Submillimeter Galaxy studies in the next decade: EAO Submillimetre Futures White Paper Series, 2019
Ran Wang, Wei-Hao Wang, David L.Clements, Haojing Yan, Yiping Ao
SS UBMILLIMETER G ALAXY STUDIES IN THE NEXT DECADE :EAO S
UBMILLIMETRE F UTURES W HITE P APER S ERIES , 2019
EAO S
UBMILLIMETRE F UTURES P APER S ERIES , 2019
Ran Wang ∗ Kavli Institute for Astronomy and Astrophysics, No. 5 Yiheyuan Road, Haidian districut, Beijing, 100871, China
Wei-Hao Wang
Academia Sinica Institute of Astronomy and Astrophysics (ASIAA), No. 1, Section 4, Roosevelt Rd., Taipei 10617, Taiwan
David L. Clements
Astrophysics Group, Blackett Laboratory, Imperial College London, London SW7 2AZ, UK
Haojing Yan
Department of Physics and Astronomy University of Missouri Columbia, MO 65211, USA
Yiping Ao
Purple Mountain Observatory, Chinese Academy of Sciences, Nanjing 210034, China A BSTRACT
Over the last two decades, the Submillimetre Common-User Bolometer Array (SCUBA) and SCUBA-2 on the James Clerk Maxwell Telescope (JCMT) achieved gread success in discovering the populationof dusty starburst galaxies in the early universe. The SCUBA-2 surveys at 450 µ m and 850 µ m setimportant constraints on the obscured star formation over cosmic time, and in combination of deepoptical and near-IR data, allows the study of protoclusters and structure formation. However, thecurrent submillimeter (submm) surveys by JCMT are still limited by area of sky coverage (confusionlimit mapping of only a few deg ), which prevent a systematic study of large samples of the obscuredgalaxy population. In this white paper, we review the studies of the submm galaxies with currentsubmillimeter/millimeter (submm/mm) observations, and discuss the important science with the newsubmm instruments in the next decade. In particular, with a 10 times faster mapping speed of thenew camera, we will expect deep 850 µ m surveys over 10 to 100 times larger sky area to i) largelyincrease the sample size of submm detections toward the highest redshift, ii) improve our knowledgeof galaxy and structure formation in the early universe. The thermal dust continuum emission at far-infrared wavelengths is an important trace of the dust and gas contents andstar forming activities in galaxies [12, 54, 22]. At high redshift, the UV and optical emission from the stellar componentis dimming dramatically. However, the hump of the thermal dust emission is shifted to the submm bands, and due to thenegative k -correction, the observing flux densities at submillieter and millimeter wavelengths do not drop with redshift.Thus, the submm/mm windows open a unique opportunity to probe active star formation and galaxy evolution towardthe earliest epoch. Over the last two decades, the submm facilities, such as the Submillimetre Common-User BolometerArray (SCUBA) and SCUBA-2 on the JCMT and the SPIRE on the Herschel
Space telescope etc., achieved greadsuccess in discovering the population of dusty starburst galaxies in the early universe [30, 22, 62].Dusty starbursts have been found into the epoch of reionization (EoR; z > . ). Some of them are found as quasarhosts and/or companions [4, 16], and yet some are found in ”blind” sub-mm/mm surveys [49, 58]. In terms of IR ∗ [email protected] a r X i v : . [ a s t r o - ph . GA ] J a n AO S
UBMM F UTURES - J
ANUARY
22, 2020luminosities, they are all ULIRGs ( L IR > L (cid:12) ) and HyLIRGs ( L IR > L (cid:12) ), which translate to dust-embeddedstar formation rates (SFRs) of > − M (cid:12) /yr . The very existence of such high-z U/HyLIRGs has importantimplications. The burst of star formation traces the most active stage of galaxy evolution. The submm sources detectedin the core of overdensity regions also probe the early evolution of galaxy clusters. In addition, The prevalence of dustat z ≈ -7 means that there must be very active star formations at even earlier epochs ( z ∼ and beyond).Bright submm/mm emission was also detected in the host galaxies of optically luminous quasars at z ∼ Hubble , Spitzer , Herschel and ground-based telescopes at multiple