AstroSat view of the NLS1 galaxy Mrk 335
aa r X i v : . [ a s t r o - ph . H E ] F e b J. Astrophys. Astr. (0000) :
AstroSat view of the NLS1 galaxy Mrk 335
Savithri H. Ezhikode , Gulab C. Dewangan and Ranjeev Misra Inter-University Centre for Astronomy and Astrophysics, Post Bag 4, Ganeshkhind, Pune 411007, India * Corresponding author. E-mail: [email protected] received : 07 Nov 2020; accepted : 24 Dec 2020
Abstract.
We present the results from the multi-wavelength monitoring observations of the Narrow-Line Seyfert 1 galaxyMrk 335 with
AstroSat . We analysed both the X-ray (SXT & LAXPC) and UV (UVIT) data of the source at twoepochs, separated by ∼
18 days. The source was in a low flux state during the observations, and the X-ray spectrawere found to be harder than usual. The presence of soft X-ray excess was identified in the observations, and thebroadband X-ray continuum was modelled with power-law and blackbody (modified by intrinsic absorption) anda distant neutral reflection component. We did not find any variability in the X-ray spectral shape or the flux overthis period. However, the UV flux is found to be variable between the observations. The obtained results fromthe X-ray analysis point to a scenario where the primary emission is suppressed and the component due to distantreflection dominates the observed spectrum.
Keywords.
AGN—NLS1—Mrk 335—X-ray—UV.
1. Introduction
Narrow-line Seyfert 1 galaxies (NLS1s) are a partic-ular class of active galactic nuclei (AGN) with someextreme properties. They show strong Fe-II emissionin the optical band along with narrow permitted linesand weak [OIII] emission (Osterbrock & Pogge, 1985;Goodrich 1989). In the X-ray band, they are generallycharacterised by enhanced spectral and flux variabilityand the presence of soft excess emission (e.g. Boller etal. 1996).Mrk 335 (RA = =+ = Uhuru , Einstein , EXOSAT , Ginga , ROSAT , ASCA , and
BeppoSAX whenit was an X-ray bright source in the sky (eg. Tananbaumet al. 1978; Halpern 1982). Its intensity dropped fromthe brightest stage to the very low flux state in 2007(Grupe et al. 2007). Later, the source remained mostlyin a low flux state though it has been reported to show-ing episodes of X-ray flaring activities (e.g. Gallo et al. 2018).Mrk 335 was extensively studied in the optical / UVand X-rays (Grupe et al. 2007, 2008, 2012; Longinottiet al. 2013; Gallo et al. 2013; Parker et al. 2014; Ko-mossa wt al. 2014; Chainakun & Young 2015; Keek& Ballantyne 2016, Sarma et al. 2015, Gallo et al.2015, Wilkins et al. 2015). The X-ray spectra obtainedwith
Ginga , ASCA , BeppoSAX , XMM-Newton and
NuS-TAR showed evidences for reflection and warm absorp-tion features in the source (Nandra & Pounds 1994;George et al. 2000; Leighly 1999; Ballantyne et al.2001; Parker et al. 2014; Longinotti et al. 2013, 2019;Ezhikode et al. 2020). Various attempts to model theX-ray spectra of Mrk 335 in the past suggested the pos-sibility of changes in the geometry of the corona lead-ing to the state changes in the source (e.g. Gallo et al.,2013; Wilkins et al., 2015; Gallo et al., 2015; Gallo L.,2018). Signatures of soft excess has been persistentlyseen in the X-ray spectra of the source (e.g., Bianchiet al. 2001; Grupe et al. 2001; Grupe et al. 2008;Chainakun & Young, 2015; Gallo et al. 2015).The source is also known to show considerablevariability in the optical / UV band which was found tobe correlated and uncorrelated with the X-ray variabil-ity at various epochs (Buisson et al. 2017; Gallo et al.2018). The long-term monitoring of Mrk 335 with
Swift in optical–UV–X-ray bands and the observations withother telescopes revealed the properties of the source at © Indian Academy of Sciences 1
J. Astrophys. Astr. (0000) : di ff erent phases of its variability (Buisson et al. 2017;Tripathi et al. 2020). AstroSat (Singh et al. 2014, Agrawal 2017) moni-tored Mrk 335 at two epochs in 2017. We present theresults from these multi-wavelength observations in X-ray and UV bands. We studied the X-ray spectral fea-tures by modelling the soft and hard X-ray spectra withdi ff erent models. The details of observations and dataprocessing are given in § § §
4. and § §
2. Observations
Mrk 335 was observed simultaneously in the X-rayand UV bands with Soft X-ray Telescope (SXT: Singhet al., 2017), Large Area X-ray Proportional Counter(LAXPC: Yadav et al. 2016; Antia et al. 2017),Cadmium Zinc Telluride Imager (CZTI: Rao et al.2017; Bhalerao et al. 2017; Vadawale et al. 2016),and Ultra Violet Imaging Telescope (UVIT: Tandon etal. 2017(a,b)) onboard
AstroSat on October 31, 2017(Obs 1) and November 18, 2017 (Obs 2). Here, we usethe data from SXT, LAXPC, and UVIT observations.The details of these observations are given in Table 1.and Table 2, in the subsequent sections. The data usedfor the study are available at the Astrobrowse archivehandled by Indian Space Science Data Centre (ISSDC).
