See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/263471785

Initial mass function of intermediate mass black hole seeds

Article  in  Monthly Notices of the Royal Astronomical Society · June 2014

DOI: 10.1093/mnras/stu1280 · Source: arXiv

CITATIONS READS

47 207

4 authors, including:

Andrea Ferrara Bin Yue

Scuola Normale Superiore di Pisa Scuola Normale Superiore di Pisa
430 PUBLICATIONS   14,104 CITATIONS    24 PUBLICATIONS   371 CITATIONS   

SEE PROFILE SEE PROFILE

Dominik R. G. Schleicher
Universidad de Concepcion
165 PUBLICATIONS   2,734 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Pop III star formation using GRADSPH-KROME View project

Deuterium fractionation in dense cores View project

All content following this page was uploaded by Andrea Ferrara on 28 June 2014.

The user has requested enhancement of the downloaded file.

Mon. Not. R. Astron. Soc. 000, 1–?? (2012) Printed 1 May 2014 (MN LATEX style file v2.2)

Initial mass function of intermediate mass black hole seeds

A.
1
Ferrara1,6 , S. Salvadori2 , B. Yue1 , D. Schleicher3
Scuola Normale Superiore, Piazza dei Cavalieri 7, I-56126 Pisa, Italy
2 Kapteyn Astronomical Institute, Landleven 12, 9747 AD Groningen, The Netherlands
3 Institut für Astrophysik, Georg-August-Universität Göttingen, Friedrich-Hund-Platz 1, D-37077 Göttingen, Germany
4 Kavli Institute for the Physics and Mathematics of the Universe (WPI), Todai Institutes for Advanced Study, the University of Tokyo

1 May 2014

ABSTRACT
We study the Initial Mass Function (IMF) and hosting halo properties of Intermediate
Mass Black Holes (IMBH, 104−6 M ) formed inside metal-free, UV illuminated atomic
cooling haloes (virial temperature Tvir > 104 K) either via the direct collapse of the
gas or via an intermediate Super Massive Star (SMS) stage. These IMBHs have been
recently advocated as the seeds of the supermassive black holes observed at z ≈ 6. We
achieve this goal in three steps: (a) we derive the gas accretion rate for a proto-SMS
to undergo General Relativity instability and produce a direct collapse black hole
(DCBH) or to enter the ZAMS and later collapse into a IMBH; (b) we use merger-tree
simulations to select atomic cooling halos in which either a DCBH or SMS can form
and grow, accounting for metal enrichment and major mergers that halt the growth of
the proto-SMS by gas fragmentation. We derive the properties of the hosting haloes
and the mass distribution of black holes at this stage, and dub it the “Birth Mass
Function”; (c) we follow the further growth of the DCBH by accreting the leftover
gas in the parent halo and compute the final IMBH mass. We consider two extreme
cases in which minihalos (Tvir < 104 K) can (fertile) or cannot (sterile) form stars
and pollute their gas leading to a different IMBH IMF. In the (fiducial) fertile case
the IMF is bimodal extending over a broad range of masses, M ≈ (0.5 − 20) × 105 M ,
and the DCBH accretion phase lasts from 10 to 100 Myr. If minihalos are sterile, the
IMF spans the narrower mass range M ≈ (1 − 2.8) × 106 M , and the DCBH accretion
phase is more extended (70 − 120 Myr). We conclude that a good seeding prescription
is to populate halos (a) of mass 7.5 < log(Mh /M ) < 8, (b) in the redshift range
8 < z < 17, (c) with IMBH in the mass range 4.75 < (log M• /M ) < 6.25.
Key words: cosmology — star formation — black hole physics — galaxies: high-
redshift

1 INTRODUCTION tion (Bromm et al. 2001a; Schneider et al. 2002; Omukai

et al. 2005; Dopcke et al. 2011) to the so-called PopII star
Along with the formation of first stars (Bromm & Yoshida formation mode we observe at lower redshifts and locally.
2011), the appearance of black holes (Volonteri & Bellovary
2012) is one of the most remarkable events occurring well
>
within the first cosmic billion year (redshift z ∼ 6). These
two types of astrophysical objects likely had a very strong
impact during cosmic evolution (Loeb et al. 2008; Petri et al. For these reasons, and others discussed in the paper,
2012; Park & Ricotti 2012; Jeon et al. 2012; Tanaka et al. the stellar and black hole populations do not evolve inde-
2012; Maiolino et al. 2012; Valiante et al. 2012) due to their pendently, but their formation and relative abundances are
radiative and mechanical energy/momentum injection in the interwoven and regulated by physical processes that main-
surrounding interstellar medium of the host galaxy and into tain an efficient cross-talk between them. Among these pro-
the intergalactic medium, thus drastically changing the sub- cesses, suppression of H2 molecules by UV (> 0.755 eV)
sequent galaxy formation history. Massive stars also disperse and Lyman-Werner (LW, 11.2 − 13.6 eV) photons is often
their metals (Madau et al. 2001; Johnson et al. 2013a; Sal- indicated as the most important one. Molecular hydrogen
vadori et al. 2014) in the gas, irreversibly polluting the sites is in fact a key species for gas cooling and fragmentation
of future star and black hole formation, causing a transi- in the primeval Universe. On the one hand, once the in-

