3C 120 Disk/Corona vs. Jet Variability in X-rays

Abstract

The 3C120 (Mrk 1506, UGC 03087, Mrk 9014) is a type 1 Seyfert (Sy1)/broad-line radio galaxy (BLRG) with intriguing variable jet activity featuring “dip” and “outburst” phases. Significant X-ray observational datasets have been collected for 3C120 by INTEGRAL, XMM-Newton, SWIFT, Suzaku, and other X-ray observational facilities. The overall X-ray spectrum of 3C 120 is too soft for typical radio-loud AGN, likely due to both variable spectral shape and jet contamination. Separating the “jet base” and nuclear (disc/corona) counterparts in the X-ray spectrum of 3C 120 can provide us with the possibility to investigate its variability in a more detailed way. Our objectives are to estimate separately the time variations of the accretion disc/corona and SSC/IC jet emission counterparts in the 3C 120 X-ray spectra and to analyze the physical state of the nucleus during different phases. Here, we attempt to use the connections between the synchrotron radio- and X-ray SSC/IC jet spectra and their photon indices and the dependence between the nuclear continuum and Fe-K iron luminescent line emission near 6.4 keV to separate the nuclear and jet base contributions to the total X-ray continuum. Using the X-ray observational dataset of 3C 120, we obtained separated fluxes that were interpreted as originating from the nucleus (disc/corona) and non-thermal SSC/IC jet base contributions. After this component separation, we identified the accretion disc/corona and jet states during different phases and compared them with the “jet/disk cycle” (Lohfink) and “magnetic plasmoid reconnection” (Shukla/Manheim) models.

1. Introduction

Since 1989, when Kellermann et al. [1] formulated the radio-to-optical brightness relation-based ($R=L_R/L_V$) radio dichotomy criterion for active galactic nuclei (AGN) (here, $L_R$ is the AGN luminosity at 5 GHz and $L_V$ is its luminosity in blue optical light near 689 THz), we have known that the majority of AGN belong to one of two classes, i.e., radio-loud (RL, $R\ge 10$) or radio-quiet (RQ, $R\le 1$), and that there are 10–20 times more RQ AGN known than RL ones. Later, radio-moderate (RM) AGN (also referred to as radio-intermediate (RI)), radio-transitional (RT), or radio-intermittent (RI again) AGN were detected in a series of works [2,3,4]. The terminology of AGN radio classes is not yet fully agreed upon, so we prefer to use the terms radio-moderate (RM) and radio-transitional (RT) to avoid the use of the ambiguous abbreviation RI. RM AGN have moderate values of $R=1-10$, whereas RT AGN demonstrate signs of both RL and RQ classes, being variable with unstable, intermittent, and recurrent behaviors typical of RL objects in a period of time, but looking similar to typical RQ objects during another period [5,6]. The 3C 120 is one of the RT AGN with a variable radio class, as it demonstrates behavior typical for RL or RQ AGN during different periods of observations [7]. In [8], the 3C 120 was even identified as a “schizophrenic” AGN because its observational characteristics change from those typical for RQ Sy1 AGN to those of a blazar or radio galaxy, and back again. Understanding the nature of such objects can help us answer what physical differences in the structure of the “central engine” of AGN lead to the observed differences in their radio properties.

The 3C 120 host is a nearby lenticular/disc S0 BLRG with the redshift z = 0.0336 [9]. Initially, it was identified as a galaxy under the name MCG +01-12-009 in the morphological catalog of galaxies [10]. Its spatial structure was investigated by means of the VLBI techniques for the first time in [11]. VLBI observations analyzed in 1999 [12] showed that the radio morphology of 3C 120 was unstable, and that its core was dominated by a single one-sided jet during some periods. The virial mass of the central BH of 3C 120 was estimated at around 5.7 ± 2.7·10${}^7$ M${}_{\odot }$ [13] using the reverberation mapping technique.

Extensive X-ray observations of 3C 120 have revealed that its X-ray spectrum is more Seyfert-like than blazar-like. Combined XMM-Newton and RXTE observations have shown a reflection fraction of $R\sim 0.5$, neutral luminescent iron lines Fe-K, and a thermal bremsstrahlung component with a 0.3–0.4 keV temperature that is clearly detectable [14], while similar results with a lower reflection fraction ($R\approx 0.2$) were obtained from Chandra observations [15]. Iron emission Fe-K lines were also detected in Suzaku observations of 3C 120 [16]. Additionally, at higher energies of up to 500 keV, the nucleus of 3C 120 demonstrates rather “RQ-like” spectral behaviors, with the values of high-energy exponential cut-offs above 100 keV [17], likely due to the effect of jet contamination. However, in [18], it was reported that the value of the high-energy exponential cut-off was found to be 83${}_{-8}^{+10}$ keV, together with the 25 ± 3 keV value of the electronic temperature within the comptonization models compTT, compPS, which is typical for an RL AGN. The jets are mostly in a quiescent state, being subdominant in X-rays, but mainly ’not zero-compatible’ [19]. The composition of the jets is likely to be electron–positron pair-dominated [20], with the particle energy density dominating over the magnetic field energy density at the jet base [19].

As shown in [21,22], there is a significant correlation between the dips and flares on the radio, optical, and X-ray light curves, which can be considered as a sign of the effect of the “jet base“ in the direct vicinity of the “central engine” in AGN. As estimated in [23], based on Fermi and VLBI observations of 3C 120, the sources of the synchrotron radio-emission and the inverse Compton (IC)/synchrotron-self-Compton (SSC) $\gamma$-ray emission are located approximately within ∼0.24 pc and ∼0.13 pc around the central massive black hole of 3C 120, respectively. The Fermi observations of 3C 120 and a sample of other similar objects have revealed their variability on a 6-month timescale at high $\gamma$ energies [24]. Thus, the jet base is located in the vicinity of the central supermassive black hole (BH) of the AGN, and that is why it is not spatially resolvable from the accretion disc and its corona. Their X-ray spectra are mixed in the total one, which will likely have the typical cut-off value significantly above 100 keV, as the values for jet IC/SSC spectra are usually on the GeV level. Therefore, the indirect methods of separating jet and nuclear counterparts are essential for investigating unstable RT AGN, such as 3C 120.

Two main concurring models explain the variability of 3C 120. The first model proposed by Lohfink [7] interprets the wideband variability observed in 3C 120 in terms of the so-called “disk-jet cycle” shown schematically in the upper panel of Figure 1.

The cycle runs from the state of “quiescence” when the disk is “full” (i.e., reaches its inner margin near the central black hole’s innermost stable circular orbit (ISCO) radius or the outer margin of the weak jet base) with no jet (or a weak but stable jet) at the center. Through the turbulent phase, when turbulence arises from the unstable innermost parts of the AGN, it arrives at the outburst state when the jet appears or strengthens significantly. At this point, the central parts of the accretion disk and corona are disrupted, and their matter is pushed away with the powerful jet flow. After reaching its maximum power, the jet gradually exhausts until its initial state, and the stage of disk refilling begins. The physical causes of such a cycle remain open to investigation.

The second model of AGN variability, called the “plasmoid magnetic reconnection”, was proposed by Shukla and Mannheim in [25] for another object, 3C 279, with a similar mechanism proposed to explain the variability of 3C 120 in [26]. The basic assumption of this model lies in the existence of an outer magnetically turbulent zone quite far away from the central black hole, at a distance of about ${10}^3{R}_g$, where enveloped plasmoids can appear and possibly grow if there is an additional fostering factor, such as an external shock wave. Such plasmoids can grow by swallowing smaller plasmoids until the so-called “monster plasmoid” is formed and begins to move inward towards the AGN central black hole. When the monster plasmoid reaches the AGN jet, the magnetic field of the plasmoid reconnects with the jet magnetic field, giving rise to a “jet flare” when the jet is curved and amplified for some period of time. Afterward, the jet exhausts and regains its initial shape. The scheme of this model is shown in the lower panel of Figure 1.

Here, we compare the results of our spectral fitting for 3C 120 with the “jet-disk cycle” and “plasmoid magnetic reconnection” models. To separate the “jet base” spectral component from that of the “disc/corona” one, we use the same two recipes we proposed in [27] for AGN 3C 111. The method is based on the spectral properties of the jet radio synchrotron and X-ray synchrotron self-Compton (SSC) or inverse Compton (IC) emissions. As it was shown in [28], the jet emission spectrum extends from the lower radio frequencies to higher observed energies in the form of the two-humped curve. The interesting property of this curve is that both the upper (IC/SSC) and lower (synchrotron) “knees” have the same photon indices. In case we have the appropriate observational data, we can apply the same power-law photon index value to fit the X-ray SSC/IC and radio synchrotron emission spectra of the jet base. In Section 3.1, we demonstrate the applicability of this method to determine the jet spectral counterparts.

