THE REMARKABLE γ-RAY ACTIVITY IN THE GRAVITATIONALLY LENSED BLAZAR PKS 1830–211

PKS 1830–211 is a high redshift blazar (z = 2.507; Lidman et al. 1999) gravitationally lensed by a spiral galaxy at z = 0.886 (Wiklind & Combes 1996), as showed by the two radio lobes located 1'' apart from each other (Lovell et al. 1998). The lensed counterparts were observed also in near-IR (NIR) and optical energy bands by the Hubble Space Telescope and the Gemini Observatory (Courbin et al. 2002). The source was observed in X-Rays by both XMM-Newton and Chandra, allowing us to study the complicated soft X-Ray behavior in detail, probably due to absorption (Dai et al. 2008). Moreover, the source presents a soft γ-ray energy spectrum (Γ = 2.56) as detected by EGRET (Hartman et al. 1999) and recently confirmed by the Fermi Large Area Telescope (LAT; Abdo et al. 2010).

The modeling of the spectral energy distribution (SED) of this blazar (De Rosa et al. 2005) led to the tentative classification of PKS 1830–211 as an MeV blazar (the inverse Compton peak lies below 100 MeV; Sikora et al. 2002), with the caveat that the entire data set was not simultaneous. In particular, the SED was interpreted in the standard framework of a one-zone leptonic model where the seed photons responsible for inverse Compton peak should be originated within the torus surrounding the central engine.

The source has been extensively monitored in hard X-Rays by the Imager on Board the International Gamma-Ray Astrophysics Laboratory Satellite (INTEGRAL/IBIS) for a seven-year-long period, during which no significant variability has appeared on either short (daily) or longer (monthly) timescales (Zhang et al. 2008).

Since its launch, AGILE (Tavani et al. 2009) has detected one γ-ray flare (F(E > 100 MeV) =(160 ± 50) × 10⁻⁸ photons cm⁻² s⁻¹) of PKS 1830–211 in 2009 October (Striani et al. 2009), which resulted in an enhancement by a factor of three compared with the one-week average flux detected before the flare (F(E > 100 MeV) = 40 × 10⁻⁸ photons cm⁻² s⁻¹).

One year later, Fermi/LAT detected a significant enhancement of the γ-ray emission F(E > 100 MeV) = (520 ± 110) × 10⁻⁸ photons cm⁻² s⁻¹ and F(E > 100 MeV) = (1400 ± 500) × 10⁻⁸ photons cm⁻² s⁻¹ on one-day and 6 hr timescales, respectively (Ciprini 2010), resulting in increases by factors of about 12 and 35 with respect to the average value reported by Abdo et al. (2010). Moreover, AGILE detected a prolonged γ-ray activity two days after the flare reported by Fermi (Donnarumma et al. 2010).

In this Letter, we present the results obtained by multifrequency monitoring of PKS 1830–211 from the NIR to γ-ray energy bands. We describe in detail the multi-λ data analysis and a possible interpretation of the SED of this object to reconcile the average γ-ray activity with the γ-ray enhancement. We adopt a Λ-CDM cosmology with the parameters h = 0.71, Ω_m = 0.27, and Ω_Λ = 0.73.

2.1. AGILE Observations

AGILE/GRID data were analyzed using the Build-20 software. Well-reconstructed γ-ray events were selected using the FM3.119 filter. All the events collected during the passage in the South Atlantic Anomaly were rejected. We filtered out Earth-albedo γ-rays, rejecting photons coming from a circular region of radius 80° and centered on the Earth. A one-month-long γ-ray light curve (two-day time bin, see Figure 1) across the γ-ray flare reported by Fermi/LAT was produced between 2010 October 8 00:00 UT and November 8 00:00 UT (MJD 55477−55509) using the standard AGILE maximum likelihood procedure. We obtain a mean flux on the whole integration time of F(E > 100 MeV) = (200 ± 29) × 10⁻⁸ photons cm⁻² s⁻¹, which is a factor of about four higher than the flux reported in the first Fermi catalog. We note that this γ-ray enhancement represents a 5σ excess with respect to its value in steady state as detected by AGILE (F(E > 100 MeV) =(32 ± 8) × 10⁻⁸ photons cm⁻² s⁻¹; F. Verrecchia et al. 2011, in preparation).

We extracted the γ-ray spectrum between $100\, \rm MeV\ {\rm and} 1\, \rm GeV$, discarding the energy channels above 1 GeV due to poor statistics. The AGILE spectrum is well fitted by a power law with a photon index of 2.05 ± 0.17 (all the errors reported in this Letter are at 1σ, unless otherwise stated).

