Far-IR/Submillimeter Spectroscopic Cosmological Surveys: Predictions of Infrared Line Luminosity Functions for z<4 Galaxies
Luigi Spinoglio, Kalliopi M. Dasyra, Alberto Franceschini, Carlotta Gruppioni, Elisabetta Valiante, Kate Isaak
aa r X i v : . [ a s t r o - ph . C O ] O c t Draft version August 25, 2018
Preprint typeset using L A TEX style emulateapj v. 6/6/04
FAR-IR/SUBMILLIMETER SPECTROSCOPIC COSMOLOGICAL SURVEYS:PREDICTIONS OF INFRARED LINE LUMINOSITY FUNCTIONS FOR z < Luigi Spinoglio , Kalliopi M. Dasyra , Alberto Franceschini , Carlotta Gruppioni , Elisabetta Valiante and Kate Isaak Draft version August 25, 2018
ABSTRACTStar formation and accretion onto supermassive black holes in the nuclei of galaxies are the twomost energetic processes in the Universe, producing the bulk of the observed emission throughout itshistory. We simulated the luminosity functions of star-forming and active galaxies for spectral linesthat are thought to be good spectroscopic tracers of either phenomenon, as a function of redshift. Wefocused on the infrared (IR) and sub-millimeter domains, where the effects of dust obscuration areminimal. Using three different and independent theoretical models for galaxy formation and evolution,constrained by multi-wavelength luminosity functions, we computed the number of star-forming andactive galaxies per IR luminosity and redshift bin. We converted the continuum luminosity counts intospectral line counts using relationships that we calibrated on mid- and far-IR spectroscopic surveysof galaxies in the local universe. Our results demonstrate that future facilities optimized for survey-mode observations, i.e., the Space Infrared Telescope for Cosmology and Astrophysics (SPICA) andthe Cerro Chajnantor Atacama Telescope (CCAT), will be able to observe thousands of z> II ], [O I ], [O III ], [C II ], in a half-square-degree survey, with onehour integration time per field of view. Fainter lines such as [O IV ], [Ne V ] and H (0-0)S1 will beobserved in several tens of bright galaxies at 1 < z <
2, while diagnostic diagrams of active-nucleus vsstar-formation activity will be feasible even for normal z ∼ Subject headings:
Galaxies: evolution, active, starburst, Seyfert - Techniques: imaging spectroscopy INTRODUCTION
Tremendous progress in infrared (IR) astronomy wasmade in the last decade, largely driven by the launch ofthe
Spitzer and
Herschel space telescopes. Nonetheless,several basic questions remain unanswered in the fields ofobservational cosmology and galaxy formation and evo-lution, emphasizing the need for future missions. Theseregard our incomplete knowledge on how the primordialgas collapses to form new stars, on the different modes ofstar formation, and on the potentially coeval growth ofblack holes and galaxies. Below, we elaborate on each ofthese turn. We argue that fine-structure and molecularlines observable in IR lines can help us address them, andwe present a set of simulations that support the need forfuture survey-oriented facilities that can make a strongimpact on studies of galaxy evolution.The collapse of the primordial gas at very high redshiftsthat led to the formation of the first stars and galaxies isthought to have occured via H line emission, which acts Istituto di Fisica dello Spazio Interplanetario, INAF, Via Fossodel Cavaliere 100, I-00133 Roma, ItalyElectronic address: [email protected] Laboratoire AIM, CEA/DSM-CNRS-Universit´e Paris Diderot,Irfu/Service dAstrophysique, CEA Saclay, F-91191 Gif-sur-Yvette,France Observatoire de Paris, LERMA (CNRS:UMR8112), 61 Av. del´ Observatoire, F-75014, Paris, France Dipartimento di Astronomia - Universit´a di Padova, Vicolodell´ Osservatorio 5, 35122 Padova, Italy Osservatorio Astronomico di Bologna - INAF, Via Ranzani 1,40127, Bologna, Italy Department of Physics and Astronomy, University of BritishColumbia, 6224 Agricultural Road, Vancouver, BC V6T 1Z1 ESA Research & Scientific Support Department - ESTEC, Ke-plerlaan 1, 2200 AG Noordwijk, The Netherlands as a very effective cooling mechanism in low-metallicityand low- temperature ( < K) environments (Wise &Abel 2007; Obreschkow & Rawlings 2009). Even thoughH remains a main coolant soon after the epoch of reion-ization, the direct detection of H gas at very high z isyet to be achieved observational point of view.Recent studies of the global infrared continuum andmolecular gas properties of galaxies in the local and in-termediate/high redshift Universe suggest that mergersand non- or weakly-interacting star-forming galaxies fol-low two separate Kennicutt-Schmidt relations with simi-lar exponents, but different normalizations (Genzel et al.2010; Daddi et al. 2010; Gracia-Carpio et al. 2011).At high- z , both galaxy mergers and accretion of cold gasvia cooling flows have been suggested as sufficient mecha-nisms to produce IR luminosities > L ⊙ (Powell et al.2011), unlike in the local Universe where ultraluminousIR galaxies (ULIRGs) are predominantly associated withmergers of comparable mass spirals (Dasyra et al. 2006).The parameters that determine the mode of star forma-tion and control its efficiency are not yet understood.While Spitzer and
Herschel helped us to identify and de-termine the star-formation rates of ULIRGs at z .
3, thebulk of star formation at these redshifts is thought to takeplace in galaxies of lower luminosity (Perez-Gonzalez etal. 2005; Reddy et al. 2010; Ly et al. 2011). Char-acterizing the star formation in such systems is essentialfor understanding the formation of present-day ellipticalsand downsizing (Cowie et al. 1996).Studies of local massive galaxies and theoretical mod-els suggest that most spheroids host massive black holes(Richstone et al. 1998; Fabian 1999). The observed cor-relations between the masses of these black holes and the Spinoglio et al.luminosities and stellar velocity dispersions of their hostgalaxies (Kormendy & Richstone 1992; Magorrian et al.1998; Ferrarese & Merrit 2000; Tremaine et al. 2002;Ferrarese & Ford 2000; Shankar et al. 2009; Gultekinet al. 2009) are remarkable, given the vastly differentscales that they involve. The enormous difference be-tween the black hole Schwarzschild radius and the char-acteristic radius of the bulge indicates that these rela-tions possibly reflect the coeval formation of the two ina common gravitational potential. To date, this scenariohas been primarily tested through the comparison of thestar-formation rate vs the black hole accretion rate as afunction of look-back time (Treu et al. 2004; Marconi etal. 2004; Merloni et al. 2004; Shim et al. 2009; Grup-pioni et al. 2011). Indeed, both activities have beenfound to peak at comparable redshifts, between z ∼ z sources with unambiguous evidence forhot dust–that is independent of the chosen star-forminggalaxy template–can lead to small, biased samples of lu-minous AGN. Less strict criteria can lead to incompleteAGN samples that are contaminated by starburst galax-ies. This is frequently the case for AGN selected basedon mid-IR color-color diagrams (Lacy et al. 2004; Sternet al. 2005; Barmby et al. 2006). The other unambigu-ous signature of a hard, AGN-related radiation field isthe line emission from ionic species that require & µ m, which however was only seen instacked spectra of z & Spitzer (Dasyra etal. 2009).Observing new FIR-selected galaxy samples, or sam-pling the peak of the IR SEDs of presently known sam-ples, will not only enable us to address such open ques-tions, but to also obtain a more coherent view of high- z galaxy populations. The thousands of 24 µ m -selected,IR-bright galaxies discovered by Spitzer (Papovich et al.2004; Rigby et al. 2004; Fadda et al. 2004; LeFloc’het al. 2004; Houck et al. 2005; Huang et al. 2009),the hundreds of mm galaxies (e.g. Smail et al. 1997;Hughes et al. 1998; Barger et al. 1998; Scott et al.2002; Borys et al. 2003; Webb et al. 2003; Coppin etal. 2006), (e.g. Greve et al. 2004; Laurent et al. 2005;Scott et al. 2008; Perera et al. 2008; Austermann et al. 2010), the tens of thousands far-IR and sub-mm galaxiesdetected at high-redshifts by the
Herschel observatory(e.g. Berta et al. 2010; Oliver et al. 2010), and thethousands of near IR dropouts (e.g. Caputi et al. 2005;Daddi et al. 2005; Papovich et al. 2006) known to dateare often treated as physically disconnected populations.Spectroscopic surveys with future missions or facilitieswill help us address the above-mentioned points. Ex-pected to be at least one order of magnitude more sen-sitive than their predecessors, future spectrographs willobserve ionic and molecular gas lines, in addition to thedust broadband features and their underlying continua,for large galaxy samples over wide fields of view. In thespirit of demonstrating what can be done in the future inthe IR/submm domain in the field of galaxy evolution, wepresent simulations of the number of galaxies that will beobserved in key IR lines and features at each z range withthe SPace IR telescope for Cosmology and Astrophysics(SPICA) and the Cerro Chajnantor Atacama Telescope(CCAT). We focus on SPICA and CCAT as these areconceived to be large blind survey machines, unlike theJames Webb Space Telescope (JWST) and the AtacamaLarge Millimeter Array (ALMA), that are optimized fordeep single-source observations.The paper is organized as follows. In Section 2we present the simulations, and the physical assump-tions that go into them, including the adopted line-to-bolometric luminosity conversion functions that we mea-sured from observations of local galaxies. The main re-sults of the simulations, i.e., the number of galaxies andAGN that can be detected for each ionized or moleculargas line with z for a given flux limit, are presented in Sec-tion 3, followed by a discussion in Section 4. Throughoutthe paper we will adopt the standard cosmological modelwith Ω M = 0.27, Ω Λ =0.73, H =71 km s − Mpc − . SIMULATING SPECTROSCOPIC SURVEYS IN THE IR