3. X-ray Analysis
Data reduction
The Level-2 data products for SXT and LAXPC obser-vations were obtained from the Level-1 data using theprocessing pipelines. The SXT observations were per-formed in the photon counting (PC) mode. We used sxtpipeline b (Release Date: 2019-01-04) for re-ducing the Level-1 SXT data. The pipeline producedcleaned event list for each orbit. These event lists werethen merged using the sxtevtmerger tool in Julia. Themerged event list in each observation was used to cre-ate high-level science products using xselect . We usedthe software L axpc S oft for processing LAXPC data.From the Level-2 event file and the GTI file created, wegenerated the light curves and spectra with the varioustasks in the tool. Since the exposure time for the ob-servations are less than that necessary for obtaining agood signal to noise data, and the source was in a lowflux state, the data quality is found to be poor. 3.2 Light curves
We created the light curves in both soft and hard X-ray bands. SXT light curves for the two observationswere generated in the 0.7 − ff erent time bins using xs - elect . We also generated LAXPC light curves in the en-ergy range of 4–20 keV for various bin sizes. Since thesource is very faint in the hard X-ray band, we used thespecific laxpcsoft code for faint source background forlight curve generation. Fig. 1 shows the 0.7–7 keV SXTand 4–20 keV LAXPC 20 light curves for the two ob-servations, created for a time bin of 500 s. We checkedthe variability of the light curves using the ftool lcstats and found no variability in SXT. Though the LAXPClight curves showed significant variability in terms ofthe fraction RMS amplitude, a similar variability pat-tern was observed in the background lightcurve. Hence,the variability seen in the net LAXPC light curves inFig.1 is not intrinsic to the source.3.3 Spectral Analysis
The SXT source spectra were extracted from circularregions of 16 arcmin radius, whereas the blank skyspectrum was used for the background. We used thebackground spectrum and the response files providedby the SXT-POC team. The Ancillary Response Func-tion (ARF) file corrected for vignetting, PSF and expo-sure was generated using the latest module released on2019 July 18, sxteefmodule v
02. The rmf file for grade0–12 was used for the analysis. The spectra were alsogrouped so that we can apply χ statistic. As the SXTresponse is not well characterised below 0.7 keV, theregion was ignored in the analysis.The LAXPC background spectra were created withthe faint source code mentioned above. Since the back-ground is more stable for LAXPC 20, we used onlyLAXPC 20 spectra for the analysis. Here, we ignoredbelow 4 keV and above 20 keV as the regions weredominated by background. The SXT and LAXPC spec-tra at the two epochs are shown in Fig. 2.Both the SXT and LAXPC 20 spectra were anal-ysed simultaneously to characterise the broadband con-tinuum of the source. The spectral analysis was doneusing xspec version 12.9. To account for the shift inSXT response, gain command in xspec was used with o ff set parameter fixed at 0.02. The model constant wasused to take care of the cross normalisation betweenSXT and LAXPC 20. Also, a systematic error of 3% isapplied while fitting.The X-ray spectral analysis was started by jointlyfitting the SXT and LAXPC 20 spectra in the hardX-ray band (2 −
20 keV) with an absorbed power-law ( tbabs × powerlaw ) model. The Galactic col- . Astrophys. Astr. (0000) : Table 1 . The details of
AstroSat observations of Mrk 335. The quoted count rates are background subtracted values in the0.3–8 keV band for SXT and in the 4–20 keV band for LAXPC 20 (LXP 20).Observation Exposure (ks) Count Rate (counts / s)Number ID Date SXT LXP 20 SXT (10 − ) LXP 20Obs 1 9000001654 31 / / ± ± / / ± ± . . . . C oun t R a t e ( c oun t s / s ) Obs 1 SXT(0.7−7 keV)LXP 20 (4−20 keV)0 10 C oun t R a t e ( c oun t s / s ) Time (s)
Bin time: 500.0 s . . . C oun t R a t e ( c oun t s / s ) Obs 2SXT (0.7−7 keV)LXP 20 (4−20 keV)0 10 C oun t R a t e ( c oun t s / s ) Time (s)
Bin time: 500.0 s
Figure 1 . X-ray light curves for the two observations (Obs 1 & Obs 2) binned for 500 s. SXT (black) and LAXPC 20(LXP 20: red) light curves are extracted from 0.7–7 keV and 4–20 keV bands, respectively. − − . . r m a li z ed c oun t s s − k e V − Energy (keV)Obs 1SXT LXP 20 1 102 5 20 − − . . r m a li z ed c oun t s s − k e V − Energy (keV)Obs 2SXT LXP 20
Figure 2 . SXT (0.7–7 keV: black) and LAXPC 20 (4–20 keV: red) spectra of Mrk 335 at the two epochs, Obs 1 and Obs 2.
