SLAC-PUB-14862 The 60-month all-sky BAT Survey of AGN and the Anisotropy of Nearby AGN M. Ajello1, D. M. Alexander2, J. Greiner3, G. M. Madejski1, N. Gehrels4 and D. Burlon3 ABSTRACT Surveys above 10keV represent one of the the best resources to provide an unbiased census of the population of Active Galactic Nuclei (AGN). We present the results of 60months of observation of the hard X-ray sky with Swift/BAT. In this timeframe, BAT detected (in the 15–55keV band) 720 sources in an all- sky survey of which 428 are associated with AGN, most of which are nearby. Our sample has negligible incompleteness and statistics a factor of ∼2 larger over similarly complete sets of AGN. Our sample contains (at least) 15 bona-fide Compton-thick AGN and 3 likely candidates. Compton-thick AGN represent a ∼5% of AGN samples detected above 15keV. We use the BAT dataset to refine the determination of the LogN–LogS of AGN which is extremely important, now thatNuSTARpreparesforlaunch,towardsassessingtheAGNcontributiontothe cosmic X-ray background. We show that the LogN–LogS of AGN selected above 10keVisnowestablishedtoa∼10%precision. Wederivetheluminosityfunction of Compton-thick AGN and measure a space density of 7.9+4.1×10−5Mpc−3 for −2.9 objects with a de-absorbed luminosity larger than 2×1042erg s−1. As the BAT AGN are all mostly local, they allow us to investigate the spatial distribution of AGN in the nearby Universe regardless of absorption. We find concentrations of AGN that coincide spatially with the largest congregations of matter in the local (≤85Mpc) Universe. There is some evidence that the fraction of Seyfert 2 objects is larger than average in the direction of these dense regions. Subject headings: cosmology: observations – diffuse radiation – galaxies: active X-rays: diffuse background – surveys 1Kavli Institute for Particle Astrophysics and Cosmology, Department of Physics and SLAC National Accelerator Laboratory, Stanford University, Stanford, CA 94305, USA 2Department of Physics, Durham University, Durham DH1 3LE, UK 3Max Planck Institut fu¨r Extraterrestrische Physik, P.O. Box 1603, 85740, Garching, Germany 4NASA Goddard Space Flight Center, Greenbelt, MD 20771, USA Submitted to Astrophysical Journal Work supported in part by US Department of Energy contract DE-AC02-76SF00515. – 2 – 1. Introduction There is a general consensus that the cosmic X-ray background (CXB), discovered more than40yearsago(Giacconietal.1962),isproducedbyintegratedemissionofActiveGalactic Nuclei(AGN).Indeed, below∼3keVsensitiveobservationswithChandraandXMM-Newton have directly resolved as much as 80% of the CXB into AGN (Worsley et al. 2005; Luo et al. 2011). However, above 5keV, due to the lack of sensitive observations, most of the CXB emission is at present unresolved. Population synthesis models have successfully shown, in the context of the AGN unified theory (Antonucci 1993), that AGN with various level of obscuration and at different redshifts can account for 80–100% of the CXB up to ∼100keV (Comastri et al. 1995; Gilli et al. 2001; Treister & Urry 2005). In order to reproduce the spectral shape and the intensity of the CXB, these models require that Compton-thick AGN (N ≥ 1.4 × 1024cm−2) contribute ∼10% of the total CXB intensity. With such heavy H absorption Compton-thick AGN have necessarily to be numerous, comprising perhaps up to 30–50% of the AGN population in the local Universe (e.g. Risaliti et al. 1999). However, it is still surprising that only a very small fraction of the population of Compton-thick AGN has been uncovered so far (Comastri 2004; Della Ceca et al. 2008a, and references therein). Studies of AGN are best done above 10keV where the nuclear radiation pierces through the torus for all but the largest column densities. Focusing optics like those mounted on NuSTAR and ASTRO-H (respectively, Harrison et al. 2010; Takahashi et al. 2010) will allow us to reach, for the first time, sensitivities ≤10−13erg cm−2 s−1 above 10keV permitting us to resolve a substantial fraction of the CXB emission in this band. Given their small field of views (FOVs) those instruments will need large exposures in order to gather reasonably large AGN samples. Because of their good sensitivity the AGN detected by NuSTAR and ASTRO-H should be at redshift ∼1, but, due to the small area