c 2012 RAS
2 Ferrara et al.
tensity1 of the LW flux raises above a critical threshold, luted gas, either brought by minor mergers or smooth ac-
?
Jν,c (Machacek et al. 2001; Fialkov et al. 2012) gas can- cretion from the IGM.
not cool and form stars, and consequently stellar-mass black Metals would enhance the cooling rate driving thermal
holes, in minihalos that have virial temperature Tvir < 104 instabilities finally fragmenting the gas into clumps which
K. On the other hand, it is only when larger, metal-free, cannot be accreted as their angular momentum is hard to
atomic-cooling (Tvir ∼ >
104 K) halos are illuminated by a dissipate. The second stopping process could be a major
•
sufficiently strong LW flux Jν,c > Jν,c (Loeb & Rasio 1994; merger that generates vigorous turbulence, again disrupting
Eisenstein & Loeb 1995; Begelman et al. 2006; Lodato & the smooth accretion flow onto the central proto-SMS star.
Natarajan 2006; Shang et al. 2010; Johnson et al. 2012; Note that at the rates discussed above it only takes 105 yr to
Regan & Haehnelt 2009; Agarwal et al. 2012; Latif et al. build a 104 M SMS. If these events indeed occur, the star
2013a) that the minimal condition for the formation of a stops accreting and rapidly evolves toward very hot Zero
“direct collapse black hole” (DCBH) can be met2 . The pre- Age Main Sequence (ZAMS) SMS emitting copious amonts
? •
cise values of Jν,c and Jν,c depend on radiative transfer, of UV photons clearing the remaining halo gas out of the po-
chemistry and spectral shape of the sources and they are tential well. After a very brief lifetime (< 1 Myr) the SMS
only approximately known; however there is a broad agree- dies and leaves behind a comparable mass IMBH. If instead
•
ment that Jν,c ?
 Jν,c = 30 − 1000 in units of 10−21 erg the star can continue to grow, it will finally encounter a
−1 −2 −1 −1
s cm Hz sr . General Relativity (GR) instability that will induce a rapid,
Recent numerical simulations and stellar evolution cal- direct collapse into a DCBH, i.e. without passing through
culations have provided strong support in favor of the direct a genuine stellar phase. The two cases differ dramatically,
collapse model. In a cosmological framework, Latif et al. as virtually no ionizing photons are produced in the sec-
2013a have shown that strong accretion flows of ≈ 1 M yr−1 ond case. Therefore the newly formed DCBH will find it-
can occur in atomic cooling halos illuminated by strong ra- self embedded in the gas reservoir of the halo and start ac-
diation backgrounds. As stellar evolution calculations by crete again. This accretion phase, similar to the quasi-stellar
(Hosokawa et al. 2013) and (Schleicher et al. 2013) suggest phase advocated by Begelman et al. (2008), remains highly
only weak radiative feedback for the resulting protostars, obscured and it is only in the latest phases (several tens of
these calculations were followed for even longer time, sug- Myr after the DCBH formation) that the DCBH will be able
gesting the formation of ' 105 M black holes (Latif et al. to clear the remaning gas photo-ionizing and heating it. The
2013b). Their accretion can be potentially enhanced in the DCBH at that point has finally grown into a fully-fledged
presence of magnetic fields, which may suppress fragmenta- IMBH.
tion in the centers of these halos (Latif et al. 2014). We will investigate in detail all these steps by using a
Even if these minimal conditions (metal-free, atomic combination of analytical and numerical methods to finally
•
cooling halos illuminated by a Jν,c > Jν,c UV field) for derive the IMF of the IMBH. This quantity is crucial to
the formation of a DCBH are met, little is known on: (i) understand the formation of supermassive black holes and
the existence of an upper mass limit of DCBH host halos; the evolution of the black hole mass function with time. It
(b) the duration of the DCBH formation/growth phase; (c) also bears important implications for observations tuned to
the final DCBH mass function. A fourth important ques- search early black hole activity.
tion, not addressed here, concerns the final fate (e.g. inclu- Throughout the paper, we assume a flat Universe with
sion in a super-massive black-hole, ejection from the host) cosmological parameters given by the PLANCK13 (Planck
of this intermediate (M• = 104−6 M ) black hole popula- Collaboration et al. 2013) best-fit values: Ωm = 0.3175,
tion. These questions are at the core of a large number of ΩΛ = 1 − Ωm = 0.6825, Ωb h2 = 0.022068, and h = 0.6711.
cosmological and galaxy formation problems and therefore The parameters defining the linear dark matter power spec-
the quest for solid answers is very strong. Additional mo- trum are σ8 = 0.8344, ns = 0.9624.
tivations come from a possible interpretation of the near-
infrared cosmic background fluctuations and its recently de-
tected cross-correlation with the X-ray background (Cap-
pelluti et al. 2013), which might imply that an unknown
faint population of high-z black holes could exist (Yue et al. 2 PHYSICS OF DCBH FORMATION
2013a;Yue et al. 2013b). After the early pioneering studies (Iben 1963; Chan-
As we will show, answering the above questions requires drasekhar 1964; Loeb & Rasio 1994; Shapiro & Shibata
a detailed description of the mass accretion and merger his- 2002), the interest in the evolution of supermassive stars
tory of the atomic halos that satisfy the minimal conditions (SMS) has recently received renewed attention in the context
described above. The process starts with the growth of a of DCBH formation. In particular, (Hosokawa et al. 2012a;
proto-SMS star inside metal-free atomic cooling halos em- Latif et al. 2013a; Johnson et al. 2013b; Hosokawa et al.
bedded in a strong LW radiation field. The growth, fed by a 2013) research has focused on the previously unexplored
high accretion rate, typically ≈ 0.1 M yr−1 , can be blocked cases of very rapid mass accretion, Ṁ = 0.1 − 10 M yr−1 .
by at least two type of events: the first is accretion of pol- In the following we summarize the current understanding of
the aspects of SMS evolution that are relevant to the present
work.
1 Unless differently stated, we express the field intensity in the The rate at which the proto-SMS accretes gas from
usually adopted units J = J21 × 10−21 erg s−1 cm−2 Hz−1 sr−1 . the surroundings plays a key role in its evolution and in
2 DCBH formation has however to pass through an intermediate particular on the stellar radius-mass relation. This can be
stellar-like phase, as discussed later. appreciated by comparing two characteristic evolutionary