In Section 2, we describe the observational data treated in this work. The basic presumptions of these recipes are described in Section 3.1. In Section 3.2 and Section 3.3, we perform our spectral analysis and compare the results obtained using the recipes mentioned above. In Section 4, we discuss our results, and in Section 5, we draw out our conclusions.

2. Observations and Data Reduction

In our study of 3C 120, we analyzed the X-ray observations performed by XMM-Newton/EPIC, Suzaku/XIS, SWIFT, INTEGRAL, and the Planck observations in radio waves available in the public data archive HEASARC. In this section, we describe the reduction performed to the initial data to obtain the spectra.

All of the soft-to-moderate X-ray data were split into parts corresponding to individual short-time observations (by Suzaku/XIS, XMM-Newton/EPIC and SWIFT/XRT) and fitted individually, to trace out the changes in the spectral shape and the fractions of disc/corona/jet counterparts.

2.1. Planck

We utilized the publicly available Planck spectrum of 3C120 in the 24–240 GHz range from the HEAVENS web page1 from the early release compact source catalog (ERCSC).

2.2. XMM-Newton/EPIC

3C 120 was observed four times by the XMM-Newton mission. The LOG of these observations is shown in

Table A1. The XMM-Newton/EPIC observational LOG. The data events from various data merged to produce one common spectrum, shown in one cell; the results of merging them are marked as “total”.

obsID	Obs.	Obs. Time,	Total EPIC
	Date	ks	Flux, cts
0109131101	2002-09-06	11.9	262,349
0152840101	2003-08-26	123.5	1,275,340
0693781601	2013-02-06	52.9	1,271,861
0693782401	2013-02-08	27.7	743,677
total	2013-02	80.6	2015538

(Appendix A).

The first two XMM observations of 3C 120, processed in 2002 and 2003, were reduced separately to obtain two complex EPIC spectra for these periods. The spectra obtained from the last two observations, performed shortly after one another in February 2013, were merged into one spectrum.

During all four observational periods, all EPIC cameras were being operated in a partial window mode.

We processed the EPIC data using the XMM SAS version 18.0 software package. The primary data reduction was performed using the standard SAS epproc and emproc chains, taking into account only the single- and double-photon events with the PATTERN ≤ 4 option. We excluded bad pixels and events near the CCD edge using the FLAG filter with FLAG = 0. To extract the source and background counts, we used the 30${}^{″}$-radii circular source-centered regions and the empty regions on the same CCD chip with (50–70)${}^{″}$-radii, respectively. When creating the spectra, the periods of significant flaring activity were excluded from the cleaned event files using the tabgtigen task. The ancillary files and response matrices were constructed using the standard SAS chains arfgen and rmfgen.

2.3. Swift

The Swift facility’s Burst Alert Telescope (BAT) and X-ray telescope (XRT) observed 3C 120 during the period from 2007 to 2020. We processed the Swift/BAT event and survey data using the BAT sub-packages to obtain spectra in the 14–195 keV energy range for every available period of 3C 120 observations (procedures batbinevt, batupdatephakw and batphasyserr) of the NASA HEASoft version 6.27.22. The whole recipe can be found on the UK Swift Science Data Centre web page3. The corresponding response matrices were generated using the batdrmgen procedure of the same sub-package. To create spectra from detector plane histograms of the BAT survey data, the batdph2pha wasapplied4.

Table A2. SWIFT XRT and BAT observation LOG.

Obs.	XRT Exp.,	XRT Flux	XRT Mode	BAT Exp.	BAT Flux
Date	ks	cts		ks	cts
2007	10.7	623	pc	9.5	0
2008	26.4	1411	pc	26.4	0
2011	14.2	1910	pc	14.2	0
2012/01	14.2	1712	pc	14.2	290
2012/02	9.7	1808	wt	9.2	650
2012/03	14.3	1852	wt	14.3	106
2012/04	6.7	975	pc/wt	6.7	0
2012/12	9.7	3074	pc/wt	4.3	0
2013/01	18.3	2044	pc/wt	4.4	200
2013/02	22.8	2046	pc/wt	14.8	490
2013/03	10.8	5798	pc/wt	10.9	90
2014/10	4.2	679	wt	4.3	0
2016/04	9.3	1272	pc/wt	9.5	110
2016/07	8.8	960	wt	8.8	150
2016/08	16.1	2015	pc/wt	16.3	680
2016/09	10.9	1500	wt	10.5	300
2016/10	10.2	1600	wt	12.3	480
2016/11	12.2	1600	wt	11.9	520
2016/12	12.0	1500	wt	11.5	360
2017/01	11.7	1991	wt	11.4	1390
2017/02	13.2	2021	wt	12.9	1570
2017/03	12.8	2200	wt	11.2	1550
2018/03	1.0	330	pc	1.0	100
2020	26.0	4246	wt	27.6	420

(Appendix A) contains the LOG of the Swift/BAT and XRT observations. We merged the closed-in-time observational data (i.e., within a monthly period) and presented them in the same table cell. The total exposure times, count rates, and signal-to-noise ratios for the resulting merged observational periods are also shown.

The Swift/XRT data were reduced using the online software provided by the Department of Physics and Astronomy of the University of Leicester (XROA)5. We utilized the single-pass centroid method with a maximum of 10 attempts and a search radius of 6 arcmin to determine the X-ray source position. Similar to the BAT observations, for the XRT data, the event collections of several close-in-time observations performed within the same months were merged.

2.4. Suzaku

The Suzaku mission performed six observations of 3C 120, four in February–March 2006 and two in February 2012. The LOG of all the Suzaku/XIS observations of 3C 120 is presented in

Table A3. Suzaku/XIS observation LOG. The data events from various data shown in one cell were merged to produce one common spectrum, marked as “total”.

obsID	Obs.	Obs. Time,	Total Flux
	Date	ks	cts
700001010	2006-02-09	41.9	448,401
700001020	2006-02-16	41.6	350,606
700001030	2006-02-23	40.9	417,900
700001040	2006-03-02	40.9	360,368
Total		165.3	1,577,275
706042010	2012-02-09	183.0	1,581,225
706042020	2012-02-14	118.1	1,020,452
Total		301.1	2,601,677

(Appendix A).

We reduced the Suzaku/XIS spectra by means of the online multi-mission web-interface UDON2 (Universe via Darts ON-line) provided by the Institute of Space and Astronautical Science (ISAS) by JAXA (Japan Aerospace Exploration Agency). The UDON2 web page and the archived Suzaku observational data can be found publically6. The resulting spectra obtained by UDON2 for all available XIS spectrometers during the same observation were merged into one using the add_asca_spec routine included in the HEASoft 6.26 software package. After this, we merged all the spectra obtained during February–March 2006 into one file with a total exposure time of 165.3 ksec and the two spectra obtained in February 2012 into another spectral file with a total exposure time of 301.1 ksec.

2.5. INTEGRAL

The INTEGRAL/ISGRI observations of 3C 120 were performed during 2005–2020; the observation LOG is shown in

Table A4. INTEGRAL/ISGRI observation LOG.

