Detection and Fundamental Applications of Individual First Galaxies
a r
X
i
v
:a
s t
r o
-
p
h
/
6
1
1
0v
1
3
1
D
e
c
2
5
Detection and Fundamental Applications of Individual First Galaxies Renyue Cen 1ABSTRACT First galaxies formed within halos of mass M =107.5?109M ⊙at z =30?40in the standard cold dark matter (CDM)universe may each display an extended hydrogen 21-cm absorption halo against the cosmic microwave background with a brightness temperature decrement of δT =?(100?150)mK at a radius 0.3≤r ≤3.0comoving Mpc,corresponding to an angular size of 10?100arcseconds.A 21-cm tomographic survey in the redshift shell z =30?40(at 35?45MHz),which could be carried out by the next generation of radio telescopes,is expected to be able to detect millions of ?rst galaxies and may prove exceedingly pro?table in enabling (at least)four fundamental applications for cosmology and galaxy formation.First,it may yield direct information on star formation physics in ?rst galaxies.Second,it could provide a unique and sensitive probe of small-scale power in the standard cosmological model hence physics of dark matter and in?ation.Third,it would allow for an independent,perhaps “cleaner”characterization of interesting features on large scales in the power spectrum such as the baryonic oscillations.Finally,possibly the most secure,each 21-cm absorption halo is expected to be highly spherical and faithfully follow the Hubble ?ow.By applying the Alcock-Paczy′n ski test to a signi?cant sample of ?rst galaxies,one may be able to determine the dark energy equation of state with an accuracy likely only limited by the accuracy
with which the matter density can be determined independently.
Subject headings:galaxies -radio -intergalactic medium -cosmology:theory
1.Introduction
It is of wide interest to detect and understand the ?rst generation of galaxies,expected to form in the redshift range z =30?50in the standard CDM universe (Spergel et al.2003).
Extensive literatures on 21-cm properties of neutral hydrogen in the dark ages and during cosmological reionization have long focused on large-scale ?uctuations of the intergalactic neutral hydrogen and global spectral features (e.g.,Hogan &Rees 1979;Scott &Rees 1990).In this Letter we point out a unique feature possessed by the ?rst individual galaxies of mass 107.5?109M ⊙formed at z =30?40—a large hydrogen 21-cm absorption halo against the
cosmic microwave background (CMB).Each 21-cm absorption halo has a size 10′′?100′′with a brightness temperature decrement of δT =?(100?150)mK at 35?45MHz,which
could serve as a visible proxy for each galaxy that otherwise may be undetectable.The next generation of radio telescopes,such as LOFAR,may be able to detect such a signal.A range of fundamental applications is potentially possible with a redshift (i.e.,21-cm tomographic)survey of the ?rst galaxies in the redshift shell z =30?40,which may hold the promise to revolutionize the ?eld of cosmology and shed illuminating light on dark matter,dark energy and in?ation physics.Throughout,a standard (Wilkinson Microwave Anisotropy Probe)WMAP-normalized CDM model is used (unless indicated otherwise):?M =0.31,Λ=0.69,?b =0.048,H 0=69km/s/Mpc,n s =0.99and σ8=0.90.
https://www.wendangku.net/doc/c217119691.html,rge 21-cm Absorption Halos of First Galaxies
A ?rst-generation galaxy is expected to emit UV and X-ray radiation,
each carving out an H II region of size (ignoring recombination):r HII ～43(M h 0.1)1/3(f esc 8×10
4)1/3kpc comoving,where M h is the halo mass,c ?the star formation e?ciency,f esc the ionizing photon escape fraction into the intergalactic medium (IGM),and N p the number of hydrogen ionizing photons produced by each baryon formed into stars,(～104.5?5for a massive metal-free Population III IMF;Bromm,Kudritzki,&Loeb 2001).Hard X-ray photons (≥1keV)produced escape deep into the IGM with a distance of ～500?1000comoving megaparsecs,building an X-ray background.Sandwiched between small H II regions and the X-ray sea sits a quite large Ly αscattering region (Loeb &Rybicki 1999),resulting in a four-layer structure,as depicted in Figure 1.
