Broadband variability and correlation study of 3C 279 during flare of 2017-2018
aa r X i v : . [ a s t r o - ph . H E ] J a n Draft version January 15, 2020
Preprint typeset using L A TEX style emulateapj v. 01/23/15
BROADBAND VARIABILITY AND CORRELATION STUDY OF 3C 279 DURING FLARE OF 2017-2018
Raj Prince Raman Research Institute, Sadashivanagar, Bangalore 560080, India
Draft version January 15, 2020
ABSTRACTA multiwavelength temporal and spectral analysis of flares of 3C 279 during November 2017–July 2018are presented in this work. Three bright gamma-ray flares were observed simultaneously in X-ray andOptical/UV along with a prolonged quiescent state. A “harder-when-brighter” trend is observed inboth gamma-rays and X-rays during the flaring period. The gamma-ray light curve for all the flaresare binned in one-day time bins and a day scale variability is observed. Variability time constrains thesize and location of the emission region to 2.1 × cm and 4.4 × cm, respectively. The fractionalvariability reveals that the source is more than 100% variable in gamma-rays and it decreases towardsthe lower energy. A cross-correlation study of the emission from different wavebands is done usingthe DCF method, which shows a strong correlation between them without any time lags. The zerotime lag between different wavebands suggest their co-spatial origin. This is the first time 3C 279has shown a strong correlation between gamma-rays and X-rays emission with zero time lag. A singlezone emission model was adopted to model the multiwavelength SEDs by using the publicly availablecode GAMERA. The study reveals that a higher jet power in electrons is required to explain thegamma-ray flux during the flaring state, as much as, ten times of that required for the quiescent state.However, more jet power in magnetic field has been observed during the quiescent state compared tothe flaring state.
Keywords: galaxies: active – galaxies: jets – gamma rays: galaxies – quasars: individual (3C 279) INTRODUCTION
Blazars are a class of active galactic nuclei whosejets are oriented close to the observer’s line of sight(Urry & Padovani 1995). In general, blazars exhibithighly luminous and rapidly variable non-thermal contin-uum emission across the entire electromagnetic spectrumextending from radio to very high energy gamma-ray.A wide range of variability time across the whole elec-tromagnetic spectrum is found in most of the blazars.A time scale of variability ranging from minutes toyears is inferred for blazars (Aharonian et al. 2007;Raiteri et al. 2013). Blazars are generally classified intotwo categories viz., BL Lac objects (BL Lacs) and flatspectrum radio quasar FSRQ, depending upon their op-tical spectra. BL Lacs shows a very weak or no emissionline in their optical spectra while on the other hand FS-RQs are known for their strong, broad emission lines.The highly energetic phenomenon inside the blazars aredetected as strong and spectacular flares across the entireelectromagnetic spectrum, with rapid variability. Therehave been several studies on blazar to understand thebroadband flaring activity, but the origin of fast vari-ability is not well understood. The observed broadbandspectral energy distribution (SED) of blazar shows twopeaks, one extending from radio to optical, and anotherranging from X-ray to gamma-rays. The low energy radi-ation from radio to UV/X-ray is caused by synchrotronradiation of relativistic electrons accelerated inside thejet. In leptonic models, the high energy gamma-ray ra-diation is caused by inverse Compton (IC) scattering ofsoft target photons originating in synchrotron radiation(SSC; Sikora et al. 2009) or external photon fields (EC; [email protected]
Dermer et al. 1992; Sikora et al. 1994).3C 279 has been classified as a FSRQ at redshift, z of0.536 (Lynds et al. 1965), and is a well-studied blazarin its class. The mass of the black hole was estimatedin the range of (3–8) × M ⊙ by Woo & Urry (2002),Gu et al. (2001), and Nilsson et al. (2009) from the lu-minosity of broad optical emission lines, the width ofthe H β lines, and from the luminosity of the host galaxyrespectively. Fermi-LAT is continuously monitoring 3C279 since 2008 and along with that it is also monitoredby different other facilities in X-ray, optical, and ra-dio. The blazar 3C 279 has been studied in great ex-tent in past throughout the entire electromagnetic spec-trum. The previous multi-wavelength studies shows thatthe gamma-ray emission region in 3C 279 lies close tothe base of the jet, at sub-parsec scale (Hayashida et al.2015; Paliya 2015a; Paliya et al. 2016). However, therecent study by Patino-Alvarez et al. (2019) found anevidence of high energy emission emitted at much largerdistance from the core, at ∼
42 pc. Further, multi-wavelength study on this source will help the communityto probe the location of the gamma-ray emission regionalong the jet axis.The correlation study between optical polarization(degree/angle) and gamma-rays give strong evidencefor the synchrotron and Compton model, and suggestthat the jet structure is not axisymmetric (Abdo et al.2010). The above correlation also suggest the presenceof helical magnetic field component (Zhang et al.2015). The rate of change of the polarization anglewith fast variability time suggests a compact emissionregion, close to the central black hole, located at adistance along the jet. A detailed correlations studybetween optical and gamma-ray has been done for asample of blazars by Cohen et al. (2014), and theyhave found a time lag of 110 days between opticaland gamma-ray emission. A similar study has beendone by Hayashida et al. (2012) for 3C 279 and theyhave noticed a lag of 10 days between optical andgamma-ray emission. They also found that the X-rayand gamma-ray emissions are not well correlated, andthe nature of X-ray emission in blazar is still unclear. Acorrelation study between radio and gamma-rays is alsodone by Pushkarev et al. (2010), and they noticed thatin a sample of 183 blazars most of them show that theradio flares lag the gamma-ray flare. A similar result wasfound in case of Ton 599 by Prince (2019), where a lagof 27 days was noticed between radio and gamma-rays.Keeping consistency with results from earlier studies,here I try to investigate the possible correlation betweendifferent wavelength during the 2017-2018 flare of 3C 279. MULTIWAVELENGTH OBSERVATIONS AND DATAANALYSIS
Fermi-LATFermi -LAT is a pair conversion γ -ray Telescope sen-sitive to photon energies between 20 MeV to higherthan 500 GeV, with a field of view of about 2.4 sr(Atwood et al. 2009). The LAT’s field of view coversabout 20% of the sky at any time, and it scans the wholesky every three hours. NASA launched the instrument in2008 into a near earth orbit. Fermi -LAT is continuouslymonitoring 3C 279 since 2008 August. The standard datareduction and analysis procedure has been followed. Ihave chosen a circular region of radius 10 ◦ around thesource of interest during the analysis. This circular re-gion is also known as region of interest (ROI). Furtheranalysis procedure is the same as given in Prince et al.(2018). I have analyzed the Fermi -LAT data for 3C 279from Nov 2017 to Jul 2018 and found that source hasshown three significant flares in these nine months.