Data	Obs.	Obs. Time,	Data	obs.	Obs. Time,	Data	Obs.	Obs. Time,
rev	Date	ks	rev	Date	ks	rev	Date	ks
0338	2005-07	1.1	1378	2014-01	30.0	2070	2019-03	10.0
0343	2005-08	79.0	1379	2014-01	63.0	2076	2019-04	13.0
0344	2005-08	183.0	1439	2014-07	3.4	2121	2019-08	175.0
0346	2005-08	112.0	1441	2014-08	102.3	2123	2019-08	168.0
	total	375.1	1442	2014-08	171.0	2124	2019-08	113.5
0483	2006-09	49.0	1444	2014-08	15.0	2126	2019-08	14.0
0487	2006-10	63.0	1445	2014-08	3.5	2127	2019-08	10.0
	total	112.0	1446	2014-08	13.0	2136	2019-09	13.0
0530	2007-01	17.0	1447	2014-08	75.0	2137	2019-09	20.5
0532	2007-02	57.8	1456	2014-08	5.3	2143	2019-10	175.0
0538	2007-03	64.6	1457	2014-09	49.0	2148	2019-10	20.0
0543	2007-03	25.5	1459	2014-09	15.0		total	731.5
0548	2007-04	36.0	1463	2014-10	6.5	2187	2020-01	153.5
0588	2007-08	4.0		total	552.0	2188	2020-01	189.5
0589	2007-08	70.0	1503	2015-02	92.4	2189	2020-02	180.0
0590	2008-08	57.8	1504	2015-02	8.8	2194	2020-02	6.6
0591	2008-08	57.8	1505	2015-02	10.0	2198	2020-02	22.4
	total	373.5	1520	2015-03	31.0		total	552.0
0650	2008-02	22.2	1523	2015-03	33.0	2210	2020-03	19.2
0656	2008-02	68.0	1524	2015-03	37.0	2212	2020-04	19.3
0659	2008-03	52.5	1528	2015-03	16.0	2217	2020-04	33.0
	total	142.7	1575	2015-08	20.0		total	52.3
1073	2011-07	6.5	1577	2015-08	54.0	2257	2020-07	75.6
1075	2011-08	1.1	1581	2015-09	45.5	2258	2020-08	73.6
1077	2011-08	1.0	1584	2015-09	30.0	2259	2020-08	67.2
	total	8.6	1593	2015-09	34.0	2260	2020-08	83.2
1252	2013-01	1.2		total	411.7	2261	2020-08	52.5
1254	2013-01	1.1	1644	2016-02	59.5	2263	2020-08	13.2
1256	2013-01	23.0	1660	2016-03	21.0	2265	2020-08	59.4
1257	2013-01	13.0	1664	2016-04	28.0	2266	2020-08	59.4
1258	2013-02	135.0		total	108.5	2267	2020-08	89.6
1259	2013-02	135.0	1781	2017-02	3.5		total	484.1
1260	2013-02	3.3	1794	2017-03	17.0	2269	2020-09	79.2
1264	2103-02	23.3	1801	2017-04	13.0	2270	2020-09	73.5
1270	2013-03	15.0		total	33.5		total	152.7
1275	2013-03	1.3	1921	2018-02	3.3
1278	2013-04	10.0	1925	2018-03	3.3
1279	2013-04	40.0	1930	2018-03	9.7
1281	2013-04	5.6	1941	2018-04	20.0
1316	2013-07	18.0	1978	2018-07	1.0
1317	2013-07	25.0	1987	2018-08	8.3
1326	2013-08	23.0	1991	2018-09	6.5
1329	2013-09	6.7	1996	2018-09	6.5
1330	2013-09	19.2	2010	2018-10	13.0
1331	2013-09	59.5	2011	2018-10	25.6
1332	2013-09	59.5		total	123.7
1333	2013-09	11.0
1338	2013-09	24.5
	total	453.2

(Appendix A). The total ISGRI exposure time of the dataset used was 3.96 Ms (including all of the observations when the object was at an angle less than ${10}^{\circ }$ off-axis). To reduce the INTEGRAL IBIS/ISGRI dataset, we used standard recipes from Offline Scientific Analysis (OSA) software, version 11.1. All of the spectra were extracted individually for every science window and then summed up into the sets corresponding to the specific periods of observation times. Moreover, 3C 120 was detected by ISGRI up to the ∼500 keV energy level.

3. Data Analysis

3.1. Distinguishing Nuclear and Jet Base Spectral Components

The main difference between RL and RQ AGN spectra in hard X-rays lies in the presence of the jet base-induced component in the spectra of RL AGN [29]. Another difference is related to the conditions in the central engine, such as a larger inner radius of an accretion disk and a lower coronal temperature in an RL AGN compared to an RQ one [30]. The jets are bright in the radio range, which enables us to distinguish between the jet base and nuclear counterparts in the RL AGN spectra in X-rays as well. We use the same spectral component distinguishing method used in our previous work [27] for 3C 111 and in [31] for a sample of blazars. In [27], we proposed two ways to determine the jet base and primary nuclear counterparts in the X-ray spectrum.

The first way is based on the specific properties of a typical jet spectrum, which consists of two principal components: the low-energy component induced by synchrotron emission of the electrons composing the jet base (from radio to UV), and the high-energy component from UV to the highest $\gamma$-rays due to the IC (inverse Compton) or SSC (synchrotron self-Compton) jet base emissions. The important property of such a spectrum is that it has the same photon index ${\Gamma }_{jet}$ in the radio wave range as in the X-rays [28].

The jet base (of a $\mathsf{\mu }$pc size) is composed of ultra-relativistic charged particles that are ejected from the central region of the AGN at a relativistic bulk velocity. After being accelerated at the wavefront of a shock inside the jet, these particles emit synchrotron radiation that is visible in a wide energy range, from radio to UV. In the local thermodynamic equilibrium, these charged particles follow a power-law distribution with a spectral index of ${\gamma }_e$, which generates self-absorbed synchrotron emissions with the following spectral energy distribution (SED):

$$F_E\propto {(E/E_0)}^{5/2}·[1-exp(-{(E/E_0)}^{-\frac{{\gamma }_e-1}{2}})],$$

(1)

where $E_0>E$ is the break frequency at which an optical depth is $\tau =1$.

If we consider the energy range above the break energy of our object, i.e., $E>>E_0$, this formula can be approximated with the simple power-law dependency with the slope $\alpha =({\gamma }_e-1)/2=\Gamma +1$, where $\Gamma$ is the photon index [28,32,33]:

$$F_E=F_0{(E/E_0)}^{-\Gamma }.$$

(2)

Similarly, within the energy range underlying the break energy we can approximate Formula (1) with the power-law dependency again, but with the spectral index 5/2 (i.e., $\Gamma =1.5$). The non-equilibrium configuration of the jet base with the turbulent magnetic field was considered in [34,35,36]; they showed that the peak segments of the SED can sometimes be enlarged to form a plateau, but the “knees” have the same slopes. The slope of the “upper knees” can vary with the electron density or energy distribution, but the “lower knee” is expected to have a persisting slope within the models described above. Taking all of this into account, we can roughly consider both the synchrotron and SSC/IC counterparts of the jet base spectra, consisting of the three following segments:

• Absorbed segment or “lower knee”, power-law with the slope near 1.5;
• Peak or plateau segment near the break energy;
• Transparent segment or “upper knee”, power-law with the slope $\Gamma$.

The slopes of the “lower knees” of synchrotron and SSC/IC spectra are the same, as well as the synchrotron and SSC/IC “upper knees” slopes.

However, before applying this method to a specific object, it is necessary to check its applicability within the working energy intervals of the observational facilities providing the data for the investigation of the object. In our case, we have the 20–240 GHz radio window of Planck, the 3–10 keV energy range of Swift/XRT, XMM-Newton/EPIC, and Suzaku/XIS, and the 15–500 keV range of INTEGRAL/ISGRI and Swift/BAT. To see if we can apply our method, we demonstrate in Figure 2 the SED of 3C 120 within the one-zone SSC model by [37], constructed using the Tramacere SED tool (accessible online at https://www.isdc.unige.ch/sedtool/PROD/SED.html, accessed on 15 December 2021). We plotted the 24–240 GHz Planck working range and 0.1–1000 keV energy range to show that they both cover the “lower knees” areas of synchrotron and SSC/IC jet emissions. Thus, if we assume this model for 3C 120 SED, we can suppose the persistent value of the “lower knee” slope in the radio and X-rays and apply this when fitting the jet base X-ray SSC/IC and radio synchrotron emission spectra. However, as seen from this figure, the “transitional” area, or even the fragment of the “upper knee”, of the SSC/IC jet emission can extend into the 0.1–1000 keV range. For the description of the jet base counterpart in the X-rays, we apply a broken power-law model with a fixed lower photon index and break energy, and a variable upper photon index.

In this view, to separate the contributions of the corona/disk and the jet base, we use the equality of the synchrotron and SSC/IC “lower knee” slopes (${\Gamma }_{lower}^{fr}={\Gamma }_{radio}$), obtained from the Planck 24–240 GHz spectrum of 3C 120. The other two parameters of the X-ray jet base spectral model, the break energy and the upper photon index, are left unfrozen.

The second additional method we are trying here is based on the link between the continuum emission of the corona/disk and the Fe K$\alpha$ emission line near 6.4 keV. This line is formed in the accretion disc that is irradiated by the corona and is expected to be correlated with the corona/disk continuum [38]. To understand how this emission line is linked with the primary nuclear continuum, we show a schematic view of the AGN “central engine” and the emission components in Figure 3.

Figure 2. The radio-to-X-ray data and SED of 3C 120 within the one-zone SSC model by [37], constructed using the Tramacere SED tool. The black circles correspond to observational points from various radio facilities, such as Planck, APEX, ATCA, PACO, SiMPlE, Metsahovi, OVRO, RATAN, UMRAO, and VLA, as well as from Effelsberg, IRAM, KVA, and Xinglong in the IR/optical range, Swift/UVOT in UV rays, and ROSAT/RASS and SWIFT (BAT and XRT) in soft to hard X-rays [39]. The red circles correspond to the Fermi/LAT spectrum. The blue lines (solid, dashed and dotted) show models with different spectral indices. The Planck range is marked by a thick green line, while the X-ray range corresponding to Swift/XRT, XMM-Newton/EPIC, and Suzaku/XIS is marked by red.