The IGM in the vicinity of a galaxy interacts with ionizing UV and soft X-ray as well as near Ly αphotons emanating from the galaxy,which can be computed.The most important and relevant physical processes are (1)the interaction between neutral gas and near Ly αphotons emitted by the central host galaxy,which couples the spin temperature of the IGM to its kinetic temperature (Wouthuysen 1952;Field 1958),and (2)the interaction between neutral gas and ionizing UV and soft X-ray photons emanating from the host galaxy,which provides a heating source for the otherwise cold IGM up to some small radius.In the absence of heating the kinetic temperature T k of the IGM would be equal to
Fig.1.—shows a four-layer structure around an individual galaxy at very high redshift.The inner-most central region(inside the red circle)is the virialized region where star formation occurs with a typical size of about1kpc comoving.The next region,enclosed by the blue circle,is the ionized region with a typical size of order40?200kpc comoving.Exterior to the ionized region lies the Lyαscattering region shown in magenta for the inner(heated)part and in green for the outer(cold)part,where Lyαphotons can strongly couple the21-cm spin temperature to the kinetic temperature of the gas;the inner part shown in magenta is signi?cantly heated by UV and soft X-ray photons emanating from the host galaxy,while most of the region shown in green remains cold at a temperature set by the general IGM. Outside the green circle is the general IGM that is only a?ected collectively by cumulative X-ray background(as well as a Lyαbackground),as symbolically circumscribed by the dashed black circle.
T IGM≈18 1+z
= ∞0ηLν
dt
,where y c≡C10T k is the collisional coupling coe?cient with the 1+yα+y c
collisional de-excitation rate C10=4
In the expression for the Lyαcoupling coe?cient yα=P10T?
8π
λ2αAαφ(ν)being the cross section for Lyαscattering (MMR),whereλα=1.216×10?5cm is the wavelength of the Lyαline,Aα=6.25×108s?1 is the spontaneous Einstein coe?cient for Lyαline andφ(ν)is the normalized Lyαline (Voigt)pro?le with φ(ν)dν=1.Most of the Lyαscattering is accomplished by UV photons slightly on the blue side of the Lyαthat redshift into Lyαresonant line due to the Hubble expansion(note that?ν/ν～10?3due to Hubble expansion at r～1Mpc comoving),not the intrinsic Lyαline photons that escape from the host galaxy and redshift to the damping wing(Madau,Meiksin,&Rees1997;MMR).Additional physical processes that were not treated previously,including higher-order Lyman lines that result in cascade in two-photon emission,?ne structure of Lyαresonance and spin-?ip scattering,introduce corrections of order unity to Pα(CM;Hirata2005;Chuzhoy&Shapiro2005)but all these corrections terms are insigni?cant for our case,and we only apply the relatively large correction term S c(～1.5)as shown in Figure4of CM due to a spectral shape change near Lyα.The observed brightness temperature increment/decrement against the CMB is
δT=41(1+?)x H(T s?T cmb
0.02
)(
0.15
31
)1/2mK,(1)
where?is gas overdensity relative to the mean,x H neutral hydrogen fraction,
T cmb=2.73(1+z)K CMB temperature and other symbols have their usual meanings.
Figure2shows the pro?le ofδT for four cases of halo masses with each choice of IMF.Let us examine each of the four regions(sketched in Figure1)with respect to21-cm observations.Inside the virial radius(the red circle in Figure1)the gas is overdense with δ≥100and a positive large-amplitude emission signal may result,if a signi?cant amount of neutral hydrogen gas exists within.However,the size of this regions falls below0.1”and its signal is unlikely to be detectable in the foreseeable future.The H II region(inside the blue circle in Figure1)is ionized henceδT=0.In the region outside the Lyαscattering region(exterior to the green region in Figure1)the spin temperature of the IGM has been progressively attracted to the temperature of the CMB with gradually weakening coupling to the gas kinetic temperature by atomic collisions,producing a small but non-negligible 21-cm absorption signal at the redshift of interest(z～30?40)(Loeb&Zaldarriaga2004).
