Unraveling the Innermost Jet Structure of OJ 287 with the First GMVA + ALMA Observations

We carried out high-resolution VLBI observations toward OJ 287 at 3.5 mm with GMVA on 2017 April 2. These observations mark the first VLBI observations with the phased ALMA, which consists of 37 ALMA antennas and is equivalent to a 70 meter dish (Event Horizon Telescope Collaboration et al. 2019a). The participating stations also include eight Very Long Baseline Array (VLBA) stations and five European stations (Effelsberg, IRAM-30 m, Metsähovi, Onsala, Yebes-40m). The on-source time was around 375 minutes between UT 17 and UT 7 the next day (April 3).

Most stations had good or typical weather conditions during the observation, except for the VLBA Maunakea and Pie Town stations, which resulted in few fringe detections with limited signal-to-noise ratio (S/N) on the baselines to these two stations. No fringes were found on the baselines to Metsähovi, due to a faulty backend setup. All data were recorded in full polarization mode, with most stations recording on a circular polarization basis, while the ALMA data were converted from a mixed linear–circular basis to circular polarization mode using PolConvert (Martí-Vidal et al. 2016). The Yebes-40 m telescope recorded only left-handed circular polarization (LCP). The bandwidths and frequency ranges recorded are not the same for all stations. ³⁶ Only the common frequency ranges among all participating stations are used in later processing.

Data correlation was performed with the DiFX correlator (Deller et al. 2007) at the Max-Planck-Institut für Radioastronomie in Bonn, Germany. The final correlated data have a total bandwidth of 232 MHz, which was further divided into four 58 MHz intermediate frequency (IF) bands.

The postcorrelation data set was then processed with the ParselTongue (Kettenis et al. 2006) ${ \mathcal A }{ \mathcal I }{ \mathcal P }{ \mathcal S }$ (Greisen 2003) interface for fringe fitting and a priori amplitude calibration. We first performed parallactic angle correction with the ${ \mathcal A }{ \mathcal I }{ \mathcal P }{ \mathcal S }$ task clcor ³⁷ , and manual phase calibration using short segments of data to remove the instrumental phase offset between the different IFs. We then performed a global fringe fitting of the data using the task fring, with a solution interval of 10 s and subintervals down to 2 s, and by integrating over the whole 232 MHz bandwidth and averaging parallel-hand polarizations (RR and LL).

The (u,v)-coverage toward OJ 287 for all baselines with fringe detections is shown in Figure 1. We note that the participation of ALMA has provided an increase in the north–south resolution by a factor of ∼3 for observations of OJ 287. ALMA has also significantly improved fringe detection due to its high sensitivity (see Figure 2) with the maximum fringe S/N reaching ∼350 at baselines longer than 1.5 Gλらむだ.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

A priori amplitude calibration was performed in ${ \mathcal A }{ \mathcal I }{ \mathcal P }{ \mathcal S }$ with the task apcal, by multiplying the system temperatures (T_sys) and gain curves of each antenna. Opacity corrections were applied to stations that measure system temperatures with the noise diode method (VLBA and Effelsberg). For ALMA, IRAM-30 m, and Yebes-40 m, the T_sys measurements were performed using the hot/cold method, and therefore already included opacity correction. The ALMA T_sys values have also taken into account the phasing efficiencies derived during the quality assurance and PolConvert processes (e.g., Goddi et al. 2019). The cross-hand phase and delay offsets of the reference station were calibrated using the ${ \mathcal A }{ \mathcal I }{ \mathcal P }{ \mathcal S }$ procedure vlbacpol.

After the ${ \mathcal A }{ \mathcal I }{ \mathcal P }{ \mathcal S }$ calibration, the data were averaged in time (with an interval of 15 s) and frequency (with all channels within each IF averaged) for further processing.

