EDP Sciences
Free Access
Issue
A&A
Volume 564, April 2014
Article Number A9
Number of page(s) 10
Section Extragalactic astronomy
DOI https://doi.org/10.1051/0004-6361/201322897
Published online 28 March 2014

© ESO, 2014

1. Introduction

Table 1

AGN detected by H.E.S.S. as of September 2013.

Since the discovery of their extragalactic origin fifty years ago (Schmidt 1963), the class of astrophysical sources called active galactic nuclei (AGN) has been a prime target for astronomers that observe the sky from radio wavelengths to very-high-energy γ rays (VHE, E > 100  GeV). AGN are thought to host super-massive black holes (typical mass of 108−9M) surrounded by an accretion disk, with a fraction of them showing two-sided jets. To unify the various subclasses of AGN, a scheme to sort them based on their orientation with respect to the observer’s line of sight has been proposed since the 1990s (Urry & Padovani 1995). Objects whose jets are closely aligned with the line of sight are called blazars. They fall into two source classes, broad-line-spectrum sources called flat spectrum radio quasars (FSRQs), and BL Lac objects (hereafter BL in tables), which show faint lines or featureless spectra.

Active galactic nuclei, in particular blazars, are the most numerous objects detected at high energy (HE, 100  MeV < E < 100  GeV), where all-sky surveys can be performed with pair-conversion detectors onboard satellites such as the Fermi Large Area Telescope (LAT, Atwood et al. 2009). The Second LAT AGN Catalog, hereafter 2LAC, comprises 886 off-plane (i.e. above a Galactic latitude of 10°) point-like sources associated with AGN that were detected in two years of operation (Ackermann et al. 2011). AGN constitute a third of the sources known at VHE, despite a coverage biased toward Galactic sources. With the fast decrease of fluxes with increasing energy, observations at VHE are mostly performed with ground-based imaging atmospheric Cherenkov telescopes (IACT), which have a field of view (FoV) of a few degrees but an effective area on the order of 105   m2. Their current sensitivity prevents an all-sky scan in a reasonable amount of time, and IACT observations must be pointed to targets of interest (see e.g. Dubus et al. 2013 for a discussion of the capabilities of next generation instruments). Targeted AGN are selected based on their radio and X-ray spectra (Stecker et al. 1996; Perlman 2000; Costamante & Ghisellini 2002) as well as based on their HE flux extrapolated to VHE (Tavecchio et al. 2010).

The High Energy Stereoscopic System (H.E.S.S., Aharonian et al. 2006a) has significantly contributed to the expansion of the class of VHE AGN, with the detection of 23 objects, including 20 discoveries, among 56 known sources of this type as of the end of the year 20131. The H.E.S.S. experiment is located in the Khomas Highland, Namibia (23°1618′′ S, 16°3001′′E) at an altitude of 1800 m above sea level. In its first phase, this experiment was an array of four identical telescopes with cameras composed of 960 photomultipliers and segmented reflectors paving a reflective area of 107 m², for an equivalent diameter of 12 m. Most of the AGN detected with H.E.S.S. belong to the BL Lac class, as shown in Table 1, with the exception of the two nearby radio galaxies of Fanaroff-Riley I type (FR I) Centaurus A and M 87, the FSRQ PKS 1510-089, and the blazar candidate HESS J1943+213, which is located in the Galactic plane. In addition to constraining the radiative processes responsible for the γ-ray emission (for detailed discussions, see, e.g., the references in Table 1), the VHE spectra of these objects can also serve cosmological purposes, as shown with the constraints (Aharonian et al. 2006b; Mazin & Raue 2007) and indirect measurement (Abramowski et al. 2013d) of the extragalactic background light (EBL). With peak intensities in the optical and far-infrared bands, the EBL is composed of the integrated emission of stars and galaxies as well as of the reprocessing of UV-to-optical light by dust. The EBL can hardly be measured directly, although it is the second-most intense diffuse radiation in the Universe after the cosmic microwave background.