Swift-XRT/UVOT
Swift-XRT/UVOT is the space-based telescope whichobserves the galactic as well as extragalactic sources in X-rays, Opticals, and UV simultaneously. Fermi detectedblazars could also be observed by Swift-XRT/UVOTtelescope as a monitoring program as well as a time ofopportunity (ToO) program. The blazar 3C 279 was ob-served by Swift-XRT/UVOT when it was flaring dur-ing 2017-2018, and the details of the observations arepresent in Table 1. I processed the XRT data by us-ing the task ‘ xrtpipeline ’ version 0.13.2, which producesthe cleaned event files for each observation. While re-processing the raw data, I have used the latest calibrationfiles (CALDB version 20160609). The cleaned event filesare produced only to the Photon Counting (PC) modeobservations. The task ‘ xrtpipeline ’ is used to select thesource and the background region. Circular regions of ra-dius 20 arcseconds around the source and slightly awayfrom the source (fewer photon counts) are chosen for thesource and the background regions, respectively. TheX-ray spectra were extracted in ’ xselect ’ and used as in-put spectra in ’
Xspec ’ for modeling. A simple power-law https://fermi.gsfc.nasa.gov/ssc/data/analysis/documentation/ Table 1
Table shows the log of the observations from Swift-XRT/UVOTtelescope during the flaring state (MJD 58050 – 58350).Obs-ID Exposure (ks)00035019201 1.900035019203 1.800035019204 2.000035019206 0.500035019210 0.900035019211 0.900035019213 0.600035019214 0.900035019218 1.100035019219 1.000035019220 1.200035019221 2.100035019222 1.700035019224 1.800035019225 1.600035019227 2.500035019228 0.200035019229 0.700035019230 2.000035019231 2.000035019232 2.000035019233 2.100035019234 1.500035019235 0.800035019236 1.600035019237 0.500035019238 1.500035019239 1.500035019240 1.500035019241 1.100035019242 1.100030867052 1.000030867054 1.400030867055 0.400030867056 1.100030867057 1.000030867058 1.000030867059 1.000030867060 1.100030867061 0.500030867062 1.200030867063 0.600030867064 1.1 model with the galactic absorption column density n H =1.77 × cm − (Kalberla et al. 2005) is used to modelthe XRT spectra.Simultaneous to XRT, the Swift Ultraviolet/OpticalTelescope (UVOT, Roming et al. 2005) has also ob-served 3C 279 in all the available six filters U, V, B, W1,M2, and W2. The image is extracted by selecting a circu-lar region of 5 arcsecond around the source and a circularregion of 10 arcsecond away from the source for the back-ground. The task ’uvotimsum’ is used to sum the mul-tiple observations in the same filter at the same epoch,and further, the task ‘uvotsource’ is used to extract thesource magnitudes and fluxes. Magnitudes are correctedfor galactic extinction using R V = A V /E(B-V) = 3.1 andE(B-V) = 0.025 (Schlafly & Finkbeiner 2011) and con-verted into flux using the zero points (Breeveld et al.2011) and conversion factors (Larionov et al. 2016). Steward Optical Observatory
Steward Optical Observatory is a part of the
Fermi multiwavelength support program. It provides the op- F . − G e V Fermi-LAT
Quiescent state
Flare A Flare B Flare CLight curve of 3C 279 during 2017 Nov - 2018 Jul F − k e V Swift-XRT F O p t i c a l Swift-UVOT
UVB F U V W1M2W2 M a g Steward VR P A ( d e g ) SPOL-Polarization D o P ( % ) SPOL-Polarization angle F l u x ( J y ) SMA(230 GHz)
Time[MJD] F l u x ( J y ) OVRO(15 GHz)
Figure 1.
Multi-wavelength light curve of 3C 279 from November 2017 to July 2018. The γ -ray flux are presented in units of 10 − phcm − s − , and X-ray/UV/Optical fluxes are in units of 10 − erg cm − s − . The vertical cyan color line separates the quiescent state andflaring state. The red vertical lines corresponding to each gamma-ray flares are drawn to recognised the flare in all wavebands. tical data for the LAT-monitored blazars and also mea-sures the linear optical polarization. Archival data fromthe Steward Optical Observatory, Arizona (Smith et al.2009) has been used in this particular study.3C 279 is being continuously monitored with the SPOLCCD Imaging/Spectrometer of the Steward Observatory.Optical V-band and R-band photometric and polarimet-ric (degree of polarization and position angle) data iscollected for the flaring period of 3C 279 from November2017 to July 2018. Radio data at 15 and 230 GHz
Owens Valley Radio Observatory (OVRO;Richards et al. (2011) is also a part of
Fermi mon-itoring program. It monitors the Fermi blazars by a40-meter single disc antenna at a frequency of 15 GHz.More than 1800 Fermi blazars are being continuouslymonitored by OVRO twice a week. 3C 279 is one ofthem, and I have collected the radio data at 15 GHzfrom November 2017 to July 2018.Submillimeter Array provided the 230 GHz data(SMA) from observer center database (Gurwell et al.2007). The data is collected for the period of 9 monthsfrom November 2017 to July 2018. RESULTS AND DISCUSSIONS
I have analyzed the
Fermi -LAT, Swift-XRT/UVOTdata from November 2017 to July 2018 (MJD 58060 –MJD 58330) and the archival data from other telescopeslike OVRO, SMA, and Steward observatory are collectedfor the same period. The multiwavelength data from allthe above telescopes are used to study the flux and po-larization variability, and it is discussed in this particularsection. I have also presented a correlation study amongdifferent wavelength during the outburst period (MJD58060 – MJD 58330). A single zone emission model ischosen to perform the multiwavelength SED modeling.
