Multiwavelength monitoring of NGC 1275 over a decade: Evidence of a shift in synchrotron peak frequency and long-term multi-band flux increase
Sanna Gulati, Debbijoy Bhattacharya, Subir Bhattacharyya, Nilay Bhatt, C. S. Stalin, V. K. Agrawal
aa r X i v : . [ a s t r o - ph . H E ] J a n MNRAS , 1–12 (2020) Preprint 28 January 2021 Compiled using MNRAS L A TEX style file v3.0
Multiwavelength monitoring of NGC 1275 over a decade: Evidenceof a shift in synchrotron peak frequency and long-term multi-bandflux increase
Sanna Gulati, Debbijoy Bhattacharya, ★ Subir Bhattacharyya , Nilay Bhatt , C. S. Stalin and V. K. Agrawal Manipal Centre for Natural Sciences, Centre of Excellence, Manipal Academy of Higher Education, Manipal - 576104, India Bhabha Atomic Research Centre, Mumbai - 400085, India Homi Bhabha National Institute, Anushaktinagar, Mumbai - 400094, India Indian Institute of Astrophysics, Bangalore - 560034, India Space Astronomy Group, U R Rao Satellite Centre, Bangalore - 560017, India
Accepted XXX. Received YYY; in original form ZZZ
ABSTRACT
We carried out a detailed study of the temporal and broadband spectral behaviour of one ofthe brightest misaligned active galaxies in 𝛾 -rays, NGC 1275 utilising 11 years of Fermi , andavailable
Swift and
AstroSat observations. Based on the cumulative flux distribution of the 𝛾 -ray lightcurve, we identified four distinct activity states and noticed an increase in the baselineflux during the first three states. Similar nature of the increase in the average flux was alsonoticed in X-ray and UV bands. A large flaring activity in 𝛾 -rays was noticed in the fourth state.The source was observed twice by AstroSat for shorter intervals ( ∼ days) during the longerobserving periods ( ∼ years) state 3 and 4. During AstroSat observing periods, the source 𝛾 -rayflux was higher than the average flux observed during longer duration states. The increase inthe average baseline flux from state 1 to state 3 can be explained considering a correspondingincrease of jet particle normalisation. The inverse Comptonisation of synchrotron photonsexplained the average X-ray and 𝛾 -ray emission by jet electrons during the first three longerduration states. However, during the shorter duration AstroSat observing periods, a shift of thesynchrotron peak frequency was noticed, and the synchrotron emission of jet electrons wellexplained the observed X-ray flux.
Key words: galaxies: active — galaxies: jets — gamma-rays: galaxies — X-rays: galaxies—quasar: individual (NGC 1275)
NGC 1275 is one of the brightest nearby radio galaxies (z = . . + 𝛾 -rays. Abdo et al. (2009) reported thediscovery of high energy 𝛾 -ray emission from NGC 1275 utilisingthe first few months of observations from the Large Area Telescope(LAT) onboard the Fermi 𝛾 -ray space telescope ( Fermi ). ★ E-mail: debbij[email protected]
Evidence of variability, in both, long (Kataoka et al. 2010;Dutson et al. 2014) as well as short timescales with large flaring ac-tivities (Donato et al. 2010; Brown & Adams 2011; Ciprini 2013;Pivato & Buson 2015; Baghmanyan et al. 2017; Kushwaha et al.2017; Tanada et al. 2018; Chitnis et al. 2020; Ghosal et al. 2020)in 𝛾 -rays was noticed in this source. NGC 1275 was also de-tected by Major Atmospheric Gamma Imaging Cherenkov ( MAGIC )telescope and Very Energetic Radiation Imaging Telescope Ar-ray System (
VERITAS ) in very high energy 𝛾 -rays (Aleksić et al.2012, 2014; Benbow & VERITAS Collaboration 2015; Mirzoyan2016, 2017; Mukherjee & VERITAS Collaboration 2016, 2017;MAGIC Collaboration et al. 2018). Though earlier MAGIC obser-vations showed marginal flux variation in monthly timescales,MAGIC Collaboration et al. (2018) reported a presence of signifi-cant variation in “day-by-day” 𝛾 -ray lightcurve.NGC 1275 was studied in hard X-ray band using obser-vations from Nuclear Spectroscopic Telescope Array - NuS- © S. Gulati et al.
TAR (Tanada et al. 2018; Rani et al. 2018; Chitnis et al. 2020).Rani et al. (2018) found that the emission above 20 keV is dom-inated by a non-thermal component with possible jet origin. NGC1275 also exhibits correlated variability in different wavebands(Aleksić et al. 2014; Fukazawa et al. 2018). The broadband spec-tral energy distribution (SED) of NGC 1275 has been explainedby one-zone synchrotron self Compton model (Abdo et al. 2009;Aleksić et al. 2014; Fukazawa et al. 2018; Tanada et al. 2018) or astructured jet (Tavecchio & Ghisellini 2014).In this work, we have carried out a long-term study of NGC1275 in 𝛾 -rays utilising 11 years of Fermi -LAT observations andidentified different activity states. This source was also observedin X-rays, UV and/or optical band by
Swift -XRT and
Swift -UVOTmultiple times during
Fermi observing period. India’s first multi-wavelength astronomical observatory “
AstroSat ” (Agrawal 2006;Singh et al. 2014; Rao et al. 2016), also observed the source twice.
AstroSat is capable of observing the sky simultaneously over awide range of energies covering from near-UV (NUV) and far-UV (FUV) bands using the Ultra-Violet Imaging Telescope (UVIT;Kumar et al. 2012; Tandon et al. 2017), soft X-ray band with theSoft X-ray Telescope (SXT; Singh et al. 2017) to the hard X-rayband with the Large Area X-ray Proportional Counter (LAXPC;Yadav et al. 2016b; Antia et al. 2017) and Cadmium-Zinc-TellurideImager (CZTI; Vadawale et al. 2015; Rao et al. 2017).We have carried out construction and modelling of averagebroadband SEDs during different states as identified in the 𝛾 -raylight curve and during AstroSat observing periods to understand thelong term behaviour of this source. Details of the analysis of multi-wavelength data used in this work are given in Section 2. In Section3, we discuss our findings, followed by a conclusion in Section 4.
We used the 𝛾 -ray data from Fermi -LAT covering the period from2008-08-04 to 2019-08-04 (11 years) in the energy range from100 MeV to 100 GeV. The data reduction is performed using theFermitools version 1.2.23 The current version of
Fermi -LAT data,pass 8 P8R3, was used for analysis (Bruel et al. 2018). InstrumentResponse Function (IRF)
P8R3_SOURCE_V2 was used for sourceclass event selection. A 15 ◦ × ◦ region of interest (ROI) centredat NGC 1275 was defined, and standard cuts were applied to se-lect the good time intervals ( 𝑧 𝑚𝑎𝑥 < ◦ , DATA_QUAL > LAT_CONFIG == Fermipy version 0.19.0 (Wood et al. 2017)was used to calculate the light curve and spectra using binnedlikelihood analysis. A spatial binning of 0 . ◦ pixel − and eightlogarithmically-spaced energy bins per decade were chosen. Ourinitial model, generated using make4fglxml.py , consists of all 𝛾 -ray point sources within 20 ◦ of the ROI centre included in the4FGL-DR2 catalogue (Ballet et al. 2020) and standard templatesfor Galactic diffuse emission model ( gll_iem_v07.fits ) and isotropicdiffuse emission ( iso_P8R3_SOURCE_V2_v1.txt ) as used in thefourth Fermi catalogue (4FGL: Abdollahi et al. (2020)).