During the eight years of the first phase of H.E.S.S., some of the observations did not result in significant excesses at the position of the target or in the FoV of the telescopes. A first set of upper limits (Aharonian et al. 2005b, hereafter HUL1) on 19 AGN observed during 63 hours was published after two years of observation. A second paper (Aharonian et al. 2008b, hereafter HUL2) listed 14 upper limits based on 94 hours of observation spanning 20052007. In this paper, which follows extensive compilation efforts from previous-generation instruments such as Whipple (Horan et al. 2004) or HEGRA (Aharonian et al. 2004), 47 selected candidates are studied, with observations spanning April 2004 to December 2011, for a total live time of approximately 400 h. The candidates and the data selection are presented in Sect. 2, the event and spectral analyses are examined in Sect. 3, and the constraints on the VHE emission are discussed in Sect. 4, together with the target-selection strategy.

2. Selected candidate VHE emitters

The sample of targets consists of the AGN observed with H.E.S.S. until December 2011, for which more than an hour of corrected live time was recorded (see Sect. 3). Only objects located away from the Galactic plane, that is above a Galactic latitude of 10°, were taken into account. Neither datasets on potential or detected H.E.S.S. sources2 are included, nor those where upper limits based on the entire dataset have already been published (HUL1, HUL2). The objects listed in the 2LAC that are located in the same FoV as selected targets and are potentially associated with AGN were also studied. These criteria yield a total of 42 AGN and 5 unknown-type Fermi-LAT or EGRET sources, as listed in Table 2. Pointed observations were performed for 33 objects, while 14 are visible in the FoVs. Of these 47 targets, 39 are studied for the first time with H.E.S.S. in this paper, while eight of them (IIIZw 2, 1ES 0323+022, 3C 120, Pictor A, 1ES 1440+122, RBS 1888, NGC 7469, 1ES 2343-151) have been re-observed since the publication of HUL1 and HUL2.

Table 2

Selected extragalactic objects observed with H.E.S.S. from April 2004 to December 2011.

Table 3

Results from H.E.S.S. observations of 47 AGN.

Table 4

Spectral and temporal analysis of 47 AGN.

The redshifts of the targets were extracted from the Roma-BZCAT catalog Ed. 4.1.1 (Massaro et al. 2009), from the work of Rau et al. (2012) on AGN detected with Fermi-LAT, and from the publication by Pita et al. (2012) about VHE (candidate) emitters. The redshifts of objects not listed in these publications were individually searched for in the literature. A detailed list of references can be found in the last column of Table 2. The distant objects with z > 1 in this table were not directly targeted. 2FGL J0426.6+0509c is located in the same FoV as 3C 120. 2FGL J1959.1-4245 and 2FGL J2219.1+1805 are neighbors of PKS 2004-447 and RBS 1888, respectively. 2FGL J0505.8-0411 and 2FGL J0540.4-5415 were jointly observed with BZB J0543-5532.

The classification of the targets is also primarily based on BZCAT. For objects not listed there, the 2LAC catalog was followed, yielding two AGN of unknown type in addition to BL Lac and FSRQ objects: 1FGL J0506.9-5435 and 2FGL J0537.7-5716, called AGU in Table 2, following the 2LAC nomenclature. Following HUL1 and HUL2, 3C 120 and Pictor A are classified as Fanaroff-Riley I (FR I) and II (FR II) radio galaxies. Searching the SIMBAD database3, Seyfert 1 nuclei (Sey I) are hosted by these two objects and by NGC 7469, while PKS 1345+125 is classified as a Seyfert 2 (Sey II). To summarize, most of the targets are blazars, with 13 FSRQ and 23 BL Lac objects, including PKS 0352-686 and 1FGL J0022.2-1850 recently confirmed as BL Lac objects by Rodriguez et al. (2009) and Shaw et al. (2013), respectively.