J. Astrophys. Astr. (0000) :
Table 2 . The details of UVIT observations at the two epochs. The last two columns show the measured count rate(background subtracted) from a circular regions of 30 subpixel radius.
Band Filter Wavelength Width Exposure (s) Count Rate (counts / s)(Å) (Å) Obs 1 Obs 2 Obs 1 Obs 2NUV N242W 2418.0 785.0 1881.957 1949.968 27.30 ± ± ± ± ± ± ± ± ± ± ± ± ± ± umn density ( N H ) for the tbabs component was fixedat 3.56 × cm − obtained from the LAB survey(Kalberla et al. 2005). The fit yielded a photon in-dex ( Γ ) of less than 1 for both the observations. Tocheck the presence of intrinsic absorption, we added a ztbabs component. However, the fit did not improve,and the intrinsic equivalent Hydrogen column density( N IntH ) was not constrained. Since such flat hard X-rayspectra could be the result of intrinsic absorption andthe presence of distant reflection, we also included one xillver (Garcia et al. 2010, 2013) component. Only thenormalisation ( N xl ) and reflection fraction ( f refl ) param-eters of xillver were allowed to vary during the fit. Thephoton index of xillver component was tied to the slopeof powerlaw model. Inclination ( i ), high-energy cut-o ff ( E cut ) and the iron abundance ( A Fe ) were fixed at 30 ◦ ,300 keV and 1 (in solar abundance), respectively. Theionisation parameter ( ξ ) was set to the minimum value,log ξ =
0, to account for the reflection from neutral ma-terial. The new model yielded a marginally better fitfor both the observations with ∆ χ ∼ − . ∆ χ ∼ − . xillver component.Further, we noticed the energy range below 2 keVand found that the spectrum rises above the currentmodel. This is a clear indication of the soft excessemission. Therefore, we added bbody to model the softX-ray excess and the fit provided a blackbody temper-ature ( kT bb ) of ∼ . xillver normalisation and re-flection fraction. However, the model tbabs ( ztbabs × ( powerlaw + bbody ) + xillver ) was preferable than theone without either xillver or bbody component.Another possible reason behind the observed hardspectrum could be the presence of partial covering ab-sorption. Therefore, we also tried fitting the spectra with zpcfabs and zxipcf models. However, the fits pro-vided poor statistic and poorly constrained parameters.