Figure 2. The radio-to-X-ray data and SED of 3C 120 within the one-zone SSC model by [37], constructed using the Tramacere SED tool. The black circles correspond to observational points from various radio facilities, such as Planck, APEX, ATCA, PACO, SiMPlE, Metsahovi, OVRO, RATAN, UMRAO, and VLA, as well as from Effelsberg, IRAM, KVA, and Xinglong in the IR/optical range, Swift/UVOT in UV rays, and ROSAT/RASS and SWIFT (BAT and XRT) in soft to hard X-rays [39]. The red circles correspond to the Fermi/LAT spectrum. The blue lines (solid, dashed and dotted) show models with different spectral indices. The Planck range is marked by a thick green line, while the X-ray range corresponding to Swift/XRT, XMM-Newton/EPIC, and Suzaku/XIS is marked by red.

In S1 type AGN, such as 3C 120, the observer can detect both the direct and indirect (reflected) emissions from the “central engine.” The direct component originates from the corona and Compton reflection on the accretion disk surface, which gives rise to the broad component of emission lines, such as Fe-K$\alpha$, with specific relativistic profiles. Thus, the two corresponding components in the spectrum are clearly connected. However, the presence of light-bending due to other reflecting elements, such as the torus, can complicate the situation [40,41]. The direct corona continuum, the continuum reflected in the disk, and the line emission can be reflected again in the torus, and there can be additional lines with natural widths emitted in the torus. In the case of a clumpy or patchy torus, the reflection and reprocessing can be very complicated and significantly smear the emission line profiles [42]. Another factor that can affect the situation is the ionization level ${\chi }_i$ [43]. If the reflection fraction and ionization level remain constant, one can roughly estimate the link between the line $F^{line}$ and primary nuclear $F^{nucl}$ fluxes as linear. However, in the general case, we should consider the possible variability of the reflection fraction and ionization level, which can be induced by the inhomogeneities in the disk and/or the torus, by the temperature and magnetic field variations, or by the changes of the inner disk radius due to the jet activity. Thus, if we suppose the linear dependence between flux levels in the corona/disk continuum and Fe-K emission line, the parameters of this dependence should take into account the reflector geometry (inclination angle, reflection coefficient, etc.), ionization level, and their variations:

$$F^{nucl}=F_0(i,{\chi }_i)+\kappa (i,{\chi }_i)\ast F^{line}.$$

(3)

In our case, we cannot establish tight constraints on the inclination value to trace out any variations. However, the variations in the ionization level are clearly visible from the results of the spectral fitting. Here, we are testing if the recipe for determining the nuclear flux based on the iron line is appropriate in such a situation.

Thus, we use the equal “lower knee” slopes for radio synchrotron and X-ray SSC/IC jet base spectral component. The Planck radio spectrum was fitted with a simple power-law model with the photon index ${\Gamma }_{jet}=1.24\pm 0.05$. To fit the SSC/IC component in the X-ray spectrum, we used a broken power-law model with the break energy and the photon index of the “upper knee” left free to vary. In addition, we used the linear relationship between the 5–7 keV fluxes in the Fe-K$\alpha$ emission line and the continuum component to set up additional constraints on the nuclear disc/corona emission.

3.2. X-ray Spectrum Fitting

We consider the 3.0–300 keV energy range to avoid additional complications due to the presence of the thermal emissions of the gas surrounding the active nucleus in 3C 120, which pollutes its spectra below 3 keV. To fit the 3C 20 hard X-ray spectrum, we used the model containing the following components:

• The corona/disk power-law continuum $F\left(E\right)=K{E}^{-\Gamma }exp(-E/Ec)$ Compton-reflected by the plasma with the electronic temperature $T_e$, represented by the model xillver-comp [44,45] with the exponential high-energy cut-off by the model high-cut and absorbed mildly by the neutral medium (tbabs);
• Fe K$\alpha$ 6.4 keV emission line, represented here by the Gaussian model;
• The jet base SSC/IC emission (represented by the broken power-law with no cut-off bknpo model): $F\left(E\right)=K{E}^{-{\Gamma }_{lower}}$ if $E<E_{break}$ and $F\left(E\right)=K{E}^{-{\Gamma }_{upper}}$ for $E>E_{break}$, with the lower photon index ${\Gamma }_{lower}$ frozen to the value 1.24 obtained from the fitting of the Planck spectrum; the break energy $E_{break}$ and upper photon index ${\Gamma }_{upper}$ were left free to vary.

The fitting process was performed in the following steps:

• Fitting the 15–300 keV continuum using the cutoffpl model (with the photon index and high-energy exponential cut-off values free to vary) and the broken power-law model for the jet base SSC/IC emission (with the lower photon index frozen to the value 1.24 and break energy and upper photon index free to vary) allows us to obtain constraints on the parameters $E_{break}$, ${\Gamma }_{upper}$, $\Gamma$ (photon index of the nuclear spectrum), and $E_C$ (high energy exponential cut-off of the primary emission of the nucleus);
• Fitting the 3–300 keV continuum, excluding the 5–7 keV range where the Fe-K emission lines are visible, using the model that includes the jet and the nuclear continuum emission: (xillver-comp($\Gamma ,x{i}_i,k{T}_e$)∗high-cut($E_c$) + bknpo(${\Gamma }_{lower}^{fr},E_{break},{\Gamma }_{upper}$))∗tbabs($N_H$)∗const with ${\Gamma }_{fr}^{jet}$ = 1.24 and the other parameters used in the previous step (E${}_{break}$, ${\Gamma }_{upper}$, $\Gamma$, E${}_C$ frozen, or confined within the error level to the values obtained in the previous step). Having performed this step, we obtain the values of the thermal (electronic temperature T${}_e$) and ionization (potential ${\chi }_i$) parameters of the nuclear continuum, as well as the proper absorption and the fluxes corresponding to the jet and nuclear components;
• The 5–7 keV range fitting using the Gaussian emission line plus the jet and nuclear continuum models with the parameters frozen to the values obtained in the previous steps. The latter gives us the emission line parameters: line energy $E_{line}$ and width ${\sigma }_{line}$, equivalent width EW, in a 5–7 kev line flux.

We fit the proper and Galactic absorption $N_{H0}$ using the tbabs model, including the Galactic absorption as the lowest possible fitting value at the level of 1.94 × 10${}^{21}$ cm${}^{-2}$ [46]. The inter-calibration constants were included in the simultaneous modeling of softer (XMM-Newton/EPIC, Suzaku/XIS, or SWIFT/XRT) and harder (INTEGRAL/ISGRI or Swift/BAT) spectra. Their values were always in the range of 0.6–3.5; they are shown in the last column of

Table 1. The datasets and the best-fit parameters for the high-energy part of the spectrum above 15 keV. The table shows the jet spectrum parameters, $E_{break}$ and ${\Gamma }_{upper}$, and the inter-calibration constants, K${}_{\gamma }$ for the $\gamma$-ray spectra. The last row in the table: variances.

N${}_{\_}^{\mathit{o}}$	Dates	Dataset	${\mathbf{\Gamma }}_{\mathit{upper}}$	${\mathit{E}}_{\mathit{break}}$, keV	K${}_{\mathit{\gamma }}$
1 ${}^{+}$	09/2002	EPIC	-	-	-
2	08/2003	EPIC + JEM-X + ISGRI	3.0 ± 1.5	189 ± 86	0.97 ± 0.10 (ISGRI) 0.75 ± 0.15 (JEMX)
3	02/2006	XIS + JEM-X + ISGRI	3.6 ± 1.5	105${}_{-15}^{+52}$	1.2 ± 0.4 (ISGRI) 1.0 ± 0.2 (JEMX)
4	03–07/2007	XRT + BAT + ISGRI	-	-	1.03 ± 0.19 (ISGRI) 1.0 ± 0.2 (BAT)
5	07–08/2008	XRT + BAT + ISGRI	2.1 ± 0.5	36 ± 23	0.8 ± 0.1 (ISGRI) 1.1${}_{-0.3}^{+0.2}$ (BAT)
6	08/2011	XRT + BAT + ISGRI	<1.15	90${}_{-50}^{+270}$	0.95 ± 0.06 (ISGRI) 0.8 ± 0.2 (BAT)
7	01/2012	XRT + BAT	3.4 ± 2.0	67${}_{-23}^{+84}$	2.0${}_{-1.0}^{+0.5}$
8	02/2012	XIS + XRT + BAT	>3.0	>105	0.62 ± 0.08
9	03/2012	XRT + BAT	2.2 ± 0.5	50 ± 13	0.8${}_{-0.2}^{+0.3}$
10	04/2012	XRT + BAT	-	-	1.3 ± 0.4
11	12/2012–03/2013	XRT + BAT + ISGRI	-	-	0.9 ± 0.2 (ISGRI) 0.6${}_{-0.1}^{+0.2}$ (BAT)
12	02/2013	EPIC + BAT + ISGRI	-	-	1.03 ± 0.18 (ISGRI) 0.83 ± 0.18 (BAT)
13	10/2014	XRT + ISGRI	3.4 ± 0.9	25${}_{-4}^{+6}$	1.2 ± 0.2
14	04/2016	XRT + BAT + ISGRI	3.6 ± 1.3	37 ± 10	0.6${}_{-0.2}^{+0.3}$
15	07–09/2016	XRT + BAT + ISGRI	4.0 ± 1.2	39 ± 14	0.7 ± 0.1 (BAT), 1.0 ± 0.2 (ISGRI)
16	11–12/2016	XRT + BAT + ISGRI	0.1${}_{-0.5}^{+0.8}$	124${}_{-36}^{+97}$	0.9${}_{-0.2}^{+0.3}$ (BAT) 1.3 ± 0.6 (ISGRI)
17	01–03/2017	XRT + BAT + ISGRI	2.1 ± 0.7	37 ± 18	1.6 ± 0.4 (BAT) 0.9${}_{-0.3}^{+0.5}$ (ISGRI)
18	12/2020	XRT + BAT + ISGRI	-	-	0.53 ± 0.09 (BAT) 1.9${}_{-0.6}^{+0.4}$
19	11/2021	XRT + BAT	-	-	2.4 ± 1.1
20	01/2022	XRT + BAT	-	-	0.5 ± 0.2
Variance			2.21	12.04	-