Table 5

Comparison of the high-energy extrapolation from the 2LAC with H.E.S.S. upper limits.

thumbnail Fig. 1

Left: distribution of the detection significances, σ, for the 47 candidates using 1σ-wide bins. Error bars indicate the square root of the number of events in each bin, and the black line is a normal distribution of 47 events centered on zero and of unit width. Right: cumulative distribution function of the detection significance and of the normal distribution. The maximum distance between the distributions is shown as a double-headed arrow.

Open with DEXTER

thumbnail Fig. 2

Left: HE γ-ray spectrum and VHE upper limit on the emission of BZB J0543-5532 as measured by Fermi-LAT and H.E.S.S. The EBL-absorbed HE extrapolation based on the 2LAC is shown with a dashed line. Right: HE γ-ray spectrum and VHE upper limit on the emission of 2FGL J0334.3-3728. For this object, a fiducial redshift of 0.3 is assumed in the extrapolation.

Open with DEXTER

3. Analysis and results

The observation conditions and the results of the event analyses are listed in Table 3. The H.E.S.S. telescopes are usually pointed with an offset angle of 0.5−0.7° (wobble mode) when observing extragalactic sources. Higher offset values occur in Table 3 because sources can be in the same FoV as a scheduled target source. The comparably (with other IACTs) large FoV of 5° of H.E.S.S. telescopes allows for reliable spectral reconstruction up to an offset of at least 2° (offset value in Abramowski et al. 2013f), close to the maximum offset values that are listed in Table 3. The observation time, shown in the second column, is corrected for the decrease in acceptance due to an increasing offset from the centre of the cameras. This correction results in a shorter acceptance-corrected live time, as shown in Col. 5.

The data that pass standard quality criteria (good weather, stability of the instrument, as in Aharonian et al. 2006a) were analyzed with Model++ Standard Cuts (de Naurois & Rolland 2009), corresponding to a selection criterion on the image charge of 60 photo electrons. A cross-check was performed with a multivariate analysis described in Ohm et al. (2009). The results of the analysis of the 47 targets described in the following were derived with a single pipeline, associated to the Model ++ analysis. The analysis energy threshold4, shown in Col. 6, depends on the average zenith angle of the observations (Col. 3) and on the offset from the center of the cameras (Col. 4). The number of ON-target (Col. 7) and OFF-target events (Col. 8) was measured above the threshold energy in regions of 0.1° radius. The normalization α of the OFF events, shown in Col. 9, is a relative exposure normalization factor between the ON and OFF regions, within the Reflected background modeling method (Aharonian et al. 2006a; Berge et al. 2007). The excess, defined as ON−α × OFF, and its significance, calculated using Eq. (17) in Li & Ma (1983), are shown in the last two columns of Table 3. No significant deviation from zero is observed, with values in the range [−2.2σ;2.4σ].

The distribution of the detection significance is compared in Fig. 1 with a normal distribution of 47 events, centered on zero and of unitary standard deviation. The deviations of the data distribution from the normal distribution were quantified using a Kolmogorov-Smirnov test. The highest value of the absolute difference between the cumulative probability distributions reaches 0.17, with a p-value for a normal distribution of 12%, equivalent to a 1.5 Gaussian standard deviation. An Anderson-Darling test yields a similar result, with a p-value for a normal distribution of 10%. These tests do not indicate a collective excess of events within the sample of source candidates.

As in HUL1 and HUL2, the spectral analysis was performed assuming a power-law spectrum with photon index Γ = 3, close to values observed for the sources listed for instance in Table 1. Upper limits on the integral fluxes above the threshold energies were computed at the 99.9% confidence level, according to the statistical method of Rolke et al. (2005). The limits shown in Col. 4 of Table 4 were converted into Crab units (C.U., Col. 5) using the power-law spectrum measured by Aharonian et al. (2006a), with a photon index Γ = 2.63 and flux at 1 TeV Φ0 = 3:45×10-11 cm-2 s-1 TeV-1.

