TY - JOUR
AU1 - Sahakyan,, N.
AU2 - Gasparyan,, S.
AB - Abstract We present the γ-ray observations of the flat-spectrum radio quasar PKS 1441+25 (z = 0.939), using the Fermi large Area Telescope data accumulated during 2015 January – December. A γ-ray flare was observed in January 24, when the flux increased up to (2.22 ± 0.38) × 10−6 photon cm−2 s−1 with the flux-doubling time-scale being as short as ∼1.44 d. The spectral analysis shows that from 2015 April 13 to April 28, the MeV-to-GeV photon index has hardened and changes in the range of Γ = (1.73 − 1.79) for most of the time. The hardest photon index of Γ = 1.54 ± 0.16 has been observed on MJD 57 131.46 with 11.8σ which is not common for flat-spectrum radio quasars. For the same period the γ-ray spectrum shows a possible deviation from a simple power-law shape, indicating a spectral cutoff at Ecut = 17.7 ± 8.9 GeV. The spectral energy distributions during quiescent and flaring states are modelled using one-zone leptonic models that include the synchrotron, synchrotron self Compton and external inverse Compton processes; the model parameters are estimated using the Markov Chain Monte Carlo method. The emission in the flaring states can be modelled assuming that either the bulk Lorentz factor or the magnetic field has increased. The modelling shows that there is a hint of hardening of the low-energy index (∼1.98) of the underlying non-thermal distribution of electrons responsible for the emission in 2015 April. Such hardening agrees with the γ-ray data, which pointed out a significant γ-ray photon index hardening on 2015 April 13 to 28. radiation mechanisms: non-thermal, galaxies: active, gamma-rays: galaxies, quasars: individual: PKS1441+25 1 INTRODUCTION Recent observations in the γ-ray band (≥100 MeV) show that the extragalactic γ-ray sky is dominated by emission from blazars – an extreme class of active galactic nuclei (AGNs) which have jets that are forming a small angle with respect to the line of sight (Urry & Padovani 1995). Blazars are known to emit electromagnetic radiation in almost all frequencies that are currently being observed, ranging from radio to very high energy (VHE; > 100 GeV) γ-ray bands. The broad-band spectrum is mainly dominated by non-thermal emissions produced in a relativistic jet pointing towards the observer. Due to small inclination angle and large bulk motion, the emission from blazars is affected by relativistic beaming which has enormous effects on the observed luminosities. Indeed, the observed luminosity (Lobs) is related to the emitted luminosity (⁠|$L_{\rm em}^\prime$|⁠) as |$L_{\rm obs}=\delta ^{3+\alpha }\:L_{\rm em}^\prime$|⁠. If so, the observed luminosity can be thus amplified by a factor of thousands or even more (usually δ ≥ 10). Such amplification makes it possible to detect emission even from very distant blazars. A key feature of the non-thermal emission from blazars is the distinct variability at all frequencies (with different variability time-scales – from years down to a few minutes). The shortest variability time-scales are usually observed for the highest energy band; an example is the minute scale variability of PKS 2155-304 (Aharonian et al. 2007) and IC 310 (Aleksic et al. 2014) which implies that the emission is produced in a very compact region. Therefore, by observing blazars one gets a unique chance to investigate the jet structure on sub-parsec scales. By their emission line features blazars are commonly grouped as BL Lacertae objects (BL-Lacs) and Flat-Spectrum Radio Quasars (FSRQs; Urry & Padovani 1995). BL Lacs have weak or no emission lines, while FSRQs have stronger emission lines. The difference in the emission-line properties of FSRQs and BL Lacs may be connected with that in the properties of accretion in these objects (Ghisellini, Maraschi & Tavecchio 2009). The multiwavelength observations of blazars have shown that their spectral energy distribution (SED) has two broad non-thermal peaks – one at the IR/optical/UV/X-ray and the other at the higher-energy (HE; > 100 MeV) γ-ray band. The low-energy peak is believed to be due to the non-thermal synchrotron emission of relativistic electrons while the origin of the second component is still debated. One of the most widely accepted theories for the second peak is that it is produced from Inverse Compton (IC) scattering of low energy synchrotron photons (Synchrotron Self Compton; SSC) (Ghisellini, Maraschi & Treves 1985; Maraschi, Ghisellini & Celotti 1992; Bloom & Marscher 1996) which often successfully explains the emission from BL-Lacs (Finke, Dermer & Böttcher 2008). Besides, the photons from the regions outside the jet may serve as seed photons for IC scattering – External Compton (EC) models which are used to model the emission from FSRQs. The external photon field can be dominated either by the photons reflected by broad line region (BLR; Sikora, Begelman & Rees 1994) or by photons from a dusty torus (Błażejowski et al. 2000; Ghisellini & Tavecchio 2009). Domination of one of the components mostly depends on the localization of the emitting region; for example, if the energy dissipation occurs within BLR then the observed HE emission is mostly due to IC scattering of BLR reflected photons, otherwise, if the emitting region is far from the central source, then the IC scattering of torus photons will dominate. SSC and EC models assume that the emission is produced by the same population of electrons, though up to now it is not clear whether it is produced in the same part of the jet or by different electron populations. Alternatively, the HE emission can be explained by the interaction of energetic protons; e.g. a significant fraction of the jet power goes for acceleration of protons so that they reach the threshold for pion production (Mücke & Protheroe 2001; Mücke et al. 2003). The majority of the blazars detected in VHE γ-ray band are high-frequency-peaked BL Lacs for which the synchrotron bump is in the UV/X-ray bands. In addition to BL Lacs, there are also 5 FSRQs detected in the VHE γ-ray band which is rather surprising, since the BLR structure of these objects, which is rich in optical-UV photons, makes these environments strongly opaque to VHE γ-rays (Liu & Bai 2006; Poutanen & Stern 2010). Moreover, FSRQs have a relatively steep photon index in the energy range of >100 MeV as was observed with the Fermi Large Area Telescope (Fermi LAT) which does not make them as strong emitters of VHE γ-ray photons. Detection of FSRQs in the VHE γ-ray band is challenging for the near-black hole dissipation scenarios; it assumes that the γ-rays are most likely produced farther from the central source, outside the BLR, where the dominant photon field is the IR emission from the dusty torus. Typically, the temperature of torus photons ∼ 103 K is lower than that of the photons reflected in the BLR ∼ 105 K, and, in principle, VHE photons with energy up to ∼1 TeV can escape from the region. Thus, the observations of FSRQs in VHE γ-ray band provide an alternative view of blazar emission as compared to BL Lacs. Moreover, since FSRQs are more luminous than BL Lacs, they could, in principle, be observed at greater distances. Indeed, the farthest sources detected in the VHE γ-ray band are the FSRQs at a redshift of z ≥ 0.9 [e.g. PKS 1441+25 (Abeysekara et al. 2015; Ahnen et al. 2015) and S3 0218+35 (Ahnen et al. 2016)]. That is why FSRQs are ideal for estimation of the intensity of extragalactic background light (EBL) through the absorption of VHE photons when they interact with the EBL photons (Coppi & Aharonian 1999; Madau & Phinney 1996). Amongst FSRQs, PKS 