Dusty Quasars at High Redshifts
aa r X i v : . [ a s t r o - ph . GA ] J un Dusty Quasars at High Redshifts
Daniel Weedman and Lusine Sargsyan ABSTRACT
A population of quasars at z ∼ νL ν (7.8 µ m) that includes unobscured, partially obscured, and obscured quasars.Quasars are classified by the ratio νL ν (0.25 µ m)/ νL ν (7.8 µ m) = UV/IR, assumedto measure obscuration of UV luminosity by the dust which produces IR lumi-nosity. Quasar counts at rest frame 7.8 µ m are determined for quasars in theBo¨otes field of the NOAO Deep Wide Field Survey using 24 µ m sources with op-tical redshifts from the AGN and Galaxy Evolution Survey (AGES) or infraredredshifts from the Spitzer
Infrared Spectrograph. Spectral energy distributionsare extended to far infrared wavelengths using observations from the
Herschel
Space Observatory Spectral and Photometric Imaging Receiver (SPIRE), andnew SPIRE photometry is presented for 77 high redshift quasars from the SloanDigital Sky Survey. It is found that unobscured and obscured quasars have simi-lar space densities at rest frame 7.8 µ m, but the ratio L ν (100 µ m)/ L ν (7.8 µ m) isabout three times higher for obscured quasars compared to unobscured, so thatfar infrared or submm discoveries are dominated by obscured quasars. Quasarsource counts for these samples are determined for comparison to the number ofsubmm sources that have been discovered with the SCUBA-2 camera at z ∼ L ν (100 µ m)/ L ν (7.8 µ m) results together with the Bo¨otes 7.8 µ m counts,and we find that only ∼
5% of high redshift submm sources are quasars, includ-ing even the most obscured quasars. Illustrative source counts are predicted toz = 10, and we show that existing SCUBA-2 850 µ m surveys or 2 mm surveyswith the Goddard-IRAM Superconducting 2 Millimeter Observer (GISMO) sur-vey camera should already have detected sources at z ∼
10 if quasar and starburstluminosity functions remain the same from z = 2 until z = 10.
Subject headings: quasars: general— infrared: galaxies — galaxies: active—galaxies: high redshift— galaxies: evolution— galaxies: starburst Cornell Center for Astrophysics and Planetary Science, Cornell University, Ithaca, NY 14853, USA;[email protected]
1. Introduction
As infrared and submillimeter observational capabilities developed over the past twodecades, the census of dusty sources in the extragalactic universe increased dramatically forredshifts z &
2. Such sources are crucial for gaining a full description of formation andevolution for galaxies and quasars in the universe, because optically derived surveys are sub-ject to severe selection effects when the rest frame ultraviolet is affected by dust extinction.High redshift, dusty sources unknown from optical surveys were initially found at z & µ m surveys with the Submillimeter Common User Bolometric Array (SCUBA) camera(Smail, Ivison and Blain 1997; Chapman et al. 2005), then with 24 µ m surveys and follow-up spectroscopy with the Spitzer
Space Telescope (Houck et al. 2005; Yan et al. 2005), morerecently (Eisenhardt et al. 2012) among 12 µ m and 22 µ m sources found by the Wide-FieldInfrared Survey Explorer (WISE) and in far infrared surveys (Casey et al. 2012; Dowell et al.2014) with the Herschel
Space Observatory Spectral and Photometric Imaging Receiver(SPIRE).Based initially on the
Spitzer surveys, a population of ”Dust Obscured Galaxies”(DOGs) was defined (Dey et al. 2008), and a scenario was developed to explain their for-mation and evolution. To summarize simply, the assembly of the earliest massive galaxiesis characterised by extensive dust formation arising in the short lived, initial stellar popula-tions. The remnants of these populations lead to formation of supermassive black holes whichpower luminous active galactic nuclei (AGN) observed as dust obscured quasars. Eventu-ally, radiation pressure from the quasars expels the dust, leading to the optically observablequasars whose apparent luminosity peaks at z ∼ ∼ ∼
10 using submillimeter or millimeter sur-veys. Our motive is to provide comparisons to the rapidly improving sensitivity of submil-limeter and millimeter surveys with SCUBA-2 at 450 µ m and 850 µ m (Geach et al. 2013;Roseboom et al. 2013; Barger et al. 2014) and with the Goddard-IRAM Superconducting2 Millimeter Observer (GISMO) survey camera (Staguhn et al. 2014), together with the 3 –capability of measuring high redshifts of dusty sources with submillimeter interferometers.Already, for example, the [CII] 158 µ m emission line has been measured in sources with 4 < z < ∼ νL ν (0.25 µ m)/ νL ν (7.8 µ m) = UV/IR. Al-though this is an observational classification independent of the interpretion, we describethe ratio as a measure of ultraviolet obscuration by dust and use it to define the three cate-gories of quasars. For all categories, luminosity functions and source counts are normalizedto the dust continuum luminosity νL ν (7.8 µ m) at rest frame 7.8 µ m to minimize effects ofextinction for optically obscured quasars. This particular wavelength is used because it is alocalized spectral maximum for quasars heavily absorbed by the 9.7 µ m silicate feature andallows a uniform comparison between obscured quasars with large extinction and silicateabsorption, and the unobscured optical quasar samples with little extinction and silicateemission. Among AGN, this measure of dust luminosity correlates well with hard X-rayluminosity, black hole mass, and high ionization emission line luminosity (Weedman et al.2012). The obscured quasars with 9.7 µ m silicate absorption, unknown from optical surveys,are those discovered in surveys with the Spitzer
Infrared Spectrograph (IRS, Houck et al.2004).To predict detections at longer wavelengths, we determine the far infrared spectralenergy distributions (SEDs) of obscured and unobscured quasars using observations withSPIRE. This includes previously published results for obscured quasars together with our ownnew SPIRE photometry of 77 unobscured quasars from the quasar catalog (Schneider et al.2010) of the Sloan Digital Digital Sky Survey (SDSS, Gunn et al. 1998).These empirical results for dusty quasars are used to produce quasar source counts forcomparison to the number of sources that have been discovered with SCUBA-2 at z ∼
2. Toillustrate an example of future discovery possibilities, we determine the number of quasarsthat should be seen for 9.5 < z < >
2. Eventual comparison of this prediction with observationswill allow a measure of whether the formation rate of luminous quasars and the mix ofquasars and starbursts changed between 2 . z .
10. We determine luminosities throughoutusing H = 74 km s − Mpc − (Riess et al. 2011), Ω M =0.27, and Ω Λ =0.73. 4 –
2. Dusty Quasar Populations
Quasar surveys at optical wavelengths naturally favor those quasars which are luminousin the rest frame ultraviolet and which have broad emission lines for classification and redshiftmeasurement. These ”type 1” quasars dominate classical samples (Carswell and Smith 1978;Lewis et al. 1979; Osmer 1982; Schmidt and Green 1983; Marshall et al. 1984; Boyle et al.1988) and extensive recent surveys such as the SDSS quasar catalog (Schneider et al. 2010)and the AGN and Galaxy Evolution Survey (AGES, Kochanek et al. 2012). Type 1quasars are presumed to show unobscured quasars whose intrinsic ultraviolet and emissionline luminosities are not affected by dust extinction. By contrast, extensive observationalstudies of type 2 quasars (Willott et al. 2000; Alexander et al. 2003; Zakamska et al. 2004;Martinez-Sansigre et al. 2006; Hickox et al. 2007) are interpreted as showing partially ob-scured quasars, in which the broad line region and intrinsic ultraviolet continuum are notobserved. These interpretations arise as an extension of the Seyfert 1 and Seyfert 2 ac-tive galactic nucleus (AGN) classifications, originally defined spectroscopically based on thepresence or absence of broad hydrogen emission lines (Khachikian and Weedman 1974) andsubsequently interpreted within the ”unified theory” as arising from orientation effects thatcould obscure the broad line region (Antonucci 1993).Our goal is the definition of a quasar population that is not biased by extinction effects,ranging from unobscured quasars through the DOGs population. To achieve this, we clas-sify quasars quantitatively based on the ultraviolet to infrared luminosity ratio. FollowingVardanyan et al. (2014), we use the rest frame ratio UV/IR = νL ν (0.25 µ m)/ νL ν (7.8 µ m).This parameter is chosen because of reasons given earlier for νL ν (7.8 µ m), and because νL ν (0.25 µ m) is determined spectroscopically for SDSS quasars (Shen et al. 2011). Cate-gories are chosen that cover UV/IR for all quasars, and we assume based on previous workthat this ratio is controlled primarily by the amount of extinction that suppresses νL ν (0.25 µ m). In the following discussions, we group quasars into three categories based on empiricaldeterminations of UV/IR: obscured quasars with log UV/IR < -1.8, partially obscured with-1.8 < log UV/IR < > ∼ µ m silicate feature also corre-lates well with obscured and unobscured classifications and the UV/IR ratio. The presenceof silicate absorption means there must be cooler dust between the observer and the hotterdust responsible for the infrared continuum; sources with the smallest values of UV/IR, the 5 –DOGS, have measurable redshifts only because of strong silicate absorption. Observing sil-icate emission means that the hotter side of the clouds is directly observed, implying littleextinction. This interpretation is consistent with observations of silicate strengths and thecorrelation with type 1 and type 2 AGN classifications (e.g. Hao et al. 2005; Imanishi et al.2007; Hao et al. 2007; Weedman et al. 2012) and with dusty torus models (Shi et al. 2006;Ramos-Almeida et al. 2011; Efstathiou et al. 2014).DOGs have also been extensively studied in Spitzer photometric surveys at variouswavelengths and large samples of obscured and unobscured quasars were defined using colorsfrom the
Spitzer
Infrared Array Camera (IRAC; Fazio et al. 2004). A comprehensivesummary of the history and definition of these photometric samples is in Chen et al. (2015).Sources with power law continua extending through the IRAC bands are interpreted as AGN(Brand et al. 2006; Donley et al. 2007; Bussmann et al. 2009; Melbourne et al. 2012) andsometimes called ”power law DOGS”. These are contrasted to sources having a photometricpeak within the IRAC bands, sometimes called ”bump” DOGS, which is interpreted as arisingfrom the rest frame 1.8 µ m absorption in stellar atmospheres (Simpson and Eisenhardt1999). The AGN DOGs overlap in characteristics with many of the Compton thick, obscuredX-ray sources (Brand et al. 2008; Polletta et al. 2008; Fiore et al. 2008; Bauer et al. 2010).As verified below, the AGN DOGS generally show the 9.7 µ m silicate absorption featurewhen IRS spectra are available. Conversely, sources chosen photometrically from Spitzer surveys as ”bump” sources consistently show PAH features in IRS spectra (Weedman et al.2006c; Farrah et al. 2008; Desai et al. 2009; Fiolet et al. 2010).All individual quasars or AGN which are discussed in this paper are summarized inFigure 1 showing dust luminosities νL ν (7.8 µ m) to illustrate the range of redshifts and mid-infrared dust luminosities encompassed in our analysis. Sources in this Figure are classifiedbased on the spectroscopic silicate criterion, so unobscured quasars are those with silicateemission and obscured quasars are those with silicate absorption. The systematic differencesin νL ν (7.8 µ m) between the two samples of high redshift quasars (SDSS/WISE unobscured, Spitzer