Multiwavelength Light Curves
Blazar 3C 279 is known for its chaotic variability andactive flaring behavior across the entire electromagnetic(EM) spectrum. The source was reported to be in a flar-ing state across the entire EM spectrum between 2017-2018. The multiwavelength light curve of 3C 279 duringthe flaring episode MJD 58060 – MJD 58330 is shown inFigure 1. The first panel shows the one day binning of
Fermi -LAT data in the energy range of 0.1 - 300 GeV.Our analysis shows the source was in flaring state duringend of 2017 to mid 2018. The high activity started atthe end of 2017 and followed by a flaring state defined as“Flare A”. After “Flare A”, the source was observed foraround two months in a low state with small fluctuationsin the flux value. 3C 279 again went to higher state andspent around two weeks in flaring state labeled as “FlareB”. After one month period of “Flare B”, the flux againstarted rising, and source attained a full flaring state la-beled as “Flare C”. The red bold vertical lines are drawnin Figure 1 corresponding to each gamma-ray flare. Justbefore the “Flare A”, the source has been observed in along low flux state. I have chosen 50 days period betweenMJD 58060–58110 when the source flux is very low and http://james.as.arizona.edu/ psmith/Fermi/ constant over a long time. This period defined as “qui-escent state” showed by cyan color data points in Figure1 and separated by flaring state by a cyan color verti-cal line. The average flux observed during this period is0.73 × − ph cm − s − .Swift-XRT/UVOT has also monitored the source dur-ing the gamma-ray outburst. The X-rays, optical, andUV light curves are shown in panels 2, 3, & 4 of Figure 1.3C 279 has gone through the strong X-ray flaring episodecorresponding to each gamma-ray flare. In Optical andUV band 3C 279 has shown the flaring behavior corre-sponding to “Flare A” and “Flare C”, while “Flare B”of gamma-ray is missing in Optical and UV because ofthe unavailability of UVOT observations.The 3C 279 has also been monitored by Steward Ob-servatory in optical V and R band. In panel 5 of Figure1, I have plotted the Steward data, and it can be seenthat the source is highly variable in both the bands. The“Flare A” of gamma-ray is followed by the flare in bothV and R band. The “Flare B” is observed close in timeat optical (Steward Observatory) and gamma-ray, while“Flare C” is not seen in Steward Observatory because ofpoorly sparsed V and R band data points.In the 6th and 7th panel of Figure 1, I have plottedthe optical degree of polarization (DoP) and polarizationangle (PA) from Steward Observatory. Huge variationsare seen during the flaring period. DoP varies from 4%–22% in 12 days of span (MJD 58130–58142) and on theother hand a slow change is seen in polarization anglefrom 25 ◦ –60 ◦ .The radio light curve from SMA and OVRO observa-tory at 230 and 15 GHz are shown in the last two panelof Figure 1. The radio data at 230 GHz from SMA ob-servatory shows that the source is variable in this energyband, and high state radio flux has been noticed dur-ing “Flare A” and “Flare B”. It is observed that thereare not much variations in OVRO flux during the flaringperiod of gamma-ray/X-ray/Optical/UV. Variations in Gamma-ray
In Figure 2, the gamma-ray flares are plotted sepa-rately along with the corresponding photon spectral in-dex. The “Flare A” is shown in the leftmost panel ofFigure 2. The source starts showing the activity at MJD58115 with a small rise in flux. The small fluctuationscontinued until the source went to the higher state, wherethe flux rose above the quiescent state flux value withinthe 10 days of time interval (MJD 58129–58140). Theflux at MJD 58129.5 is 1.89 × − ph cm − s − and al-most after 7 days the flux rose up to 22.24 × − phcm − s − at MJD 58136.5 and the photon spectral indexbecame harder changing from 2.43 to 2.19, respectively.Just after three days from the peak, the flux started de-creasing and within 10 days, it attained the low flux stateat MJD 58149.5 with a flux of 1.08 × − ph cm − s − and spectral index 2.31.The middle panel of Figure 2 represents the light curveof “Flare B”. The activity was found to commence atMJD 58215 just after two months of “Flare A”. Theflux started rising very slowly from MJD 58216.5 withflux 1.35 × − ph cm − s − and in span of two weeks itachieved the maximum flux value of 19.06 × − ph cm − s − at MJD 58228.5. The spectral index became harderfrom 2.84 to 2.04. Within one day the flux dropped from Table 2
Table shows the variability time estimated from equation (1) forall the different flares. The flux F and F are in units of 10 − phcm − s − and t & t are in MJDs .Flares F F t t t var (days)Flare A 1.89 3.91 58129.5 58130.5 1.43 ± ± ± (19.06–7.49) × − ph cm − s − and then the sourcewent through small fluctuations in flux for 20 days be-fore reaching the quiescent state. The quiescent stateflux noticed at MJD 58249.5 is 1.00 × − ph cm − s − .The “Flare C” is shown in the rightmost panel of Fig-ure 2. As soon as “Flare B” ends, the flux startedfluctuating again above the quiescent state flux value(0.27 × − ph cm − s − ) and it continues till one weekand shows a little rise in flux (3.36 × − ph cm − s − ).After spending two days in a high flux state, the sourcecomes back to low flux state, and after five days it againrises to higher flux state. The maximum flux achievedduring high flux state is 13.31 × − ph cm − s − atMJD 58271.5.One day bin light curve shown in Figure 2 are usedto estimate the variability time, which can be defined bythe following equation (Zhang et al. 1999), t var = F + F t − t | F − F | (1)where F and F are the fluxes measured at time t and t . The above equation (1) is used to scan the 1 day binlight curve (Figure 2) to find the variability time. Theshortest variability time is found to be order of days inall the three flares and the details are shown in Table 2.In Figure 3, the gamma-ray flux and the photon spec-tral index during the flaring period are plotted togetheralong the x and y-axis, respectively. A “harder-when-brighter” trend of source is seen here. Similar trendis also seen previously for 3C 279 by Hayashida et al.(2012), Hayashida et al. (2015), & Paliya (2015a). Thespectral hardening during the flaring state can predictthe possibility of detection of high energy photons andconsequently can shift the IC peak of the SED to higherenergy. A strong correlation between spectral harden-ing during flare and detection of high energy photons areshown by many authors (Britto et al. 2016; Shah et al.2019). Variations in X-ray
