Dominic J. Benford, S. Harvey Moseley, David T. Leisawitz, Eli Dwek,Michael J. Amato, Johannes G. Staguhn†, John C. MatherNASA / Goddard Space Flight Center † SSAI/GSFC

On the Detection of Dust at HighRedshifts with SIRCE and SAFIR

FUTURE MISSION CONCEPTS

SIRCE: Survey of InfraRed Cosmic Evolution

SIRCE is a mission concept developed atNASA/GSFC for a spaceborne observato-ry that can:• Map the sky at high spatial resolution(~10’’) at far-infrared wavelengths(~100-500µm)• Achieve sensitivity high enough todetect very distant objects• Cover enough sky to enable statisti-cally significant discoveries• Integrate until images are limited bythe confusion of the source distribution.

The SIRCE mission can be built withexisting technologies as a MIDEX-scalemission. SIRCE is envisioned as a 2m-class cryogenically-cooled telescopewith high sensitivity imaging arrays. Such a telescope can findtens of thousands of z>7 galaxies, directly measuring star forma-tion activity back to an era unreachable by existing telescopes.The core wavelengths of 100µm≤λ≤500µm are chosen to enablethe measurement of photometric redshifts and hence the character-ization and classification of the galaxies.

SAFIR: Single Aperture Far-Infrared Observatory

The SAFIR Observatoryis a mission in NASA’splan, recommended bythe NAS as a follow-onto JWST (previouslyNGST). At NASA/GSFC,we have developed aconceptual design forthe SAFIR observatorybased on JWST’s cur-rent design (Figure 8).A detailed analysis has been made of the changes and technolo-gies necessary to produce SAFIR. Crucial technologies requiringinnovation include lightweight deployable optics, cryogen-freecooling of optical elements and instruments to <4K (Figure 9), andlarge arrays of sensitive detectors for wavelengths of20µm<λ<800µm. As an example of the sensitivity of SAFIR(Figure 10), it will be able to detectthe rest frame 10µm+ thermal emis-sion from dust in ultraluminous galax-ies at z=10 and beyond.

With the recent detection by the WMAP mission of the epochof reionization at 11≤z≤30, we know when the first starsformed. Shortly thereafter, the dust enrichment of theUniverse began, when the relic radiation field was at a temper-ature of 35-85K. Present submillimeter observations have suc-ceeded in detecting the dust in galaxies out to redshifts ofz≈5, covering the majority of the star formation history of the

Universe. Future space-based missions can probe this timewith great sensitivity. A far-infrared survey mission, SIRCE, isbeing studied with the goal of characterizing the evolution ofdust emission out to z≈7. Pushing the boundary to z=10 oreven z=20 to trace the origin of dust enrichment is a challeng-ing prospect, but may be possible with the future SingleAperture Far-InfraRed (SAFIR) Observatory.

Finding these rare dust-enriched galaxies at high redshifts willrequire a sky coverage and sensitivity unavailable fromground-based observatories. The missions we describe hereuse the advantage of the very dark natural sky backgroundaccessible from space in the 100µm-500µm wavelengthregion. Extrapolating from expected results from SIRTF, thesemissions will open up the high-redshift universe in the far-IR.

DUST AND GALAXY EVOLUTION

The Cosmic Infrared Background (CIRB) is the integral of the lightfrom all sources at all distances (Figure 1). Much of this lightcomes from ultraluminous infrared galaxies (ULIRGs), but somefraction arises in AGN and from normal galaxies. The energyreleased by the formation of stars and in regions around AGNs isabsorbed and reemitted by dust. Half the total luminosity in theUniverse is emitted at infrared wavelengths, much of it at~100µm. The fraction of dust emission was higher in the past thanit is today, implyingthat dusty galaxiesproduce a greaterportion of the lumi-nosity at high red-shifts. A completepicture of star forma-tion and AGN activityin the Universe can beobtained only whenfar-infrared observa-tions reach the abilityto probe to high red-shifts comparable tothat at shorter wave-lengths. Finally, thedust emission from the earliest objects – those present at or justafter the epoch of reionization at z≈20 – will appear at wavelengthsof ~1mm.

Determination of the cosmic star formation rate history, thegrowth of cosmic structure, and the accompanying energy releaserequires direct observations of the sources that dominate the lumi-nosity of the early Universe, which were previously seen only as acomponent of the CIRB.

In order to image to a given sensitivity limit, telescopes need bothangular resolution and collecting area. At radio wavelengths,widely separat-ed, small tele-scopes areneeded toachieve both; inthe optical, asingle mirrorsuffices. Thedividing line isnear 100µm, asshown at right.

This confusionlimit is the keylimitation pre-venting thedetection of dustat high red-shifts.