4. UVIT Analysis
The source was observed with UVIT at both NUV andFUV wavelengths. Four filters (in PC mode) were usedfor both NUV and FUV observations at the first epoch.In the second observation, only two FUV filters wereused as the instrument stopped working during thattime. The filter information and other details of theexposures are given in Tabel 2. The Level-2 data, al-ready processed with the latest pipeline UVIT L evel -2P ipeline (UL2P) version 6.3 by the POC were availableat the archive. We used these data sets for further anal-ysis.We carried out the photometry on the combined im-age in each filter. The combined images are obtainedfrom the Level-2 data created using aspect correctiondone with VIS or NUV data. To do the photometry, wechose a circular region of radius 30 subpixels ( ∼ ff er from saturation. Hence, we followed the proce-dure described in Tandon et al. (2017, 2020) to cor-rect for the e ff ect. We note that the NUV N242W fil-ter records a total count rate of ∼
30 counts / s wherethe above-mentioned saturation correction is not valid.Hence, we do not use the data from N242W filter forfurther analysis. The background regions were selectedfrom di ff erent circles with radii of 60 subpixels. Theaverage background count rates were then subtractedfrom the saturation corrected values. We obtained theAB magnitude ( m AB ) from the count rates and the zero- . Astrophys. Astr. (0000) : −6 −5 −4 −3 k e V ( P ho t on s c m − s − k e V − ) Obs 1 powerlaw bbody xillver r a t i o Energy (keV) 10 −6 −5 −4 −3 k e V ( P ho t on s c m − s − k e V − ) Obs 21 102 5012 r a t i o Energy (keV)
Figure 3 . The Spectral fitting plots for the model tbabs ( ztbabs × ( powerlaw + bbody ) + xillver ) in the 0.7–20 keV for thetwo observations. The unfolded spectra and the model are shown in the upper panels and the ratio of data to model areshown in the lower panels. The solid lines in the upper panels represent the overall model whereas the dotted lines deonte theindividual model components powerlaw , bbody and xillver . The SXT data and model are shown in black colour while thoseof LAXPC are shown in red colour. point (ZP) magnitudes and calculated the correspond-ing flux density F λ (Tandon et al., 2017, 2020) in eachfilter. We have also applied the Galactic extinction cor-rection for the estimated F λ values using Cardelli et al.(1989) relation for R V = V =
5. Spectral Energy Distribution
A comprehensive modelling of the UV to X-ray spec-tral energy distribution (SED) of AGN can unveil thegeometry of the central emitting regions and physicsrelated to the variability mechanisms. We have UV ob-servations of Mrk 335 with various filters in NUV andFUV bands (see Table 2). But modelling the accretiondisc emission from these photometric data is compli-cated as many components, like host galaxy and emis-sion lines, can contribute to the observed flux at thesewavelengths. Moreover, most of these observations area ff ected by saturation. Though we corrected the sourcecount rates for the saturation e ff ects (as mentioned inthe previous section), there could still be uncertaintiesassociated with the flux estimation. Also, it is di ffi -cult to derive the source flux free from the host galaxycontamination. To account for these uncertainties, weadded a systematic of 5–10% and fitted the UV–X-raySED. The UV spectra were created by converting theflux (corrected for the Galactic reddening) in each filter(given in Table. 5) using the task flx xsp .We used the xspec model optxagnf (Done et al.,2012) to fit the broadband SED of the source. Themodel can describe the emissions form accretion disc together with the soft and hard X-ray components.Here, we show the example of SED fitting for Obs 1since there are more FUV data points for this obser-vation. We began with the analysis of X-ray spectraby replacing powerlaw + bbody components with optx-agnf in the best-fit model tbabs ( ztbabs × ( powerlaw + bbody ) + xillver ). Further, we added the FUV spec-tra for the filters F148Wa, F154W, F169M & F172M,and included the model zreddedn to correct for intrin-sic reddening. The parameter E(B-V) for zredden wasobtained from the intrinsic column density N IntH usingthe relation given by Bessell (1991). The parametersof xillver for the X-ray part were fixed at the best-fitvalues, whereas the component was not used for UVspectra. The cross normalisation constant for both theNUV and FUV spectral groups were tied to that of theSXT spectrum. We notice that fitting the FUV–X-raySED resulted in a reasonable χ of 91.78 for 71 de-grees of freedom when a systematic error of 5% wasapplied. The fit provided a hard X-ray photon index of ∼ N H of 2 . × cm − (fixed). Theother parameters obtained from the fit are Eddingtonratio ∼
1, coronal radius R cor ∼ R g and the fractionof power below R cor emitted as the hard X-ray com-ponent f pl ∼ .
9. The temperature and optical depthof the soft X-ray component are ∼ ∼ R g , below which the energy is dis-sipated as Comptonised emission. However, this is a J. Astrophys. Astr. (0000) : preliminary analysis and the errors on parameters arenot obtained. A detailed and systematic study of themulti-wavelength SED of Mrk 335 with
AstroSat datawill be done later.
6. Results
As expalined in Sec. 3.3, we tried fitting the X-ray spec-tra with di ff erent models. The best-fit parameters forthese models are given Table 3. Fig. 3 shows the var-ious spectral fitting plots for Obs 1 and Obs 2. Wefound that tbabs(ztbabs × (powerlaw + bbody) + xillver) better fits the data in the 0.7–20 keV band than the othermodels. We estimated the unabsorbed flux in di ff erentenergy bands using the cflux convolution model. Theobtained flux values are provided in Table 4.Both observations are found to be intrinsicallyabsorbed with N IntH ∼ × cm − . The sourcedoes not show significant X-ray spectral or flux vari-ability between the observations. The X-ray spectraseem to be harder in both the observations with Γ ∼ . .