To calculate the average flux for the 11 years dataset, we begin withan initial automatic optimisation of the ROI by iteratively fitting https://fermi.gsfc.nasa.gov/ssc/data/analysis/user/ the sources using the optimize method of Fermipy . It is recom-mended by the
Fermipy developers to run this method at the startof the analysis “to ensure that all parameters are close to their globallikelihood maxima” . After that, both normalisations and spectralparameters of the sources within 5 ◦ and only normalisations of thesources lying within 12 ◦ of the ROI centre were left to vary. Wefreeze the spectral parameters for sources having Test Statistic (TS) < − . Following Meyer et al. (2019), the normalisa-tions of the Galactic and isotropic diffuse backgrounds, includingthe spectral index of the Galactic diffuse background template, wereleft free during the fit. A TS map was generated using findsource tool of Fermipy to search for additional point sources, that are notpresent in the 4FGL-DR2 catalogue. No new sources were detectedwith TS ≥ 𝑑𝐹𝑑𝐸 = 𝑁 (cid:18) 𝐸𝐸 (cid:19) − 𝛼 and log-parabola 𝑑𝐹𝑑𝐸 = 𝑁 (cid:18) 𝐸𝐸 𝑏 (cid:19) − 𝛼 − 𝛽 log (cid:16) 𝐸𝐸𝑏 (cid:17) models.Here, 𝑑𝐹𝑑𝐸 and 𝑁 are the differential flux and normalisationfactor, respectively in the unit of photon cm − s − MeV − . 𝐸 is theenergy, 𝐸 and 𝐸 𝑏 are the scale and break value, respectively inthe unit of MeV. 𝛼 and 𝛽 are the spectral parameters. The sourcewas considered to be detected if its TS >
25, which correspondsto ∼ 𝜎 confidence (Mattox et al. 1996). The source spectrum isconsidered significantly curved if two times the difference in log-likelihood value for log-parabola and log-likelihood value for power-law ( 𝑇 𝑆 𝑐𝑢𝑟 𝑣𝑒 ) is greater than 16 (Acero et al. 2015). The averagespectrum of NGC 1275, utilising the 10 years of
Fermi observationsis reported to be significantly curved (Ballet et al. 2020). Utilisingthe 11 years data set, we also noticed a significant curvature in thesource spectrum.
The monthly averaged 𝛾 -ray light curves in the energy bands 100MeV to 100 GeV, 100 MeV to 1 GeV, and 1 GeV to 100 GeV werecomputed using lightcurve tool of Fermipy . The best fit modelobtained for the 11 years dataset considering a power-law spectrumfor NGC 1275 was used as an input model. While constructingthe 𝛾 -ray lightcurve in 100 MeV to 100 GeV energy band, bothnormalisations and spectral parameters of the sources within 3 ◦ andonly normalisations of sources within 3 ◦ to 12 ◦ from ROI centrewere left free to vary in the input model of each time bin. For 𝛾 -raylightcurves in 100 MeV to 1 GeV energy band and 1 GeV to 100GeV energy band, only normalisations of the sources within 12 ◦ from ROI centre were left free to vary in the input model for eachtime bin. For all the three light curves, the normalisations of theGalactic and isotropic diffuse emission models were left free, andthe spectral index of the Galactic emission was frozen to the 11years averaged value. A signature of increase in the baseline fluxwas noticed from 2008 to 2017 (Fig 1-a). This was followed by alarge, broad flare until 2019. https://fermipy.readthedocs.io/en/latest/ MNRAS000
The monthly averaged 𝛾 -ray light curves in the energy bands 100MeV to 100 GeV, 100 MeV to 1 GeV, and 1 GeV to 100 GeV werecomputed using lightcurve tool of Fermipy . The best fit modelobtained for the 11 years dataset considering a power-law spectrumfor NGC 1275 was used as an input model. While constructingthe 𝛾 -ray lightcurve in 100 MeV to 100 GeV energy band, bothnormalisations and spectral parameters of the sources within 3 ◦ andonly normalisations of sources within 3 ◦ to 12 ◦ from ROI centrewere left free to vary in the input model of each time bin. For 𝛾 -raylightcurves in 100 MeV to 1 GeV energy band and 1 GeV to 100GeV energy band, only normalisations of the sources within 12 ◦ from ROI centre were left free to vary in the input model for eachtime bin. For all the three light curves, the normalisations of theGalactic and isotropic diffuse emission models were left free, andthe spectral index of the Galactic emission was frozen to the 11years averaged value. A signature of increase in the baseline fluxwas noticed from 2008 to 2017 (Fig 1-a). This was followed by alarge, broad flare until 2019. https://fermipy.readthedocs.io/en/latest/ MNRAS000 , 1–12 (2020) ultiwavelength monitoring of NGC 1275 In a previous work, Tanada et al. (2018) studied 𝛾 -ray vari-ability of NGC 1275 utilising ∼ Fermi observations(2008 − 𝛾 -ray light curveutilising 11 years of Fermi observations based on the increase inthe baseline flux nature. To define different states, we calculatedthe cumulative flux of the monthly averaged 𝛾 -ray light curve. Aconstant slope in cumulative lightcurve corresponds to insignificantvariation in flux. Whereas, an increase/decrease in slope indicatesa rise/fall in baseline flux. Visual inspection of the cumulative fluxdistribution suggests the presence of four distinct regions/states.The first three states exhibit linear feature with increase in slopes,whereas, a significant non-linearity was noticed beyond that (statefour). To constrain the boundaries of these states, data from shadedregions, as shown in Fig 1-b, were fitted with linear functions. Theseregions were chosen in a way to avoid edge and/or transition effect.The mid-point of the monthly bin that contains the intersection ofthe fitted linear functions in the first and the second intervals isconsidered as the upper boundary of state 1 (S1). Similarly, themid-point of the monthly bin that contains the intersection of thefitted linear functions in the second and the third intervals is consid-ered as the upper boundary of state 2 (S2). A noticeable deviationfrom the fitted linear function in interval 3 is considered as the upperboundary of state 3 (S3). Boundaries of these states are indicatedin Fig 1-a. Beyond state 3, a significant non-linearity was noticedin the cumulative flux, first a sharp increase followed by a gradualdecrease. This region represents the state 4 (S4). The details of theboundaries of these states are mentioned in Table 1. Fig 1-c showsthe variation of the monthly averaged spectral parameter 𝛼 of thissource. The epoch A and epoch B as defined in Tanada et al. (2018)overlap with S1-S2 and S2-S3, respectively. 𝛾 -ray spectral analysis was carried out in different activity statesto study the behaviour of the source. The average flux in the fouractivity states as defined in the 𝛾 -ray light curve was calculatedfollowing the methodology used for spectral analysis of the entiredata set (Sec. 2.1.1). To calculate the average flux in an activitystate, the spectral index of the Galactic diffuse background templatewas kept frozen to the 11 years averaged value. For all the fourstates, it was found that the log-parabola spectral model was stronglypreferred over the power-law model. The best fit values of spectralparameters are given in Table 1. The spectral parameter 𝛼 of thesource in all four states lies in the range 2 . − .