1441+25 is one of the most distant sources detected so far at z = 0.939 (Shaw et al. 2012). In April 2015 both VERITAS and MAGIC collaborations announced the detection of VHE γ-rays from PKS 1441+25 (with up to 250 GeV photons; Mirzoyan 2015; Mukherjee 2015). A strong emission from the source had been detected from 2015 April 20 to 27. During the same period, the source had been also observed with the telescopes Swift and NuSTAR. The origin of the multiwavelength emission from PKS 1441+25 observed in April is modelled assuming the emission region is beyond the BLR, and the emission in the VHE γ-ray band is mostly due to the IC scattering of the dusty torus photons (Abeysekara et al. 2015; Ahnen et al. 2015). Moreover, the large distance to PKS 1441+25 allowed to indirectly probe the EBL absorption at redshifts up to z ∼ 1 with the help of ground-based γ-ray instruments. In the theoretical interpretation of the multiwavelength emission from blazars, the size/location of the emitting region, magnetic field and electron energy distribution are uncertain. Only during flaring periods some of the unknown parameters can be constrained based on the observations in different bands. The observations of PKS 1441+25 during the bright period in 2015 April by different instruments provide us with data on the maximums of the emitting components (Swift UVOT/ASAS-SN and Fermi LAT) as well as on the transition region between these components in the energy range from 0.3 to 30 keV (Swift XRT and NuSTAR; Abeysekara et al. 2015). Similar data (up to HE γ-ray band) are available also from the observations carried out on 2015 January 06 to 28, which is the period of the large flare that was observed with Fermi LAT. Thus, by modelling the emission in these two periods and estimating the parameter space that describes the underlying particle distribution responsible for the emission through the Markov chain Monte Carlo (MCMC) technique, one can investigate and explore particle acceleration/emission processes and jet properties in these two significant flaring periods which are crucial for understanding the origin of the flares. This motivated us to have a new look at the origin of the multiwavelength emission from PKS 1441+25, using currently available data from Swift, NuSTAR and Fermi LAT. This paper is structured as follows. The results of the spectral and temporal analysis of the Fermi LAT data are presented in Section 2. The broad-band SED modelling with MCMC technique is presented in Section 3 and discussion and conclusions are presented in Section 4. 2 FERMI LAT DATA ANALYSIS The large Area Telescope on board the Fermi satellite is a pair-conversion telescope sensitive to γ-rays in the energy range from 20 MeV to 300 GeV. It constantly scans the whole sky every 3 h already for more than eight years. More details about Fermi LAT can be found in Atwood et al. (2009). In this paper, for spectral analysis we use the publicly available data acquired in the periods from 2015 January 06 to 28 and from 2015 April 15 to 26. These two periods have been picked, because they are contemporaneous with the Swift XRT observations of the source (Abeysekara et al. 2015). The data were analysed with the standard Fermi Science Tools v10r0p5 software package released on 2015 May 18 available from the Fermi Science Support Center.1 The latest reprocessed PASS eight events and spacecraft data are used with the instrument response function P8R2_ SOURCE_ V6. We have downloaded photons in the energy range from 100 MeV to 100 GeV from a region of interest defined as a circle of a 20° radius centred at the γ-ray position of PKS 1441+25 (RA, Dec.) = (220.996, 25.039) (Acero et al. 2015). Only the events with higher probability of being photons (evclass=128, evtype=3) have been considered in the analysis. A cut on the zenith angle of 90° is applied to reduce contamination from the Earth-limb γ-rays produced by cosmic rays at their interaction with the upper atmosphere. The model file, describing the region of interest, contains point sources from the Fermi LAT third source catalogue (Acero et al. 2015, 3FGL) within 25° from the target, as well as contains Galactic gll_ iem _ v05_ rev1 and isotropic iso_source_v05 diffuse components. All point-source spectra were modelled with those given in the catalogue, allowing the photon index and normalization of the sources within 20° to be free in the analysis. Also, the normalization of diffuse background components was not fixed. 2.1 Spectral analysis In order to find the best matches between spectral models and events, an unbinned likelihood analysis is performed with gtlike. The PKS 1441+25 spectrum has been initially modelled as a power-law function where the normalization and the power-law index are taken as free parameters. The best-fitting parameters obtained with gtlike analysis are presented in Table 1 and the corresponding spectrum is shown in Fig. 1 (blue and red data for January and April, respectively). The spectrum is calculated by separately running gtlike for five energy bands equal on a log scale. Figure 1. Open in new tabDownload slide The γ-ray spectrum of PKS 1441+25 above 100 MeV averaged over the Fermi LAT observations in January (blue) and April (red). Figure 1. Open in new tabDownload slide The γ-ray spectrum of PKS 1441+25 above 100 MeV averaged over the Fermi LAT observations in January (blue) and April (red). Table 1. The best parameters obtained with gtlike for power-law modelling. For each time period, photon flux in the range 0.1 − 100 GeV, photon index and detection significance are presented. Parameter name . Blue . Red . Flux (photon cm−2 s−1) (5.89 ± 0.30) × 10−7 (3.63 ± 0.36) × 10−7 α 1.99 ± 0.04 1.74 ± 0.06 TS 2174 910 Parameter name . Blue . Red . Flux (photon cm−2 s−1) (5.89 ± 0.30) × 10−7 (3.63 ± 0.36) × 10−7 α 1.99 ± 0.04 1.74 ± 0.06 TS 2174 910 Open in new tab Table 1. The best parameters obtained with gtlike for power-law modelling. For each time period, photon flux in the range 0.1 − 100 GeV, photon index and detection significance are presented. Parameter name . Blue . Red . Flux (photon cm−2 s−1) (5.89 ± 0.30) × 10−7 (3.63 ± 0.36) × 10−7 α 1.99 ± 0.04 1.74 ± 0.06 TS 2174 910 Parameter name . Blue . Red . Flux (photon cm−2 s−1) (5.89 ± 0.30) × 10−7 (3.63 ± 0.36) × 10−7 α 1.99 ± 0.04 1.74 ± 0.06 TS 2174 910 Open in new tab The fluxes presented in Table 1 significantly exceed the averaged flux given in 3FGL (≈1.28 × 10−8 photon cm−2 s−1) (Acero et al. 2015). The photon index estimated in 2015 January is consistent with the value reported in 3FGL α = 2.13 (averaged over four years of observations); however, a relative hardening of α = 1.74 ± 0.06 is observed in April, which is rarely observed for FSRQs. Moreover, we note an indication of deviation of the power-law model with respect to the data above 10s of GeV energies observed in April (red bowtie plot in Fig. 1). In order to check for a statistically significant curvature in the spectrum, an alternative fit of the power law with an exponential cut-off function in the form of |${\rm d}N/{\rm d}E\sim E_{\gamma }^{-\alpha }\:\times Exp(-E_{\gamma }/E_{{\rm cut}})$| is done, which results in α = 1.56 ± 0.1 and Ecut = 17.7 ± 8.9 GeV (black bowtie plot in Fig. 1). The power-law and cut-off models are compared with a log likelihood ratio test: the TS is twice the difference in the log likelihoods, which gives eight for this case. Note that the TS probability distribution can be approximated by a χ2 distribution with 1 degree of freedom (dof) corresponding to the difference of the dof between the two functions. The results give P(χ2) = 0.0046, which again indicates a deviation from a simple power-law function. The best-fitting cut-off power-law function is shown as a black bowtie line in Fig. 1. However, 2.8 σ is not a high enough significance to claim for a statistically significant curvature although it is as high as 3.86 σ if the data collected during the whole month of April are considered. 