Simultaneous Optical-to-X-Ray Spectrum of OJ 287 During Lowest X-Ray State: Synchrotron-Soft Tail and Harder X-Ray Spectrum ^†

Abstract

The X-ray spectrum of OJ 287 has exhibited diverse variations with broadband spectral behavior representative of all the spectral classes of blazars. These changes have been explained either via new emission components or as the sum of the jet synchrotron and its inverse Compton part. In the current work, we focus on the systematic spectral investigation of the lowest X-ray state recorded by the Swift facility to understand X-ray spectral changes. Considering optical-to-X-ray observations jointly, we found a power-law optical–UV spectrum with a photon spectrum of 2.71 ± 0.03 extending to X-ray energies. Accounting for this contribution in X-rays, we inferred a power-law photon X-ray spectrum of 1.22 ± 0.20 that improves to 1.29 ± 0.06 when considering other observations with similar X-ray spectra. An extended optical–UV spectrum with an associated low hard X-ray spectrum is further strengthened by the natural explanation of another optical–UV state of similar flux with a very different optical–UV-to-X-ray spectrum by its synchrotron and this hard X-ray spectrum. This is the hardest reported X-ray spectrum (0.3–10 keV), consistent with the Swift-BAT X-ray spectrum. We further found that this X-ray spectrum can reproduce most of the flat X-ray spectra when combined with the corresponding optical–UV continuum during the low and intermediate flux states, strengthening the synchrotron as the primary driver of most of the X-ray spectral changes in the LBL state of the source. Compared with the sharp steepening/cutoff of the optical–UV spectrum during bright phases, the inferred extended spectrum implies a comparatively larger emission region and could be associated with large-scale jet emission. The optical–UV spectrum implies a high-energy power-law particle spectrum of ∼4.4, while X-ray implies a hard low-energy particle spectrum of 1.3–1.6 that alternatively can also result from a higher lower-energy cutoff in the particle spectrum.

1. Introduction

OJ 287 is an optically bright BL Lacertae (BLL) type object at a cosmological redshift of z = 0.306, characterized by a non-thermal continuum-dominated optical spectrum with very weak emission line features reported only during its very faint optical brightness states [1,2,3]. Together with flat spectrum radio quasars (FSRQs), they are called blazars—active galactic nuclei hosting large-scale powerful relativistic jets directed roughly towards the Earth. Blazars are characterized by a highly variable continuum observed in the entire electromagnetic (EM) spectrum from radio to GeV-TeV gamma rays with a characteristic broad, bi-modal spectral energy distribution ($\nu F_{\nu }$ vs. $\nu$; e.g., [4,5,6]). The temporal continuum variation is primarily erratic and variable at all times, from decades or longer to minutes, accessible within the sensitivities of currently existing facilities (e.g., [7]).

The low-energy emission component of the bi-modal broadband SED extends from radio up to X-rays, peaking in between infrared (IR) and X-ray energies is widely accepted to be the synchrotron emission from relativistic electrons within the jet. This is strongly favored by the observed behavior, e.g., the non-thermal nature of the radio and optical spectra and strong and variable polarization (e.g., [8,9,10,11,12]). The origin of the high-energy part is contentious, proposed to be either via inverse Comptonization (IC)—a natural successor given relativistic electrons and prominent soft photon fields comprising accretion disk photons, broad emission line region photons (BLR), IR torus photons, etc.—or hadronic processes or a combination of both (e.g., [13,14] and references therein). Claims of the detection of neutrinos from the direction of a few of these sources support hadronic processes, and modeling implies a sub-dominant hadronic contribution at MeV-GeV energies, bounded basically by the detected neutrinos’ flux [13,14].

A remarkable apparent feature of the bi-modal broadband SED is that the frequency at which the low-energy emission component peaks rarely changes despite the observed strong flux variations, most of which are often accompanied by a change in the continuum spectrum in different observational bands. This stability combined with the physical understanding of the low-energy part of the SED has led to a physical process-based classification of blazars into low- (LSP: LBL + FSRQs), intermediate- (ISP/IBL), and high-synchrotron peaked blazars (HSP/HBL; [4,15]).