1. An increase ofaverage 𝛾 -ray flux was noticed from S1 to S4. 𝛾 -ray SEDs wereconstructed for 8 logarithmically-spaced energy bins per decade in100 MeV to 100 GeV energy band using the sed tool of Fermipy for these four states. The spectral index in each energy bin wasfrozen to the power-law approximation (local index) to the shape ofthe global spectrum, while normalisation was left to vary.
AstroSat observed NGC 1275 on 2017 January 12-13 underTarget of Opportunity (ToO: hereafter referred to as AS1) whichfalls in the later phase of the S3 state. NGC 1275 was observedagain with
AstroSat on 2017 September 26-27 (referred to as AS2in the rest of the paper) which falls at the beginning of the S4 state.The 𝛾 -ray analysis during AS1 and AS2 was carried out followinga similar methodology as used in S1-S4 states. However, due tolimited photon statistics, a simple power-law was considered tomodel the source spectrum during these AS1 and AS2 states. The normalisations and spectral parameters of sources within 3 ◦ , onlynormalisations of sources within 3 ◦ to 12 ◦ from ROI centre and,normalisations of the Galactic and isotropic emission were left freeto vary. We freeze the normalisations and spectral parameters forsources having TS < − . While constructing the 𝛾 -ray SED during AS1 and AS2, twoenergy bins per decade in the full energy band were considered.Due to low photon statistics, the data in the last two energy bandsin the 𝛾 -ray SED was merged into a single energy band. The fluxand spectral parameters during AS1 and AS2 observing periodsare given in Table 1. The 𝛾 -ray flux during AS1 and AS2 (2 daysof observation) is higher than the average flux during the longerduration states (a few years of observation). For X-rays, we used data from the
Swift
X-ray Telescope (
Swift -XRT;Burrows et al. 2005) that covers the energy range for 0 . −
10 keVas well as SXT and LAXPC onboard
AstroSat . The energy rangesof SXT and LAXPC are 0 . − −
80 keV, respectively.
Swift -XRT
Though there was no continuous monitoring of this source in the X-ray band, a significant number of pointed observations were carriedout by
Swift -XRT. Fukazawa et al. (2018) analysed archival
Swift -XRT data of NGC 1275 from 2007-2015. They derived flux for eachobservation ID and calculated the light curve. In this work, we havecalculated the light curve in a similar way utilising all the archival
Swift -XRT photon counting (PC) mode observations covering S1,S2 and S3 states. Only one observation is present in S1, whichis also studied by Fukazawa et al. (2018). There are 16, and 50 PCmode observations present during S2 and S3 state, respectively. Theanalysis results of 14 and 4 PC mode observations during S2 andS3, respectively were given in Fukazawa et al. (2018). During theAS1 state, there are no Swift -XRT PC mode observations available.However, one
Swift -XRT observation in windowed timing (WT)mode (Obs ID: 00031770015, exposure time: ∼
239 s) was presentduring this period, which was used in this work. We processedthe XRT data using the xrtdas (v 3.4.1) package distributed under heasoft (v 6.24). The task xrtpipeline (v 0.13.4) was used to cleanand calibrate level 1 data files with standard filtering criteria andusing calibration files caldb (v 20190910). The source spectrumwas binned to have 20 counts per bin using grppha . The XRTspectra were fitted in xspec (v 12.10.0c).For PC mode data, the source spectrum was derived consider-ing an annular region from 12 ′′ − ′′ . The central 12 ′′ region isblocked since the PC mode data suffers from pile-up events. Back-ground spectra were extracted in 60 ′′ − ′′ from the centre toaccount for the cluster emission. Considering the significant differ-ence in the response functions of PC mode and WT mode of Swift -XRT below 1 . . − . https://heasarc.gsfc.nasa.gov/cgi-bin/W3Browse/w3browse.pl https://swift.gsfc.nasa.gov/analysis/xrt_swguide_v1_2.pdf MNRAS , 1–12 (2020)
S. Gulati et al. F l u x ( - ph / c m / s e c ) -2-10123456 (a)(b)(c) C u m u l a t i v e f l u x ( - ph / c m / s e c ) α MJD
Figure 1. (a): Monthly averaged 𝛾 -ray light curve ( MeV-
GeV) for NGC 1275 utilising years of Fermi observation (2008-2019). (b): Plot of cumulativeflux with time. Data in shaded regions was considered for fitting linear function. The best fit linear functions in the three shaded regions are represented bythe dotted line (slope: . × − ph/cm /sec/day), dashed line (slope: . × − ph/cm /sec/day) and dash-dotted (slope: . × − ph/cm /sec/day) linerespectively. The vertical lines in (a) and (b) define the boundary of the different states. (c) Variation of the monthly averaged spectral parameter 𝛼 for NGC1275 utilising years of Fermi observation (2008-2019).
Table 1. 𝛾 -ray flux and spectral indices for different activity statesInterval Start Date End Date Flux ( × − ) alpha beta TScurveMJD MJD ph/cm /sec(1) (2) (3) (4) (5) (6) (7)State 1 (S1) Aug 05 ( ) June 11 ( ) . ± .
07 2 . ± .
02 0 . ± .
01 19 . State 2 (S2)
June 11 ( ) Feb 25 ( ) . ± .
08 2 . ± .
01 0 . ± .
007 79 . State 3 (S3)
Feb 25 ( ) July 04 ( ) . ± .
06 2 . ± .
008 0 . ± .
005 212 . State 4 (S4)
July 04 ( ) Aug 05 ( ) . ± .
18 2 . ± .
02 0 . ± .
01 129 . AS1
Jan 12 ( ) Jan 14 ( ) . ± . . ± . - . AS2
Sep 26 ( ) Sep 28 ( ) . ± . . ± . - . MNRAS000
Sep 26 ( ) Sep 28 ( ) . ± . . ± . - . MNRAS000 , 1–12 (2020) ultiwavelength monitoring of NGC 1275 respectively. The Galactic hydrogen column density ( 𝑁 𝐻 ) frozento 1 . × cm − (Yamazaki et al. 2013; Tanada et al. 2018). Wefix the temperature and abundance of the “apec” model to 4 . .
65 solar, respectively (Fukazawa et al. 2018) and estimate thenormalisation of the “apec” model in each observation. The “apec”normalisation was then frozen to the median value of 0 . , andthe AGN flux and photon index was derived in each observation ID.AGN flux was derived using “cflux” routine in XSPEC . No significantvariation in photon indices was noticed during S2 and S3 states. Thesource flux exhibits noticeable variations in both S2 and S3 stateswith a fractional variance of 23% ±
7% and 29% ± 𝛾 -rays, a signatureof average flux enhancement without any significant change in thephoton index from S1 to S3 was observed.The WT mode data during the AS1 period was analysed fol-lowing the methodology described in Fukazawa et al. (2018). Thespectra was extracted within 0 . ′ of NGC 1275. The correspond-ing extraction region is 36 × in the sky. All the PCmode data was used to extract the background spectra from the samesky region as that of WT mode spectra after excluding the central36 ×
36 arcsec , and subtracted from the WT mode spectra. A 3%systematic error was considered during the spectral analysis of WTmode data. Parameters of the “apec” model namely, temperature,abundance, and normalisation were fixed to 4 . .
60 solar and0 . Swift -XRT light curves of NGC 1275. Fromvisual inspection of their light curves, for the data in common, wefound good agreement between our results and that of Imazato et al.(2021).