2) (consistent within errorbars) for Obs 1(Obs 2)while the reflection parameters are not properly con-strained. The total flux in the 2–20 keV is roughly2 × − erg cm − s − , and in the 0.7–2 keV band it isaround 6 × − erg cm − s − .Unlike the X-ray observations, UV emission fromthe source shows variability in both FUV and NUVbands. The net count rate and flux in each filter arementioned in Table 5. The flux variability (except forthe FUV filters F169M, F172M and the NUV filterN245M) are shown in Fig. 5. In order to check if thevariability is an instrument artefact, we obtained thelight curves of a star in both NUV and FUV images(since the stars were too faint in NUV N279, F154Wand F148Wa filters we did not obtain the count rate forthose exposures). The net count rates for the star seemto be non-variable in the NUV band showing that thevariability shown by the source is real.6.1 Comparison with other observations
Mrk 335 has been observed at optical, UV and X-raywavelengths with various missions. Here, we give abrief summary of the analysis of some of these dataand compare those with the results from our
AstroSat observations.
Swift has been monitoring Mrk 335 foryears in X-rays and optical / UV. We analysed one
Swift observation close to
AstroSat observations as there areno observations strictly simultaneous with that of
As-troSat . We retrieved XRT and UVOT data taken onNovember 03, 2017 (almost three days after Obs 1 andtwo weeks before Obs 2). This near-simultaneous
Swift observation (ID: 00033420140) has an exposure time −6 −5 −4 −3 k e V ( P ho t on s c m − s − k e V − ) Obs 1
FUV SXT LXP 20 r a t i o Energy (keV)
Figure 4 . Broadband FUV–X-ray SED of Mrk 335 withSXT, LAXPC 20 and UVIT (FUV filters: F148Wa, F154W,F169M & F172M) data from Obs 1. The data are modelledwith optxagnf and xillver , modified by both Galactic andintrinsic absorption and reddening. of only ∼ xrtpipeline andextracted the spectrum and light curve from a circu-lar region of radius 30 pixels. The background regionof 50 pixels radius circle was also selected from thesame image. We generated the XRT light curve in the0.3–10 keV band and did not find any variability. Thenet count rate of the XRT observation is around 0.07counts s − (0.3–10 keV). Since the data quality is notgood, we grouped the spectrum for minimum 5 countsper bin and used cstat while fitting. Modelling the 0.3-10 keV spectrum with T Babs × zT Babs × powerlaw gave a photon index of about 1.8, softer than that ob-tained for the fit of combined SXT and LAXPC spectrain the 0.7–20 keV band. The corresponding fit-staticis cstat / dof = /
9. When a xillver component wasadded the fit-static reduced to 1.46 / N HInt = . . − . × cm − , N pl = . + . − . × − , N xl = . + . − . × − and f refl was not constrained. Addinga bbody , with kT bb fixed at 0.1 keV did not change thefit-statistic any more. However, the spectrum remainedsteeper with Γ > .
3. We also fitted the spectrum with
T Babs × zT Babs ( bbody + powerlaw ) model which re-sulted in a cstat / dof of 1.35 / N HInt < . × cm − , Γ < . N pl = . + . − . × − , kT bb = . + . − . keV, N bb = . + . − . × − . Here, theblackbody normalisation is lower than the results from AstroSat whereas the temperature remains similar. Inthe optical / UV band, the observation was made with . Astrophys. Astr. (0000) :
Table 3 . Best-fit parameters for the spectral fits in the 0.7–20 keV band for
AstroSat (SXT & LAXPC 20) observationstaken on 31 / / / / xillver parameters that kept fixed while fitting are i = ◦ , A Fe = ξ =
0, and E cut =
300 keV.
Energy Model Parameter Obs 1 Obs 2Range2–20 keV tbabs × powerlaw Γ + . − . + . − . N pl (10 − ) 1.64 + . − . + . − . constant 0.50 + . − . + . − . χ / dof 28.74 /
39 31.68 / tbabs × ztbabs × powerlaw N IntH (10 cm − ) < . < . Γ + . − . + . − . N pl (10 − ) 1.88 + . − . + . − . constant 0.49 + . − . + . − . χ / dof 28.56 /
38 30.28 / tbabs ( ztbabs × powerlaw + xillver ) N IntH (10 cm − ) < . + . − . Γ + . − . + . − . N pl (10 − ) 3.91 + . − . + . − . f refl > . > N xl (10 − ) > . > + . − . + . − . χ / dof 23.66 /
36 21.83 / tbabs ( ztbabs × powerlaw + xillver ) N IntH (10 cm − ) < . < . Γ + . − . + . − . N pl (10 − ) 2.56 + . − . + . − . f refl > .