We performed the imaging and self-calibration of the data independently with three different imaging softwares: DIFMAP, eht-imaging, and SMILI. DIFMAP is the software commonly used for the conventional CLEAN method for interferometric imaging (Shepherd et al. 1995). It interactively establishes a collection of point-source models from the inverse Fourier transform of the visibilities, i.e., the dirty map. CLEAN windows, which define the regions to search for CLEAN components, are used during our imaging process. Phase-only self-calibration is performed after each step of cleaning. Amplitude and phase self-calibration is performed once a good fit to the visibilities has been established through multiple steps of cleaning and phase self-calibration. We repeat the clean and self-calibration loops several times during our imaging process by gradually decreasing the solution interval of the amplitude and phase self-calibration. On the other hand, the regularized maximum likelihood (RML) methods, employed by the eht-imaging (Chael et al. 2016, 2018) and SMILI (Akiyama et al. 2017) libraries, reconstruct images by minimizing an objective function, which is a weighted combination of χかい² of the data and various regularizer terms. The data terms may include closure quantities (closure phases and amplitudes; e.g., Thompson et al. 2017), visibility amplitudes, and complex visibilities. Common regularizers include the maximum entropy (e.g., Chael et al. 2018), the ℓ₁-norm (e.g., Honma et al. 2014; Akiyama et al. 2017), the total variation (TV), and the total squared variation (TSV) of the brightness (e.g., Kuramochi et al. 2018). With RML methods, it is possible to achieve an angular resolution a few times finer than the nominal interferometric beam (e.g., Akiyama et al. 2017; Event Horizon Telescope Collaboration et al. 2019b). During our imaging process with eht-imaging and SMILI, we started with a Gaussian prior image and reconstructed images with only the closure quantities, or a combination of closure quantities and low-weighted visibility amplitudes. After a few iterations of imaging and self-calibrating, we included full complex visibilities in the optimization process, further constraining the reconstructed images. To determine the best set of regularizer combinations, we surveyed a range of different weights of each regularizer, in a total of ∼128 combinations, and selected the one that resulted in the best fit to the closure quantities.

After the imaging of the total intensity, we estimated the instrumental polarimetric leakage (known as D-terms) for each station, using the self-calibrated data of OJ 287. This process was carried out independently with two pipelines: the ${ \mathcal A }{ \mathcal I }{ \mathcal P }{ \mathcal S }$ task lpcal and the eht-imaging library, each based on a particular self-calibrated data set generated during the total intensity imaging process, i.e., DIFMAP and eht-imaging, respectively. Both approaches provide consistent values of D-terms. Details of the leakage calibration are described in the Appendix. The polarization imaging of the lpcal processed data was carried out with DIFMAP. With eht-imaging, the imaging was performed iteratively with the D-terms calculation. The calibration of the absolute orientation of the EVPAs was performed through comparison with the ALMA array data (Goddi et al. 2021).

We also carried out nonimaging analysis of the data to measure the properties of the jet. We performed circular Gaussian model fitting to the SMILI self-calibrated visibility data with DIFMAP. The results indicated that the jet structure could be represented by four Gaussian components. We labeled the components following the convention described in Gómez et al. (2022). The total flux, size, and position offset with respect to the core (the component at the southeastern end of the jet; see Section 3 below) of all the components are listed in Table 1. The uncertainties of the fitted parameters are derived following the equations outlined in Nair et al. (2019).

Table 1. Model Fitting Parameters of OJ 287 with GMVA + ALMA on 2017 April 2

Comp	S	r	PA	FWHM	${T}_{b}^{\mathrm{obs}}$
Name	(Jy)	(μみゅーas)	({°})	(μみゅーas)	(1010 K)
(1)	(2)	(3)	(4)	(5)	(6)
C0	1.25 ± 0.13	0.0	0.0	31.1 ± 2.4	21.2 ± 4.1
C1	0.96 ± 0.10	63.9 ± 2.4	−41.5 ± 2.2	28.9 ± 2.4	18.8 ± 3.9
C2a	2.49 ± 0.25	172.9 ± 2.4	−40.0 ± 0.8	44.4 ± 2.5	20.6 ± 3.1
C2b	0.51 ± 0.05	212.7 ± 3.5	−47.7 ± 0.9	42.5 ± 4.6	4.7 ± 1.2

Note. Columns from left to right: (1) Component ID; (2) Flux density; (3) Radial distance from the core component (C0); (4) Position angle; (5) Component FWHM; and (6) Observed brightness temperature.

Download table as:  ASCIITypeset image

Figure 3 shows the total intensity maps of OJ 287 obtained with our GMVA + ALMA observations, achieving the highest angular resolution to date of the source at the wavelength of 3.5 mm. The imaging results are consistent across different imaging methods (CLEAN and RML). Under the nominal resolution, the jet appears to consist of three major features, extending along the southeast to northwest direction. We denote the three features as the components C0, C1, and C2, as shown in the bottom right panel of Figure 3.