AstroSat
The Level 1 SXT data was analysed using “sxtpipeline” of the SXTsoftware “as1sxtlevel2-1.4b” . The clean events were merged us-ing the “sxtpyjuliamerger_v01” . A circular region of 15 ′ radiuswas used as a source region, and for the background, the instru-ment team provided “SkyBkg_comb_EL3p5_Cl_Rd16p0_v01.pha” file was used. The ancillary response file was created by using “sx-teefmodule_v02” . A 2% systematic error was considered duringspectral analysis.LAXPC data was analysed using the analysis software “laxpc_soft” package (May 19, 2018 version) available at the AstroSat
Science Support Cell . Standard procedures were used The “apec” normalisation, in units of − 𝜋 [ 𝐷 𝐴 ( + 𝑧 )] ∫ 𝑛 𝑒 𝑛 𝐻 𝑑𝑉 , where 𝐷 𝐴 is the angular diameter distance to the source in cm, 𝑛 𝑒 and 𝑛 𝐻 are theelectron and Hydrogen densities in cm − , respectively (followed throughoutthe paper). http://astrosat-ssc.iucaa.in/?q=data_and_analysis to reduce the Level 1 data (Yadav et al. 2016a; Antia et al. 2017).The spectra for the source and background in the energy range of4 . − . . .
05 for energy grouping factor. Since thesize of the emission region is large, the systematic error of LAXPCis expected to be high. We considered a 3% systematic error forLAXPC spectral analysis.Due to large PSF of
AstroSat -SXT ( ∼ ′ ) and AstroSat -LAXPC (1 ◦ × ◦ ), the observed X-ray flux of NGC 1275 sufferssignificant contamination from the Perseus cluster. Therefore, to ob-tain intrinsic emission from the nucleus of NGC 1275, it is essentialto constrain cluster parameters. For SXT and LAXPC analysis, thecluster abundance was frozen to the value of 0 .
42 solar, which wasobtained by averaging the data from Fig. 6 of Churazov et al. (2003)over a region of 15 ′ radius. This value of abundance is consistentwith other studies (Schmidt et al. 2002; Nishino et al. 2010). Thetemperature and abundance of the “apec” model, were constrainedutilising simultaneous Swift -XRT observation.As during
Swift -XRT analysis data below 1 . . − . ∼
33 ks), the photon index and the unabsorbedflux of the power-law component were kept frozen to the valuesderived from the simultaneous
Swift -XRT observation to estimatethe temperature (5 . ± .
09 keV) and normalisation (0 . ± . ∼
17 ks), the temperature and normalisationwere kept frozen to the best fit values obtained during the AS1observing period, and the unabsorbed flux from NGC 1275 wasobtained (Table 4).During LAXPC analysis in the AS1 period (exposure time ∼
50 ks), the photon index and the unabsorbed flux of the power-law component (emission from NGC 1275) were kept frozen tothe values derived from simultaneous
Swift -XRT observation in4 . − . . ± . . ± .
02) of the “apec” modelwere obtained. During the AS2 observing period (exposure time ∼
16 ks), the “apec” temperature and normalisation were kept frozento the values obtained during the AS1 period. Due to low photonstatistics, the photon index of the source could not be constrainedwith LAXPC observation. Therefore, the power-law photon indexwas kept frozen to the best fit value obtained during simultaneousSXT observation. The unabsorbed flux from NGC 1275 in 4 . − . MNRAS , 1–12 (2020)
S. Gulati et al.
Table 2: Summary of
Swift -XRT PC mode observations . ★ marks the observation IDsanalysed by Fukazawa et al. (2018).State Sequence No. Date Exposure Time Flux (in × − erg/cm /sec) Photon index(seconds)S1 ∗ . ± . . ± . S2 ∗ . ± . . + . − . ∗ . ± . . ± . ∗ . ± . . + . − . ∗ . ± . . ± . . ± . . + . − . ∗ . + . − . . + . − . ∗ . ± . . ± . ∗ . + . − . . ± . ∗ . ± . . ± . ∗ . + . − . . ± . ∗ . + . − . . ± . ∗ . + . − . . + . − . ∗ . + . − . . + . − . ∗ . ± . . ± . S3 ∗ . ± . . ± . ∗ . ± . . ± . ∗ . + . − . . ± . . ± . . + . − . ∗ . + . − . . + . − . . ± . . ± . . + . − . . + . − . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . + . − . . + . − . . ± . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . ± . . ± . . ± . . + . − . . ± . . + . − . . + . − . . + . − . . ± . . ± . . ± . . + . − . . + . − . . + . − . . ± . . + . − . . ± . . ± . . ± . . + . − . . + . − . . + . − . . ± . . + . − . . ± . . + . − . . ± . . + . − . . ± . . + . − . . ± . . ± . . ± . . ± . . ± . . + . − . . ± . . ± . . ± . . ± . . ± . . + . − . . ± . . ± . . + . − . . + . − . . ± . . ± . . ± . . + . − . . + . − . MNRAS000
Swift -XRT PC mode observations . ★ marks the observation IDsanalysed by Fukazawa et al. (2018).State Sequence No. Date Exposure Time Flux (in × − erg/cm /sec) Photon index(seconds)S1 ∗ . ± . . ± . S2 ∗ . ± . . + . − . ∗ . ± . . ± . ∗ . ± . . + . − . ∗ . ± . . ± . . ± . . + . − . ∗ . + . − . . + . − . ∗ . ± . . ± . ∗ . + . − . . ± . ∗ . ± . . ± . ∗ . + . − . . ± . ∗ . + . − . . ± . ∗ . + . − . . + . − . ∗ . + . − . . + . − . ∗ . ± . . ± . S3 ∗ . ± . . ± . ∗ . ± . . ± . ∗ . + . − . . ± . . ± . . + . − . ∗ . + . − . . + . − . . ± . . ± . . + . − . . + . − . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . ± . . + . − . . + . − . . ± . . + . − . . + . − . . + . − . . + . − . . + . − . . + . − . . ± . . ± . . ± . . + . − . . ± . . + . − . . + . − . . + . − . . ± . . ± . . ± . . + . − . . + . − . . + . − . . ± . . + . − . . ± . . ± . . ± . . + . − . . + . − . . + . − . . ± . . + . − . . ± . . + . − . . ± . . + . − . . ± . . + . − . . ± . . ± . . ± . . ± . . ± . . + . − . . ± . . ± . . ± . . ± . . ± . . + . − . . ± . . ± . . + . − . . + . − . . ± . . ± . . ± . . + . − . . + . − . MNRAS000 , 1–12 (2020) ultiwavelength monitoring of NGC 1275 Table 3.
Swift -XRT analysis results for different activity statesActivity state F1.6-10.0 keV 𝑎 Γ 𝑏 State 1 . ± . . ± . State 2 . ± . . ± . State 3 . ± .
09 1 . ± . AS1 . + . − . . + . − . 𝑎 − erg cm − s − 𝑏 Table 4.
AstroSat -SXT and LAXPC analysis results during AS2 periodInstrument Energy range F 𝑎 Γ 𝑏 𝜒 /dof(keV)SXT . − . . ± . . ± . . / LAXPC . − . . ± . . 𝑐 . / 𝑎 Unabsorbed flux in units of − erg cm − s − 𝑏 Photon index of power-law model 𝑐 Photon index kept frozen to the best fit value obtained duringsimultaneous SXT observation.
Table 5.