A search for variability, one of the characteristic properties of AGN, was performed by fitting a constant function to the flux estimates derived on a night-by-night time scale, as in HUL1 and HUL2. The modified Julian dates of observation for which at least one ON-event is recorded are given in the last column of Table 4 and the χ2 probabilities for a constant fit (with Nnights − 1 degrees of freedom) are shown in Col. 5. With χ2 probabilities higher than 10%, no flaring event is detected on the nightly time scale. A search on shorter time scales is ruled out by the small statistics in each temporal bin, and larger bins would result, for some of the targets, in a number of points that is too small to lead such a study.

4. Discussion

Among the 47 candidates, four blazars have been detected by other IACTs. The BL Lac object RBS 0413 has been discovered by the VERITAS Collaboration (Aliu et al. 2012) at the 1% C.U. level, in agreement with the upper limit of 2.2% C.U. set in this study. 1ES 1218+30.4, detected by the MAGIC (Albert et al. 2006) and VERITAS (Acciari et al. 2009, 2010) Collaborations, is a known variable BL Lac object, with reported fluxes between 6% and 20% C.U. These are on the order of the upper limit of 8.0% measured above the comparably high energy threshold of 1.4 TeV. The VHE flux of the FSRQ 3C 279 has been measured at the 0.5% C.U. level by the MAGIC Collaboration (Aleksić et al. 2011) and is compatible with the 1.2% C.U. derived here. The last detected BL Lac object in the list of targets is 1ES 1440+122, with a flux of 1% C.U. (Benbow et al. 2011) that matches the upper limit derived in this paper.

The upper limits on 3C 120 and NGC 7469 are a factor of two higher than those derived in HUL1, despite a doubled amount of data. This can be related to background fluctuations, with negative detection significances of ~−2σ in HUL1 and ~+ 2σ upward fluctuations observed in this study.

For the other targets, that is 41 among the 47 AGN, the upper limits derived in Table 4 are the strongest reported to date5, with values down to 0.4% C.U. These upper limits are compared with the HE flux reported in the 2LAC, extrapolated above the threshold energy of H.E.S.S., I2LAC(>Eth), without taking into account absorption by the EBL. Since H.E.S.S. observed the sky for a longer period than Fermi-LAT, the 2LAC spectra are not strictly simultaneous with the VHE constraints. The comparison of the Fermi-LAT extrapolated fluxes and of the H.E.S.S. upper limits is thus based on the assumption that the 2LAC spectra are representative of the average HE emission. This assumption is corroborated by 2LAC studies of FSRQs and BL Lac objects that show an average fractional variance of the flux (square root of the normalized excess variance) on the order of 0.55 ± 0.10, that is fluxes that vary on average within ±55%, and also by a rather short duty cycle for high flux events (above 1.5 standard deviation), with a most probable value for the duty cycle on the order of 5% to 10%. The targets for which I2LAC(>Eth) is at least twice as high6 as the H.E.S.S. upper limit are listed in Table 5. Sources detected with other IACTs as well as the distant 2FGL J0537.7-5716 (z = 1.55) are not included in the list.

The extrapolated fluxes of these sources are higher than the H.E.S.S. upper limits, indicating curvature in their spectra. The curvature can have an intrinsic (i.e. related to the underlying emitting particles) and extrinsic (i.e. due to the EBL absorption) origin. To constrain the origin of this curvature, the Fermi-LAT fluxes, φ2LAC(E), were extrapolated taking into account the best-fit EBL model derived with the H.E.S.S. data, corresponding to the optical depth of Franceschini et al. (2008), τFR08(E,z) scaled up by a factor α0 = 1.27 (Abramowski et al. 2013d). The EBL-absorbed extrapolations are thus computed as . Targets for which only an upper limit on the redshift was available were assumed to lie at z = 0.3, roughly corresponding to the peak of the distribution for BL Lac objects in the 2LAC. For 1FGL J0022.2-1850, the EBL-absorbed extrapolation derived using a redshift of 0.77 does not exceed the H.E.S.S. upper limit.