The computation of the number of objects that futuremissions will detect in each line was a two-fold process.First, we used three galaxy evolution models that arebased on galaxy counts and luminosity functions in sev-eral bands of the
Spitzer , AKARI and
Herschel missionsto predict the integrated 8-1000 µ m IR luminosity func-tion at different z bins. We have considered three dif-ferent models with the aim of basing our conclusions onsolid results, and at the same time quantifying the rangeof the predictions. We then computed correlations be-tween the IR luminosity and the luminosities of variousfine-structure lines, molecular lines, and dust emissionfeatures, using samples of local galaxies with completemid- and far-IR spectroscopic coverage. This enabledus to transform the continuum luminosity functions intoline luminosity functions up to z< Predicting continuum luminosity functions with theFranceschini et al. (2010) model
As a first approach, we adopted the model developedby Franceschini et al. (2010). This is a backwardevolution model that fits essentially all available datafrom
Spitzer , ISO, COBE, and SCUBA. Moreover, it in-cludes constraints from preliminary results of the
Her-schel Space Observatory surveys (presented in the Sci-ence Demonstration Phase papers of Oliver et al. 2010;IR spectroscopic cosmological surveys 3
Fig. 1.—
IR (8 − µ m) continuum luminosity function con-structed from the Franceschini et al. (2010) model. Berta et al. 2010; Glenn et al. 2010; Gruppioni et al.2010; Nguyen et al. 2010).The model accounts separately for normal spiral galax-ies, actively star-forming galaxies, and for type 1 and 2AGNs. The actively star-forming population is furthersplit into two galaxy classes, i.e., moderate-luminosityluminous infrared galaxies (LIRGs) and high-luminosityULIRGs. The LIRGs have typical IR luminosities ≃ L ⊙ and the ULIRGs ≃ L ⊙ . They are treatedas two galaxy classes, with different luminosity functions,evolution rates, and spatial clustering properties.The broad-line AGNs–type 1 Seyferts and quasars–were modelled adopting luminosity functions and evolu-tion rates consistent with those observed in optical andmid-IR surveys (e.g. Rush, Malkan & Spinoglio 1993;Spinoglio et al. 1995). As detailed in Franceschini et al.(2010), important constraints on the infrared evolutionproperties of type 1 AGNs have been inferred from a flux-limited sample of 24 µm -selected sources with completespectroscopic classification, as reported by Rodighiero etal. (2010). These objects are easily identified by theirflat spectral shapes over the optical through IR wave-length range (Spinoglio, Andreani & Malkan 2002).Narrow-line type 2 AGNs are instead much more dif-ficult to disentangle from starburst galaxies. Thereforetheir statistical properties and incidence among the IRpopulation at high redshifts are still essentially unknown.At the present stage of knowledge, the model simply con-siders type 2 AGNs as a fraction of both low-luminosityand high-luminosity starbursts. In an attempt to con-strain such fractions, a follow-up analysis by Frances-chini et al. (2011, in preparation) has compared the star-formation and stellar-mass assembly histories of galaxies,and found that the two are consistent with each other if10% of the LIRG and 30% of the ULIRG objects aredominated by obscured (type 2) AGN accretion. These are the fractional contributions for type 2 AGNs adoptedin the present work. It is clear from the above that thismodelling might easily give results inconsistent with thevalue derived from both predictions by the AGN unifi-cation scheme and observations (e.g., from the SLOANand FIRST surveys, Lu et al. 2010) of a ∼ z range considered:0. < z < < z < < z < < z < < z < < z < Fig. 2.—
IR luminosity function from Gruppioni et al. (2011)for various galaxy populations, i.e., normal and starburst galaxies,low luminosity AGN, type 1 AGN and type 2 AGN.
Predicting continuum luminosity functions with theGruppioni et al. (2011) model
Gruppioni et al. (2011) have also developed a back-ward evolution model fitting the main constraints pro-vided by IR/sub-mm surveys in the 15 µ m to 500 µ mrange. In the mid-IR (MIR), it uses data from bothISO and Spitzer that are available in the literature, at15 µ m from the ELAIS-S1 (Gruppioni et al. 2002),HDF-N, HDF-S and Marano fields (Elbaz et al. 1999),ultradeep lensed (Metcalfe et al. 2003), Lockman Deepand Shallow (Rodighiero et al. 2004), and at 24 µ mfrom the GOODS (Papovich et al. 2004) and SWIREsurveys (Shupe et al. 2008). In the FIR, we use datafrom Herschel , including those of the very recent PACSEvolutionary Probe Survey (Berta et al. 2010; Grup-pioni et al. 2010), the
Herschel -ATLAS Survey (Ealeset al. 2010; Clements et al. 2010), and the
Herschel
Multi-tiered Extra-galactic Survey (Oliver et al. 2010;Vaccari et al. 2010). The model uses the classical ap-proach of evolving a local luminosity function in luminos-ity and/or density with z , with a different evolutionary Spinoglio et al.description for galaxies and AGN. What differentiatesthis model from others is the spectral energy distribu-tion (SED) scheme that it uses to distinguish betweenthe different IR-bright galaxy populations. The schemeis based on a large spectroscopic study of MIR-selectedsources (Gruppioni et al. 2008), and it divides sourcesinto five (instead of four) broad SED classes. These arethe normal spiral galaxies, the starburst galaxies, the ob-scured type 2 AGNs, the unobscured type 1 AGNs, andthe objects containing a low luminosity (LL) AGN. Inother words, the authors of this model have put partic-ular emphasis on determining the AGN contribution atlow luminosities at all redshifts. The contribution of thevarious galaxy populations to the IR luminosity functioncan be seen in Figure 2. Fig. 3.—
IR luminosity function from Valiante et al. (2009).
Predicting continuum luminosity functions with theValiante et al. (2009) model
The third model that we used is the backward evolu-tion model of Valiante et al. (2009). It was developedusing
Spitzer and SCUBA observations, and it has beenvery successful in predicting
Herschel results (Berta etal. 2010; Altieri et al. 2010; Glenn et al. 2010; Oliveret al. 2010). This model allows us to take into accountgalaxies that are not ’pure’ starburst or ’pure’ AGN, andfor which the ratio between IR lines might not be theones expected assuming ’pure’ SEDs. This is becausethe model considers all infrared galaxies as a single pop-ulation, assuming that starbursts and AGN coexist. Itthen uses an empirical relation to assign to each galaxythe fraction of the IR luminosity that is powered by theAGN for its given luminosity and redshift. This rela-tion was derived using a complete sample of local IRASgalaxies, and extrapolated to high z using Spitzer andSCUBA observations. Figure 3 shows the IR luminosityfunction predicted by this model for each z bin. Fig. 4.—
Critical density for collisional de-excitation vs. ioniza-tion potential of IR fine-structure lines.
Converting continuum luminosity functions to lineor feature luminosity functions.
The galaxy number counts per redshift and bolomet-ric IR luminosity bin that are predicted by each modelneed to be converted into line luminosity functions, inorder to estimate the number of objects that will bedetectable in various lines, and to assess whether sev-eral of the open questions presented in Section 1 canbe addressed. For this purpose, we derived correla-tions between line and continuum luminosities, includ-ing lines for which such correlations were not previ-ously available in the literature, e.g., that of [SiII] forboth AGN and star-forming galaxies. We examined thePAH feature at 11.25 µ m, the purely rotational H (0-0)S1 line at 17.03 µ m, and the [NeII] 12.8 µ m, [NeV]14.3 µ m, [NeIII] 15.5 µ m, [SIII] 18.7 µ m, [NeV] 24.3 µ m,[OIV] 26 µ m, [SiII] 34.8 µ m, [OIII] 52 µ m, [NIII] 57 µ m,[OI] 63 µ m, [OIII] 88 µ m, [NII] 121.90 µ m, [OI] 145.52 µ m,and [CII] 157.74 µ m fine-structure lines. These lines covera wide parameter space of the critical density vs. ioniza-tion potential diagram (see Fig. 4), tracing different as-trophysical conditions: from photodissociation regions,to stellar/HII regions, to the AGN and coronal line re-gions (Spinoglio & Malkan 1992). This makes the com-bination of their ratios useful for the creation of AGNvs star-formation diagnostic diagrams (e.g. Spinoglio &Malkan 1992; Genzel et al. 1998; Dale et al. 2006;Smith et al. 2007).For lines at wavelengths shorter than 35 µ m, we usedthe complete, 12 µ m-selected sample of local Seyfertgalaxies (Tommasin et al. 2008, 2010) and the Bernard-Salas et al. (2009) sample of starburst galaxies to cali-brate the line luminosities to L IR . These samples havebeen extensively observed in the MIR with the IRS spec-trometer (Houck et al. 2004) onboard Spitzer (Werner etIR spectroscopic cosmological surveys 5al. 2004), and the
Spitzer spectra have been reduced andanalysed in a consistent way. For the starburst galaxies,we excluded all objects for which there was evidence forthe presence of an AGN from the literature or from thedetection of [NeV] (see Table 1 of Bernard-Salas et al.2009). For the long- wavelength lines, we used the het-erogeneous sample of local galaxies compiled by Brauheret al. (2008) containing all observations collected bythe LWS spectrometer (Clegg et al. 1996) onboard ISO(Kessler et al. 1996). The IR luminosities of the galaxiesof our sample have been computed from the IRAS fluxes,using the formula of L IR representing the total mid- andfar-infrared luminosity (Sanders & Mirabel 1996). Allluminosities are in units of 10 erg s − .Using least-squares fitting, we obtained the followingrelations for the Seyfert galaxies, log( L PAH11 . ) = (0 . ± .