IRS obscured) arise primarily from differences in survey areas. The SDSS/WISEsample covers a large sky area of > , whereas the obscured quasars arise onlywithin ∼
10 deg , so the smaller area survey does not reach the rare but more luminoussources within the larger survey. This figure also shows that the highest infrared luminositiescontinue to the highest redshifts observed, with no turndown at any redshift yet found, aresult described in more detail for SDSS/WISE quasars in Vardanyan et al. (2014). 6 –Fig. 1.— The νL ν (7.8 µ m) (erg s − ) distribution with redshift of individual sources used inthis paper for determining ratios of far infrared to νL ν (7.8 µ m) luminosity. Triangles arelow redshift silicate emission AGN with far infrared luminosities from IRAS, asterisks arelow redshift silicate absorption AGN with far infrared luminosities from IRAS, diamonds arehigh redshift SDSS/WISE quasars with new SPIRE photometry in Table 2, and squares arehigh redshift silicate absorption quasars discovered by Spitzer
IRS having published SPIREphotometry (Melbourne et al. 2012; Sajina et al. 2012). For SDSS/WISE detections, an em-pirical IRS template is used to transform observed frame f ν (22 µ m) to rest frame νL ν (7.8 µ m). As explained in the text, the total infrared dust luminosity L IR is empirically de-termined from νL ν (7.8 µ m) as log [ L IR / νL ν (7.8 µ m)] = 0.51 in low redshift AGN withsilicate emission, log [ L IR / νL ν (7.8 µ m)] = 0.80 in low redshift AGN with silicate absorp-tion, log [ L IR / νL ν (7.8 µ m)] = 0.41 for the high redshift silicate emission quasars, and log[ L IR / νL ν (7.8 µ m)] = 0.68 for the high redshift silicate absorption quasars. 7 – The original definition of DOGs in Dey et al. (2008) assumed that small UV/IR ratiosarise because of dust extinction, a conclusion based primarily on the presence of silicate ab-sorption in the original DOG quasar samples as proof of intervening dust. Subsequent studyconfirmed that extinction was indeed the best explanation for the general DOG population,rather than intrinsic differences in SEDs (Penner et al. 2012). These obscured DOG quasarswere first found using
Spitzer
IRS spectroscopy (Houck et al. 2005; Yan et al. 2005) thatdiscovered optically faint quasars having redshifts measureable only from the 9.7 µ m silicateabsorption feature. Subsequently, the DOGs were defined by Dey et al. as having observedinfrared to optical flux density ratios f ν (24 µ m)/ f ν ( R ) > R - [24] >
14 (Vega mag-nitudes). Adopting that a [24] magnitude of zero corresponds to 7.3 Jy, the DOG definitionmeans that any source having f ν (24 µ m) > R >
R >
24. We define these magnitudes as”optically faint”.The obscured quasars defining the DOGs were initially found among sources identifiedin 24 µ m surveys using the Spitzer
MIPS instrument (Rieke et al. 2004), primarily of theBo¨otes field of the NOAO Deep Wide Field Survey (NDWFS, Jannuzi and Dey 1999) andthe
Spitzer
First Look Survey (FLS, Fadda et al. 2006). To assemble our present sum-mary of obscured quasars, we have utilized all sources with IRS spectra within the Bo¨otesfield (Houck et al. 2005; Weedman et al. 2006; Melbourne et al. 2012) and the FLS field(Yan et al. 2007; Sajina et al. 2007; Dasyra et al. 2009; Weedman et al. 2006b) meeting thephotometric criteria of f ν (24 µ m) > R >
24. Although redshifts and spectroscopicidentifications of the silicate absorption feature had previously been identified in most cases,we reexamined all spectra using the improved spectral extractions in the CASSIS spectralatlas (Lebouteiller et al. 2011) . We also measured rest frame f ν (7.8 µ m) in all sources fromthe CASSIS spectra.Our total sample of obscured quasars is given in Table 1, with the Bo¨otes sourcesreproduced from Vardanyan et al. (2014). Because the Spitzer quasars derive initially from24 µ m surveys, there is a strong redshift selection for silicate absorption sources when the 7.8 µ m continuum peak is near 24 µ m in the observed frame. To accommodate this selection,we describe luminosity functions and source counts only within the range 1.8 < z < > http://cassis.sirtf.com. The Cornell Atlas of Spitzer IRS Spectra (CASSIS) is a product of the InfraredScience Center at Cornell University. > Spitzer
IRS given in Table1). Bottom spectrum is average observed rest frame spectrum of 65 silicate absorbed lowredshift AGN from Sargsyan et al. (2011) used as comparisons for far infrared SEDs. Spectraare normalized to peak f ν (7.8 µ m) and displaced by 0.5 units of f ν . 9 –Table 1. Published photometry from Herschel
SPIRE is also included in this Table for thediscussion of far infrared SEDs which follows below.Other than the DOG criterion, the most important selection to be applied is to assurethat we identify DOGs which are powered by the AGN of a quasar, without having a sig-nificant contribution to dust luminosity from a starburst component. We base this decisionalso on an IRS spectroscopic criterion. Many studies have shown that the strength of thepolyclyclic aromatic hydrocarbon (PAH) features with rest frame wavelengths 6 µ m < λ < µ m are a measure of the starburst component (e.g. Genzel et al. 1998; Laurent et al. 2000;Brandl et al. 2006; Desai et al. 2007; Veilleux et al. 2009; Tommasin et al. 2010; Wu et al.2010; Sargsyan et al. 2011; Stierwalt et al. 2013). For high redshift sources, the PAH featureused is at 6.2 µ m. A classification generally adopted is that any source with rest frameEW(6.2 µ m) < µ m is dominated by AGN luminosity.Because of the poor S/N of many sources in our present study, it is not realistic toapply a rigorous criterion for EW(6.2 µ m) to each source to determine the classification, butno source included in Table 1 has a measurable 6.2 µ m feature that exceeds the spectralnoise, and the upper limits are significantly smaller than 0.1 µ m. This is illustrated by theaverage spectrum of all sources in Table 1, shown in Figure 2, for which the EW(6.2 µ m)= 0.018 µ m. This small EW is evidence that the sample is indeed dominated by “pure”AGN. The average spectrum also illustrates the 7.8 µ m peak flux density that is measuredand shows for comparison the low redshift, silicate absorption AGN used as local analogues.The distinctive difference between an obscured quasar with silicate absorption and a sourcewith PAH emission is illustrated below in section 5.1. We note also that 26 of the Bo¨otessources in Table 1 which define our sample of obscured quasars are photometrically classifiedby Melbourne et al. (2012), and 23 of 26 are ”power law” DOGS, with only 3 classed as”bump” sources.The limiting UV/IR for obscured quasars cannot be determined using monochromaticwavelengths because rest frame ultraviolet flux densities are measured only with broad band R and I filters, and which filter is closer to rest frame 0.25 µ m depends on redshift; effectivewavelengths are ∼ µ m and 0.80 µ m . For the redshift interval 1.8 < z < µ m is 0.7 µ m < λ < µ m. The magnitudesof the brightest IRS obscured quasars are ∼
24 in either filter, corresponding to 0.77 or 0.61 µ Jy for R or I . Taking the average as representing the brightest obscured quasar ( R =24) and comparing to the faintest f ν (7.8 µ m) in Table 1 ( ∼ νL ν (0.25 µ m)/ νL ν (7.8 µ m)] < -1.8. All obscured quasars have values of UV/IR smallerthan this. 10 – The largest sample of quasars that are unobscured, optically bright, and having measur-able νL ν (7.8 µ m) are those within the SDSS (Schneider et al. 2010). Their dust luminosities νL ν (7.8 µ m) can be determined using the WISE 22 µ m photometry for SDSS quasars to-gether with a template spectrum to transform observed frame 22 µ m to rest frame 7.8 µ m (Weedman et al. 2012; Vardanyan et al. 2014). The template is determined from IRSspectra of SDSS quasars (Deo et al. 2011) which are characterized by silicate emission, indi-cating that these quasars are unobscured. The νL ν (0.25 µ m) of SDSS quasars are tabulated(Shen et al. 2011), so UV/IR can be determined. The results in Vardanyan et al. show thatall SDSS quasars in the redshift interval we use have log UV/IR > νL ν (7.8 µ m) thanSDSS because AGES utilizes the Bo¨otes photometry going to 0.3 mJy at 24 µ m whereas theSDSS/WISE quasars only reach 2 mJy at 22 µ m . The partially obscured quasars are, by definition, intermediate between the obscuredDOGs and the unobscured type 1 quasars defined above. This defines their -1.8 < log UV/IR < R = 21, but type 2 has median R = 23, and the faint limits ofthe two samples are also shifted by ∼ f ν (24 µ m) amongtype 1 and type 2, this result demonstrates that median log UV/IR for the type 2 samples issystematically smaller by about 0.8 than for type 1, so that the type 2 represent a partiallyobscured sample with representative log UV/IR ∼ -0.6.
3. Far Infrared Luminosities of Dusty Quasars
A major purpose of this paper is to determine the source counts expected at submil-limeter and millimeter wavelengths for the full quasar population including all classes of ob- 11 –scuration. These results are enabled by observations of quasars with the SPIRE instrument(Griffin et al. 2010) on the
Herschel
Space Observatory (Pilbratt et al. 2010) to determinefar infrared luminosities. In this section, we compare results for unobscured quasars usingnew observations to results for obscured quasars using previously published results in theBo¨otes and FLS fields.Although we are determining the far infrared luminosities of sources classified spec-troscopically as quasars, we note the extensive previous studies summarized in Chen et al.(2015) that attribute the far infrared luminosity from many AGN and quasars to dust reradi-ation from a starburst component. These conclusions rely on SED template libraries whichshow that starbursts have stronger far infrared luminosity than AGN (Chary and Elbaz2001; Dale and Helou 2002; Assef et al. 2010; Elbaz et al. 2011; Wardlow et al. 2011). Theclassification of infrared SEDs in these libraries arises primarily from the PAH features orfrom emission line ratios that correlate with PAH features. For these reasons, using theabsence of PAH features to define obscured AGN or quasars is consistent with the approachof previous studies based on SEDs. For sources without PAH features, there is no spectro-scopic evidence to attribute the far infrared luminosity to a starburst. More detailed effortsto deconvolve starburst and AGN far infrared components based on SEDs also show good cor-relations between PAH strength and the starburst luminosity component (Feltre et al. 2013;Hill et al. 2014). In either case, it could never be proven based on any criterion whether asource is starburst or AGN if sources are so obscured that all spectroscopic indicators fromemission line ratios or PAH strengths are hidden (Polletta et al. 2008; Farrah et al. 2007;Desai et al. 2007; Kirkpatrick et al. 2012; Leipski et al. 2014).