wavelengths and their redshifts can be determinedby photometric observations at multiple bands. Alternatively, a small fraction of sources have been confirmed withspectroscopic measurements and narrow-band imaging. Due to large extinction, observations at optical and/or near-infrared are difficult and only bright sources can be detected at high redshifts. For some sources with large amount ofdust, they are very likely to be invisible with current optical facilities in a limit observing time.The mapping capability of the single dish telescopes at submm wavelengths are also becoming important in tracingoverdensity and structure formation in the early universe. Clusters of galaxies are the most massive bound structures inthe local universe. They are largely dominated by ‘red and dead’ elliptical galaxies, and the oldest and most massiveelliptical galaxies lie at their cores (eg. [57]). Studies of the stellar populations of these massive galaxies reveal them tobe old, with inferred formation redshifts z > and with the bulk of their stellar populations forming over a short timescale. This would imply that the progenitors of the massive elliptical galaxies in the cores of clusters must have formedin major starbursts.Finding galaxy clusters in formation at these high redshifts is a difficult task. The standard methods of cluster detectioninclude X-ray and Sunyaev Zel’dovich (SZ) observations of the hot intracluster medium, and the search for red sequencegalaxies in the optical and near-IR. All of these methods fail for forming galaxy clusters; the first two since youngsystems are yet to Virialize and thus lack a significant intracluster medium that can be detected by X-rays or SZ, andthe last because the galaxies making up the cluster are still star forming and thus do not lie on the red sequence. Thehigh star formation rates (SFRs) of forming giant elliptical galaxies, however, in principle allow us to search for theseobjects in the far-IR, since, like starbursts in the local universe, they should be luminous at these wavelengths. Recentresults using Herschel and
Planck data in the far-IR and submm have begun to find candidate protoclusters in this way.
Submm surveys of galaxies were carried out with JCMT in the last two decade (Table 1), e.g., the SCUBA programwhich discovers the first submm galaxy sample at high redshift, the SCUBA-2 Cosmology Legacy Survey (S2CLS, [22]),The Hawaii SCUBA-2 Lensing Cluster Survey [29], the Submillimeter Perspective on the GOODS Fields (SUPERGOODS, [10]), the SCUBA-2 Large eXtragalactic Survey (S2LXS ), S2COSMOS/eS2COSMOS , and SCUBA-2Ultra Deep Imaging EAO Survey (STUDIES, [67]). The SCUBA-2 surveys image the submm sky to a 1 σ noise level of . ∼ mJy, discovering the extreme starburst systems with FIR luminosities on orders of to L (cid:12) . Theseobservations, together with the data from Spitzer and Herschel at shorter wavelengths, measure the submm sourcenumber densities and update the star formation rate density [30, 14, 24, 52, 22, 67], revealing that the submm galaxiescontribute significantly to the cosmic star formation history over a wide range of redshift [33, 52].To study the SFH in the early Universe, observations by space telescopes at far-infrared and ground facilities at submmbecome an important tool to detect highly dust-obscured galaxies. However, due to the poor resolution and confusion UBMM F UTURES - J
ANUARY
22, 2020level of the space telescopes, the observations will only select bright sources and therefore largely underestimate theSFR at high redshifts. Ground-based facilities like JCMT/SCUBA2 can provide better sensitivity and resolution incomparison with space facilities like