As shown in Figure 1, the source remained at a flux level of ∼200 × 10⁻⁸ photons cm⁻² s⁻¹ on a one-month timescale. A clear enhancement emerged on October 14 (MJD 55483) that lasted about four days. The two-day binned AGILE data showed that the source reached its maximum of F(E > 100 MeV) ∼(500 ± 130) × 10⁻⁸ photons cm⁻² s⁻¹ on October 15 (MJD 55484.75 ± 0.7); the estimate of this epoch accounts for the bias introduced by the choice of the starting time (T_start) and the binning adopted to obtain the light curve. This was done by operating four translations of T_start (by 0, 0.5, 1, and 1.5 days) in the light curve.

Zoom In Zoom Out Reset image size

We also extracted the γ-ray spectrum during the four-day flare, obtaining a power-law photon index Γ = 2.40 ± 0.28. Because of the poor statistics on this short duration, it was not possible to infer whether any hardening or softening occurred during this flare, the photon indices being compatible within 1σ. In Section 3, we will focus on the one-month γ-ray enhancement and on the four-day flare detected between October 14 and 18 (horizontal lines in Figure 1).

2.2. X-Ray Observations

2.3. Optical Monitoring: SMARTS Consortium

Here, we discuss the SED modeling of the blazar PKS 1830–211 during the high γ-ray state observed between 2010 October and November. Multifrequency observations were activated to follow the γ-ray flare covering NIR–optical (SMARTS), and soft and hard X-Ray energy bands (Swift/XRT and INTEGRAL/IBIS, respectively). The main results obtained during this multifrequency campaign are that the high and unusual activity recorded in γ-rays seems to have no significant counterpart at lower frequencies. In detail, the simultaneous NIR–optical and soft and hard X-Ray emissions of this source did not follow the significant changes observed in γ-rays. As for the NIR and optical emissions, we derived only upper limits for this source that are indeed vary faint, especially in the optical energy bands (Courbin et al. 2002). Nevertheless, by comparing the derived upper limits in the J band (red downward arrow in Figure 2) with the NIR data reported by Courbin et al. 2002 and Meylan et al. 2005, we can exclude variation in the thermal + non-thermal components greater than a factor of ∼1.5. The soft and hard X-Ray emissions also exclude variation greater than a factor of 1.6 and 1.5, respectively. It is worth noting that the observed variations of the SED rule out the hypothesis that the γ-ray emission was connected to macrolensing since its effects would be energy independent. On the other hand, the chromaticism of the SED variability may suggest that microlensing from stars in the lensing galaxy may cause the observed γ-ray variability (see, e.g., Torres et al. 2003). In order to investigate this possibility, we simulated the transition of a source with different dimensions (Jovanović et al. 2008) across the microlensing magnification pattern for the lens system (taking the convergence k = 0.158 and shear γ = 0.096 as given in Winn et al. 2002). By assuming a γ-ray emitting region of 10¹⁵ cm (observer frame), we found an amplification by a factor of tens on a timescale of years. Even worse is the case of the γ-ray flare observed by Fermi on a 6 hr timescale, which is a factor of 35 higher than the steady state, requiring high amplification in a very short time. Shorter timescales of microlensing can be obtained with a very compact source (on the order of 10¹³ cm), but we cannot further reduce the γ-ray emitting region; otherwise, it would result in an increase in γ-ray opacity due to pair production. Therefore, we found that the observed variability is more likely intrinsic to this blazar.

Zoom In Zoom Out Reset image size

We complemented the low-frequency data with the non-simultaneous Planck observations (30, 100, and 217 GHz) taken from the Early Release Compact Source Catalogue (ERCSC) extracted from the first all-sky survey.¹⁷ In Figure 2 we report all the observed multifrequency data and the SED modeling, taking into account the achromatic magnification due to the gravitational lensing, assumed to be one order of magnitude (De Rosa et al. 2005).

We model three γ-ray states of the source. The first one is the average state reported by De Rosa et al. (2005) and consistent with the average γ-ray state reported in Abdo et al. (2010; ∼40 × 10⁻⁸ photons cm⁻² s⁻¹; see black points in Figure 2); the second is the one-month enhancement showing a mean flux of ∼200 × 10⁻⁸ photons cm⁻² s⁻¹ (blue points in Figure 2); the third is the peak of the γ-ray emission (October 14–18; hereafter "flare"), which is a factor of ∼2–3 higher than the one-month average flux (red points in Figure 2). We note that the multifrequency behavior of the source, i.e., the lack of correlated variability between the low (NIR–optical bands, X-Rays) and high (γ-rays) energy portions of the SEDs, disfavors the one-zone leptonic model for this event. We account for this evidence, assuming two relativistic electron populations with the energy density described by a broken power law in the Lorentz factor domain (γ) with break equal to γ_b:

$$\begin{equation} n_{e}(\gamma)=\frac{K\gamma _{b}^{-1}}{(\gamma /\gamma _{b})^{a_{l}}+(\gamma /\gamma _{b})^{a_{h}}}\,, \end{equation} \tag{ 1 }$$

where a_l and a_h are the pre- and post-break electron distribution spectral indices, respectively. We assume that the two electron populations contained a random magnetic field B. In Table 1 we summarize the values of the model parameters for the two components.