To assemble far infrared luminosities of luminous, unobscured quasars at high redshifts,we selected sources from the SDSS quasar catalog for new observations with SPIRE photom-etry. Our selection of unobscured sources proposed for SPIRE cycle 2 observations was madeby comparing the SDSS quasar catalog (Schneider et al. 2010) with the first WISE data re-lease (Wright et al. 2010) using a criterion . ′′ for source identification. This resulted in9424 SDSS/WISE quasars detected at 22 µ m . The observed fluxes were scaled to f ν (restframe 7.8 µ m) by using SDSS redshifts combined with an empirical spectral template wedetermined using IRS spectra of type 1 AGN and SDSS quasars; this template is illustratedand defined in Weedman et al. (2012) and Vardanyan et al. (2014). Initially, we chose themost infrared luminous 25 quasars in each redshift interval of 0.5 for 1.5 < z < Herschel mission ended.Our observations of individual sources were made with the SPIRE small map mode .Photometry was analyzed using the Herschel
Interactive Processing Environment (HIPE)v11.1.0 and the SPIRE Small Map Mode User Reprocessing Script . Photometry of sourceswas done with SUSSExtractor point source extraction (Savage and Oliver 2007). Thesetechniques were used so that our results were derived in similar fashion to those of HerMES,which we use below for comparison to obscured quasars already published. The signal tonoise (S/N) threshold was set at 3, and full width half maximum (FWHM) for the pointspread function (PSF) were taken as 18.2 ′′ , 24.9 ′′ and 36.3 ′′ for 250 µ m, 350 µ m, and 500 µ m. A detection is assumed to be real if the distance between the SPIRE source and theSDSS coordinate is < ′′ for 250 µ m, < ′′ for 350 µ m, and < ′′ for 500 µ m (i.e. distancesof one FWHM of the PSF). If no source is detected within these criteria, an upper limit of25 mJy is assumed at all wavelengths. Results of the new photometry are in Table 2. Thephotometry we report gives the measured fluxes of sources at the positions listed. We applyno statistical corrections for faint, underlying background sources that might artifically boostthe observed fluxes (e.g. B´ethermin et al. 2012), because the flux limits we use exceed by 4 σ the background confusion noise (Nguyen et al. 2010). Far Infrared luminosities of obscured quasars from the Bo¨otes and FLS survey fieldsare given in Table 1. These were determined by Melbourne et al. (2012) for Bo¨otes andSajina et al. (2012) for FLS using the
Herschel
Multi-tiered Extragalactic Survey withSPIRE (HerMES; Oliver et al. 2010).Our SEDs are all normalized to L ν (7.8 µ m), because this wavelength defines the lumi-nosity functions and quasar counts we use. Figures 3 and 4 show the SEDs of the unobscuredand obscured quasar samples, as determined by the SPIRE observations. For comparisonwith the luminous, high redshift quasars, we also include low redshift AGN which have bothIRS spectra for classification and measures of f ν (7.8 µ m) as well as photometry with theInfrared Astronomical Satellite (IRAS) to give fluxes at ∼ µ m . These AGN are listedin Sargsyan et al. (2011).One of our goals for the SPIRE observations is to test the simple expectations of the http://herschel.esac.esa.int/Docs/SPIRE/html/spire-handbook.html http://herschel.esac.esa.int/hipe/
13 –unified model, explaining obscured and unobscured quasars as differing only in orientationof a dusty torus that can obscure ultraviolet luminosity. In this interpretation, the infraredSEDs should not show differences between obscured and unobscured quasars if extinctiondoes not affect the infrared. Detailed considerations of radiative transfer effects allow thatthe overall dust content, or covering factors, may be intrinsically different, however, such thatthe most obscured sources have larger covering factors and are not systematically obscuredonly because of orientation (Levenson et al. 2007; Thompson et al. 2009; Elitzur 2012).The comparison of our results between Figure 3 and Figure 4 show obvious differencesin the SEDs of unobscured quasars compared to obscured quasars. Compared to unobscuredsources in Figure 3, the obscured sources in Figure 4 have more far infrared luminosity (at ∼ µ m) relative to the mid infrared 7.8 µ m luminosity where the SEDs are normalized.This is true also for the lower luminosity AGN. How do we interpret this result?Consider first the difference in L ν (100 µ m)/ L ν (7.8 µ m) between unobscured quasars(Figure 3) and obscured quasars (Figure 4). The median ratio log L ν (100 µ m)/ L ν (7.8 µ m)for unobscured quasars is 0.6 but is 1.05 for obscured quasars, a difference of about 2 σ compared to the dispersions within the ratios for each class. This means the unobscuredquasars appear to have a smaller fraction of cool dust (seen in the far infrared) comparedto hot dust; if this difference is intrinsic rather than an orientation effect, it means thatthere are real differences in the dust distribution between obscured and unobscured quasars.The observed differences could be explained as arising only from orientation, however, ifobscuration is so great that obscured quasars suffer extinction of the continuum at 7.8 µ m compared to 100 µ m .The summary by Draine (1989) shows that the extinction at 7.8 µ m from silicate absorp-tion alone is about 20% of extinction at the peak of 9.7 µ m silicate absorption, as measuredin magnitudes. The average spectrum of the obscured quasars used for our sample (Figure 2)has a silicate feature that absorbs about 50% of the continuum, corresponding to extinctionof 0.75 mag at peak extinction. This implies extinction of ∼ µ m.The difference in log L ν (100 µ m)/ L ν (7.8 µ m) between unobscured and obscured quasars inFigures 3 and 4 is 0.5, or 1.3 magnitudes, which is much larger that the estimated extinctionof 0.15 mag.This result implies that extinction effects arising from orientation do not explain thedifferences between unobscured and obscured. However, silicates are not the only source ofextinction for the continuum. The prototype highly absorbed AGN is IRAS F00183-7111for which Spoon et al. (2004) illustrate various other absorption features near 7.8 µ m . Thefeatures are normalized to the observed local continuum at 7.8 µ m without any estimatesof the actual extinction at 7.8 µ m, so it is feasible that these absorptions from ices and 14 –hydrocarbons suppress the 7.8 µ m continuum by the additional ∼ one mag. needed to ex-plain the differences in L ν (100 µ m)/ L ν (7.8 µ m) ratios. In this case, orientation effects alonecould explain why obscured quasars appear to have relatively higher far infrared luminosi-ties. In sum, we cannot confidently conclude as yet that the differences in SEDs betweenunobscured and obscured quasars are caused by any effect other than differential extinctionat 7.8 µ m compared to 100 µ m, because of the complex absorptions near 7.8 µ m .There also are differences in both unobscured and obscured samples in log L ν (100 µ m)/ L ν (7.8 µ m) for luminous quasars compared to local AGN. Figure 3 shows that thehigh luminosity, unobscured quasars have log L ν (100 µ m)/ L ν (7.8 µ m) that is smaller by 0.5than the median ratio for lower luminosity AGN. For the obscured quasars and AGN, thedifference between high luminosity quasars and lower luminosity AGN is 0.25. These differ-ences can be attributed primarily to selection effects, because the quasars in both sampleswere selected based on brightnesses near rest frame 7.8 µ m, so their selection favors sourceshaving smaller L ν (100 µ m)/ L ν (7.8 µ m) compared to local AGN whose selection was notbiased by 7.8 µ m luminosities.An alternative interpretation of the systematic differences between obscured and unob-scured quasar samples might invoke luminosity dependence in the L ν (100 µ m)/ L ν (7.8 µ m)ratio, because the SDSS/WISE unobscured quasars are systematically more luminous bya factor of ∼
10 than the Bo¨otes obscured quasars (Figure 1). We rule out this interpre-tation because it is not evident among the AGN. These have similar luminosities betweenobscuration categories (Figure 1), but comparison of Figures 3 and 4 shows that the L ν (100 µ m)/ L ν (7.8 µ m) ratio is a factor of 2.2 larger for the obscured AGN compared to unobscured.This is similar to the factor of 2.8 for the difference between obscured and unobscured quasarsat much higher luminosities.The dispersions in the L ν (100 µ m)/ L ν (7.8 µ m) ratios in Figures 3 and 4 are a measureof the intrinsic variations in the ratio of cool dust to hot dust within sources. Such variationscan arise for many reasons, but it is useful to measure the extent of these variations. Toestimate intrinsic dispersions, the dispersions produced by observational uncertainties inmeasures of both f ν (100 µ m) and f ν (7.8 µ m) need to be removed.For the unobscured SDSS/WISE quasars in Figure 3, the uncertainties noted in Table 2include ±
15% for WISE 22 µ m fluxes and ±
25% for typical SPIRE fluxes. In addition, thereis additional uncertainty of ∼ ±
15% at typical redshifts in using the template that trans-forms observed frame f ν (22 µ m) to rest frame f ν (7.8 µ m), as described in Vardanyan et al.(2014). Adding these uncertainties quadratically leads to an overall dispersion expected fromobservational uncertainty alone of ± σ range of a factor of 3, or ± L ν / L ν (7.8 µ m) for unobscured AGN and quasars. Triangles arelow redshift silicate emission AGN in Sargsyan et al. (2011) with far infrared luminositiesfrom IRAS photometry and L ν (7.8 µ m) from IRS spectra. Diamonds are high redshiftSDSS/WISE quasars with new SPIRE photometry in Table 2, and circles are these highredshift quasars with upper limits. Large cross is the median and one sigma dispersion withinrest wavelength range 80 µ m to 110 µ m for the high redshift quasars, including limits. Thincurve is the median for silicate emission AGN; thick curve is the most luminous ULIRGSED from Herschel photometry in Symeonidis et al. (2013), normalized at 100 µ m to theobserved median of the SDSS/WISE quasars. Long, thick vertical line is the rest wavelengthfor source with z = 2.1 at observed frame 850 µ m (SCUBA-2); short, thick vertical line isthe rest wavelength for source with z = 2.1 at observed frame 450 µ m (SCUBA-2). Long,thin vertical line is the rest wavelength for source with z = 10 at observed frame 850 µ m;short, thin vertical line is rest wavelength for source with z = 10 at observed frame 2 mm(GISMO). Luminosity ratios at these rest wavelengths are taken from thick curve. Error baris observational uncertainty in ratio log L ν / L ν (7.8 µ m) for individual SDSS/WISE quasars. 16 –Fig. 4.— Observed results for L ν Lv (100 um ) /Lv (7 . um )/ L ν (7.8 µ m) for obscured AGN andquasars. Asterisks are low redshift silicate absorption AGN in Sargsyan et al. (2011) with farinfrared luminosities from IRAS photometry and L ν (7.8 µ m) from IRS spectra. Squares arehigh redshift obscured quasars discovered by Spitzer
IRS in Table 1, and circles are quasarswith limits. Large cross is the median and one sigma dispersion within rest wavelength range80 µ m to 110 µ m for the high redshift quasars, including limits. Thin curve is the medianfor silicate absorption AGN; thick curve is the most luminous ULIRG SED from Herschelphotometry in Symeonidis et al. (2013), normalized at 100 µ m to the observed median ofthe high redshift quasars. Long, thick vertical line is the rest wavelength for source with z= 2.1 at observed frame 850 µ m (SCUBA-2); short, thick vertical line is the rest wavelengthfor source with z = 2.1 at observed frame 450 µ m (SCUBA-2). Long, thin vertical line is therest wavelength for source with z = 10 at observed frame 850 µ m; short, thin vertical line isrest wavelength for source with z = 10 at observed frame 2 mm (GISMO). Luminosity ratiosat these rest wavelengths are taken from thick curve. Error bar is observational uncertaintyin ratio log L ν / L ν (7.8 µ m) for individual quasars. 17 –in the L ν (100 µ m)/ L ν (7.8 µ m) ratio (cross in Figure 3). From these comparisons of obser-vational uncertainties and observed dispersions, we conclude that the intrinsic variation inthe ratio of cool dust to hot dust is ∼ ± f ν (7.8 µ m) is the 10% uncertainty in measurement of the IRS spectrum. Combining this with the25% SPIRE uncertainty gives a total observational uncertainty in L ν (100 µ m)/ L ν (7.8 µ m)of ± ± ∼ ± µ m, the observed SEDs in Figures 3 and 4 for luminous quasars need to beextended. We do this by adopting the SED of the most luminous ULIRGs determined by Herschel (Figure 17 of Symeonidis et al. 2013) and normalizing to the 100 µ m luminosityof the quasars. These extended SEDs are shown as thick curves in Figures 3 and 4. Thesecurves are used below to determine values of L ν ( λ )/ L ν (7.8 µ m) for λ the rest frame wave-length corresponding to submm and mm observations at different observed wavelengths andredshifts. L IR for Obscured and Unobscured Quasars Having full SEDs allows the determination of total infrared luminosities L IR . For thelocal AGN, the total infrared luminosity L IR reradiated by absorbing dust can be deter-mined as defined by Sanders and Mirabel (1996) using IRAS fluxes, whereby f IR = 1.8 x10 − [13.48 f ν (12) + 5.16 f ν (25) + 2.58 f ν (60) + f ν (100)], for f IR in erg cm − s − and IRASflux densities in Jy. (This relation also includes an estimated contribution from longer wave-lengths.) For AGN, Sargsyan et al. (2011) found that log [ L IR / νL ν (7.8 µ m)] = 0.51 ± L IR / νL ν (7.8 µ m)] = 0.80 ± µ m and 100 µ m for the high redshift quasars, assumingthat shorter wavelengths which are unobserved retain the same ratios to f ν (7.8 µ m) andthat f ν (60 µ m) = f ν (100 µ m) for the quasars. The result for the high redshift silicateemission quasars is log [ L IR / νL ν (7.8 µ m)] = 0.41 and for silicate absorption quasars is log[ L IR / νL ν (7.8 µ m)] = 0.68. Applying these transformations to the νL ν (7.8 µ m) luminosityfunctions for obscured and unobscured quasars with 1.8 < z < L IR (L ⊙ ) for unobscured quasars (diamonds) and obscuredquasars (squares) with 1.8 < z < − scaledfrom νL ν (7.8 µ m) space densities in Vardanyan et al. (2014) using relations log [ L IR / νL ν (7.8 µ m)] = 0.41 for unobscured quasars and log [ L IR / νL ν (7.8 µ m)] = 0.68 for obscured quasars,determined from Figures 3 and 4 as described in text. The envelopes encompass statisticaluncertainties ± √ N for N the number of quasars > L in this redshift interval and are shownonly for luminosities that include quasars observed within the ∼ Bo¨otes survey field;no extrapolations of luminosity functions to fainter sources than observed have been applied. 19 –in νL ν (7.8 µ m), the different corrections to L IR mean that the obscured quasars dominatethe bolometric luminosity function.One of the questions we are asking is what fraction of high redshift, high luminosityDOGS are powered by quasars, compared to the fraction powered by starbursts. This isfundamental to deciding if luminous, obscured submm sources trace SFR in the early uni-verse. The Herschel
HerMES survey yielded an independent estimate of DOG luminosityfunctions (Calanog et al. 2013). We compare space densities of detected sources reported inthis survey within the interval 1.5 < z < of the HerMES survey, Calanog et al. report (their Table 2) 31 sources inthis redshift interval having log L IR > ⊙ , which yields a result of 1400 DOGS Gpc − in this luminosity range. Transforming to the value we adopt for H would be equivalent tolog L IR > ∼
800 obscured quasars Gpc − . Their L IR for these sources are derived in different manner than ours by assuming various spectraltemplates so results for the luminosity functions are independent. Given the uncertaintiesentering this comparison, these space densities are similar, which indicates that for the mostluminous DOGS, the high redshift examples are dominated by DOG quasars rather thanby DOG starbursts. This result cautions against using DOG samples having no spectralclassification as indicators of SFR. The dominance of quasars in the HerMES DOG studyprobably arises because they are found using a 24 µ m criterion, which selects in favor ofhotter dust.