003 0 – 0(?) N xl (10 − ) > . > . + . − . + . − . χ / dof 47.57 /
68 45.28 / tbabs ( ztbabs × ( powerlaw + bbody ) + xillver ) N IntH (10 cm − ) 0.89 + . − . + . − . Γ + . − . + . − . N pl (10 − ) 3.47 + . − . + . − . kT bb (keV) 0.08 + . − . + . − . N bb (10 − ) 4.23 + . − . + . − . f refl + . − . + . − . N xl (10 − ) > . > . + . − . + . − . χ / dof 39.46 /
66 39.37 / tbabs × ztbabs × ( powerlaw + bbody ) N IntH (10 cm − ) < . < . Γ + . − . + . − . N pl (10 − ) 1.81 + . − . + . − . kT bb + . − . + . − . N bb (10 − ) 1.40 + . − . + . − . constant 0.49 + . − . + . − . χ / dof 43.76 /
68 46.21 / J. Astrophys. Astr. (0000) :
Table 4 . The total unabsorbed X-ray flux for the model tbabs ( ztbabs × ( powerlaw + bbody ) + xillver ) from the two AstroSat observations and the near-XRT observation.
Energy range Flux (10 − erg cm − s − )(keV) Obs 1 Obs 2 XRT0.7 – 20 2.14 + . − . + . − . ± . ± . + . − . + . − . + . − . + . − . + . − . + . − . + . − . + . − . + . − . + . − . + . − . Table 5 . Results from the analysis of UVIT data. The saturation corrected net count rates with the corresponding magnitudeand flux density ( F λ ) for each filter are given in the last tree columns. F λ values are corrected for the Galactic reddening. Filter Zero Point Count Rate (counts / s) m AB (magnitude) F λ (10 − erg cm − s − )(magnitude) Obs 1 Obs 2 Obs 1 Obs 2 Obs 1 Obs 2N245M 18.452 +/ -0.005 21.84 ± ± +/ -0.010 15.024 +/ -0.007 2.15 ± ± +/ -0.010 18.14 ± ± +/ -0.011 14.932 +/ -0.012 1.98 ± ± +/ -0.010 3.75 ± ± +/ -0.026 15.020 +/ -0.024 1.76 ± ± ± ± ± +/ -0.012 15.070 +/ -0.014 5.55 ± ± ± ± ± +/ -0.015 15.119 +/ -0.015 4.84 ± ± ± ± ± ± ± ± ± ± only uvw2 filter. We estimated the flux of the sourceusing the task uvotproducts . For this, we selected asource region of 5” radius circle and background cir-cles of larger radius at di ff erent regions from the ob-served images. The background subtracted flux for uvw2 filter is 2 . ± . × − erg cm − s − Å − (notcorrected for Galactic reddening).The X-ray spectrum of Mrk 335 is complex tobe modelled with low-quality data from AstroSat and
Swift . We notice that when the X-ray spectra from XRTand
AstroSat (SXT & LAXPC 20) observations werefitted in the same energy range of 0.7–10 keV with sim-ple models like ( tbabs × zT Babs × powerlaw ), the pa-rameters agree well within errorbars. The discrepancyarises when we fit the broadband X-ray spectrum, in-cluding LAXPC data.We also carried out a preliminary analysis of oneof the XMM-Newton observations of Mrk 335 takenin 2019 January. The EPIC-pn spectrum (net expo-sure ∼
66 ks) showed soft excess emission and a broadiron emission line. We fitted the spectrum in the 0.3–10 keV band with an absorbed (Galactic and intrin-sic) power-law, blackbody and a redshifted broad Gaus-sian component. The fit yielded a photon index of ∼ N HInt < xillver component wasadded, the fit improved significantly, and the photon in- dex increased to around 1.4 while N H remained uncon-strained. The X-ray (2–10 keV) flux from the obser-vation is found to be decreased roughly by a factor of5–6 as compared to the AstroSat observations, but Γ isconsistent within error bars.