DETECTABILITY OF REDSHIFTED DUST

Shown at right is the fluxdensity of thermal dustemission from a typicalultraluminous infraredgalaxy as a function of itsredshift. The spectralmodel of Dwek et al. hasbeen used. At achievableflux density limits (1mJyat 100µm, roughly 1000times more sensitive thanIRAS), thousands of dustysources at z>7 can be dis-covered – if they exist.

If we calculate the highest redshift detected galaxy, by comparingthe confusion limit to the expected flux, as a function of the wave-length and telescope diameter, we find that detecting ULIRGs atz~5 requires a ~2m diameter mirror (Figure 4).

Shown in Figure 5 are the fluxes of a set of template galaxies red-shifted until they become too faint to resolve. The paths are gen-erally well separated, enabling reliable photometric estimates ofredshift and hence luminosity.

In order to detect galaxies at redshifts approaching the epoch ofreionization – z>10 – the flux sensitivity must be better, of order100µJy at the longest submillimeter wavelengths. It can be shown(Figure 6) that this requires a 10m-class far-IR telescope, and thatit will require high sensi-tivity (and hence cooledoptics and large formatdetector arrays) to per-mit blind searches forthese unknown galaxies.If these galaxies aredetectable primarilythrough their far-infrared emission, it isunlikely that their posi-tions will be known priorto such a telescope.

1.0E-6

1.0E-5

1.0E-4

1.0E-3

1.0E-2

1.0E-1

1.0E+0

1.0E+1

1.0E+2

1.0E+3

1.0E-2 1.0E-1 1.0E+0 1.0E+1

Flu

x D

ensi

ty (

Jy)

Redshift

50µm

100µm

250µm

500µm

850µm

10.10.01

1

0.1

0.01

0.001

500µ

m F

lux

Den

sity

(Jy)

250µm Flux Density (Jy)

0.001

0.01

0.1

1

1 0.1 0.01 0.001

1

0.1

0.01

0.001

500µ

m F

lux

Den

sity

(Jy)

100µm Flux Density (Jy)

0.001

0.01

0.1

110 L

10 L

10 L

10 L�

�

�

�10

11

12

1310

0µm

Flu

x D

ensi

ty (J

y)

250µm Flux Density (Jy)

0.001

0.01

0.1

10.01 0.1 1

500µ

m Flu

x De

nsity

(Jy)500µm Flux Density (Jy)

100µm Flu

x Dens

ity (Jy

) 250µm Flux Density (Jy)

0.001

0.01

0.1

1

0.001

0.01

0.1

1

0.001

0.01

0.1

1

0.01

0.1

1

Figure 3. Flux density of a ULIRG as a functionof redshift illustrates the sensitivity of submil-limeter wavelengths to high redshift galaxies

Figure 4. Maximum detectable redshiftof galaxies, given the confusion limit,as a function of luminosity, observedwavelength and mirror diameter.

Figure 5. Paths of template galaxiesobserved in three far-IR bands as they red-shift until becoming fainter than the confu-sion limit.

Figure 7. Cutaway view ofSIRCE’s cryogenically cooled

telescope.

Figure 8. Concept for a SAFIR, a 10m-class,4K-cooled, far-IR (20µm–800µm) observatory.

Figure 9. Thermal control for SAFIR con-cept, providing for a 4K telescope.

10

10

10

10

10

10

5

6

7

8

9

10

1 10 100 1000

Sens

itivi

ty

Wavelength (µm)

Sensitivity in 10,000s Integration(in Jy·Hz 5σ sensitivity)

NGSTSAFIR

SIRCE

SPIRIT

SIRTF

SPECS

ALMA

Herschel

1µJy

1mJy

Figure 10. SAFIR and SIRCE sensi-tivity compared to other missions.

1E-10

1E-9

1E-8

1E-7

1E-6

1E-5

1E-4

1E-3

1E-2

1E-1

1E+0

0.1 1 10

Con

fusi

on L

imit

(Jy)

Mirror Diameter (m)

50µm

100µm

250µm

500µm

850µm

Figure 6. Confusion-limited flux of galaxies as afunction of wavelength and telescope diameter.

Figure 1. Extragalactic background light (Hauser &Dwek 2001). DIRBE & FIRAS measured the λ>1µmCIRB; two peaks are known, at ~1µm and ~100µm.

1nJy

10nJy

100nJy

1µJy

10µJy

100µJy

1mJy

10mJy

100mJy

1Jy

10 100 1000

Sens

itivi

ty L

imit

(Jy; 1

σ in

10,

000s

)

Wavelength (µm)

TelescopeDiameter:1m2m4m

8m

16m32m

Diameter Limited

CollectingAreaLimited

Photon LimitConfusion LimitTotal

Figure 2. Sensitivity of telescopes as a function of diameter;longer wavelengths need angular resolution whereas shorterwavelengths require collecting area.