07) log( L IR ) − (2 . ± .
21) (1)with R = 0 . , n = 69 , χ = 7 . L [NeII]12 . ) = (0 . ± .
06) log( L IR ) − (3 . ± .
18) (2)with R = 0 . , n = 87 , χ = 8 . L [NeV]14 . ) = (0 . ± .
08) log( L IR ) − (3 . ± .
25) (3)with R = 0 . , n = 81 , χ = 14 . L [NeIII]15 . ) = (0 . ± .
07) log( L IR ) − (3 . ± .
24) (4)with R = 0 . , n = 87 , χ = 15 . L (H )17 . ) = (0 . ± .
05) log( L IR ) − (3 . ± .
15) (5)with R = 0 . , n = 76 , χ = 5 . L [SIII]18 . ) = (0 . ± .
07) log( L IR ) − (3 . ± .
21) (6)with R = 0 . , n = 70 , χ = 7 . L [NeV]24 . ) = (0 . ± .
08) log( L IR ) − (3 . ± .
24) (7)with R = 0 . , n = 71 , χ = 10 . L [ OIV ]25 . ) = (0 . ± .
08) log( L IR ) − (2 . ± .
25) (8)with R = 0 . , n = 83 , χ = 14 . L [SIII]33 . ) = (0 . ± .
05) log( L IR ) − (3 . ± .
17) (9)with R = 0 . , n = 75 , χ = 5 . L [SiII]34 . ) = (1 . ± .
06) log( L IR ) − (3 . ± .
20) (10)with R = 0 . , n = 72 , χ = 7 . where for each relation the Pearson coefficient R, thenumber of considered objects n and the computed χ are given. For the starburst galaxies, the corresponding L IR is computed by fitting a single-temperature dust emis-sivity model ( ǫ ∝ ν − ) to the flux in all four IRAS bands, andshould be accurate to ±
5% for dust temperatures in the range 25- 65 K. We notice that the IR luminosities, as defined above, aremodel-dependent, and therefore could introduce some systemat-ics. However these do not affect the derived relations, as they arewithin the given errors. relations are: log( L PAH11 . ) = (1 . ± .
11) log( L IR ) − (3 . ± .
32) (11)with R = 0 . , n = 14 , χ = 1 . L [NeII]12 . ) = (1 . ± .
14) log( L IR ) − (3 . ± .
40) (12)with R = 0 . , n = 14 , χ = 2 . L [NeIII]15 . ) = (1 . ± .
18) log( L IR ) − (4 . ± . R = 0 . , n = 15 , χ = 3 . L (H )17 . ) = (1 . ± .
14) log( L IR ) − (5 . ± .
42) (14)with R = 0 . , n = 15 , χ = 2 . L [SIII]18 . ) = (1 . ± .
15) log( L IR ) − (3 . ± .
45) (15)with R = 0 . , n = 15 , χ = 3 . L [OIV]25 . ) = (1 . ± .
24) log( L IR ) − (5 . ± .
74) (16)with R = 0 . , n = 12 , χ = 1 . L [SIII]33 . ) = (1 . ± .
10) log( L IR ) − (3 . ± .
29) (17)with R = 0 . , n = 15 , χ = 1 . L [SiII]34 . ) = (1 . ± .
09) log( L IR ) − (3 . ± .
25) (18)with R = 0 . , n = 15 , χ = 0 . The best-fit solution is shown in Figure 5 for each of thepopulations. Considering the sum of the populations, wederived the following generic relations: log( L PAH11 . ) = (0 . ± .
06) log( L IR ) − (2 . ± .
18) (19)with R = 0 . , n = 83 χ = 9 . L [NeII]12 . ) = (0 . ± .
06) log( L IR ) − (3 . ± .
20) (20)with R = 0 . , n = 101 , χ = 17 . L [NeIII]15 . ) = (1 . ± .
07) log( L IR ) − (3 . ± . R = 0 . , n = 102 , χ = 24 . L (H )17 . ) = (1 . ± .
05) log( L IR ) − (4 . ± .
16) (22)with R = 0 . , n = 91 , χ = 10 . L [SIII]18 . ) = (0 . ± .
06) log( L IR ) − (3 . ± .
20) (23)with R = 0 . , n = 70 , χ = 7 . L [OIV]25 . ) = (0 . ± .
11) log( L IR ) − (3 . ± .
34) (24)with R = 0 . , n = 95 , χ = 36 . L [SIII]33 . ) = (0 . ± .
05) log( L IR ) − (3 . ± .
14) (25)with R = 0 . , n = 90 , χ = 7 . L [SiII]34 . ) = (1 . ± .
05) log( L IR ) − (3 . ± .
16) (26)with R = 0 . , n = 87 , χ = 8 . For the far-IR lines (Fig. 6), we obtain: log( L [OIII]51 . ) = (0 . ± .
10) log( L IR ) − (2 . ± .
31) (27)with R = 0 . , n = 16 , χ = 2 . L [NIII]57 . ) = (0 . ± .
10) log( L IR ) − (2 . ± .
32) (28)with R = 0 . , n = 10 , χ = 0 . L [OI]63 . ) = (0 . ± .
03) log( L IR ) − (2 . ± .
10) (29)with R = 0 . , n = 109 , χ = 9 . L [OIII]88 . ) = (0 . ± .
10) log( L IR ) − (2 . ± .
30) (30)with R = 0 . , n = 55 , χ = 12 . L [NII]121 . ) = (1 . ± .
04) log( L IR ) − (3 . ± .
11) (31)with R = 0 . , n = 100 , χ = 13 . L [OI]145 . ) = (0 . ± .
06) log( L IR ) − (3 . ± .
17) (32)with R = 0 . , n = 46 , χ = 10 . L [CII]157 . ) = (0 . ± .
03) log( L IR ) − (2 . ± .
07) (33)with R = 0 . , n = 217 . χ = 42 . Spinoglio et al.
Fig. 5.—
Correlations between the various feature and line luminosities and the far-IR luminosity for the Seyfert galaxies of the complete12 µ m galaxy sample (Tommasin et al. 2008, 2010), and for the pure starburst galaxies of the sample of Bernard-Salas et al. (2009).The dotted, broken and solid lines represent the least-squares fit of the data of the Seyfert, the pure starburst galaxies and all galaxiespopulations together, respectively. Figures (c) and (g) have only the Seyfert galaxies, because the [NeV] lines are not detected in starburstgalaxies. IR spectroscopic cosmological surveys 7
Fig. 6.—
Correlations between the [OIII]52 µ m, [NIII]57 µ m, [OI]63 µ m, [OIII]88 µ m, [NII]122 µ m, [OI]145 µ m and [CII]158 µ m luminosityand the far-IR luminosity for the galaxies observed with the ISO-LWS spectrometer (Brauher et al. 2008). For all correlations the hypothesis that the variablesare unrelated can be rejected at a level of significancewhich is always less that 10 − . Wu et al. (2010) pre-sented the relation between the total infrared luminosityand the PAH emission band at 11.25 µ m for AGN andstarburst galaxies of the 24 µ m flux limited intermediateredshift ( < z > ∼ .
