Welcome to Scribd. Sign in or start your free trial to enjoy unlimited e-books, audiobooks & documents.Find out more
Download
Standard view
Full view
of .
Look up keyword
Like this
1Activity
0 of .
Results for:
No results containing your search query
P. 1
Joseph Silk and Martin J. Rees- Quasars and galaxy formation

Joseph Silk and Martin J. Rees- Quasars and galaxy formation

Ratings: (0)|Views: 23|Likes:
Published by Tuoma2

More info:

Published by: Tuoma2 on Jan 28, 2012
Copyright:Attribution Non-commercial

Availability:

Read on Scribd mobile: iPhone, iPad and Android.
download as PDF, TXT or read online from Scribd
See more
See less

01/28/2012

pdf

text

original

 
Astron. Astrophys. 331, L1–L4 (1998)
ASTRONOMYANDASTROPHYSICS
 Letter to the Editor 
Quasars and galaxy formation
Joseph Silk
1
and Martin J.Rees
2
1
Institute of Astronomy, Cambridge, UK, Institut d’Astrophysique de Paris, France,and Departments of Astronomy and Physics, University of California, Berkeley, CA 94720, USA
2
Institute of Astronomy, Cambridge, UKReceived 9 October 1997 / Accepted 12 December 1997
Abstract.
The formation of massive black holes may precedethe epoch that characterises the peak of galaxy formation, ascharacterizedbythestarformationhistoryinluminousgalaxies.Hence protogalactic star formation may be profoundly affectedby quasar-like nuclei and their associated extensive energeticoutflows. We derive a relation between the mass of the cen-tral supermassive black hole and that of the galaxy spheroidalcomponent, and comment on other implications for galaxy for-mation scenarios.
Key words:
galaxy formation: supermassive black holes quasars: outflows
1. Introduction
It is generally assumed that the first objects to form in the uni-verse were stars. However this is by no means assured. Thereis general agreement among theorists that, if cosmic structuresformhierarchically(‘bottomup’),asincolddarkmatter(CDM)models, the first baryonic clouds have masses in the range10
5
10
6
,
and that the characteristic mass subsequentlyrises. But it is actually not obvious that these clouds would un-dergo fragmentation into stars. Conditions in primordial cloudsdifferfromthoseprevailinginconventionalstar-formingcloudsthat the fate of nearby molecular clouds is not a reliable guide.For example, in the absence of magnetic flux, cloud collapsemay have been far more catastrophic than is the case at thepresentepoch.Amassivediskcouldhaverapidlysheditsangu-lar momentum via non-axisymmetric gravitational instabilities,andbecomesodenseandopaquethatitcontinuedtoevolveasasingleunit.Atveryhighredshifts,theinefficiencyofatomicandmolecular cooling via H and H
2
excitations is compensated by
Send offprint requests to
: J. Silk 
Comptoncooling;Comptondragprovidesanadditionalmecha-nismfortransferringangularmomentumandallowingcollapse.This outcome seems no less likely, a priori, than the alterna-tive evolutionary pathways found in the literature, according towhichprimordialcloudsfragmentintostarswithaninitialmassfunction that varies between being bottom-heavy, top-heavy oreven normal (that is, solar neighbourhood-like), depending onthe observations that are being interpreted. The quasar distribu-tion tells us directly that at least some massive black holes formearly. Indeed, the quasar comoving density peaks at
z >
2, andonly declines at
z >
3 (Shaver et al. 1996); on the other hand,the peak of galaxy formation occurs at
z
1
.
5 (Madau et al.1996; Connolly et al. 1997), although there is uncertainty aboutthe effects of extinction in leading to an underestimate of thegalaxy luminosity at high redshift.In fact, the known quasars, or their dead counterparts, arelikelytobewithinthecoresofatleast30percentofthesegalax-ies. To see this, note that combining the integrated density inquasar light with the assumption that quasars radiate at or nearthe Eddington limit yields an estimate of typical dead quasar(or black hole) mass (Soltan 1982; Chokshi & Turner 1992) as10
7
10
8
M
.
Whether one actually could observe an AGNcomponent in high redshift galaxies depends sensitively on theadopted lifetime of the active phase: higher redshift helps. Theobserved correlation between massive black holes and dynami-cally hot galaxies then suggests that most hot galaxies, amount-ing to of order a third of all galaxies in terms of stellar content,couldcontainsuchamassiveblackhole(Faberetal.1996).Onenote of caution would therefore be that dynamically hot galax-iesprobablyformsystematicallyearlierthanmostgalaxies,andthat if starbursts characterize their birth, existing high redshiftsamples of such objects may be incomplete.Nevertheless, while the case remains ambiguous, we aresufficiently motivated by the possible implications of a causalconnection between quasars and galaxy formation to explore inthisnotetheconsequencesofacosmogonicalscenarioinwhichthe first objects to form, at some highly uncertain efficiency, are
      L      E      T      T      E      R
 