Zoom In Zoom Out Reset image size

Component C0, which lies at the southern end of the jet, is compact and shows the highest brightness temperature (Table 1). This feature is more likely to be the VLBI core at 3.5 mm. The component C2 has the highest flux density among the three components. This feature shows complex substructures under the fine resolution of the RML images (Figure 3, top middle and top right panels). We see hints of the jet bending and extending toward the western direction downstream of C2. This bend is more obvious in the lower-frequency maps which are more sensitive to the extended lower-brightness regions despite the lower angular resolutions (e.g., Cohen 2017; Jorstad et al. 2017). The downstream jet is largely resolved out and not well constrained in our high-resolution images, because of their steep spectra and extended structure. Our higher-resolution images reveal for the first time the twisted morphology of the innermost, ultracompact jet region. The first bending occurs between C0 and C1, with the jet axis gradually changing from north to northwest (clockwise). We also see hints of a subsequent bending occurring downstream of C1, where the jet axis turns in the counterclockwise direction.

The three-component structure is also consistent with the recent 22 GHz RadioAstron space–ground VLBI observations of OJ 287 made at a similar resolution (Gómez et al. 2022). However, a PA difference of ∼50° in the inner jet can be found when comparing with the RadioAstron image obtained in 2014. Such a difference could be attributed to the variation in the PA over ∼ 3 yr. A detailed analysis of the inner jet PA variation on a yearly scale, and a comparison with theoretical predictions, will be presented in a forthcoming paper (G-Y. Zhao et al., in preparation).

In order to quantify the PA evolution along the jet, we fit the jet ridgeline on the eht-imaging map. First, we transform the image to polar coordinates centered on the jet origin, then slice it transversely. For each slice, we store the flux density peak position, then transform them back to Cartesian coordinates. Thus, we obtain a collection of positions tracing the jet axis between C0 and C2. The results are presented in Figure 4, where we also show a sketch tracing the conical structure of C2 in the figure. The jet axis near C0 extends along a PA of ∼−15°, decreases to ∼−50° at C1, and then starts to increase again near C2. A similar trend can also be found in the SMILI image.

Zoom In Zoom Out Reset image size

We investigate the brightness temperature of the OJ 287 jet using two independent approaches. (1) We calculate the observed brightness temperature of each Gaussian component from the model fitting results, using the following equation (e.g., Tingay et al. 2002):

$$\begin{eqnarray}&&{T}_{{\rm{b}}}^{\mathrm{obs}}=1.22\times {10}^{12}\displaystyle \frac{S}{{\theta }_{\mathrm{obs}}^{2}{\nu }^{2}},\end{eqnarray} \tag{ 1 }$$

where S is the component flux density in Jy, θしーた_obs is the size of the emitting region in mas, and νにゅー is the observing frequency in GHz. (2) We calculate the minimum and maximum brightness temperature directly from the visibilities, using the method described in Lobanov (2015). The model fitting results, which are listed in Table 1, show the observed brightness temperature of the jet components at 86 GHz ranges from 10¹⁰ to 10¹¹ K. This is in agreement with the values calculated from the visibility amplitudes, as shown in Figure 5. The brightness temperature values agree quantitatively with the typical values at the same frequency band (e.g., Lee et al. 2008; Nair et al. 2019). The 86 GHz brightness temperatures are about one order of magnitude lower compared to those at 22 GHz obtained from the RadioAstron results (Gómez et al. 2022). This can be attributed to differences in the intrinsic brightness and opacity between the two frequencies.

Zoom In Zoom Out Reset image size

We estimate the intrinsic brightness temperature, ${T}_{{\rm{b}}}^{\mathrm{int}}$, by (e.g., Gómez et al. 2016):

$$\begin{eqnarray}&&{T}_{{\rm{b}}}^{\mathrm{int}}=\left(1+z\right){\delta }^{-1}{T}_{{\rm{b}}}^{\mathrm{obs}},\end{eqnarray} \tag{ 2 }$$