14) sample of 5MUSES (Helou etal. 2011, in preparation). Their result is comparableto ours. By inverting our relations we derive a slope of1.05 ±
011 for the Seyfert galaxies and 0.85 ± ± ± RESULTS
Number counts per spectral line
To compare and visualize our results for the threegalaxy-evolution models, we need to adopt a line-detection sensitivity curve as a function of wavelength λ ,an integration time, and a field of view for the simulatedobservations. For this purpose, we opt to use numbersrelevant to future missions or facilities. For the FIR do-main, we use the sensitivity curve proposed for SPICA’sfar infrared instrument (SAFARI), while for the submil-limeter domain, we use the sensitivity curve that has Spinoglio et al. Fig. 7.—
Number of objects detected per spectral line (and per object type, when applicable) in an hour-long 0.5 deg survey withSPICA SAFARI. been estimated for a R=1000 resolution spectrometer atthe focal plane of the CCAT telescope. Details on theseinstruments are presented in the Appendix. Moreover,we selected a common integration time of 1 hour and atotal field of view to be covered by our simulated surveyof 0.5 deg . For an instrument, such as SAFARI, with a2 ′ × ′ field of view, this corresponds to 450 hours of in-tegration time, to be compared to 4.5 hours for a CCATspectrometer, assuming a field of view of 20 ′ × ′ .The resulting number of AGN and starbursts that willbe detectable in each line as a function of z with SA-FARI is presented in Tables 1 and 2, respectively, forthe Franceschini et al. (2010) model. The same resultsare presented in Tables 3 and 4 for the Gruppioni et al.(2011) model. The number of detectable galaxies basedon the Valiante et al. (2009) model is presented in Ta-ble 5. We note that for the sake of completeness, andto assist further planning and designing of new instru-mentation, we also present in Tables 1-5 the predictednumber of sources that are detectable in the low − Z binsfor the short-wavelength lines and those in the high- Z bins for the long-wavelength lines, even lines outside thenominal SAFARI spectral range. We have simply as-sumed a flat extrapolation of the SAFARI sensitivitiesto shorter and longer wavelengths. Tables 6, 7, and 8present the simulation results for CCAT.A basic result of this analysis is that the total numberof detectable objects agrees to within a factor of 2 − z ranges, and that at least a thousandgalaxies will be simultaneously detected in four lines at5 σ over a half square degree. A comparison of the outputof the three models is plotted in Fig. 7 for SAFARI. Asurvey of the assumed sensitivity will detect bright lines(e.g., [O I ] and [O III ]) and PAH features in thousandsof galaxies at z>
1. Hundreds of z> IV ] line, and several tens of z> V ] and H . For H in particular,this number corresponds to a lower limit. Our models donot account for an increase of the H emission efficiencyas a cooling mechanism, or for the H mass contentto increase with increasing z and decreasing metallic-ity. Neither do the applied continuum-to-line luminosityrelations include sources of extremely high L( H )/ L IR ratios, associated with shock fronts due to galaxy colli-sions or AGN feedback mechanisms (Cluver et al. 2010;Ogle et al. 2010). Our line detectability results are sensible when com-pared with predictions from local galaxy templates.To make this comparison, we used four objects withwell-determined MIR and FIR spectra. These areNGC1068, a prototypical Seyfert 2 AGN, NGC4151, awell-studied type 1 AGN, the prototypical moderate-luminosity starburst M82, and a starburst-dominatedULIRG, IRAS17208-0014. Their line intensities aretaken from Alexander et al. (2000); Spinoglio et al.(2005), Sturm et al. (1999); Spinoglio et al. (1997),Farrah et al. (2007); Brauher et al. (2008) and ForsterSchreiber et al. (2001); Colbert et al. (1999), respec-tively for the four templates. We scaled the bolometricIR luminosity of all systems to 10 L ⊙ , and show whichlines can be observed as a function of z in Figs. 8 and9. This basic comparison confirms that faint lines can bedetected in z ∼ IV ] will be observable in z ∼ II ] out to z ∼ z . We only make this comparison for the Franceschiniet al. (2010) model, as the three models do not differin terms of their IR-luminosity classification, and as thetotal number of sources is comparable in all cases. Wefind that for the brightest lines, such as [Si II ], SAFARIwill be able to observe LIRGS even at z>
3. However, forthe typical [Ne II ] and [O III ] 52 µ m line luminosities,the transition from LIRGs to ULIRGs will occur at z ∼ V ] and H S1 lines at z ∼
1. CCAT willbe highly complementary to SPICA, as it will be able toobserve the [O
III ] 88 µ m line at z> II ], an important coolant of the interstellar medium,at all z<
5. At 3 < z < σ level in a 0.5 deg survey. Line luminosity function predictions in theIR/submm
We present in Fig. 12 and 13 the predicted luminos-ity functions of AGN and starburst galaxies for each lineand feature in the Franceschini et al. (2010) model. It isclear from the figure that the AGN line luminosity func-tions (dashed lines in the figure) for the lowest redshiftrange (0. < z < Fig. 8.— a) Predictions of lines observable with SPICA/SAFARI, based on local templates, scaled to an intrinsic luminosity of 10 L ⊙ .Selected diagnostic lines are shown as a function of z for a type 1 AGN (NGC4151) and a type 2 AGN (NGC1068). For comparison, weoverplot the 5 σ (in 1 hour) sensitivity threshold of SPICA-SAFARI as a function of wavelength (see Appendix). For completeness, we alsoreport the sensitivities for Herschel
PACS (Poglitsch et al. 2010), for the Mid-Infrared Instrument (MIRI, Wright et al. 2004), which willbe onboard the James Webb Space Telescope (Gardner et al. 2006) and which will operate from 5 to 28 µ m, and for the Mid-ir Camera andSpectrograph (MCS) planned for SPICA (Wada & Hirokazu 2010). b) Same as panel a) for a moderate-luminosity prototypical starburst(M82), and a starburst-dominated ULIRG (IRAS17208-0014).
Fig. 9.—
Same as Fig. 8, for long-wavelength lines, where the complementarity between SPICA and CCAT is clearly shown. For theadopted sensitivities of CCAT we refer to the Appendix.
Fig. 10.—
Prediction of the number of sources of a 0.5 deg spectroscopic survey with SAFARI based on Franceschini et al. (2010),giving the number of detectable starburst galaxies (divided by L IR ) and AGN (divided by obscuration) at the 3 σ level. The adopted lineflux sensitivities as a function of wavelength are given in the Appendix. The left panels correspond to the AGN predictions, the rightpanels to the starburst predictions. As for the former, the number of type 2 AGNs associated with the LIRG and the ULIRG populationsare shown separately (in black and red, respectively). nosity functions, at an average < z > of 0.03. The spacedensities of starburst galaxies are expected to be higherthan those of AGN, for any line, except for the [OIV]line, which–as is well known–is much fainter in starburstgalaxies. The volume densities of AGN drop faster with z than those of starbursts when traced by the [Ne II ] and[Ne III ] lines, possibly due to a saturation of the lines inAGN of high luminosities. The total number of AGNdetected in the same lines is one to two orders of mag- nitude lower than that detected for starbursts at any z .The number of sources detected in [O III ] 88 µ m is com-parable for AGN and starbursts at all z . This result mostlikely differs from that for [Ne III ] which comes from anion of comparable ionization potential to O