c 2012 RAS, MNRAS 000, 1–??
 The IMF of black holes seeds 3
timescales: the Kelvin-Helmholtz (KH), be initially characterized by tKH  tacc , such inequal-
ity will be reversed as soon as its contraction begins. As
GM?2
tKH ≡ , (1) a result, the accreting envelope phase can last consider-
R∗ L∗ ably longer than tacc . Schleicher et al. (2013) preliminarly
and the accretion, find that the accreting phase could continue until M? =
M? 3.6 × 108 (Ṁ / M yr−1 ) M . Beyond this point the system
tacc ≡ , (2) will evolve into a main-sequence SMS, eventually collapsing
Ṁ
into a black hole4 .
timescales. In the early phases of the evolution, even for very If this phase can be prolonged up to such high masses is
small values of Ṁ , the strong inequality tKH  tacc holds, a question that needs further investigation. There are hints
i.e. the time scale on which the star radiates its gravita- from ongoing calculations (Hosokawa et al. 2013; K. Omukai,
tional energy is much longer than the time on which mass is private communication) that when M? approaches 105 M
added to the star by accretion. Hence, the star grows almost for accretion rates ∼>
0.1 M yr−1 , the star enters a contrac-
adiabatically. As free-free absorption, providing the neces- tion phase (i.e. similarly to what happens for lower accretion
sary opacity to radiative losses, depends on temperature as rates at smaller masses and possibly due to a H− opacity
κ ∝ ρT −3.5 , while the stellar temperature increases with drop); however, at this stage numerical difficulties do not
mass, at some point tKH ≈ tacc . After this stage, the sub- allow to confirm this hypothesis. If this will turn out to be
sequent evolution of the star depends on the accretion rate. the case, the SMS final mass will be limited to M? ≈ 105 M
For sufficiently low values, Ṁ ∼<
10−2 M yr−1 the star en- as the surrounding gas will be prevented from accreting by
ters a contraction phase during which temperatures become radiative feedback connected with the increased Tef f . As
sufficiently high to ignite hydrogen burning and the star en- we will see later, the growth of proto-SMS is limited anyway
ters the ZAMS (for Ṁ ∼ <
10−3 M yr−1 this occurs when the by either general relativistic instabilities or cosmological ef-
star has reached M? ≈ 50M ) on the standard metal-free fects (as for example the accretion of polluted gas, see Sec.
effective temperature-mass relation (Bromm et al. 2001b) 5) to masses ∼ <
3 × 105 M , so the above difficulties do not
 0.025 represent a major source of uncertainty on the final results.
5 M?
Tef f = 1.1 × 10 K . (3) It is important to note that although D and H nuclear
100M
burning can be ignited relatively early (e.g. at 600-700 M
Note that the weak mass dependence implies that massive for Ṁ = 0.1 M yr−1 ) during the evolution, the associated
metal-free stars are very hot and therefore produce copi- energy production is however always subdominant compared
ous amounts of ultraviolet radiation that rapidly ionizes and to the luminosity released via Kelvin-Helmholtz contraction.
clears out the remaining surrounding envelope. Thus, radia- Therefore it does not sensibly affect the proto-SMS evolution
tive feedback effects rapidly quench the further growth of and final mass. The same conclusion is reached by Montero
stars once they reach the ZAMS. et al. (2012) who performed general relativistic simulations
The situation is drastically different for accretion rates of collapsing proto-SMS with and without rotation, includ-
exceeding Ṁ ' 10−2 M yr−1 , as pointed out by Hosokawa ing thermonuclear energy release by hydrogen and helium
et al. (2012a) and Schleicher et al. (2013): even when tKH burning.
has become much shorter than tacc , the stellar radius con- The radius-mass relation for an accreting supermassive
tinues to increase following very closely the mass-radius re- protostar can be obtained in a simple manner following the
lation: method outlined in Schleicher et al. (2013). As we have dis-
 1/2  1/2 cussed above, as long as Ṁ ∼ >
10−2 M yr−1 during the ac-
M? M?
R? = 2.6 × 103 R ≡ R0 , (4) cretion phase, the star radiates at a luminosity ≈ LE . By
100M 100M introducing the mass, radius and time non-dimensional vari-
while the effective temperature is thermostated3 to rela- ables
tively low values, Tef f ≈ 5000 K. The star then emits at
m = M/M , (7)
a rate close to the Eddington luminosity,
 
4πGmp c M r = R/R0 , (8)
LE = M = L0 (5)
σT M
where L0 = 1.5 × 1038 erg s−1 , as inferred from τ = t/t0 , (9)

L? = 4πR?2 σTef
4
f, (6) where t0 is defined starting from the KH time as
2
combined with eq. 4. GM m m m
tKH = ≡ t0 = 0.316 yr , (10)
Why does the radius continue to grow even when the R0 L0 r r r
KH time has become shorter than the accretion timescale?
a shell of accreted gas that forms at time5 τi = m/ṁ with
Schleicher et al. (2013) have analyzed this question in de-
initial mass m and radius ri = m1/2 (see eq. 4), starts to
tail. The key point is that tKH ∝ R?−1 : this implies that
even if at the time of formation a shell of mass M might
4 Obviously the star can grow only as long as there is sufficient
3 The thermostat is provided by the strong temperature sensi- gas to accrete in the host halo. This point will be considered in
tivity of H− opacity; such effect is missed if electron scattering is the next Section.
considered as the only source of opacity, e.g. Begelman (2010). 5 Note that time is simply related to mass according to t = M/Ṁ .

c 2012 RAS, MNRAS 000, 1–??
4 Ferrara et al.
dress this possibility. As today, though, there is no sign of
such transitions for stars that have masses up to 5 × 104 M .
Second, the proto-SMS might instead become general rela-
tivistically (GR) unstable, directly collapsing into a black
hole before reaching the ZAMS. Finally, an earlier stop to
the proto-SMS growth can be imposed when the two neces-
sary conditions (metal-free gas, strong UV background) for
the direct collapse of the gas into a black hole cease to be
valid. Whether and for how long these conditions hold can
only be ascertained from a cosmological analysis that we
defer to the next Section. In the following we analyze the
constraints posed by the most stringent local condition, i.e.
GR gravitational instability.

2.1 Non rotating proto-SMS

A non-rotating proto-SMS becomes gravitationally unsta-
ble (Chandrasekhar 1964, Montero et al. 2012) when the
gas equation of state (EOS) cannot be made stiff enough
to compensate for the de-stabilizing general relativistic ef-
fects. Mathematically, this happens when the adiabatic in-
dex drops below the critical value
4 Rs
Figure 1. Internal mass-radius structure of accreting super- Γc = + 1.12 , (14)
3 R?
massive protostars of different final mass M? = 105−8 M ; for
each mass three different values of the accretion rate Ṁ = where Rs = 2GM? /c2 is the Schwarzschild radius of the
10, 1, 0.1 M yr−1 are reported from the uppermost to the low- star. The de-stabilizing role of the relativistic term is evi-
ermost set of curves. Also shown are the regions corresponding to
dent. This condition is easily translated into a critical cen-
general relativistic instability for (grey) a non-rotating (T = 0)
tral density6 at which a spherical star becomes unstable to
and (green) for a maximally rotating (T = 0.009|W |) proto-SMS,
along with the zero-age main sequence (ZAMS) relation. No hy- radial perturbations:
drostatic equilibrium is possible beyond 108 M .  µ 3  M −7/2
?
ρc = 1.214 × 1018 g cm−3 , (15)
0.59 M
contract according to the local KH timescale. At a generic
which corresponds to a mass-radius relation condition for
time t the stellar radius is given by
GR instability expressed in the non-dimensional units intro-
dr r2 duced above
=− , (11)
dτ m
rc < 4.05 × 10−10 m3/2 . (16)
or
Z r Z τ
dr dt The unstable region is shown as a grey area in Fig. 1. As
=− , (12)
ri r2 τi m seen from there, depending on the accretion rate stars above
a certain mass can become GR unstable and collapse into
finally yielding the solution a black hole. By equating eq. 16 and eq. 13 (in the limit
m τ  τi ) we can determine the proto-SMS mass upper limit7
r(m) = . (13)
m1/2 + [(τ (m? ) − τi (m))]
 2/3
Ṁ?
The evolution or r as function of M for various values of Ṁ <
M? ∼ 8.48 × 105 M . (17)
M yr−1
is shown in Fig. 1. The radius grows proportionally to m1/2
for the inner shells, but flattens out in the envelope whose Thus, for an accretion rate of 0.15 M yr−1 (typical of
size relative to the stellar radius containing 90% of the stellar atomic cooling halos) a proto-SMS will collapse into a DCBH
mass increases with Ṁ , becoming as large as 104 R in the when its mass reaches ≈ 2.4 × 105 M . Interestingly, this
most extreme case (M? , Ṁ ) = (108 M , 10 M yr−1 ). mass limit is comparable to the mass possibly marking the
The previous results imply that in principle the SMS transition to ZAMS according to ongoing stellar evolution
final mass could be extremely large, provided a sufficiently 1D numerical calculations.
large halo gas reservoir is present to feed it. However, the
growth might be hindered by at least three factors. The
first is the possible transition to the ZAMS (shown in Fig. 6 We assume a fully ionized H+He gas and adopt a He abundance
1). In this case, as already discussed above, radiative feed- Y = 0.2477 (Peimbert et al. 2007); this gives a mean molecular
back of UV photons from the now hot stellar surface will weight µ = 0.59.
quench accretion. Although not expected, our simple treat- 7 Note that the dynamical time (Gρ )−1/2  t
c KH for M? >
ment cannot exclude that. Detailed numerical simulations 108 M . Hence no hydrostatic equilibrium is possible beyond this
accounting for the opacity evolution will be required to ad- mass.

