Search for an extended emission around blazars with the MAGIC telescope

Very high energy gamma rays coming from extra-galactic sources can interact with intergalactic radiation fields. This process may result in electromagnetic cascades with the following cycle: the production of electron-positron pairs and then secondary gamma-rays due to inverse Compton scattering. Since electrons and positrons will be scattered in the intergalactic magnetic field, under certain conditions their radiation may be redirected towards the observer. Thus one can anticipate that the secondary gamma-ray emission may produce an apparent extended halo around the source. MAGIC is an Imaging Atmospheric Cerenkov Telescope located on Canary island of La Palma at Roque de los Muchachos Observatory (2200 m.a.s.l). Various source sizes and extended emission profiles within $1^\circ$ diameter have been studied by using dedicated Monte Carlo simulations for the MAGIC telescope. We present results of the study of a possible extended emission for Mrk 421 and Mrk501 done with the MAGIC telescope.

💡 Research Summary

The paper investigates the possible existence of an extended γ‑ray halo around two well‑studied blazars, Mrk 421 and Mrk 501, using data from the MAGIC Imaging Atmospheric Cherenkov Telescope (IACT). The theoretical motivation stems from the interaction of very‑high‑energy (VHE) γ‑rays with the intergalactic background light (EBL) and the cosmic microwave background (CMB). These interactions produce electron‑positron pairs, which in turn up‑scatter low‑energy photons via inverse Compton scattering, generating secondary γ‑rays. If the intergalactic magnetic field (IGMF) is extremely weak (∼10⁻¹² G), the pairs are only slightly deflected, so the secondary photons retain a direction close to that of the primary γ‑ray. Consequently, a cascade can develop over cosmological distances, and a fraction of the secondary photons may be redirected into the observer’s line of sight, forming an apparent extended emission (a “halo”) around the point‑like source.

Detecting such a halo requires excellent angular resolution. MAGIC, with its 17 m reflector and fast photodetectors, provides the lowest energy threshold among current IACTs (∼300 GeV for the analysis presented). The authors introduced a novel method for reconstructing the arrival direction of each shower: a Random Forest (RF) based DISP estimator. Traditional DISP calculations rely on simple Hillas parameters (image length, width, and size) and a proportionality to ellipticity. The RF DISP incorporates both geometric and timing information (e.g., the time gradient along the image major axis) through multidimensional decision trees, achieving a 20‑30 % improvement in the width of the θ² distribution (the squared angular distance between reconstructed and true source positions). This translates into a point‑spread function (PSF) with a 41 % containment radius of about 0.1°, sufficient to resolve halos with radii of a few tenths of a degree.

Monte Carlo simulations were performed to model mixed sources consisting of a point‑like component plus an extended component with flat surface brightness (dN/dθ² = const) and characteristic radii of 0.1°, 0.2°, and 0.3°. The cumulative θ² distributions for these models show that when the extension is smaller than the PSF the distributions are indistinguishable from a pure point source, while larger extensions produce clearly broader θ² profiles.

The observational data set comprises:

• Mrk 501: 26 h of ON data taken in April–May 2008 in ON/OFF mode, with a contemporaneous 50 h OFF sample for background estimation. The source was in a low state (∼15 % of the Crab Nebula flux).
• Mrk 421: 32 h of wobble‑mode data collected between December 2007 and January 2009, with the source offset by 0.4° from the camera centre. The source was in a high state (∼1.4 Crab).

Only events with image size > 400 photo‑electrons were retained, ensuring good angular reconstruction and an effective energy threshold of ∼300 GeV. The θ² distributions for both blazars were compared with those obtained from Crab Nebula data (used as a point‑like reference because its spectral index closely matches those of the blazars). The χ² per degree of freedom values (10.7/11 for Mrk 421 and 3.8/11 for Mrk 501) indicate no statistically significant deviation from a point‑like source.

Upper limits on the flux of any extended component were derived for several assumed halo profiles (flat, and power‑law surface brightness with indices β = 0, ‑1, ‑2) and radii up to 0.3°. At the 95 % confidence level, the extended flux is constrained to be less than 4 % of the Crab flux for Mrk 501 and less than 7 % of the Crab flux for Mrk 421, at the analysis threshold of 300 GeV.

The authors conclude that, with the current single‑telescope configuration and the improved RF DISP reconstruction, no halo is detected around either blazar. However, the method demonstrates a substantial gain in angular resolution, which will be further enhanced in MAGIC’s stereoscopic mode (two telescopes separated by 85 m). The stereoscopic observations are expected to reduce the PSF and improve background rejection, thereby increasing the sensitivity to faint, compact halos. Looking ahead, next‑generation facilities such as the Cherenkov Telescope Array (CTA), with anticipated angular resolutions of ∼2 arc minutes, will be capable of probing much smaller halo extensions and lower flux levels, potentially revealing the elusive intergalactic cascade signatures predicted by theory.