4. Quasar Counts based on Dust Luminosities
Our objective in this section is to compare source counts, observed and predicted, forall three UV/IR categories of dusty quasars within different redshift ranges and at differentobserving wavelengths, from mid-infrared to millimeter. Because our SEDs scale to L ν (7.8 µ m), source counts are first established using observed f ν at rest wavelength 7.8 µ m. Sourcecounts at submillimeter and millimeter wavelengths are then predicted by scaling the farinfrared SEDs from section 3.2. 20 – < z < Obscured quasars discovered with the IRS based on the 9.7 µ m silicate absorptionfeature are found at redshifts 1.5 . z . µ m peak within the 24 µ m Spitzer
MIPS survey band causes the sample to cluster within 1.8 < z < f ν (24 µ m) > R >
24. Our goal is to determinesource counts for obscured quasars meeting these two criteria. Because IRS spectra werenot obtained of all sources defined by these flux limits, corrections need to be determinedfor incompleteness in the spectroscopic selections. We determine these corrections by theratio of sources having IRS spectra compared to the total number of sources meeting thephotometric criteria.Selection criteria varied between the FLS and Bo¨otes spectroscopic surveys. For theFLS, a variety of photometric criteria were used including IRAC colors and optical R magas bright as 19 (Sajina et al. 2007) whereas the Bo¨otes spectroscopy used only 24 µ m andoptical criteria because the primary goal was to understand the optically faintest sources. Asa result, the Bo¨otes sources contain many more obscured quasars despite the smaller overallspectroscopic sample size. From the FLS survey, there are 11 obscured quasars in Table 1within 1.8 < z < R &
24 are more reliable. For these reasons, we use only the Bo¨otesobscured quasars to determine incompleteness corrections and statistical uncertainties forthe obscured quasars.The distribution of the Bo¨otes survey in R and [24 µ m] Vega magnitude is illustrated inFigure 1 of Dey et al. (2008). There are 85 sources having R >
24 and f ν (24 µ m) > , of which 53 are included in the spectroscopic samples summarizedin Houck et al. (2005), Weedman et al. (2006), and Bussmann et al. (2009). This gives acorrection of 1.6 for incompleteness. (This is similar to the factor of 1.8 previously reportedby Weedman et al. (2006) as the ratio of total/observed Bo¨otes sources based on a selectioncriterion of f ν (24 µ m) > I > ± N . for N the number of sources in a bin, the surfacedensities of obscured Bo¨otes quasars with 1.8 < z < < z < f ν at rest wavelength 7.8 µ m forunobscured and obscured quasars, corrected for incompleteness as described in text. Enve-lope with thin line shows unobscured AGES quasars in Bo¨otes with optical redshifts. Errorbars with squares and envelope with thick line are obscured quasars in Bo¨otes with IRSredshifts from silicate absorption having R >
24, listed in Table 1. Lengths of error bars andsizes of envelopes show statistical uncertainties ± √ N for N the total number of observedquasars > f ν (7.8 µ m). Counts are shown only for quasars observed within the Bo¨otes surveyfield; no extrapolations of counts to brighter or fainter sources have been applied. Sourcecounts at any other observed wavelength or redshift can be predicted by scaling f ν ( λ )/ f ν (7.8 µ m) from Figures 3 and 4 with λ the rest wavelength corresponding to λ observed /(1+z), as infollowing figures. As discussed in text, intermediate quasars that are partially obscured areestimated as equal in number to the obscured quasars shown. Total counts for all quasarsis the sum of all three samples, shown as the single thick line, with statistical uncertaintyshown by the error bar. 22 –area, although slightly smaller at 7.7 deg . AGES is also based on an infrared selectioncriteria, f ν (24 µ m) > µ m flux densitiesarising from dust luminosities can be compared directly to the obscured quasar counts. Forquasars, AGES reaches I < contain 2070MIPS-selected quasars for which optical spectra were obtained of 1991 (Table 3 of Kochaneket al.) for a spectroscopic survey completeness of 96%. Almost all are type 1 quasars;Figure 7 of Hickox et al. (2007) shows that less than 2% of sources with spectra are type2. Assuming that 98% of sources are type 1 with a completeness correction of 1.04 resultsin an overall correction to counts for unobscured, type 1 quasars of only 1.02 times numberof sources with spectra. Taking the quasars from the AGES catalog having 1.8 < z < µ m is close to rest frame 7.8 µ m at the redshifts of interest.)The results in Figure 6 show that the obscured and unobscured Bo¨otes quasars are verysimilar in number for 1.8 < z < ∼ f ν (22 µ m) ∼
13 mJy, or f ν (restframe 7.8 µ m) ∼
17 mJy with our assumed template. The median R band flux densities are ∼ µ Jy (about magnitude 23), which gives representative log UV/IR ∼ -2.4, meeting ourdefinition of obscured quasars. Similar space densities for these obscured ELIRGS comparedto unobscured quasars were determined by Assef et al. (2015) at the highest luminositieswithin 2.0 < z < < log UV/IR < < z < ∼ R >
24 for obscured quasars. From the FLS survey area, Table1 includes 11 obscured sources with 1.8 < z < R >
24. There are an additional 11IRS observed quasars within 1.8 < z < < R <
24 which are not included in Table 1 (MIPS numbers 8226, 268,8251, 521, 509, 22204, 16080, 16152, 22482, 15949, and 16113.) Although statistics are small,this equal number indicates that the sample of partially obscured quasars (
R <
24) is thesame as the obscured, optically faint sample (
R > µ m . The sum of quasar counts shown in Figure 6 includes all quasars, therefore, withequal contributions from unobscured, partially obscured, and obscured quasars as definedby UV/IR.
5. Detections with Submm and mm Observations
The evolution of SFR in the universe is tracked primarily by the evolution of sourceswith cool dust (Chary and Elbaz 2001; Dale and Helou 2002; Assef et al. 2010; Elbaz et al.2011; Wardlow et al. 2011), invoking the assumption that dust luminosity at & µ m arisescompletely from star formation. We want to test this assumption by determining how manyquasars contaminate the submm samples, based on the empirical determinations given aboveof far infrared luminosities and space densities for all categories of dusty quasars. In whatfollows, we determine the expected submm counts for these quasar populations and comparewith sources actually known from the SCUBA-2 surveys (Chapman et al. 2005; Barger et al.2014; Roseboom et al. 2013; Geach et al. 2013). < z < Although previous analyses have concluded from optical spectral classifications, X-rayobservations and SED considerations that the submm source surveys contain few AGN(Chapman et al. 2005; Alexander et al. 2005), obscured and partially unobscured quasarswould be difficult to identify in these ways so could have been overlooked as contributing 24 –to submm source counts. This is emphasized, for example, by Alexander et al. who findthat only ∼
8% of SCUBA sources have observed X-ray fluxes consistent with quasars butthat the fraction can increase to ∼
80% if absorbed X-rays are assumed in Compton thick,dusty sources. This ambiguity is our main reason for comparing expected counts of theknown dusty quasar population to actual submm detections. Combining the rest frame 7.8 µ m counts shown in Figure 6 with the SEDs in Figures 3 and 4 allows predictions of quasarcounts that should be observed in submm source counts within 1.8 < z < µ m and 450 µ m wavelengths of SCUBA-2 are shown in Figures 7 and 8.From these Figures, it is seen that the expected submm counts from quasars are domi-nated by the obscured quasars, which are the sources which would not have been identifiedamong spectroscopic redshifts of submm sources. This emphasizes why quantitative com-parisons of expected quasars with observed counts are important. For specific comparisonswith submm surveys, we use the Barger et al. (2014) survey at 850 µ m with SCUBA-2 andthe Roseboom et al. (2013) 450 µ m survey with SCUBA-2.The expected counts at 850 µ m, determined with our empirical 7.8 µ m source counts inBo¨otes with no extrapolations, nearly reach the 2 mJy limit of the faintest 850 µ m survey,GOODS-N in Barger et al. (2014). The predicted counts for all quasars are ∼ − > < z < µ m . Barger et al. find five 850 µ m sourceswith spectroscopic 1.8 < z < brighter than 2 mJy, or a density of 45deg − . If estimated photometric redshifts are added, there are 3 more sources for a totaldensity of 72 deg − , ten times more than the expected number of quasars. These submmdetections are actually lower limits because the flux density limit is somewhat brighter overparts of the field.Of course, these results suffer from small number statistics, but they certainly confirmthat the 850 µ m surveys are indeed dominated by starbursts, as previously concluded byothers. The infrared classification of quasars and starbursts based on the strength of PAHemission also confirms the dominance of starbursts in submm samples. Of the submm sourcesobserved with the Spitzer
IRS by Pope et al. (2008) and Mene´ndez-Delmestre et al. (2009),at least 80% show PAH features. The most useful result of our analysis is that the heavilyobscured quasar population, not known before
Spitzer , is not a significant contaminant forthe 850 µ m surveys. The main difference between this result and our conclusion in section 3.3that high redshift DOGS detected by Herschel
SPIRE are dominated by quasars probablyarises because the DOG selection is based on 24 µ m, which selects for the hotter dust ofquasars.Comparison to the SCUBA-2 450 µ m survey (Roseboom et al. 2013) gives even largerdifferences between observed counts and quasar counts, although the 450 µ m redshifts are 25 –Fig. 7.— Expected quasar counts within 1.8 < z < µ m (SCUBA-2) for quasars scaled from f ν (7.8 µ m) counts in Figure 6 using SEDs in Figures3 and 4. Diamonds are unobscured quasars scaled from AGES optical survey, and squaresare optically faint, obscured quasars with IRS redshifts from silicate absorption. Range ofcounts encompasses statistical uncertainties in the source counts from Figure 6. Small errorbar shows estimate for the partially obscured quasars scaled as described in text, having7.8 µ m counts from Figure 6 the same as obscured quasars but assuming SEDs the same asunobscured quasars in Figure 3. Total counts for all quasars is the sum of all three samples,shown as the single thick line, with statistical uncertainty shown by the thick error bar.Current SCUBA-2 850 µ m detection limit is ∼ < z < µ m (SCUBA-2) for quasars scaled from f ν (7.8 µ m) counts in Figure 6 using SEDs in Figures3 and 4. Diamonds are unobscured quasars scaled from AGES optical survey, and squaresare optically faint, obscured quasars with IRS redshifts from silicate absorption. Range ofcounts encompasses statistical uncertainties in the source counts from Figure 6. Small errorbar shows estimate for the partially obscured quasars scaled as described in text, having7.8 µ m counts from Figure 6 the same as obscured quasars but assuming SEDs the same asunobscured quasars in Figure 3. Total counts for all quasars is the sum of all three samples,shown as the single thick line, with statistical uncertainty shown by the thick error bar.Current SCUBA-2 450 µ m detection limit is ∼ µ m and detection limit 20 mJy for comparison to SPIRE surveys such as inTable 1. 27 –only photometric. The 450 µ m detection limit is 6 mJy, to which 19 separate sources werefound within 210 arcmin having photometric redshifts 1.8 < z < − in this redshift interval. From Figure 8, we would expect only ∼ − to this limit. The large difference in densities of 450 µ m compared to 850 µ m sources ispuzzling and seems to arise in part because of photometric redshift estimates, and in partbecause of a large variance in overall submm source densities between the survey fields.For example, 19/69 of the 450 µ m sources in Roseboom et al. are assigned redshifts 1.8 < z < µ m sources with spectroscopic redshiftsin Chapman et al. (2005). However, only 8/49 of the 850 µ m sources in Barger et al. (2014)have spectroscopic or photometric redshifts in this interval, with 23 sources assigned noredshift. If these no redshift sources have comparable fractions within 1.8 < z < µ m sources, this indicates that the Barger et al. results underestimateby about a factor of two the real number of sources within 1.8 < z < ∼
150 deg − , about 1/2 the estimate from the 450 µ m survey.Total counts at 850 µ m and 450 µ m also differ by about this same factor. For anyredshifts .
