Gamma-Ray Flares from Mrk421 in 2008 observed with the ARGO-YBJ experiment
aa r X i v : . [ a s t r o - ph . H E ] J u l Mem. S.A.It. Vol. 75, 282 c (cid:13) SAIt 2008
Memorie della
Gamma-Ray Flares from Mrk421 in 2008observed with the ARGO-YBJ experiment
G. Di Sciascio on behalf of the ARGO-YBJ Collaboration
INFN – Sez. di Roma Tor Vergata, Viale della Ricerca Scientifica 1, I-00133 Roma, Italye-mail: [email protected]
Abstract.
In 2008, the blazar Mrk421 entered in a very active phase and was one of thebrightest sources in the sky at TeV energies, showing strong and frequent flaring. Wesearched for γ -ray emission at energies E > ± . × − γ cm − s − for E γ > Key words.
Galaxy: Mrk421 – Galaxy: TeV gamma-ray flares – Gamma rays: observations
1. Introduction
Mrk421 ( z = / TeV energies. X-raysare generally attributed to synchrotron radi-
Send o ff print requests to : G. Di Sciascio ation from high energy electrons, while theorigin of the γ -ray emission is more un-certain. Possibilities include inverse Comptonscattering of synchrotron (Synchrotron Self-Compton, SSC) or ambient photons (externalCompton, EC) o ff a single electron population(Sambruna et al. 1996; Ghisellini et al. 1998;Fossati et al. 2008). Nevertheless, an alterna-tive hadronic model ( γ -rays from proton syn-chrotron (Mucke et al. 2003)) for Mrk421 isnot ruled out yet.Multiwavelength observations are thekey for understanding the blazar phe-nomenon. Indeed, it is generally true thathadronic models are in trouble to re-produce the observed highly correlated i Sciascio: Mrk421 with ARGO-YBJ 283 X-ray / TeV variability, that strongly sup-ports the SSC models (Fossati et al. 2008;Fidelis & Iakubovskyi 2008; Wagner 2008).The ARGO-YBJ experiment is presentlythe only wide-field of view γ -ray telescopeable to detect AGN TeV flaring activity ona few day period. In this paper we reporton the monitoring of Mrk421 performed withARGO-YBJ in 2008. We will give special at-tention to the 2008 June flaring activity be-cause an extraordinary set of simultaneousmeasurements covering 12 decades of energy,from optical to TeV gamma rays, is available(Donnarumma et al. 2009).
2. The ARGO-YBJ experiment
The ARGO-YBJ detector, located at theYangBaJing Cosmic Ray Laboratory (4300 ma.s.l., Tibet, P.R. China), is the only exper-iment exploiting the full coverage approachat very high altitude. The detector is consti-tuted by a central carpet ∼ ×
78 m , madeof a single layer of Resistive Plate Chambers(RPCs) with ∼
92% of active area, enclosedby a partially instrumented guard ring that ex-tends the detector surface up to ∼ × . The apparatus has a modular structure,the basic data acquisition element being acluster (5.72 × ), divided into 12 RPCs(2.80 × each). Each chamber is readby 80 strips of 7 ×
62 cm (the space pixel),logically organized in 10 independent pads of56 ×
62 cm representing the time pixel of thedetector. The RPCs are operated in streamermode with a standard gas mixture (Argon 15%,Isobutane 10%, TetraFluoroEthane 75%), theHigh Voltage settled at 7.2 kV ensures an over-all e ffi ciency of about 96% (Aielli et al. 2006).The full detector is composed of 153 clus-ters for a total active surface of ∼ .All events giving a number of fired padsN pad ≥ N trig in the central carpet within atime window of 420 ns are recorded. The spa-tial coordinates and the time of any fired padare used to reconstruct the position of theshower core and the arrival direction of the pri-mary, as described in (Di Sciascio et al. 2007;Di Sciascio et al. 2008). Since 2007 November the full detector isin stable data taking at the multiplicity triggerthreshold N trig ≥
20 with a duty cycle ∼
3. Data analysis
The dataset for the analysis of Mrk421 pre-sented in this paper contains all showers withN pad ≥