Swift-XRT light curve of all the flares for 2-10 keVis shown in Figure 4 along with the photon spectral in-dex. The observations done by Swift-XRT are poorlysparsed, and as a result, there is no clear indication ofrising and decaying part of the flares. The maximum fluxachieved during “Flare A” in X-ray is (6.03 ± × − erg cm − s − with spectral index 1.23 ± ± × − erg cm − s − and the spec-tral index noticed as 1.22 ± ± × − erg cm − s − , and the correspondingspectral index found to be 1.35 ± ± ± Spectral Ananlysis
The γ -ray spectral energy distributions (SEDs) havebeen generated for all the three flares and one quies-cent state between 0.1 – 300 GeV, from the likelihoodanalysis. The spectral data points are plotted in Fig-ure 5, I have used three different spectral models to fitthe SEDs data points. The spectral models are Power-law (PL), Logparabola (LP), and simple broken Power-law (BPL), and corresponding expressions are given inPrince et al. (2018). The presence or absence of curva-ture in the gamma-ray SED plays a crucial role in con-straining the location of the emission region. A breakin the gamma-ray spectrum is expected when the emis-sion region is within the broad line region (BLR), be-cause of photon- photon pair production. Earlier studyby Liu & Bai (2006) has shown that the BLR region isopaque to photons of energy >
20 GeV , and as a re-sult, a curvature or break can be seen in the gamma-rayspectrum above 20 GeV. The curvature in the gamma-ray spectrum can be justified by estimating the TS curve .According to Nolan et al. (2012) the TS curve can be de-fined as TS curve = 2(log L(LP/BPL) - log L(PL)), whereL represents the likelihood function. The model param-eters and the value of TS curve are mentioned in Table 3.The model which have a large positive value of TS curve is considered as the best fit to the SEDs data pointsand it suggests the presence of a spectral cut-off. FromTable 3, it is very much clear that the log-parabola spec-tral model better explains the gamma-ray SED data. Ihave also noticed that the break energy found in bro-ken power-law fit is constant irrespective of the differentflares. This finding is also consistent with the previousstudy on this source, by Paliya (2015a), and on 3C 454.3by Abdo et al. (2011). There is also an alternative wayto explain this curvature/cut-off. A cut-off in the energyspectrum can also occur when there is already a cut-offin the energy distributions of the particles.A strong break is seen in the gamma-ray spectrumwhile fitting with BPL; similar break has also been seen F . − G e V Time [MJD+58110] −3.5−3.0−2.5−2.0 Γ F . − G e V Time [MJD+58215] −3.2−3.0−2.8−2.6−2.4−2.2−2.0−1.8 Γ F . − G e V Time [MJD+58250] −3.0−2.5−2.0−1.5−1.0 Γ Figure 2.
Gamma-ray light curve for all the three flares separately with corresponding photon spectral index. The y-axis of the upperpanels are in units of 10 − ph cm − s − . -1 F GeV [ ×10 −6 ph cm −2 s −1 ] −3.0−2.8−2.6−2.4−2.2−2.0−1.8−1.6 Γ Figure 3.
The gamma-ray photon spectral index correspondingto flux observed during flare shows a “harder-when-brighter” trend. F − K e V A B C
Time [MJD] Γ Figure 4.
X-Ray light curve for all the observed flares. The fluxesare in units of 1.0 × − erg cm − s − . Lower panel represent thecorresponding photon spectral index. A “harder-when-brighter”trend is also seen here. before for other FSRQ like 3C 454.3 by Abdo et al.(2011) and also for 3C 279 during flare of 2014 and 2015(Paliya 2015a; Paliya et al. 2015b). In table 3, dur-ing all the flares the BPL photon index (Γ ) before thebreak (E break = 1 GeV) is ≤
2, which suggest an increas-ing slope, and after the break the photon spectral index(Γ ) is ≥
2, indicates a falling spectrum. We know thatthe gamma-ray energy spectrum is governed by inverse Compton (IC) scattering and my result shows it peaksaround 1 GeV which lies in the energy range of LAT (0.1-300 GeV). Since the break energy is almost constant forall the flares, it is possible that the observed shape isa reminiscence of the electron energy distribution of theemitting electrons.
Fractional variability (F var ) Blazars are a particular class of AGN that show strongchaotic variability at all frequencies. The variability ismore evident during the flaring period and the flare pro-files mostly depend on the particle injection, particle ac-celeration, and energy dissipation in the jets of blazars.The variability amplitude can be estimated by measur-ing all kind of jet parameters like the magnetic field inthe jets, viewing angle of the jet, particle density in-side the jet, and finally the efficiency of the accelerationprocess involved within the jet. The challenge to deter-mine the variability amplitude across the energy band re-quires a good quality of data. 3C 279 is a well-observedsource across the entire electromagnetic spectrum, andthat makes it possible to estimate the variability ampli-tude. The fractional root mean square (rms) variabilityparameter (F var ) is a tool to determine the variabilityamplitude and that is introduced by Edelson & Malkan(1987); Edelson et al. (1990). The relation given inVaughan et al. (2003) is used to determine the frac-tional variability (F var ), F var = r S − σ r (2) err ( F var ) = s(cid:16)r N . σ r F var (cid:17) + (cid:16)r σ N . r (cid:17) (3)where, σ XS = S – σ , is called excess variance, S isthe sample variance, σ is the mean square uncertaintiesof each observation and r is the sample mean.The fractional variability estimated across the entireelectromagnetic spectrum is shown in Figure 6, and thevalues are presented in Table 4. It is seen that the sourcevariability is more than 100% in γ -ray, followed by X-rayat more than 60%, and then followed by UV and Opti-cal, where the source variability is more than 50%. Inradio at 15 and 230 GHz, the source is very less variable(less than 10%) during this particular gamma-ray flar-ing period. The F var is 1.201 in γ -ray, 0.580 in UVM2- -1 Energy (GeV) 10 -11 -10 -9 E d N / d E ( e r g c m − s − )
3C 279 (Flare A)
PL LP BPL -1 Energy (GeV) 10 -11 -10 -9 E d N / d E ( e r g c m − s − )
3C 279 (Flare B)
PL LP BPL -1 Energy (GeV) 10 -10 -9 E d N / d E ( e r g c m − s − )
3C 279 (Flare C)
PL LP BPL -1 Energy (GeV) 10 -11 -10 E d N / d E ( e r g c m − s − )
3C 279 (Quiescent State)
PL LP BPL
Figure 5.
