SEXSI X-Ray Analysis and
6.4 X-ray Selected SEXSI Sources Viewed in the Infrared
Our Spitzerdata covers 250 hard-X-ray-selected SEXSI sources; 140 of these have spectro- scopic redshifts. A significant fraction of the sources in the catalog do not have coverage in all of the IRAC bands; our preliminary analysis here will focus on the sources with de- tection in all four IRAC bands. Because of the relative depths of the IR data compared to the X-ray data, close to all of the SEXSI sources have IR counterparts in one or more of the bands. 170 SEXSI sources were observed in all four IRAC bands, and of those 146 have MIPS coverage. Of the 4-band IRAC-observed sources 60% have optical spectroscopic redshifts, and of those 91% are detected in all four IRAC bands, with a corresponding MIPS detection rate of 70%. Overall, over 97% of SEXSI sources observed at 3.6µm are detected, and the detection rate for the other IRAC bands are all above 90%, while the overall MIPS detection rate is 75%.
Table6.2.SEXSI-SpitzerCatalog CXOSEXSI∆α∆δS2−10HRRclasszlog(Lx)log(NH)f3.6f4.5f5.8f8.0f24 [00][00][10−15][µJy][µJy][µJy][µJy][µJy] (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15) J090954.0+541752-0.5-0.85.670.7122.07ELG1.10143.423.5139.2±93.2 J090955.5+5418131.7-1.25.130.42>23.90ELG1.10243.423.1215.6±103.8 J090959.6+5424580.50.54.660.7122.93unid178.2±301.0 J091004.7+5420501.0-1.33.60-0.3921.25unid210.4±88.5 J091008.0+5414010.11.53.73-0.2022.58ELG0.81143.020.816.6±1.717.4±1.89.9±2.4<15.8<158.5 J091008.2+541524-1.11.414.400.5621.94ELG0.67043.422.950.6±5.143.6±5.1143.0±75.2 J091009.7+5415070.00.04.75-0.06>23.90unid<10.0<15.8<10.0<15.8<158.5 J091012.7+5412051.10.215.200.40>23.90ALG0.79243.622.515.6±1.623.2±2.425.8±3.554.0±6.2 J091014.3+541255-1.1-0.260.80-0.3519.24unid149.2±14.9245.7±24.6365.8±36.8639.0±64.1<158.5 J091017.4+5417570.62.14.80-0.7312.79star0.0000.00.05096.1±509.7±1996.6±200.4812.7±125.2 J091018.6+541316-2.4-0.16.46-0.32>23.90unid<10.0<15.8<10.0<15.8341.1±155.9 J091020.7+541848-0.9-0.95.65-0.3319.64BLAGN2.79444.322.443.1±4.377.5±8.1503.3±88.0 J091021.9+5415290.30.12.890.3523.44ELG0.48242.422.218.6±1.926.3±2.738.3±4.339.1±4.9<158.5 J091022.2+542007-0.3-0.55.690.2822.37ELG0.89843.322.8259.7±62.4 J091023.3+541358-0.4-0.514.40-0.5320.68NLAGN0.49943.10.082.9±8.3112.4±11.3128.5±13.1184.5±18.7444.7±133.1 J091023.7+541514-1.0-1.17.920.3225.34unid16.1±1.619.9±2.120.8±2.816.9±3.3<158.5 J091026.8+541241-0.1-0.160.500.0418.78NLAGN0.37343.421.9558.2±55.9620.8±62.2570.7±57.3773.4±77.5 J091027.0+542054-0.2-0.217.50-0.3221.58BLAGN1.63844.321.419.4±2.032.2±3.352.8±5.988.7±9.3268.6±49.9 J091028.5+542320-0.81.41.910.1323.68BLAGN1.96043.523.317.9±1.915.8±15.829.5±5.033.8±5.569.2±53.6 J091028.6+5416340.1-0.11.73>0.7322.91ELG0.90742.724.040.7±4.116.5±3.197.0±54.8 J091028.9+541523-0.40.024.60-0.3822.03BLAGN0.64743.620.041.5±4.251.2±5.263.5±6.678.0±8.4351.5±88.1 J091030.9+541914-0.3-1.07.79-0.3523.88unid54.7±5.5122.0±12.5888.8±120.5 J091031.4+542157-0.1-1.11.590.0123.61unid27.2±2.725.3±2.621.9±2.917.8±2.975.2±40.0 J091031.7+5420240.1-0.12.78-0.2025.61unid10.0±1.013.6±1.511.9±2.112.8±2.979.3±34.8 J091031.9+5423420.4-1.33.270.2825.37unid68.9±6.958.3±5.952.8±6.147.7±5.5<158.5 J091032.9+5412460.10.35.47-0.1719.63ELG0.59542.820.3175.3±17.6116.8±11.793.0±9.8109.8±11.2478.6±263.9 J091033.2+541526-0.20.43.020.60>23.90unid47.2±4.759.2±6.041.5±4.515.8±3.5<158.5 J091034.2+542408-2.2-0.72.41-0.13>23.90ELG0.34042.021.617.9±1.816.3±1.722.4±3.7<15.8110.5±39.7 J091034.5+5419180.50.55.310.45>23.90unid35.5±3.6<10.0<158.5 J091037.8+5415432.32.16.65-0.3623.59unid7.2±0.9134.3±44.2 