The BL Lacertae object OJ 287 is an emblematic nearby active galactic nucleus (AGN) with a redshift of z = 0.306 (Stickel et al. 1989), well known for its 12-year quasi-periodic outbursts in the optical regime (e.g., Sillanpää et al. 1988;Villata et al. 1998). These quasi-periodic variations have been interpreted as evidence of a supermassive black hole binary (SMBHB) system where the secondary black hole, in a highly eccentric orbit around the central black hole, modulates the jet emission as it interacts with the primary's accretion disk. This model was used to explain the quasi-periodic variability (Lehto & Valtonen 1996;Valtaoja et al. 2000;Valtonen et al. 2008), which aligns with theoretical models of an accreting SMBHB based on general relativistic magnetohydrodynamics simulations, which account for dynamic spacetime effects in accreting binary systems (Farris et al. 2012;Gold et al. 2014a,b;Gold 2019;Paschalidis et al. 2021). Supporting the binary scenario, Britzen et al. (2018) identified a 23-year jet precession period that was later corroborated by Britzen et al. (2023), who linked spectral energy distribution (SED) states to the precession phase. On the other hand, the presence of an ultra-massive primary black hole has been put into question by multi-frequency observations and a lack of predicted outbursts (Komossa et al. 2023a,b) in the years 2021 and 2022. The authors raising these doubts showed that the data favor more periodic outbursts with a period of 11.5 ± 1 yr, most recently observed in 2016-2017, and they estimated the mass of the primary to be 10 8 M ⊙ .
A23, page 2 of 24 Variations in the jet position angle can also be explained by alternative scenarios that do not require a SMBHB system. For instance, the precession of a single misaligned accretion disk around a single central supermassive black hole (SMBH) (Mizuno et al. 2012), or a warped accretion disk, not perfectly aligned with the black hole spin axis (Liska et al. 2018). Beyond these mechanisms, several internal processes can generate similar observational signatures in AGN jets. For example, jet instabilities play a fundamental role in AGN jet phenomenology.
Two types of instabilities are mainly encountered in AGN jets: Kelvin-Helmholtz (K-H) instabilities and current-driven instabilities (CDI). The K-H instabilities, which arise from velocity shear between the jet and the ambient medium, can develop in kinetically dominated jets (e.g., Perucho et al. 2004;Hardee 2007). These instabilities can generate helical perturbations that manifest as twisted structures when projected onto the plane of the sky (Perucho et al. 2012;Vega-García et al. 2019). Additionally, CDI kink instabilities can develop in strongly magnetized jets (e.g., Nakamura et al. 2007;Mizuno et al. 2012), further contributing to jet wiggling and bending. These various instability modes, coupled with recollimation shocks and magnetic field compression, can drive internal shocks and turbulence, potentially explaining the observed variability in AGNs (Marscher 2014;Jorstad et al. 2022) and playing a crucial role in particle acceleration mechanisms (Sironi et al. 2015). In fact, OJ 287 constitutes an ideal laboratory for investigating particle acceleration mechanisms, as it has an emission spectrum stretched up to teraelectron-volt energies (e.g., Mukherjee & VERITAS Collaboration 2017;Lico et al. 2022). It also provides a critical platform of testing the validity of different launching scenarios of AGN jets, as theoretical models suggest that relativistic jets are produced by accreting SMBH and driven by their dynamically important magnetic fields, which can be twisted by the ergosphere (Blandford & Znajek 1977) or by the differential rotation of the black hole's accretion disk (Blandford & Payne 1982).
Very long baseline interferometry (VLBI) observations at the highest possible angular resolution are the ideal method for probing the innermost regions of AGN jets. This can be achieved by either increasing the observing frequency or extending the baselines to include space-based antennas. Indeed, with an apogee of approximately 350 000 km, space VLBI observations with RadioAstron have been capable of imaging blazar jets with unprecedented resolution, on the order of a few tens of microarcseconds (e.g., Gómez et al. 2016;Fuentes et al. 2023). During its operation, RadioAstron observed OJ 287 on several epochs, yielding the highest angular resolution image obtained for this source (Gómez et al. 2022). Most recently, space-based VLBI imaging of OJ 287 with RadioAstron at 22 GHz, together with multi-epoch Very Long Baseline Array (VLBA) observations at 43 GHz, revealed a ribbon-like inner-jet morphology and multiyear swings of the jet position angle that are consistent with a rotating helical jet structure on parsec scales (Traianou et al. 2025).