The gamma-ray SEDs produced for all the flares and quiescent state. The data points are fitted with three different modelsand the modeled parameters are shown in Table 3.
Table 3
Parameters obtained from the spectral analysis fit, for the different models PL, LP, and BPL, by using the Likelihood analysis method.TS curve is estimated with respect to the TS value of the PL fit.PowerLaw (PL)Activity F . −
300 GeV
Γ TS TS curve (10 − ph cm − s − )Quiescent state 0.67 ± ± ± ± ± ± ± ± α β Quiescent state 0.64 ± ± ± ± ± ± ± ± ± ± ± ± Γ E break Quiescent state 0.65 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± Table 4
Fractional variability in various wavebands are estimated for thetime interval MJD 57980 to 58120.Waveband F var err(F var ) γ -ray 1.201 0.008X-ray 0.660 0.035U 0.524 0.008B 0.520 0.006V 0.509 0.008W1 0.548 0.008M2 0.580 0.007W2 0.557 0.007OVRO (15 GHz) 0.032 0.002SMA (230 GHz) 0.089 0.010 Frequency (Hz) F r a c t i o n a l v a r i a b ili t y a m p li t u d e Radio Optical/UV X-ray Gamma-ray
Figure 6.
Fractional variability for various wavebands are plottedwith respect to their frequency. band, 0.520 in optical U-band and 0.032 in radio at 15GHz, and 0.089 at 230 GHz. It is found that F var isincreasing with energy, suggesting that a large numberof particles are injected in the jet, resulting in high en-ergy emission. Similar behavior of fractional variability isalso seen for other FSRQ like Ton 599 by Prince (2019),where he found a trend of large fractional variability to-wards higher energies. This is not always the case, and anopposite trend was also noticed in the past. An oppositetrend is reported by Bonning et al. (2009), where vari-ability amplitudes decrease towards shorter wavelength(IR, Optical, and UV), which suggests the presence ofsteady thermal emission from the accretion disk.
Correlation Studies
In Figure 1, all the light curves are plotted together,and the visual inspection suggests that the flares in γ -ray, X-ray, Optical, and UV band are mostly correlated.The radio light curve at 15 GHz does not show any flarefor this particular period, while the light curve at 230GHz shows a rise in flux corresponding to “Flare A” and“Flare B”. The gamma-ray “Flare C” detected in X-rays,and Optical/UV have not seen in radio at 230 GHz. Thecross-correlation study of flux variations in different en-ergy bands are presented in this section. The correlationsstudy help to reveal the location of the emission regionin different energy bands with respect to each other. Astrong and good correlation between two energy emis-sions indicates their co-spatial origin. A positive and negative time lag between two energy bands with strongcorrelation coefficient suggest the different emission re-gion for different waveband emissions and the time lagcan be used to estimate the separation between them(Fuhrmann et al. 2014). The correlation study is doneby using the discrete correlations function (DCF) for-mulated by Edelson & Krolik (1988). When the twolight curves, suppose LC1 and LC2 are correlated then apositive time lag between these two light curves impliesthat the LC1 is leading the LC2, and a negative time lagmeans the opposite. The results of DCFs are shown inFigure 7, Figure 8 & Figure 9 for “Flare A” & “Flare C”,for various waveband combinations.A cross-correlation study between two light curves (dif-ferent energy band) by
DCF needs sufficient data pointsin each light curve. The “Flare B” fails to fulfil this cri-teria since it has good quality data points only in thegamma-ray light curve. It is therefore not possible tocross-correlate two different energy band for “Flare B”.Significant correlation has been noticed between gamma-ray and optical band (U, B, & V) emissions with zerotime lag (within the time bin 2.5 days). The correla-tion coefficient for “Flare A” is found to be 0.50 ± ± ± ± ± ± ± ± ± ± ± ± ± ± MODELING THE MULTIWAVELENGTH SED
The simultaneous observation of 3C 279 in different en-ergy bands during its 2017–2018 flaring period providesan opportunity to gain further insight into its multiwave-length properties. In this work, theoretical modeling ofthe observed SED is done using the publicly availablecode GAMERA (Hahn 2015) which solves the time-dependent transport equation and calculates the propa-gated electron distribution.This electron spectrum finally used as an input to esti-mate the synchrotron, synchrotron self-Compton (SSC),and inverse Compton (IC) emissions. The transport http://joachimhahn.github.io/GAMERA −20 −15 −10 −5 0 5 10 15 20Time lag (days)−1.0−0.50.00.51.0 D C F Flare A: Gamma-ray vs U-band −20 −15 −10 −5 0 5 10 15 20Time lag (days)−1.0−0.50.00.51.0 D C F Flare A: Gamma-ray vs B-band −20 −15 −10 −5 0 5 10 15 20Time lag (days)−1.0−0.50.00.51.0 D C F Flare A: Gamma-ray vs V-band −20 −15 −10 −5 0 5 10 15 20Time lag (days)−1.0−0.50.00.51.0 D C F Flare A: Gamma-ray vs W1-band −20 −15 −10 −5 0 5 10 15 20Time lag (days)−1.0−0.50.00.51.0 D C F Flare A: Gamma-ray vs M2-band −20 −15 −10 −5 0 5 10 15 20Time lag (days)−1.0−0.50.00.51.0 D C F Flare A: Gamma-ray vs W2-band
Figure 7.
DCFs shown for “Flare A” for all the possible combinations: γ vs. Swift-U, B, V, W1, M2, W2 band. The significance shownin cyan color is 95%. −20 −15 −10 −5 0 5 10 15 20Time lag (days)−1.5−1.0−0.50.00.51.01.5 D C F Flare C: Gamma-ray vs U-band −20 −15 −10 −5 0 5 10 15 20Time lag (days)−1.5−1.0−0.50.00.51.01.5 D C F Flare C: Gamma-ray vs B-band −20 −15 −10 −5 0 5 10 15 20Time lag (days)−1.5−1.0−0.50.00.51.01.5 D C F Flare C: Gamma-ray vs V-band −20 −15 −10 −5 0 5 10 15 20Time lag (days)−1.5−1.0−0.50.00.51.01.5 D C F Flare C: Gamma-ray vs W1-band −20 −15 −10 −5 0 5 10 15 20Time lag (days)−1.5−1.0−0.50.00.51.01.5 D C F Flare C: Gamma-ray vs M2-band −20 −15 −10 −5 0 5 10 15 20Time lag (days)−1.5−1.0−0.50.00.51.01.5 D C F Flare C: Gamma-ray vs W2-band
Figure 8.