It is the Lyαscattering region that is of most interest here.In the inner part of the Lyαscattering region(shown in magenta in Figure1)the IGM is signi?cantly heated by UV and soft X-ray photons to exceed the CMB temperature,while its spin temperature
Fig. 2.—The top panel shows theδT pro?les for halos with M h=(106,107,108,109)M⊙, respectively,from top to bottom,at z=30as a function of comoving radius,with the adopted SAL IMF(SAL,see text).The top x-axis is in units of arcseconds.The division mass between large halos and minihalos at z=30is～1.4×107M⊙.For each chosen value of M h two snapshots are shown,at t=1.6×106yrs(solid curves)and t=6.5×106yrs(dotted curves),where t=6.5×106yrs is the lifetime of the least massive stars with M=25M⊙in the chosen SAL IMF.The bottom panel is similar to the top panel but for the adopted VMS IMF,and the two snapshots are at t=5.5×105yrs(solid curves)and t=2.2×106yrs (dotted curves),where t=2.2×106yrs is the lifetime of the stars with M=200M⊙.Also shown as two long-dashed curves in the bottom panel are two cases for M h=(108,109)M⊙, respectively,with soft X-ray(hν>100eV)intensity arti?cially boosted by a factor of10, which should be compared to the respective cases shown as blue curves located slightly to their left.The star formation e?ciency is assumed to be0.10for large halos and0.001for minihalos(Abel et al.2002).Escape fraction f esc=0.01is used,although the results depend very weakly on it.
is very strongly coupled to its kinetic temperature by Lyαscattering.As a result,the
inner radial region at0.01?0.04Mpc/h comoving for minihalos and0.04?0.4Mpc/h
comoving for large halos displays an emission signal against CMB with an amplitude of
δT～30mK(a shark dorsal?n-like feature in Figure2).Going outward(shown in green
in Figure1),the soft X-ray heating abates(because the cumulative optical depth to these
photons increases quickly)but the Lyαscattering remains strong,up to a distance of about
10Mpc/h comoving.Consequently,a strong21-cm absorption signal against the CMB
with an amplitude ofδT=?(100?150)mK at～35?45MHz(for z=30?40)on a scale
of0.3?3Mpc/h comoving,corresponding to an angular scale of10′′?100′′,is produced for large halos.This is the21-cm absorption halo—a unique and strong feature for the
large?rst galaxies.We note that the absorption signal cast by minihalos(the two top sets
of thin curves in each panel in Figure2)is relatively weak due to a combination of low
mass and low star formation e?ciency.We will therefore focus on large halos for practical
observability purposes.
Besides stars,no other soft X-ray source in the galaxy is assumed to concurrently exist.
We shall examine the validity of this assumption in detail.The density of the interstellar
medium plays an important role for some of the potentially relevant processes considered
here and it is assumed to be n(z)=n0(1+z)3,where local interstellar density n0=1cm?3.
This assumption should hold in hierarchical structure formation model for the following
reasons.The mean gas density scales as(1+z)3and halos at low and high redshift in
cosmological simulations show similarities when density and length are measured in their
respective comoving units(e.g.,Navarro,Frenk,&White1997;Del Popolo2001).The
spin parameters(i.e.,angular momentum distribution)of both high and low redshift halos
have very similar distributions peaking at a nearly identical valueλ～0.05(Peebles1969;
White1984;Barnes&Efstathiou1987;Ueda et al.1994;Steinmetz&Bartelmann1995;
Cole&Lacey1996;Bullock et al.2001).Thus,cooling gas in galaxies at low and high
redshift should collapse by a similar factor before the structure becomes dynamically stable
(e.g.,rotation support sets in),resulting in interstellar densities scaling as(1+z)3.Direct
simulations(Abel et al.2002;Bromm et al.2002)suggest a gas density of103?104cm?3 by the end of the initial free fall for minihalos at z～20,verifying this simple analysis.
We will estimate each of several possible types of soft X-ray emission sources in turn.