The inferred dominance of the leptonic component, supported also by the modeling of candidate neutrino blazars, e.g., the work of Gao et al. [13], implies a highly correlated flux variability as well as spectral properties between the EM bands constituting the two humps in the SED. Thus, for simultaneous continuum variation indicating emission from the same region, the spectrum of the low-energy hump is related to that of the high-energy hump. For LBLs/LSPs, to which OJ 287 belongs (e.g., [4]), this implies a direct connection between the optical–UV (synchrotron) and the MeV–GeV gamma-ray spectra [16,17]. The former, being purely of synchrotron origin, provides a direct tracer of the underlying particle spectrum. Combining this with the simultaneous X-ray spectrum additionally allows us to explore the extent of the high-energy part of the particle spectrum. This makes the simultaneous optical-to-X-ray spectrum an excellent observable to probe the high-energy particle spectrum, free from any artifact/complications that could arise if using gamma-ray emission, e.g., multiple IC fields contributing to the high-energy hump, steepening introduced by the onset of the Klein–Nishina regime, extra-galactic background light, etc. The particle spectrum, being the fundamental entity of non-thermal processes, has important implications beyond blazars’ high energy emission or emission regions, e.g., cosmic rays, astro-particle physics, etc.

Among blazars, OJ 287 has been one of the best-monitored sources at radio and optical wavelengths since its discovery due to its frequent variability and brightness in radio and optical bands ([18] and references therein). Subsequently, following the claim of it being a binary supermassive black hole (SMBH) system based on apparent recurring optical outbursts every ∼12 years [19], it became a target of coordinated monitoring. The first turn-out of optical flares, as per the prediction, strengthened further monitoring. Radio observations also indicate a periodicity of ∼22 years and instead argue the recurring flares are from a precessing jet (e.g., [20,21] and references therein), though studies also argue otherwise (e.g., [22]). Overall, both interpretations still have many issues. The former claims these outbursts to be thermal, but the explanation of observed polarization (e.g., [23]) and optical–UV spectral features do not support this [18,24], nor the lack of the most recent expected flare (e.g., [25], but see [26,27]). In the latter interpretation, flares at different bands are expected to be achromatic with broadly similar temporal profiles [3,20,21], contrary to observations. In general, the characteristic feature of jet emission is erratic rapid and strong flux variations which are normally believed to be due to shocks, etc. ([28], and references therein). These findings and outcomes have made OJ 287 a potential astrophysical laboratory not only for understanding and exploring jet and possibly accretion physics but also far nano-Hertz gravitational waves [29], general relativity and BHs [29,30], tidal disruption events (TDE; [3]), etc.

In terms of spectral features of OJ 287, X-ray has been the most dynamic band. Apart from many distinct spectral features reported in OJ 287 that have been argued to be associated with new components like a thermal-emission-like feature consistent with ∼10¹⁰ M_⊙ [18,31], a super soft X-ray spectral state [32,33] with HSP/HBL-like SED [34] implying jet emission that has been argued to be due to shocks by Kapanadze et al. [28] while Huang et al. [3] contends it as a TDE, a likely iron absorption line feature [33], a thermal-like soft X-ray excess [35], etc. Apart from these clear spectral changes indicating additional emission components, spectral changes have been reported in its normal LSP states, which have been argued to be likely due to new emission components [36]. However, many studies have demonstrated that the optical–UV synchrotron spectrum can significantly modify the X-ray spectrum (e.g., [37,38]). Thus, simultaneous optical–UV and X-ray monitoring are essential to probe the driver or cause behind the X-ray spectral changes.

Its frequent and strong activities, even at X-rays, have made OJ 287 one of the best-monitored blazars by the transient facility: The Neil Gehrels Swift Observatory that autonomously tracks sudden bright activity. The facility is currently the best facility for simultaneous1 optical to X-ray coverage of broadband transient systems. These observations have been investigated extensively to understand OJ 287 behavior. For example, over a decade timescale, Siejkowski and Wierzcholska [36] reported that OJ 287 is more variable at optical–UV than at X-rays, and optical–UV spectral shape does not exhibit any apparent trend with the optical–UV brightness. X-ray, on the other hand, shows a harder-when-brighter trend in general. The study reported an apparent anti-correlation between spectral shape at X-ray and optical–UV, and also the observation of a flat X-ray spectrum for the first time, arguing it to be either the sum of synchrotron and IC or a new emission component, likely of hadronic origin. More recent studies with dense coverage under the MOMO [25,32,33] project have reported strong X-ray spectral changes, with the focus mostly on flux/brightness variability and overall spectral trends (see also [3,18,28]) while joint consideration is mostly for broadband SEDs (e.g., [31,34,38,39,40,41]). Overall, almost all these studies exploring spectral variations have focused on optical–UV and X-ray separately, or, if considered jointly, only a few observations have been considered mainly in the context of broadband SEDs (e.g., [38]) with gamma-ray spectra from a much longer duration.