c 2012 RAS, MNRAS 000, 1–??
 The IMF of black holes seeds 5
2.2 Rotating proto-SMS
If the star is rotating, this has a stabilizing effect and can
hold up the collapse. In this case the expression for the adi-
abatic index must be modified (Janka 2002) as follows:
2(2 − 5η) Rs
Γc = + 1.12 , (18)
3(1 − 2η) R?
which of course gives the correct limit (eq. 14) if the rota-
tional to gravitational energy ratio, η = T /|W | → 0. For a
maximally and rigidly rotating n = 3 polytrope, Baumgarte
& Shapiro (1999) find that η approaches the universal value
0.009 and that the instability criterion simply becomes:
 
R?
= 321, (19)
Rs c
where we have defined the stellar radius as the equatorial
one, ≈ 3/2 times the polar radius. We thus obtain an equa-
tion analog to eq. 16 for the critical radius:
rc < 5.26 × 10−7 m. (20)
By equating eqs. 16 and 20 we find that rotation increases
the stability of stars with masses M?r > 1.7 × 106 M (green
region in Fig. 1), while at lower masses thermal pressure
alone is sufficient to stabilize the star.
The results obtained in this Sec. can then be summa-
rized by the following formulae giving the stability condi-
tions8 for a proto-SMS:
 2/3
Ṁ?
M? ∼<
8.48 × 105 M (T = 0), (21)
M yr−1 Figure 2. Schematic view of the DCBH formation and growth
scenario discussed in Sec. 3.
 
Ṁ?
<
M? ∼ 6.01 × 105 M (T 6= 0, M? > M?r ). (22)
M yr−1
proto-SMS that are more massive than the above limits will by Reisswig et al. (2013) who studied the three-dimensional
inevitably collapse to form a DCBH. general-relativistic collapse of rapidly rotating supermassive
Likely, the entire proto-SMS mass will be finally locked stars. In the following, therefore, we will make the assump-
into the DCBH. This is confirmed by the results presented tion that M• ≈ M? .
in Montero et al. (2012) who performed general relativis-
tic simulations of collapsing supermassive stars with and
without rotation and including the effects of thermonuclear
energy released by hydrogen and helium burning. They find
that at the end of their collapse simulation (t ≈ 105 s) of 3 COSMOLOGICAL SCENARIO
a proto-SMS of mass 5 × 105 M , a black hole has already
> So far we have built a physical framework for the formation
formed and its apparent horizon contains a mass ∼ 50%
of the total initial mass. Analytical arguments discussed by of IMBH occurring either via GR instability of a proto-SMS
Baumgarte & Shapiro (1999) and later refined by Shapiro followed by accretion, or as the end point of the evolution
& Shibata (2002) reach a similar conclusion, indicating that of a more standard SMS. We now aim at embedding such
≈ 90% of the proto-SMS mass actually ends up into the a framework in a cosmological scenario to derive the global
DCBH, leaving a bare 10% of matter in an outer region, population properties of these potential SMBH seeds. As the
possibly a circumstellar disk. This high collapse efficiency is first stars form in minihalos at very high redshifts (Naoz
essentially a result of the highly concentrated density pro- et al. 2006; Salvadori & Ferrara 2009; Trenti & Stiavelli
file of n = 3 polytropes. Similar conclusions are reached 2009) their cumulative UV radiation boosted the intensity of
> ?
the LW background to values ∼ Jν,c for which star forma-
8
tion in newly born minihalo is quenched. Precisely quantify-
In principle these conditions should be complemented with the
ing the level of suppression is difficult as it depends of several
one expressing the ability of the SMS to maintain its extended ac-
fine-grain details, although reasonable attempts have been
creting envelope during the growth. This relation has been derived
by Schleicher et al. (2013): M? ∼ < 3.6 × 108 M (Ṁ /M yr−1 )3 , made (Ahn et al. 2009; Xu et al. 2013). For these reasons,
 ? 
or M? = 1.2 × 106 M for an accretion rate of 0.15 M yr−1 . and to bracket such uncertainty, we will assume that (a) all
However, in practice, this condition is met only when the star is minihalos with mass above a certain threshold correspond-
already GR unstable for the accretion rates > 0.1 M yr−1 con- ing to a virial temperature Tsf ≈ 2000 K, or (b) none of
sidered here, and therefore we will disregard it in the following. them, form stars, i.e. radiative feedback is either moderate