L2 J. Silk & M.J. Rees: Quasars and galaxy formation
supermassive black holes. We discuss how, during subsequentmergers, the holes could grow, and exert a feedback on starformation.
2. Model
2.1. Early Growth
We suppose that the initial black holes form via a coherent col-lapse. This probably implies
bh>
10
6
βM 
,
, with
β 
1
.
Formationoflowermassholeswouldbelessefficient,foratleasttwo reasons. Primordial clouds of mass less than
10
9
M
arereadily disrupted by supernova-driven winds (Dekel and Silk 1986). Given the observed efficiency of black hole formation,the formation of black holes of mass below
10
6
is likelyto be inhibited. Moreover if the precursor object forms a super-massivestar,therewouldbesubstantialmassloss.SincetypicalfirstgenerationcloudsofprimordialCDMhavemassesoforder10
6
αM 
where
α
1
,
the total mass going into black holeswould be significant even if they formed in only a small pro-portion of clouds. Of course it is also possible that primordialclouds form smaller black holes which subsequently merge asthe hierarchy develops. However this process involves two ad-ditional stages of inefficiency (via formation and merging), andwe regard it as an improbable pathway. For a typical
L
galaxytoformfromhierarchicalmergingofprimordialcloudsandcon-tain a supermassive black hole, we require that the efficiency
at which the supermassive black hole formed (or equivalently,the inefficiency of fragmentation in primordial clouds) satisfies
>
10
5
β.
How small can
be in order for the consequences to beof interest? One observes today that many, if not all, galax-ies contain central supermassive black holes, and that
bh
=2
×
10
3
sph
,
where
bh
is the black hole mass and
sph
isthe mass of the spheroidal component (Magorrian et al. 1997;Fordetal.1997;vanderMarel1997).Oncesupermassiveblack holes are formed, the final black hole mass is enhanced dur-ing the hierarchical merging process, when dynamical frictionand dissipative drag on gas can drive supermassive black holesinto the center of the developing protogalaxy. Mergers providea continuing supply of gas, and gas dissipation and accretionfeed the central black hole. The most detailed numerical sim-ulations of protogalaxy collapse with cosmological initial con-ditions that have hitherto been performed demonstrate that an-gular momentum transfer is highly effective (cf. Navarro andSteinmetz 1997). Baryons collapse to form a dense, massivecentral clump at the resolution limit of the simulations, ratherthan acentrifugally-supported discon agalacticscale. This lineof thought at least supplies a motivation for exploring (and con-straining) the hypothesis that black hole formation and growth,rather than star formation, characterizes the earliest stages of galaxyformation.Wewillarguethatthevalueof 
self-regulatesso as to approximately satisfy the observed correlation.
2.2. Quasar Winds
Massiveblackholes,wheneverfuelledatasufficientrate,woulddisplay quasar-like activity. For brevity we term such objects‘quasars’ – noting, however, that the events we are discussingmay occur at higher redshifts than the typical observed quasars.An explosion model whereby outflows from early quasar-likeobjects led to cooled shells which fragmented into galaxies wasoriginally developed by Ikeuchi (1981;
cf.
also Ostriker andCowie 1981); this idea has, however, fallen from favour as theprincipalmodeforgalaxyformationbecausepost-shockComp-ton cooling would lead to excessive spectral distortions of thecosmic microwave background spectrum. We consider here the(more localised) effects on the gas within the protogalaxy inwhich the quasar is embedded.The effect of a protogalactic wind may be estimated asfollows. We model a protogalaxy as an isothermal sphere of cold dark matter that contains gas fraction
gas
with den-sity
ρ
=
σ
2
/
2
πGr
2
, constant velocity dispersion
σ
, and mass
(
< r
) = 2
2
/G.
In massive halos,
σ
2
m
p
/
3
k
=4
×
10
6
K
σ/
300kms
1
2
.
A sufficiently intense wind fromthe central quasar can sweep up the gas into a shell, and push itoutwards at constant velocity
v
s
=
w
L
Edd
8
π
2
G
gas
σ
2
1
/
3
.
In this expression, the mechanical (
i.e.
wind) luminosity istaken to be a fraction
w
of the Eddington luminosity
L
=4
πGcM 
bh
κ
1
= 1
.
3
×
10
46
8
ergss
1
,
Note that
w
˙
out
v
2
w
/L
E
=
1
(
v
w
/c
)
2
0
.
01
,
where
=
L
E
/
˙
inf 
c
2
is the radiation efficiency.Expulsion of this shell requires that its velocity should ex-ceed the escape velocity from the protogalaxy:
i.e.
v
s
> σ.
Thecondition for this to be the case is that
bh
> ασ
5
κG
2
c
= 8
×
10
8
γ 
(
σ/
500kms
1
)
5
M
,
(1)where
σ
500
σ/
500kms
1
and
γ 
1
32
π
3
w
/f 
gas
1
.
If for example
γ 
1, black holes could in principle eject allthe material from their host galaxies when their masses exceed
10
7
.
If the situation were indeed spherical, then, even if the central source switched off, the outflow would continue toexpand into the intergalactic medium for up to a Hubble timebefore stalling due to the ambient pressure. The shell velocitydecreases via an explosive outflow according to
v
s
= 330(
62
/
0
.
05
)
1
/
5
(1 +
z
)
3
/
10
kms
1
,
for an explosion energy, equal to the kinetic energy at break-out, of 10
62
62
ergs and an intergalactic medium density equalto 5 percent of the Einstein de Sitter density (with
0
=60kms
1
Mpc
1
), as compared to the binding energy of thegas in a massive protogalaxy of 
10
60
10
61
erg. The ambientpressure is not high enough to halt the shell, initially movingat breakout at a velocity of 
500kms
1
, until
z
<
1
.
The
      L      E      T      T      E      R
 