In the current work, we focus on a detailed spectral investigation of the simultaneous optical to X-ray spectrum of the observed lowest X-ray flux state of OJ 287 by the Swift observatory. The work is organized as follows. In Section 2, we present the details of the data and reduction procedure. Section 3 presents the analysis and results reporting an optical–UV synchrotron spectrum extending to X-ray energies and the hardest X-ray spectrum of the source during the lowest X-ray flux state. Section 4 presents the implications of this on particle acceleration scenarios and emission mechanisms, followed by a summary in Section 5.

2. Data Reduction

The Neil Gehrels Swift Observatory [42] is a space-based facility with three primary payloads— the Ultraviolet Optical Telescope (UVOT; 2.3–6.1 eV), X-ray Telescope (XRT; 0.3–10 keV), and the Burst Alert Telescope (BAT; 15–150 keV) capable of simultaneously observing optical to hard X-ray band over a wide range of source brightness states. In this work, we have used all the available pointed XRT and UVOT data of OJ 2872 between MJD 53510 and MJD 59940. The dense Swift follow-up since mid-2015 has been carried out by Komossa et al. [32,44]3.

UVOT: The UVOT data of the source were reduced following the methodology adopted by Kushwaha et al. [39]. We used a $5^{″}$ circular region for the source and an annular source-free region of inner and outer radii of ${25}^{″}$ and ${45}^{″}$ respectively for the background, both centered on the source location. We then executed the HEASOFT (v6.30.1) tasks uvotsum and uvotsource to extract the flux density in each filter. The extracted flux densities were then corrected for reddening using an $E(B-V)=0.0241$ with the extinction law of Cardelli et al. [45].

XRT: For the X-ray spectral study, we used the scientific grade spectra and respective ancillary files provided by the UKSSDC4. The underlying pipeline corrects for pile-up in both photon counting (PC) and window timing (WT) mode data following the procedure detailed in Evans et al. [46] and subsequent ongoing updates. For OJ 287, none of the WT mode data require pile-up treatment (rate $<<100$ c/s), while PC mode data having a count rate ≳ 0.5 c/s were marked for pile-up correction.

To find the best description of the X-ray spectrum, we followed the approach adopted by Kushwaha et al. [34,39] using F-test statistics to choose the best-fit description between the absorption (tbabs) modified power-law (PL; $dN/dE\sim E^{-\mathsf{\Gamma }}$) and log-parabola (LP; $dN/dE\sim E^{-\alpha -\beta log\left(E\right)}$) models. We used a F-test value of 0.01, below which LP is preferred and vice-versa. We binned each source spectral file with a minimum of one count per bin and performed the model fitting (0.3–10 keV) using the $Cash-statistics$ with background (W Statistic in $XSPEC$)—the most widely used statistics in the photon starved high energy astrophysics and recently demonstrated to have the capability of a goodness-of-fit indicator [47]. During the model fitting, we kept all the parameters free initially but froze the nH value to the Galactic column density of $2.4\times {10}^{20}$ cm⁻² [48] in the direction of the source whenever the fit value was equal or below it. Finally, for spectral study, we corrected all the extracted X-ray SEDs for the nH absorption.

3. SED Analysis and Results

The optical to X-ray light curve and the respective spectral behavior extracted from the Swift facility are shown in Figure 1. The UVOT spectral index is extracted assuming a PL model fitted to the reddening corrected flux data in SHERPA [49]. The source has undergone a strong flux variation of >100 in the X-ray band (0.3–10 keV) and ≳15 in UVOT bands between the minimum and maximum during this period. Overall, in terms of flux variations, cross-correlation, which provides a statistical quantification of variation between two different time series, shows that the optical–UV and X-rays variations are simultaneous with a peak at zero lag. An interesting behavior is the concurrent low-flux state in optical–UV and X-ray, marked in the cyan-colored band in Figure 1 (MJD ∼ 54,160–54,180), with the lowest X-ray flux within the cadence, statistics, and instrument sensitivity. Later, though optical–UV has even gone below this level, the X-ray has not. Another point to be noted is that there are fewer X-ray points with the LP model in Figure 1 compared to our previous works, (e.g., Kushwaha [6]). The difference is due to the F-test probability value used −0.01 here compared to 0.05 in our earlier works [6,34]. We chose a tighter value because the low-flux states—the focus of our study, favor the PL5 model.