Herschel . However, due to a relatively low mapping speed the statistic results stillsuffer large uncertainties by the cosmic variance from small surveyed fields. Large area of deep surveys are highlyrequired in future.In addition, to fully understand the early cosmological star formation history, we will need large samples of high-z dustystarbursts selected in a systematic manner. The most promising method to construct such samples is through the use ofFIR/sub-mm colors in blind, un-biased surveys. The typical cold-dust emission (as being heated by star formation) ingalaxies has its peak at around rest-frame 100 µ m, and the color selection is to utilize this characteristics. The so-called“500 µ m riser” technique is an implementation tailored for the BLAST and the Herschel /SPIRE bands [48, 51]: a dustygalaxy at z > should have red colors in 250, 350 and 500 µ m bands because its SED is “rising” through these threebands as the peak is redshifted to ∼ µ m and redder wavelengths. An extension of this technique to higher redshiftsis the “850/870 µ m riser” method, where a redder band at 850 or 870 µ m is added to the selection [50].A growing number of protoclusters and protocluster candidates are being found using far-IR and submm techniques.Casey [7] summarises results on five specific protoclusters at < z < . with spectroscopic confirmation, as wellas a further three candidates at higher redshifts. Meanwhile, there is an increasing number of candidate protoclusterswith estimated redshifts around 2 to 3 emerging from work on the Herschel and
Planck surveys, including at least 27candidates emerging from investigating
Planck sources lying within the large area
Herschel surveys [23], [11], and 228resulting from colour selection from the
Planck [44] all sky survey. The nature of these sources is currently unclearsince nearly all the sources lack the spectroscopic followup necessary to confirm, or otherwise, their protocluster nature.One of the
Planck candidates has been found to be a superposition of two galaxy overdensities at different redshifts (1.7and 2.0, [21]) while photometric redshifts suggest that several others are genuine protoclusters ([11], [7]). Over the nextten years the spectroscopic study of these candidates should allow considerable progress in confirming their nature andstudying the detailed physics of the processes driving the star formation in their member galaxies.Theoretically, protoclusters at z ∼ Herschel and
Planck (see eg.[23]). At higher redshifts, theoretical predictions suggest a different picture, with the cores of eventual protoclustersshowing high rates of star formation on scales of a few 100kpc at z ∼ [10]. The observational situation at higherredshift is somewhat confused, however. Starbursting cluster cores may have been seen at z ∼ , rather lower thanthe predicted redshift, in followup observations of very red sources from Herschel [39] and the South Pole Telescope[37], but protocluster candidates at z ∼ selection through Ly α emission by HyperSuprime Cam [26] show structures,including far-IR luminous sources, extended on scales of 10 Mpc instead. It is thus clear that much remains to be learntabout the early stages of galaxy cluster formation.In summary, the submm/mm bands are unique in tracing obscured star formation and galaxy evolution over the cosmictime. The great success of the existing JCMT/SCUBA and SCUBA-2 survey proved that the 850 µ m band on MaunaKeais the most efficient window for deep imaging to detect the submm population at high redshift. However, the currentsubmm surveys by JCMT are limited by the small area of sky coverage (Table 1). e.g., The deep SCUBA-2 surveys(S2CLS, S2COSMOS) cover only ≤ deg of sky area. For comparison, much large sky area are aleardy covered withdeep optical, near-infrared, and radio observations (e.g., Stripe 82 of 300 deg ). The lack of deep submm data in theseregion prevent a systematic study of large samples of the obscured galaxy population. Thus, the new JCMT 850 µ mcamera with a 10 times faster mapping speed becomes an urgent request for developing large submm surveys. We would like to propse wide field surveys at 850 µ m using the new camera on the JCMT when the instrument isavailable in 2022. The 10 times faster mapping speed of the new 850 µ m camera will allow submm mapping overhundred deg of sky area at mJy sensitivity level in the next decade. The fields covered by the deep X-ray, optical,near-infrared, and radio observations will have the highest priority for such submm surveys. As we discussed above,the previous submm/mm surveys with sky coverages of 100 to 1000 deg , such as the HerS and HerMes Surveysat 250 µ m, 350 µ m, and 500 µ m, the SPT survey at 1.4mm, 2mm, and 3mm, are insufficient in both sensitivity andangular resolution, to fully recover the obscured star formation population toward the highest redshift and identify thecounterpart of galaxy samples and protocluster members discovered in deep optical and near-IR observations. The new3AO S UBMM F UTURES - J