First,let us estimate soft X-ray emission from supernova remnants.Assuming the standard
cooling curve(Sutherland&Dopita1993)we?nd that at z=30a supernova blastwave
with an initial explosion energy of5×1052erg(for a star of mass200M⊙;e.g.,Heger
&Woosley2002)would enter its rapid cooling phase at a temperature of2.7×107K.
This implies that the energy emitted at～100?300eV from the cooling shell is about
7%.A200M⊙mass would release2.5×1054erg total energy due to nuclear burning,out
of which0.3%is released in photons at100?300eV for our adopted radiation spectrum.
Therefore,the ratio of total photon energy from the supernova remnant to that from the
star is0.4?0.5.Thus,for the VMS IMF,stellar soft X-ray appears to dominate over
that from its supernova remnant.The soft X-ray contribution from supernova remnant
cooling increases relatively compared to that from the star itself with decreasing stellar
mass and we estimate that the overall contribution from the two components may become
comparable for the SAL IMF case,averaged over time.Second,we will examine X-rays
produced from cooling of supernova-accelerated relativistic electrons by CMB photons via
inverse Compton(IC)process(e.g.,Oh2001).For adiabatic shocks,as is appropriate in
our case,the IC spectral energy distribution has a two-power-law form:Lν∝constant at E
redshift.Then the ratio of energy from IC to that from stars is found to be10?4?10?3 in the100?300eV band,depending on the exact upper energy cuto?(assuming10%of supernova explosion energy is utilized to accelerate relativistic electrons in shocks).Clearly, contribution to soft X-rays from IC process is unimportant.Third,X-ray binaries during the relatively short lifetime of massive stars may be rare,for top-heavy IMFs of concern here.We can make an estimate based on the calculation by Rappaport,Podsiadlowski
&Pfahl(2005),who give an ultra luminous X-ray binary formation rate of3×10?5per
supernova.It is clear that even if each X-ray binary is able to release as much energy as
in a supernova explosion itself and all in the soft X-ray band,the resulting contribution
will be less than a fraction of a percent of that from stars.Fourth,stellar mass black holes
(BH)of～10?100M⊙may be produced in signi?cant numbers with a top-heavy IMF
as well as a central galactic BH.It seems that stellar BH accretion is likely signi?cantly
suppressed and small due to feedback e?ect from stars on surrounding gas(e.g.,Mori,
Umemura,&Ferrara2004;Alvarez,Bromm,&Shapiro2005).A concomitant contribution
of soft X-rays from central BH accretion in the lifetime of a200M⊙star is approximately
(M BH/M?)×(1/0.007))×(t?/t E)×(f BH,SX/f?,SX)=0.6f BH,SX(M BH/M?/0.003),after
inserting soft X-ray emission fraction for a200M⊙stellar spectrum of f?,SX=0.003,
stellar lifetime t?=2.2×106yrs and Eddington time t E=4.4×108yrs,where f BH,SX
is the energy fraction released by the BH accretion in the soft X-ray band(100?300eV).
Thus,if the ratio of black hole mass to(bulge)stellar mass follows the local Magorrian
(Magorrian et al.1998)relation,then,unless most of the accretion energy is released in the
soft X-ray band,contribution from central BH accretion to the soft X-ray band is relatively
small.Fifth,thermal bremsstrahlung emission from gravitational shock heated gas is likely
negligible due to a low gas temperature(T～104K).Finally,soft X-rays from massive?rst
stars themselves are thought to be produced by stellar winds and quite uncertain.Recent
work suggests that the winds hence soft X-ray emission from metal-free stars are expected
to be insigni?cant(e.g.,Krticka&Kubat2005).In summary,soft X-rays from neglected, possible sources other than that from the stellar photospheres would,at most,make a modest correction to what is adopted in our calculation.To ascertain our conclusion,we compute a case with the amplitude of soft X-ray intensity at hν≥100eV arti?cially raised by a factor of10and do not?nd any signi?cant e?ect that would qualitatively change our results(Figure2).The reason is that the IGM quickly becomes optically thick to a few 100eV soft X-ray photons at～1Mpc comoving.Therefore,our results should be quite robust.