Swift -UVOT flux values in different UVOT filters for differentactivity statesFilters S1 𝑎 S2 𝑎 S3 𝑎 AS1 𝑎 V − . ± . . ± . − B − . ± . . ± . − U − . ± . . ± . − UVW1 . ± . . ± . . ± . − UVM2 − . ± . . ± . − UVW2 − . ± . . ± . . ± . 𝑎 Flux in units of − erg cm − s − Hz − Table 6.
AstroSat -UVIT analysis resultsFilter 𝜆 𝑚𝑒𝑎𝑛 (Å) Flux 𝑎 CaF2-1 (F1) . ± . Silica (F5) . ± . NUVB15 (F2) . ± . NUVB4 (F5) . ± . 𝑎 Flux in − erg cm − s − Hz − For data in the UV/optical bands, we used both
Swift -UVOT andUVIT onboard
AstroSat . From
Swift -UVOT, we have observationsin 𝑈 , 𝐵 , 𝑉 , 𝑈𝑉𝑊 𝑈𝑉𝑊
2, and
𝑈𝑉 𝑀 were analysed using different tasks,which are a part of heasoft (v 6.24), and 20170922 version of caldb . uvotimsum task was used to merge the different observa-tions during a particular epoch.For photometry, a source region with radius ∼ point spreadfunction was chosen to reduce cluster/host galaxy contribution. Acircular source region with 3 ′′ radius centred at the source positionand an annular background region with inner and outer radii of 15 ′′ https://heasarc.gsfc.nasa.gov/cgi-bin/W3Browse/w3browse.pl and 20 ′′ was used. uvotsource task was used to get the source mag-nitude. The Galactic extinction was calculated using Cardelli et al.(1989) and Schlafly & Finkbeiner (2011). The extinction corrected 𝐴𝐵 magnitudes were then converted to flux (erg cm − s − ). DuringS2 and S3 states, the source was observed in all the Swift -UVOT fil-ters. In the state S2, the exposure times for 𝑉 , 𝐵 , 𝑈 , 𝑈𝑉𝑊 𝑈𝑉 𝑀
𝑈𝑉𝑊 ∼ ∼ ∼ ∼
11 ks, ∼ ∼
17 ks, respectively. For the state 3, the exposure times for 𝑉 , 𝐵 , 𝑈 , 𝑈𝑉𝑊 𝑈𝑉 𝑀
𝑈𝑉𝑊 ∼ ∼ ∼
23 ks, ∼ ∼
23 ks, ∼
30 ks seconds respectively. However, during S1 andAS1 states, the source was only observed in
𝑈𝑉𝑊 ∼ 𝑈𝑉𝑊
Swift -UVOT light curves of NGC 1275 were recently pre-sented by Imazato et al. (2021). As they have corrected for the hostgalaxy contribution to the observed UV/optical emission, it is likelythat the flux values quoted in Table 5 are marginally brighter thanthat of Imazato et al. (2021).Similar to the X-ray and 𝛾 -ray bands, an increase in the averagesource flux was noticed from S1 to S3. Utilising the fluxes obtainedin three optical and three UV filters and considering a power-lawspectral shape, averaged energy spectral indices of optical and UVfluxes were derived in S2 and S3 states. An energy spectral indexof 1 . ± . . ± .
01 was obtained in the optical band forthe S2 and S3 states respectively. Similarly, in the UV band, energyspectral index of 1 . ± . . ± . AstroSat -UVIT under AS2 ob-servation in FUV filters: CaF2-1 (exposure time ∼ ∼ ∼ ∼ .
3) was used to carry out standard photometry using iraf . Acircular aperture of 5 pixels ( ∼ . ′′ ) and background region of15 −
20 pixels ( ∼ . ′′ − ∼ . ′′ ) was used for photometry. Thederived magnitudes were converted into fluxes (Tandon et al. 2017)and corrected for Galactic extinction. The estimated flux was cor-rected for the chosen aperture size using Table 11 of Tandon et al.(2020). The Galactic extinction corrected fluxes in four UVIT filtersduring AS2 observing period are given in Table 6. Similar to the long-term variability as observed in blazars (e.g.,Bhattacharya et al. 2013, 2017), Abdo et al. (2009) noticed that the 𝛾 -ray flux observed during the first few months of Fermi obser-vations was nearly a factor of ten higher than the EGRET fluxupper limit suggesting a presence of flux variability of a muchlonger timescale in NGC 1275. Dutson et al. (2014) reported thepresence of a decade timescale variability in this source. Based onthe first year of
Fermi observations of NGC 1275, Kataoka et al. IRAF is distributed by the National Optical Astronomy Observatory,which is operated by the Association of Universities for Research in As-tronomy (AURA)under a cooperative agreement with the National ScienceFoundationMNRAS , 1–12 (2020)
S. Gulati et al. F l u x ( - e r g / c m / s e c ) Time (MJD)State-2 average 0.511.522.533.54 55360 55400 55440 55730 55770 P ho t on i nde x Time (MJD)State-2 average
Figure 2.
Swift -XRT light curve and variation of photon index for NGC 1275 for state 2. Left panel: Variation in . − keV flux. Right panel: Variation inphoton index. The solid horizontal lines in both plots represent the weighted average and dashed horizontal lines represent the error in weighted average. F l u x ( - e r g / c m / s e c ) Time (MJD)State-3 average 0.511.522.533.544.5 56500 57100 57300 57700 57800 57900 P ho t on i nde x Time (MJD)State-3 average
Figure 3.
Swift -XRT light curve and variation of photon index for NGC 1275 for state 3. Left panel: Variation in . − keV flux. Right panel: Variation inphoton index. The solid horizontal lines in both plots represent the weighted average and dashed horizontal lines represent the error in weighted average. (2010) noticed a monthly timescale variability of the flux andthe spectral indices. Short-term variability on hours and sub-weektimescales was also noticed in this source (Baghmanyan et al. 2017;Brown & Adams 2011). The hour-timescale variability reported byBaghmanyan et al. (2017) suggests an extremely compact emittingregion. Short term ( ∼ hours) variability in the optical band was alsonoticed in a few other 𝛾 -ray detected misaligned active galaxies(Bhattacharya et al. 2019).Tanada et al. (2018) studied the 𝛾 -ray variability of NGC 1275utilising ∼ Fermi observations and defined two epochs(epoch A and epoch B) in the light curve based on the fluctuationsin the spectral indices. In the present study, we have defined 4activity states based on the increase in the baseline flux naturenoticed in the monthly averaged 𝛾 -ray light curve of NGC 1275as described in Section 2.1.2. As represented in the top panels ofFig. 4, no significant correlation was noticed between the monthlyaveraged 𝛾 -ray flux and the spectral parameter 𝛼 during S1, S2,and S3 states with Pearson correlation coefficient value 0 . − . − . ∼ − . 𝛾 -ray fluxes in 100 MeV- 1 GeV and 1 GeV - 100 GeV energy bands, which is representedin bottom panels of Fig. 4. Similar to the findings of the flux-spectral parameter correlation study, no significant correlation wasnoticed between the flux and the hardness ratio values during S1, S2,and S3 states with the Pearson correlation coefficient value − . .
1, and 0 .
2, respectively. However, a significant correlation wasfound during S4 with Pearson correlation coefficient 0 .
