FGL J0515.6-4556	FSRQ	78.9385	-45.9453	0.194
VER 0521+211	4FGL J0521.7+2112	BL Lac	80.445	21.213	0.108
CRATES J0531-4827	4FGL J0532.0-4827	BL Lac	83.002	-48.461	0.8116
OG 050	4FGL J0532.6+0732	FSRQ	83.1625	7.5454	1.254
PKS 0537-441	4FGL J0538.8-4405	BL Lac	84.7098	-44.0858	0.89
PKS 0537-286	4FGL J0539.9-2839	FSRQ	84.9762	-28.6655	3.104
TXS 0552+398	4FGL J0555.6+3947	FSRQ	88.901	39.788	2.365
TXS 0646-176	4FGL J0648.4-1743	FSRQ	102.1187	-17.7348	1.232
MG2 J071354+1934	4FGL J0713.8+1935	FSRQ	108.482	19.5834	0.54
S5 0716+714	4FGL J0721.9+7120	BL Lac	110.488	71.341	0.31
4C 14.23	4FGL J0725.2+1425	FSRQ	111.32	14.4205	1.039
PKS 0727-11	4FGL J0730.3-1141	FSRQ	112.5796	-11.6868	1.591
PKS 0736+01	4FGL J0739.2+0137	FSRQ	114.8251	1.6179	0.1894
BZU J0742+5444	4FGL J0742.6+5443	FSRQ	115.671	54.727	0.72
B2 0748+33	4FGL J0752.2+3313	FSRQ	118.055	33.232	1.9352
PKS 0805-07	4FGL J0808.2-0751	FSRQ	122.0647	-7.8527	1.837
TXS 0827+243	4FGL J0830.8+2410	FSRQ	127.701	24.173	0.9388
S5 0836+71	4FGL J0841.3+7053	FSRQ	130.3515	70.895	2.172
PMN J0852-5755	4FGL J0852.5-5755	BCU	133.1614	-57.9249	…
OJ 287	4FGL J0854.8+2006	BL Lac	133.7036	20.1085	0.3056
3EG J0903-3531	4FGL J0904.5-3513	BCU	136.14	-35.231	…
PKS 0903-57	4FGL J0904.9-5734	BCU	136.2216	-57.5849	0.695
PKS B0906+015	4FGL J0909.1+0121	FSRQ	137.292	1.3599	1.0232
S5 0917+62	4FGL J0921.6+6216	FSRQ	140.418	62.271	1.4576
PKS 0920-39	4FGL J0922.7-3959	FSRQ	140.692	-39.987	0.591
S4 0954+65	4FGL J0958.7+6534	BL Lac	149.6969	65.5652	0.368
PKS 1004-217	4FGL J1006.7-2159	FSRQ	151.6934	-21.989	0.331
4C +40.25	4FGL J1023.1+3949	FSRQ	155.7982	39.8043	1.2545
S4 1030+61	4FGL J1033.9+6050	FSRQ	158.4643	60.852	1.4082
PMN J1038-5311	4FGL J1038.8-5312	FSRQ	159.6694	-53.1954	1.45
S5 1044+71	4FGL J1048.4+7143	FSRQ	162.1151	71.7266	1.15
4C +01.28	4FGL J1058.4+0133	BL Lac	164.6234	1.5663	0.8936
Mrk 421	4FGL J1104.4+3812	BL Lac	166.1138	38.2088	0.0301
PMN J1123-6417	4FGL J1123.5-6418	BCU	170.889	-64.311	…
PKS 1127-14	4FGL J1129.8-1447	FSRQ	172.5294	-14.8243	1.184
B2 1144+40	4FGL J1146.9+3958	FSRQ	176.74	39.977	1.088
Ton 599	4FGL J1159.5+2914	FSRQ	179.8826	29.2455	0.7247
1ES 1215+303	4FGL J1217.9+3007	BL Lac	184.467	30.1168	0.129
TXS 1219+285	4FGL J1221.5+2814	BL Lac	185.382	28.2329	0.102
PKS B1222+216	4FGL J1224.9+2122	FSRQ	186.2269	21.3796	0.4338
3C 273	4FGL J1229.0+0202	FSRQ	187.2779	2.0524	0.1583
ON 246	4FGL J1230.2+2517	BL Lac	187.5587	25.302	0.14
J123939+044409	4FGL J1239.5+0443	FSRQ	189.885	4.728	1.7606
PKS 1244-255	4FGL J1246.7-2548	FSRQ	191.695	-25.797	0.633
3C 279	4FGL J1256.1-0547	FSRQ	194.0465	-5.7893	0.5362
OP 313	4FGL J1310.5+3221	FSRQ	197.6194	32.3455	0.9959
GB6 B1310+4844	4FGL J1312.6+4828	BCU	198.1806	48.4753	0.501
PKS 1313-333	4FGL J1316.1-3338	FSRQ	199.0333	-33.6498	1.21
PKS 1335-127	4FGL J1337.6-1257	FSRQ	204.4158	-12.9569	0.539
B3 1343+451	4FGL J1345.5+4453	FSRQ	206.3882	44.8832	2.542
1406-076	4FGL J1408.9-0751	FSRQ	212.2353	-7.8741	1.5
OQ 334	4FGL J1422.3+3223	FSRQ	215.6266	32.3862	0.6816
PKS 1424-41	4FGL J1427.9-4206	FSRQ	216.9846	-42.1054	1.522
H 1426+428	4FGL J1428.5+4240	BL Lac	217.1359	42.6725	0.129
PKS 1454-354	4FGL J1457.4-3539	FSRQ	224.3613	-35.6528	1.424
PKS 1502+106	4FGL J1504.4+1029	FSRQ	226.1041	10.4942	1.838