Since we are concerned with gas of relatively low temperature～20K,heating by a cumulative(hard)X-ray background may become relevant at some redshift.We estimate when this may happen in the CDM model.While an X-ray background may be generated by a variety of processes,black hole accretion at the centers of galaxies are thought to be the most dominant(e.g.,Ricotti&Ostriker2003;Kuhlen&Madau2005),estimated as follows.Let us suppose the energy extraction e?ciency from black accretion isαand a fraction f x of the released energy is in the form of hard X-rays.Then,the X-rays may collectively heat up the IGM temperature at most(ignoring Compton cooling and assuming all X-ray photons in the background are consumed by the IGM)by an increment
?T xray=1.1(f coll
0.1
)(
α
0.003
)(
f x
Fig. 3.—shows cumulative halo mass functions(solid curves)at redshift z=(20,30,40), respectively,from top to bottom.Each solid curve is broken into two parts with a thick portion corresponding to large halos with e?cient atomic cooling and a thin portion corresponding to minihalos with molecular cooling only.The corresponding dashed curves are cumulative c?f coll for the three cases,assuming star formation e?ciency of c?(large)=0.1 for large halos and of c?(mini)=0.001for minihalos(Abel et al.2002).The horizontal
dotted line indicates c?f coll=10?7(see equation2).
Another critical issue is whether heating of IGM by Lyαphotons is important.In a recent accurate calculation based on Fokker-Planck approximation,Chen&Miralda-Escud′e (2004;CM)show that the heating rate by Lyαphotons is much lower than previous estimates(MMR).We recast their important result(equation17of CM and using Figure3 of CM),the heating rate per hydrogen atom and per Hubble time,β,in the following way:
β≡Γc
P th
)K(3)
at z=30,where k B is the Boltzmann constant,H is the Hubble constant and P th=2.4×10?11(1+z
P th =1.For the regime of interest where we see the strong21-cm
absorption signal(Figure2)we?nd Pα
Fig. 4.—shows the density of galaxies versus the maximum absorption cross section(i.e., in the plane of the large circle centered on the galaxy perpendicular to the line of sight) withδT
Figure4shows the density of galaxies versus the maximum absorption cross section (i.e.,in the plane of the large circle centered on the galaxy perpendicular to the line of sight) withδT
First,a characteristic sharp fall-o?at5?10square arcseconds and a characteristic peak of the number of21-cm absorption halos is expected,as seen in Figure4due to a lower star formation e?ciency in(and low mass of)minihalos,as suggested by available simulations(Abel et al.2002).This should yield direct information on physics of cooling and star formation in?rst galaxies,which may be unobtainable otherwise by any other means in the foreseeable future.Note that the left cuto?and the peak density are functions of c?and g(IMF)(noting the di?erences between SAL and VMS cases in Figure4),where g(IMF)denotes dependence on the properties of IMF such as stellar lifetime and spectrum.
A full parameter space exploration will be given in a separate paper with more detailed treatments.We expect that c?and g(IMF)may be determined separately,when jointly analyzed with the density of absorption halos in the context of the standard CDM model.
Second,we see that the density of strong21-cm absorption halos depends strongly on n s,as testi?ed by the large di?erence(a factor of～50)between solid and dotted curves in Figure4.One may then obtain a constraint on n s,which is made possible because the e?ect due to di?erence in IMF may be isolated out,as discussed above,thanks to the features in the density of absorption halos(e.g.,peak location and sharp fall-o?at the low end).Let us estimate a possible accuracy of such measurements.At n～10?6h3Mpc?3 one would?nd0.1million galaxies in the redshift shell between z=28and z=32,giving a relative fraction(Poisson)error of0.3%.By comparing the solid and dotted curves in Figure4,we?nd that a constraint on n s with?n s=0.01(～3σ)may be achieved.This may have the potential to discriminate between in?ationary theories(e.g.,Liddle&Lyth 1992;Peiris et al.2003).In addition,the constraint placed on the temperature(or mass) of dark matter particles or running of the spectral index may be still tighter,because a signi?cant,?nite dark matter temperature or a running index tends to suppress small-scale power exponentially thus amplify the e?ects.The high-sensitivity constraint on small-scale power is a?orded by the physical fact that we are dealing with rare≥5?6σpeaks in the matter distribution.