6, supportinga ‘harder when brighter’ scenario. Tanada et al. (2018) found acorrelation between the 𝛾 -ray flux and the hardness ratio duringone of the flares in epoch A of their study. However, in epoch B oftheir study, which overlaps with S2 and S3 states of this work, nosignificant correlation was noticed.A strong positive correlation between the optical and the 𝛾 -rayfluxes was noticed by Aleksić et al. (2014) between October 2009to February 2011. Fukazawa et al. (2018) reported the first positivecorrelation between fluxes in the X-ray and the 𝛾 -ray energy bandsduring 2013 − 𝛾 -ray, X-ray and UV/optical fluxes averaged over S1, S2 and S3states without any appreciable change in the spectral indices. MNRAS000
6, supportinga ‘harder when brighter’ scenario. Tanada et al. (2018) found acorrelation between the 𝛾 -ray flux and the hardness ratio duringone of the flares in epoch A of their study. However, in epoch B oftheir study, which overlaps with S2 and S3 states of this work, nosignificant correlation was noticed.A strong positive correlation between the optical and the 𝛾 -rayfluxes was noticed by Aleksić et al. (2014) between October 2009to February 2011. Fukazawa et al. (2018) reported the first positivecorrelation between fluxes in the X-ray and the 𝛾 -ray energy bandsduring 2013 − 𝛾 -ray, X-ray and UV/optical fluxes averaged over S1, S2 and S3states without any appreciable change in the spectral indices. MNRAS000 , 1–12 (2020) ultiwavelength monitoring of NGC 1275 State 1 I nde x ( α ) State 1 H a r dne ss R a t i o Flux (10 -7 ph/cm /sec) State 2 I nde x ( α ) State 2 H a r dne ss R a t i o Flux (10 -7 ph/cm /sec) State 3 I nde x ( α ) State 3 H a r dne ss R a t i o Flux (10 -7 ph/cm /sec) State 4 I nde x ( α ) State 4 H a r dne ss R a t i o Flux (10 -7 ph/cm /sec) Figure 4.
Variation of monthly averaged 𝛾 -ray flux with photon index (top panel in each sub-figure) and monthly averaged 𝛾 -ray flux with hardness ratio(bottom panel in each sub-figure) for the four states identified in the 𝛾 -ray light curve. -12-11-10 9 11 13 15 17 19 21 23 25 27 Log ν F ν ( e r g s / c m / s ) Log ν (Hz) Fermi-S1Swift-UVOT-S1Fermi-S2Swift-UVOT-S2Fermi-S3Swift-UVOT-S3One-zone SSC-S1One-zone SSC-S2One-zone SSC-S3
Figure 5.
SED of NGC 1275 during different activity states. The dotted, solidand dashed lines represents the total (Sync+SSC) SED model contributionfrom S1, S2, and S3 respectively. The bow-ties in X-ray band for S1, S2, andS3 are from
Swift -XRT. The downward triangles are for upper limit values.
Usually one zone synchrotron self Compton (SSC) jet modelhas been used to explain the broadband SEDs of NGC 1275(e.g., Abdo et al. 2009; Kataoka et al. 2010; Aleksić et al. 2014;Tanada et al. 2018). However, Tavecchio & Ghisellini (2014) con-sidered a “spine-layer” scenario to explain the broadband SED ofthis source during the MAGIC campaign (Aleksić et al. 2014).From the broadband SED modelling of the 𝛾 -ray flaring andquiescent states of NGC 1275, Tanada et al. (2018) suggested thatflux changes in epoch A were caused by the injection of the high-energy electrons in the jet, while a change of Doppler factor couldexplain the observed flux variations in epoch B. Aleksić et al. (2014)derived the parameters, which are in the typical range found for BLLacs (except for bulk Lorentz factor), from the broadband SEDmodelling of simultaneous observations in two campaigns (Octo-ber 2009-February 2010 and August 2010-February 2011) of NGC1275. They suggested that NGC 1275 could be a misaligned BLLac with large jet inclination angle and small bulk Lorentz factor.Alternatively, it might be more aligned with smaller jet inclinationangle and higher bulk Lorentz factor.In the present work, we have studied the average behaviourof the source at its various activity states. Broadband SEDs of thesource were constructed and modelled during S1, S2 and S3 states.Due to the lack of adequate multi-band data, averaged broadbandSED study could not be carried out during the S4 state. We alsoconstructed and modelled the broadband SEDs during AS1 and AS2 MNRAS , 1–12 (2020) S. Gulati et al. -12-11-109 11 13 15 17 19 21 23 25 27
AS1
Log ν F ν ( e r g s / c m / s ) Log ν (Hz) FermiSwift-UVOTSwift-XRT -12-11-109 11 13 15 17 19 21 23 25 27
AS2
Log ν F ν ( e r g s / c m / s ) Log ν (Hz) FermiAstroSat-UVITAstroSat-SXTAstroSat-LAXPC
Figure 6.
SED of NGC 1275 during
AstroSat observing periods. The dashed line represents the contribution from synchrotron emission and dash-dotted linerepresents the contribution from SSC emission. The solid line represents the total (Sync+SSC) SED model contribution. The downward triangles are for upperlimit values.
Table 7.
Model parameters for the SEDParameter Symbol State 1 State 2 State 3
AstroSat -AS1
AstroSat -AS2Minimum electron Lorentz factor 𝛾 . . . Maximum electron Lorentz factor 𝛾 . × . × . × . × . × Break Lorentz factor 𝛾 𝑏 . × . × . × . × . × Normalisation of particle spectrum 𝑁 . × . × . × . × . × Particle spectral index (before break) 𝑝 . . . . . Particle spectral index (after break) 𝑝 . . . . . states. During AS1, data from UVW2 filter of
Swift -UVOT,
Swift -XRT and
Fermi -LAT were used to model the broadband SED. Dur-ing AS2, data from four filters of
AstroSat -UVIT,
AstroSat -SXT,
AstroSat -LAXPC and
Fermi -LAT were used to model the broad-band SED. We considered a single zone leptonic model that includessynchrotron and SSC emission by non-thermal relativistic electronsin relativistically moving emission region (blob) in AGN jet, asdescribed in Bhattacharyya et al. (2018). The energy distributionof relativistic jet electrons was considered as a broken power-law,which is given by 𝑁 ( 𝛾 ) = 𝑁 (cid:16) 𝛾𝛾 𝑏 (cid:17) − 𝑝 for 𝛾 ≤ 𝛾 ≤ 𝛾 𝑏 = 𝑁 (cid:16) 𝛾𝛾 𝑏 (cid:17) − 𝑝 for 𝛾 𝑏 ≤ 𝛾 ≤ 𝛾 where, 𝛾 , 𝛾 , and 𝛾 𝑏 are minimum, maximum and breakLorentz factors whereas 𝑝 and 𝑝 are the particle spectral indicesbefore and after the break Lorentz factor.For modelling of SEDs in all the activity states, the employedjet inclination angle ( 𝜃 ) was 20 ◦ , consistent with previous studies onbroadband SED of this source ( e.g., Abdo et al. 2009; Aleksić et al.2014; Tanada et al. 2018; Fukazawa et al. 2018). However, thisvalue of 𝜃 is much smaller than that inferred from radio observations( 𝜃 = ◦ − ◦ ; Walker et al. 1994 and 𝜃 = ◦ − ◦ ; Fujita & Nagai2017). Other derived physical parameters for all the activity statesare the blob radius of 𝑅 = . × cm, the magnetic field of 𝐵 = .
08 G and the bulk Lorentz factor
Γ = .