surveyed, very few if any will be at much lower redshift. All-sky surveys, like those performed by Swift/BAT and INTEGRAL above 10keV are very effective in making a census of nearby AGN, thus providing a natural extension to more sensitive (but with a narrower FOV) missions. Here we report on the all-sky sample of AGN detected by BAT in 60months of exposure. Our sample comprises 428 AGN detected in the whole sky and represents a factor of ∼2 improvement in number statistics when compared to previous complete samples (e.g. Burlon et al. 2011). In this paper, we present the sample and refine the determination of the source count distribution and of the luminosity function of AGN. This is especially important considering the upcoming launch of NuSTAR (scheduled for March 2012) as it allows us to make accurate predictions for the expected space densities of distant AGN. We also use the BAT sample to investigate the spatial distribution of AGN in the local Universe. We leave for an upcoming publication the follow-up of all new sources – 3 – using 2–10keV data and the determination of the absorption distribution. This paper is organized as follows: the BAT observations are discussed in § 2, while § 3.1 and § 3.2 discuss respectively the source count distribution and the luminosity function of AGN.In§4wepresentameasurementoftheover-densityofAGNinthelocalUniverse,while in § 5 the prospects for the detection of AGN by NuSTAR are discussed in the framework of the BAT observations and population synthesis models. Finally, § 6 summarizes our findings. Throughout this paper, we assume a standard concordance cosmology (H =71km 0 s−1 Mpc−1, Ω =1-Ω =0.27). M Λ 2. Properties of the Sample The Burst Alert Telescope (BAT; Barthelmy et al. 2005) onboard the Swift satellite (Gehrels et al. 2004), represents a major improvement in sensitivity for imaging of the hard X-ray sky. BAT is a coded mask telescope with a wide field of view (FOV, 120◦×90◦ partially coded) aperture sensitive in the 15–200keV range. Thanks to its wide FOV and its pointingstrategy, BATmonitorscontinuouslyupto80%oftheskyeverydayachieving, after several years, deep exposures across the entire sky. Results of the BAT survey (Markwardt et al. 2005; Ajello et al. 2008a; Tueller et al. 2008) show that BAT reaches a sensitivity of ∼1mCrab1 in 1Ms of exposure. Given its sensitivity and the large exposure already accumulated in the whole sky, BAT is an excellent instrument for studying populations whose emission is faint in hard X-rays. For the analysis presented here we use 60months of Swift/BAT observations taken be- tween March 2005 and March 2010. Data screening and processing was performed according to the recipes presented in Ajello et al. (2008a) and Ajello et al. (2008b). The chosen energy interval is 15–55keV. The all-sky image is obtained as the weighted average of all the shorter observations. The final image shows a Gaussian normal noise and we identified sourcecandidatesasthoseexcesseswithasignal-to-noise(S/N)ratio≥5σ. Thefinalsample comprises720sourcesdetectedall-sky. Identificationoftheseobjectswasperformedbycross- correlating our catalog with the catalogs of Tueller et al. (2008), Cusumano et al. (2010), Voss & Ajello (2010), and Burlon et al. (2011). Whenever available we used the newest optical identifications provided by Masetti et al. (2008), Masetti et al. (2009), and Masetti et al. (2010). Of the 720 all-sky sources only 37 (i.e. ∼5%) do not have a firm identification. This small incompleteness does not change when excluding or including the Galactic plane. Of the 720 objects, 428 are identified with AGN. This represents an improvement of a factor 11mCrab in the 15–55 keV band corresponds to 1.27×10−11erg cm−2 s−1 – 4 – >2 in the number of detected AGN with respect to previous complete samples (e.g. Ajello et al. 2009b; Burlon et al. 2011). Cusumano et al. (2010) recently reported on the sample of sources detected by BAT in 58months of observations. Their catalog is constructed using three energy bands and selecting ≥4.8σ excesses in any of the three bands. As such their catalog is larger than the one presented here. However, for the scope of this and future analyses (e.g. a follow-up work of that presented in Burlon et al. 2011) it is important to have a clean sample whose selection effects are well understood and can be accounted for during the analysis. Fig. 1 shows the sky coverage of the BAT survey. It is apparent that the limiting flux is ∼0.45mCrab (∼5.5×10−12erg cm−2 s−1) and that the BAT survey becomes complete (for the whole sky) for source fluxes ≥1mCrab. The sensitivity scales nicely with the inverse of the square root of the exposure time as testified by the limiting sensitivity of 0.6mCrab reached in 36months of observations (Ajello et al. 2009a). 