ANUARY
22, 2020Figure 1: The spectral energy distribution (SED) of a dusty star forming galaxy at redshift from z=1 to 8. We use theSED of M82 from the SWIRE library as a template [47]. The SEDs are normalized at an 850 µ m observing flux densityof 4mJy, which are the typical 4 σ detection limit for JCMT/SCUBA-2 surveys. The vertical lines shows the observingwindows at 250 µ m, 350 µ m, 500 µ m, 850 µ m, 1.1 mm, 1.4 mm, 2 mm. The 850 µ m window is marked as a red line,which samples the peak of the dust continuum bump at the highest redshift.4AO S UBMM F UTURES - J
ANUARY
22, 2020 ■ Dannerbauer et al. 2014 found an excess of DSFGs associated with the Spiderweb cluster [z = 2.2] over a 3 Mpc radii. At least 5 associated with the cluster. ■ Clements, Greenslade et al. 2016. Overdensity of 14 SCUBA2 sources around a z = 3.26 lensed galaxy H12-00. ■ Oteo et al. 2017. Overdensty of 8 LABOCA sources which further resolved into 17+ Z spec = 4.001 DSFGs !
And yet….
Figure 2: From [39]: A candidate protocluster core at a redshift of 4.001 discovered through colour selection from
Herschel and followup submm imaging. On the left is a LABOCA image of a very red
Herschel source selection fromits 250 to 350 to 500 µ m colour. The Herschel source breaks up into several 870 µ m sources. In the centre is an ALMAimage of the central region of this clump of sources, showing that they too break up into numerous submm sources.These sources have spectroscopic redshifts measured by ALMA. On the right is the brightest of these sources which, inhigher reslution ALMA data, is found to be a pair of submm sources. This is consistent with the picture of [10] forprotocluster core formation, but it is found at z = 4 rather than the predicted z ∼ for this stage of cluster formation.For details see [39] and references therein.850 µ m continuum survey will largely increase the sample size of submm sources, and combined with data at otherwavelengths, will significantly improve our knowledge of galaxy and structure evolution in the early universe. High-z dusty starbursts are rare. Depending on the exact color criteria adopted, the surface density of 500 µ m risers inthe HerMES and the H-ATLAS areas is ∼ deg − or less. While there is not yet sufficient statistics of 850/870 µ mrisers, there is evidence that they are even rarer (e.g. Duivenvoorden et al. 2018). By providing an increase of ∼ × in the mapping speed than SCUBA-2, the planned JCMT new wide-field camera will be the most powerful tool in oursearch for dusty starbursts at z > and beyond. Furthermore, the increase in mapping speed will bring JCMT to thesame league as the new millimeter camera, TolTEC, on the LMT [5]. Combining wide-area submm/mm data fromJCMT and LMT will greatly boost our capability to study the high-redshift dusty population in terms of both samplesize and robust sample selection. Large populations of quasars and active galactic nuclei (AGN) are discovered from the optical and near-IR surveys.Previous submm/mm observations discovered a small fraction of the quasar host galaxies that are still efficiently formingstars. However, the connection between star formation activity and AGN properties are still unclear. In particular,the limit sky coverage of previous submm/mm observations cannot provide enough sample in different bins of AGNactivities for such studies at high redshift. The new 850 µ m survey will increase the sample size of submm observedquasars by factors of > a few tens and