0. These values arein the range considered in previous studies (e.g., Abdo et al. 2009;Aleksić et al. 2014; Tanada et al. 2018; Fukazawa et al. 2018).While modelling the S1, S2 and S3 states, the maximum value of the particle Lorentz factor and the break Lorentz factors weregrossly estimated from the observed photon spectrum so that themaximum observed photon frequency and the spectral turn overin the IC hump could be generated. The minimum Lorentz factorwas adjusted so that the SSC process could explain the X-ray flux.Nevertheless, with the given data none of these parameters couldbe constrained well for S1, S2 and S3 states.While modelling of the SEDs during AS1 and AS2, the primaryapproach followed to estimate the spectral parameters remained thesame. However, for both the states, the soft and hard X-ray spectraappeared in the falling edge of the synchrotron peak. Therefore,the highest frequency of the synchrotron peak and the approxi-mate turnover frequency of the SSC peak were used to estimate themaximum and the break Lorentz factors of the particle spectrum.Nevertheless, they could not be constrained well. Since the spec-tra moved towards higher energies, the minimum Lorentz factor ofthe electron distributions was pushed to higher values as comparedto the states S1 - S3. The values of the minimum Lorentz factorof electrons comparable to what was obtained for states S1 - S3would result in more radiation power in the lower frequency region.Therefore, the orders of magnitude of the Lorentz factors that weobtained from the modelling of the SEDs are relatively consistent,even though they are not fully constrained.The fitted SEDs during S1, S2, and S3 states are shown inFig 5. SEDs during
AstroSat observing periods are shown in Fig 6.Due to large error in X-ray flux and photon index values during S1and AS1 states, we used the derived flux in four bands 1 . − . . − .
0, 5 . − . . − . XSPEC to construct X-ray SED as shown in Fig 5 and Fig 6. Theresults of our SED fitting are given in Table 7.
MNRAS , 1–12 (2020) ultiwavelength monitoring of NGC 1275 Our findings from the SED modelling of S1, S2, S3, AS1 andAS2 states are given below. • The optical spectral indices during S2 and S3 are steeper thanthat of X-ray, and 𝛾 -rays. Also, the observed flux in the optical bandis significantly higher than the predicted flux from the synchrotronemission of the jet electrons. The observed nature of optical flux andspectra suggests that the optical emission might have originated ex-ternal to the jet, probably from the accretion disk/BLR, host galaxyas well as from the cluster. • The 𝛾 -ray emission was well explained by the SSC emission.The UV emission is marginally higher than that predicted by thesynchrotron emission process. • During S1, S2, and S3 states, the flux in the X-ray band wasexplained by the SSC emission. However, during AS1 and AS2,X-ray emission was explained by the synchrotron process. • An increase in jet particle normalisation with no significantvariation in other parameters was also noticed during S1, S2, andS3 states. While studying long-term X-ray behaviour of NGC 1275during 2006-2015, Fukazawa et al. (2018) speculated that increasein the electron density could be one of the possible explanations ofthe observed long-term X-ray flux increase. • Unlike S1, S2, and S3 states where energy distribution of jetelectrons was very close to a single power-law, a broken powerenergy distribution of jet electrons, with a much flatter spectrum,was noticed during both AS1 and AS2 observing periods. Also, asignificant increase in the minimum and maximum electron Lorentzfactors was noticed during
AstroSat observing periods. The SEDsof the source during AS1 and AS2 show significant changes ascompared to the SEDs during S1, S2 and S3 states. First, the overallluminosity of the source increased. Second, the X-ray spectra, asobtained from XRT, SXT and LAXPC, changed the slope duringboth AS1 and AS2 as compared to the S1, S2 and S3 states. Finallythe peak of the SED in synchrotron as well as Compton hump shiftedto the higher energies in AS1 and AS2. Therefore the slope of theX-ray spectra during
AstroSat observing periods indicated that itwas part of the synchrotron hump. To model such features in theSEDs of AS1 and AS2 states, it was necessary to increase the valuesof 𝛾 and 𝛾 . The values of 𝛾 were constrained by both the X-rayand the 𝛾 -ray spectra. In this work, based on the variation in the 𝛾 -ray baseline flux,we identified four activity states of this source. We also observedthe source twice with AstroSat during its high 𝛾 -ray activity state.We found three distinct states (S1, S2, S3) with increase in the 𝛾 -ray baseline flux for ∼ 𝛾 -ray flare. We determined theboundaries of these activity states by fitting linear functions tothe cumulative flux distribution. A correlation study between the 𝛾 -ray flux and the spectral nature of the source was carried out.Also, broadband SEDs were constructed and modelled utilisingobservations from Fermi -LAT,
Swift and
AstroSat . We conclude: • An increase of the 𝛾 -ray baseline flux with no appreciablechange in averaged spectral properties was noticed during S1 toS3 state. Similar behaviour was also noticed in the UV and X-raybands. An increase in the jet particle normalisation could explainthis observed feature. • No significant correlation was noticed between the 𝛾 -ray fluxand the spectral parameter/hardness ratio during S1, S2, and S3 states. However, a hint of correlation was noticed during the S4state. • Based on the first two years of
Fermi observationsBrown & Adams (2011) reported that “NGC 1275 appeared to mi-grate from the FR I radio galaxy to the BL Lac object region” duringlarge 𝛾 -ray flare. While explaining the steeper X-ray spectrum dur-ing 2010 flare of this source, Fukazawa et al. (2018) proposed thatthe X-ray emission might have synchrotron origin unlike SSC in nor-mal state. In this work, we found that during S1, S2, and S3 states,which represent the long term averaged behaviour of this source,X-ray emission was well explained by the SSC process. However,during AS1 and AS2 observing periods, there is an evidence of anincrease in the synchrotron peak frequency and the X-ray emissionis explained by the synchrotron emission of jet electrons.Long term study of NGC 1275 in this work provides a better un-derstanding of the underlying emission mechanism during variousactivity states. ACKNOWLEDGEMENTS
We thank the anonymous referee for his/her constructive commentsthat helped us to improve the manuscript considerably. The author(s)thank Ranjeev Misra for discussions regarding
AstroSat -LAXPCand overall X-ray analysis, and Gulab Dewangan for discussion re-garding
AstroSat -SXT analysis. The author(s) thank Jayashree Roy,Bitan Ghosal, Anil Tolamatti, and Ashish Devaraj (UVIT-POC) fortheir useful discussions. The author(s) acknowledge the financialsupport of Indian Space Research Organisation (ISRO) under
As-troSat archival Data utilization program. This publication uses thedata from the
AstroSat mission of the ISRO, archived at the ISSDC.This work has been performed utilising the calibration data-basesand auxiliary analysis tools developed, maintained and distributedby
AstroSat -SXT team with members from various institutions inIndia and abroad. This work has made use of public
Fermi -LATdata obtained from the
Fermi
Science Support Center (FSSC), pro-vided by NASA Goddard Space Flight Center. This research hasmade use of the NASA/IPAC Extragalactic Database (NED), whichis operated by the Jet Propulsion Laboratory, California Institute ofTechnology, under contract with the National Aeronautics and SpaceAdministration. This research has made use of data and/or softwareprovided by the High Energy Astrophysics Science Archive Re-search Center (HEASARC), which is a service of the AstrophysicsScience Division at NASA/GSFC and the High Energy AstrophysicsDivision of the Smithsonian Astrophysical Observatory. This re-search has made use of the XRT Data Analysis Software ( xrtdas )developed under the responsibility of the ASI Science Data Cen-ter (ASDC), Italy. Manipal Centre for Natural Sciences, Centre ofExcellence, Manipal Academy of Higher Education (MAHE) isacknowledged for facilities and support.