III , becauseof the single relationship used to convert the line luminos-ity to L IR for long wavelength lines. This is even truefor the [O I ] 63 µ m line, which also has a high criticaldensity for collisional de-excitation, ∼ cm − , mak-IR spectroscopic cosmological surveys 11 Fig. 11.—
Prediction of the number of sources of a 0.5 deg spectroscopic survey based on Franceschini et al. (2010), givingthe number of detectable galaxies at the 3 σ level with CCAT. Theadopted line flux sensitivities as a function of wavelength are re-ported in the Appendix. Fig. 12.—
Predicted line luminosity functions of [NeV]14.3 µ m,[NeIII]15.5 µ m, [NeV]24.3 µ m, [OIV]26 µ m, [OIII]52 µ m and[OIII]88 µ m, for SAFARI. Dashed lines correspond to the predic-tions for AGN, while solid lines correspond to the predictions forstarbursts. Where available, the comparison with the observedlocal LF of AGNs from Tommasin et al. (2010) is given. ing it bright in AGN. The contamination of the [O IV ]AGN luminosity functions from starbursts is minimal allthe way through z =4. The volume density of activelyaccreting black holes as traced by [O IV ] and [Ne V ] in- Fig. 13.—
Predicted line luminosity functions of PAH 11.25 µ m,[NeII]12.8 µ m, H µ m, [SiII]34.8 µ m , [NIII]57 µ m and [OI]63 µ m,for SAFARI. Dashed lines correspond to the predictions for AGN,while solid lines correspond to the predictions for starbursts.Where available, the comparison with the observed local LF ofAGNs from Tommasin et al. (2010) is given. creases up to 1 . z . z =4.The same applies for tracers of star-formation, reproduc-ing the suggested coevolution of black hole growth andstellar mass build-up. DISCUSSION: NEW PARAMETER-SPACE COVERAGEBY SPICA AND CCAT
Which of the questions raised in the Introductionwill the proposed future telescopes address? The linespredominantly emitted by ions in star-forming com-plexes, like [Si II ], [C II ], and [Ne II ], will be detected in L IR > L ⊙ systems at least out to z ∼
2. They will alsobe detected in 10 L ⊙ < L IR < L ⊙ galaxies at leastout to z ∼
1. This result indicates that star-formationtracers will be detected and compared for the sourcesthat are mainly responsible for the formation of present-day ellipticals. The creation of line-ratio diagnostic di-agrams, and the comparison of the line to IR contin-uum luminosities in hundreds of sources will help us, incombination with imaging data, to further address thestar-formation bimodality, and to obtain a more coher-ent picture of the intermediate/high- z IR-bright popula-tions. Several sources will be detected with CCAT evenpast the peak of star-formation activity at z> II ] line, which will help us constrain the shapeof Lilly-Madau diagrams at such high redshifts.To date, the detection of FIR fine-structure lines hasonly been achieved in lensed z> Herschel
PACS detected the [O
III ] 52 µ m line with a flux of 9 × − Wm − in IRAS F10214+4724 at z =2.28 (Sturm etal. 2010). In a z =1.32 source, MIPS J142824.0+352619,[O III ] 52 µ m and [O I ] 63 µ m were detected with fluxesof 3.7 and 7.8 × − Wm − , respectively (Sturm et al.2 Spinoglio et al.2010). The [O III ] 88 µ m line was detected at the CaltechSubmillimeter Observatory (CSO) with the ZEUS spec-trometer (Ferkinhoff et al. 2010b) in APM 08279+5255at z =3.9 and SMM J02399-0136 at z =2.8 (Ferkinhoff etal. 2010a), with fluxes of 2.68 and 6.04 × − Wm − .All four systems are lensed galaxies, with magnificationfactors estimated to be in the range 2.4 −
90 (Egami etal. 2000; Ao et al. 2008; Riechers et al. 2009; Ivisonet al. 2010). Such experiments were not possible forunlensed galaxies with the present-day missions, leavingample room for new discoveries for SPICA and CCAT.The recent detection of [NII]122 µ m (Ferkinhoff et al.2011) and [CII]158 µ m (Stacey et al. 2010) lines in afew high redshift galaxies with the ZEUS spectrometer(Ferkinhoff et al. 2010b) at the CSO shows that theselines can be much brighter than in local galaxies. The[NII]122 µ m line to FIR luminosity ratio is 2-10 timeshigher in the two observed galaxies (H1413+117 andSMMJ02399-0136) than in the local galaxies that we usedfor deriving the line to continuum relations adopted forour predictions. Similarly the [CII]/FIR luminosity ra-tio in the 1 < z < line flux and mass indicates that detectionof H -bright galaxies at z> × − W m − at a typical tem-perature of 300K, yields an H mass of 3 × M ⊙ at z =6. Cosmological simulations indicate that z ∼ to cool the ISM, or unless there is aconsiderable number of sources with an AGN jet-ISM in-teraction that leads to very high L( H )/ L IR ratios (Ogleet al. 2010), our conclusion will not hold at z ∼
10. Atthat z , the same temperature and S1 flux correspond toa mass of 10 M ⊙ . Even though the H (0-0)S0 lineat 28.03 µ m will be in the spectral range of CCAT at z>
6, its detection will be equally hard: assuming an S0flux of 2 × − W m − and an ortho-to-para ratio of 3,the minimum detectable H mass would be ∼ M ⊙ at z =6. SPICA will thus be unique for H studies seekingthe end of the reionization era.Another parameter space unique to SPICA will bethe detection of resolved [O IV ] 25.89 µ m emission upto z ∼
4. The calibration of the widths of high-ionizationMIR lines, as probes of the narrow-line-region kinemat-ics, to the black hole mass was attempted with
Spitzer for sources out to z =0.3 (Dasyra et al. 2008, 2011).It might be useful for the mid-infrared instrument ofJWST, which will be able to observe the [Ne V ] 14.32 µ mline out to z ∼ z> z > SUMMARY
We used three galaxy formation and evolution mod-els, constrained by luminosity functions from multi-wavelength observations, to predict the number of star-forming and active galaxies that are observable in any IRluminosity bin with look-back time. We converted thenumber counts of galaxies per IR luminosity to numbercounts of galaxies per line luminosity using several (new)line-to-continuum conversion relations, built upon localAGN and starburst galaxy samples. We compute our re-sults for an hour-long integration/FoV half-square-degreesurvey, and for the sensitivity values of SPICA SAFARIand CCAT. Their anticipated values are 2.5 × − Wm − at 160 µ m and 1.1 × − W m − at 620 µ m, re-spectively. These telescopes/instruments were selectedbecause they are designed to be survey machines, ableto perform large blind cosmological surveys in reason-able integration times. We find that SAFARI will de-tect thousands of z> II ] and [Ne II ], and sev-eral tens of 1 < z < IV ], [Ne V ] and H (0-0)S1. For the bright lines, nor-mal galaxies will be observed out to z ∼
1, LIRGs outto z ∼
2, and ULIRGs to even higher z . This meansthat studies of the ionized gas properties in the galax-ies that form the present day massive ellipticals will befeasible. AGN/star-formation diagnostic diagrams willbe obtained for different classes of IR-bright galaxies at1 < z <
2, which will enable us to not only look fora redshift evolution of line ratios, but also for a lumi-nosity evolution within each z range. Further tests ofthe black hole growth - galaxy build-up coevolution sce-nario will be performed, as the creation of accretion-ratefunctions and mass functions will be determined out to z ∼