3, the ULIRG curve in Figures 3 and 4 shows that observed frame 450 µ m ob-servations should see flux densities about 3 times brighter for the same ULIRGS seen inobserved frame 850 µ m observations. Yet, the 450 µ m survey reports 69 sources > , or 1200 deg − , compared to the 850 µ m result of 49 sources > , or 440 deg − , so the surface density of 450 µ m sources is nearly 3 times larger. Thisimplies either a large incompleteness in the 850 µ m results to the assumed 2 mJy limit, ora large cosmic variance in the survey fields.Regardless of the explanation of differences between 850 µ m and 450 µ m surveys, wecan conclude that the quasars in our known populations are responsible for only between2% and 10% of known submm sources with 1.8 < z < µ m. The SPIRE surveys for sources with 1.8 < z < f ν (7.8 µ m ) to SPIRE rest frame wavelengths using the SEDs in Figures 3 and 4 and combiningwith counts from Figure 6 gives the result in Figure 8. The SPIRE detection limit is takenas 20 mJy by comparison to Bo¨otes sources observed in Table 1. The predicted number ofdetections of ∼ − for absorbed quasars compares well with the 12 obscured quasars inTable 1 detected in Bo¨otes. The result in Figure 8 also indicates that we would not expectany SPIRE detections of unobscured quasars (counts of unobscured quasars extrapolate todensities approaching zero at the 20 mJy limit required) so we predict that no AGES quasars 28 –in Bo¨otes are detected by SPIRE. Spitzer
Starbursts
This large excess of submm sources compared to dusty quasars initially seems surprisingbecause the number of high redshift obscured quasars is comparable to the number of highredshift PAH sources (classified as starbursts) in the
Spitzer
IRS surveys. We examine inmore detail, therefore, whether the submm surveys based on detecting the rest frame farinfrared continuum reveal the same starburst population as the
Spitzer
IRS surveys whichdetect PAH features. We can test this only within a redshift interval similar to that usedfor the obscured quasars, because the
Spitzer photometric surveys at 24 µ m that revealstarbursts heavily favor redshifts at which the strong 7.7 µ m PAH feature is within the 24 µ m band, similarly to the redshift selection for the DOGS peaking at 7.8 µ m. For thiscomparison, the FLS spectral surveys are most useful instead of the Bo¨otes surveys becausethe FLS surveys are not constrained to faint optical magnitudes, and high redshift starburstsare not necessarily DOGS that would be fainter than R ∼
24. An example of such a sourcefound within the FLS survey is illustrated in Figure 9.To make this test, we reexamined all FLS sources included in the SPIRE measures bySajina et al. (2012) for which a PAH detection is mentioned. Within the range 1.8 < z < µ m , we scale PAH luminosities to the peak atrest frame 7.7 µ m , νL ν (7.7 µ m). As measured in CASSIS, all of these sources have f ν (7.7 µ m ) > f ν (7.7 µ m ) of 1.5 mJy at z = 2.0, log νL ν (7.7 µ m) = 45.8 (erg s − ) or12.2 (L ⊙ ). Using local starbursts with IRS spectra and IRAS fluxes, Sargsyan et al. (2011)calibrate log L IR / νL ν (7.7 µ m) = 0.74. From Kennicutt (1998) calibrating star formationrate (SFR) to total luminosity, log (SFR) = log L IR - 9.76, for SFR in M ⊙ yr − and L IR inL ⊙ . Luminosities log νL ν (7.7 µ m) > ⊙ for the PAH feature correspond, therefore, tolog L IR > ⊙ , or SFR > ⊙ yr − .The SPIRE 350 µ m flux densities in Sajina et al. allow a measure of the far infraredluminosity of these PAH sources compared to the f ν (7.7 µ m) from CASSIS. For these 11sources, we find that log f ν (115 µ m )/ f ν (7.7 µ m ) = 1.2 ± µ m for the average z of 2.03. The flux density observed by SPIRE can bescaled to that observed by SCUBA-2 850 µ m using the long wavelength SED for luminousULIRGS shown in Figure 3 or 4. The resulting SCUBA-2 flux density that should be observed 29 –Fig. 9.— Downloaded CASSIS spectrum for source MIPS22530 from FLS survey. Solid curvewith points is the optimal spectrum determined by CASSIS. Shading indicates uncertaintieswithin individual spectral pixels. Rest frame wavelengths at adopted redshift shown at top.The PAH 7.7 µ m feature dominates the spectral flux, and a distinct 6.2 µ m feature is alsoseen which can be used for starburst classification. 30 –at 850 µ m (275 µ m rest frame for z = 2.1) would be ∼ f ν (7.7 µ m) of 1.5mJy. Using this scaling, the expected count of SCUBA-2 sources can be determined fromthe observed count of PAH sources in the FLS survey.To compare with SCUBA-2 surveys, the FLS survey has to be corrected for incom-pleteness. This correction is given in Dasyra et al. (2009) who state that the sample is 57%complete (to limits including our relevant PAH flux densities) over an area of 2.8 deg . Ap-plying this correction to the 11 sources detected gives a density of 7 ± − within 1.8 < z < f ν (7.7 µ m, rest frame) > f ν ratios, thiscorresponds to f ν (275 µ m, rest frame) > µ m. Onlyvery few SCUBA-2 sources are so bright. In Barger et al. (2014), there are only 2 sourcesabove 6 mJy in 400 arcmin within 1.8 < z < ±
13 deg − .(Uncertainties in these count densities are scaled by N − . for N the number of actual sourceswhich were found.) Within the large statistical uncertainties that arise because of the fewsources detected in either FLS or SCUBA-2 surveys, the results for PAH sources overlapthe results for submm sources. This indicates that similar starbursts are detected at highredshift with these independent methods, but the uncertainties are too large for a definitiveconclusion about whether precisely the same populations are detected. The best route to afinal test will be to observe SCUBA-2 flux densities for numerous PAH sources detected byIRS. This can determine if similar SFR densities are measured with both techniques. Nev-ertheless, the comparison of these results for Spitzer
PAH starbursts and SCUBA-2 submmstarbursts confirms that a large excess of dusty starbursts compared to dusty quasars shouldbe expected at high redshifts, as observed for the submm sources. < z < As quasar discoveries continue to higher and higher redshifts, understanding the exis-tence of the supermassive black holes required to produce their luminosity becomes increas-ingly puzzling (e.g. Volonteri 2012; Feng et al. 2014; Toft et al. 2014). Quasars found to thehighest redshifts seen so far (5 < z .
7) are also dusty (Venemans et al. 2015; Wu et al.2015). There is potential to push dusty quasar discoveries to extreme redshifts, z & µ m detections cannot reach the same lumi-nosity limits at these redshifts, having the observing band close to the rest frame luminosity 31 –peak also makes SCUBA-2 competitive for mapping the very high redshift universe.Sources detectable with GISMO or SCUBA-2 represent the best opportunity to con-strain the very high redshift quasar population because redshifts are so large that any opticaldiscoveries ( < µ m) are precluded by intergalactic absorption below the Lyman limit. AsStaguhn et al. and Dwek et al. emphasize, detections of z > ∼ L ν / L ν (7.8 µ m), Figure 10 shows the quasar source counts for 9.5 < z < µ m for these redshifts (not currentlyfeasible). In Figure 11, these counts are converted to observed frame 850 µ m and 2 mmusing the counts in Figure 10 and the SEDs in Figures 3 and 4. The results in Figure 11indicate that if quasar and starburst luminosity functions continue unchanged to such highredshifts, a source at z ∼
10 may already have been found within existing 2 mm GISMO or850 µ m SCUBA-2 surveys!At the GISMO survey limit of 0.5 mJy, Figure 11 indicates that & − are expected within 9.5 < z < ∼
20 implies 100 sources deg − that should be found within9.5 < z < and contains 5 unidentified sources, so it is feasible based on theseestimates that one of these could be at z ∼ µ m SCUBA-2 surveys are even more promising. The deepest surveylimit is 2 mJy, at which Figure 11 shows an expected quasar density (dominated by obscuredquasars) of 2 deg − . If this is scaled by the factor of 20 to include luminous starbursts, asubmm source with 9.5 < z < . The 850 µ m survey field in Barger et al. (2014) already covers 0.1deg , so at least one of the optically unidentified sources already found in that field shouldbe at such a redshift if quasar and starburst luminosity functions have not changed from z ∼ < z <
6. Summary and Conclusions
A population of quasars including all categories of dust obscuration is determined for1.8 < z < µ m survey fields with redshifts fromthe AGES survey or from the Spitzer
IRS. Luminosities are normalized to rest frame dustluminosities νL ν (7.8 µ m), which provides the best comparison among unobscured quasarswith 9.7 µ m silicate emission and obscured quasars with 9.7 µ m absorption. Obscurationis quantitatively classified by the ratio UV/IR = νL ν (0.25 µ m)/ νL ν (7.8 µ m); unobscuredquasars have log UV/IR > < log UV/IR < < -1.8. Quasar counts based on rest frame f ν (7.8 µ m) within the flux densitylimits of available 24 µ m surveys are given for 1.8 < z < ∼ µ m are determined using Herschel
SPIRE photometry ofobscured and unobscured quasars (Figures 3 and 4). New SPIRE photometry is presentedfor 77 unobscured quasars from the SDSS extending to z = 5. It is found that the ratio L ν (100 µ m)/ L ν (7.8 µ m) is about three times higher for obscured quasars compared to unobscured;the median ratio log [ L ν (100 µ m)/ L ν (7.8 µ m)] for unobscured quasars is 0.6 but is 1.05 forobscured quasars, a difference of about 2 σ compared to the dispersions in the ratios. Aftercorrecting for observational uncertainties, the intrinsic variation in the ratio of cool dust tohot dust is ∼ ±
40% within each category. Results mean that obscured quasars appear tohave a larger fraction of cool dust compared to hot dust, but we cannot determine if thisis intrinsic or is caused by differential extinction at 7.8 µ m for the obscured quasars. The L ν (100 µ m)/ L ν (7.8 µ m) ratio is less by about a factor of two for quasars compared to localAGN of the same obscuration class.Using the far infrared SEDs together with the quasar counts at 7.8 µ m, total quasarcounts within 1.8 < z < µ m and 850 µ m wavelengthsfor comparison to the submm sources that have been discovered with SCUBA-2 (Figures 7and 8). It is found that only ∼