We estimate the contributions of both electron populations to synchrotron emission and to synchrotron self-Compton plus inverse Compton from external seed photons (EC) originating in a dusty torus, in the accretion disk, and in the broad-line region (BLR; see Sikora et al. 2002; Vittorini et al. 2009). In particular, we assume (1) the emission from a torus with temperature T = 10³ K and 1.26 × 10¹⁹ cm from the central engine, (2) the direct emission from the disk with T = 1.5 × 10⁴ K and luminosity L_disk = 6 × 10⁴⁵ erg s⁻¹ and 5 × 10¹⁶ cm from the second population, and (3) the emission from the BLR and R_BLR = 5 × 10¹⁷ cm from the central engine and reprocessing 10% of the disk luminosity.

In this scenario, the average γ-ray state is produced through EC on dusty torus, disk, and BLR seed photons mainly by the first population, resulting in a jet power of P¹_jet = 10⁴⁶ erg s⁻¹ (black line in Figure 2). The one-month γ-ray enhancement, a factor of four higher than the average flux, has required the additional contribution of the second electron population, which is characterized by a smaller size, higher electron density and higher γ_b with respect to the first electron population (Table 1). On the basis of these assumptions, the γ-ray enhancement (blue points in Figure 2) is then produced by EC of the second population on the photon density field while moving inside the BLR with Γ = 15, transporting a jet power of P²_jet = 3 × 10⁴⁶ erg s⁻¹. This modeling is in agreement with the source fading at its average flux level when the blob is moving outside the BLR (Δt ∼ R_BLR(1 + z)/cδ ≳ 30 days). Following this description, the synchrotron and inverse Compton emissions are dominated by the first population, thus accounting for the moderate variation inferred by the NIR–optical and X-Ray observations. On the other hand, the "flare," which is a factor of 12 greater than the average flux and a factor of ∼2–3 greater than the one-month activity, requires a "local" enhancement of the photon density field by a factor of 3 (red line in Figure 2), likely due to a blob–cloud interaction (Araudo et al. 2010). In this case we derived a jet power of P^flare_jet = 6 × 10⁴⁶ erg s⁻¹. The lack of correlated variabilities in the optical, X-Rays, and γ-rays during the flare prevented us from explaining this event as associated with changes in the electron population in terms of injection/acceleration mechanisms.

We emphasize that the γ-ray behavior of PKS 1830–211 recorded by AGILE and Fermi/LAT between 2010 October and November is rare as probed by the lack of similar variability since 2007. Similarly, no hard X-Ray variability has been observed during seven years of monitoring by INTEGRAL/IBIS. In conclusion, we can attribute the properties of PKS 1830–211 to the ones of "γ-ray only flaring" blazars showing substantial variations, especially in the γ-ray energy range above 10–100 MeV. These evidences strongly support our interpretation of a "steady" electron population, filling the jet below the BLR, that is responsible for the average γ-ray emission recorded by both EGRET and Fermi/LAT more than 10 years apart.

It is worth noting that the inspection of the γ-ray light curve during a longer period across the October–November enhancement did not show any evidence of the time delay between the emissions of the two lensed images A and B as measured in the radio maps (26⁺⁵ _{− 4} days; Lovell et al. 1998). We can assume the lack of the delay between A and B only if the flux ratio of the two components is ∼1. If it was lower than 1, we are prevented from drawing any conclusions because the emission would have occurred below the AGILE sensitivity level. We also note that microlensing could play a role in explaining the possible lack of echo in the γ-ray light curve. In fact, in some lensed quasars (Blackburne et al. 2006), different flux ratios in the lensed images were detected between the optical and X-Rays. This dependency of the flux ratio on the energy has been interpreted as due to microlensing, thus justifying different amplifications as a function of the emitting region size (see also Jovanović et al. 2008). On the basis of this evidence, we can consider that the lack of the γ-ray echo may be due to different flux ratios between the two components in radio and γ-ray energy bands, i.e., it might be that one component (due to the flux anomaly) has a negligible contribution in γ-rays. Consequently, intrinsic variation in combination with microlensing can also be considered.

We thank the anonymous referee for helpful comments and suggestions. We warmly thank the INTEGRAL/IBIS survey team for providing us the seven-year hard X-Ray spectrum of PKS 1830–211. Research was partially supported by ASI grant No. I/040/10/0. L.C.P and S.S. are supported by the Serbian Ministry of Science (project 176001).