Third,clustering of galaxies may be computed using such a survey containing potentially hundred of thousands to millions of galaxies in a comoving volume of size ～100Gpc3(for the redshift shell z=28?32).Both the survey volume and the number
of observable galaxies within are large enough to allow for accurate determinations of the correlations of?rst galaxies,particularly on large scales.It may then provide an independent,perhaps“cleaner”characterization of interesting features in the power spectrum such as the baryonic oscillations,with the advantage that they are not subject to subsequent complex physical processes,including cosmological reionization,gravitational shock heating of the IGM and complex interplay between galaxies and IGM,which in turn might introduce poorly understood biases in galaxy formation.A comparison between clustering of?rst galaxies and local galaxies(e.g.,Eisenstein et al.2005)will provide another,high-leverage means to gauge gravitational growth and other involved processes between z=30to z=0.
Finally,the scale(～1Mpc comoving)of the21-cm absorption halo signals is much greater than the nonlinear scale and virial radius(both～1kpc comoving).Thus,the IGM region in the21-cm absorption halo is expected to closely follow the Hubble?ow.Since near Lyαphotons(between Lyαand Lyβ)are not subject to absorption by hydrogen (and helium)atoms whose distribution might be complex,they escape into the IGM in
a spherical fashion.Additionally,since the dependence on?is linear(see equation2), density inhomogeneities are likely to average out(to zero-th order)and results do not depend sensitively on uncertain linear density?uctuations in the IGM.Although there might exist“pores”in the domain of21-cm absorption halo due to?uctuations in local IGM temperatures which may be caused by local shock heating due to formation of minihalos,the overall e?ect is likely negligible,because the mass fraction contained in all halos down to a mass as small as M h=105M⊙is about10?4at z=30.Furthermore,at n=10?6h3Mpc?3the mean separation between the galaxies is100Mpc/h,much larger than the size of Lyαscattering regions of size～1Mpc/h,so overlapping of the latter should be very rare(taking into account the known fact that they are strongly clustered in the standard cosmological model with gaussian random?uctuations;Mo&White1996). For these reasons,each21-cm absorption halo is expected to be highly spherical in real space.Therefore,21-cm absorption halos are ideal targets to apply the Alcock-Paczy′n ski (1979)test.Accurate measurements of angular size?θand radial depth?v for a sample of galaxies would yield a sample of d A(z)H(z),where d A(z)and H(z)are the angular diameter distance and Hubble constant,respectively,both of which are,in general,functions of?M, w(≡p/ρ)and k,with w describing the equation of state for dark energy and k being the curvature of the universe.As an example,let us assume that?M(≈0.3)has been
?xed exactly by independent observations and k=0and that w≈?1.Then one obtains |dln[d A(z)H(z)]/dw|=0.45at z=30(Huterer&Turner2001).Let us suppose a relative measurement error on each individual d A(z)H(z)is20%,then with ten thousand galaxies,
√
one could obtain a highly accurate constraint on w with?w=20%/0.45/
Likely,the accuracy of w determined by this method may eventually be limited by the accuracy with which?M(and k)can be determined by independent observations,due to the degenerate nature.We stress that this method is valid for each individual?rst galaxy and una?ected by uncertainties,for example,in the precise abundance of such galaxies.
In post-survey analyzes one faces the practical issue of extracting the wanted signals from the raw data,whose amplitude is expected to be dominated by foreground radio sources,including galactic synchrotron radiation,galactic and extragalactic free-free emission,and extragalactic point sources(e.g.,Di Matteo,Ciardi,&Miniati2004).While seemingly daunting,it has already been shown that signals of the amplitude proposed here may be recovered with relatively high?delity,when one takes into account the expected, potent di?erences in the spectral and angular properties between the21-cm signal and foreground contaminants(e.g.,Zaldarriaga,Furlanetto,Hernquist2004;Santos,Cooray, &Knox2005;Wang et al.2005).Since the21-cm absorption halos are expected to be rather regular and simple,one might be able to signi?cantly enhance the signal by using additional techniques,such as matched?lter algorithm,in combination with foreground “cleaning”methods.Finally,the amount of data in such a high spatial and frequency resolution3-dimensional survey will be many orders of magnitude larger than that of https://www.wendangku.net/doc/c217119691.html,putational challenges for analyzing it will be of paramount concern and most likely demand new and innovative approaches.