significantly improve the statistics.The detection of dust continuum at 850 µ m will intermediately constrain the dust mass and star formation rate in thequasar host galaxies. The sky coverage of large quasar samples will allow detections of the stacking signals with objectsin bins of redshifts, SMBH masses, and quasar luminosities. This will address the evolutionary connection between starformation and SMBH accretion over a wide range of redshift. Based on the assumption of dust-to-molecular gas ratio,the gas mass could also be constrained. The gas fraction in quasar hosts is an important parameter to address the effectof AGN feedback (e.g., [27, 55]). i.e., a low gas fraction compared to the normal star forming galaxies may indicatethat the molecular gas are removed and the star formation are suppressed due to the AGN power. In addition, More5AO S UBMM F UTURES - J
ANUARY
22, 2020 implies a relatively top-heavy stellar mass function indense regions.2. Between z ∼ z ∼ – (cid:58) M . This allows protocluster galaxies to growat a total SFR of about - (cid:58) M for a prolongedperiod of time, which contributes to about 65% of thetotal stellar mass seen in present-day clusters. Dependingon the magnitude of the initial overdensity, someprotoclusters may already contain signi fi cant group- orcluster-sized cores near the end of this epoch. These coreswould be the fi rst regions to show evidence of galaxyquenching or dense intracluster gas.3. After z ∼ M å contained in protoclustercores starts to increase as the cluster is being assembled.The fraction of SFR in the cores lags behind its massgrowth, implying that the growth of the cores is achievedmainly by incorporating externally formed stars frominfalling galaxies. The violent gravitational collapse ( Figures 2 and 3 ) proceeds in an inside-out manner asthe inner shells of a centrally peaked overdensityturnaround before the outer shells. In this epoch, galaxyquenching is enhanced through multiple channels,including gravitational heating, AGN feedback, grouppre-processing, and various types of satellite quenching processes like starvation, ram-pressure stripping, andtidal disruption.Due to the extreme hierarchical nature of cluster assembly,the far majority of the stars in present-day clusters formed inthe extended protocluster regions, mainly during the secondphase outlined above. Only 15% of the stellar mass formed “ in situ ” in cores ( this calculation takes into account mass lossduring stellar evolution ) , which makes the core halos under-representative during the main epoch of cluster ( galaxy ) growthat cosmic noon.
6. Discussion
Based on two recent SAMs, we have demonstrated in thisLetter that the fraction of the cosmic SFR density associatedwith the formation of present-day clusters is as high as 20% at z = z =
10. Protocluster galaxies are thus anearly dominant population at Cosmic Dawn, and remainsigni fi cant at Cosmic Noon.We outlined three stages that describe the early history ofcluster formation, which began with an inside-out growth phasefrom z
10 to z ∼
5, followed by an extended star formationphase at z ∼ – z ∼ – ( ) Figure 5.
Middle panel: average total SFR per protocluster. Bottom panel: The fractions of the total SFR ( black curves ) and stellar mass ( red curves ) occurring in thecore halo. The dashed lines at high redshift in the bottom panel illustrate the dependence of the upturn of core dominance on the limiting galaxy stellar mass. The riseand fall in the total SFR of protoclusters and the reversed trend found for the cores motivates the three-stage scenario for cluster formation illustrated in the top panel. The Astrophysical Journal Letters, ( ) , 2017 August 1 Chiang et al. Figure 3: From [10]: predictions for the evolution of protoclusters showing the core formation phase at high redshiftand the growth phase at z = 2 − . For details see [10] and references therein.quasar-starburst systems will be discovered at high redshift. These objects will be the targets for further ALMA andJWST observations to image the dust, gas, and