J. Silk & M.J. Rees: Quasars and galaxy formation L3
shell radius is
R
s
= 4
.
6(
62
/
0
.
05
)
1
/
5
(1 +
z
)
6
/
5
Mpc
,
and thefragment mass is (Ostriker and Cowie 1981)
= 7
×
10
10
a
410
1
/
562
4
/
50
.
1
(1 +
z
)
9
/
5
,
where 10
a
10
kms
1
is the sound velocity within the shell.How effective this expulsion would actually be, depends onthegeometryoftheoutflowandonthedegreeofinhomogeneityof the protogalactic gas. Realistically, the outflow may be di-rectional (probably bipolar) rather than spherically-symmetric.Some gas may survive and even be overpressured by doubleradio lobes to form stars (Begelman and Cioffi 1989)It is even possible (Natarajan et al 1997) that swept-up ma-terial,aftercoolingdownintoadenseshell,mayfragmentintoaclass of dwarf galaxies; these would differ from normal dwarfsin not being embedded in dark halos, and consequently be moreliable to disruption by massive star formation. Hence a high-mass black hole has two contrasting effects. It inhibits star for-mation in its host halo by blowing gas out; on the other hand,theejectedgasmayeventuallypileupinacoolshellthatbreaksup into small galaxies.
2.3. Protogalaxy Core
The implications of (1) are that a massive hole, if it continuesto emit at close to the Eddington rate, could expel gas com-pletely from its host galaxy. (The amount of gas that has to beaccreted is in itself negligible in this context: each unit mass of gas accreted into the hole can release enough energy to expel of order(
c/v
)
2
timesitsownmassfromthefarshallowerpotentialwell of the halo.) However, the fuelling demands a continuingaccumulation of gas in the centre (probably supplied by hier-archical merging). We can therefore interpret (1) as setting anupperlimittothemassofaholethatcanexistinagalaxywherestar formation can proceed efficiently.Expression (1) gives a relation between a black hole and itssurrounding halo. If a significant fraction of the Eddington lu-minosity emerged in ‘mechanical’ form, this gives the criterionthat the energy liberated on the dynamical timescale of the halois equal to the gravitational binding energy. The same expres-sion can be derived by a different argument. The maximum rateat which gas can be fed towards the centre of a galaxy (as theoutcome of mergers, etc) is
σ
3
/G,
where
σ
is the velocitydispersion in the merged protogalactic system. A quasar couldexpel all this gas from the galactic potential well on a dynam-ical timescale if its mass exceeded a critical value obtained byrequiring that
σ
5
/G
= 4
πGcM 
crbh
/κ.
Thisisindistinguishablefromtheobservedrelationbetweenthemassesofcentralholesandthoseoftheirhostspheroids(e.g.Magorrian et al. 1997). The observed relation has considerablescatterbuthintsatadependenceofblackholeonspheroidmassthat rises more rapidly than linearly: e.g. the Milky Way has acentral black hole mass of 2
.
5
×
10
6
,
whereas M87, with aspheroid mass that is only
100 larger than that of the MilkyWay, has a central black hole of mass 4
×
10
9