When taking into account the EBL absorption, all but two of the HE extrapolations lie below the H.E.S.S. upper limits, indicating that no intrinsic curvature is required to explain the observed spectral break. BZB J0543-5532 is an exception, with a VHE upper limit a factor of two lower than the EBL-absorbed extrapolation. A straight power-law extrapolation of the intrinsic emission is therefore rejected, suggesting an intrinsic break in the photon spectrum. This curvature is also suggested by the marginal agreement between the H.E.S.S. upper-limit and the high-energy end of the spectrum from the Fermi-LAT Catalog of Sources Above 10 GeV (1FHL, The Fermi-LAT Collaboration 2013). Similar conclusions can be drawn for 2FGL J0334.3-3728, though with smaller statistics from the 1FHL and under the assumption that the object is nearby (z < 0.3). Tighter constraints on the distance of this source and an increased coverage with Fermi-LAT and H.E.S.S. will allow for more definite conclusions on the intrinsic emission of the source.

With the launch of Fermi-LAT, the AGN observation strategy at VHE has partly shifted from a target selection based on radio and X-ray fluxes towards a selection based on extrapolations of HE spectra. It should be noted nonetheless that, based only on the latter criterion, a fourth of the sources listed in Table 1 would not have been discovered. High-frequency-peaked BL Lac objects such as PKS 0548-322, SHBL J001355.9-185406, 1ES 1312-423, 1ES 0229+200, and 1ES 0347-121 are indeed not listed in the 2LAC because of a hard but faint HE emission.

Broadband multiwavelength strategies prove to be of critical importance in such cases. As discussed in Costamante & Ghisellini (2002) and illustrated in Fig. 3, bright TeV BL Lac objects tend to have bright X-ray and radio counterparts. The latter criterion is not sufficient, however, as it tends to discard FSRQs with their low X-ray fluxes (low-energy component peaking in the optical-infrared band) and as X-ray bright objects, such as RBS 421 or 3C 120 studied in this paper, do not necessarily show bright TeV counterparts. Good HE-based candidates do not necessarily cluster in the upper-right corner of Fig. 3 either, as shown by the relatively low radio flux of BZB J0543-5532 and the low X-ray flux of 2FGL J0334.3-3728.

The extension of the population of AGN detected at VHE and the discovery of new types of sources will be a primary task of the future Cherenkov Telescope Array, CTA (Sol et al. 2013; Reimer & Böttcher 2013). A best-suited target selection will account both for multiwavelength information from radio, X-ray, optical, and HE instruments, and for the charting effort led by previous and current-generation VHE instruments.

thumbnail Fig. 3

X-ray flux in the 0.1-2.4 keV band vs radio flux at 1.4 GHz for objects listed in the Roma BZCAT Catalog. 50 of the 56 AGN detected at VHE (as of the end 2013) are listed in the BZCAT with detected X-ray (ROSAT) and radio (NVSS/SUMSS) emission. 25 of the 47 objects studied in this paper are shown, with a selection biased toward X-ray-bright BL Lac objects. Based on ROMA BZCAT and TeVCat.

Open with DEXTER

5. Conclusion

A large sample of AGN has been observed with H.E.S.S. since 2002, resulting in the discovery of more than a third of the known extragalactic VHE emitters. Observations of 47 targets without significant excess were selected and upper limits on their integral fluxes were computed. For 41 of these objects, the upper limits derived in this paper are the strongest to date.

No significant flaring event was detected during the ~400 h of observation of the 47 targets. The straight extrapolation of the HE emission is challenged by the VHE upper limit for a dozen objects. For all but two of them, this spectral curvature can be accounted for by the interaction of γ rays with the EBL.