${}^{+}$ Due to the absence of high-energy (ISGRI or BAT) observations for this period, we used a simple power-law model with a frozen photon index to fit the jet emission. For cases where only BAT data were available at high energies, we left the photon index for the nucleus free to vary when fitting the total 3–300 keV spectrum, in order to obtain better constraints on the parameter.

. The lower and upper errors indicated below correspond to the 90% level of confidence (i.e., $\Delta {\chi }^2$ = 2.71). The standard error calculation algorithm of the HEASoft Xspec software package is as follows: Each non-frozen parameter is varied within its allowed limits until the value of the fit statistic ${\chi }^2/d.o.f.$ is minimized, which is equal to the last value of the fit statistic determined by the fitting command plus the indicated $\Delta {\chi }^2$ (i.e., 2.71 in our case). At the end of each table containing the fitting parameters, we also added the variance of the parameter calculated using the following formula:

$$Var\left(P\right)=\frac{1}{N}\sum _{i=1}^N\frac{{(P_i-<P>)}^2}{err{P}_i^2}.$$

where P and $errP$ are the fitting parameter and error bar, and $<P>$ is its mean value.

The spectral modeling was performed by means of the Xspec v.12.12.1f (the HEASoft software package, version 6.30.1). To investigate variability, we combined the Swift/BAT and INTEGRAL spectra from observational time intervals that overlapped with shorter intervals of observations by XMM-Newton/EPIC, Suzaku/XIS, or SWIFT/XRT facilities. The most interesting spectra, along with the fitting spectral models, are shown in Figure A1 in Appendix B.

Initially, all of the observational data were divided into 28 observational periods of monthly duration. However, it was noticed that during the period covering July 2016 to March 2017, the behavior of the primary nuclear counterpart in the spectrum was similar to the photon index at the level of 1.6–1.7, with no detected exponential cut-off at high energies, and the Fe-K$\alpha$ emission line was poorly constrained due to insufficient statistics for a high-quality line fit. Therefore, those observational periods, namely December 2012–March 2013, July–September 2016, November–December 2016, and January–March 2017, were merged into longer ones to obtain better constraints on the fitting parameters. This resulted in a set of 20 observational periods described in Table 1. The datasets involving the combination of softer (XMM-Newton/EPIC, SUZAKU/XIS, or SWIFT/XRT) and harder (INTEGRAL/ISGRI or SWIFT/BAT) X-ray spectra were collected during the same periods.

As we can see from Table 1, the values of the photon index corresponding to the “upper knee” of the jet base IC/SSC emissions were mostly compatible with the range of values between 2.0 and 2.8, obtained by [47] from the Fermi/LAT observations of 3C 120. The exceptions were only the two periods with numbers 6 (08/2011) and 16 (11–12/2016) when they were zero-compatible; this can be interpreted as a result of the appearance of the “plateau” in the SSC/IC segment of the jet base emission.

In

Table 2. The 3–300 keV continuum best-fit parameters. The first column: observational period; the second to fifth columns: the fitting parameters: photon index of the primary disk/corona emission, the excess of the absorbing column density above the Galactic absorption N${}_H$, the high-energy exponential cut-off in the corona/disk spectrum E${}_c$, the electronic temperature of the coronal plasma kT${}_e$, the ionization potential ${\chi }_i$; the sixth column: ${\chi }^2$/degrees-of-freedom. The last row: variances.

N${}_{\_}^{\mathit{o}}$	$\mathbf{\Gamma }$	N${}_{\mathit{H}}$, 10${}^{22}$ cm${}^{-2}$	E${}_{\mathit{c}}$, keV	kT${}_{\mathit{e}}$, keV	lg${\mathit{\chi }}_{\mathit{i}}$	${\mathit{\chi }}^2/\mathit{d}.\mathit{o}.\mathit{f}.$
1	2.0 ± 0.3	2.0${}_{-1.4}^{+1.7}$	unc.	10${}_{-2}^{+28}$	3.4${}_{-0.1}^{+0.3}$	239.1/247
2	1.67 ± 0.18	<0.01	33${}_{-11}^{+29}$	9.5 ± 0.2	4.1 ± 0.1	1567.5/1257
3	1.68${}_{-0.29}^{+0.42}$	<0.03	22${}_{-9}^{+32}$	8.2 ± 0.3	3.6 ± 0.1	1170.7/970
4	1.36 ± 0.16	1.8 ± 1.2	23${}_{-8}^{+12}$	8.1 ± 0.5	3.5${}_{-0.2}^{+0.9}$	51.7/44
5	1.59${}_{-0.2}^{+0.4}$	<0.84	44${}_{-27}^{+1389}$	8.8 ± 0.8	3.8${}_{-0.4}^{+1.0}$	195.6/207
6	1.44${}_{-0.14}^{+0.18}$	2.1 ± 0.9	23${}_{-6}^{+8}$	8.4 ± 0.5	3.7${}_{-0.3}^{+1.0}$	144.8/120
7	1.85 ± 0.19	<2.1	71${}_{-56}^{+68}$	29${}_{-20}^{+28}$	>3.65	101.2/119
8	2.09 ± 0.03	<0.09	121${}_{-20}^{+30}$	36${}_{-13}^{+10}$	3.9 ± 0.1	2457.1/2035
9	2.9 ± 0.4	<5.0	unc.	<6.0	<2.1	101.8/80
10	1.97 ± 0.18	<3.7	>51	>7.0	3.4${}_{-0.3}^{+1.3}$	44.1/45
11	2.0 ± 0.13	0.9 ± 0.5	>326	>24.0	3.8 ± 0.2	284.3/265
12	1.3 ± 0.1	<0.11	27${}_{-3}^{+17}$	9.1 ± 0.3	3.8 ± 0.1	162.6/146
13	3.3 ± 0.5	<14.8	25${}_{-4}^{+6}$	<20.0	3.1${}_{-2.7}^{+0.5}$	232.4/124
14	1.31 ± 0.04	<1.2	unc.	47 ± 15	3.9 ± 0.2	195.2/200
15	>2.07	<4.5	11${}_{-6}^{+108}$	<48	2.0${}_{-1.2}^{+0.1}$	280.2/256
16	2.06 ± 0.15	<0.53	>37	>8.6	3.7 ± 0.2	161.5/133
17	1.75 ± 0.09	<2.0	unc.	>6.6	3.7 ± 0.3	90.8/90
18	1.28${}_{-0.14}^{+1.68}$	<1.6	2.1 ± 0.9	<3.6	1.7 ± 0.4	132.9/130
19	>1.2	<5.3	>1.9	<60.0	>2.2	90.3/124
20	>1.7	16.2${}_{-7.0}^{+4.0}$	unc.	<3.0	>3.8	138.7/133
Var.	28.5	3400.6	344.2	111.8	11.0	-

, we demonstrate the 3–300 keV continuum parameters obtained in the second step of our fitting, namely the photon index $\Gamma$, the high-energy exponential cut-off $E_c$ of the corona/disk continuum, the electronic temperature of the coronal plasma kT${}_e$, the absorbing column density excess in comparison with the Galactic absorption N${}_H$, ionization potential ${\chi }_i$, and the ${\chi }^2$/d.o.f. The mean value of the electronic temperature is 17${}_{-4}^{+10}$ keV, which is in good agreement with the value of 25 ± 2 keV obtained by [18] from the NuSTAR observations of 3C 120; the temperature itself demonstrates the obvious signs of significant variability, as well as the ionization potential, with variance values of 111.8 and 11.0, correspondingly. The mean values of the photon index and exponential high-energy cut-off in the primary nuclear spectrum are 1.92 ± 0.12 and 72${}_{-6}^{+97}$ keV, which are also compatible with the values of 1.87 ± 0.02 (for the photon index) and 83${}_{-8}^{+10}$ keV (for the high-energy exponential cut-off) obtained in [18] from the NuSTAR observations within the error levels; the mean value of the photon index of the primary nuclear continuum emission is also compatible with the value of $\Gamma ={2.08}_{-0.06}^{+0.07}$, obtained for 3C 120 in [48]. The higher value of the electronic temperature of the plasma reported in [48], namely $k{T}_e={176}_{-23}^{+24}$ keV, is probably due to their model not including the jet base spectral component.