2.2 Temporal analysis In order to investigate the size of the γ-ray emitting region, light curves with different time binning are generated. A characteristic time-scale for flux variation τ would limit the (intrinsic) size of the emission region to R ≤ c × δ × τ/(z + 1). Thus, it is crucial to do a variability analysis in order to distinguish between different emission processes. The light curve of PKS 1441+25 for the period from 2015 January to December has been calculated by the gtlike tool, applying the unbinned likelihood analysis method. (0.1 − 100) GeV photons from a region with a 10° radius centred on the position of PKS 1441+25 are used in the analysis with the appropriate quality cuts applied as in the previous case. During the analysis, in order to reduce the uncertainty in the flux estimations, in the model file the photon indices of all background sources are fixed to the best guess values. Two different sets of light curves are calculated, considering the power-law index of PKS 1441+25 as being fixed and then as free. Since no variability is expected for the background diffuse emission, the normalization of both background components is also fixed to the values obtained for the whole time period. The γ-ray light curve of PKS 1441+25 obtained with one-day and three-days binning is presented in Fig. 2 (a) (blue and green data, respectively). In the light curve there can be identified several periods when the flux was in high as well as in quiescent states. A major increase of the γ-ray flux had been detected in the period from 2015 January 21 to 28, with a daily averaged maximum of (1.55 ± 0.18) × 10−6 photon cm−2 s−1 observed on 2015 January 25. Unfortunately, the peak flare of January 25 was not observed by Swift. The γ-ray photon index evolution in time in a three-day long binning is shown in Fig. 2 (c) with green data (three-day long binning is used since the photon index uncertainties are less than in one-day binning). During the flaring period the photon index is 1.9 − 2.0. Also an increase in the flux can be noticed around January 22 which lasted just one day. In order to check if this brightening is statistically significant, light curves with denser time sampling (half a day and 4 h) are generated. However, the corresponding flux increase is within the uncertainty of the surrounding bins, while the peak of the flux around January 25 is present in both light curves. In addition, a substantial increase in the γ-ray flux was observed in April, from June to about July 15, mid August and around October–November; but the maximum flux intensity was lower as compared with that observed during the strong γ-ray outburst of January 21 to 28 (Fig. 2). The active state in April is the period when PKS 1441+25 was observed by MAGIC on MJD 57 130–57 139 and VERITAS on MJD 57 133–57 140 (Abeysekara et al. 2015; Ahnen et al. 2015). The γ-ray light curve with three-day binning shows that, between MJD 57 125.56–57 140.64 (from 2015 April 13 to April 28), the γ-ray photon index is significantly harder, Γ = (1.73 − 1.79). It implies that during the observations in the VHE band the source was in a state characterized by a hard γ-ray photon index in the MeV-to-GeV range. Figure 2. Open in new tabDownload slide The γ-ray light curve of PKS 1441+25 from 2015 January to December (a). The bin intervals correspond to 1 day (blue data) and 3 days (green data). The light curve obtained by adaptive binning method assuming 20 per cent of uncertainty is presented in red (b). The change of photon index for 3-day binning (green) and with adaptive binning method are shown in (c). Figure 2. Open in new tabDownload slide The γ-ray light curve of PKS 1441+25 from 2015 January to December (a). The bin intervals correspond to 1 day (blue data) and 3 days (green data). The light curve obtained by adaptive binning method assuming 20 per cent of uncertainty is presented in red (b). The change of photon index for 3-day binning (green) and with adaptive binning method are shown in (c). Next, in order to investigate the flux changes in time, and in particular in the flaring periods, the light curves have been generated by an adaptive binning method. In this method, the time bin widths are flexible and chosen to produce bins with constant flux uncertainty (Lott et al. 2012). This method allows detailed investigation of the flaring periods, since at times of a high source flux, the time bins are narrower than during lower flux levels therefore the rapid changes of the flux can be found. In order to reach the necessary relative flux uncertainty, the integral fluxes are computed above the optimal energies (Lott et al. 2012) which correspond to E0 = 215.4 MeV in this case. Adaptively binned light curves in the 215 MeV–300 GeV energy range with 20 per cent and 15 per cent uncertainties have been generated. Flare is present in both light curves. The light curve with 20 per cent flux uncertainty at each bin is presented in Fig. 2(b) with red colour. It confirms all the features visible in the constant-bin-width light curve, but also allows us to investigate fast variability during high-flux states in greater detail. The first flare episode occurred during MJD 57 043.30–57 049.38, when the time width was less than ∼15 h. A strong flaring period is observed around January 24 and 25. The flux peak of (1.14 ± 0.24) × 10−6 photon cm−2 s−1 was observed on January 24 at 22:35 pm in a bin with a half-width of 3.1 h. The analysis of the data acquired in the mentioned period on the energies of >100 MeV results in a flux of (2.22 ± 0.38) × 10−6 photon cm−2 s−1, which is the highest photon flux detected from this source. The data analysis for the entire flaring period (January 21–28) resulted in a flux of (1.05 ± 0.06) × 10−6 photon cm−2 s−1 and a photon index of ∼1.98 ± 0.04 (Sahakyan & Gasparyan 2017). After MJD 57 049.38, PKS 1441+25 was in its quiescent state, and the data should be accumulated for more than a day to reach 20 per cent uncertainty. Then from MJD 57 109.89 to MJD 57 143.91, PKS 1441+25 was again in its active state which was characterized by emission with a significantly hardened γ-ray photon index. Starting from MJD 57 126.70 to MJD 57 141.93, the photon index of PKS 1441+25 hardened and reached ≤1.9 most of the time. Measured within a few hours, the photon index kept varying from Γ = 1.73 to Γ = 1.91. The hardest photon index of Γ = 1.54 ± 0.16 was observed on MJD 57 131.46 with 11.8σ and the data being accumulated for ≈29 h. Other periods, when PKS 1441+25 was bright enough to be detected on sub-day scales, are MJD 57 177.38–57 199.76 and MJD 57 243.02–57 249.39. For the rest of the time the source was in its quiescent state and the data should be accumulated for a few days or even longer in order to detect the source. The analysis of the light curve with the new adaptive binning method for the first time allowed us to investigate the flaring activity of PKS 1441+25 with a sub-day resolution and to perform detailed investigation of the flux and photon index changes. Furthermore, to derive the flare doubling time-scales and understand the nature of the January flare, the light curve is fitted with an exponential function in the form of (Abdo et al. 2010b) \begin{eqnarray} F(t)= F_{\rm {\rm c}}+F_0\times \left({\rm e}^{\frac{t-t_{\rm 0}}{t_{\rm r}}}+{\rm e}^{\frac{t_{\rm 0}-t}{t_{\rm d}}}\right)^{-1}, \end{eqnarray} (1) where t0 is the time of the maximum intensity of the flare (F0), Fc is the constant level present in the flare, tr and td are the rise and decay time constants, respectively. The fit shows that the flare is best explained when t0 = 57 048.25 ± 0.18, tr = 1.92 ± 0.3, td = 0.72 ± 0.1 and F0 = (22.6 ± 1.4) × 10−7 photon cm−2 s−1. The fit of the flaring period is shown in Fig. 3 with a dashed red line. Using this technique, it is also possible to estimate the shortest time variability (flux doubling) defined by τ = 2tr,d, corresponding to τ = 1.44 d which is used to put an important constraint on the radiative region size. We note that the previous PKS 1441+25 γ-ray emission studies with the Fermi LAT data that covered only the period in April did not allow to properly estimate the γ-ray emitting region size, while here the analysis of the flaring period in January allowed to constrain the flare doubling time which is necessary for constraining the γ-ray emission region size. Figure 3. Open in new tabDownload slide The light curve's sub interval that covers a major flaring period. The red dashed line shows the flare fit with equation (1). Figure 3. Open in new tabDownload slide The light curve's sub interval that covers a major flaring period. The red dashed line shows the flare fit with equation (1). 3 BROAD-BAND SED MODELLING It is hard to make theoretical modelling of the observed broad-band SED because the structure of the central region of blazars is complex and the exact localization of emitting regions is unknown. The observed fast variability indicates compactness of the emitting region but its localization remains an open problem. Along the jet, the emission can be produced in different zones and depending on the distance from the central black hole different components can contribute to the observed emission (Sikora et al. 2009). 3.1 Broad-band SED The broad-band SEDs of PKS 1441+25 for different periods are shown in Fig. 4 where with red and blue colours are the SED observed in January and April, respectively, while the archival data from ASI science data centre2 are shown with grey colour. We note that during the high states, the second emission peak increased by intensity and shifted to HEs. This kind of change has already been observed during the flaring state of 4C+49.22 (Cutini et al. 2014) and PKS 1510-089 (Abdo et al. 2010a). During the flaring periods, the low-energy component's intensity increased as compared with the quiescent state; the increase in April exceeded that one observed in January (although the power-law photon index in the X-ray band (≈2.3) had been relatively constant during both observations). More evident and drastic is the change of the peak intensity of the low-energy component; from January to April it increased by nearly an order of magnitude and as compared with the quiescent state it increased ≥15 times. On the contrary, the peak of the second component (in the HE γ-ray band) is relatively constant, only the photon index in the MeV-GeV energy range is harder during the observations in April. The Compton dominance of the source is stronger and evident during the flaring periods which suggests that the density of the external photon fields significantly exceeds the synchrotron photon density (Uext/Usyn ≫ 1). Figure 4. Open in new tabDownload slide The broad-band SED of PKS 1441+25 for January (red), April (blue) and for the quiescent state (grey). The blue, red and grey lines are the models fitting the data with the electron spectrum given by equa-tion (2) for January, April and for the quiescent state, respectively. The model parameters are presented in Table 2. The UV-X-ray and VHE γ-ray data observed in January and April are from Abeysekara et al. (2015) and HE γ-ray data (Fermi LAT) are from this work. Figure 4. Open in new tabDownload slide The broad-band SED of PKS 1441+25 for January (red), April (blue) and for the quiescent state (grey). The blue, red and grey lines are the models fitting the data with the electron spectrum given by equa-tion (2) for January, April and for the quiescent state, respectively. The model parameters are presented in Table 2. The UV-X-ray and VHE γ-ray data observed in January and April are from Abeysekara et al. (2015) and HE γ-ray data (Fermi LAT) are from this work. Such a strong amplification of the emission from blazars can be explained by means of introducing changes in the emission region parameters, e.g. in the magnetic field, emitting region size, bulk Lorentz factor and others, and/or particle energy distribution. In principle, all the parameters describing the emitting region can be changed at the same time if the flares are due to a global change in the physical processes in the jet, which also affect the jet dynamics and properties. However, usually, the change in one or two parameters is enough to explain the flares. An interesting study of the flaring activity in FSRQs as a result of changes in different parameters has been investigated in Paggi et al. (2011). Namely, the emission spectra evolution as a function of changes in different parameters (e.g. bulk Lorentz factor, magnetic field, accretion rate, etc.) is investigated. In the case of PKS 1441+25, during its flaring periods, both the low energy and HE components increased several times. The increase of the second component is most likely due to moving of the emitting region outside its BLR. In principle, there are two possibilities: i) either the emitting region moves faster due to increasing bulk Lorentz factor and leaves the BLR; or ii) the bulk Lorentz factor is unchanged and only the emitting region is moving beyond the BLR. In the first case, since the external photon density in the comoving frame of the jet depends on the Doppler boosting factor, a strong increase in the Compton dominance will be observed. We note that the change of the bulk Lorentz factor will also affect the low-energy component. In the second case, the flaring activity is due to the change of the location of the emitting region and due to the magnetic field amplification. Additional increase of the magnetic field from January to April is also evident when the low-energy component kept increasing [this corresponds to the case shown in Fig. 1(b) in (Paggi et al. 2011)]. Accordingly, we discuss two possibilities. First, we assume that δ has increased from 10 in the quiescent to 18 in the flaring periods and then we assume that it was constant (δ = 18) in both periods. These values are below and above the estimated mean bulk Lorentz factor of FSRQs obtained from the analysis of a large sample of γ-ray emitting FSRQs (Ghisellini & Tavecchio 2015). The emission region size can be estimated through the observed variability time-scale τ = 1.44 d implying that Rb ≤ δ c τ/(1 + z) ≈ 3.5 × 1016 cm when δ = 18 and Rb = 1.92 × 1016 cm, when δ = 10. 