J091037.9+541608-0.3-0.25.48-0.2622.42ELG1.10243.421.249.9±5.037.5±4.9289.6±67.7 J091038.5+542025-0.1-0.24.45-0.3221.12unid60.9±6.155.3±5.669.9±7.388.3±9.1454.0±70.5 J091039.2+541320-1.2-2.17.27-0.09>23.90unid<10.0<15.8<10.0<15.8218.8±83.0 J091039.8+542032-0.5-1.48.90-0.2921.93BLAGN2.45044.422.4<10.0<15.8<10.0<15.8643.8±92.0 J091040.0+5422590.30.14.370.0120.76ELG0.25241.921.733.8±3.435.1±3.622.5±3.450.4±5.581.0±41.4 J091040.5+542033-0.2-0.91.650.14>23.90unid11.6±1.216.0±1.725.0±3.1<15.8<158.5 J091040.8+542006-0.2-0.42.42>0.7422.38ELG1.09743.123.767.1±6.768.4±6.957.7±6.187.6±9.1639.5±86.2 J091041.4+541945-0.1-0.426.30-0.4021.38BLAGN0.78643.821.494.0±9.469.3±7.074.5±7.883.3±8.7362.4±61.4 J091041.9+5421271.0-0.52.75-0.2621.13BLAGN0.59842.621.749.4±5.036.3±3.746.3±5.037.1±4.4113.0±45.6 J091041.9+542340-0.2-0.610.50-0.3523.02BLAGN1.63844.121.811.9±1.214.8±1.6<10.025.9±3.4<158.5 J091042.6+541618-0.7-0.82.190.40>23.90unid4.6±0.8253.2±60.9 J091042.7+5420080.10.11.670.31>23.90unid41.0±4.135.9±3.721.9±2.929.2±3.7<158.5 J091042.7+542034-0.3-0.41.460.0522.59ELG1.10842.922.653.3±5.351.9±5.239.9±4.467.8±7.2234.5±43.8 J091043.8+541724-0.4-0.44.610.52>23.90unid20.2±2.020.2±2.114.5±2.324.3±3.949.6±34.8
Figure 6.1 shows a color-color diagram of all Spitzer 4-band detected IRAC sources (small black dots) with larger symbols indicating the SEXSI 2 – 10 keV sources. SEXSI sources with optical-spectroscopic classification or continuum-only spectra have symbols indicated by the legend; sources with no optical spectroscopic information are marked with a blue square. The dotted lines indicate the Stern et al.(2005a) AGN selection wedge.
The majority (62%, or 68/109) of the X-ray selected AGN lie in the Stern et al.(2005a) selection wedge. Considering the spectroscopic sample, the AGN wedge contains 86%
(30/35) of the BLAGN, 67% (8/12) of NLAGN, and 28% (8/28) of the ELG, as shown in Figure 6.1. Considering the errors on the IRAC fluxes, all but two of the BLAGN are consistent with the wedge (one of these is the lowest-luminosity BLAGN in the sample, with log(Lx) = 42.6) and all of the NLAGN are consistent with the wedge selection.
6.4.1 X-ray Luminosity Dependence
Figure6.2illustrates the wedge selection for four 2 – 10 keV luminosity ranges, showing that X-ray luminosity strongly affects the fraction of sources which appear as AGN in the mid-IR.
The entire lowest luminosity subsample (Lx <1043erg s−1) lie either outside of the wedge or near the outskirts of the wedge, while the fraction within the wedge increases monotonically with X-ray luminosity. At the highest X-ray luminosities (Lx >1043.5erg s−1), all but two of the sources are consistent with the wedge. Table 6.3 tabulates the fraction of X-ray sources in the wedge for the four plotted X-ray luminosity ranges. The X-ray luminosities used for binning the sources are the intrinsic, absorption-corrected luminosities described in§ 5.5; the results are not significantly different if we use observed luminosities.
These results are consistent with the notion that for low-luminosity (unobscured) AGN, 2 – 10 keV X-ray surveys will find sources missed by infrared selection techniques: for lower-luminosity AGN the mid-IR is less likely to be energetically dominated by the AGN- component than for higher-luminosity AGN. Our results are consistent with those ofDonley et al.(2007), who use deeper X-ray data over a smaller field of view from theChandraDeep Field North and find that a small fraction of low-Lx sources fit their mid-IR power-law AGN selection criterion.