104 103 2 g e d 102 10 ×10-12 0 2 4 6 8 10 12 14 16 18 flux [erg s-1 cm-2] Fig. 1.— Sky coverage of the BAT survey for the 15–55keV band and for sources detected all-sky above the 5σ level. 2.1. Jet-dominated and Disk-dominated Objects Jet-dominated AGN (radio galaxies and blazars) constitute a ∼15% fraction of the BAT samples (e.g. see Ajello et al. 2009b). This is confirmed also here where 67 (out of Table 1. The 428 AGN detected by BAT a . SWIFTNAME R.A. Decl. PositionError Flux S/N ID Typeb Redshift PhotonIndex LogLX (J2000) (J2000) (arcmin) (10−11 cgs) J0004.2+7018 1.050 70.300 6.551 0.76 5.5 2MASXJ00040192+7019185 AGN 0.0960 2.04±0.46 44.2 J0006.2+2010 1.571 20.168 3.276 1.06 6.5 Mrk335 Sy1 0.0254 2.60±0.32 43.2 J0010.4+1056 2.622 10.947 2.281 1.85 11.0 QSOB0007+107 BLAZAR 0.0893 2.23±0.20 44.6 J0018.9+8135 4.732 81.592 4.720 0.94 6.5 QSOJ0017+8135 BLAZAR 3.3600 2.51±0.52 48.3 J0021.2-1908 5.300 -19.150 4.773 0.92 5.1 1RXSJ002108.1-190950 AGN 0.0950 1.96±0.45 44.3 J0025.0+6826 6.264 68.436 4.721 0.78 5.6 IGRJ00256+6821 Sy2 0.0120 1.66±0.33 42.4 J0033.4+6125 8.351 61.431 4.274 1.01 7.3 IGRJ00335+6126 AGN 0.1050 2.46±0.28 44.5 J0034.6-0423 8.651 -4.400 6.165 0.92 5.2 2MASXJ00343284-0424117 AGN 0.0000 1.79±0.43 ··· J0035.8+5951 8.965 59.852 2.095 2.05 14.8 1ES0033+59.5 BLAZAR 0.0860 2.74±0.18 44.6 J0038.5+2336 9.648 23.600 5.132 1.01 6.2 Mrk344 AGN 0.0240 1.80±0.59 43.1 J0042.8-2332 10.701 -23.548 3.068 2.52 14.7 NGC235A Sy2 0.0222 1.90±0.11 43.4 aThefulltableisavailableintheonlineversionofthepaper. – bAGNaresourceslackinganexactopticalclassification. 5 – – 6 – 104 103 102 1] -s 10 g r e 4 1 4 0 Sy1 - Sy1.5 1 x [10-1 Sy1.8 - Sy2 L RGs and Blazars LINERs 10-2 AGN Compton-thick AGN 10-3 Flux limit 10-4 10-3 10-2 10-1 1 z Fig. 2.— Position on the luminosity-redshift plane of the 428 AGN detected by BAT in the 15–55keV band. The color coding reflects the optical classification reported in Tab. 1. AGN are sources lacking an exact optical classification. The black squares mark the position of the Compton-thick AGN reported in Tab. 2. Note that their luminosities were not corrected for absorption (see text for details). The dashed line shows the flux limit of the BAT survey of 5.5×10−12erg cm−2 s−1. the 428 AGN) are classified as either radio-galaxies or blazars. The remaining 361 AGN are associated with objects optically classified as Seyfert galaxies (323 objects) or with nearby galaxies (38 sources), through the detection of a soft X-ray counterpart, for which an optical classification is not yet available. The full sample is reported in Tab. 1. For all the sources, k−corrected L luminosities were computed according to: X F L = 4πd2 X (1) X L(1+z)2−ΓX where F is the X-ray energy flux in the 15–55keV band and Γ is the photon index. This X X assumes that the source spectra are adequately well described by a power law in the 15– – 7 – 55keV band in agreement with what found by Ajello et al. (2008b), Tueller et al. (2008), and Burlon et al. (2011). Unless noted otherwise (i.e. § 3.3) luminosities are not corrected for absorption along the line of sight since this correction is different than unity (in the 15–55keV band) only for Compton-thick AGN (see Fig. 11 in Burlon et al. 2011) and does not introduce any apparent bias in any of the results shown in the next sections. Fig. 2 shows the position of the 428 sources in the luminosity-redshift plane for the different optical classifications reported in Tab. 1. The BAT AGN sample spans almost 8 decades in luminosity and includes sources detected from z≈0.001 (i.e. ∼4Mpc) up to z≈4. It is also evident that Seyfert galaxies dominate the low-luminosity part of the sample, while blazars and radio-galaxies dominate the high-luminosity part of the sample. The increased exposure of BAT allows us to detect fainter AGN with respect to previous