where δでるた stands for the Doppler factor. We adopt the value of the latest estimate based on the proper motion of moving components from the VLBA-BU-BLAZAR monitoring program, δでるた = 8.6 ± 2.8 (Weaver et al. 2022). This gives the intrinsic brightness temperature values ${T}_{{\rm{b}},{\rm{C}}0}^{\mathrm{int}}$ = $(3.2\pm 1.7)\,\times {10}^{10}\ {\rm{K}},{T}_{{\rm{b}},{\rm{C}}1}^{\mathrm{int}}=(2.8\pm 1.4)$ × ${10}^{10}\ {\rm{K}},{T}_{{\rm{b}},\ {\rm{C}}2{\rm{a}}}^{\mathrm{int}}=(3.1\pm 1.4)$ × 10¹⁰ K, and ${T}_{{\rm{b}},\ {\rm{C}}2{\rm{b}}}^{\mathrm{int}}=(0.7\pm 0.4)\times {10}^{10}\ {\rm{K}}$, for each component, respectively. These values fall below the equipartition value of ∼ 5 × 10¹⁰ (Readhead 1994), indicating possible magnetic dominance in the innermost jet. However, this is quite uncertain, as the errors in the Doppler factor and brightness temperature values are large.

We perform polarimetric imaging of the instrumental polarization calibrated data with CLEAN and eht-imaging independently. The corresponding images are shown in Figure 6. Our images show that the overall degree of polarization of OJ 287 is ∼8%, which is in quantitative agreement with the ALMA array results of 8.8% presented in Goddi et al. (2021). The EVPAs extend mostly along the mean jet axis, which suggests that the magnetic field in the jet has a predominant toroidal component. Again, the image reconstructed by eht-imaging shows a fine structure, because of the super-resolution that is naturally achieved by the forward-modeling method. However, even in the CLEAN image, which is convolved with the nominal beam, we see a remarkable polarimetric structure in the inner jet. The overall structure is consistent between the two images that have been reconstructed independently with different approaches. We notice that the apparent difference in the fractional polarization between the two maps is due to the fact that the CLEAN images are convolved with the nominal beam. The overall degree of polarization and the EVPA distributions agree well between the two images.

Zoom In Zoom Out Reset image size

Of the several jet components, C0 shows the lowest fractional polarization of ∼5%, as measured from the eht-imaging map. This further supports this component being the jet core, which is usually depolarized (e.g., Lister & Homan 2005). C1 exhibits a high level of polarization (∼16%), which indicates that the magnetic field is more ordered in this region. C2 shows a conspicuous polarimetric structure, which can be further divided into two subcomponents, with the EVPAs lying perpendicular to each other. The EVPAs in the upper subcomponent also lie perpendicular to the direction along which the brightness extends, while in the bottom subcomponent, they lie nearly in parallel. The degrees of polarization are ∼7% and ∼13% in the upper and lower subcomponents, respectively. These substructures are clearly seen in both the eht-imaging and CLEAN images.

Our GMVA + ALMA observations have revealed a remarkable structure of the inner jet of OJ 287 because of the improved (u,v)-coverage and high sensitivity. In particular, the C2 component shows a complex conical structure in both total and linearly polarized intensity, and a bimodal distribution in the EVPAs. Previous multi-epoch observations have shown that this component is nearly stationary (e.g., Hodgson et al. 2017; Jorstad et al. 2017; Lico et al. 2022). In the following, we discuss the possible nature of this component.

Oblique shocks could result from the jet striking a cloud of interstellar media. Under the precessing jet model, this would naturally occur for some period, as the jet sweeps through the ambient material. Since the location of C2 coincides with where the jet bends, the northeastern section of C2 could be interpreted as an oblique shock on one side of the jet. The oblique shock is in a plane making a small angle to the jet boundary on the northeast side. The flow is then bent by the shock toward the west. The magnetic field could become compressed to strengthen the component nearly parallel to the jet. Therefore, the EVPAs on the northeast side are roughly perpendicular to the jet. The southwestern section of C2 could then just be the main jet after the bend, with the magnetic field transverse to the jet direction at that point, as usual for a BL Lac object.