Since our focus is the lowest X-ray flux state and synchrotron emission has been shown to affect the X-ray spectrum [35,37,38], we ordered the Obs-IDs6 first by X-ray flux (increasing) and then filtered all the Obs-IDs with best-fit X-ray flux within $3\sigma$ of the lowest X-ray flux (resulting in 4 Obs-IDs). We then looked for the Obs-ID with the lowest flux in UVOT bands (Obs-ID: 00030901002).

As optical–UV is synchrotron emission, to get the spectral index, we first fitted a reddening modified PL model to the UVOT SED using the ${\chi }^2$ statistic in XSPEC. For this, we converted the image files to the respective spectrum files using the HEASOFT task uvot2pha and then performed the fitting. We then extrapolated the best–fit optical–UV spectrum and examined it vis-a-vis the corresponding X-ray spectrum, as shown in Kushwaha [50]. We found that the UVOT PL fit (${\mathsf{\Gamma }}_{UVOT}=2.71\pm 0.03$) can extend well into the X-rays. Thus, to get the uncontaminated X-ray spectral parameter, we jointly modeled the optical-to-X-ray spectrum with a two-PL model (redden × powerlaw + tbabs × powerlaw) with the reddening modified PL for optical–UV synchrotron contribution to X-ray energies. We found a harder X-ray photon spectral index of ${\mathsf{\Gamma }}_X=1.22\pm 0.20$7 compared to ${\mathsf{\Gamma }}_X\sim 1.5$ when fitting X-ray alone.

To further improve the constraint, we jointly modeled all the observation Obs-IDs (7 Obs-IDs8) having similar X-ray spectral indices (hard; ${\mathsf{\Gamma }}_X$: 1.2–1.3) in the joint UVOT+X-ray fitting. This resulted in an improved constraint of ${\mathsf{\Gamma }}_X=1.29\pm 0.06$. We re-checked the result using only ${\chi }^2$-statistic for optical–UV and X-ray with X-ray binned to a minimum of 20 counts per bin. We obtained a consistent result (within $1\sigma$; ${\mathsf{\Gamma }}_X=1.16\pm 0.08$; ${\chi }^2/dof$9 = 67.4/61), except that the lowest X-ray flux state has to be excluded due to insufficient counts. The best-fit model from the lowest X-ray flux state, along with data, is shown in Figure 2. This is the hardest-ever reported X-ray spectrum (0.3–10 keV) for OJ 287 to the best of our knowledge.

We then explored the optical to X-ray spectrum of other low and intermediate X-ray flux states, most of which have a relatively flatter X-ray spectrum ($\mathsf{\Gamma }$∼ 1.7–2.1), and found that this inferred hard X-ray spectrum without or with a variation of normalization along with the respective optical–UV spectrum can naturally explain a majority of the observed flat X-ray spectrum of the LBL phase of the source. A subset of such Obs-IDs, representative of the different X-ray spectra exhibited by the source, are shown in Figure 2 and Figure 3. We also found that a few are indeed different spectrally. These and the observation Obs-IDs requiring a change of normalization of the inferred hard X-ray spectrum are marked with ‘∗’ in Figure 3.