40 and zenith angle θ < ◦ .A 20 ◦ × ◦ sky map in celestial coordi-nates (right ascension and declination) with0.1 ◦ × ◦ bin size, centered on the sourcelocation, is filled with the detected events.The background is evaluated with the timeswapping method (Alexandreas et al. 1992). N ”fake” events are generated for each detectedone, by replacing the measured arrival timewith new ones. These times are randomly se-lected within a 3 hours wide bu ff er of recordeddata. Swapping the time means swapping theright ascension, keeping unchanged the decli-nation. A new sky map (background map) isbuilt by using 10 such fake events for each realone, so that the statistical error on the back-ground can be kept small enough.To maximize the signal to noise ratio, thebins are then grouped over a circular area ofradius ψ , i.e. every bin is filled with the con-tent of all the surrounding bins whose centeris closer than ψ from its own center. Whenthe Point Spread Function of the detector isa Gaussian with r.m.s. σ , the opening an-gle ψ containing 71.5% of the events maxi-mizes the signal to background ratio for a pointsource with a uniform background, and it isequal to 1.58 · σ . The values of ψ are 1.9 ◦ , 0.9 ◦ and 0.5 ◦ respectively for N pad ≥
40, 100 and300, in agreement with MonteCarlo simula-tions (Vernetto et al. 2009a).Finally, the integrated background map issubtracted from the corresponding integratedevent map, thus obtaining the ”source map”.For each bin of this latter map, the significancewith respect to the background is calculated.This analysis procedure has been testedwith the Crab Nebula, the standard candle forVHE astronomy. At the Yangbajing latitude theCrab culminates at zenith angle θ culm = ◦ and it is observable every day for 5.8 hours
84 Di Sciascio: Mrk421 with ARGO-YBJ
Energy (TeV) d N / d E ( c m - s - T e V - ) -16 -15 -14 -13 -12 -11 -10 -9 -8 -1 Fig. 1.
The Crab Nebula energy spectrum measuredin 2008 by ARGO-YBJ compared with the results ofsome other detectors. with a zenith angle θ < ◦ . The Crab Nebulahas been observed from 2007 November to2009 March, for a total of 424 days on-source,obtaining a signal with a statistical significanceof 7 standard deviations for N pad ≥
40. The cor-responding photon median energy is 1.1 TeV.In this analysis no events selection or γ / hadrondiscrimination algorithm has been applied. Theaverage number of gamma rays detected perday in the observational window centered onthe source position is 155 ± ff erential spectrum dN / dE = K · E − α , we simulated a source in the sky fol-lowing the diurnal path of the Crab, and eval-uated the number of events expected in the 3N pad bins 40-99, 100-299 and ≥ ff er-ent values of K and α (10 − < K < − photonscm − s − TeV − and 1.5 < α < dN / dE = (3 . ± . × − E − . ± . (1)in fair agreement with other experiments (seeFig. 1). Energy (TeV) d N / d E ( c m - s - T e V - ) -13 -12 -11 -10 Fig. 2.
The Mrk421 energy spectrum measured byARGO-YBJ from 2008 day 41 to 180, when thesource was in active state. The line represents thebest fit to the data.
4. Mrk421 analysis
The same analysis has been performed forMrk421. This source culminates at the ARGO-YBJ location at zenith angle θ culm = . ◦ , andit is observable every day for 6.38 hours with azenith angle θ < ◦ .Here the Mrk421 data collected from 2007day 311 to 2008 day 366 are presented. Usingthe same method adopted for the Crab Nebula,we evaluated the Mrk421 spectrum from 2008day 41 to day 180, where the X-ray fluxshowed the most intense flares. In this period(755 observation hours) the observed signalhad a statistical significance of 6.1 standarddeviations. The di ff erential spectrum (photonscm − s − TeV − ) that best fits to data is (seeFig. 2) dN / dE = (7 . ± . × − E − . ± . e − τ ( E ) (2)where the exponential factor e − τ ( E ) takes intoaccount the absorption of gamma rays in theExtragalactic Background Light, with τ ( E )given in (Primack et al. 2005). The integralflux above 1 TeV is (4.9 ± . × − photonscm − s − , about twice the Crab Nebula flux.The median energies corresponding to the 3N pad bins 40-99, 100-299 and ≥
300 are re- i Sciascio: Mrk421 with ARGO-YBJ 285
Fig. 3.