Conical shock waves can be formed when there is a pressure imbalance between the jet plasma and the ambient medium. The properties of shocks in relativistic jets have been explored by numerical and semidynamical simulations. Gomez et al. (1995) carried out relativistic hydrodynamics simulations of a parsec-scale jet surrounded by an ambient medium with constant or decreasing pressure. The simulations confirmed the existence of stationary components associated with recollimation shocks. Gómez et al. (1997) simulated the interactions of standing shocks and relativistically moving perturbations propagating down the stable jet, and found that the shock could enhance the emission of the moving feature and the stationary component could be temporarily "dragged" downstream. Further simulations of the interactions between recollimation shocks and traveling shocks are presented in Fromm et al. (2016), based on the observations presented in Fromm et al. (2011, 2013a, 2013b, 2015), for the particular case of CTA 102. Various configurations of the upstream magnetic field components have also been included in subsequent numerical simulations (e.g., Broderick & McKinney 2010; Porth et al. 2011; Fuentes et al. 2018). In particular, Mizuno et al. (2015) studied kinematically dominated jets with different magnetic field configurations, including axial, toroidal, and helical, based on a relativistic magnetohydrodynamics (RMHD) simulation code. Fuentes et al. (2018) characterized the properties of recollimation shocks in RMHD simulations of jets at the parsec scale as a function of the dominant type of energy: internal, kinetic, or magnetic. By solving the radiative transfer equations for synchrothron radiation, using these simulations as inputs, they analyzed the total intensity and linear polarization signatures imprinted in the stationary components associated with these shocks. Fuentes et al. (2021) extended the analysis to RMHD jet models threaded by helical magnetic fields with larger magnetic pitch angles, and also explored the effect of different nonthermal particle populations on the polarimetric properties of stationary features and the overall observed synchrotron radiation.

On the other hand, Cawthorne & Cobb (1990) established a semidynamical model, assuming only that the shock front is emitting, and found that conical shock waves could result in polarization angles that were either parallel or perpendicular to the jet axis. This model also only considered random magnetic fields in the upstream jet. In Cawthorne (2006), a poloidal magnetic field component was added to the model, and the results can well explain the observed polarization of the knot K1 in 3C 380. Furthermore, Cawthorne et al. (2013) extended this model to include a paired collimating and decollimating shock, and the predicted EVPAs could successfully describe the observational results of the BL Lac object 1803 + 784.

Comparing our observational results of C2 with the numerical and semidynamical studies, we find that the conical shape of the emitting region is quite consistent between our observations and the simulations. Numerical simulations predict a series of stationary shocks along the jet that can be triggered by a pressure imbalance between the jet and the external medium. The reason that we find only one conical-shaped component is most likely the adiabatic expansion of the jet. As also shown in Gomez et al. (1995), with decreasing pressure downstream of the jet, the intensity of the stationary components gradually decreases and the separation between the components increases, so the downstream shocks may be too faint and become undetectable at our observing frequency. Regarding the polarized emission, the semidynamical simulations show different EVPA distributions across the cone. However, the EVPA pattern is more symmetric with respect to the cone axis. The numerical simulations also show that the EVPA pattern will depend on the upstream magnetic field configuration and the viewing angle (e.g., Mizuno et al. 2015; Gómez et al. 2016; Fuentes et al. 2021). Fuentes et al. (2021) pointed out that jets with a large magnetic pitch angle, i.e., threaded by a helical magnetic field dominated by its toroidal component, can exhibit a bimodal EVPA distribution around recollimation shocks for small viewing angles. This EVPA configuration could imply a sign flip of the Stokes ${ \mathcal Q }$ parameter that leads to an EVPA flip, which then results in a dip in the linearly polarized emission, as we observe in the C2 component from the reconstructed polarimetric images.

Alternative to the standing shock scenario, the observed properties of the C2 component could be a result of geometric effects due to the bending of the jet axis toward the line of sight. With a decreasing viewing angle, the enhanced Doppler boosting could amplify the emission in this region and make C2 the brightest component in the inner jet. If the viewing angle becomes smaller than the jet opening angle, the bimodal distribution of the EVPAs could be produced by the existence of helical magnetic fields in the jet, as the direction of the projected magnetic field is different across the component (Fuentes et al. 2021). This scenario is supported by previous observations that revealed the existence of bending around C2 (e.g., Hodgson et al. 2017; Jorstad et al. 2017; Gómez et al. 2022). However, it is difficult to explain the conical shape of the emission region with this assumption.