4.Conclusions
It is shown that a?rst galaxy hosted by a halo of mass M=107.5?109M⊙at
z=30?40possesses a large21-cm absorption halo against the CMB with a brightness temperature decrementδT=?(100?150)mK and an angular size of10′′?100′′.A21-cm tomographic survey of galaxies in the redshift shell at z=30?40may detect millions of galaxies and may yield critical information on cosmology and galaxy formation.A successful observation may need an angular resolution of≤1′′,a spectral resolution of≤4kHz,and a sensitivity of≤10mK at35?45MHz.LOFAR appears poised to be able to execute this unprecedented task,at least for the high end of the distribution.
At least four fundamental applications may be launched with such a survey,which could potentially revolutionize cosmological study and perhaps the?eld of astro-particle physics. First,it may provide unprecedented constraint on star formation physics in?rst galaxies, for there is a proprietary sharp feature related to the threshold halo mass for e?cient atomic cooling.Second,it may provide a unique and sensitive probe of the small-scale power in the cosmological model hence physics of dark matter and in?ation,by being able to,for
example,constrain n s to an accuracy of?n s=0.01at a high con?dence level.Constraints on the nature of dark matter particles,i.e.,mass or temperature,or running of index could be still tighter.Third,clustering of galaxies that may be computed with such a survey will provide an independent set of characterizations of potentially interesting features on large scales in the power spectrum including the baryonic oscillations,which may be compared to local measurements(Eisenstein et al.2005)to shed light on gravitational growth and other involved processes from z=30to z=0.Finally,the21-cm absorption halos are expected to be highly spherical and trace the Hubble?ow faithfully,and thus are ideal systems for an application of the Alcock-Paczy′n ski test.Exceedingly accurate determinations of key cosmological parameters,in particular,the equation of state of the dark energy,may be ?nally realized.As an example,it does not seem excessively di?cult to determine w to an accuracy of?w～0.01,if?M has been determined to a high accuracy by di?erent means. If achieved,it may have profound rami?cations pertaining dark energy and fundamental particle physics(e.g.,Upadhye,Ishak,&Steinhardt2005).
If a null detection of the proposed signal is found,as it might turn out,implications may be as profound.It might be indicative of some heating and/or reionizing sources
in the early universe(z=30?200)that precede or are largely unrelated to structure formation,possibly due to yet unknown properties of dark matter particles or dark energy. Alternatively,star formation and/or BH accretion in?rst galaxies may be markedly
di?erent from our current expectations.
I thank Dr.Daniel Schaerer for helpful information on Pop III stars.This research is supported in part by grants AST-0206299,AST-0407176and NAG5-13381.