stellar components. These systems will be the key examples to probe theearly growths of the SMBH and their host galaxies. Large samples of SMGs have been detected with the JCMT/SCUBA-2 (e.g., [22]), and follow-up observationswith ALMA have mapped some SCUBA-2 sources. Machine-learning algorithm can efficiently identify the likelycounterparts at optical/NIR for the SCUBA-2 sources by using ALMA observations as a training sample [1]. A newbolometer camera at JCMT with a mapping speed of 10 times faster than the current SCUBA-2 can be used to finisha much wider field with multiple optical/NIR archival data in an efficient way. Together with the machine-learningmethod, this can well constrain the cosmic SFH and significantly reduce the cosmic variance for the measurements.However, it is difficult to efficiently search for high redshift sources even with ALMA. Currently, lacking of identifiedhigh redshift sources will largely underestimate their corresponding SFRs at z >
5. ALMA observations show someSCUBA-2 sources without any optical/NIR counterparts. The latter could be good high-redshift candidates. Thedeep continuum observations with single dishes can reveal a large sample of SMGs and the follow-up high resolutionobservations with the interferometers like ALMA can locate their accurate positions. Together with archival multi-wavelength data and possible follow-up deep observations at optical/NIR with large optical/NIR telescopes or on-going6AO S
UBMM F UTURES - J
ANUARY
22, 2020Table 1: Submillimeter and millimeter surveysSurvey Sky Area Wavelength 1 σ rms Referencedeg µ m mJy beam − SCUBA/JCMT SHADES 0.2 850 2 [12]SCUBA/JCMT HDF 0.0025 850 0.45 [30]SUPER GOODS 0.125 850/450 0.28/2.6 [13]Hawaii SCUBA-2 Lensing Cluster Survey 0.137 450 4.4 c [29]S2CLS 5 850 1.2 [22]S2COSMOS/eS2COSMOS 2 850 0.9 [56]STUDIES 0.08 450 0.55 [67]S2LXS 10 850 2 aHerMes 380 250/350/500 5.2 ∼ ∼ ∼ d ∼ ∼ d a Geach et al. M17BL001; b http://toltec.astro.umass.edu/science_legacy_surveys.php c Before lensing amplificationcorrection. d See also Beelen et al. https://lpsc-indico.in2p3.fr/Indico/event/1765/session/9/contribution/52facility like TMT, one can constrain the cosmic SFH at high redshift. The machine-learning method is also helpful toidentify the likely high-redshift candidates without any optical/NIR counterparts. It may provide informative clues orconstraints on the cosmic SFHs at high redshift when ultra-deep optical/NIR observations are currently not availableyet.To compare with the cosmology simulation, we always require a large survey area to reduce the cosmic variance andto find some extreme sources or extreme environments, which are important astrophysical laboratories. The numberdensity of massive galaxies at high redshifts can be use to test different galaxy evolution models, which predict massivegalaxies decline very rapidly at z> µ m camera at JCMT, one canefficiently select high redshift candidates based on the color criterion. Ultra-deep radio surveys at LOFAR and SKAprecursors can be important for getting cross identifications and accurate positions due to their superior sensitivities andlarge field of views in comparison to the submm surveys. The far-IR/submm based protocluster surveys discussed above were not designed for protocluster work. They havefound protocluster candidates with total SFRs > , M (cid:12) /yr, with individual member galaxies forming stars atrates of 100s to 1,000s of M (cid:12) /yr, but it is likely that they are seeing only the peak of the luminosity function both ofprotoclusters and of galaxies within protoclusters because of the relatively limited sensitivity of these surveys to sourcesat such high redshifts. Meanwhile, dependence on selection at short submm wavelengths, typically 350, 500 or 550 µ mfor the Herschel and
Planck surveys, hampers studies of the highest redshift protoclusters at z ≥ . At the same time,7AO S UBMM F UTURES - J
ANUARY