3.2 Theoretical modelling We attempt to fit the SEDs in the high states of January and April as well as in the quiescent state. Even if a quiescent state SED is constrained with non-simultaneous data, its modelling provides an insight into the dominant physical processes which are constantly present in the jet but are covered by the flaring components during the high states. We modelled the PKS 1441+25 SED for high and quiescent states in the framework of single-zone leptonic models that include the synchrotron, SSC and EC processes. The emission region (the ‘blob’), assumed to be a sphere with a radius of R which is moving with a bulk Lorentz factor of Γ, carries a magnetic field with an intensity of B and a population of relativistic electrons. The blob velocity makes a small angle with respect to the line of sight, so the emission is amplified by a relativistic Doppler factor of δ. The energy spectrum of the population of electrons in the jet frame, which is responsible for the non-thermal emission is assumed to have a broken power-law shape: \begin{eqnarray} N_{\rm e}^{\prime }(E_{{\rm e}}^\prime )= \left\lbrace \begin{array}{@{}l@{\quad }l@{}}N_0^\prime \left(\frac{E_{{\rm e}}^\prime }{m_{\rm e}c^2}\right)^{-\alpha _1} & E_\mathrm{e,min}^\prime \leqslant E_{{\rm e}}^\prime \leqslant E_\mathrm{br}^\prime \\ N_0^\prime \left(\frac{E_{{\rm br}}^\prime }{m_{\rm e}c^2}\right)^{\alpha _2-\alpha _1} \left(\frac{E_{{\rm e}}^\prime }{m_{\rm e}c^2}\right)^{-\alpha _2} & E_\mathrm{br}^\prime \leqslant E_{{\rm e}}^\prime \leqslant E_{{\rm e},{\rm max}}^\prime \end{array}\right. , \end{eqnarray} (2) where |$N_{0}^\prime$| is connected with the total electron energy |$U_{\rm e}=\int _{\rm E_{min}^\prime }^{\rm E_{\rm max}^\prime }E_{{\rm e}}^\prime N_{\rm e}(E_{{\rm e}}^\prime ){\rm d}E_{\rm e}^\prime$|⁠, α1 and α2 are the low and high indexes of electrons correspondingly below and above the break energy |$E_{\rm br}^\prime$|⁠, and |$E_{\rm min}^\prime$| and |$E_{\rm max}^\prime$| are the minimum and maximum energies of electrons in the jet frame, respectively. The electron spectrum given in equation (2) is naturally formed from the cooling of relativistic electrons (Kardashev 1962; Inoue & Takahara 1996). The low-energy (from radio to optical/X-ray) emission is due to the synchrotron emission of electrons with an energy spectrum as given by equation (2) in a homogeneous and randomly oriented magnetic field. For the quiescent state we assume the energy dissipation occurs close to the central source region and it is explained as an IC scattering of synchrotron photons (SSC). Instead the high state emission is dominated by that from a region well outside the BLR in order to avoid the strong absorption of photons with energies ≥ 100 GeV [similar assumptions have been already made in (Abeysekara et al. 2015; Ahnen et al. 2015)]. In this case the dominant external photon field is the IR radiation from the dusty torus which, as we assume, has a blackbody spectrum with a luminosity of LIR = η Ldisc [η = 0.6, (Ghisellini et al. 2009)] and a temperature of T = 103 K and fills a volume that for simplicity is approximated as a spherical shell with a radius of RIR = 3.54 × 1018 (Ldisc/1045)0.5 cm (Nenkova et al. 2008). The disc luminosity is estimated from the BLR luminosity, Ldisc = 10 × LBLR ≈ 2 × 1045 erg s−1 (Xiong & Zhang 2014). 3.3 Fitting technique In order to constrain the model parameters more efficiently, we employed the MCMC method, which enables to derive the confidence intervals for each model parameter. For the current study we have modified the naima package (Zabalza 2015) which derives the best-fit and uncertainty distributions of spectral model parameters through MCMC sampling of their likelihood distributions. The prior likelihood, our prior knowledge of the probability distribution of a given model parameter and the data likelihood functions are passed on to the emcee sampler function for an affine-invariant MCMC run. In addition, there are multiple simultaneous walkers which improve the efficiency of the sampling and reduce the number of computationally expensive likelihood calls. We run the sampling with 64 simultaneous walkers, for 100 steps of burn-in and 100 steps of run. In the parameter sampling, the following expected ranges are considered: 1.5 ≤ (α1, 2) ≤ 10, |$0.511\:{\rm MeV}\le E^\prime _{({\rm br},\:{\rm min},\:{\rm max})}\le 1\:{\rm TeV}$|⁠, and N0 and B are defined as positive parameters. The synchrotron emission is calculated using the parametrization of the emissivity function of synchrotron radiation in random magnetic fields presented in Aharonian, Kelner & Prosekin (2010) while the IC emission is computed based on the monochromatic differential cross-section of Aharonian & Atoyan (1981). 3.4 SED modelling and results The results of SED modelling are shown in Fig. 4 with the corresponding parameters in Table 2. The radio emission is due to the low-energy electrons which are accumulated for longer periods, that is why, the radio data are treated as an upper limit for the purposes of our modelling. To have an indication of a change in the energetic contents of the jet, as well as of changes in the radiating particle distribution, first we try to fit the SED in a quiescent state which is modelled assuming two different Doppler boosting factors. The grey solid line in Fig. 4 shows the synchrotron/SSC emission assuming that the jet Doppler boosting factor is δ = 10, and the grey dashed line is the case of δ = 18. In case of δ = 10, as the emitting region size is as small as Rb = 1.92 × 1016 cm, the magnetic field should be as strong as B = 0.19 G to account for the observed data, while at δ = 18 the magnetic field is much weaker, B = 0.046 G. Also, the underlying electron distribution for the case of δ = 10 is characterized by a slightly higher break (2.83 GeV versus 1.11 GeV) in order to account for the observed emission. Table 2. Model parameters. . Parameter . Quiescent . Quiescent . January . April . Doppler factor δ 10 18 18 18 Normalization of electron distribution |$N_0^\prime \times 10^{48}\:{\rm eV^{-1}}$| |$10.68_{-2.64}^{+3.09}$| |$43.44_{-7.76}^{+6.59}$| |$23.83^{+8.11}_{-7.32}$| |$6.12^{+1.67}_{-1.56}$| Low-energy electron spectral index α1 2.14 ± 0.04 |$2.09_{-0.04}^{+0.03}$| |$2.10_{-0.05}^{+0.04}$| 1.98 ± 0.03 High-energy electron spectral index α2 |$3.39_{-0.14}^{+0.27}$| 3.38 ± 0.06 3.46 ± 0.06 3.64 ± 0.01 Minimum electron energy |$E^\prime _{\rm min}$| (MeV) |$1.84_{-1.23}^{+1.75}$| |$286.37_{-25.39}^{+30.64}$| |$1.97^{+0.31}_{-0.34}$| |$4.16^{+1.00}_{-1.86}$| Break electron energy |$E^\prime _{\rm br}$| (GeV) |$2.83_{-0.31}^{+0.51}$| |$1.11_{-0.12}^{+0.14}$| |$1.62^{+0.23}_{-0.15}$| |$3.11^{+0.15}_{-0.23}$| Maximum electron energy |$E^\prime _{\rm max}$| (GeV) |$46.27_{-13.76}^{+49.74}$| |$82.32_{-17.14}^{+13.47}$| |$127.82^{+26.74}_{-24.75}$| |$202.79^{+21.2}_{-14.6}$| Magnetic field B [G] 0.19 ± 0.013 0.046 ± 0.002 |$0.11^{+0.005}_{-0.004}$| |$0.18^{+0.009}_{-0.006}$| Jet power in magnetic field LB × 1043 erg s−1 0.49 0.31 1.71 4.51 Jet power in electrons Le × 1045 erg s−1 2.11 4.07 9.60 4.47 . Parameter . Quiescent . Quiescent . January . April . Doppler factor δ 10 18 18 18 Normalization of electron distribution |$N_0^\prime \times 10^{48}\:{\rm eV^{-1}}$| |$10.68_{-2.64}^{+3.09}$| |$43.44_{-7.76}^{+6.59}$| |$23.83^{+8.11}_{-7.32}$| |$6.12^{+1.67}_{-1.56}$| Low-energy electron spectral index α1 2.14 ± 0.04 |$2.09_{-0.04}^{+0.03}$| |$2.10_{-0.05}^{+0.04}$| 1.98 ± 0.03 High-energy electron spectral index α2 |$3.39_{-0.14}^{+0.27}$| 3.38 ± 0.06 3.46 ± 0.06 3.64 ± 0.01 Minimum electron energy |$E^\prime _{\rm min}$| (MeV) |$1.84_{-1.23}^{+1.75}$| |$286.37_{-25.39}^{+30.64}$| |$1.97^{+0.31}_{-0.34}$| |$4.16^{+1.00}_{-1.86}$| Break electron energy |$E^\prime _{\rm br}$| (GeV) |$2.83_{-0.31}^{+0.51}$| |$1.11_{-0.12}^{+0.14}$| |$1.62^{+0.23}_{-0.15}$| |$3.11^{+0.15}_{-0.23}$| Maximum electron energy |$E^\prime _{\rm max}$| (GeV) |$46.27_{-13.76}^{+49.74}$| |$82.32_{-17.14}^{+13.47}$| |$127.82^{+26.74}_{-24.75}$| |$202.79^{+21.2}_{-14.6}$| Magnetic field B [G] 0.19 ± 0.013 0.046 ± 0.002 |$0.11^{+0.005}_{-0.004}$| |$0.18^{+0.009}_{-0.006}$| Jet power in magnetic field LB × 1043 erg s−1 0.49 0.31 1.71 4.51 Jet