REFERENCES
Abel,T.,Bryan,G.L.,&Norman,M.L.,2002,Science,295,93
Alcock,C.,&Paczy′n ski,B.1979,Nature,281,358
Alvarez,M,Bromm,V.,&Shapiro,P.R.2005,astro-ph/0507684
Barnes,J.,&Efstathiou,G.1987,ApJ,319,575
Bromm,V.,Coppi,P.S.,&Larson,R.B.2002,ApJ,564,23
Bromm,V.,Kudritzki,R.P.,&Loeb,A.2001,ApJ,552,464
Bullock,J.S.,Dekel,A.,Kolatt,T.S.,Kravtsov,A.V.,Klyping,A.A.,Porciani,C.,& Primack,J.R.2001,ApJ,555,240
Cen,R.2003,ApJ,591,12
Chen,X.,&Miralda-Escud′e,J.2004,ApJ,602,1(CM)
Chuzhoy,L.,&Shapiro,P.R.2005,astro-ph/0512206
Cole,S.,&Lacey,C.1996,MNRAS,281,716
Del Popolo,A.2001,MNRAS,325,1190
Di Matteo,T.,Ciardi,B.,&Miniati,F.2004,MNRAS,355,1053
Eisenstein,D.J.,et al.2005,ApJ,633,560
Elvis,M.,et al.1994,ApJS,95,1
Field,G.B.1958,Proc.IRE,46,240
Field,G.B.1959,ApJ,129,551
Gnedin,N.Y.2000,ApJ,535,L75
Heger,A.,&Woosley,S.E.2002,ApJ,567,532
Hirata,C.M.2005,astro-ph/0507102
Hogan,C.J.,&Rees,M.J.1979,MNRAS,188,791
Huterer,D.,&Turner,M.S.2001,Phys.Rev.D.64,123527
Jenkins,A.,et al.2001,MNRAS,321,372
Krticka,J.,&Kubat,J.2005,astro-ph/0509171
Kuhlen,M.,&Madau,P.2005,MNRAS,363,1069
Liddle,A.R.,&Lyth,D.H.1992,Phys.Lett.B,291,391
Loeb,A.,&Rybicki,G.B.1999,ApJ,524,527
Loeb,A.,&Zaldarriaga,M.2004,Phys.Rev.Lett.92,211301
Madau,P.,Meiksin,A.,&Rees,M.J.1997,ApJ,475,429,(MMR)
Magorrian,J.,et al.1998,AJ,115,2285
Mo,H.,&White,S.D.M.1996,MNRAS,282,347
Mori,M.,Umemura,M.,&Ferrara,A.2004,PASA,21,232
Navarro,J.F.,Frenk,C.S.,&White,S.D.M.1997,ApJ,490,493
Oh,S.P.2001,ApL,553,499
Oh,S.P.,Nollett,K.M.,Madau,P.,&Wasserburg,G.J.2001,ApL,562,L1 Peebles,P.J.E.1969,ApJ,155,393
Peebles,J.P.E.1993,“Physical Cosmology”,Princeton University Press. Peiris,H.V.,et al.2003,ApJS,148,213
Press,W.H.,&Schechter,P.L.1974,ApJ,215,703
Pritchard,J.R.,&Furlanetto,S.R.2005,astro-ph/0508381
Qian,Y.-Z.&Wasserburg,G.J.2002,ApJ,567,515
Rappaport,S.A.,Podsiadlowski,Ph.,&Pfahl,E.2005,MNRAS,356,401 Ricotti,M.,&Ostriker,J.P.2004,MNRAS,352,547
Rottgering,H.2003,astro-ph/0309537
Santos,M.G.,Cooray,A.,&Knox,L.2005,ApJ,625,575
Schaerer,D.2002,A&A,382,28
Scott,D.,&Rees,M.J.1990,MNRAS,247,510
Sheth,R.K.,&Tormen,G.1999,MNRAS,308,119
Shull,J.M.,&Van Steenberg,M.E.1985,ApJ,298,268
Spergel,D.N.,et al.2003,ApJS,148,175
Steinmetz,M.,&Bartelmann,M.1995,MNRAS,272,570
Sutherland,R.S.,&Dopita,M.A.1993,ApJS,88,253
Tan,J.C.,&McKee,C.F.2004,ApJ,603,383
Tumlinson,J.,Venkatesan,A.,&Shull,J.M.2004,ApJ,612,602 Ueda,H.,Shimasaku,K.,Suginohara,T.,&Suto,Y.1994,PASJ,46,319 Umeda,H.&Nomoto,K.2003,Nature,422,871
Upadhye,A.,Ishak,M.,&Steinhardt,P.J.2005,Phys.Rev.D,72,063501 Wang,X.,Tegmark,M.,Santos,M.,&Knox,L.2005,astro-ph/0501081 White,S.D.M.1984,ApJ,286,38
Wouthuysen,S.1952,AJ,57,31
Yu,Q.,&Tremaine,S.2002,MNRAS,335,965
Zaldarriaga,M.,Furlanetto,S.R.,&Hernquist,L.2004,ApJ,608,622 Zygelman,B.2005,ApJ,622,1356