Plot of the RXTE / ASM X-ray and ARGO-YBJ TeV gamma-ray count rates of Mrk421. Theline shows the best linear fit. spectively: E = + . − . , 1.74 + . − . and 4.2 + . − . TeV.The observed gamma ray rate appears tobe correlated with the X-ray rate measuredby the All Sky Monitor detector aboard theRXTE satellite in the 1.5-12 keV energy range(RXTE web site), as can be seen in Fig. 3where the X-ray and ARGO-YBJ TeV gamma-ray count rates are shown for all the simultane-ous measurements in 2008. The excess eventsobserved by ARGO-YBJ, averaged over 10days, refer to showers with N pad ≥ ffi cient is 0.64.Since Mrk421 is known to vary on di ff erenttime scales, the source has been studied dur-ing 1, 10 and 30 days (Vernetto et al. 2009b).For this analysis we have considered the datataken in the period 2007 day 311 - 2009 day89. No excess has been observed on a dailyscale. Concerning the 10 scale, we observed anexcess at 4.6 s.d. in the time interval 2008 days161 - 170, during a strong X-ray flare. Lookingfor 30 days excesses, the search has been car-ried out by shifting the 30 days intervals insteps of 10 days. We found several excessesfrom Mrk421 with significances between 4 and5 s.d., in particular in the intervals: 2008 days 1 - 30, 71 - 100, 81 - 110, 91 - 120, 141 - 170,when several X-ray flares have been observed. A set of simultaneous measurements cover-ing 12 decades of energy, from optical toTeV gamma rays, was performed during thestrong flaring activity in the first half of June2008 by di ff erent detectors: WEBT (opticalR-band), SWIFT (UV, soft and hard X-rays),RXTE / ASM (soft X-rays), AGILE (hard X-rays and gamma rays) and the Cherenkov tele-scopes VERITAS and MAGIC (VHE gammarays) (Donnarumma et al. 2009).In this period two flaring episodes werereported, the first one on June 3 - 8, ob-served from optical to TeV gamma rays, thesecond one, larger and harder, on June 9- 15, observed from optical to 100 MeVgamma rays. Using this multi-frequency data,in (Donnarumma et al. 2009) the authors de-rived the SED for June 6, that shows the typicaltwo humps shape, in agreement with the SSCmodel. According to the authors, the secondhump intensity (that reached a flux of about 3.5Crab units at energy E >
400 GeV) seems to in-dicate that the variability is due to the harden-ing / softening of the electron spectrum, and notto the increase / decrease of the electron den-sity. Their model predicts for the second flare aVHE flux about a factor 2 larger with respectto the first one. Unfortunately there were noVHE data included in their multi-wavelengthanalysis after June 8 because the moonlighthampered the Cherenkov telescopes measure-ments.The VHE observation was actually madeby the ARGO-YBJ experiment, that since 2007November is performing a continuous mon-itoring of Mrk421. Fig. 4 shows the rate ofevents with N pad ≥
100 observed by ARGO-YBJ during the first half of June, averaged over3 days, compared with the X-ray flux measuredby RXTE / ASM.A second flare has been detected with astatistical significance of 3.2 standard devia-tions during the interval 11-13 June. The sig-nificance increases to 4.2 s.d. with a suitableevents selection (Vernetto et al. 2009a). Fig. 5
86 Di Sciascio: Mrk421 with ARGO-YBJ
Fig. 4.