c 2012 RAS, MNRAS 000, 1–??
6 Ferrara et al.
or extremely effective9 . We will refer to these two possibili- consequence, in this case the mass of the IMBH that forms
ties as the “fertile” or “sterile” minihalo cases, respectively. at the end of the brief (≈ Myr) stellar lifetime is the same as
Following the growth of cosmic structures, atomic- the SMS. The main difference between these two IMBH for-
cooling halos start to appear as a result of accretion and mation channels is that for the GR instability channel, the
merging of minihalos. In the fertile minihalo case a frac- birth mass function of DCBHs is modified during the sub-
tion of them are born polluted; if minihalos are instead sequent feedback-regulated growth. This does not happen if
sterile, atomic halos are metal-free by construction as no the proto-SMS reaches the ZAMS, as already explained. For
stars/metals have been produced at earlier epochs. In both this reason we will compute the IMF of IMBH in two steps:
cases a fraction of them will be located in regions in which first computing the DCBH IMF (Sec. 5), and then modifying
•
Jν > Jν,c and therefore they are candidate IMBH forma- it to account for the additional feedback-regulated growth.
tion sites. Recent studies (Ahn et al. 2009; Yue et al. 2013b) What could cause the gas accretion rate to drop-off be-
have shown that during cosmic dawn, large spatial UV field fore the proto-SMS has become GR unstable? There are
intensity fluctuations existed and persisted for long times. several potential show-stopper effects that could come into
High illumination regions are then found near the peak of play. The first is that major galaxy mergers, in contrast with
the field intensity. the smoother accretion of small lumps of matter, are likely
Our main goal here is to determine the Initial Mass to dramatically perturb the smooth accretion flow onto the
Function of the IMBH once they are allowed to form by proto-SMS. An additional effect of the merger could be that
the environmental conditions discussed above. We will not the shock-induced electron fraction enhances the cooling.
attempt here to quantify (as done e.g. in Dijkstra et al. This mechanism was initially proposed by Shchekinov &
2008) how many among the unpolluted, atomic-cooling halos Vasiliev (2006) and Prieto et al. (2014), and recently con-
•
reside in Jν > Jν,c regions: this multiplicative (to first-order) firmed by a numerical simulation in the absence of radiative
function is only required to determine the number density backgrounds (Bovino et al. 2014).
of such objects. Such an assumption is equivalent to state Secondly, the gas brought by the merging halos (or col-
that newly-born Tvir > 104 K halos reside in highly biased lected from the intergalactic medium) can be already pol-
regions where the field is sufficiently intense; moreover, if the luted with heavy elements. As a result, clump formation
first few among them manage actually to form IMBH, the following metal-cooling fragmentation of the gas is likely to
radiation field of the latter will greatly amplify Jν , triggering drastically quench the accretion rate onto the proto-SMS,
the birth of additional IMBH (Yue et al. 2013b). limiting its growth.
The formation of IMBH in atomic halos starts with an Finally, the halo could be very gas-poor as a result of
isothermal, coherent collapse centrally accumulating the gas gas ejection by supernova explosions occurred in the pro-
at rates comparable or larger than the thermal accretion genitor halos. In conclusion, even if the minimal conditions
rate, for IMBH formation in a give halo are met, the hidden and
3/2 quiet growth of the proto-SMS finally leading to a DCBH
π2 3

T via GR instability is hindered by a number of effects. All
Ṁi ≈ cs = 0.162 M yr−1 . (23)
8G 104 K these possible physical paths are graphically summarized in
As simple as it is, this formula is in remarkable agreement Fig. 2.
with the results of most recent and complete simulations of The main challenge of the problem is to consistently fol-
the collapse of atomic halos. For example, by analyzing 9 low the growth of a proto-SMS inside an atomic halo within
such halos extracted from a cosmological large-eddy simula- a cosmological context, i.e. following the history of the par-
tion, Latif et al. (2013a) found a very similar accretion rate ent halo as it merges with other halos and accretes gas from
of 0.1 − 1 M yr−1 , measured at z = 15, with little depen- the intergalactic medium. We accomplish this by using a
dence on the galactocentric radius. merger tree approach as described in the following.
As long as this high accretion rate can be maintained,
the proto-SMS continues to grow. Because of its low effec-
tive temperature, radiative feedback is unable to stop the 4 MERGER TREES
halo gas from accreting. Eventually, the proto-SMS hits the
GR unstable boundary shown in Fig. 1 and collapses into In order to quantitatively investigate the above scenario we
a DCBH in a very short time (about 105 s). Once formed, follow the merger and mass accretion history of dark matter
the DCBH will continue to grow increasing its birth mass by halos and their baryonic component. To this aim we use
accreting the gas leftover (if any) in their parent atomic halo the data-calibrated merger-tree code GAMETE (GAlaxy
finally becoming an IMBH. This feedback-regulated growth MErger Tree and Evolution, Salvadori et al. 2007), which
is a complex process and we will discuss it separately in Sec. has been developed to investigate the properties of present-
6. However, if during this accretion phase the rate for any day ancient metal-poor stars. The code successfully re-
reason drops below ≈ 0.1 M yr−1 , then the proto-SMS be- produces the metallicity-luminosity relation of Milky Way
gins to contract and evolves into the ZAMS SMS. Given the (MW) dwarf galaxies, the stellar Metallicity Distribution
corresponding high effective temperature (eq. 3), the SMS Function (MDF) observed in the Galactic halo, in classical
luminosity exerts a sufficient radiation pressure on the sur- and ultra-faint dwarfs (Salvadori & Ferrara 2009), and the
rounding gas. Hence accretion is completely halted. As a properties of very metal-poor Damped Lymanα Absorbers
(Salvadori & Ferrara 2012). Here we only summarize the
main features of the code, deferring the interested reader to
9Strictly speaking even in hypothesis (b) a small number of halos the previous papers for details.
must form anyway to provide the UV radiation field. GAMETE reconstructs the possible merger histories of