power in electrons Le × 1045 erg s−1 2.11 4.07 9.60 4.47 Open in new tab Table 2. Model parameters. . Parameter . Quiescent . Quiescent . January . April . Doppler factor δ 10 18 18 18 Normalization of electron distribution |$N_0^\prime \times 10^{48}\:{\rm eV^{-1}}$| |$10.68_{-2.64}^{+3.09}$| |$43.44_{-7.76}^{+6.59}$| |$23.83^{+8.11}_{-7.32}$| |$6.12^{+1.67}_{-1.56}$| Low-energy electron spectral index α1 2.14 ± 0.04 |$2.09_{-0.04}^{+0.03}$| |$2.10_{-0.05}^{+0.04}$| 1.98 ± 0.03 High-energy electron spectral index α2 |$3.39_{-0.14}^{+0.27}$| 3.38 ± 0.06 3.46 ± 0.06 3.64 ± 0.01 Minimum electron energy |$E^\prime _{\rm min}$| (MeV) |$1.84_{-1.23}^{+1.75}$| |$286.37_{-25.39}^{+30.64}$| |$1.97^{+0.31}_{-0.34}$| |$4.16^{+1.00}_{-1.86}$| Break electron energy |$E^\prime _{\rm br}$| (GeV) |$2.83_{-0.31}^{+0.51}$| |$1.11_{-0.12}^{+0.14}$| |$1.62^{+0.23}_{-0.15}$| |$3.11^{+0.15}_{-0.23}$| Maximum electron energy |$E^\prime _{\rm max}$| (GeV) |$46.27_{-13.76}^{+49.74}$| |$82.32_{-17.14}^{+13.47}$| |$127.82^{+26.74}_{-24.75}$| |$202.79^{+21.2}_{-14.6}$| Magnetic field B [G] 0.19 ± 0.013 0.046 ± 0.002 |$0.11^{+0.005}_{-0.004}$| |$0.18^{+0.009}_{-0.006}$| Jet power in magnetic field LB × 1043 erg s−1 0.49 0.31 1.71 4.51 Jet power in electrons Le × 1045 erg s−1 2.11 4.07 9.60 4.47 . Parameter . Quiescent . Quiescent . January . April . Doppler factor δ 10 18 18 18 Normalization of electron distribution |$N_0^\prime \times 10^{48}\:{\rm eV^{-1}}$| |$10.68_{-2.64}^{+3.09}$| |$43.44_{-7.76}^{+6.59}$| |$23.83^{+8.11}_{-7.32}$| |$6.12^{+1.67}_{-1.56}$| Low-energy electron spectral index α1 2.14 ± 0.04 |$2.09_{-0.04}^{+0.03}$| |$2.10_{-0.05}^{+0.04}$| 1.98 ± 0.03 High-energy electron spectral index α2 |$3.39_{-0.14}^{+0.27}$| 3.38 ± 0.06 3.46 ± 0.06 3.64 ± 0.01 Minimum electron energy |$E^\prime _{\rm min}$| (MeV) |$1.84_{-1.23}^{+1.75}$| |$286.37_{-25.39}^{+30.64}$| |$1.97^{+0.31}_{-0.34}$| |$4.16^{+1.00}_{-1.86}$| Break electron energy |$E^\prime _{\rm br}$| (GeV) |$2.83_{-0.31}^{+0.51}$| |$1.11_{-0.12}^{+0.14}$| |$1.62^{+0.23}_{-0.15}$| |$3.11^{+0.15}_{-0.23}$| Maximum electron energy |$E^\prime _{\rm max}$| (GeV) |$46.27_{-13.76}^{+49.74}$| |$82.32_{-17.14}^{+13.47}$| |$127.82^{+26.74}_{-24.75}$| |$202.79^{+21.2}_{-14.6}$| Magnetic field B [G] 0.19 ± 0.013 0.046 ± 0.002 |$0.11^{+0.005}_{-0.004}$| |$0.18^{+0.009}_{-0.006}$| Jet power in magnetic field LB × 1043 erg s−1 0.49 0.31 1.71 4.51 Jet power in electrons Le × 1045 erg s−1 2.11 4.07 9.60 4.47 Open in new tab The emission in flaring periods is modelled assuming that the HE emission is entirely due to the IC scattering of external photons (Fig. 4). In all calculations the absorption due to the EBL was taken into account using a model from Franceschini, Rodighiero & Vaccari (2008) since a strong absorption is evident at ≥100 GeV (red dashed line in Fig. 4). In both periods the HE electron spectral index is within the range of α2 ∼ (3.46 − 3.64) which is required to explain the UV-X-ray data with a photon index of ≈2.3. The lack of low-energy data makes the precise estimation of the low-energy electron index harder. Only the Swift XRT/NuSTAR data from the observation of the transition region between low- and high-energy components allows to define the parameters Emin and α1. The low-energy electron index is in a typical range expected from shock acceleration theories, α1 ≈ 2. As distinct from the quiescent state, in order to explain the flaring activities, both, the electron distribution and the magnetic field should be varying. We note that the magnetic field required for modelling of flaring periods, (B ≥ 0.11 G), is weaker than the one estimated in the quiescent state in case of δ = 10 (B ∼ 0.19 G). Since the synchrotron emission depends on the total number of emitting electrons Ne, δ and magnetic field strength B, in case of smaller δ (and emitting region size) the magnetic field should be stronger. Instead, when δ is constantly equal to 18 in both states, the magnetic field should be nearly ∼2.4 and ∼3.9 times stronger in January and April, respectively, in order to explain the observed data. As the synchrotron photon density is proportional to B2, the increase in the magnetic field strength resulted in the observed increase of the synchrotron flux by a factor of 5.7 and 15.3 (Fig. 4). In the modelling of the SEDs observed in the flaring periods of January and April, the magnetic field should be changed in accordance with the increase in the low-energy component. Since the emission in the HE band is dominated by the IC scattering of external photons, this component remains stable during those periods [this corresponds to Fig. 1(b) Paggi et al. (2011)]. The electron spectra obtained during the fit of SEDs in quiescent and flaring states are shown in Fig. 5. It is clear the evolution of the electron spectra during the quiescent and high states. The low-energy indexes of the underlying electron distribution in the quiescent state (δ = 10) are softer as compared with the flaring period (April). The total electron energy for modelling the emission in the quiescent period, when δ = 18, is almost of the same order as that in the flaring periods, which is expected, as the magnetic field is weaker, most of the jet energy is carried by particles. During the flaring periods, there are evident changes also in the underlying electron distribution. The electron distribution best describing the data observed in April hints at i) hardening of the low-energy index; ii) a higher break at ∼3.1 GeV and maximum energies of ∼203 GeV. Ebr and Emax are expected to shift, as the γ-ray spectrum observed in April is slightly inclined towards HEs, as compared with the January spectrum (see Fig. 4). However, due to the large uncertainties in the estimations, especially for α1 (since the data in between 100 keV and 100 MeV are missing), no definite conclusions can be drawn. For a statistically significant claim for hardening, there are required additional data in the energy range characterizing the rising part of the low- and high-energy components, which will allow to constrain α1 with higher confidence. We note, however, that the significant hardening of the γ-ray emission observed in April [Fig. 2(c)] supports and strengthens the assumptions on the hardening of the low-energy electron index. Figure 5. Open in new tabDownload slide The electron spectra (broken power laws) obtained from the fit of the quiescent and flaring states of PKS 1441+25. Details on the parameter values are given in Table 2. Figure 5. Open in new tabDownload slide The electron spectra (broken power laws) obtained from the fit of the quiescent and flaring states of PKS 1441+25. Details on the parameter values are given in Table 2. Similar modelling of the SED of PKS 1441+25 observed in April has been already done in Abeysekara et al. (2015) and Ahnen et al. (2015), but it was done in a different manner. For example, in Ahnen et al. (2015) the low-electron energy index is fixed to be α1 = 2, a value expected from strong shock acceleration theories, while in our case all the parameters can vary in the fitting procedure. After having observed the hardening of the γ-ray photon index in April, we believe that exact estimation of α1 is important. Moreover, possible hardening or softening of α1 would point out the acceleration processes in the jet. However, the main difference in the modelling presented here, as compared with the previous ones, is the size of the emitting region (blob). They used larger blob size, 5 × 1016 cm, in Ahnen et al. (2015) and 4 × 1017 cm in Abeysekara et al. (2015). In our case, the modelling of the January flare time profile allowed us to constrain the emitting region size by Rb ≤ 3.5 × 1016 (δ/18) cm. Another difference with the previously reported parameters is that in our case the electron energy density is nearly 100 times higher than the magnetic field energy density. In Abeysekara et al. (2015) Ue/UB = 1.5, which is related to the fact that much bigger emitting region size is used. We note that in Ahnen et al. (2015), where the considered blob size is similar to our case, they also found that Ue/UB ≥ 10. Moreover, in our case the radius of the IR torus is derived from a different scaling law, which can cause additional difference in the estimation of the total energy. Despite using different approaches and parameters as compared with those used in the previous modellings, we note, that the main parameters for the underlying electron distribution obtained during April are similar to the previously reported values. 