4. Discussion

We found an extended optical–UV synchrotron spectrum (${\mathsf{\Gamma }}_{UVOT}=2.71\pm 0.03$), continuing into X-rays during a period with a concurrent dip in optical–UV and X-ray flux (MJD: 54,160–54,180; cyan band in Figure 1), corresponding to the lowest recorded X-ray flux within the Swift cadence. Accounting for this contribution at X-rays, we found a hard X-ray spectrum with photon spectral index ${\mathsf{\Gamma }}_X\sim 1.15-1.3$ (0.3–10 keV; ref Figure 2), the hardest ever reported from the study of simultaneous optical to X-ray spectrum of OJ 287. This is in contrast to the bright/active state optical to X-ray SEDs, where the comparison of the PL extrapolation of the optical–UV synchrotron spectrum to the X-ray data generally requires a cutoff or sharp steepening of the optical–UV spectrum in the (far) UV region (e.g., [24,31,38]). This inference of an extended optical–UV spectrum during low-flux optical–UV states is further strongly supported by our finding of another optical–UV flux state having roughly similar flux in one of the UVOT bands (W1) but the associated optical–UV spectrum, as well as the X-ray spectrum, is significantly different (Figure 2; ID: 00030901123) with X-ray flux almost three times higher than the lowest one and the X-ray spectrum can be reproduced by the sum of the extrapolated optical–UV spectrum and the inferred hard X-ray spectrum.

The inferred hard X-ray spectrum is consistent with the spectral range reported in the hard X-ray band from the Swift-BAT data (50–300 keV; [43]). It should, however, be noted that the latter spectrum is derived by using data collected over almost 12 years (105 months). Thus, for highly variable sources like blazars, the BAT spectrum will include different spectral states, biased towards the most photon-contributing state. Further investigation of the spectral changes in our study reveals that a majority of the flat X-ray spectra during the low and intermediate X-ray flux states can be naturally explained by simply adding the inferred hard X-ray spectrum with the respective optical–UV spectrum either originally or with a variation of the normalization (refer to Figure 2 and Figure 3). A few cases, however, indeed have different spectrum (refer to Figure 3).

Blazar optical continuum is well-understood to be synchrotron, and thus, the observed spectrum directly traces the underlying particle spectrum. A similar relation, though a bit more involved, holds for the IC spectrum i.e., X-ray, if it is away from the peak and IC happens in the Thomson regime. For a power-law particle distribution of spectral index p ($\mathrm{N}\left(\mathrm{E}\right)\sim {\mathrm{E}}^{-\mathrm{p}}$; E ≡ particle energy), the observed radiation energy spectrum is $\mathrm{f}\sim {\mathrm{E}}_{\mathrm{ph}}^{-(\mathrm{p}-1)/2}$, where $E_{ph}$ is the observed photon energy. OJ 287 being an LBL (LSP BLL) source, the X-ray is a good tracer of the low-energy part, while the optical–UV directly traces the high-energy part of the broken power-law particle distribution required for the broadband SED modeling (e.g., [38]). The reported optical to X-ray spectrum corresponding to the lowest X-ray flux thus implies a hard low-energy particle spectrum of 1.3–1.6 and a high-energy spectrum of ∼4.4.

For blazars synchrotron spectrum, the observed frequency (${\nu }_{obs}$) is related to the rest frame via ${\nu }_{obs}=\delta /(1+z){\gamma }^2{\nu }_L$ where ${\nu }_L$ is the Larmor frequency, $\gamma$ is the electron Lorentz factor, and $\delta$ is the Doppler factor associated with the motion of the emission region. Further, ${\nu }_L=eB/2\pi m_{e}c$ depends solely on the magnetic field, B. Though the jet is expected to have a wide range of magnetic fields; for SEDs, the magnetic field associated with the dominant emission component will dictate the observed behavior. The most general broadband spectral state of OJ 287 is the LSP SED, and available literature records of SEDs do not indicate any significant change in the synchrotron peak of the LSP/LBL state. It should be noted that new spectral features e.g., thermal-like [31], an extremely soft X-ray spectral state [32] associated with an HSP/HBL-like SED [3,34] as well as soft X-ray excess [35] spectral states have been seen during or following its 2015 outbursts associated with the acclaimed ∼12-year optical outburst. However, all these were additional emission components. Further, the lowest concurrent optical–UV to X-ray spectral state and other SEDs reported here are all associated with the LBL/LSP state. Thus, for a given B and $\delta$, the extent of the optical–UV spectrum in the LSP spectral states depends on the Lorentz factor (energies) of the emitting particles. Since X-ray lies at the tail of the high-energy synchrotron spectrum, the combined optical to X-ray spectrum thus provides a potential direct tracer of the extent of the high-energy particle spectrum. The finding of an extended optical–UV synchrotron spectrum reported here clearly indicates a much extended high-energy particle spectrum compared to the bright X-ray phases of the source (e.g., [31,37,38]). On the other hand, the natural explanation of most of the flat X-ray spectra associated with the low and intermediate X-ray flux states of the LBL/LSP state of the source by adding the high-energy synchrotron tail to the lowest hard X-ray spectrum further strengthens claims/findings that most X-ray spectral changes are due to optical–UV synchrotron spectrum [37,38].