B2 1504+37	4FGL J1506.1+3731	FSRQ	226.5397	37.5142	0.6715
PKS 1510-089	4FGL J1512.8-0906	FSRQ	228.2106	-9.1	0.36
PKS 1514-24	4FGL J1517.7-2422	BL Lac	229.4242	-24.3721	0.049
B2 1520+31	4FGL J1522.1+3144	FSRQ	230.5416	31.7373	1.489
PG 1553+113	4FGL J1555.7+1111	BL Lac	238.9294	11.1901	0.36
PKS B 1622-297	4FGL J1626.0-2950	FSRQ	246.5251	-29.8575	0.815
1633+382	4FGL J1635.2+3808	FSRQ	248.817	38.14	1.814
Mrk 501	4FGL J1653.8+3945	BL Lac	253.4676	39.7602	0.0329
GB6 J1700+6830	4FGL J1700.0+6830	FSRQ	255.022	68.504	0.301
B3 1708+433	4FGL J1709.7+4318	FSRQ	257.4212	43.3124	1.027
PMN J1717-5155	4FGL J1717.6-5154	FSRQ	259.403	-51.909	…
S4 1726+45	4FGL J1727.4+4530	FSRQ	261.8652	45.511	0.717
1730-130	4FGL J1733.0-1305	FSRQ	263.263	-13.086	0.902
OT 355	4FGL J1734.3+3858	FSRQ	263.5857	38.9643	0.97
S4 1749+70	4FGL J1748.6+7005	BL Lac	267.1368	70.0974	0.77
OT 081	4FGL J1751.5+0938	BL Lac	267.8867	9.6502	0.322
S5 1803+78	4FGL J1800.6+7828	BL Lac	270.1903	78.4678	0.68
S4 1800+44	4FGL J1801.5+4404	FSRQ	270.3846	44.0728	0.663
PKS 1824-582	4FGL J1829.2-5813	FSRQ	277.3017	-58.232	1.531
PKS 1830-211	4FGL J1833.6-2103	FSRQ	278.416	-21.0615	2.507
CGRaBS J1849+6705	4FGL J1849.2+6705	FSRQ	282.317	67.0949	0.657
PKS B1908-201	4FGL J1911.2-2006	FSRQ	287.7902	-20.1153	1.119
TXS 1923+123	4FGL J1925.7+1227	BCU	291.4201	12.4606	…
PKS 1954-388	4FGL J1958.0-3845	FSRQ	299.4992	-38.7518	0.63
1ES 1959+650	4FGL J2000.0+6508	BL Lac	299.9994	65.1485	0.047
S5 2007+77	4FGL J2005.5+7752	BL Lac	301.3792	77.8787	0.342
PKS 2023-07	4FGL J2025.6-0735	FSRQ	306.4194	-7.598	1.388
PKS 2032+107	4FGL J2035.4+1056	FSRQ	308.8431	10.9352	0.601
PKS 2052-47	4FGL J2056.2-4714	FSRQ	314.0682	-47.2466	1.489
PMN J2141-6411	4FGL J2141.7-6410	BCU	325.4435	-64.1874	0.959
OX 169	4FGL J2143.5+1743	FSRQ	325.8981	17.7302	0.211
PKS 2144+092	4FGL J2147.1+0931	FSRQ	326.7923	9.4963	1.113
PKS 2142-75	4FGL J2147.3-7536	FSRQ	326.803	-75.6037	1.139
PKS 2149-306	4FGL J2151.8-3027	FSRQ	327.9813	-30.4649	2.345
PKS 2155-304	4FGL J2158.8-3013	BL Lac	329.7169	-30.2256	0.1167
PKS 2155-83	4FGL J2201.5-8339	FSRQ	330.5802	-83.6366	1.865
4FGL J2202.7+4216	4FGL J2202.7+4216	BL Lac	330.6804	42.2778	0.0668
CTA 102	4FGL J2232.6+1143	FSRQ	338.1517	11.7308	1.032
PKS 2233-148	4FGL J2236.5-1433	BL Lac	339.142	-14.5562	0.325
TXS 2241+406	4FGL J2244.2+4057	FSRQ	341.053	40.9538	1.171
PMN J2250-2806	4FGL J2250.7-2806	BL Lac	342.6854	-28.1109	0.525
3C 454.3	4FGL J2253.9+1609	FSRQ	343.4906	16.1482	0.859
PKS 2255-282	4FGL J2258.1-2759	FSRQ	344.5248	-27.9726	0.926
PKS B2258-022	4FGL J2301.0-0158	FSRQ	345.2832	-1.9679	0.777
B2 2308+34	4FGL J2311.0+3425	FSRQ	347.7722	34.4197	1.817
PKS 2320-035	4FGL J2323.5-0317	FSRQ	350.8831	-3.2847	1.411
1ES 2322-409	4FGL J2324.7-4041	BL Lac	351.1861	-40.6804	0.174
PKS 2326-502	4FGL J2329.3-4955	FSRQ	352.337	-49.928	0.518
PMN J2345-1555	4FGL J2345.2-1555	FSRQ	356.3019	-15.9188	0.621
1ES 2344+514	4FGL J2347.0+5141	BL Lac	356.7702	51.705	0.044
PKS 2345-16	4FGL J2348.0-1630	FSRQ	357.0109	-16.52	0.576
Notes:
Column 1 is the name of the object;
Column 2 is the name of Fermi 4FGL;
Column 3 is the type of source (BL Lac or FSRQ);
Column 4 and Column 5 are right ascension (RA) and declination (Dec) coordinates;
Column 6 is the redshift of the source.