DCF shown for “Flare C” for all the possible combinations: γ vs. Swift-U, B, V, W1, M2, W2 band. The significance shownin cyan color is 95%. −30 −20 −10 0 10 20 30Time lag (days)−1.5−1.0−0.50.00.51.01.5 D C F Flare A: Gamma-ray vs X-ray −30 −20 −10 0 10 20 30Time lag (days)−1.5−1.0−0.50.00.51.01.5 D C F Flare C: Gamma-ray vs X-ray
Figure 9.
Gamma-ray vs X-ray DCF are Shown for “Flare A” and “Flare C”. The significance shown in cyan color is 95%. ∂N ( E, t ) ∂t = Q ( E, t ) − ∂∂E (cid:16) b ( E, t ) N ( E, t ) (cid:17) (4)where, Q ( E, t ) is the injected electron spectrum, and N ( E, t ) is the spectrum achieve after the radiative loss.The radiative loss due to synchrotron, SSC, and ICare represented by b ( E, t ). GAMERA takes care ofthe inverse Compton in the Klein-Nishina regime fromBlumenthal & Gould (1970). The transport equationdoes not have the diffusive loss term as it is insignificantcompared to the radiative loss of the electrons. The log-parabola spectral model is chosen as the input injectedelectron spectrum, motivated from the gamma-ray spec-tral analysis. To model the multiwavelength SEDs, mymodel considers a single spherical emitting zone or blobwhich is moving down the jet along the jet axis with aLorentz factor, Γ, and Doppler factor, δ . The externalphoton field required for external Compton emission isbelieved to be dominated by the broad line region (BLR)photons, particularly in FSRQ like 3C 279. The BLRphoton density in the comoving frame is given by, U ′ BLR = Γ η BLR L disk πcR BLR (5)where the η BLR represents the fraction of disk emis-sion processed in BLR, I assume it to be typically 2%(Pittori et al. 2018), R
BLR is the size of the BLR, L disk denotes the disk luminosity, and c is the speed of lightin vacuum. The photon energy density in BLR is onlya fraction η BLR ∼ . BLR =10 L / d, (Ghisellini & Tavecchio 2009), where L d, isthe accretion disk luminosity in units of 10 erg s − .For L disk = 2 × erg s − (Pian et al. 1999), the sizeof BLR is estimated to be R BLR = 1.414 × cm.The minimum Doppler factor ( δ min ) during the flarecan be estimated from the γγ opacity arguments andby estimating the highest energy photon. The minimumDoppler factor can be calulated as (Dondi & Ghisellini1995; Ackermann et al. 2010), δ min ∼ = (cid:20) σ T d L (1 + z ) f x ǫ t var m e c (cid:21) / (6)which assumes that the optical depth of a photon ( τ γγ )with energy ǫ = E /m e c to the γγ interaction is 1. Theluminosity distance is denoted as d L (=3.1 Gpc), σ T isthe Thompson scattering cross section, E is the high-est photon energy detected during the flare, t var is thevariability time, and f x is the X-ray flux in 0.3-10 keV.Here, the values of E , f x , and t var are estimated aroundthe same time period and the values are found to be27 GeV at MJD 58228.45, 3.63 × − erg cm − s − atMJD 58228.38, and 1.14 days at MJD 58228.50 respec-tively. The minimum Doppler factor is found to be δ min = 10.7. The location of the gamma-ray emission regioncan be estimated by assuming the bulk Lorentz factor Γ= δ min = 10.7, then the location can be defined as d ∼ c Γ t var /(1+ z ). It is found that the gamma-ray emissionregion is located at distance of 4.40 × cm from the central SMBH down the jet. This value is comparable tothe size of the BLR and hence conculded that during theemission of high energy photon (27 GeV) the gamma-ray emitting region must have been located at the outerboundary of the BLR.I have also considered the contribution of the ac-cretion disk photons in the EC emission. The pho-ton energy density in the comoving frame is defined as(Dermer & Menon 2009), U ′ disk = 0 . R g l Edd L Edd πcz Γ (7)where, R g is known as gravitational radius, and l Edd =L disk /L Edd is the Eddington ratio. Here z representsthe distance of blob from the SMBH and which is esti-mated to be 4.40 × cm. For black hole mass M BH =2.51 × M ⊙ (Wu et al. 2018), the gravitaional radiusfound to be R g = 3.72 × cm. In this study, I havenot considered dusty torus as a external target photonfield since there is no observational evidence. However,the contribution of NIR/optical/UV photons emitted bydisk and dusty torus based clouds irradiated by a spine-sheath jet could be important (Finke 2016; Gaur et al.2018; Breiding et al. 2018) for the EC emission in somecases.There are numbers of parameters that GAMERA usesas a input to model the multiwavelength emissions. Thespectral index ( α , β ), minimum and maximum ( γ min , γ max ) energies, magnetic field inside the blob (B), andluminosity in injected electrons are the parameters thathave been optimized to obtain the best model fit tothe SEDs. The BLR photon density ( U ′ BLR ), BLRtemperature = 10 K (Peterson 2006), accretion diskphoton density ( U ′ disk ), disk temperature = 2.6 × K(Dermer & Menon 2009) are kept fixed while modelingthe SEDs.The size of the blob can also be estimated by using thevariability time and the minimum Doppler factor by therelation, R ≤ ct var δ (1 + z ) − (8)where t var is the observed flux variability time (1.14days). The size of the emitting blob is estimated to beR = 2.1 × cm. However, during the SED modelingDoppler factor and the size of the blob are optimized tobest fit value.A successful SED modeling is performed for all thethree flares and one quiescent state. The best fit modelparameters are shown in Table 5. The multiwavelengthSEDs modeled with GAMERA are presented in Figure10. The low energy synchrotron peak is constrained bythe optical and UV emission and the high energy Comp-ton peak by gamma-ray data points. The X-ray emissionobserved by Swift-XRT constrains the SSC peak. Moremagnetic field value is needed to explain the optical/UVemission in quiescent state, which suggest that the syn-chrotron process is more dominant here compared to theflaring state. The maximum electron energy found dur-ing all the states are in very much agreement though it islittle higher during the flaring state. The total jet poweris estimated by the following relation,P jet = π Γ r c (U e + U B + U p ) (9)1where Γ is the Lorentz factor, r is the size of the blob.The energy density in electrons, magnetic field, and coldprotons are represented by U e , U B , and U p . Here the jetcomposition consists of an equal number of non-thermalelectrons and cold protons. The calculated jet powersin all components are shown in Table 5. The jet poweris calculated for Γ=15.5 (Jorstad et al. 2005) and r =4.64 × cm. It is found that the magnetic field hasmore jet power during the quiescent state