As we can see from our spectral fits during the periods of observation and the plots of their time dependencies shown in Figure A6 and Figure A7 (Appendix D), the parameters of high-energy cut-off and photon index of the primary nuclear emission are highly variable with time, the values of their variances are 344.2 and 28.5; it is noticeable that the values of the upper knee photon index of the jet and that of the primary nuclear continuum are very close to each other. The latter significantly complicates the situation with the jet and nuclear spectral component recognition. It can be seen from Table 2 that the ionization level is, on average, within the range of 3.–4. However, during several periods, it decreased significantly to a level between 2 and 2.2 (or below). We observed three of such periods (9, 15, and 18), with ∼52–54 months period between them.

In

Table 3. Fe K$\alpha$ line emission parameters. The first column: the observational period; the second to fourth columns: the line energy, width, and EW; the fifth column: the null-hypothesis probability of the spectral model with no emission line in comparison with the model with the Gaussian line near the 6.4 keV rest frame added. The last row: variances.

N${}_{\_}^{\mathit{o}}$	${\mathit{E}}_{\mathit{line}}$, keV	${\mathit{\sigma }}_{\mathit{line}}$, keV	EW, keV	${\mathit{P}}_{\mathit{null}}$, %
1	6.38${}_{-0.05}^{+0.07}$	<0.15	0.06 ± 0.01	0.6
2	6.41 ± 0.02	0.05 ± 0.02	0.04 ± 0.01	1.7 × 10${}^{-39}$
3	6.38 ± 0.03	0.08 ± 0.02	0.04 ± 0.01	9.5 × 10${}^{-33}$
4	6.4${}^{fr}$	<0.15	0.08 ± 0.07	43.7
5	6.25 ± 0.18	<0.2	0.07 ± 0.06	3.5
6	6.52${}_{-0.13}^{+0.12}$	<0.2	0.09 ± 0.06	15.5
7	6.33 ± 0.20	<0.2	0.04 ± 0.03	76.0
8	6.38 ± 0.02	0.06 ± 0.02	0.05 ± 0.01	1.0 × 10${}^{-58}$
9	6.41 ± 0.11	<0.13	0.05 ± 0.03	46
10	6.4${}^{fr}$	<0.07	<0.09	75
11	6.4 ± 0.09	<0.12	0.06 ± 0.02	0.59
12	6.41${}_{-0.03}^{+0.02}$	<0.2	0.042 ± 0.003	5.9 × 10${}^{-6}$
13	6.4${}^{fr}$	<0.2	<0.05	81.8
14	6.42 ± 0.16	<0.3	0.15 ± 0.02	19.6
15	6.4${}^{fr}$	<0.09	<0.001	99.5
16	6.33 ± 0.15	<0.17	0.07 ± 0.01	31.8
17	6.28${}_{-0.08}^{+0.42}$	<0.1	0.05 ± 0.05	42.7
18	6.26 ± 0.15	<0.15	0.06 ± 0.05	29.6
19	6.4${}^{fr}$	<0.2	<0.02	93.5
20	6.4${}^{fr}$	<0.2	<0.07	75.8
Var.	0.65	0.25	3.67	-

, we present the fitting parameters of the 6.4 keV Fe-K$\alpha$ emission line, including the line energy (second column), width (third column), and equivalent width (EW) calculated using the Xspec standard procedure eqwidth (last column). As expected, the variance values of the line energy and width are lower than 1, indicating that these parameters are constant. The flux level and EW, with their variances of 2.27 and 3.67, respectively, show signs of variability during the considered period. The EW values during the observational periods were mostly in the range of 0.03–0.05 keV, typical for RL AGN [49], and compatible with this range within the error levels. However, there was one period (N${}_{-}^o$14, April 2016) when it was significantly higher, at a level of 0.15 keV. Such values of the Fe-K$\alpha$ equivalent width are common in RQ AGN [49], which is why we can presume the zero-compatible level of the X-ray jet base emission during this period.

The jet base and corona/disk 5–7 keV continuum fluxes and the 5–7 flux in the Fe-K$\alpha$ emission line, calculated using this spectral model, are shown in the

Table 4. The 5–7 keV range fluxes for the three spectral components. The first column: the observational period; the second to fourth columns: the jet base, corona/disk, and the Fe-K$\alpha$ emission line fluxes calculated within the spectral fitting model. The fifth and sixth columns: the corona/disk and jet base fluxes calculated using the dependence between the Fe-K$\alpha$ and corona/disk continuum fluxes. The last row: variances.

N${}_{\_}^{\mathit{o}}$	${\mathit{F}}_{5-7}^{\mathit{jet}\ast }$	${\mathit{F}}_{5-7}^{\mathit{nucl}\ast }$	${\mathit{F}}_{5-7}^{\mathit{line}\ast }$	F${}_{\mathit{nucl},\mathit{calc}}$	F${}_{\mathit{jet},\mathit{calc}}$
1 ${}^{†}$	1.1 ± 0.3	9.2 ± 0.3	0.3 ± 0.1	5.4${}_{-4.7}^{+4.9}$	<9.6
2	1.0 ± 0.1	9.6 ± 0.1	0.2 ± 0.1	3.6${}_{-3.0}^{+7.0}$	<10.0
3 ${}^{†}$	4.60 ± 0.02	4.55 ± 0.02	0.14 ± 0.02	2.5${}_{-2.2}^{+6.7}$	<8.8
4	2.1 ± 0.5	7.4 ± 0.5	<0.9	>10.0	<10.0
5	3.6 ± 0.2	6.9 ± 0.2	0.4 ± 0.2	7.2${}_{-6.6}^{+3.5}$	<10.1
6	2.1 ± 0.2	7.4 ± 0.2	0.4 ± 0.2	7.2${}_{-6.6}^{+2.4}$	<9.0
7 ${}^{†}$	2.1 ± 0.2	6.9 ± 0.2	<0.3	<9.0	<9.0
8 ${}^{†}$	5.30 ± 0.03	6.80 ± 0.03	0.15 ± 0.03	2.7${}_{-2.3}^{+9.4}$	<11.7
9	6.7 ± 0.2	4.1 ± 0.2	0.3 ± 0.2	5.4${}_{-5.1}^{+5.6}$	<10.5
10 ${}^{†}$	1.7 ± 0.5	7.0 ± 0.5	<0.5	<9.3	<9.0
11 ${}^{†}$	1.3 ± 0.3	10.2 ± 0.3	<0.4	<11.8	<11.8
12	3.00 ± 0.04	8.40 ± 0.04	0.24 ± 0.01	4.3${}_{-3.6}^{+8.1}$	<11.7
13	6.2 ± 0.3	2.8 ± 0.3	<0.4	<9.3	<9.3
14	<1.0	10.3 ± 0.5	0.6 ± 0.5	10.8${}_{-10.5}^{+1.0}$	<11.3
15 ${}^{†}$	7.6 ± 0.2	2.0 ± 0.2	<0.2	<9.8	<9.8
16 ${}^{†}$	3.0 ± 0.2	8.2 ± 0.2	0.4 ± 0.2	7.2${}_{-6.6}^{+4.2}$	<10.8
17	3.4 ± 0.2	7.5 ± 0.2	0.2 ± 0.1	3.6${}_{-3.0}^{+7.2}$	<10.5
18 ${}^{†}$	7.9 ± 0.2	1.8 ± 0.2	<0.3	<9.9	<9.9
19 ${}^{†}$	7.8 ± 0.4	0.8 ± 0.4	<0.5	<9.0	<9.0
20 ${}^{†}$	2.3 ± 0.6	4.4 ± 0.6	<0.7	<7.3	<7.3
Var.	404.8	650.1	2.27	-	-

* 10⁻¹² erg/cm²·s. ^† During the phases marked with this the photon index Γ of the primary nuclear emission (shown in the Table 2, the second column) was compatible with the mean value of the photon index Γ_jet of the jet base emission.

.

3.3. Nuclear Continuum Flux vs. Fe-K$\alpha$ Line Flux

In this subsection, we estimate the jet and nuclear 5–7 keV fluxes using the results of the Fe-K$\alpha$ emission line fitting, and the emission line fluxes calculated in the previous subsection.