c 2012 RAS, MNRAS 000, 1–??
 The IMF of black holes seeds 7
a MW-size dark matter halo via a Monte Carlo algorithm the host halo (including that brought by mergers) remains
based on the Extended Press-Schechter theory (Salvadori metal-free, the proto-SMS grows at a rate set by Ṁ∗ until
et al. 2007), tracing at the same time the star formation (SF) it eventually becomes GR unstable (eq. 17).
along the hierarchical trees with the following prescriptions: However, GR instability is not necessarily the final
(i) the SF rate is proportional to the mass of cold gas in fate for the proto-SMS. In fact, the protostar growth can
each galaxy, and to the SF efficiency ∗ ; (ii) in minihalos ∗ stop because of two distinct physical processes (a) a major
is reduced as H2 = 2∗ [1 + (Tvir /2 × 104 K)−3 ]−1 due to the merger event, i.e. a collision with a halo of comparable mass,
ineffective cooling by H2 molecules; (iii) Population II stars M1 /M2 = [0.5, 2], and (b) pollution from heavy elements
form according to a Larson IMF if the gas metallicity exceeds carried by a merging halo or acquired from the IGM. For
the critical value, Zcr = 10−5±1 Z (Schneider et al. 2006), reasons already explained both events are likely to suppress
here assumed Zcr = 10−6 Z . At lower metallicity, PopIII gas accretion onto the proto-SMS. Therefore we assume that
stars form with reference mass m∗ = 25M and explosion major mergers and/or accretion of polluted gas stop ac-
energy ESN = 1051 erg consistent with faint SNe (Salvadori cretion and lead to a zero-age main sequence SMS, which
& Ferrara 2012). The chemical evolution of the gas is simul- shortly after will collapse into an IMBH. To account for these
taneously traced in haloes and in the surrounding MW envi- events we keep track along the tree of major mergers and
ronment by including the effect of SN-driven outflows, which mergers with “killer” halos, i.e. halos enriched with heavy
are controlled by the SN wind efficiency. The metal filling elements by previous episodes of star formation. When one
factor, QZ = VZtot /Vmw , is computed at each z by summing of these termination events occurs, the proto-SMS growth in
the volumes of the individual metal bubbles around star- that halo is stopped. We store the final masses of both SMS
forming haloes, VZtot , and where Vmw ≈ 5(1 + z)−3 Mpc3 , and DCBH along with all the information on their lifetime,
is the proper MW volume at the turn-around radius (Sal- cosmic formation epoch and parent halo properties. Finally,
vadori et al. 2014). The probability for newly formed halos we average the results over 50 random realizations of the
to reside in a metal enriched region is then computed as MW-analog halo merger tree.
P (z) = [1 − exp(QZ )]/Qδ>δc , where Qδ>δc (z) is the vol-
ume filling factor of fluctuations with overdensity above the
critical threshold, δ > δc = 1.686, for the linear collapse 5 BIRTH MASS FUNCTION
(Miralda-Escudé et al. 2000). The latter quantity describes
the abundance of high-density regions, in which metals first We are now ready to derive what we call the “birth mass
penetrate (Tornatore et al. 2007; Pallottini et al. 2014). Ob- function”. This is the mass distribution including the newly
jects in enriched (primordial) regions are assigned an initial formed DCBH originating from GR instability of a proto-
metallicity Zvir = ZGM /[1 − exp(QZ )] (Zvir = 0), where SMS, and the black holes corresponding to the end point of
ZGM is the average metallicity of the MW environment. the SMS evolution. The final IMF of IMBH seeds (Sec. 6)
As primordial composition (Z < Zcr ) halos in the needs to additionally account for the subsequent feedback-
merger tree cross the Tvir = 104 K threshold, we postu- regulated growth of DCBH.
late that a proto-SMS can form in each of them and follow The birth mass function will be presented for the two
its growth in the following manner. We assume that the limiting cases of fertile (representing the fiducial case) and
proto-SMS is fed by an accretion rate, Ṁ∗ = max(Ṁi , Ṁe ), sterile minihalos. The fertile case assumes that all mini-
that is the maximum between the “internal” accretion rate halos with Tvir > Tsf = 2000 K can form stars when
from the host halo gas, Ṁi , and the “external” accretion z > 10. At lower redshifts Tsf slowly increases up to the
rate due to minor10 merger events, Ṁe . Ṁi is computed as value Tsf ≈ 2 × 104 K reached at z ≈ 6, when MW envi-
the thermal accretion rate (eq. 23). The external rate Ṁe is ronment is reionized (Salvadori et al. 2014). This empirical
taken as the ratio between the gas mass of the merging halo functional form catches the essence of the increasing ability
and the dynamical friction time-scale, tmerge , between the of the LW radiation to suppress star formation in halos as
two colliding objects. We compute tmerge from the classic its intensity climbs and it is calibrated on a detailed com-
Chandrasekhar formula (Mo et al. 2010): parison with the dwarf galaxy population of the MW halo
(Salvadori & Ferrara 2009). The sterile minihalo case, in-
ζx stead, assumes that minihalos never form stars. These two
H(z)tmerge = 0.234 (24)
ln(1 + x2 ) cases are meant to bracket the uncertain role of radiative
where x = M1 /M2 is the ratio of the merging halo masses feedback in suppressing star formation via H2 destruction
(with M1 > M2 ), and ζ is the circularity parameter encoding in these small systems. Note that in the first case atomic
the eccentricity of the orbit decay, which we randomly select halos resulting from the merger of smaller progenitors can
in the interval ζ = [0.1, 0.25] following Petri et al. (2012). be polluted with heavy elements when they form; if instead
This prescription allows the occurrence of supra-thermal ac- minihalos are sterile, atomic halos are all born unpolluted.
cretion rates, consistently with the numerical simulation re-
sults of Mayer et al. (2010) (see also Ferrara et al. (2013) for
5.1 Fertile minihalos
a critical discussion) that show that following merger events
the gas rapidly loses angular momentum and is efficiently The main results for this case are depicted in Fig. 3. In
funneled towards the nuclear region. As long as the gas in the left panel we show the masses of DCBH and SMS,
and of their hosting halos (Mh ) as a function of their for-
mation redshift for a single realization of the merger tree.
10For reasons explained later in this Section, we stop the proto- For comparison purposes, the mass of all atomic halos in
SMS growth after a major merger. the merger tree at different redshifts are also shown (gray

c 2012 RAS, MNRAS 000, 1–??
8 Ferrara et al.

Figure 3. Left: As a function of their formation redshift we show the mass of: (i) halos with Tvir > 104 K (gray open squares), (ii) halos
hosting a DCBH or SMS (green squares), (iii) DCBH (black circles), and (iv) SMS whose growth has been halted by a major merger
event (red exagones) or by metal-pollution (yellow exagones). The results are shown for a single MW merger history Right: Mass of gas
left in the halos after a DCBH (gray) or SMS (yellow-orange) formation as a function of the DCBH/SMS mass. Contours refer to 68%,
95%, 99.7% confidence levels (50 realizations).

points). The halos hosting DCBH or SMS span a well-

defined and narrow range of masses Mh ≈ (2 − 10) × 107 M
during the entire formation epoch, 8 < z < 17. They are
low-mass systems, which at any given redshift have roughly
the minimum virial temperature required for atomic cool-
ing, Tvir ≈ (1 − 1.3) × 104 K. While the chances to remain
unpolluted are relatively high for these small systems (30
such halos out of a total of 36, i.e. ≈ 83%), this probability
drops rapidly for more massive Tvir > 1.3 × 104 K objects,
which form via merging of smaller progenitors that have al-
ready formed stars. These results are more quantitatively
illustrated in the upper panel of Fig. 4, which shows the
probability distribution function (PDF) of host halo masses.
The mass distribution of DCBH hosts (gray histogram) has
an almost symmetric distribution with a pronounced peak
around Mh ≈ (3.5 − 5) × 107 M and rapidly declines to-
wards the tails of the distribution. Most of the halos host-
ing DCBH (> 80% of the total) are low-mass objects with
Mh ≈ (2.5 − 6.3) × 107 M . Moreover, the external accretion
occurs at a subdominant rate with respect to internal one,
Ṁe = [0.0004 − 0.16] 6 Ṁi . This implies that most of the
time the proto-SMS accretes gas at the thermal accretion Figure 4. Mass probability distribution function of DCBH (dot-
rate set by the halo virial temperature (eq. 23), and hence ted histogram) or SMS (yellow shaded histogram) host halos for
comparable for all of them, Ṁ∗ ≈ (0.162 − 0.57) M yr−1 . the fertile (upper panel) and sterile (lower panel) minihalo cases.
Since the final mass of DCBHs is entirely determined by Ṁ∗ The results are averaged over 50 MW merger histories, and the
(see eq. 17), it follows that also the DCBHs span a very nar- ±1σ errors are shown.
row range of masses, 2.5 × 105 M ∼ < <
M• ∼ 4.5 × 105 M , as
can be appreciated by inspecting Fig. 3 (black points).
In the same Figure we show the mass of SMS (filled pollution is the dominant process stopping the proto-SMS
exagones) whose growth has been blocked before reaching growth (22 out of 24 SMS share this origin). Moreover, the
the GR instability because of: (i) a major merging event typical masses of SMS are smaller than DCBH, although
(red symbols), or (ii) a minor merger with a metal polluted they span a larger range, MSM S ≈ (3 − 45) × 104 M . This
halo (yellow symbols). It is clear from the Figure that metal is due to the stochastic nature of the merging/accretion