samples. Indeed, theaveragefluxoftheSeyfert-likeAGNdecreasedby∼20%whencomparingittothesample of AGN reported in Burlon et al. (2011). Since the average redshift in the two samples is very similar, this translates into a larger number of low-luminosity AGN. 2.2. Compton-thick AGN Hard X-ray selected samples are among the best resources to uncover Compton-thick AGN which are otherwise difficult to detect. A detailed measurement of the absorbing column density of all the AGN in this sample is beyond the scope of this paper and left for a future publication. However, in order to determine the likely candidates, it is possible to cross-correlate our source list with catalogs of Compton-thick AGN. Our AGN catalog contains all the 9 Compton-thick AGN reported by Burlon et al. (2011) and 6 additional Compton-thick AGN reported in the list of bona-fide objects of Della Ceca et al. (2008a). There 3 additional sources which are labeled as Compton-thick candidates by Della Ceca et al. (2008a) (see their Table 2) which are also detected in this sample. The full list of 18 known Compton-thick AGN contained in this sample is reported in Tab. 2. It is clear that the number of (likely) Compton-thick AGN is doubled with respect to the sample of Burlon et al. (2011) and that Compton-thick AGN represent a ‘steady’ 5% fraction (i.e. ∼18/361) of AGN samples selected above 10keV. The redshift distribution of Compton-thick AGN is also different than that of the whole AGN sample.The median redshift of the Compton-thick AGN of Tab. 2 is 0.010 while that one of the entire AGN sample is 0.029. Compton-thick AGN can be detected by BAT only within a distance of ∼100Mpc beyond which the strong flux suppression caused by the Compton-thick medium limits the capability of BAT to detect these objects. – 8 – We also checked if any of the remaining 3 bona-fide Compton-thick AGN (or the 20 remaining candidates) reported in Della Ceca et al. (2008a) lie just below the reliable BAT detection threshold. None of the remaining sources in the above lists exhibits a significance larger than 3.5σ in our analysis. This means that none of these sources are likely to be detectable by BAT in a deeper survey. The main consequence is however that the new Compton-thick objects that will appear in the BAT samples will be new (i.e. previously un-studied) sources. A few might already be present in this sample and this aspect will be investigated in a follow-up study. – 9 – Table 2. Known Compton-thick AGN detected in the BAT sample. Unless written explicitly the values of the absorbing column density come from Burlon et al. (2011). NAME Type Reshift R.A. Decl. N H (J2000) (J2000) (1024 cm−2) NGC 424 Sy2 0.011588 17.8799 -38.0944 1.99 NGC 1068 Sy2 0.003787 40.7580 -0.0095 >10 NGC 1365a,b Sy1.8 0.005460 53.4442 -36.1292 3.98 CGCG 420-015 Sy2 0.029621 73.3804 4.0600 1.46 SWIFT J0601.9-8636 Sy2 0.006384 91.1972 -86.6245 1.01 Mrk 3 Sy2 0.013509 93.9722 71.0311 1.27e UGC 4203a,c Sy2 0.013501 121.0585 5.1217 >1.00e NGC 3079 Sy2 0.003720 150.4701 55.6978 5.40 NGC 3281 Sy2 0.010674 157.9743 -34.8571 1.96e NGC 3393 Sy2 0.012500 162.1000 -25.1539 4.50 NGC 4939 Sy1 0.010374 196.1000 -10.3000 >10e NGC 4945 Sy2 0.001878 196.3726 -49.4742 2.20e Circinus Galaxy Sy2 0.001447 213.3828 -65.3389 4.30e NGC 5728 Sy2 0.009467 220.6916 -17.2326 1.0 ESO 138-1 Sy2 0.009182 253.0085 -59.2386 1.5e,f NGC 6240 Sy2 0.024480 253.3481 2.3999 1.83 NGC 6552a,d Sy2 0.026550 270.0981 66.6000 >1.00e NGC 7582 Sy2 0.005253 349.6106 -42.3512 1.10 aPart of the sample of candidate Compton-thick objects in Della Ceca et al. (2008a). bNGC 1365 is a complex source that shows a column density that can vary from LogN ≈ 23 to ≥ 24 on timescales of ∼10hr (Risaliti et al. 2009b). Ac- H cording to Risaliti et al. (2009a) the source has an absorber with LogN ≈ 24.6 H which covers ∼80% of the source. cUGC 4203 (also called the ‘Phoenix’ galaxy) is known to exhibit changes in the absorbing column density from the Compton-thin to the Compton-thick regime (see e.g. Risaliti et al. 2010). dReported to be Compton-thick by Reynolds et al. (1994), and Bassani et al. (1999). – 10 – eFor the value of the absorbing column density see Della Ceca et al. (2008a) and references therein. fPiconcelli et al. (2011) reports that this source might be absorbed by LogN ≥25. H