The relativistic particle spectrum has two main observables—the spectral index and the extent (energy limits; particle $\gamma$ referred above) of the spectrum. The spectral index is related to the particle acceleration processes, and the extent of the spectrum is related to the size of the acceleration region—the Hillas energy criteria [51] and probably radiative cooling (an often used Ansatz e.g., Inoue and Takahara [16]). The reported extended high-energy spectrum during low-flux states thus implies a comparatively larger acceleration/emission region than the flaring states that, in general, indicate a steepening or cutoff. The inference is consistent with short variability timescales during flares, indicating a compact emission region. Indirectly, the larger acceleration/emission region could be the large-scale jet emission.

Since the reported low hard X-ray spectrum is the sum of the synchrotron tail and IC, mainly by the low-energy part of the particle distribution, a harder spectrum at X-ray may not be a direct reflection of harder particle spectrum as such hard photon spectrum can also result from a higher lower-energy cutoff of the underlying particle spectrum [16]. The latter is consistent again with a larger acceleration/emission region argued above. These two scenarios are demonstrated in Figure 4 via SED modeling, where we have reproduced the optical-to-X-ray emission by changing only the low-energy particle spectral index, p and low-energy cutoff of the particle distribution, ${\gamma }_{min}$ in the low-state (State-3) model parameters of Kushwaha et al. [17]10. In case the harder photon spectrum is due to the particle spectrum, then the lower limit of the hard particle spectrum indicates magnetic reconnection as the most likely driver (e.g., [52]), while the upper limit appears consistent with the shock scenario [53].

Though OJ 287 has been studied extensively at all wavelengths, and especially at optical and X-rays, almost all spectral studies focusing on the latter have individually explored optical–UV and X-rays (e.g., [25,33,36,39], and references therein) or the X-ray flux have been comparatively higher (e.g., [24,38]). Given the blazar broadband emission, a consideration of all likely processes is a must to gain further insight into the source behavior and, thus, the jet physics, extra/hidden emission components, etc.

5. Summary and Conclusions

Our investigation of simultaneous optical to X-ray observations (spectra) of the BL Lacertae object OJ 287 with a focus on the lowest X-ray state revealed a power-law optical–UV spectrum with a photon spectral index ${\mathsf{\Gamma }}_{UVOT}=2.71\pm 0.03$ that extends into X-rays. Accounting for this contribution to 0.3–10 keV X-ray spectrum results in a harder power-law photon spectral index of ${\mathsf{\Gamma }}_X\sim 1.15-1.3$. This is harder than the reported hardest spectrum (${\mathsf{\Gamma }}_X\sim 1.5$) considering only the X-ray data and is consistent with the hard X-ray spectrum reported from the Swift-BAT data (20–100 keV). Our claim is verified by the natural explanation of another low optical–UV flux state having a similar flux level but with very different optical–UV and X-ray spectra as the sum of the synchrotron, and this inferred hard X-ray spectrum. Since the optical–UV directly traces the underlying high-energy particle spectrum, our finding implies a particle PL spectrum of ∼4.4, while the X-ray implies a harder PL spectrum of 1.3–1.6. Alternatively, the harder spectrum in the latter case could result from a comparatively higher lower-energy cutoff under the standard shock acceleration process.

The reported hard X-ray can naturally explain most of the observed flat X-ray spectra when combined with the corresponding optical–UV spectrum. Our finding further strengthens claims in the literature that most of the flat X-ray spectra of the source in the LBL/LSP spectral phase are driven by the optical–UV synchrotron spectrum extending to X-ray energies.

The extended optical–UV spectrum implies an extended high-energy particle spectrum and thus, a relatively large acceleration/emission region size per the Hillas criteria. The higher low-energy spectral cutoff is also consistent with the inference of a larger emission region. The larger size suggests that the low-state emission could be associated with the large-scale jet.