c 2012 RAS, MNRAS 000, 1–??
 The IMF of black holes seeds 9
processes, which can quench the growth of the proto-SMS tion. Such a quantity is shown as a function of DCBH/SMS
at different stages. We can also note that the formation mass in Fig. 3 (right panel). Although DCBH masses span
epoch of SMS is shifted towards lower redshifts with re- a very small range the gas mass can vary by more one order
spect to DCBH, z ≈ (14 − 8). Indeed, the probability to of magnitude Mg ≈ (3 − 50) × 105 M . This gas can be po-
merge/accrete metal enriched gas is very low at z > 15, tentially accreted by DCBH. Thus, as we will discuss in the
when only a few halos have formed stars and QZ 6 0.002. next Section, the final mass of IMBH seeds will crucially de-
However, it gradually increases at lower redshift becoming pend on the subsequent accretion phase and feedback effects.
≈ 1 at z 6 9, when the growth of all proto-SMS is stopped SMS are instead able to evacuate the gas that is not quickly
because of metal pollution. Due to this delay the hosting turned into stars. Halos residing within 68% confidence level
halos of SMS are typically more massive than DCBH hosts have Mg = fb (Ωb /ΩM )Mh with fb ≈ 0.2. Such a reduced gas
as seen in Fig. 4. fraction with respect to the cosmological value is the result
All these findings can be better interpreted by inspect- of the previous star-formation activity and SN feedback pro-
ing Fig. 5, where the comoving number density of DCBHs cesses, occurred in their progenitors. Only a few halos that
(black points), SMS (yellow/red exagones), and halos with form at z ≈ 20, and correspond to rare high-σ density fluc-
different physical properties (all, Tvir > 104 K, unpolluted), tuations, are found to have fb ≈ 1. The halos hosting SMS
are shown as a function of redshift. It is evident that the cover roughly the same Mg range as DCBH hosts. The bulk
number density, n, of unpolluted atomic halos (blue trian- of SMS hosts, however, are more gas rich than DCBH hosts.
gles), differently from the other curves does not monotoni- This is because SMS typically form in more massive halos
cally increase with time. Instead n gently grows from z ≈ 20 (see Fig. 4), which therefore contain more gas.
to z ≈ 13, reaches a maximum, and then slowly decreases
when metal pollution starts to dominate. The number den-
sity of both DCBH and SMS are tightly connected with this 5.2 Sterile minihalos
evolution. From z = 20 to z ≈ 17 the amount of DCBHs In the second case, we consider the other extreme possibil-
increases steeply, tightly following the rise of unpolluted ha- ity in which the UV flux is sufficiently intense to completely
los, while SMS are very rare, n 6 0.02Mpc−3 . At lower z the suppress star formation in minihalos. As we commented al-
steepness of the curve progressively decreases, becoming flat ready, strictly speaking this case corresponds to an unphys-
at z ≈ 9.5, while the SMS density progressively increases, ical situation as at least some stars must form in order to
gradually approaching the DCBH value. Below z ≈ 8, also produce the required radiation field (unless some other UV
the formation of SMS is stopped, and n becomes constant. source is present, as for example dark matter annihilation).
At this z, metals have already reached the high density re- The sterile minihalo scenario requires that the fraction of
gions in which halos form, QZ ≈ 0.2 > Qδ>δc , making the baryons converted into stars in these systems is negligible.
onset of proto-SMS formation impossible. The final num- With this hypothesis and caveat in mind we can an-
ber density of DCBHs and SMS are expected to be roughly alyze the results of the merger trees and the predicted
the same, n ≈ 7 Mpc−3 . We recall that this number has properties of DCBH/SMS and of their hosting halos. From
been obtained assuming that all host halos reside in regions Fig. 7 we highlight two major differences with respect to
•
in which Jν > Jν,c , and therefore represent a strong upper the previous case. First, the formation era of DCBH/SMS
limit. stretches towards lower redshifts, z ≈ 7. Second, the prob-
In Fig. 6 we show the mass probability distribu- ability to merge with an already polluted halo strongly de-
tion functions of DCBH (gray histogram) and SMS (yel- creases. Both these effects are simply a consequence of the
low histogram) normalized to the total number of objects lack of an early metal enrichment, impying that all atomic
(DCBH+SMS), i.e. what we call the ”Birth Mass Function”. halos are unpolluted at birth. In these objects therefore star-
The mass distribution of DCBH exhibits a peak roughly formation, and the subsequent metal-enrichment, is only ac-
at the low-mass end, M• ≈ 2.5 × 105 M , and monotoni- tivated if a major merger event induces a vigorous fragmen-
cally declines towards higher masses. The lower limit of the tation of the gas, thus stopping the proto-SMS growth.
PDF is populated by objects accreting at the thermal rate, An inspection of the right panel of Fig. 5 further il-
Ṁ∗ ≈ 0.162, corresponding to Tvir ≈ 104 K halos, the most lustrates these points. The number density of unpolluted
common DCBH hosts. This sharp low-mass cut is set by GR atomic halos and Tvir > 104 K halos exactly overlap down
instability (eq. 17). On the other hand, more massive DCBH to z ≈ 15. At lower z, however, the two functions start to
form in unpolluted halos with higher Tvir , that therefore are slowly deviate. At z ≈ 10 the number density of unpol-
much less common. This explaines the rapid downturn of the luted atomic halos, i.e. the sites for SMS/DCBH formation,
distribution. reaches the maximumand then rapidly declines since metal
The PDF of SMS has a very different, roughly sym- pollution start to dominate. The maximum value is almost
metric shape, displaying a wide plateau in the mass interval one order of magnitude larger than found fertile minihalos.
M ≈ (0.8 − 2.2) × 105 M ; it then rapidly declines towards As a consequence, the final number density of DCBH is much
lower/higher masses. The decreasing number of SMS with larger, n ≈ 65 Mpc−3 .
masses < 8×104 M depends on the limited number of merg- Despite of the drastically different conditions between
ing/accretion (driving the proto-star evolution towards the the fertile and sterile cases, the properties of DCBH/SMS
ZAMS) occurring on timescales equal to 8 × 104 M /Ṁ∗ ≈ and of their hosting halos are surprisingly similar. This is
0.1 Myr. The PDF decline at M > 2.5 × 105 M is due to evident both from the mass probability distribution func-
the same processes decribed for DCBH. tions shown (Fig.4) and in the birth mass function (Fig. 6).
Another quantity that we can derive from the previous Nevertheless there is a remarkable difference, that can be
analysis is the gas left in the halo after DCBH or SMS forma- noted by comparing the right panels of Fig. 7 and Fig. 3. If