Event Horizon Telescope (EHT) observations at 1.3 mm have also significantly advanced our understanding of SMBHs and their relativistic jets, culminating in the groundbreaking capture of the first images of a black hole in M87 (Event Horizon Telescope Collaboration 2019a,b,c,d,e,f, 2021a,b, 2023), hereafter M87* Papers I-IX) and Sgr A * (Event Horizon Telescope Collaboration 2022a,b,c,d,e,f, 2024a,b, hereafter Sgr A * Papers I-VIII) and demonstrating that black holes with masses ranging from mil-lions to billions of solar masses can be consistently described by the Kerr metric. Building on this foundational work, EHT observations of a number of AGNs have provided a crucial window into the physics of jet launching and initial collimation in extragalactic radio jets at scales down to 10-100 gravitational radii, encompassing the processes of jet launch and its initial collimation (e.g., Kim et al. 2020;Janssen et al. 2021;Issaoun et al. 2022;Jorstad et al. 2023;Paraschos et al. 2024;Baczko et al. 2024;Röder et al. 2025).
In April 2017, we conducted the inaugural 1.3 mm VLBI observations of OJ 287 with the EHT in order to probe its structure at scales corresponding to the hypothesized presence of a SMBHB system. These observations were part of an extensive multiwavelength campaign including additional longer wavelength VLBI observations from both ground-and space-based facilities such as RadioAstron (Gómez et al. 2022) and the Global Millimeter VLBI Array (GMVA; Zhao et al. 2022) and alongside observations in optical, UV, and X-ray wavebands (e.g., Komossa et al. 2021a;EHT MWL Science Working Group et al. 2021, on OJ 287 and M 87, respectively). Concurrently, the independent Multiwavelength Observations and Modelling of OJ 287 (MOMO) project, which started in 2015, has provided a framework for these integrated studies (e.g. Komossa et al. 2017Komossa et al. , 2021bKomossa et al. , 2023a)). MOMO provides high-cadence optical, UV, X-ray, and MWL single-dish radio observations and their interpretation. OJ 287 has also been regularly monitored with the VLBA at 43 GHz and 15 GHz for over two decades as part of the BEAM-MEfoot_0 (Jorstad & Marscher 2016;Jorstad et al. 2017;Weaver et al. 2022) and MOJAVEfoot_1 (Lister et al. 2018) monitoring programs, respectively.
This work is organized as follows: In Section 2, we describe the EHT observations, data reduction, and imaging techniques used to reconstruct the total intensity and polarization structure of the source. Section 3 presents the results, detailing the detected variability in total intensity and polarization over a fiveday timescale, the characterization of jet features, and the measured apparent motions. In Section 4, we discuss the implications of these findings in the context of relativistic jet physics, focusing on the role of K-H instabilities, shocks, and a helical magnetic field in explaining the observed variability. We introduce a model that accounts for the rapid polarization swings and outline how future observations with improved time sampling could further test these predictions. In Section 5, we provide a summary of our results. We note that for all the calculations, we adopt a flat ΛCDM cosmology with H 0 = 67.4 km s -1 Mpc -1 , Ω m = 0.315, and Ω Λ = 0.685 (Planck Collaboration VI 2020). At the redshift of OJ 287, this corresponds to a luminosity distance of 1.642 Gpc, an angular scale of 4.65 pc mas -1 , and an apparent speed of 7.25 c for a proper motion of 1 µas day -1 .