DATA AVAILABILITY
This work has made use of public
Fermi -LAT data avail-able at https://fermi.gsfc.nasa.gov/cgi-bin/ssc/LAT/LATDataQuery.cgi . This research has made use of data andsoftware provided by the High Energy Astrophysics ScienceArchive Research Center (HEASARC) available at https://heasarc.gsfc.nasa.gov/docs/software/lheasoft/ , Swift data available at https://heasarc.gsfc.nasa.gov/cgi-bin/W3Browse/w3browse.pl and the NASA/IPAC Extragalactic
MNRAS , 1–12 (2020) S. Gulati et al.
Database (NED). This publication has also made use ofthe data from the
AstroSat mission of the ISRO, archivedat the ISSDC ( https://astrobrowse.issdc.gov.in/astro_archive/archive/Home.jsp ). AstroSat data will be shared onrequest to the corresponding author with the permission of ISRO.
REFERENCES
Abdo A. A., et al., 2009, ApJ, 699, 31Abdollahi S., et al., 2020, ApJS, 247, 33Acero F., et al., 2015, ApJS, 218, 23Agrawal P. C., 2006, Advances in Space Research, 38, 2989Aleksić J., et al., 2012, A&A, 539, L2Aleksić J., et al., 2014, A&A, 564, A5Antia H. M., et al., 2017, ApJS, 231, 10Asada K., Kameno S., Shen Z.-Q., Horiuchi S., Gabuzda D. C., Inoue M.,2006, PASJ, 58, 261Baghmanyan V., Gasparyan S., Sahakyan N., 2017, ApJ, 848, 111Ballet J., Burnett T. H., Digel S. W., Lott B., 2020, arXiv e-prints,p. arXiv:2005.11208Benbow W., VERITAS Collaboration 2015, in 34th International CosmicRay Conference (ICRC2015). p. 821 ( arXiv:1508.07251 )Bhattacharya D., Misra R., Rao A. R., Sreekumar P., 2013, MNRAS,431, 1618Bhattacharya D., Mohana A K., Gulati S., Bhattacharyya S., Bhatt N.,Sreekumar P., Stalin C. S., 2017, MNRAS, 471, 5008Bhattacharya D., Gulati S., Stalin C. S., 2019, MNRAS, 483, 3382Bhattacharyya S., et al., 2018, MNRAS, 481, 4505Brown A. M., Adams J., 2011, MNRAS, 413, 2785Bruel P., Burnett T. H., Digel S. W., Johannesson G., Omodei N., Wood M.,2018, arXiv e-prints, p. arXiv:1810.11394Burrows D. N., et al., 2005, Space Sci. Rev., 120, 165Buttiglione S., Capetti A., Celotti A., Axon D. J., Chiaberge M., MacchettoF. D., Sparks W. B., 2010, A&A, 509, A6Cardelli J. A., Clayton G. C., Mathis J. S., 1989, ApJ, 345, 245Chitnis V., Shukla A., Singh K. P., Roy J., Bhattacharyya S., Chandra S.,Stewart G., 2020, Galaxies, 8, 63Churazov E., Forman W., Jones C., Böhringer H., 2003, ApJ, 590, 225Ciprini S., 2013, The Astronomer’s Telegram, 4753Donato D., Wood D., Cheung C. C., 2010, The Astronomer’s Telegram,2737Dutson K. L., Edge A. C., Hinton J. A., Hogan M. T., Gurwell M. A., AlstonW. N., 2014, MNRAS, 442, 2048Falco E. E., et al., 1999, PASP, 111, 438Fujita Y., Nagai H., 2017, MNRAS, 465, L94Fukazawa Y., et al., 2018, ApJ, 855, 93Ghosal B., et al., 2020, New A, 80, 101402Godet O., et al., 2009, A&A, 494, 775Humason M. L., 1932, PASP, 44, 267Imazato F., Fukazawa Y., Sasada M., Sakamoto T., 2021, ApJ, 906, 30Kataoka J., et al., 2010, ApJ, 715, 554Khachikian E. Y., Weedman D. W., 1974, ApJ, 192, 581Kumar A., et al., 2012, in Proc. SPIE. p. 84431N ( arXiv:1208.4670 ),doi:10.1117/12.924507Kushwaha P., Sinha A., Misra R., Singh K. P., de Gouveia Dal Pino E. M.,2017, ApJ, 849, 138MAGIC Collaboration et al., 2018, A&A, 617, A91Mattox J. R., et al., 1996, ApJ, 461, 396Meyer M., Scargle J. D., Blandford R. D., 2019, ApJ, 877, 39Mirzoyan R., 2016, The Astronomer’s Telegram, 9689, 1Mirzoyan R., 2017, The Astronomer’s Telegram, 9929, 1Mukherjee R., VERITAS Collaboration 2016, The Astronomer’s Telegram,9690, 1Mukherjee R., VERITAS Collaboration 2017, The Astronomer’s Telegram,9931, 1 Nishino S., Fukazawa Y., Hayashi K., Nakazawa K., Tanaka T., 2010, PASJ,62, 9Pivato G., Buson S., 2015, The Astronomer’s Telegram, 8219Rani B., Madejski G. M., Mushotzky R. F., Reynolds C., Hodgson J. A.,2018, ApJ, 866, L13Rao A. R., Singh K. P., Bhattacharya D., 2016, Space Research Today,196, 30Rao A. R., Bhattacharya D., Bhalerao V. B., Vadawale S. V., Sreekumar S.,2017, Current Science, 113, 595Schlafly E. F., Finkbeiner D. P., 2011, ApJ, 737, 103Schmidt R. W., Fabian A. C., Sanders J. S., 2002, MNRAS, 337, 71Singh K. P., et al., 2014, in Proc. SPIE. p. 91441S, doi:10.1117/12.2062667Singh K. P., et al., 2017, Journal of Astrophysics and Astronomy, 38, 29Tanada K., Kataoka J., Arimoto M., Akita M., Cheung C. C., Digel S. W.,Fukazawa Y., 2018, ApJ, 860, 74Tandon S. N., et al., 2017, AJ, 154, 128Tandon S. N., et al., 2020, AJ, 159, 158Tavecchio F., Ghisellini G., 2014, MNRAS, 443, 1224Vadawale S. V., Chattopadhyay T., Rao A. R., Bhattacharya D., BhaleraoV. B., Vagshette N., Pawar P., Sreekumar S., 2015, A&A, 578, A73Vermeulen R. C., Readhead A. C. S., Backer D. C., 1994, ApJ, 430, L41Walker R. C., Romney J. D., Benson J. M., 1994, ApJ, 430, L45Wood M., Caputo R., Charles E., Di Mauro M., Magill J., Perkins J. S.,Fermi-LAT Collaboration 2017, International Cosmic Ray Conference,301, 824Yadav J. S., et al., 2016a, ApJ, 833, 27Yadav J. S., et al., 2016b, in Proc. SPIE. p. 99051D, doi:10.1117/12.2231857Yamazaki S., et al., 2013, PASJ, 65, 30This paper has been typeset from a TEX/L A TEX file prepared by the author.MNRAS000