Upper panel: rate of events with N pad ≥ / ASM. shows the 6 ◦ × ◦ sky map around the sourceposition in these 3 days, after applying theevent selection. For every 0.1 ◦ × ◦ bin, themap gives the value of the statistical signifi-cance of the excess of events inside the circu-lar window of radius 0.9 ◦ centered on that bin.Donnarumma et al. evaluate a theoretical SEDcurve for the first flare fitting the observationsmade on June 6 from optical up to VHE en-ergies. During the second flare, they reporteda higher photon flux from soft X-rays to 100MeV gamma rays. From these data they pre-dict a flux at E > · − photonscm − s − corresponding to about 7 Crab units,and they model a SED curve with the InverseCompton hump slightly shifted towards higherenergies.Fig. 5 shows the event rate observed byARGO-YBJ (in 18.2 hours of measurement)compared with the rate expected from a sourcespectrum given by the theoretical SED, for anyN pad interval. The agreement is good. A simi-lar analysis is made for the first flare, integrat-ing our data on June 5-7 (17.9 hours of mea-surement). The observed signal has a signifi- Minimum pad multiplicity E ve n t s / d ay June 11-13June 5-7 -3-2-10123-3 -2 -1 0 1 2 3 -3-2-101234deg d e g Fig. 5.
Event rates observed by ARGO-YBJ as afunction of the minimum pad multiplicity on June5-7 and June 11-13 (respectively triangles and cir-cles) compared with expected rates according to theDonnarumma et al. model for the same two periods(respectively dashed and solid lines). The inset rep-resents the sky map around the Mrk421 position onJune 11-13, obtained for events with N pad ≥ ◦ . cance of ∼ ≥ ± · − , (3.55 ± · − and (2.49 ± · − erg − cm s), togetherwith all the measurements in the optical-TeVrange, and the theoretical SED for the twoflares. The measurements appear in fair agree-ment with the expected emission. i Sciascio: Mrk421 with ARGO-YBJ 287 -2 0 2 4 6 8 10 12 14-11.0-10.5-10.0-9.5-9.0-8.5 ARGO-YBJ
VERITASGRIDSuperAGILEBATXRTASMWEBTGASPUVOT period 2period 1 l og [ F () e r g - c m s ] log [ /eV] -10 -9 E F ( E ) e r g c m - s - Fig. 6.
SED measured by ARGO-YBJ on June11-13 (large blue filled circles) together with thedata of other experiments, obtained during the firstflare (open circles), and the second one (filled cir-cles). The curves represent the SEDs modeled in(Donnarumma et al. 2009) for the first flare (dashedline) and the second one (solid line). The insetshows a zoom on the ARGO-YBJ data.
5. Conclusions
In summary, Mrk421 has been continuouslymonitored with ARGO-YBJ during 2008,showing a VHE flux twice the Crab Nebulalevel from day 41 to 180, when the source wasin active phase, and decreasing afterwards.ARGO-YBJ observed a flare on 5-7 June,with a flux about 3.5 Crab units, in agreementwith the VERITAS observation. A second flarehas been discovered with a statistical signif-icance of 4.2 s.d., during the interval 11-13June, with a flux about 7 Crab units.ARGO-YBJ measured the spectra ofMrk421 above 0.8 TeV during the second flarecompleting a multiwavelength campaign fromoptical to TeV energies.For the first time an air shower array wasable to detect gamma-ray flaring activity atsub-TeV energies on a few days period.
Acknowledgements.
We are grateful to the au-thors of (Donnarumma et al. 2009), in particular toMarco Tavani and the AGILE team, for helpful dis-cussions and for providing us full details on theMrk421 broadband data relative to the publishedanalysis.
References
Aielli, G. et al. 2006, NIM A562, 92Alexandreas, D.E. et al. 1992, NIM A311,350Di Sciascio, G. et al. 2007, in Proc.30th Int. Cosmic Ray Conf. (Merida)(arXiv:0710.1945)Di Sciascio G. et al. 2008, in Proc.Vulcano Workshop 2008, Vulcano, Italy(arXiv:0811.0997)Donnarumma, I. et al. 2009, ApJ, 691, L13Fidelis, V.V and Iakubovskyi D.A. 2008,Astron. Rep., 52, 526Fossati, G. et al. 2008, ApJ, 677, 906Gaidos, J.A. et al. 1996, Nature, 383, 319Ghisellini, G. et al. 1998, MNRAS, 301, 451Mucke, A. et al. 2003, Astropart. Phys., 18,593Primack, J.R., Bullock, J.S., & Sommerville,R.S. 2005, AIP Conf. Proc., 745, 23Punch, M. et al. 1992, Nature, 358, 477RXTE / ASM public datahttp: // xte.mit.edu / asmlc //