The EHT observed OJ 287 on three nights during the 2017 campaign on April 5, 10, and 11. Two of those days, April 5 and 10, offer sufficient (u, v) coverage to fully model the source structure in total intensity and in linear polarization. Additionally, the coverage is similar enough to reliably compare results between days (see Figure 1). The source was observed with the array consisting of seven telescopes located at five geographic sites: the
We now apply the previously described model (see also Figure 11) to analyze the time-variable structure observed in OJ 287, focusing particularly on the evolving EVPAs of its components. One of the most restrictive observational constraints is that components P1 and P2 exhibit EVPA rotations in opposite directions-counterclockwise for P1 and clockwise for P2.
When computing the observed synchrotron polarization, it is necessary to account for the characteristic swing in the polarization angle, as discussed previously in the seminal work by Blandford & Königl (1979) (see also Lyutikov et al. 2003). For a magnetic field, B, in the source frame, whose orientation is specified by the angles η and ψ, as shown in Figure 12, the observed polarization angle, ξ, is given by (Blandford & Königl 1979;Lyutikov et al. 2003)
where βc is the velocity of the plasma and θ is the angle between β and the observer's direction, n. Note that ξ is measured from the projected jet axis and is positive in the clockwise direction. We could then compute how the polarization angle, ξ, evolves over time in our model. To do this, we considered a cylindrical jet (neglecting a small opening angle for simplicity) threaded by a helical magnetic field with a given pitch angle, α b , defined as the angle between the magnetic field and the jet axis. The helical magnetic field, measured in the source frame and normalized to unit strength, is expressed in Cartesian coordinates as
where φ is the phase of the helical field.
To determine how φ evolves over time, we consider a helical K-H instability propagating along the jet surface with a projected wavelength of θ λ ∼ 100 µas, as estimated previously. For simplicity, we neglect the rotation of the K-H instability over the 5day time span of our observations. A plane-perpendicular shock travels along the jet with an apparent velocity β app . The interaction of these moving shocks with the K-H instabilities gives rise to components P1 and P2, as observed (see Figure 11), and determines the evolution of φ and, consequently, the observed polarization angle ξ over time.
The phase of the helical magnetic field, φ, as a function of the observed time, t obs , is given by
where φ 0 is an arbitrary initial phase. Substituting this phase into the helical magnetic field expression (Equation ( 4)). Using Equation (3), where η = arcsin(-B y ) and ψ = arctan(B x /B z ), we could obtain the observed polarization angle as a function of time.
Figure 12 shows the simulated time evolution of the polarization angle for components P1 and P2, assuming that the jet in OJ 287 is threaded by a predominantly toroidal magnetic field and is observed at an angle of approximately 1/Γ, with Γ being the jet's bulk flow Lorentz factor. Due to light aberration, this viewing angle makes the line of sight nearly perpendicular to the jet axis in the plasma frame, leading to EVPAs that are more closely aligned with the jet axis for a predominantly toroidal magnetic field. (See the figure caption for the specific values used in our simulation.) By examining Figure 12, we observe that, by assuming different initial phases in the magnetic field probed by each component, our model qualitatively reproduces the most salient observational feature: the opposite-direction rotation of the polarization angle in components P1 and P2. This effect occurs when the line of sight in the plasma frame exceeds the pitch angle of the magnetic field. Otherwise, a progressive evolution of the EVPAs is expected as the interaction of the component with the helical instability probes different phases of the helical magnetic field. As discussed in Blandford & Königl (1979), a rapid swing in the polarization angle is expected when β ∼ (cos θsin θ tan ψ), which can be tested in the future with more continuous and densely time sampled monitoring of the source.
Although our model can roughly reproduce the observed evolution of the polarization angle in P1 and P2, we note that it is not intended to exactly match the specific values of the observed EVPAs, as any combination of the Lorentz factor and viewing angle satisfying θ = 1/Γ would lead to similar rotations. Additionally, we have made several simplifications, such as assuming a perfectly straight jet with a perfectly helical instability, which is unlikely to hold in reality. Moreover, the model naturally accounts for the more rapid polarization rotation observed in the faster moving component.