compared tothe flares. However, the jet power required in electronsand protons is much higher in case of flaring state com-pared to the quiescent state.The previous study on flares of 3C 279 suggest that mostof the time the emission region is located outside the BLR(Dermer et al. 2014; Yan et al. 2015; Vittorini et al.2017). However, the brightest flare of 3C 279 during2015 June reported by (Paliya 2015a), demands a com-pact emission region with high photon density and suchregions can not be far from the central source. All ear-lier studies on this flare (June 2015) (Hayashida et al.2015; Ackermann et al. 2016; Pittori et al. 2018) foundthe emission region to be located within or at the bound-ary of the BLR. Here, in this study too, similar result isobtained. A compact emission region can be inferred,which is located within the boundary of the BLR.January 2018 flare of 3C 279 was also studied byShah et al. (2019), where they have chosen BLR andIR photons as the target photons to describe the multi-wavelength SEDs. They concluded that both cases canexplain the broadband SEDs of various flaring states.However, the parameters found for EC/IR is more ac-ceptable than EC/BLR process. Two-zone leptonic anda single-zone lepto-hadronic SED modeling is performedby Paliya et al. (2015c) to describe the broadband SEDsof 3C 279 during the flare of 2013 December. However,my result shows that the single emission zone is enoughto explain the broadband SEDs.3C 279 is studied by Hayashida et al. (2012) using thebroadband data from radio to gamma-ray for the first twoyear (2008-2010) of Fermi operation. They found a lag of10 days between optical and gamma-ray emission duringthe flaring episode. However, they also argue that the X-ray emission does not correlate with optical or gamma-ray emission. On the other hand my study exhibit agood correlation among optical, X-ray and gamma-rayemission. They also have plotted the optical polarization(degree and angle) along with the gamma-ray flare andfound huge swing in the polarization angle correspond-ing to the gamma-ray flare (Period D). During the SEDmodeling, they constrained the location of the emissionregion based on the observed change in the optical polar-ization angle. A huge swing in the polarization angle canbe interpreted as the precession in the jet, which allowthe arbitrary location of the emission region within theBLR. My study also shows a huge variation in the opti-cal polarization angle during the flaring episode (Flare Aand Flare B). Considering the polarization angle swingis caused by the precession of jet, I have chosen the sin-gle emission zone within the boundary of the BLR, tomodel the broadband SED, which is also supported bythe value estimated from the variability time.Comparing the SEDs corresponding to quiescent andflaring state, I observed that the flux has increased during flares across the entire electromagnetic spectrum (Figure10). The major flux change is observed in gamma-rayband between quiescent and flaring state. However, thechange in the gamma-ray flux among the different flares(A, B, & C) are small. A relatively lesser flux change isobserved in optical/UV and X-ray bands among all thestates. I have varied a few parameters to explain thesechanges and the parameters are magnetic field, minimumand maximum energy of electrons. I found that almostten times more jet power in electron is required to explainthe gamma-ray flux observed during flares compared tothe quiescent state. SUMMARY
Fermi -LAT data was collected between 2017 Novem-ber to 2018 July and three bright flares were observed.To differentiate among each other the flares have beennamed as “Flare A”, “Flare B”, and “Flare C”. A longlow flux states period was observed just before the “FlareA”. This low flux states is defined as quiescent state andtime period of 50 days is chosen to represent it. A si-multaneous observation in X-ray and Optical/UV wasalso done by Swift-XRT/UVOT telescope. The archivaldata from Steward observatory for optical V and R bandhas also been collected for the whole period, and a hugevariation is seen in the degree of polarization and polar-ization angle. The archival radio data from OVRO andSMA are also collected for the entire flaring period.The day scale variability has been seen in one-day binlight curve, which constrains the size of the emission re-gion to 2.1 × cm. The emission region is located ata distance of 4.40 × cm, which lies at the boundaryof the BLR (R BLR =1.414 × cm). A “harder-when-brighter” trend were also observed in both gamma-rayand X-ray, which predicts detection of high energy pho-ton during the high state.The gamma-ray spectral data points are fitted withthree different spectral models PL, LP, and BPL. TheTS curve obtained for all the fits suggest that LP is thebest model to describe the gamma-ray photon spec-trum. Further, I have estimated the fractional variabilityamong different wavebands, and it is observed that thefractional variability is increasing towards the higher en-ergy.The discrete cross-correlation (DCF) study has beenperformed to find out the possible correlation in variouswavebands emission. The results show a good and strongcorrelation in all possible combinations like: gamma-rayvs. Optical/UV, gamma-ray vs. X-rays. In author’sknowledge this is the first time that a strong correlationwith zero time lag between gamma-ray, X-rays, and op-tical/UV are seen for 3C 279. A strong correlation andzero time lag between optical/UV, X-rays, and gamma-ray emissions suggest their co-spatial origin. This out-come provides an impetus to choose single emission zoneto explain the multiwavelength emission in SED mod-eling. GAMERA is used to model the multiwavelengthemission, and the parameters like magnetic field, injectedelectron spectrum, minimum and maximum energy of in-jected electrons have been optimized to get a good fit tothe SEDs data points. So this study suggests that asingle-zone model can also be good enough to explainthe multi-waveband emissions from one of the brightest
Fermi blazar called 3C 279.2 ν (Hz) -14 -13 -12 -11 -10 -9 -8 ν F ν ( e r g c m − s − ) Sync SSC ECBLR DiskTotal period = 40 daysFlare-A ν (Hz) -14 -13 -12 -11 -10 -9 -8 ν F ν ( e r g c m − s − ) Sync SSC ECBLR DiskTotal period = 35 daysFlare-B ν (Hz) -14 -13 -12 -11 -10 -9 -8 ν F ν ( e r g c m − s − ) Sync SSC ECBLR DiskTotal period = 30 daysFlare-C ν (Hz) -14 -13 -12 -11 -10 -9 ν F ν ( e r g c m − s − ) Sync SSC ECBLR DiskTotal period = 50 daysQuiescent State
Figure 10.