Figure 6.1 Color-color diagram of sources with IR-detections in all four IRAC bands. All 4-bandSpitzersources are indicated by small black dots. Sources with a 2 – 10 keV X-ray de- tection have a larger symbol overlaid; the specific symbol indicates the optical-spectroscopic classification. The dotted lines demarcate the AGN selection wedge introduced by Stern et al. (2005a). The 0 ≤ z ≤ 3 color tracks for two non-evolving galaxy templates from Devriendt et al. (1999) are illustrated; the large filled circles indicate z = 0. A starburst galaxy is illustrated with the track of M82 (dashed line) and NGC 4429, an S0/Sa galaxy with a star-formation rate ∼4000×lower, is indicated with a solid line
Figure 6.2 IRAC color-color diagram of SEXSI 2 – 10 keV sources, split by absorption- corrected intrinsic X-ray luminosity. As the X-ray luminosity grows, a higher fraction of sources fall solidly into theStern et al. (2005a) infrared-AGN selection wedge
Table 6.3. Wedge selection of SEXSI X-ray sources for several luminosity ranges
Lmin< Lx< Lmax
Sample log(Lmin) log(Lmax) Wedge Total %
All classes — 43.0 4 13 31%
BLAGN 1 2 50%
NLAGN 0 1 0%
ELG 3 10 30%
All classes 43.0 43.5 10 19 53%
BLAGN . 4 4 100%
NLAGN 2 2 100%
ELG 4 13 31%
All classes 43.5 44.0 11 18 61%
BLAGN 10 13 77%
NLAGN 1 2 50%
ELG 0 3 0%
All classes 44.0 — 21 25 84%
BLAGN 15 16 94%
NLAGN 5 7 71%
ELG 1 2 50%
6.4.2 Dependence on R-band Magnitude
Figures 6.3 and 6.4 show this same color-color diagram but split at Rcut= 22; Figure 6.3 shows Spitzer and SEXSI sources with bright optical counterparts and Figure 6.4 shows those fainter than Rcut = 22.1 Tabulated results comparing the wedge selection shown in these plots are presented in Table 6.4; the results are similar usingRcut= 22.5.
These data show a few strong trends. Only 5% of optically bright 4-band IRAC sources lie in the wedge area, while 20% of the optically fainter 4-band IRAC sources do. If we consider the subset of these sources that have 2 – 10 keV detections, we see that the 67%
of the optically bright X-ray sources lie in the wedge and 59% of the optically faint X- ray sources are in the wedge. The X-ray sources in general are more likely to fall in the wedge than a typical IRAC source, consistent with the notion that the wedge preferentially identifies AGN. An interesting trend evident from this sample is the small fraction ofR <22 X-ray non-detected wedge sources: of the 40 IRAC sources in the wedge with R < 22, 31 (78%) are X-ray detected. In contrast, the corresponding optically faint (R >22) sample are only detected by Chandra 31% (37/121) of the time. This trend is consistent with
1Note that the total number ofSpitzer-only sources (small black dots) from Figures6.3and6.4is smaller than the number of sources shown in Figure 6.1. This is because Figures 6.3and 6.4require an R-band detection, while Figure 6.1 includes optically undetected sources as well as IRAC sources outside of the optical FOV
Figure 6.3 IRAC color-color diagram of sources with optical counterparts brighter than R= 22. The symbols and galaxy tracks are the same as in Figure 6.1
.
Figure 6.4 IRAC color-color diagram of sources with optical counterparts fainter thanR= 22. The symbols and galaxy tracks are the same as in Figure 6.1
Table 6.4. Wedge selection of IRAC sources and SEXSI X-ray sources split at Rcut= 22
Bright:R < Rcut Faint:R > Rcut
All 4-band IRAC SEXSI X-ray All 4-band IRAC SEXSI X-ray
Sample Rcut Wedge Total % Wedge Total % Wedge Total % Wedge Total %
All 22.0 40 759 5% 31 46 67% 121 596 20% 37 63 59%
BLAGN 22.0 22 25 88% 8 10 80%
NLAGN 22.0 3 4 75% 5 8 71%
ELG 22.0 1 10 10% 7 18 39%
the idea that many of the X-ray non-detectedSpitzer AGN are obscured, resulting in faint optical counterparts and X-ray non-detections.
The final three lines of Table 6.4split the population by spectroscopic source type; the columns for Spitzer sources are blank here because we do not have spectroscopic followup of the non-X-ray-detected Spitzer AGN candidates. The final line of Table 6.4 shows the population of emission-line galaxies, which do not show optical spectroscopic emission lines typical of AGN, but have X-ray luminosities that indicate that a strong active nucleus is present. Of the optically bright ELG, only one of ten sources lie within the wedge (and that one is near the boundary), while closer to 40% of the eighteenR >22 ELG lie in the wedge.
This trend is different from that seen with the BLAGN and NLAGN, where a similar (e.g., similarly high) fraction of sources lie in the wedge, independent ofR magnitude. This may indicate that for optically bright ELG the starburst component of the SED is brighter than the AGN component pushing the source away from the AGN-defined wedge, while for some of the optically fainter sources the AGN component is strong in the infrared leading to wedge selection. This trend is explained by the trend that higher X-ray luminosity sources tend to fall in the wedge: as ELG are identified at higher redshift they will be fainter in R and also have higher average Lx due to selection effects, and thus preferentially fall in the wedge as compared to low-luminosity, low-z sources. Splitting the wedge sources by redshift confirms this assertion. Overall, regardless of R-magnitude, the ELG tend to fall outside of the wedge.