Alternative models have explained blazar variability as the result of orientation changes of emitting regions in a twisting jet, which modulate the Doppler factor over time (e.g., Raiteri et al. 2017). In contrast, in our case the jet is straight and viewed at a constant angle, and the observed variability can be explained without invoking time-dependent Doppler factors. Instead, the EVPA rotations and apparent non-ballistic motions arise naturally from the interaction between plane perpendicular shocks and a helical Kelvin-Helmholtz instability in a jet threaded by a helical magnetic field.
On parsec scales, space-VLBI imaging with RadioAstron at 22 GHz, combined with multi-epoch VLBA observations at 43 GHz, revealed a ribbon-like inner jet in OJ 287 and multi-year swings of the jet position angle, consistent with a rotating helical Kelvin-Helmholtz pattern (Traianou et al. 2025), as envisioned in our model. Our EHT observations probe the jet much closer to its origin with tens of µas resolution at 230 GHz and follow its evolution over only a few days in April 2017. Together these data sets provide complementary perspectives, with RadioAstron and the VLBA tracing the parsec-scale multi-year evolution and the EHT capturing the rapid day-scale dynamics at the innermost jet base.
Previous observations of rapid optical polarization rotations in BL Lac by Marscher et al. (2008) inferred the presence of a helical magnetic field in the jet's collimation and acceleration zone. However, the lack of angular resolution prevented a direct imaging of this effect, leaving its interpretation dependent on time-resolved polarization variability. Similarly, Cohen et al. (2018) and Cohen & Savolainen (2020) reported changes in the direction of EVPA rotation in integrated polarization measurements of OJ 287. Their model attributes these changes to the superposition of a variable polarized component on top of a steady jet component, which can produce apparent rotations in opposite directions when combined. In contrast, our EHT observations at an angular resolution of 18 µas for the first time is capable of directly resolving two distinct jet components, each exhibiting EVPA rotations in opposite directions. This spatially resolved evidence rules out the model based on integrated polarization components and provides direct support for the existence of a helical magnetic field in the jet's collimation and acceleration zone. The ability to directly resolve these structures with A23, page 14 of 24 EHT's unprecedented angular resolution offers new constraints on jet formation, magnetic field evolution, and the role of instabilities in shaping AGN jets at their earliest stages.
We report on the first 1.3 mm observations of the candidate SMBH OJ 287 conducted with the EHT, which achieved an exceptional angular resolution of 18 µas. These observations reveal notable changes in both structural and polarization properties within just five days-the shortest interval over which such variability has been spatially resolved in this source.
The total intensity images reveal a twisted ridgeline structure extending northwest, with jet features displaying apparent superluminal motions reaching velocities up to ∼22 c. These rapid structural variations are also clearly visible in the evolving morphology of the jet ridgeline observed between the two epochs.
Polarimetric imaging identifies three prominent polarized components within the jet, and they show significant evolution in their EVPAs across the five-day interval. One component, located approximately 200 µas northwest from the core, exhibits a radial polarization signature consistent with a recollimation shock.
The observations directly resolve two innermost jet components exhibiting opposite directions of EVPA rotation: the faster-moving component (∼17.4 c) experiences counterclockwise rotations at ∼3.7 • per day, whereas the slower-moving component (∼10.2 c) rotates clockwise at ∼2.5 • per day. This result provides the first spatially resolved confirmation of a helical magnetic field threading the jet's collimation and acceleration zones, which were previously inferred but not directly imaged in AGN jets.
The observed polarization variability is best explained by propagating shocks interacting with K-H instabilities, illuminating different phases of a threaded helical magnetic field. This interpretation rules out scenarios based solely on the blending of unresolved polarized components, emphasizing the crucial role of ordered magnetic field structures in driving the observed EVPA swings.
In addition, these millimeter-wavelength observations establish a compelling link to the historically observed highly variable optical polarization behavior in OJ 287. They suggest a common physical mechanism driving polarization variability across different wavelengths.
https://www.bu.edu/blazars/BEAM-ME.html
https://www.cv.nrao.edu/MOJAVE