.Wethereforehypothesizethat,inthemergerprocessleadingtotheformationoftypicalgalaxies,theholemassstabilisesnearthe critical value. Hierarchical merging, augmented by continu-ingblackholegrowth,helpsmaintaintheblackhole-to-spheroidmass ratio at the self-regulation level. We suggested in Sect.1,however, that in the first bound systems gas may accumulateinto a single compact unit and evolve into a supermassive hole.If single black holes are indeed favoured over star formation inthis way, the holes in these first systems would be far above the‘critical’ value appropriate to such small halos with low veloc-ity dispersions: in other words
bh
crbh
.
Star formationwould then be inhibited (except in a disc close to the centralquasar) until, via mergers, the halos had become large enoughto bring the mass of the actual central hole below the criticalmass (which, as we have seen, grows faster than linearly withhalo mass). Thereafter, the scaling would be maintained.If. after mergers had formed high-mass halos, the centralhole were still above the critical mass, the implications wouldbe ominous for galaxy formation. Such a system would end upas a supermassive black hole embedded in a low surface bright-ness galaxy. Around the central host quasar, one would expecttofindacavityofhotgas.Comptoncoolingathighredshiftwilltend to quench any associated extended radio emission. If largeradio sources were responsible for the hot gas bubbles, then thegeometry is not a sensitive issue. Indeed, double lobes are aseffective as spherical outflows and are more reminiscent of thegeometryoftheassociatedhotgasbubbleswiththesimilarener-getics of reported Sunyaev-Zeldovich decrements that have noapparent associated galaxy cluster but with possibly associatedquasars (Jones et al. 1997; Partridge et al. 1997). Natarajan etal. 1997) argue that one might expect to detect a shell of newlyformed Magellanic irregular-type galaxies at the periphery of such bubbles.
3. Implications for galaxy formation and suppression
Near the hole, the ionizing radiation certainly suppresses starformation in normal molecular clouds, which, if of density
n
,survive only at a distance greater than
10(
L
46
/n
)
1
/
2
kpcfromthecentralquasar.Theultravioletfluxis effective at destroying
2
molecules produced via
for-mation to much larger distances, e.g. the
photodissociationrate is
10
10
L
46
r
21
s
1
at
r
1
Mpc from the quasar whereastherateofthecompeting
2
formationprocessis
10
9
n
s
1
(Tegmarketal.1997).Thiswouldtendtosuppressformationof dwarf galaxies out to
0
.
3
n
1
/
2
Mpc, If the radiation extendsto the x-ray band, there is however a narrow regime where thehard ionizing flux, by maintaining a fraction of free electronsdeep inside the cloud, may actually enhance star formation bystimulating
formation (Haiman, Rees and Loeb 1996).There are several further noteworthy consequences of thehypothesis that supermassive black holes form within the firstsubgalactic structures that virialise at high redshift, and are inplacebeforemostgalacticstarshaveformed.AGNactivitygen-eratesoutflowsthatcaninteractdynamicallywiththesurround-ing protogalactic gas as well as provide a possible early flux of hard ionizing photons. Star formation in the accretion disk sur-rounding the broad emission line region of the AGN is likely to
      L      E      T      T      E      R

You're Reading a Free Preview

Download
scribd
/*********** DO NOT ALTER ANYTHING BELOW THIS LINE ! ************/ var s_code=s.t();if(s_code)document.write(s_code)//-->