4 DISCUSSION AND CONCLUSIONS In this paper, we present the results of the study of the γ-ray emission from PKS 1441+25 during 2015 January–December. The data from the observations of a bright GeV flare in January allow us to estimate the emitting region size whereas the modelling of the broad-band SED of PKS 1441+25 in January and April provided a chance to probe into the physical process during the flaring periods. The γ-ray light curve generated with an adaptive binning method shows that the source entered its high activity state around MJD 57 043.3 and the flux reached its maximum on January 24, when, within a few hours, the flux increases up to Fγ(>100 MeV) = (2.22 ± 0.38) × 10−6 photon cm−2 s−1. During this γ-ray brightening the fit of the flare profile shows a slow rise and a fast decay trend with the shortest variability (flux doubling) time being τd = 1.44 d. The rise of the flare can be attributed to the shock acceleration, whereas the decay phase cannot be explained by cooling of particles. Indeed, for the electrons that emit γ-rays with εγ = 1 GeV, as measured in the observer frame, the corresponding cooling time-scale would then be ∼(3 me c/σT uIR) × (εIR(1 + z)/εγ)0.5 (Saito et al. 2013) which corresponds to 0.47 d in this case. This time-scale is shorter than the observed e-folding decay time-scales of the flares implying that the observed flux decrease is related to the processes other than radiative losses. After the flare on MJD 57 049.38, the source is in its quiet state and the next increase in the flux is observed starting from MJD 57 109.89. Even if during this period, the flux amplitude is lower than the one that was observed in January, an interesting modification of the γ-ray emission spectrum is observed. First, the γ-ray photon index hardened during MJD 57 126.70–57 141.93, it was ≤1.9. This period coincides with the one when VHE γ-rays from PKS 1441+25 were detected. The hardest γ-ray photon index, Γ = 1.54 ± 0.16, has been observed on MJD 57 131.46 with a convincingly high detection significance of 11.8σ. This photon index is unusual for FSRQs which are with an averaged photon index of 2.4 in the third Fermi LAT AGN catalogue [see Fig. 8 of Ackermann et al. (2015)]. This photon index is even harder than the index of B3 1151+408 (Γ = 1.77) which has the hardest photon indexes in the clean sample of Fermi LAT detected FSRQs. Although, hard photon indexes have been occasionally observed during rapid flaring events in FSRQs (Pacciani et al. 2014). The observed hardening was perhaps related to the emission of new energetic particles that were either injected into the emitting region or re-accelerated. Next, the data analysis covering only the period in April shows that the γ-ray flux hints at a spectral curvature and a power law with an exponential cut-off model is preferred over the simple power-law modelling assuming a break around Ecut = 17.7 ± 8.9 GeV with a significance of 2.8σ. Although the low statistics do not allow to claim for a statistically significant curvature in the spectrum, the γ-ray photon index observed in the VHE γ-ray band (∼5.4, which corresponds to an intrinsic index of 3.4 after correction for the EBL) strongly supports the presence of a break or a cutoff in the PKS 1441+25 spectrum around tens of GeV. Most likely, this break is defined by the break present in the radiating electron spectrum rather than is caused by the absorption within BLR (Poutanen & Stern 2010, otherwise the photons with >100 GeV would be strongly absorbed). The origin of multiwavelength emission: The SEDs observed during quiescent and flaring states are modelled using one-zone leptonic models and the model parameters are estimated using the MCMC method. The HE γ-ray emission observed in the flaring states can be explained by IC scattering of IR photons from the dusty torus whereas the SSC model gives a satisfactory representation of the data observed during the quiescent state. The flares observed in January and April can be explained assuming there are changes in the bulk Lorentz factor or in the magnetic field. If the emitting region leaves the BLR region due to the increase of the bulk Lorentz factor (from δ = 10 to δ = 18), the Compton dominance will increase as it has been observed in the γ-ray band. Indeed in the flaring states, the IC to synchrotron luminosities ratio Lγ/Lsyn ≈ 200 and ≈28 in January and April, respectively as compared with that in the quiescent state Lγ/Lsyn ≈ (2 − 4). At the same time, the increase in the low-energy component indicates that the magnetic field also increased between the flares in January and April (Paggi et al. 2011). On the other hand, if the bulk Lorentz factor is unchanged (δ = 18), only the change in the emitting region location and amplification of the magnetic field can explain the multifrequency behaviour observed during the flares. It is possible to distinguish between these two scenarios, provided there are data in the hard X-ray or soft γ-ray band, as the modelling with δ = 18 predicts a higher flux in the hard X-ray band than when δ = 10 is assumed (grey dashed and solid lines in Fig. 2). Such data are missing in this case, making it hard to give exact interpretation of the origin of the flare. Anyway, physically reasonable parameters are used in both of these scenarios. When comparing the electron parameters required for the modelling of the SEDs in January and April, we find a hint of possible hardening of the low-energy electron index in April. We note, however, that no definite conclusions can be drawn since α1 is poorly constrained (due to missing or non-sufficient data). For all that, the April hardening of the γ-ray photon index in the MeV-GeV energy region supports our assumptions on hardening of the power-law index of the underlying electron distribution. Jet Energetics: The jet power in the form of magnetic field and electron kinetic energy are calculated by |$L_{B}=\pi c R_b^2 \Gamma ^2 U_{B}$| and |$L_{{\rm e}}=\pi c R_b^2 \Gamma ^2 U_{{\rm e}}$|⁠, respectively, and are given in Table 2. The jet power in the electrons changes in the range (4.5 − 9.6) × 1045 erg s−1 during the flares, while in the quiescent state it is of the order of (2.1 − 4.1) × 1045 erg s−1. Assuming one proton per relativistic electron (e.g. Celotti & Ghisellini 2008; Ghisellini 2013), the total kinetic energy in the jet is Lkin = 8.02 × 1047 erg s−1 and Lkin = 1.35 × 1047 erg s−1 for January and April, respectively. The maximum γ-ray flux during the period of high activity is (2.22 ± 0.38) × 10−6 photon cm−2 s−1 which corresponds to an isotropic γ-ray luminosity of Lγ = 1.22 × 1049 erg s−1 (using a distance of dL ≈ 6112.8 Mpc). Likewise, the γ-ray