4, for both obscured and unobscured black holes usingthe [O IV ] line. Over large areas, SAFARI could also beable to detect H at z &
6, and help constrain the end ofthe reionization era. In the light of the current findingsat ground-based submillimeter telescopes of substantiallybrighter fine structure lines in high redshift galaxies com-pared to local galaxies, we are even more confident thatCCAT will be unique for studying the star formation his-tory of galaxies back to very high redshift.We acknowledge input from Takao Nakagawa, PI ofthe SPICA Mission, and from Peter Roelfsema, FrankHelmich and Bruce Swinyard, PI and members of theSAFARI Consortium, respectively. We also acknowledgeGordon Stacey, Simon Radford, Jason Glenn and Ric-cardo Giovanelli for information on the CCAT projectand its planned instrumentation. We thank Matt Malkanwho commented on this manuscript and Scott Douglas,Nicola Sacchi, Silvia Tommasin, Anna Di Giorgio, JohnScige Liu and Erina Pizzi for assisting us in improvingthis document. We also thank the anonymous refereefor a very thorough and constructive report. This workis based on observations made with the
Spitzer
SpaceIR spectroscopic cosmological surveys 13Telescope which is operated by the Jet Propulsion Lab-oratory and Caltech under a contract with NASA. K.M. D. acknowledges support by the European Commu- nity through a Marie Curie Fellowship (PIEF-GA-2009-235038) awarded under the Seventh Framework Pro-gramme (FP7/2007-2013).
APPENDIX
DESCRIPTION OF THE SPICA MISSION AND THE CCAT FACILITY TO BE USED FOR FUTURE,BLIND COSMOLOGICAL SURVEYS IN THE IR AND SUBMM
The deep cosmological surveys undertaken by ISO (Kessler et al. 1996),
Spitzer (Werner et al. 2004), AKARI(Murakami et al. 2007; Goto et al. 2010), WISE (Wright et al. 2010), and
Herschel (Pilbratt et al. 2010) will haveproduced catalogues containing the fluxes of many tens of thousands of IR-bright sources by the early 2020’s. Thesecatalogues will provide excellent targets to be followed up by ALMA (Brown, Wild & Cunningham 2004; Wootten2008), JWST (Gardner et al. 2006), SPICA (Swinyard et al. 2009), and CCAT (Sebring 2010). Among the listedfacilities, JWST and ALMA will be more suited to deep follow-up spectroscopy of known targets. SPICA and CCATwill be suited for performing blind large-scale spectro-photometric surveys, because of the wide field of view (of severalarcminutes squared) of their instruments. Covering different wavelength ranges, these two instruments will be highlycomplementary. In the rest of the Appendix, we provide details of their present-day design concepts.SPICA is a proposed JAXA-led astronomical mission with suggested contributions by European, Korean, andpossibly US institutions, and with a launch date planned in the early 2020s (Swinyard et al. 2009). With a 3-mmirror that is actively cooled to < µ m spectral range with unprecedented sensitivity. It will offer an improvementin raw photometric sensitivity with respect to Herschel of two orders of magnitude in the FIR. In the MIR, it willextend the capabilities of JWST with an uninterrupted spectral coverage in the range 5-38 µ m. The Mid InfraredCamera and Spectrometer (MCS) planned for SPICA (Wada & Hirokazu 2010; SPICA Study Team Collaboration2010) is an integral field unit with a field of view of 12 ′′ × ′′ at 10-20 µ m and 12 ′′ × ′′ at 19.5-36.1 µ m. Its5 σ , 1 hour sensitivity is in the range 2-2.5 × − W m − (Wada & Hirokazu 2010). Observing capability in theFIR is provided for SPICA by SAFARI (Swinyard et al. 2009; SPICA Study Team Collaboration 2010). Proposedby a consortium of European institutes (with Canadian and Japanese participation), SAFARI is an imaging FourierTransform Spectrometer (FTS) with a field of view of 2 ′ × ′ . The FTS provides an instantaneous spectral coverage ofthe 34 - 210 µ m wavelength range and spectral resolution modes with λ /∆ λ of 2000 (at 100 µ m), ∼ few hundred, or evenas low as 20 < λ /∆ λ <
50. With its sensitive superconducting transition edge sensor (TES) detectors (Khosropanah etal. 2010), SAFARI will offer a factor of ∼
100 increase in raw sensitivity in the continuum, and ∼
15 in high-resolutionmode spectroscopy compared to
Herschel
PACS (Poglitsch et al. 2010). The improvement in spectral mapping speedover PACS will be of more than 100 at λ /∆ λ ∼ × − W/ √ Hz for the SAFARI detectors, the 5 σ , 1 hour detection limits are predicted to be 4.16, 2.58, 1.89 and 2.48 × − Wm − for the four planned spectral bands centred at 48, 85, 135 and 160 µ m, respectively, for an unresolved line at λ /∆ λ ∼ . The designof the CCAT telescope is currently undergoing changes to increase the field of view from 20 ×
20 arcminutes square to1 square degree (Sebring 2010). This would substantially increase the survey capability of the telescope. CCAT willcarry out spectroscopic surveys of submm galaxies, using multi-object versions of broadband direct-detection gratingspectrometers such as Z-Spec (Bradford et al. 2004) and ZEUS (Ferkinhoff et al. 2010b), now in use at the CSO.Conceptual development indicates that spectrometers capable of observing 10-100 objects simultaneously while span-ning multiple atmospheric windows will be feasible (Stacey et al. 2006) . Sensitivity estimates based on the 25 mtelescope CCAT telescope with a 10 µ m rms surface on Cerro Chajnantor (5600 m elevation) for a spectrometer witha resolution of 1000 (5 σ , 1 hour) that include wavelength-dependent, typical precipitable water vapor corrections are:2.0, 1.9 and 1.0 × − Wm − at 200 µ m, 230 µ m and 291 µ m respectively, 1.8, 1.3 and 1.1 × − Wm − at 350 µ m,450 µ m and 620 µ m respectively and 3.8 and 2.6 × − Wm − at 740 µ m and 865 µ m respectively (Stacey, G. 2011,private comm.). REFERENCESAlexander, T. et al. 2000, ApJ, 536, 710Altieri, B., et al. 2010, A&A, 518, L17Ao, Y., Weiß, A., Downes, D., Walter, F., Henkel, C., Menten, K.M. 2008, A&A, 491, 747Austermann, J.E., et al. 2010, MNRAS, 401, 160Barger, A.J., et al. 1998, Nature, 394, 248Barmby, P. et al. 2006, ApJ, 642, 126Bernard-Salas, J., Spoon, H. W. W.; Charmandaris, V. et al. 2009,ApJS, 184, 230Berta, S., et al. 2010, A&A, 518, L30 Borys, C., Chapman, S., Halpern, M., Scott, D. 2003, MNRAS,344, 385Brown, R.L., Wild, W. & Cunningham, C. 2004, AdSpR, 34, 555Bradford, C.M., et al. 2004, SPIE, 5498, 257Brauher, J.R., Dale, D.A., Helou, G. 2008, ApJS, 178, 280Caputi, K. I., Dunlop, J. S., McLure, R. J., Roche, N. D., 2005,MNRAS, 361, 607Clegg, P.E. et al 1996, A&A, 315, L38Clements, D.L., et al. 2010, A&A, 518, L8Cluver, M. E., et al. 2010, ApJ, 710, 248Colbert, J. et al. 1999, ApJ, 511, 721