In Figure 4, we arranged equivalent widths of the Fe K$\alpha$ line vs. integrated 5–7 keV jet flux $F_{5-7}^{jet}$ to demonstrate that the dependency between them is inverse; for the case of 3C 111, the larger equivalent widths correspond to the smaller $F_{5-7}^{jet}$ values. The highest value of the equivalent width derived for 3C 120 is about 150 eV (14th period) and it corresponds to zero-compatibly or an extremely low level of the jet base emission. This is also in good agreement with the typical values of the EW of RQ AGNs [49].

Now, let us consider the relation between the X-ray continuum in the 5–7 keV range and flux in the Fe K$\alpha$ line. Namely, we suppose that the total integrated (5–7 keV) observed flux $F_{5-7}^{line}$ in the Fe K$\alpha$ line is proportional to the total nucleus (disc/corona) flux in the 5–7 keV range: $F_{5-7}^{nucl}=\kappa F_{5-7}^{line}$.

We used the results of our spectral fitting in order to derive these coefficients. There was a zero-compatible jet emission period, namely the 14th one. We considered this period to estimate the coefficient’s value of $\kappa ={18}_{-15}^{+90}$. With this coefficient, we applied it to the data to derive the Fe-K$\alpha$ line fluxes to derive the nuclear and jet base fluxes within the 5–7 keV energy range, as shown in Table 4.

Using the linear coefficient found in this recipe, we calculated the values of the nuclear flux for each observation and compared these values with those obtained from the spectral fitting. However, it was observed that the error bars for the fluxes derived with this recipe for 3C 120 were too high to use successfully with the available data for this object.

3.4. Correlation of Parameters

To trace out possible links between the spectral parameters, we calculated the correlations between the fitting parameters ${\Gamma }_{upper}$, $E_{break}$, $\Gamma$, $E_c$, $N_H$, $k{T}_e$, lg${\chi }_i$, EW, and 5–7 keV fluxes corresponding to the jet base, Fe-K$\alpha$ emission line, and nuclear counterparts. The correlation between parameters X and Y was calculated using the following formula:

$$corr(X,Y)=\frac{{\sum }_{i=1}^N(X_i-<X>)(Y_i-<Y>)}{\sqrt{{\sum }_{i=1}^N{(X_i-<X>)}^2{\sum }_{j=1}^N{(Y_j-<Y>)}^2}}$$

where $<X>$ is the mean value of the X parameter. The positive (+) and negative (−) errors of the correlation coefficients were calculated using the formulae:

$$\Delta cor{r}^{+}(X,Y)=\frac{{\sum }_{i=1}^N(X_i-<X>)\Delta Y_i^{+}+(Y_i-<Y>)\Delta X_i^{+}}{\sqrt{{\sum }_{i=1}^N{(X_i-<X>)}^2{\sum }_{j=1}^N{(Y_j-<Y>)}^2}}+$$

$$+\frac{{\sum }_{i=1}^N{\sum }_{j=1}N(X_i-<X>)(Y_j-<Y>)((Y_j-<Y>)\Delta X_i^{-}+(X_i-<X>)\Delta Y_i^{-}}{{\left({\sum }_{i=1}^N{(X_i-<X>)}^2{\sum }_{j=1}^N{(Y_j-<Y>)}^2\right)}^{3/2}}$$

$$\Delta cor{r}^{-}(X,Y)=\frac{{\sum }_{i=1}^N(X_i-<X>)\Delta Y_i^{-}+(Y_i-<Y>)\Delta X_i^{-}}{\sqrt{{\sum }_{i=1}^N{(X_i-<X>)}^2{\sum }_{j=1}^N{(Y_j-<Y>)}^2}}+$$

$$+\frac{{\sum }_{i=1}^N{\sum }_{j=1}N(X_i-<X>)(Y_j-<Y>)((Y_j-<Y>)\Delta X_i^{+}+(X_i-<X>)\Delta Y_i^{+}}{{\left({\sum }_{i=1}^N{(X_i-<X>)}^2{\sum }_{j=1}^N{(Y_j-<Y>)}^2\right)}^{3/2}}$$

where $\Delta X_i^{+}$ is the positive direction error in the ith observation.

The results are shown in

Table 5. Correlations between parameters. Not included in the Table: the correlation between the ${\Gamma }_{upper}$ and 5–7 keV emission line flux $F_{5-7}^{line}$ is −0.48 ± 0.13.

Param.	${\mathit{E}}_{\mathbf{break}}$	$\mathbf{\Gamma }$	${\mathit{E}}_{\mathit{C}}$	${\mathbf{kT}}_{\mathit{e}}$	${\mathit{N}}_{\mathit{H}}$	lg${\mathit{\chi }}_{\mathit{i}}$	EW	${\mathit{F}}_{5.-7.}^{\mathbf{jet}}$	${\mathit{F}}_{5.-7.}^{\mathbf{nucl}}$
${\Gamma }_{upper}$	0.0${}_{-0.07}^{+0.08}$	0.24${}_{-0.06}^{+0.08}$	0.33${}_{-0.79}^{+0.21}$	0.36 ± 0.21	0.08${}_{-0.54}^{+0.46}$	−0.06 ± 0.24	−0.3${}_{-0.18}^{+0.25}$	0.27${}_{-0.15}^{+0.13}$	−0.31${}_{-0.11}^{+0.18}$
$E_{break}$	1	−0.27${}_{-0.08}^{+0.18}$	0.21${}_{-0.58}^{+0.32}$	0.25${}_{-0.10}^{+0.09}$	−0.5${}_{-0.22}^{+0.56}$	0.39${}_{-0.32}^{+0.12}$	−0.07${}_{-0.09}^{+0.20}$	−0.21${}_{-0.36}^{+0.14}$	0.38${}_{-0.16}^{+0.09}$
$\Gamma$	−0.27${}_{-0.08}^{+0.18}$	1	0.05${}_{-0.07}^{+0.01}$	−0.06${}_{-0.08}^{+0.17}$	0.33${}_{-0.36}^{+0.25}$	−0.47${}_{-0.08}^{+0.17}$	−0.58${}_{-0.04}^{+0.36}$	0.48${}_{-0.21}^{+0.28}$	−0.46${}_{-0.34}^{+0.27}$
$E_C$	0.21${}_{-0.58}^{+0.32}$	0.05${}_{-0.07}^{+0.01}$	1	0.99${}_{-0.25}^{+0.01}$	−0.2${}_{-0.04}^{+0.08}$	0.28${}_{-0.04}^{+0.13}$	0.22${}_{-0.06}^{+0.05}$	−0.29 ± 0.07	0.41${}_{-0.06}^{+0.26}$
$k{T}_e$	0.25${}_{-0.10}^{+0.09}$	−0.06${}_{-0.08}^{+0.17}$	0.99${}_{-0.25}^{+0.01}$	1	−0.22${}_{-0.01}^{+0.21}$	0.32${}_{-0.13}^{+0.01}$	0.23${}_{-0.21}^{+0.32}$	−0.3 ± 0.07	0.42 ± 0.05
$N_H$	−0.5${}_{-0.22}^{+0.56}$	0.33${}_{-0.36}^{+0.25}$	−0.2${}_{-0.04}^{+0.08}$	−0.22${}_{-0.01}^{+0.21}$	1	0.02 ± 0.2	−0.23 ± 0.13	0.03${}_{-0.26}^{+0.15}$	0.34${}_{-0.06}^{+0.22}$
lg${\chi }_i$	0.39${}_{-0.32}^{+0.12}$	−0.47${}_{-0.08}^{+0.17}$	0.28${}_{-0.04}^{+0.13}$	0.32${}_{-0.13}^{+0.01}$	0.02 ± 0.2	1	0.28 ± 0.17	-0.74${}_{-0.02}^{+0.3}$	0.65${}_{-0.44}^{+0.07}$
EW	−0.07${}_{-0.09}^{+0.20}$	−0.58${}_{-0.04}^{+0.36}$	0.22${}_{-0.06}^{+0.05}$	0.23${}_{-0.21}^{+0.32}$	−0.23 ± 0.13	0.28 ± 0.17	1	−0.58 ± 0.08	0.62 ± 0.1
$F_{5.-7.}^{jet}$	−0.21${}_{-0.36}^{+0.14}$	0.48${}_{-0.21}^{+0.28}$	−0.29 ± 0.07	−0.3 ± 0.07	0.03${}_{-0.26}^{+0.15}$	−0.74${}_{-0.02}^{+0.3}$	−0.58 ± 0.08	1	−0.9${}_{-0.03}^{+0.04}$
$F_{5.-7.}^{nucl}$	0.38${}_{-0.16}^{+0.09}$	−0.46${}_{-0.34}^{+0.27}$	0.41${}_{-0.06}^{+0.26}$	0.42${}_{\pm }$0.05	0.34${}_{-0.06}^{+0.22}$	0.65${}_{-0.44}^{+0.07}$	0.62 ± 0.1	−0.9${}_{-0.03}^{+0.04}$	1
$F_{5-7.}^{line}$	−0.24 ± 0.08	−0.36${}_{-0.12}^{+0.08}$	0.01${}_{-0.23}^{+0.39}$	0.05 ± 0.07	0.34${}_{-0.27}^{+0.11}$	0.24${}_{-0.06}^{+0.12}$	0.8 ± 0.14	−0.5 ± 0.12	0.42${}_{-0.1}^{+0.08}$