luminosities in the periods of January and April were Lγ = 3.48 × 1048 erg s−1 and Lγ = 5.21 × 1048 erg s−1, respectively. Yet, at δ = 18 the total power emitted in the γ-ray band in the proper frame of the jet would be Lem,γ = Lγ/2 δ2 = 1.89 × 1046 erg s−1 during the peak flux and would change within Lem,γ = (5.38 − 8.04) × 1045 erg s−1 in January and April. These luminosity values account for only a small fraction (≤6.7 per cent) of the total kinetic energy of the jet. However, assuming that the standard radiative efficiency of the accretion disc ηdisc ∼ 10 per cent, the accretion power would be Lacc = 2 × 1046 erg s−1. Thus during the flaring period the power emitted as γ-ray photons constitutes the bulk of the total accretion power Lem,γ/Lacc ≈ 1, while in January and April it made a substantial fraction of it – Lem,γ/Lacc ≈ (0.3–0.4); this is in a good agreement with the recent results by Ghisellini et al. (2014), which showed that the radiative jet power in blazars is higher than (or of the order of) the accretion disk luminosity. The observations in both X-ray and γ-ray bands show that after the activity observed in January and April the emission from the source again enters a quiescent state. A small increase in the γ-ray flux has been observed only in 2015 June–August and October–November. Also, the UV/X-ray flux measured by Swift in 2015 May (Abeysekara et al. 2015) shows that the synchrotron component is weaker than it was in April. Thus, this indicates that the magnetic field in the emitting region started to decrease. In addition, in the γ-ray band, the flux slowly decreases down to a few times 10−7 photon cm−2 s−1 for most of the time after 2015 August, and the γ-ray photon index reaches its mean level. These point out that the emission from the blob outside the BLR region weakened, and the decrease of the Compton component shows that the emission responsible for the emission in the quiescent state (SSC) starts to dominate again. Since in this case the emission occurs close to the central source, due to the strong absorption, it is not expected to have emission of VHE γ-ray photons. The multiwavelength observations of PKS 1441+25 during the flaring periods allowed us to investigate and discuss the changes that possibly took place in the jets and caused flaring activities. However, the parameters describing the underlying electron distribution below the break are poorly constrained, because the data describing the rising part of both low and HE components are missing. It did not allow us to exactly identify the processes responsible for the acceleration of particles in the jet. However, the future possible observations of flaring periods also in other energy bands will provide a chance to investigate the dominant particle acceleration processes. Acknowledgements This work was supported by the Republic of Armenia Ministry of Education and Science State Committee of Science, in the frames of the research project No 15T-1C375. Also, this work was made in part by a research grant from the Armenian National Science and Education Fund (ANSEF) based in New York, USA. We thank the anonymous referee for constructive comments that significantly improved the paper. 1 " http://fermi.gsfc.nasa.gov/ssc/data/analysis/software/ 2 " http://tools.asdc.asi.it/SED/index.jsp REFERENCES Abdo A. A. et al. , 2010a , ApJ , 721 , 1425 Crossref Search ADS Abdo A. A. et al. , 2010b , ApJ , 722 , 520 Crossref Search ADS Abeysekara A. U. et al. , 2015 , ApJ , 815 , L22 Crossref Search ADS Acero F. , Ackermann M., Ajello M.et al. , 2015 , ApJS , 218 , 23 Crossref Search ADS Ackermann M. , Ajello M.et al. , 2015 , ApJ , 810 , 14 Crossref Search ADS Aharonian F. A. , Atoyan A. M., 1981 , Ap&SS , 79 , 321 Aharonian F. et al. , 2007 , ApJ , 664 , L71 Crossref Search ADS Aharonian F. A. , Kelner S. R., Prosekin A. Y., 2010 , Phys. Rev. D , 82 , 043002 Crossref Search ADS Ahnen M. L. et al. , 2015 , ApJ , 815 , L23 Crossref Search ADS Ahnen M. L. et al. , 2016 , A&A , 595 , A98 Crossref Search ADS Aleksic J. , Ansoldi S.et al. , 2014 , Science , 346 , 1080 Crossref Search ADS PubMed Atwood W. B. et al. , 2009 , ApJ , 697 , 1071 Crossref Search ADS Błażejowski M. , Sikora M., Moderski R., Madejski G. M., 2000 , ApJ , 545 , 107 Crossref Search ADS Bloom S. D. , Marscher A. P., 1996 , ApJ , 461 , 657 Crossref Search ADS Celotti A. , Ghisellini G., 2008 , MNRAS , 385 , 283 Crossref Search ADS Coppi P. S. , Aharonian F. A., 1999 , ApJ , 521 , L33 Crossref Search ADS Cutini S. et al. , 2014 , MNRAS , 445 , 4316 Crossref Search ADS Finke J. D. , Dermer C. D., Böttcher M., 2008 , ApJ , 686 , 181 Crossref Search ADS Franceschini A. , Rodighiero G., Vaccari M., 2008 , A&A , 487 , 837 Crossref Search ADS Ghisellini G. , 2013 , Lecture Notes in Physics, Vol. 873, Radiative Processes in High Energy Astrophysics. Springer-Verlag, Berlin Ghisellini G. , Tavecchio F., 2009 , MNRAS , 397 , 985 Crossref Search ADS Ghisellini G. , Tavecchio F., 2015 , MNRAS , 448 , 1060 Crossref Search ADS Ghisellini G. , Maraschi L., Treves A., 1985 , A&A , 146 , 204 Ghisellini G. , Maraschi L., Tavecchio F., 2009 , MNRAS , 396 , L105 Crossref Search ADS Ghisellini G. , Tavecchio F., Maraschi L., Celotti A., Sbarrato T., 2014 , Nature , 515 , 376 Crossref Search ADS PubMed Inoue S. , Takahara F., 1996 , ApJ , 463 , 555 Crossref Search ADS Kardashev N. S. , 1962 , Soviet Ast. , 6 , 317 Liu H. T. , Bai J. M., 2006 , ApJ , 653 , 1089 Crossref Search ADS Lott B. , Escande L., Larsson S., Ballet J., 2012 , A&A , 544 , A6 Crossref Search ADS Madau P. , Phinney E. S., 1996 , ApJ , 456 , 124 Crossref Search ADS Maraschi L. , Ghisellini G., Celotti A., 1992 , ApJ , 397 , L5 Crossref Search ADS Mirzoyan R. , 2015 , The Astronomer's Telegram , 7416 Mücke A. , Protheroe R. J., 2001 , Astropart. Phys. , 15 , 121 Crossref Search ADS Mücke A. , Protheroe R. J., Engel R., Rachen J. P., Stanev T., 2003 , Astropart. Phys. , 18 , 593 Crossref Search ADS Mukherjee R. , 2015 , The Astronomer's Telegram , 7433 Nenkova M. , Sirocky M. M., Nikutta R., Ivezić Ž., Elitzur M., 2008 , ApJ , 685 , 160 Crossref Search ADS Pacciani L. , Tavecchio F., Donnarumma I., Stamerra A., Carrasco L., Recillas E., Porras A., Uemura M., 2014 , ApJ , 790 , 45 Crossref Search ADS Paggi A. , Cavaliere A., Vittorini V., D'Ammando F., Tavani M., 2011 , ApJ , 736 , 128 Crossref Search ADS Poutanen J. , Stern B., 2010 , ApJ , 717 , L118 Crossref Search ADS Sahakyan N. , Gasparyan S., 2017 , AIP Conference Proceedings , 1792 , 050005 Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Saito S. , Stawarz Ł., Tanaka Y. T., Takahashi T., Madejski G., D'Ammando F., 2013 , ApJ , 766 , L11 Crossref Search ADS Shaw M. S. et al. , 2012 , ApJ , 748 , 49 Crossref Search ADS Sikora M. , Begelman M. C., Rees M. J., 1994 , ApJ , 421 , 153 Crossref Search ADS Sikora M. , Stawarz Ł., Moderski R., Nalewajko K., Madejski G. M., 2009 , ApJ , 704 , 38 Crossref Search ADS Urry C. M. , Padovani P., 1995 , PASP , 107 , 803 Crossref Search ADS Xiong D. R. , Zhang X., 2014 , MNRAS , 441 , 3375 Crossref Search ADS Zabalza V. , 2015 , Proc. of International Cosmic Ray Conference 2015 , in press Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC © 2017 The Authors Published by Oxford University Press on behalf of the Royal Astronomical Society
TI - High energy gamma-ray emission from PKS 1441+25
JF - Monthly Notices of the Royal Astronomical Society
DO - 10.1093/mnras/stx1402
DA - 2017-09-21
UR - https://www.deepdyve.com/lp/oxford-university-press/high-energy-gamma-ray-emission-from-pks-1441-25-h0ksH0N7G9
SP - 2861
VL - 470
IS - 3
DP - DeepDyve
ER -