5% of the high redshift submm sources are quasars, and 33 –Fig. 10.— Expected quasar counts within 9.5 < z < µ m for unobscured and obscured quasars if luminosityfunctions are the same for 9.5 < z < < z < f ν ( λ )/ f ν (7.8 µ m) from Figures 3 and 4 with λ the rest wavelength corresponding to λ observed /(1+z). As discussed in text, intermediatequasars that are partially obscured are estimated as equal in number to the obscured quasarsshown. Single thick line is sum for all three quasar classes with statistical uncertainty shownby error bar. 34 –Fig. 11.— Expected quasar counts within 9.5 < z < µ m (SCUBA-2 = thin lines) and 2 mm (GISMO = thick lines) for unobscured and obscuredquasars if luminosity functions for 9.5 < z < < z < µ m source counts scale from Figure 10 and SEDs scale as in Figures 3 and 4. Envelopeswithout symbols are unobscured quasars like those from AGES and envelopes with squaresare like obscured IRS quasars in Table 1. Range of envelopes includes statistical uncertaintiesin the Bo¨otes counts of these quasar populations. Current SCUBA-2 850 µ m detection limitis ∼ ∼ < z < µ m surveys or 2 mm surveys with the GISMOsurvey camera should already have detected sources at z ∼
10 in this case. This illustrativecalculation demonstrates the importance of extending the submm and mm surveys to largerareas and of obtaining redshifts for the unidentified sources in these surveys.We thank those who built the
Herschel
Observatory for the opportunity to observe withopen time. SPIRE was developed by a consortium of institutes led by Cardiff University (UK)and including University of Lethbridge (Canada); NAOC (China); CEA, LAM (France);IFSI, Univ. Padua (Italy); IAC (Spain); Stockholm Observatory (Sweden); Imperial CollegeLondon, RAL, UCL-MSSL, UKATC, University of Sussex (UK); and Caltech, JPL, NHSC,University of Colorado (USA). This development has been supported by national fundingagencies CSA (Canada); NAOC (China); CEA, CNES, CNRS(France); ASI (Italy); MCINN(Spain); SNSB (Sweden);STFC, UKSA (UK); and NASA (USA). Partial support for thiswork was provided by NASA through RSA 1489723 issued by JPL/Caltech through theNASA
Herschel
REFERENCES
Alexander D. M., Bauer, F. E., Brandt, W. N., et al., 2003, AJ, 126, 539Alexander, D. M., Bauer, F. E., Chapman, S. C., Smail, I., Blain, A. W., Brandt, W. N.and Ivison, R. J. 2005, ApJ, 632, 736Antonucci, R. 1993, ARA&A, 31, 473Assef, R. J., Kochanek, C. S., Brodwin, M., et al. 2010,ApJ, 713, 970Assef, R. J., Eisenhardt, P. R. M., Stern D., et al. 2015, ApJ, 804, 27Banados, E., Decarli, R., Walter, F., Venemans, B., Farina, E., and Fan, X. 2015, ApJ, 805,8A. J. Barger, A. J., Cowie, L. L., Chen, C.-C. et al. 2014, ApJ, 784, 9Bauer, F. E., Yan, L., Sajina, A., and Alexander, D. M. 2010, ApJ, 710, 212B´ethermin, M., LeFloc’h, E. Ilbert, O., et al. 2012, A&A, 542, A58Boyle, B. J., Shanks, T., and Peterson, B. A. 1988, MNRAS, 235, 935Brand, K., Dey, A., Weedman, D., et al. 2006, ApJ, 644, 143Brand, K., Dey, A., Desai, V., et al. 2007, ApJ, 663, 204Brand, K., Weedman, D. W., Desai, V., et al. 2008, ApJ, 680, 119Brandl, B. R., Bernard-Salas, J., Spoon, H. W. W., et al. 2006, ApJ, 653, 1129Brown, M. J. I., Brand, K., Dey, A., et al. 2006, ApJ, 638, 88Bussmann, R. S., Dey, A., Lotz, J., et al. 2009, ApJ, 693, 750Calanog, J. A., Wardlow, J., Fu, H. et al., 2013, ApJ, 775, 61.Carilli, C. L., Riechers, D., Walter, F., et al. 2013, ApJ, 763, 120Carswell R. F., and Smith, M. G. 1978, MNRAS, 185, 381Casey, C. M., Berta, S., B´ethermin, M., et al. 2012, ApJ, 761, 139Chary, R., and Elbaz, D. 2001, ApJ, 556, 562Chapman, S. C., Blain, A. W., Smail, I., and Ivison, R. J. 2005, ApJ, 622, 772 37 –Chen, C.-T. J., Hickox, R. C., Alberts, S. et al 2015, ApJ, 802, 50Dale, D. A., and Helou, G. 2002, ApJ, 576, 159Dasyra, K. M., Yan, L., Helou, G., et al. 2009, ApJ, 701, 1123Deo, R. P., Richards, G. T., Nikutta, R., et al. 2011, ApJ, 729, 108Desai, V., Armus, L., Spoon, H. W. W., et al. 2007, ApJ, 669, 810Desai, V., Soifer, B. T., Dey, A., et al. 2009, ApJ, 700, 1190Dey, A., Soifer, B. T., Desai, A., et al. 2008, ApJ, 677, 943Donley, J. L., Rieke, G. H., P´erez-Gonz´alez, P. G., Rigby, J. R., and Alonso-Herrero, A.2007, ApJ, 660, 167Dowell, C. D., Conley, A., Glenn, J. et al. 2014, ApJ, 780, 75Draine, B. T. 1989, Proceedings of the 22nd ESLAB Symposium, ed. B.H. Kaldeich, ESASP-290, 93Dwek, E., Staguuh, J., Arendt, R. G., Kovacks, A., Ting, S., and Benford, D. J. 2014, ApJ,788, 30.Eisenhardt, P. R. M., Wu, J., Tsai, C.-W., et al. 2012, ApJ, 755, 173Efstathiou, A., Pearson, C., Farrah, D. et al. 2014, MNRAS, 437, 16Elbaz, D. ,Dickinson, M., Hwang, H. S., et al. 2011, A&A, 533, 119Elitzur, M. 2012, ApJ, 747, 33Fadda, D., Marleau, F. R., Storrie-Lombardi, L. J., et al. 2006, AJ, 131, 2859Farrah, D., Bernard-Salas, J., Spoon, H. W. W., et al. 2007, ApJ, 667, 149Farrah, D., Lonsdale, C. J., Weedman, D. W., et al. 2008, ApJ, 677, 957Fazio, G. G., Hora, J. L., Allen L. E., et al. 2004, ApJS, 154, 10Feltre, D., Hatziminaoglou, E., Hern´an-Caballero, A., et al. 2013, MNRAS, 434, 2426Feng, Y., Di Matteo, T., Croft, R., and Khandai, N. 2014, MNRAS, 440, 1865Fiolet, N., Omont, A., Lagache, G. et al. 2010, A&A, 524, 33 38 –Fiore, F., Grazian, A., Santini, P., et al. 2008, ApJ, 672, 94Geach, J. E., Chapin, E. L., Coppin, E. K. et al. 2013, MNRAS, 432, 53Genzel, R., Lutz, D., Sturm, E., et al. 1998, ApJ, 498, 579Griffin, M. J., Abergel, A., Abreu, A. et al. 2010, A&A, 518, 3Hao, L., Spoon, H. W. W., Sloan, G. C., et al. 2005, ApJ, 625, L75Hao, L., Weedman, D. W., Spoon, H. W. W., et al. 2007, ApJ, 655, L77Hickox, R. C., Jones, C., Forman, W. R., et al. 2007, ApJ, 671, 1365Hill, A. R., Gallagher, S. C., Deo, R. P., Peeters, E., and Richards, G. T. 2014, MNRAS,438, 2317Hopkins, P.F., Hernquist, L., Cox, T. J., and Keres, D. 2008, ApJS, 175, 356Houck, J. R., Roellig, T. L., van Cleve, J., et al. 2004, ApJS, 154, 18Houck, J.R., Soifer, B. T., Weedman, D., et al. 2005, ApJ, 622, L105Huynh, M. T., Norris, R. P., Coppin, K. E. K., et al. 2013, MNRAS, 431, 88Imanishi, M., Dudley, C. C., Maiolino, R., Maloney, P. R., Nakagawa, T., and Risaliti, G.2007, ApJ, 171, 72Jannuzi, B. T., and Dey, A., 1999, in ”Photometric Redshifts and the Detection of HighRedshift Galaxies”, ASP Conference Series, Vol. 191, Edited by R. Weymann, L.Storrie-Lombardi, M. Sawicki, and R. Brunner. ISBN: 158381-017-X, p. 111Kennicutt, R.C. 1998, ARA&A, 36, 18Khachikian, E. Y. and Weedman, D. W. 1974, ApJ, 192, 581Kirkpatrick, A., Pope, A., Alexander, D. M., et al. 2012, ApJ, 759, 139Kochanek, C. S., Eisenstein, D. J., Cool, R. J., et al. 2012, ApJS, 200, 8Laurent, O., Mirabel, R. F., Charmandaris, V., et al. 2000, A&A, 359, 887Lebouteiller, V., Barry, D. J., Spoon, H. W. W. et al. 2011, ApJS, 196, 8Leipski, C., Meisenheimer, K., Walter F., et al. 2014, ApJ, 785, 154 39 –Levenson, N. A., Sirocky, M. M., Hao, L., Spoon, H. W. W., Marshall, J. A., Elitzur, M.,and Houck, J. R. 2007, ApJ, 654, 45Lewis D. W., MacAlpine, G. M., and Weedman, D. W. 1979, ApJ, 233, 787Marshall, H. L., Huchra, J. P., Tananbaum, H., et al. 1984, ApJ, 283, 50Mart´ınez-Sansigre, A., Rawlings, S., Lacy, M., et al. 2006, MNRAS, 370, 1479Melbourne, J., Soifer, B. T., Desai, V., et al. 2012, AJ, 143, 125Mene´ndez-Delmestre, K., Blain, A. W., Smail, I. et al. 2009, ApJ, 699, 667Narayanan, D., Dey, A., Hayward, C. C., et al. 2010, MNRAS, 407, 1701Nguyen, H. T., Schulz, B., Levenson, L., et al. 2010, A&A, 518, L5Oliver, S. J., Wang, L., Smith, A. J. et al. 2010, A&A, 518, L21Osmer, P.S. 1982, ApJ, 283, 28Penner, K., Dickinson, M., Pope, A. et al. 2012, ApJ, 759, 28Pilbratt, G. L., Riedinger, J. R., Passvogel, T. et al. 2010, A&A, 518, L1Polletta, M., Weedman, D., H¨onig, S., et al. 2008, ApJ, 675, 960Pope, A., Chary, R.-R., Alexander, D. M., et al. 2008, ApJ, 675, 1171Ramos Almeida, C. Levenson, N. A., Alonso-Herrero, A., et al. 2011, ApJ, 731, 92Rieke, G., Young, E. T., Engelbracht, C. W., et al. 2004, ApJS, 154, 25Riess, A. G., Macri, L., Casertano, S., et al. 2011, ApJ, 730, 119Roseboom, I. G., Dunlop, J. S., Cirasuolo, M. et al. 2013, MNRAS, 436, 430Sajina, A., Yan, L., Armus, L., Choi, P., Fadda, D., Helou, G., and Spoon, H. 2007, ApJ,664, 713Sajina, A., Yan, L., Fadda, D., Dasyra, K., and Huynh, M. 2012, ApJ, 757, 13Sanders, D. B., and Mirabel, I. F. 1996, ARA&A, 34, 749Sargsyan, L., Weedman, D., Lebouteiller, V., et al. 2011, ApJ, 730, 19Savage R. S. and Oliver S., 2007, ApJ, 661, 1339 40 –Schmidt, M., and Green, R. F. 1983, ApJ, 269, 352Schneider, D. P., Richards, G. T., Hall, P. B., et al. 2010, AJ, 139, 2360Shen, Y. et al. 2011, ApJS, 194, 45Shi, Y. et al. 2006, ApJ, 653, 127Simpson, C., and Eisenhardt, P. 1999, PASP, 111, 691Smail, I., Ivison, R. J., and Blain, A. W., 1997, ApJ, 490, L5H. W. W. Spoon, H. W. W., Armus, L., Cami, J. et al. 2004, ApJS, 154, 184.Staguhn, J. G., Kovacks, A., Arendt, R. G. et al. 2014, ApJ, 790, 77Stierwalt, S.,Armus, L., Surace, J. A., et al. 2013, ApJS, 206, 1Symeonidis, M., Vaccari, M., Berta, S. et al. 2013, MNRAS, 431, 2317Thompson, G. D., Levenson, N. A., Uddin, S. A. and Sirocky, M. M. 2009, ApJ, 697, 182Toft, S., Smolci´c, V., Magnelli, B., et al. 2014, ApJ, 782, 68Tommasin, S., Spinoglio, L., Malkan, M., and Fazio, G. 2010, ApJ, 709, 1257Tsai, C.-W., Eisenhardt, P. R. M., Wu, J., et al. 2015, ApJ, 805, 90Vardanyan, V., Weedman, D., and Sargsyan, L. 2014, ApJ, 790, 88Veilleux, S., Rupke, D. S. N., Kim, D.-C., et al. 2009, ApJS, 182, 628Venemans, B. P., Banados, E., Decarli, R., et al. 2015, ApJ, 801, L11Volonteri, M. 2012, Science, 337, 544Wang, R., Wagg, J., Carilli, C. L., et al. 2013, ApJ, 773, 44J. L. Wardlow, J. L., Smail, I., Coppin, K. E. K., et al. 2011, MNRAS, 415, 1479Willott, C. J., Rawlings, S., Blundell, K. M., and Lacy, M. 2000, MNRAS, 316, 449Weedman, D.W., Soifer, B. T., Hao, L., et al. 2006, ApJ, 651, 101Weedman, D. W., Le Floch, E., Higdon, S. J. U., Higdon, J. L., and Houck, J. R. 2006, ApJ,638, 613 41 –Weedman, D., Polletta, M., Lonsdale, C. J., et al. 2006, ApJ, 653, 101Weedman, D. W., Sargsyan, L., Lebouteiller, V., Houck, J., and Barry, D. 2012, ApJ, 761,184Wright, E. L., Eisenhardt, P. R. M., Mainzer, A. K., et al. 2010, AJ, 140, 1868Wright, E. L. 2006, PASP, 118, 1711Wu, Y., Helou, G. , Armus, L., et al. 2010, ApJ, 723, 895Wu, X.-B., Wang, F., Fan, X., et al. 2015, Nature, 518, 512Yan, L., Chary, R., Armus, L. et al. 2005, ApJ, 628, 604Yan, L., Sajina, A., Fadda, D., et al. 2007, ApJ, 658, 778Zakamska, N. L., Strauss, M. A., Heckman, T. M., Ivezi, Z, and Krolik, J. H. 2004, AJ, 128,1002
This preprint was prepared with the AAS L A TEX macros v5.2.