Coppin, K. et al. 2006, MNRAS, 372, 1621Cowie, L.L. et al. 1996, AJ, 112, 839Daddi, E., et al. 2005, ApJ, 631, L13Daddi, E. et al. 2010, ApJ, 714, L118Dale, D.A. et al. 2006, ApJ, 646, 161Dasyra, K. M., et al. 2006, ApJ, 638, 745Dasyra, K. M., et al. 2008, ApJ, 674, L9Dasyra, K. M., et al. 2009, ApJ, 701, 1123Dasyra, K. M., Ho, L. C., Netzer, H., Combes, F., Trakhtenbrot,B., Sturm, E., Armus, L., & Elbaz, D., ApJ, in press,arXiv:1107.3397Eales, S., et al. 2010, PASP, 122, 499Egami, E., Neugebauer, G., Soifer, B. T., Matthews, K., Ressler,M., Becklin, E. E., Murphy, T. W., Jr., Dale, D. A. 2000, ApJ,535, 561Elbaz D. et al. 1999 A&A, 351, L37Fabian, A.C. 1999, MNRAS, 308, L39Fadda, D. et al. 2004, AJ....128....1Farrah, D. et al.. 2007, ApJ, 667, 149Ferkinhoff, C., Hailey-Dunsheath, S., Nikola, T., Parshley, S.C.,Stacey, G.J., Benford, D.J., Staguhn, J.G. 2010a, ApJL, 714,L147Ferkinhoff, C., Nikola, T., Parshley, S.C., Stacey, G.J., Irwin, K.D.,Cho, H-M., & Halpern, M. 2010b, SPIE, 7741, E26FFerkinhoff, C., et al. 2011, ApJ, 740, L29Ferrarese, L. & Merrit, D. 2000, ApJ, 539, L9Ferrarese, L. & Ford, H., 2005, Space Science Rev., 116, 523Forster Schreiber, N.M., Genzel, R., Lutz, D., Kunze, D.,Sternberg, A. 2001, ApJ, 552, 544Franceschini, A., Rodighiero, G., Vaccari, M., Berta, S., Marchetti,L., Mainetti, G. 2010, A&A, 517, 74FGardner, J.P. et al 2006, SSRv, 123, 485Genzel, R., et al. 1998, ApJ, 498, 579Genzel, R. et al 2010, MNRAS, 407, 2091Gilli, R., Comastri, A., & Hasinger, G. 2007, A&A, 463, 79Glenn, J. et al. 2010, MNRAS, 409, 109Goto, T., et al. 2010, A&A, 514A, 6Gracia-Carpio J. et al. 2011, ApJ, 728, L7Greve, T.R., et al. 2004, MNRAS, 354, 779Gruppioni C. et al. 2002, MNRAS, 335, 831Gruppioni, C. et al. 2008, ApJ, 684, 136Gruppioni, C. et al. 2010, A&A, 518, L27Gruppioni, C., Pozzi, F., Zamorani, G., Vignali, C. 2011, MNRAS,in press, and arXiv:1105.1955Gultekin, K., et al. 2009, ApJ, 698, 198Houck, J.R. et al. 2004, ApJS, 154, 18Houck, J.R., et al. 2005, ApJ, 622, 105Huang, J.-S., et al. 2009, ApJ, 700, 183Hughes, D.H. et al. 1998, Nature, 394, 24Kessler, M.F. et al. 1996, A&A, 315, L27Khosropanah, P. et al 2010, Proc. SPIE, 7731, E14Kormendy, J. & Richstone, D. 1992, ApJ, 393, 559Ivison, R. J., Smail, I., Papadopoulos, P.P., Wold, I., Richard, J.,Swinbank, A.M., Kneib, J.-P., Owen,F.N. 2010, MNRAS, 404,198Lacy, M. et al. 2004, ApJS, 154, 166Laurent, G.T., et al. 2005, ApJ, 623, 742Le Floc’h, E., et al. 2004, ApJS, 154, 170Lu, Y., et al. 2010, MNRAS, 404, 176Ly, C., Malkan, M., Hayashi, M., Motohara, K., Kashikawa, N.,Shimasaku, K., Nagao, T., Grady, C. 2011, ApJ, 735, 91Magorrian, J. et al. 1998, AJ, 115, 2285Marconi, A., Risaliti, G., Gilli, R., Hunt, L.K., Maiolino, R.,Salvati, M. 2004, MNRAS, 351, 169Merloni, A., Rudnick, G., Di Matteo, T. 2004, MNRAS, 354, L37Metcalfe L. et al. 2003, A&A, 407, 791Mor, R., Netzer, H., Elitzur, M. 2009, ApJ, 705, 298Mullaney, J.R. et al. 2010, MNRAS, 401, 995Murakami, H. et al. 2007, PASJ, 59, 369Nguyen, H. T., Schulz, B., Levenson, L., et al. 2010, A&A, 518, L5Obreschkow, D. & Rawlings, S. 2009, ApJ, 696, L129 Oliver, S., et al. 2010, A&A, 518, L210Ogle, P., Boulanger, F., Guillard, P., Evans, D. A., Antonucci, R.,Appleton, P. N., Nesvadba, N., & Leipski, C. 2010, ApJ, 724,1193Papovich, C. et al. 2004, ApJS, 154, 70Papovich, C. et al. 2006, AJ, 132, 231Perera, T.A. et al. 2008, MNRAS, 391, 1227Perez-Gonzalez, P.G. et al. 2005, ApJ, 630, 82Pilbratt, G.L. et al. 2010, A&A, 518, L1Poglitsch, A. et al. 2010, A&A, 518, L2Powell, L.C., et al., 2011arXiv1102.4195PReddy N.A., et al. 2010, ApJ, 712, 1070Richstone, D. et al. 1998, Nature, 395, 14Riechers, D.A., Walter, F., Carilli, C.L., Lewis, G.F. 2009, ApJ,690, 463Rigby, J.R., et al. 2004, ApJS, 154, 160Rodighiero G. et al., 2004, A&A, 427, 773Rodighiero, G., Vaccari, M., Franceschini, A., et al., 2010, A&A,515, A8Rush, B., Malkan, M.A. & Spinoglio, L. 1993, ApJS, 89, 1Sanders, D.B. & Mirabel, I.F. 1996, ARA&A, 34, 749Scott, S.E., et al 2002, MNRAS, 331, 817Scott, K.S., et al. 2008, MNRAS, 385, 2225Sebring, T., 2010, SPIE, 7733, 59Shankar, F., Bernardi, M., Haiman, Z. 2009, ApJ, 694, 867Shim, H., Colbert, J., Teplitz, H., Henry, A., Malkan, M.,McCarthy, P., Yan, L. 2009, ApJ, 696, 785Shupe D. L. et al. 2008, AJ, 135, 1050Smail, I., Ivison, R. J., Blain, A. W. 1997, ApJ, 490, L5Smith, J.D.T. et al. 2007, ApJ, 656, 770SPICA Study Team Collaboration, 2010. SPICA Assessment StudyReport for ESA Cosmic Vision 2015-2025 Plan, ESA/SRE(2009)6 December 2009, arXiv1001.0709S.Spinoglio, L. & Malkan, M.A. 1992, ApJ, 399, 504Spinoglio, L., et al. 1997, Extragalactic Astronomy in the Infrared.Edited by G. A. Mamon, Trinh Xuan Thuan, and J. Tran ThanhVan. Paris: Editions Frontieres, p.333Spinoglio, L., Malkan, M., Rush, B., Carrasco, L., Recillas-Cruz,E. 1995, ApJ, 453, 616Spinoglio, L., Andreani, P., Malkan, M., 2002, ApJ, 572, 105Spinoglio, L., Malkan, M.A., Smith, H.A., Gonz´alez-Alfonso, E.,Fischer, J. 2005, ApJ, 623, 123Stacey, G. J. et al 2006, SPIE, 6275, 47Stacey, G. J. et al 2010, ApJ, 724, 957Stern, D. et al. 2005, ApJ, 631, 163Sturm, E. et al. 1999, ApJ, 512, 197Sturm, E. et al. 2010, A&A, 518, L36Schweitzer, M. et al. 2006, ApJ, 649, 79Swinyard, B., Nakagawa, T. et al. 2009, Experimental Astronomy,23, 193Tremaine, S. et al. 2002, ApJ, 574, 740Tommasin, S., Spinoglio, L., Malkan, M.A., Smith, H., Gonzlez-Alfonso, E., Charmandaris, V. 2008, ApJ, 676, 836.Tommasin, S., Spinoglio, L., Malkan, M., Fazio, G., 2010, ApJ, 709,1257Treu, T., Malkan, M., Blandford, R., 2004, ApJ, 615, 97Valiante, E., Lutz, D., Sturm, E., Genzel, R., Chapin, E.L. 2009,ApJ, 701, 1814Vaccari, M., et al. 2010, A&A, 518, L20Vestergaard, M., & Osmer, Patrick S, 2009, ApJ, 699, 800Wada, T. & Hirokazu, K. 2010, Proc. SPIE, 7731, E23Webb, T.M., et al. 2003, ApJ, 582, 6Werner, M.W. et al. 2004, ApJS, 154, 1Wise, J. H., & Abel, T. 2007, ApJ, 671, 1559Wootten, A. 2008, Ap&SS, 313, 9Wright, E.L. et al 2010, AJ, 140, 1868Wright, G.S. et al. 2004, Proc. SPIE 5487, 653Wu, Y. et al. 2010, ApJ, 723, 895
IR spectroscopic cosmological surveys 15
TABLE 1Number of AGN detectable in a SAFARI survey of 0.5 in IR lines as a function of redshift at 5 σ (3 σ ) in 1 hr.integration per FoV, following Franceschini et al. (2010) line/redshift 0 < z < < z < < z < < z < < z < < z < µ m) 687. (945.) † † † · · · (1.35) 60.0 (187.) ‡ [ NeII ] 12.81 µ m 152. (366.) † † · · · (0.45) · · · ( · · · ) 15.3 (52.2) ‡ [ NeV ] 14.32 µ m 43.2 (152.) † † · · · ( · · · ) · · · ( · · · ) · · · ( · · · ) 0.45 (5.85) ‡ [ NeIII ] 15.55 µ m 152. (366.) † · · · (0.45) · · · ( · · · ) 41.4 (130.) ‡ H (17.03 µ m) 10.8 (43.2) † · · · (6.75) · · · (0.45) · · · ( · · · ) · · · ( · · · ) · · · ( · · · ) · · · (7.20) ‡ [ SIII ] 18.71 µ m 43.2 (106.) † · · · (0.90) · · · ( · · · ) · · · ( · · · ) · · · ( · · · ) 46.3 (118.) ‡ [ NeV ] 24.32 µ m 26.6 (69.8) 11.2 (55.8) 14.4 (45.0) 0.90 (6.75) · · · (0.90) · · · ( · · · ) 53.1 (178.)[ OIV ] 25.89 µ m 152. (366.) 82.3 (176.) 73.8 (246.) 17.1 (54.0) 0.90 (14.8) · · · ( · · · ) 326. (857.)[ SIII ] 33.48 µ m 69.8 (210.) 122. (247.) 45.0 (174.) 6.75 (32.8) 0.90 (14.8) · · · ( · · · ) 244. (679.)[ SiII ] 34.81 µ m 210. (366.) 333. (633.) 174. (439.) 54.0 (204.) 14.8 (121.) 1.35 (11.7) 787. (1775.)[ OIII ] 51.81 µ m 464. (687.) 333. (633.) 246. (563.) 54.0 (204.) 14.8 (28.8) · · · (0.90) † ‡ [ NIII ] 57.32 µ m 106. (281.) 55.8 (176.) 14.4 (73.8) · · · (0.90) · · · ( · · · ) · · · ( · · · ) 176. (532.)[ OI ] 63.18 µ m 687. (945.) 1128. (1817.) 563. (863.) 204. (381.) 47.7 (179.) 4.50 (22.5) † ‡ [ OIII ] 88.35 µ m 687. (945.) 459. (856.) 246. (563.) 