. As one can notice, there is a correlation between the high-energy exponential cut-off and the electronic temperature in the spectral counterpart corresponding to the primary nuclear (accretion disk + corona) emission, with a coefficient of 0.99${}_{-0.25}^{+0.01}$; their time dependencies are shown in Figure A7 in Appendix D, demonstrating prominent signs of synchronized variability during the observational period). There is a strong anti-correlation between jet and nuclear fluxes (at the level of −0.9${}_{-0.03}^{+0.04}$), which can be considered a sign of the close spatial proximity of these two components (i.e., it favors the “jet/disk cycle” rather than the “reconnection” scenario). There is also moderate anti-correlation between the ionization parameter and the jet flux level (−0.74${}_{-0.02}^{+0.3}$). The 5–7 keV nuclear and jet base flux curves are shown in Figure A6 in Appendix D. Surprisingly, the correlation between the nuclear and Fe-K$\alpha$ 5–7 keV emission fluxes was not traced, being at the level of 0.42; this can be considered a result of the ionization level variability [43]. The dependence of the ionization parameter upon the time is shown in Figure 5.

We plotted the dependencies of the parameters with high correlation/anti-correlation levels in Figure 6, Figure 7 and Figure 8; the other pairs with no prominent correlation are shown in Appendix C in Figure A2–Figure A5. As we can see from Figure 6, the dependency between these parameters looks almost linear with the coefficient between these two parameters determined as 3.04${}_{-0.85}^{+0.55}$ using the minimum squares method; similar values of the coefficient 2.2–5.96 were predicted for 3C 120 in [50].

4. Discussion

We applied the two recipes, giving us the possibility to distinguish between the disk/corona and the jet base contributions to the hard X-ray spectra of the RT AGN 3C 120 proposed in our previous work [27]. We analyzed the data of the XMM-Newton, INTEGRAL, SWIFT, and Suzaku observations during the 20 periods together with the radio data observed by the Planck mission. To perform this separation with the first recipe, we assumed the same values of the photon indices for the lower knees of the X-ray IC/SSC and the radio synchrotron jet base emission spectra of 3C 120 (taking into account that the radio emissions were pure jet base synchrotron emissions). The first method shows quite accurate results, despite the fact that the Planck radio spectrum used to fix the synchrotron emission photon index was averaged over time.

The other recipe we tried to use was based on the link between the emission fluxes of the disk/corona continuum and Fe K$\alpha$ line emissions [51,52]. The relative errors of the emission line parameters were too large in most observations, including the 14th one when the jet base emissions were estimated as zero-compatible, which resulted in very rough estimates of the linear coefficient and nuclear fluxes. Additionally, the ionization level in 3C 120 appears to vary significantly over time, breaking the linearity of the link between the Fe-K line and disk/corona continuum fluxes (the variance of the ionization parameter is 11.0). This is confirmed by the results of our correlation analysis (corr($F_{5-7}^{nucl}$,$F_{5-7}^{line}$) = 0.42${}_{-0.10}^{+0.08}$ is not enough to trace out the correlation). Thus, we can see that this second recipe is rather inappropriate for the case of 3C 120 as it can give us only the lower limit on the primary nuclear emission flux. However, taking into account the close values of the photon indices of the primary nuclear emission and the jet base spectra “upper knee”, the emission line presence during some period of time can be considered as a sign of the primary nuclear emission dominating in the spectrum.

We compare our results with the “disk-jet cycle” model [7] and the “plasmoid magnetic reconnection” model [25]. Following the “disk-jet cycle” model, the structural changes observable in an RT AGN can be interpreted in terms of the five-phase jet-disc cycle:

• Silent full-disc phase: The jet is weak or absent, the disc continues into its innermost stable orbit, and an object looks similar to a classical RQ AGN with prominent emission lines and high values (i.e., above 100 keV) of exponential high-energy cut-off in the corona/disk continuum;
• Phase of instability: The inner part of the disc becomes unstable, ionized, and/or heated, turbulent, and radiatively inefficient, causing luminescent Fe-K lines to weaken. At the same time, recombination lines of strongly ionized iron, such as the 6.58 keV Fe XXII and 6.64 keV Fe XXIV, can appear. During this phase, lower values of high-energy exponential cut-off in the primary nuclear continuum (below 100 keV) can be observed, and jet appearances are still weak or even absent;
• Jet outburst: Jet power increases rapidly and significantly. During this phase, the jet emission dominates, whereas the nuclear continuum can be invisible or weak in comparison with the jet component, with the exponential cut-off below 100 keV. The emission lines are weak as well. AGN looks similar to a typical RL at this stage;
• Jet exhausting: The jet power gradually decreases to the “normal” value or even lower. During this phase, the jet emission is no longer dominant, and the nuclear continuum can be at a similar level, with the exponential cut-off below 100 keV. The emission lines are typically weak at this stage;
• Disc refilling: The accretion disc begins to refill. During this phase, the AGN can look similar to an RQ AGN again. It can be characterized by the “weakness” of all of its counterparts: the flux from the jet base is zero-compatible, the iron K lines are weak (if present), and the exponential cut-off in the primary nuclear continuum is below 100 keV.

Within this model, there are no direct correlations between the flux levels (for the jet base emission or the primary nuclear one) and high-energy exponential cut-off or electronic temperature values. As for the ionization level, it should anti-correlate with the jet base flux (as lower ionization levels can indicate that the central regions of the disc/corona are disrupted by the jet outburst). It can optionally correlate with the primary nuclear flux. The nuclear and jet base fluxes, being fed from the same source (region), are supposed to be anti-correlated.

The “plasmoid magnetic reconnection” model predicts the following four phases:

• Quiescence: The jet is weak with its normal flux values; within the outer magnetically turbulent zone, there are small-scale plasmoids that form;
• Plasmoid envelope: The “monster” plasmoid forms and begins to move inwards down to the central black hole;
• Jet flare: The “monster” reconnects to the jet, inflicting the fast and significant growth of the jet power and curving the jet;
• Jet shape regain: The jet power gradually returns to its “quiescent” values; the jet itself regains its initial (i.e., direct) shape.

Within this model, there are no direct correlations between the flux levels and the high-energy exponential cut-off or electronic temperature. There is no correlation between the jet flux and nuclear flux, but there may be a correlation with the ionization level. Additionally, a curved jet can play a role in providing seed photons for the Fe-K emission lines that are Compton-reflected from the accretion disc surface.

As we can see from Table 3, the jet flux anti-correlates with the ionization level (see Figure 7) and with the primary nuclear flux. That is why we can say that our results favor the “jet-disc cycle” model.

However, there is still an interesting question that remains unanswered: what are the physical reasons causing such structural changes in the “central engine” of an RT AGN? A possible explanation for this structural instability could include some quite exotic alternatives to a central massive black hole, such as wormholes, “magnetic tunnels” [53,54], or weakly/strongly naked singularities [55].

A stream of charged particles passing through a “magnetic tunnel” would cause the appearance of a magnetic field around it, resulting in a “silent phase”. Due to the “Lorentz ionization effect” [56], the ionization level of the disc/corona would increase along with the magnetic field of the central object during the “turbulent phase”. When the magnetic field is strong enough, it launches jets during the “jet activity phase”, which takes away the disc plasma and interrupts its inflow into the tunnel. After the interruption of the jets, the disc remains truncated and begins to refill during the “refilling phase”.

A more accurate determination of the profiles of the emission lines formed in the near vicinity of the central object of the AGN can give us the possibility to test such exotic models [57].

Another important clue to understanding the scenarios of jet formation in this object and to determine the jet composition (i.e., hadronic or leptonic) lies at the highest energy ranges, leading to key information about their spectra above 10${}^{10}$–10${}^{11}$ eV [58]. Tracing out this difference can be performed in a more effective way by combining the existing Fermi-LAT observational data with the Cherenkov Telescope Array (CTA) results [59].

5. Conclusions

• We applied the recipes proposed in [27] to separate the “jet base” and “central engine” spectral components in the case of AGN 3C 120 in the range from X-ray to soft $\gamma$-ray spectrum for 20 periods of observations by Suzaku/XIS, XMM-Newton/EPIC, SWIFT, and INTEGRAL/ISGRI. Despite the close values of the photon indices corresponding to the jet base and primary nuclear emission, the results of the distinguishing components look optimistic enough.
• Our fitting results are in good agreement with the “jet-disc cycle” model [7].
• The maximal values of the Fe K$\alpha$ line-equivalent widths obtained during the “silent” phases are in satisfactory agreement with the values typical for RQ AGN.