Table 1. Obscured Quasars Discovered by
Spitzer
IRS
Source identifier a coordinates z b f ν (7.8 µ m) c νL ν (7.8 µ m) d f ν (250 µ m) e f ν (350 µ m) e f ν (500 µ m) e ref. f J2000 mJy log erg s − mJy mJy mJy1 SST24(Bo¨otes) 142538.23+351855.1 2.28 1.3 45.76 56 48 35 12 SST24(Bo¨otes) 142611.35+351217.9 1.82 2.1 45.79 · · · · · · · · ·
23 SST24(Bo¨otes) 142622.01+345249.2 1.98 2.3 45.90 < < <
25 34 SST24(Bo¨otes) 142648.90+332927.2 1.82 3.3 45.99 < < <
25 35 SST24(Bo¨otes) 142653.23+330220.7 1.80 1.5 45.63 36 24 <
25 36 SST24(Bo¨otes) 142745.88+342209.0 3.35 4.5 46.58 · · · · · · · · ·
27 SST24(Bo¨otes) 142804.12+332135.2 2.16 1.6 45.81 19 < <
25 38 SST24(Bo¨otes) 142924.83+353320.3 2.05 1.3 45.67 < < <
25 19 SST24(Bo¨otes) 142931.36+321828.2 2.33 1.5 45.84 · · · · · · · · ·
10 SST24(Bo¨otes) 142958.33+322615.4 2.34 1.8 45.92 < < <
25 111 SST24(Bo¨otes) 143001.91+334538.4 2.46 5.8 46.46 64 55 39 112 SST24(Bo¨otes) 143004.77+340929.9 3.22 4.1 46.51 · · · · · · · · ·
213 SST24(Bo¨otes) 143025.74+342957.3 2.73 3.6 46.33 26 21 <
25 314 SST24(Bo¨otes) 143026.04+331516.3 1.83 2.2 45.81 · · · · · · · · ·
515 SST24(Bo¨otes) 143028.52+343221.3 2.15 1.9 45.88 40 37 29 416 SST24(Bo¨otes) 143109.78+342802.7 2.2 1.3 45.73 < < <
25 317 SST24(Bo¨otes) 143135.29+325456.4 1.52 4.3 45.95 60 55 35 318 SST24(Bo¨otes) 143251.89+333536.8 1.70 1.1 45.45 24 18 <
25 319 SST24(Bo¨otes) 143253.39+334844.3 2.90 1.9 46.10 · · · · · · · · ·
220 SST24(Bo¨otes) 143312.70+342011.0 2.11 2.2 45.93 19 < <
25 421 SST24(Bo¨otes) 143318.59+332127.0 2.72 1.4 45.92 · · · · · · · · ·
222 SST24(Bo¨otes) 143358.07+332607.7 1.95 1.7 45.75 20 < <
25 323 SST24(Bo¨otes) 143447.70+330230.6 1.99 2.2 45.88 96 69 56 324 SST24(Bo¨otes) 143504.12+354743.2 2.08 1.6 45.78 22 19 <
25 325 SST24(Bo¨otes) 143508.49+334739.8 2.08 3.4 46.10 < < <
25 326 SST24(Bo¨otes) 143520.75+340418.2 2.2 2.0 45.92 < < <
25 127 SST24(Bo¨otes) 143523.99+330706.8 2.59 1.3 45.85 16 < <
25 128 SST24(Bo¨otes) 143539.34+334159.1 2.5 3.7 46.28 34 22 <
25 129 SST24(Bo¨otes) 143545.11+342831.4 2.53 3.0 46.20 16 < <
25 330 SST24(Bo¨otes) 143644.22+350627.4 1.8 3.3 45.98 50 40 27 131 SST24(Bo¨otes) 143725.23+341502.4 2.04 1.9 45.84 33 34 26 332 SST24(Bo¨otes) 143807.92+341612.4 2.33 2.6 46.07 · · · · · · · · ·
233 SST24(Bo¨otes) 143808.34+341015.6 2.33 2.3 46.02 < < <
25 334 SST24(FLS) 171057.45+600745.2 2.34 3.0 46.08 · · · · · · · · · Table 1—Continued
Source identifier a coordinates z b f ν (7.8 µ m) c νL ν (7.8 µ m) d f ν (250 µ m) e f ν (350 µ m) e f ν (500 µ m) e ref. f J2000 mJy log erg s − mJy mJy mJy35 MIPS8392 171343.88+595714.5 1.81 2.0 45.76 < < <
23 736 MIPS532 171526.06+585632.7 1.51 2.4 45.66 16 17 <
23 837 MIPS8245 171536.34+593614.8 2.65 2.0 46.06 < < <
20 838 MIPS78 171538.18+592540.1 2.46 3.8 46.25 < < <
21 939 MIPS8413 171545.7+595156.4 2.23 1.2 45.71 15 < <
16 740 MIPS429 171611.81+591213.3 2.12 1.6 45.80 17 < <
17 941 MIPS42 171758.44+592816.8 2.06 5.6 46.30 < < <
24 942 MIPS22303 171848.80+585115.1 2.34 2.7 46.10 < < <
20 843 SST24(FLS) 172048.02+594320.6 2.24 1.9 45.90 · · · · · · · · ·
644 MIPS16122 172051.48+600149.1 2.00 1.5 45.72 < < <
18 845 MIPS16037 172133.83+595046.9 1.59 2.2 45.70 16 < <
20 846 MIPS15958 172324.84+592455.5 1.95 1.6 45.74 31 19 <
20 747 MIPS22548 172330.46+584544.9 2.21 1.5 45.78 < < <
26 848 SST24(FLS) 172448.65+601439.9 2.34 3.5 46.14 · · · · · · · · · a Sources with Bo¨otes numbers correspond to numbers in Table 2 of Vardanyan et al. (2014). Sources with MIPS names are fromreferences listed for sources in the
Spitzer
First Look Survey. b Redshift z measured on CASSIS spectra from fitting median template of AGN with silicate absorption from Sargsyan et al. (2011).Redshifts for Bo¨otes sources are from Vardanyan et al. (2014); redshifts for MIPS sources are newly measured. Uncertainty in z is ± c Peak flux density at 7.8 µ m determined by median of all points in spectrum between 7.7 µ m and 7.9 µ m. Relative uncertainty amongsources is ±
10% because of poor S/N of faint spectra. Absolute uncertainty of CASSIS flux calibration applied to all sources is below ± d Rest frame luminosity νL ν (7.8 µ m) in erg s − determined as νL ν (7.8 µ m) = 4 π D L [ ν /(1+z)] f ν (7.8 µ m), for ν corresponding to 7.8 µ = 74 km s − Mpc − , Ω M =0.27and Ω λ =0.73. (Log [ νL ν (7.8 µ m)(L ⊙ )] = log [ νL ν (7.8 µ m)(erg s − )] - 33.59.) e SPIRE flux densities and limits from Melbourne et al. (2012) for Bo¨otes sources and from Sajina et al. (2012) for MIPS sources. f Reference to original IRS discovery: 1 = Houck et al. (2005), 2 = Weedman et al. (2006), 3 = Bussmann et al. (2009), 4 = Brand et al.(2007), 5 = Brand et al. (2008), 6 = Weedman et al. (2006b), 7 = Dasyra et al. (2009), 8 = Sajina et al. (2007), 9 = Yan et al. (2007).