7. Discussion & Summary
We found harder X-ray spectra for both
AstroSat obser-vations of Mrk 335. A similar photon index was ob-tained when the XRT spectrum was fitted with an ab-sorbed power-law and blackbody model (0.3–10 keV),that is consistent with the result obtained by Tripathi etal. (2020). However, when a reflection component wasadded in the model, the primary power-law appearedto be softer in the XRT observation. The significanceof reflection in the source was noticed in the previ-ous studies as well. For example, Parker et al. (2019)studied the
XMM-Newton , Swift and
NuSTAR spectrataken in 2018–2019 when the source was showing anextremely low flux level in X-rays. By modelling thebroadband continuum in detail, they found that the hardX-ray spectrum is dominated by distant reflection andthe soft part by photoionised emission lines. They alsoobserved steep X-ray spectra Γ ∼ . Astrophys. Astr. (0000) : F λ ( − e r g c m − s − Å − ) Obs 1Obs 2
Figure 5 . Plot showing the variability in FUV (F148Wa &F154W) and NUV (N245M, N263M & N279M) emissionsfrom Mrk 335 between the two epochs. blackbody component. Earlier observations of Mrk 335reported partial covering absorption and relativistic re-flection in the source (Longinotti et al. 2019, Parker etal. 2019) that we could not parameterise with the
As-troSat data. In a previous study on the correlation be-tween the reflection fraction and photon index in AGN(Ezhikode et al., 2020), we analysed the
NuSTAR spec-trum of Mrk 335 observed in 2013 June. The spec-trum showed the presence of broad and narrow emis-sion lines, and we fitted the 3–79 keV spectrum usingthe models relxill and xillver . The spectrum was steepwith gamma around 2.2, and we did not see any signif-icant intrinsic absorption.A larger X-ray photon index of Γ & . ff er-ent behaviour even after including the neutral reflectionmodel (though the slope of the second observation ismarginally within this range). We also note that theX-ray spectrum gets steeper when the intrinsic obscu-ration is fixed at larger values, although the fit worsens.We obtained the confidence contour plot (see Fig. 6) forphoton index and intrinsic absorption. It is clear fromthe plot that the data could not constrain Γ well and ahigher index similar to those found in other AGN is notruled out. The X-ray emission from the source may beobscured intrinsically, and hence the distant reflectioncould be dominating the observed spectra.During AstroSat observations, Mrk 335 was in alow-flux state in the UV band as well. Our observa-tions, separated by almost 18 days, show variability inboth NUV fand FUV emissions. However, no signifi-cant variability was found between the two X-ray ob-servations. Variable UV emission using
Swift
UVOT . . P a r a m e t e r : P ho I nde x Parameter: nH (10 )cross = 39.373; Levels = 41.673 43.983 48.583Confidence contours: Chi−Squared + Figure 6 . The confidence (one, two and three sigma) contourplot of the parameters N IntH and Γ for Obs 2. observations was detected by Grupe et al. (2008) ontime-scales of days to weeks. They found the UV vari-ability to be following the XRT light curve, suggestingthe possibility of the same mechanism triggering boththe emissions.Considering the time-scale of variability in our ob-servations, X-ray reprocessing could be the origin ofthe observed UV variability in the source. However, asimilar variability is not observed in X-ray emissions,and we do not have enough monitoring observationsto confirm this. The obscuration of X-rays by cloudscould be another possibility of the observed UV vari-ability that is unrelated to X-ray emission. Detailedmodelling of the broadband SED is required to explainthe scenario. Owing to the low signal-to-noise of theX-ray spectra and uncertainties in the UV flux mea-surements, a proper modelling of the UV-X-ray SED isdi ffi cult. Hence, we do not make a definitive statementregarding the UV variability in the source. A more de-tailed characterisation of broadband X-ray continuumemissions may be carried out with future better ob-servations with AstroSat . With simultaneous filter andgrating observations with UVIT, we can also study thenature of variability in the accretion disc emission indepth.
Acknowledgements
We would like to acknowledge the anonymous refereefor the helpful comments and suggestions. We thankProf. Shyam Tandon and Mr. Prajwel Joseph for theuseful discussions on UVIT data analysis. This publi-cation uses the data from the
AstroSat mission of theIndian Space Research Organisation (ISRO), archived
J. Astrophys. Astr. (0000) : at the Indian Space Science Data Centre (ISSDC). Thiswork has used the data from the Soft X-ray Telescope(SXT) developed at TIFR, Mumbai, and the SXT POCat TIFR is thanked for verifying and releasing the datavia the ISSDC data archive and providing the necessarysoftware tools. We thank the UVIT POC at IIA, Ban-galore for the data and their support. This research hasmade use of data, software and / or web tools obtainedfrom the High Energy Astrophysics Science ArchiveResearch Center (HEASARC), a service of the Astro-physics Science Division at NASA / GSFC and of theSmithsonian Astrophysical Observatory’s High EnergyAstrophysics Division.