22, 2020the existing samples have been selected using a range of rather heterogeneous methods, making statistical assessmentsof luminosity functions and evolution rather difficult.The next decade will see a range of developments that will move these studies forward. Firstly, spectroscopic followupof the existing samples will significantly increase the number of confirmed protoclusters known at z ≥ . This willbe achieved using both mm/submm spectroscopy using ALMA and NOEMA, and optical/near-IR observations withinstruments like KMOS and MUSE. Secondly, theoretical models for these objects, which require detailed n-body-hydrocodes with high spatial resolution, will improve our insight into this population. Thirdly, large area surveys at mmwavelengths using instruments such as NIKA2 and TOLTEC will provide new candidate protoclusters for detailedexamination. It is here where JCMT will be able to contribute, since the addition of large area surveys at higherfrequency submm wavelengths will greatly enhance our ability to select candidate protoclusters from z ∼ to thehighest redshifts. The protocluster candidates we currently know have an area density of about 1 per 40 sq. degrees,so the necessary surveys will have to be large area and ideally reaching a sensitivity of ∼ µ m survey to these sensitivities possible as part of a large area Legacy Survey,which will also be useful for many other studies. At the same time a larger field submm instrument will be needed tosurvey the ∼
15 Mpc region around known protoclusters to search for starbursting galaxies in their infall region that,if the predictions of Muldrew et al. [38] are current, will subsequently fall into the clusters. Other instrumentationdevelopments, such as KIDS-based submm imaging spectrometers, able to measure redshifts for all submm sourcesin a field simultaneously, would be ideal to followup the protocluster candidates detected in these surveys, and tocharacterise the molecular and atomic gas properties of their member galaxies, allowing this population to be confirmedand analysed much more rapidly than is possible with current instruments. The JCMT thus has a huge potential forstudying rare submm emitters, such as protoclusters, using the proposed new instrumentation. µ m capability Surveys at 850 µ m will play a leading role in discovering the dusty star forming sources. However, detections at asingle submm band is insufficient to constrain the nature of the detections. Measurements of dust temperature and IRluminosity, as well as the photometric-redshift require observations at more submm/mm bands. For objects at z ≥
3, the450 µ m window samples the peak of the thermal dust emission (Figure 1), thus is critical in determine the dust SED.The Herschel data at 350 µ m and 500 µ m are confusion limited and are only available for a small fraction of the sky(HerMes, H_ATLAS, and HerS fields). The 450 µ m window at Maunakea and the 450 µ m capability of JCMT willremain unique. JCMT has an angular resolution at 450 µ m that is similar to that of the LMT in the millimeter. Therefore,JCMT can detect 450 µ m sources that are much fainter than the Herschel /SPIRE limits and the 850 µ m limit of JCMT.The imaging capability of JCMT at 450 µ m will become a very important complement for future multi-bands surveys. Itwill also provide a unique band for the color selections of the dusty objects toward the highest redshift, and for probingthe normal galaxy population at the peak epoch ( z ∼ ) of the cosmic star formation and AGN. With the new 850 µ m wide field camera, we will be able to carry out surveys that cover very wide areas and still achievevery deep sensitivities. For example, we could conduct a survey over 300 deg to rms = 2 . mJy in 2500 hours underthe Band 2/3 weather condition with matched-filter applied (assuming that τ GHz = 0 . , and that the new camera is10 times faster than that of the current SCUBA-2 in PONG3600 mode with 660 pointings). We could choose the fieldsthat have already been covered by the previous Herschel surveys and/or will be covered by the coming LMT/TolTECsurveys. The much larger survey area compared to those of the existing SCUBA-2 surveys will allow us to select