Table 2. Far Infrared Fluxes from SPIRE for High Redshift SDSS Quasars
No. SDSS identifier a z a f ν (22 µ m) b f ν (7.8 µ m) c νL ν (7.8 µ m) d f ν (250 µ m) e RA,Dec. f f ν (350 µ m) e RA,Dec. f f ν (500 µ m) e RA,Dec. f HerschelJ2000 mJy mJy log erg s − mJy arcsec mJy arcsec mJy arcsec i.d.1 030341.04-002321.9 3.18 6.2 8.3 46.82 31.9 -0.2,-1.8 <
25 0.3,56.3 <
25 56.0,-48.9 13422249692 032108.45+413220.8 2.47 12.7 14.8 46.87 <
25 686,190 <
25 609,49.6 <
25 763,-19.7 13422036143 073100.16+430740.6 3.95 4.5 6.8 46.87 <
25 1.1,-44.5 <
25 31.6,-18.4 <
25 -72.2,-44.2 13422702674 073103.12+445949.4 5.00 2.2 4.1 46.80 22.2 4.8,15.8 15.5 2.3,14.9 <
25 -38.1,-46.8 13422049595 073502.30+265911.5 1.97 27.6 28.8 46.99 84.8 -2.3,2.0 45.0 1.3,0.3 31.7 4.2,-2.7 13422703246 073733.01+392037.4 1.74 33.7 32.9 46.94 119.8 -1.7,1.4 71.7 -1.1,2.3 43.6 -0.1,2.2 13422702697 073936.25+280754.7 2.74 6.9 8.5 46.71 <
25 -70.7,-55.2 23.3 -6,2.7 37.8 -14.5,8.2 13422703238 074126.07+421530.7 1.85 16.0 16.2 46.68 67.9 -0.9,-1.3 31.5 4.7,-2 <
25 74.6,53.8 13422702689 074521.78+473436.1 3.22 11.1 14.9 47.06 40.9 1.8,-4.7 46.1 3.9,-3.9 50.0 2.1,-2.7 134226835310 074815.82+355912.2 3.36 6.4 8.9 46.88 77.9 -0.1,1.2 83.7 0.3,1.9 50.2 1.4,5.6 134227027211 074914.78+472904.1 2.04 13.8 14.6 46.72 74.4 -1.9,0.3 60.5 -0.1,3.1 <
25 -57.2,8.9 134227026612 075054.64+425219.2 1.90 22.2 22.7 46.85 <
25 38.7,11.9 <
25 -77.1,-28.3 <
25 -102,251 134227027313 075732.89+441424.6 4.17 3.1 4.9 46.76 <
25 -29.2,-10.0 <
25 62.8,10.6 <
25 105,-28.1 134227027414 080117.79+521034.5 3.24 12.3 16.6 47.12 79.8 -2.9,0.2 79.7 -0.3,0.4 56.7 -2.2,-5 134227026515 080542.39+155528.1 2.51 7.70 9.5 46.69 35.6 -1.1,3.6 <
25 -7.8,29.5 62.6 -4.7,30.6 134227031716 080849.42+521515.3 4.46 3.1 5.1 46.82 30.8 -1.5,1.1 40.9 -6,-3.5 46.2 -8.3,1.9 134227026417 081114.66+172057.4 2.30 10.8 12.2 46.73 <
25 -40.9,-21.1 <
25 -47.4,-59.3 <
25 16.7,49.7 134227031818 081207.57+052341.1 1.88 16.5 16.8 46.71 81.8 2.7,8.8 76.0 3.5,7.2 67.2 7.8,5.9 134227030919 081331.28+254503.0 1.51 89.5 80.1 47.21 188.8 1.0,-1.7 84.6 -1.8,-0.4 <
25 -98.6,35.4 134225447320 081806.87+071920.2 4.58 3.1 5.2 46.85 <
25 -69.1,4.4 <
25 63.2,58.6 <
25 74.3,-57.7 134227031021 081855.77+095848.0 3.67 6.5 9.5 46.96 47.9 -3.5,2.4 45.2 -3.5,4.2 29.3 -4.5,4.6 134227031222 081940.58+082357.9 3.21 6.3 8.4 46.82 37.3 1.7,3.4 <
25 -50.7,8.7 <
25 -47.6,0.2 134227031123 082319.65+433433.7 1.66 35.2 33.4 46.91 61.3 -2.0,0.1 40.8 -2.7,1.4 <
25 25.0,-71.6 134227027724 082450.79+154318.4 1.87 19.5 19.9 46.78 132.4 -1.0,1.2 100.5 -1.8,2.1 57.9 -2.4,2.5 134227031525 082454.02+130217.0 5.19 2.2 4.1 46.83 28.5 0.9,1.8 30.0 -1.2,3.3 43.2 -0.9,-0.8 134227031426 082548.07+095339.4 3.80 4.6 6.9 46.84 <
25 66.6,-92.2 <
25 -61.4,-47.4 <
25 19.2,-91.9 134227031327 082619.70+314847.9 3.09 6.6 8.7 46.80 34.3 -6.9,13.0 41.3 -4.9,11 36.1 0.2,0.1 134227029128 082638.59+515233.2 2.85 11.2 14.1 46.95 71.7 -1.5,0.0 55.4 -3.3,-0.8 32.2 1.8,-1.6 134227026229 082804.54+445256.9 2.07 12.1 13.4 46.69 37.7 -0.3,2.6 33.0 -2.1,4.7 <
25 3.5,152.5 134227027630 082854.70+431220.1 3.17 6.5 8.6 46.82 44.4 -2.2,1.7 <
25 53.6,-14.5 <
25 -79.3,17.9 134227027831 083103.01+523533.5 4.44 2.6 4.3 46.75 <
25 -29.2,-62.2 <
25 -56.5,33.2 <
25 30.2,-129.7 134227026132 083212.37+530327.3 4.05 4.0 6.2 46.85 36.7 -0.3,2.7 23.2 1.9,0.7 30.3 -0.7,-8.3 134227026033 083249.39+155408.6 2.42 8.25 10.0 46.68 34.7 -4.0,0.2 <
25 -16.6,-46.6 <
25 -19.2,-53.1 134227030234 083255.63+182300.6 2.27 10.3 11.6 46.70 <
25 -9.4,36.4 <
25 -8.5,33.8 27.9 -1.8,24.2 1342270300
Table 2—Continued
No. SDSS identifier a z a f ν (22 µ m) b f ν (7.8 µ m) c νL ν (7.8 µ m) d f ν (250 µ m) e RA,Dec. f f ν (350 µ m) e RA,Dec. f f ν (500 µ m) e RA,Dec. f HerschelJ2000 mJy mJy log erg s − mJy arcsec mJy arcsec mJy arcsec i.d.35 083413.90+511214.6 2.39 53.8 62.0 47.47 182.0 -0.6,1.1 124.1 -0.9,1.5 56.2 -3.0,4.5 134227025936 083417.12+354833.1 2.16 12.9 14.2 46.74 68.3 0.3,-1.5 77.3 -0.7,0.7 43.9 -0.2,0.7 134223075237 083535.69+212240.1 3.12 8.1 10.6 46.90 <
25 -39.6,-15.8 <
25 -43.3,-16.3 <
25 -52.0,-5.6 134227029638 083552.62+163343.9 4.25 9.2 14.7 47.26 41.9 -0.4,-0.4 37.2 4.9,-1.5 48.6 3.3,0.7 134227030139 083839.16+285852.7 4.36 4.3 7.0 46.95 <
25 77.3,-39.5 <
25 -97.9,-102.5 <
25 -22.9,138 134227029240 083850.15+261105.4 1.61 20.5 19.5 46.66 <
25 -35.1,-8.2 <
25 34.1,69.3 <
25 30.5,66.7 134227029341 084045.40+090809.4 4.54 2.7 4.6 46.79 <
25 50.4,-47.9 26.4 -4.5,1.7 <
25 -17.0,56.3 134227030542 084051.22+404806.7 4.42 2.5 4.1 46.73 <
25 -50.7,18.9 23.0 8,3.7 <
25 42.8,-115 134227027943 084401.95+050357.9 3.35 8.6 11.9 46.99 <
25 63.1,17.7 <
25 52.1,15.8 <
25 -137,-129 134227030644 084438.04+584825.5 4.77 2.5 4.3 46.80 <
25 61.0,-42.7 <
25 -9.6,24.3 <
25 126,-123 134227024445 084547.19+132858.1 1.88 16.4 16.7 46.71 27.5 -8.8,2.3 39.3 -15.7,1.7 43.6 -15.4,-1.9 134227030346 085010.26+593118.2 1.72 25.4 24.7 46.81 36.3 -6.6,4.0 23.6 -12.1,-15.3 <
25 215,-31 134227024347 085210.88+535948.9 4.22 4.6 7.3 46.94 39.0 12.6,-12.4 44.8 9.9,-9.8 39.8 14.5,-15.1 134227024648 085335.74+185446.5 2.15 13.3 14.5 46.76 31.4 -2.3,7.0 47.3 -6,21.5 47.4 -4.2,8.7 134227029949 085611.69+411516.8 3.68 5.5 8.0 46.89 <
25 4.4,-22.1 <
25 -39.4,21.8 <
25 -38.3,25.2 134227028050 085626.47+194137.7 2.82 30.4 38.1 47.38 <
25 -64.5,-6.1 <
25 -37.7,58.4 <
25 101,-114 134227029851 085634.92+525206.2 4.82 1.9 3.3 46.69 36.2 -13.2,5.4 49.3 -9.6,6 32.6 -4.1,1.9 134227024852 085707.94+321031.9 4.78 2.6 4.5 46.82 <
25 17.9,-21.6 <
25 17.2,19.6 <
25 -48.4,-106 134223075853 090033.50+421547.0 3.29 9.6 13.1 47.02 30.2 -3.0,2.5 35.3 0,-2 <
25 -40.0,-53.8 134227028154 090158.85+610931.7 4.08 3.5 5.4 46.79 29.1 -8.5,-5.8 31.7 -13.4,-4.2 32.8 -22.0,-4.5 134227024155 090334.94+502819.3 3.58 8.1 11.6 47.03 220.9 -0.3,-1.3 237.8 0.5,-1.7 189.9 0.7,-2.3 134225462856 090527.46+485049.9 2.69 8.2 10.0 46.76 <
25 118.8,8.3 27.7 13.3,16.4 <
25 -112,31.6 134227025757 091206.78+331109.3 3.33 7.0 9.6 46.90 57.0 -1.9,0.2 45.7 -2.6,-1.1 28.9 -14.1,1 134227028858 091301.01+422344.7 2.31 10.8 12.3 46.74 45.4 -0.8,5.4 24.5 3.7,-1.5 <
25 -31.5,-30.9 134227028259 091342.48+372603.3 2.13 12.1 13.2 46.70 36.1 -2.1,-0.6 33.1 1.6,4 37.3 -6.8,6.1 134227028460 091610.35+621326.2 2.08 21.3 22.9 46.92 119.6 -1.8,2.4 75.9 -1.6,3.3 55.1 1.2,2.5 134227023961 092058.46+444154.0 2.19 19.3 21.3 46.93 112.7 0.3,1.2 164.5 0,0 193.9 0.9,-0.8 134227025562 092819.29+534024.1 4.39 5.0 8.2 47.02 56.5 4.1,2.5 66.3 2.7,4.6 56.8 -0.1,3.5 134227024963 093554.46+525616.4 4.01 5.0 7.6 46.93 <
25 -36.6,-14.7 <
25 -36.3,-11 <
25 -38.0,176 134227025064 094056.01+584830.2 4.66 3.1 5.4 46.88 29.2 1.5,8.0 29.2 -0.8,5.2 <
25 -47.2,-27.4 134227023665 095014.05+580136.5 3.96 4.4 6.7 46.86 39.5 -14.1,-4.2 45.0 -13.1,-8.4 30.4 -10.4,-8.1 134227023566 100129.64+545438.1 1.76 15.8 16.0 46.64 38.8 -11.1,7.6 48.2 -11.9,5.1 40.1 -11.1,7.8 134227025167 101051.14+570530.8 1.96 11.9 12.8 46.63 40.6 -1.6,5.5 30.2 -7.9,13.5 <
25 -89.5,35.1 134227023368 102907.09+651024.6 2.16 12.3 13.4 46.73 26.5 -0.6,1.3 28.4 2.4,-0.1 <
25 -96.3,-64.3 1342270222
Table 2—Continued
No. SDSS identifier a z a f ν (22 µ m) b f ν (7.8 µ m) c νL ν (7.8 µ m) d f ν (250 µ m) e RA,Dec. f f ν (350 µ m) e RA,Dec. f f ν (500 µ m) e RA,Dec. f HerschelJ2000 mJy mJy log erg s − mJy arcsec mJy arcsec mJy arcsec i.d.69 140146.53+024434.7 4.44 2.9 4.7 46.79 <
25 318,-58.6 <
25 288,-251 <
25 194,-326 134220113070 141546.24+112943.4 2.56 57.1 68.0 47.56 531.8 1.1,-0.4 399.4 1,-0.5 229.3 0.5,0 134226153771 144709.24+103824.5 3.68 8.3 12.1 47.07 <
25 -309,394 <
25 -225,447 <
25 -261,513 134223615372 150424.98+102939.1 1.84 18.6 18.7 46.74 148 0.8,-2.5 <
25 53.3,-3.8 <
25 350,254 134223832373 153308.65+301820.7 4.45 2.7 4.5 46.77 37.3 2.2,-0.9 29.8 0.9,-3.6 28.7 4.9,-4.7 134226168174 160336.64+350824.3 4.46 2.4 4.0 46.73 43.7 1.2,0.1 57.2 1.9,-0.2 34.8 6.6,1.5 134224116275 161622.10+050127.7 4.87 2.2 3.8 46.76 45.5 5.6,-14.8 62.2 2.8,-13.2 41.5 -3.9,-10.7 134222956476 163411.82+215325.0 4.53 3.0 5.0 46.83 <
25 4.6,32.2 21.8 1.7,-3.2 14.8 1.3,-1.8 134223998177 172413.27+571046.7 2.83 7.5 9.4 46.77 34.5 -0.2,-8.8 44.6 2.7,-11.9 29.1 -2.0,-18.1 1342270212 a SDSS identifier and redshift from version 7 of the SDSS quasar catalog (Schneider et al. 2010). b Observed flux density at 22 µ m from the WISE All Sky Catalog available at wise2.ipac.caltech.edu/docs/release/allsky/. Zero point of 22 µ m magnitude listed in catalogtaken as 8280 mJy; typical uncertainties for sources with fluxes listed are ± c Flux density f ν (7.8 µ m) at observed wavelength corresponding to rest wavelength 7.8 µ m, determined by scaling f ν (observed 22 µ m) to f ν (rest frame 7.8 µ m) using tabulatedredshift and template spectrum of silicate emission quasars in Weedman et al. (2012). d Rest frame luminosity νL ν (7.8 µ m) in erg s − determined as νL ν (7.8 µ m) = 4 π D L [ ν /(1+z)] f ν (7.8 µ m), for ν corresponding to 7.8 µ = 74 km s − Mpc − , Ω M =0.27 and Ω λ =0.73. e Flux density from SPIRE in observed frame at wavelength listed. Median one sigma uncertainty for f ν (250 µ m) is ± f ν (350 µ m) is ± f ν (500 µ m) is ± f Offset in arcsec from SDSS coordinate of closest SPIRE source at wavelength of preceding column. SPIRE source is identified with the SDSS source and a value for f ν listedin the preceding column if the total offset distance < ′′ at 250 µ m , < ′′ at 350 µ m , and < ′′ at 500 µ m (i.e., within one FWHM of the beam size.). If no SPIRE sourceis found within these distances, flux of the SDSS quasar is listed as upper limit of <<