Multi-wavelength SEDs of all the flares and one quiescent state are presented. The optimize model parameters are shown inTable 5. ACKNOWLEDGEMENTS
This work has made use of public
Fermi data obtained fromFSSC. This research has also made use of XRT data analysissoftware (XRTDAS) developed by ASI science data center,Italy. Archival data from the Steward observatory is used inthis research. This research has made use of radio data fromOVRO 40-m monitoring programme (Richards et al. 2011)which is supported in part by NASA grants NNX08AW31G,NNX11A043G, and NNX14AQ89G and NSF grants AST-0808050 and AST-1109911. The archival data from Submil-limeter Array observatory has also been used in this study(Gurwell et al. 2007). The Submillimeter Array is a jointproject between the Smithsonian Astrophysical Observatoryand the Academia Sinica Institute of Astronomy and Astro-physics and is funded by the Smithsonian Institution and theAcedemia Sinica. RP thanks Saikat, Manami, and Gunjanfor manuscript reading and Arkadipta for helpful discussionsregarding DCF.
REFERENCES
Abdo, A. A., et al. 2010, ApJ, 721, 1425Abdo, A. A., Ackermann, M., Ajello, M., et al. 2011, ApJ, 733,L26Atwood, W. B., Abdo, A. A., et al., 2009, ApJ, 697, 1071Ackermann, M., Asano, K., Atwood, W. B., et al. 2010, ApJ, 716,1178Ackermann, M., Anantua, R., Asano, K., et al. 2016, ApJL, 824,L20Aharonian, F., Akhperjanian, A. G., et al. 2007, ApJ, 664, L71Alexander, T. 1997, in Astrophysics and Space Science Library,Vol. 218, Astronomical Time Series, ed. D. Maoz, A. Sternberg,& E. M. Leibowitz, 163 Alexander, T. 2013, arXiv:1302.1508Blumenthal, R., & Gould, G., 1970, RvMP, 42, 237Bonning, E. W., Bailyn, C., et al. 2009, ApJ, 697, L81Breeveld, A. A., Landsman, W., et al., 2011, arxiv, 1358, 373-376Breiding, P., Georganpoulos, M., & Meyer, E., 2018, ApJ, 853, 19R. J., Britto et al. 2016, ApJChatterjee R., Jorstad, S. G., Marscher, A. P., et al., 2008, ApJ,689, 79Cohen, D. P., Romani, R. W., et al. 2014, ApJ, 797, 137Dondi, L., & Ghisellini, G. 1995, MNRAS, 273, 583Dermer, C. D., Schlickeiser, R., & Mastichiadis, A. 1992, A&A,256, L27Dermer, C. D., Menon, G., 2009, High Energy Radiation fromBlack Holes (Princeton, NJ: Princeton Univ. Press)Dermer C. D., Cerruti M., Lott B., Boisson C., Zech A., 2014,ApJ, 782, 82Edelson, R. A., & Malkan, M. A., 1987, ApJ, 323, 516Edelson, R. A., & Krolik, 1988, J. H., ApJ, 333, 646Edelson, R. A., Krolik, J. H., & Pike, G, F., 1990, ApJ, 359, 86Emmanoulopoulos, D., McHardy, I. M., & Papadakis, I. E., 2013,MNRAS, 433, 907-927Finke, J., 2016, ApJ, 830, 94Fuhrmann, L., Larsson, S., Chiang, J., et al. 2014, MNRAS, 441,1899Gaur, H., Mohan, P., Wiercholska, A., & Gu, M., 2018, MNRAS,473, 3638Ghisellini, G., & Tavecchio, F., 2009, MNRAS, 397,985Gu, M., Cao, X., & Jiang, D. R. 2001, MNRAS, 327, 1111Gurwell, M. A., Peck, A. B., Hostler, S. R., Darrah, M. R., &Katz, C. A. 2007, From Z-Machines to ALMA: (Sub)MillimeterSpectroscopy of Galaxies, 375, 234Hahn, J., 2015, Proc. ICRC, 34, 917Hayashida, M., et al. 2012, ApJ, 754, 114Hayashida M., Nalewajko, K., Madejski, G. M., et al. 2015, ApJ,807, 79 Table 5
Results of fitted Multiwavelength SEDs showed in Figure 10. A LogParabola model is used as electron injected spectrum which is definedas dN/dE = N (E/E ) ( − α − β ∗ log ( E/E )) , where E is chosen as 90 MeV.Activity Parameters Symbol Values Activity period (days)BLR temperature T ′ blr × KBLR photon density U ′ blr Disk temperature T ′ disk × KDisk photon density U ′ disk × − erg/cm Size of the emission region R 4.64 × cmDoppler factor of emission region δ γ min γ max × Spectral index of injected electron spectrum (LP) α β e × erg/secjet power in magnetic field P B × erg/secjet power in cold protons P p × erg/secFlare B Min Lorentz factor of injected electrons γ min γ max × Spectral index of injected electron spectrum (LP) α β P e × erg/secjet power in magnetic field P B × erg/secjet power in cold protons P p × erg/secFlare C Min Lorentz factor of injected electrons γ min γ max × Spectral index of injected electron spectrum (LP) α β P e × erg/secjet power in magnetic field P B × erg/secjet power in cold protons P p × erg/secQuiescent state Min Lorentz factor of injected electrons γ min γ max × Spectral index of injected electron spectrum (LP) α β P e × erg/secjet power in magnetic field P B × erg/secjet power in cold protons P p ×45