avery large sample (on the order of × ) of 850 µ m risers that are candidates of very high redshift dusty starburstgalaxies. The combination of 850 µ m data with the Herschel, NIKA2, and TOLTEC data will also allow the selectionof protocluster candidates. Moreover, the large sky area will cover a significant number of optically-selected quasars.According to the Sloan Digital Sky Survey fourteen data release of quasar catalog, about 5000 quasars at 2 < z < (i.e., the peak of SMBH and galaxy evolution) are expected within 300 deg [43]. While only a small fraction of themwould be directly detected, the shear number of the undetected ones will provide enough stacking signal to obtainthe average dust continuum emission and dust mass of the hosts in different SMBH mass and quasar luminosity bins.Meanwhile, deeper surveys could be carried out over tens deg down to the 0.7 ∼ rms = 0 . mJy over 10 deg will need 720 hours in Band 2/3 weather ( τ GHz = 0 . ). This ismuch deeper than the existing Herschel surveys, allowing us to detect dusty star forming galaxies or quasar hosts withFIR luminosities around × L (cid:12) at high redshifts. In addition, an area on the order of 10 deg will greatly reducethe cosmic variance, better prob the source counts of submm sources down to 2 ∼ S/N ≈ ), recovermore obscured star forming population at high redshift, and improve our knowledge of the SFH over cosmic time.8AO S UBMM F UTURES - J
ANUARY
22, 2020
The new 850 µ m wide field camera will allow submm surveys of high-z dusty star forming systems in tens to hundredsdeg of sky area with a much improved sensitivity compared to the current SCUBA2 surveys. The observations willserve as the sub-mm counterparts of the large-scale optical, infrared, and radio surveys with future telescopes (such asthe LSST, EUCLID, and SKA et c.). This will reveal the obscured star forming over cosmic time that is not detectablein optical/IR. For galaxies at z ≥
5, the measurement at 850 µ m samples the dust SED close to the peak (see Figure1). Thus, the combination of 850 µ m data with surveys using IRAM/NIKA2 or LMT/TolTEC at longer wavelengthsand constraints/upper limits from Herschel/SPIRE at shorter wavelengths will allow the selection of large sample ofcandidates at very high-z. With the 10 times faster mapping speed using the new 850 µ m camera, this new 850 µ mcamera will enlarge the current sample of z > References [1] An F. X. et al. 2018, ApJ, 862, 101[2] Aretxaga I. et al., 2011, MNRAS, 415, 3831[3] Austermann J. E. et al., 2010, MNRAS, 401, 160[4] Bertoldi F., Carilli, C. L., Cox, P., et al. 2003, A&A, 406, L55[5] Bryan S. 2018, Atacama Large-Aperture Submm/mm Telescope (AtLAST), p.36[6] Casey C. M., et al., 2015, ApJ, 808, L33[7] Casey C. M., 2016, ApJ, 824, 36[8] Carilli C. L., Neri, R., Wang, R., et al. 2007, ApJ, 666, L9[9] Chapin E. L. et al., 2011, MNRAS, 411, 505[10] Chiang Y.-K., Overzier R. A., Gebhardt K., Henriques B., 2017, ApJ, 844, L23[11] Clements D. L., et al., 2014, MNRAS, 439, 1193[12] Coppin K. et al., 2006, MNRAS, 372, 1621[13] Cowie, L. L., et al. 2017, ApJ, 837, 139[14] Daddi E. et al. 2005, ApJ, 631, L13[15] Dannerbauer H., et al., 2014, A&A, 570, A55[16] Decarli R., Walter, F., Bañados, E. et al. 2017, Nature, 545, 457[17] Decarli R., Walter, F., Venemans, B. P. et al. 2018, ApJ, 854, 97[18] Désert F.-X. et al., Proceedings of the annual meeting of the French Society of Astronomy & Astrophysics Lyon,June 14-17, 2016, C. Reylé et al. (eds.) SF2A, pp.439 - 442, 2016[19] Dudzeviˇci¯ut˙e, U. et al. 2019, arXiv:1910.07524[20] Eales S. et al., 2010, PASP, 122, 499[21] Flores-Cacho I., et al., 2016, A&A, 585, A54[22] Geach J. E. et al. 2017, MNRAS, 465, 1789[23] Greenslade J., et al., 2018, MNRAS, 476, 3336[24] Gruppioni C. et al. 2013, MNRAS, 432, 23[25] Gullberg B. et al. 2018, ApJ, 859, 12[26] Harikane Y., et al., 2019, ApJ, in press, arXiv:1902.09555[27] Ho L. C., Darling, J.& Greene, J. E. 2008, ApJ, 681, 128[28] Hodge J. A. et al. 2016, ApJ, 833, 103[29] Hsu L.-Y., et al. 2016, ApJ, 829, 25 9AO S
UBMM F UTURES - J