Active galactic nuclei observations, which are crucial both for the understanding of the EBL and of the objects themselves, will remain a primary goal of H.E.S.S. during its second phase, H.E.S.S. II, where observations at lower energies will increase the number of detected sources and the maximum redshift accessible by Cherenkov telescopes. Extensive campaigns probing the sky down to fractions of percent of the Crab Nebula flux remain a major task of current VHE telescopes. This tremendous effort is paving the way for targeted AGN observations with CTA.


2

A source is considered as detected above a significance of 5σ, while a potential source corresponds to an extrapolated observation time needed to reach detection shorter than 40 h. The list of objects studied in this paper does not depend on the latter criterion within ±10 h.

4

Hereafter, the energy threshold is defined as the energy for which the acceptance reaches 20% of the highest value. This approach, which results in a somewhat lower threshold than the conventional definition (peak of the energy distribution of the events), corresponds to an energy bias lower than the energy resolution (see Figs. 23, 24 in de Naurois & Rolland 2009), which ensures the quality of the reconstructed spectrum.

5

Variations in the energy thresholds of different instruments that observed the same targets were taken into account when comparing upper limits. Values are also reported in Crab units in this paper for the sake of clarity.

6

A fiducial value of two corresponds to the average EBL absorption between 500GeV and 1 TeV for a source situated at z ~ 0.1 (e.g., within the modeling of Franceschini et al. 2008).

Acknowledgments

The support of the Namibian authorities and of the University of Namibia in facilitating the construction and operation of H.E.S.S. is gratefully acknowledged, as is the support of the German Ministry for Education and Research (BMBF), the Max Planck Society, the German Research Foundation (DFG), the French Ministry for Research, the CNRS-IN2P3 and the Astroparticle Interdisciplinary Programme of the CNRS, the UK Science and Technology Facilities Council (STFC), the IPNP of the Charles University, the Czech Science Foundation, the Polish Ministry of Science and Higher Education, the South African Department of Science and Technology and National Research Foundation, and by the University of Namibia. We appreciate the excellent work of the technical support staff in Berlin, Durham, Hamburg, Heidelberg, Palaiseau, Paris, Saclay, and in Namibia in the construction and operation of the equipment. This research has made use of the SIMBAD database, operated at the CDS, Strasbourg, France.

References

All Tables

Table 1

AGN detected by H.E.S.S. as of September 2013.

Table 2

Selected extragalactic objects observed with H.E.S.S. from April 2004 to December 2011.

Table 3

Results from H.E.S.S. observations of 47 AGN.

Table 4

Spectral and temporal analysis of 47 AGN.

Table 5

Comparison of the high-energy extrapolation from the 2LAC with H.E.S.S. upper limits.

All Figures

thumbnail Fig. 1

Left: distribution of the detection significances, σ, for the 47 candidates using 1σ-wide bins. Error bars indicate the square root of the number of events in each bin, and the black line is a normal distribution of 47 events centered on zero and of unit width. Right: cumulative distribution function of the detection significance and of the normal distribution. The maximum distance between the distributions is shown as a double-headed arrow.

Open with DEXTER
In the text
thumbnail Fig. 2

Left: HE γ-ray spectrum and VHE upper limit on the emission of BZB J0543-5532 as measured by Fermi-LAT and H.E.S.S. The EBL-absorbed HE extrapolation based on the 2LAC is shown with a dashed line. Right: HE γ-ray spectrum and VHE upper limit on the emission of 2FGL J0334.3-3728. For this object, a fiducial redshift of 0.3 is assumed in the extrapolation.

Open with DEXTER
In the text
thumbnail Fig. 3

X-ray flux in the 0.1-2.4 keV band vs radio flux at 1.4 GHz for objects listed in the Roma BZCAT Catalog. 50 of the 56 AGN detected at VHE (as of the end 2013) are listed in the BZCAT with detected X-ray (ROSAT) and radio (NVSS/SUMSS) emission. 25 of the 47 objects studied in this paper are shown, with a selection biased toward X-ray-bright BL Lac objects. Based on ROMA BZCAT and TeVCat.

Open with DEXTER
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.