88.2 (204.) † † † ‡ * Notes: † : outside the SAFARI spectral range; ‡ : excluding detectionsoutside the SAFARI spectral range. TABLE 2Number of starburst galaxies detectable in a SAFARI survey of 0.5 in IR lines as a function of redshift at 5 σ (3 σ ) in1 hr. integration per FoV, following Franceschini et al. (2010) line/redshift 0 < z < < z < < z < < z < < z < < z < µ m) 1023. (1519.) † † † ‡ [ NeII ] 12.81 µ m 636. (1023.) † † ‡ [ NeIII ] 15.55 µ m 115. (254.) † ‡ H (17.03 µ m) 28.8 (43.2) † · · · ( · · · ) · · · ( · · · ) 12.6 (39.1) ‡ [ SIII ] 18.71 µ m 176. (355.) † · · · (0.45) 250. (566.) ‡ [ OIV ] 25.89 µ m 2.70 (9.45) 1.80 (9.90) 1.35 (7.65) 0.45 (0.90) · · · ( · · · ) · · · ( · · · ) 6.30 (27.9)[ SIII ] 33.48 µ m 355. (636.) 495. (974.) 337. (608.) 119. (277.) 64.3 (164.) 6.30 (30.1) 1377. (2689.)[ SiII ] 34.81 µ m 482. (816.) 974. (1790.) 608. (984.) 277. (519.) 164. (339.) 30.1 (81.4) 2535. (4529.)[ OIII ] 51.81 µ m 816. (1257.) 495. (974.) 337. (783.) 72.9 (277.) 2.25 (39.1) · · · (0.90) † ‡ [ NIII ] 57.32 µ m 176. (482.) 80.5 (258.) 19.3 (101.) · · · (1.80) · · · ( · · · ) · · · ( · · · ) 276. (843.)[ OI ] 63.18 µ m 1257. (1808.) 1790. (2960.) 783. (1221.) 277. (519.) 64.3 (242.) 6.30 (30.1) † ‡ [ OIII ] 88.35 µ m 1257. (1808.) 692. (1340.) 337. (783.) 119. (388.) † † † ‡ * Notes: † : outside the SAFARI spectral range; ‡ : excluding detectionsoutside the SAFARI spectral range. TABLE 3Number of AGN detectable in a SAFARI survey of 0.5 in IR lines as a function of redshift at 5 σ (3 σ ) in 1 hr.integration per FoV, following Gruppioni et al. (2011) line/redshift 0 < z < < z < < z < < z < < z < < z < µ m) 1057.(1451.) † † † ‡ [ NeII ] 12.81 µ m 398.( 701.) † † · · · ( · · · ) 14.0 (43.2) ‡ [ NeV ] 14.32 µ m 177.( 398.) † † · · · (0.90) · · · ( · · · ) · · · ( · · · ) 0.90 (6.75) ‡ [ NeIII ] 15.55 µ m 398.( 701.) † · · · (0.45) 24.7 (95.3) ‡ H (17.03 µ m) 82.3( 177.) † · · · (0.90 ) · · · ( · · · ) · · · ( · · · ) · · · ( · · · ) 0.45 (4.05) ‡ [ SIII ] 18.71 µ m 177.( 398.) † · · · (0.45) · · · (0.45) · · · ( · · · ) 2.25 (10.3) ‡ [ NeV ] 24.32 µ m 124.( 240.) 5.40 ( 32.8) 9.90 (28.8 ) 3.60 (12.6) 0.45 (3.60) · · · ( · · · ) 143. (318.)[ OIV ] 25.89 µ m 491.( 701.) 58.0 ( 192.) 47.7 (187. ) 22.0 (64.3) 3.60 (25.2) 0.90 (4.05) 623. (1174.)[ SIII ] 33.48 µ m 240.( 491.) 111. ( 326.) 28.8 (121. ) 12.6 (39.1) 6.75 (25.2) 0.90 (4.05) 400. (1006.)[ SiII ] 34.81 µ m 491.( 701.) 512. (1076.) 121. (486. ) 64.3 (209.) 45.9 (170.) 7.65 (28.3) 1242. (2670.)[ OIII ] 51.81 µ m 814.(1184.) 512. (1076.) 188. (759. ) 64.3 (209.) 6.75 (45.9) · · · ( · · · ) 1585. (3274.)[ NIII ] 57.32 µ m 398.( 701.) 32.8 ( 192.) 9.90 (47.7 ) 0.45 (3.60) · · · (0.45) · · · ( · · · ) 441. (945.)[ OI ] 63.18 µ m 1057.(1316.) 1862. (2805.) 759. (1593.) 209. (391.) 75.6 (230.) 15.7 (52.6) † ‡ [ OIII ] 88.35 µ m 1057.(1316.) 762. (1443.) 188. (759. ) 100. (209.) † † † ‡ * Notes: † : outside the SAFARI spectral range; ‡ : excluding detectionsoutside the SAFARI spectral range. TABLE 4Number of Starburst galaxies detectable in a SAFARI survey of 0.5 in IR lines as a function of redshift at 5 σ (3 σ )in 1 hr. integration per FoV, following Gruppioni et al. (2011) line/redshift 0 < z < < z < < z < < z < < z < < z < µ m) 515. (779.) † † † ‡ [ NeII ] 12.81 µ m 326. (515.) † † ‡ [ NeIII ] 15.55 µ m 62.1 (115.) † ‡ H (17.03 µ m) 26.6 (46.3) † · · · (0.45) · · · ( · · · ) 17.1 (40.5) ‡ [ SIII ] 18.71 µ m 115. (199.) † · · · (0.45) 64.7 (144.) ‡ [ OIV ] 25.89 µ m 8.10 (15.3) 3.60 (9.90) 3.15 (11.2) 0.90 (1.80) · · · ( · · · ) · · · ( · · · ) 15.8 (38.2)[ SIII ] 33.48 µ m 199. (326.) 180. (245.) 171. (270.) 71.1 (140.) 31.5 (69.8) 5.85 (18.4) 658. (1069.)[ SiII ] 34.81 µ m 256. (412.) 245. (329.) 270. (373.) 140. (240.) 69.8 (136.) 18.4 (46.4) 999. (1536.)[ OIII ] 51.81 µ m 92.7 (219.) 180. (245.) 171. (321.) 48.1 (140.) 4.05 (19.8) · · · (0.45) 496. (945..)[ NIII ] 57.32 µ m 153. (326.) 52.2 (122.) 18.4 (64.8) 0.45 (4.05) · · · ( · · · ) · · · ( · · · ) 224. (517.)[ OI ] 63.18 µ m 637. (939.) 329. (459.) 321. (422.) 140. (240.) 31.5 (99.0) 5.85 (18.4) † ‡ [ OIII ] 88.35 µ m 637. (939.) 212. (284.) 171. (321.) 71.1 (140.) † † † ‡ * Notes: † : outside the SAFARI spectral range; ‡ : excludingdetections outside the SAFARI spectral range. TABLE 5Total number of galaxies detectable in a SAFARI survey of 0.5 in IR lines as a function of redshift at 5 σ (3 σ ) in 1 hr.integration per FoV, following Valiante et al. (2009) line/redshift 0 < z < < z < < z < < z < < z < < z < µ m) 3353 (5602) † † † ‡ [ NeII ] 12.81 µ m 1114 (1811) † † · · · ( · · · ) 42.3 (201.) ‡ [ NeIII ] 15.55 µ m 850 (1429) † · · · (4.95) 179. (507.) ‡ H (17.03 µ m) 158 (331) † · · · ( · · · ) · · · ( · · · ) · · · ( · · · ) 6.3 (24.8) ‡ [ SIII ] 18.71 µ m 158 (331) † · · · ( · · · ) · · · (1.80) · · · ( · · · ) 6.3 (26.5) ‡ [ OIV ] 25.89 µ m 638 (1114) 94.0 (252.) 136. (327.) 49.1 (211.) 15.8 (51.8) · · · (4.95) 938. (1961.)[ SIII ] 33.48 µ m 638 (1114) 251. (879.) 136. (327.) 49.1 (211.) 29.3 (142.) · · · (4.95) 1103. (2678.)[ SiII ] 34.81 µ m 1114 (1811) 879. (1761.) 505. (1074.) 322. (676.) 207. (452.) 9.9 (61.2) 3037. (5835.)[ OIII ] 51.81 µ m 2258 (4010) 879. (1761.) 505. (1512.) 212. (676.) 29.2 (142.) · · · (4.95) 3883. (8106.)[ NIII ] 57.32 µ m 850 (1811) 94.0 (392.) 28.8 (136.) · · · (11.7) · · · ( · · · ) · · · ( · · · ) 973. (2351.)[ OI ] 63.18 µ m 3353 (4762) 3149. (5211.) 1512. (2743.) 676. (1326.) 207. (639.) 19.8 (61.2) † ‡ [ OIII ] 88.35 µ m 3353 (4762) 1263. (2383.) 505. (1512.) 322. (955.) † † † ‡ * Notes: † : outside the SAFARI spectral range; ‡ : excluding detectionsoutside the SAFARI spectral range. TABLE 6Number of galaxies detectable in a CCAT survey of 0.5 in IR lines as a function of redshift at 5 σ (3 σ ) in1 hr. integration per FoV, following Franceschini et al. (2010) line/redshift 0 < z < < z < < z < < z < < z < < z < OI ] 63.18 µ m · · · ( · · · ) · · · ( · · · ) · · · ( · · · ) 0.27 (3.60) ⋄ · · · (0.27) † · · · (0.22) ‡ [ OIII ] 88.35 µ m · · · ( · · · ) 22.2 (67.5) ⋄ † ‡ NII ] 121.90 µ m 13.5 (39.0) † † ‡ § [ OI ] 145.52 µ m 0.45 (1.35) † · · · (0.09) ‡ · · · (0.25) $ · · · ( · · · ) § · · · (0.72) ¶ [ CII ] 157.74 µ m 235. (643.) † § ¶ * Notes: ⋄ : at 200 µ m band; † : at 230 µ m band; ‡ : at 291 µ m band; µ m band; $: at 450 µ m band; § : at 620 µ m band; ¶ :at 740 µ m band IR spectroscopic cosmological surveys 17
TABLE 7Number of galaxies detectable in a CCAT survey of 0.5 in IR lines as a function of redshift at 5 σ (3 σ ) in1 hr. integration per FoV, following Gruppioni et al. (2011) line/redshift 0 < z < < z < < z < < z < < z < < z < OI ] 63.18 µ m · · · ( · · · ) · · · ( · · · ) · · · ( · · · ) 2.02 (13.7) ⋄ † ‡ [ OIII ] 88.35 µ m · · · ( · · · ) 22.1 (53.5) ⋄ † ‡ NII ] 121.90 µ m 12.7 (32.5) † † ‡ § § [ OI ] 145.52 µ m 0.81 (2.02) † · · · (0.45) ‡ § ¶ [ CII ] 157.74 µ m 172. (432.) † § ¶ * Notes: ⋄ : at 200 µ m band; † : at 230 µ m band; ‡ : at 291 µ m band; µ m band; $: at 450 µ m band; § : at 620 µ m band; ¶ :at 740 µ m band TABLE 8Number of galaxies detectable in a CCAT survey of 0.5 in IR lines as a function of redshift at 5 σ (3 σ ) in 1hr. integration per FoV, following Valiante et al. (2009) line/redshift 0 < z < < z < < z < < z < < z < < z < OI ] 63.18 µ m · · · ( · · · ) · · · ( · · · ) · · · ( · · · ) · · · (26.4) ⋄ · · · (1.94) † · · · (0.09) ‡ [ OIII ] 88.35 µ m · · · ( · · · ) 17.5 (54.9) ⋄ † ‡ NII ] 121.90 µ m 13.1 (40.5) † † ‡ § [ OI ] 145.52 µ m 0.94 (1.44) † · · · (0.09) ‡ · · · (3.73) $ · · · ( · · · ) § · · · (1.66) ¶ [ CII ] 157.74 µ m 231. (639.) † § ¶ * Notes: ⋄ : at 200 µ m band; † : at 230 µ m band; ‡ : at 291 µ m band; µ m band; $: at 450 µ m band; § : at 620 µ m band; ¶ :at740 µµ