Moreover, by means of multi-epoch GMVA observations, Lico et al. (2022) identified a new jet feature in the region of C2, in a quasi-concurrent GMVA observing epoch. The authors argued that the passage of this new jet component through the stationary feature at 0.1 mas core separation (i.e., C1) triggered the high-energy outburst during 2016–2017 (Komossa et al. 2017, 2021a), including the faint VHE flare detected in February 2017 (Mukherjee & VERITAS Collaboration 2017) ³⁸ and then moved down to the C2 jet region at the time of these observations. In this scenario, the C2 component in our observations could correspond to the blending of the new feature and the standing shock. The observed bimodal distribution of the EVPAs could be due to the different polarimetric properties of the two components. A similar case was found in the core region of PKS 1510-089 during a γがんま-ray flare in 2015 (Park et al. 2019).

OJ 287 is one of the most promising candidates for harboring an SMBBH system at the center. In fact, OJ 287 is among the candidates for hosting a nano-Hz gravitational wave–emitting SMBBH system (Valtonen et al. 2021). The binary model has been successful in explaining the periodic light curves and predicting upcoming impact flares, which have been confirmed by observations within a few hours (e.g., Laine et al. 2020). The direction of the jet axis was also found to vary with time, and this could also be related to the orbital motion of the BHs (Dey et al. 2021). Models that do not require a secondary BH to explain the observed variability have also been proposed. For instance, the flux variation could be explained by viewing angle changes and the Doppler beaming effects of a precessing jet. The precession could be driven by the Lense–Thirring effect, due to misalignment between the BH spin and the accretion disk (e.g., Britzen et al. 2018; Liska et al. 2018; Chatterjee et al. 2020; Liska et al. 2021). It would also be possible for MHD instabilities (current-driven or Kelvin–Helmholtz) to produce a helical distorted jet structure (e.g., Mizuno et al. 2012; Perucho et al. 2012; Vega-García et al. 2019).

Dey et al. (2021) established a model to explain the parsec-scale jet direction variations at different frequencies, in which the jet precession is powered by the SMBBH with parameters constrained by optical observations. This model predicts that the 86 GHz jet axis should be ∼−37° around 2017 April, assuming a disk model. The PAs of the inner jet components (e.g., C1 and C2a) measured in our GMVA + ALMA observations agree well with this prediction (see Table 1). However, we note that this agreement is partially due to the observing epoch being not far apart from the 86 GHz GMVA data used to constrain the model. Furthermore, this agreement will not rule out other possible scenarios. For example, the tilted accretion could also result in the precession of the inner jet. Britzen et al. (2018) argue that the PA change observed at 15 GHz can be modeled by a jet precession combined with a nutation of the axis. The precession could be a result of Lense–Thirring effects, and a secondary BH is not always required. Furthermore, our RML images also revealed a twisted pattern of the innermost jet, which resembles a precessing jet in projection.

Future kinematic studies with multi-epoch GMVA and EHT observations will hopefully provide further insights to distinguish among different theoretical models regarding the underlying nature of the source.

Dey et al. (2021) also explored the possibility of the existence of a jet from the secondary SMBH based on the SMBBH model. With the high sensitivity and improved north–south resolution resulting from the participation of ALMA, we found no evidence for a secondary jet, even in the eht-imaging and SMILI images with super-resolution. There could be several possible reasons for such a nondetection. First, the jet is likely to be short-lived, as commented upon in Dey et al. (2021). Since the projected separation of the two SMBHs in 2017 April is ∼10 μみゅーas (Dey et al. 2018), the current image resolution is not sufficient to spatially resolve the binary system if there is no extended jet emission from the secondary SMBH. The same would apply if the secondary jet extends in a similar direction as the primary jet. If the secondary jet is present and points in a different direction, the nondetection implies that the brightness temperature of the jet must be lower than 4 × 10⁹ K, which corresponds to three times the r.m.s. level of the eht-imaging map. We note that the dynamic range of our image reconstruction is much higher than the mass ratio of the two BHs.

We further note that the GMVA + ALMA observations presented in this work are part of a multiwavelength observing campaign of OJ 287. Close-in-time observations with the EHT at 230 GHz (on 2017 April 4 and 9) and with the RadioAstron space-VLBI mission at 22 GHz (on 2017 March 7) could provide even higher angular resolutions and probe slightly different regions of the inner jet. Together with the observations at X-ray and optical bands (e.g., Komossa et al. 2017, 2020, 2021a, 2021b, 2021c), we will be able to test or obtain constraints on the physical parameters of the possible jet associated with the secondary SMBH.