References
Antia H. M., et al., 2017, ApJS, 231, 10Agrawal P. C., 2017, Journal of Astrophysics and Astron-omy, 38, 27Ballantyne, D. R., Iwasawa, K., & Fabian, A. C. 2001, MN-RAS, 323, 506Bessell M. S., 1991, A&A, 242, L17Bianchi, S., et al. 2001, A&A, 376, 77Boller, T., Brandt, W. N., & Fink, H. 1996, A&A, 305, 53Bhalerao V., et al., 2017,JAA,38, 31Buisson D. J. K., Lohfink A. M., Alston W. N., Fabian A. C.,2017, MNRAS, 464, 3194Cardelli J. A., Clayton G. C., Mathis J. S., 1989, ApJ, 345,245Chainakun P., Young A., 2015, ebha.conf, 76Done C., Davis S. W., Jin C., et al., 2012, MNRAS, 420,1848Ezhikode S. H., et al., 2020, MNRAS, 495, 3373Gallo L. C., et al., 2013, MNRAS, 428, 1191Gallo L. C., et al., 2015, MNRAS, 446, 633Gallo, L., Blue, D. M., Grupe, D., et al. 2018, MNRAS, 478,2557Gallo L., 2018, rnls.conf, 34Garcia, J., & Kallman, T. R. 2010, ApJ, 718, 695.Garcia, J., Dauser, T., Reynolds, C. S., Kallman, T. R., Mc-Clintock, J. E., Wilms, J., & Eikmann, W. 2013, ApJ, 768,146.George, I. M., Turner, T. J., Yaqoob, T., Netzer, H., Laor, A.,Mushotzky, R. F., Nandra, K., & Takahashi, T. 2000, ApJ,531, 52Goodrich R. W., 1989, ApJ, 342, 224Grupe, D., Thomas, H.-C., & Beuermann, K. 2001, A&A,367, 470 Grupe, D., Komossa, S., & Gallo, L. 2007, ApJ, 668, L111Grupe D., Komossa S., Gallo L. C., Fabian A. C., Larsson J.,Pradhan A. K., Xu D., Miniutti G., 2008, ApJ, 681, 982Grupe D., Komossa S., Gallo L. C., Longinotti A. L., FabianA. C., Pradhan A. K., Gruberbauer M., Xu D., 2012, ApJS,199, 28Halpern, J. P. 1982, Ph.D. thesis, Harvard Univ.Kalberla P. M. W., Burton W. B., Hartmann D., Arnal E. M.,Bajaja E., Morras R., P¨oppel W. G. L., 2005, A&A, 440,775Keek L., Ballantyne D. R., 2016, MNRAS, 456, 2722Komossa S., Grupe D., Saxton R., Gallo L., 2014, Proceed-ings of Swift: 10 Years of Discovery (SWIFT 10), id. 143Komossa S., Grupe D., Gallo L. C., Poulos P., Blue D., KaraE., Kriss G., Longinotti A. L., Parker M. L., and WilkinsD., 2020, A&A 643, L7Leighly, K. M. 1999a, ApJS, 125, 297Longinotti A. L. et al., 2013, ApJ, 766, 104Longinotti A. L., et al., 2019, ApJ, 875, 150Nandra, K., & Pounds, K. A. 1994, MNRAS, 268, 405Osterbrock D. E., Pogge R. W. 1985, ApJ, 297, 166Parker M. L. et al., 2014, MNRAS, 443, 1723Rao A. R., Bhattacharya D., Bhalerao V. B., Vadawale S. V.,Sreekumar S., 2017, Curr. Sci., 113, 595Singh, K. P., Tandon, S. N., Agrawal, P. C., et al. 2014, Proc.SPIE, 9144, 91441SSingh, K. P., Stewart, G. C., Westergaard, N. J., et al. 2017,JApA, 38, 29Sarma R., Tripathi S., Misra R., Dewangan G., Pathak A.,Sarma J. K., 2015, MNRAS, 448, 1541Schlegel D. J., Finkbeiner D. P., Davis M., 1998, ApJ, 500,525Tananbaum, H., Peters, G., Forman, W., Giacconi, R., Jones,C., & Avni, Y. 1978, ApJ, 223, 74Tandon S. N., et al., 2017a, Journal of Astrophysics and As-tronomy, 38, 28Tandon S. N., et al., 2017b, AJ, 154, 128Tripathi, S., McGrath, K. M., Gallo, L. C., et al. 2020, MN-RAS, 499, 1266Vadawale S. V., et al., 2016, in Space Telescopes andIn-strumentation 2016: Ultraviolet to Gamma Ray. p.99051GWilkins, D. R, et al., 2015, MNRAS, 454, 4440Yadav J. S., et al., 2016, Large Area X-ray ProportionalCounter (LAXPC) instrument onboard ASTROSAT. p.99051D, doi:10.1117 //