c 2012 RAS, MNRAS 000, 1–??
10 Ferrara et al.

Figure 5. Comoving number density evolution of: (a) all halos in the simulation (top curve), (b) halos with Tvir > 104 K (purple filled
squares), (c) unpolluted (Z < Zcrit = 10−6 Z ) halos with Tvir > 104 K (blue triangles) (d) DCBH hosts (black circles), (e) SMS hosts
(exagons) formed after (f) a major merger (orange) or (g) metal-pollution event (yellow). The errors are only shown for the total number
of halos and represent the ±1σ dispersion among different merger histories. Left: fertile minihalos case; Right: sterile minihalos case.

Figure 7. As Fig. 3 for the sterile mini-halo case.

minihalos are sterile the gas mass at the DCBH/SMS forma- 6 FEEDBACK-REGULATED GROWTH
tion is larger, Mg ≈ (1 − 30) × 106 M . At their formation,
indeed, all Tvir ≈ 104 K haloes have a gas mass fraction
close to fb ≈ 1, since no gas have been consumed within
their sterile progenitor minihalos. This implies that the sub-
sequent feedback-regulated DCBH accretion phase, will be In order to determine the IMF of the IMBH the final step is
crucial in setting the final IMF of IMBH for the two different to asses wheter they were able to accrete the gas eventually
scenarios. left at the time of DCBH/SMS formation. This is the goal
of this Section.

c 2012 RAS, MNRAS 000, 1–??
 The IMF of black holes seeds 11
by Latif et al. (2013a), who showed that the 9 candidate
DCBH host halos (all of mass M ≈ 107 M ) remarkably
follow the distribution given by eq. 25, independent on their
mass and formation redshift (note that both the mass and
redshift range are rather narrow as we have shown in the
previous Section). The core radius, rc , is comparable to the
Jeans length of the gas ∝ cs tf f as in a King profile for which
 1/2  −1/2
3cs T ρc
rc = √ = 65.5 AU,
4πGρc 104 K 10−11 g cm−3
(26)
where the core density reference value is taken from Fig. 1
of Latif et al. (2013a). The previous formula gives a core
radius in very good agreement with the simulated value.
Finally, we require that the mass contained within the outer
radius, rout , at which we truncate the distribution is equal
to Mg . This gives (in the reasonable limit rout  rc ),

Mgi GMgi
rout ≈ 2
= = 14.3 pc. (27)
4πρc rc 9c2s
Note that, due to collapse, the gas concentration increases,
Figure 6. Mass probability distribution function of DCBH seeds i.e. rout is more than 10 times smaller than the halo virial
(dotted histogram) and SMS (yellow shaded histogram) for the radius
fertile (upper panel) and sterile (lower panel) minihalos case. The  1/2  −3/2
symbols are the same of Fig. 4. Tvir 1+z
rvir = 583 pc. (28)
104 K 15

6.1 Formation via SMS Assuming that, to a first approximation, the DCBH is
at rest and that the accretion flow is close to spherical11 ,
If the transition to ZAMS occurs before the onset of GR the relevant scale for accretion is the Bondi radius,
instability (eq. 16) an SMS forms. Due to the large amount   −1
of UV photons emitted by the hot (Tef f ≈ 105 K, see eq. 3) 2GM• M• T
rB = 2 = 9.9 pc, (29)
stellar surface the radiation pressure on the remaining gas is cs,∞ 105 M 104 K
very likely to evacuate it during its lifetime, expected to be
where we denote with the subscript ∞ quantities evaluated
t? ≈ 0.007M? c2 /LE = 3 Myr, virtually independent on its
at large distances from the DCBH. Note that rB is about 2%
mass. Obviously we cannot exclude that some fraction of this
of the virial radius of a typical DCBH host halo, and rB ≈
gas can be turned into stars before this happens. Irrespective
rout , implying that the DCBH can easily drain gas from
of these details by the time the SMS collapses into an IMBH
the entire volume in which gas is present. This fact has two
there will be virtually no gas left to accrete and further
important consequences that we analyze in the following.
growth becomes impossible. In this case therefore, the IMF
The first implication of the approximate equality be-
for these type of IMBH is the birth mass function itself.
tween the Bondi and outer radius is that the initial gas den-
sity distribution will be modified by the accreting DCBH.
6.2 Formation via GR instability The rearrangement of the gas requires that the dynamical
time is shorter than the Salpeter time, i.e.
As the DCBH of mass M•i emerges from the collapse of  1/2
the proto-SMS, it will be surrounded by the remaning halo 3π M•
tf f =  ≡ tS = 4.4 × 108  yr, (30)
gas mass, Mg , that has not been previously included into 32Gρ Ṁ•
the DCBH. These two initial values are obtained from the
where we have conservatively assumed that accretion occurs
merger tree outputs, along with the total halo mass Mh
at the Eddington rate. The minimum density required to re-
(i.e. 3 and 7). We assume that during the DCBH growth
arrange the profile fast enough is ρ = 2.17 × 10−24 g cm−3 ,
phase Mh ≈ const. given the short duration of such phase.
having further assumed a standard radiative efficiency  =
In principle, all the remaining gas could be eventually incor-
0.1. As from eq. 25 we obtain that the gas density is always
porated into the DCBH unless feedback from energetic ra-
larger than the previous value we can safely assume that
diation emitted during the accretion process is able to stop
this is the case.
or reverse the accretion flow.
In order to obtain an explicit expression for the accre-
The typical density structure resulting from the isother-
tion flow density profile, let us proceed as follows. The one-
mal collapse of the halo gas prior to DCBH formation is con-
dimensional mass and momentum conservation equations for
stituted by a central (adiabatic) core in which the collapse
is stabilized, and an outer envelope where ρ ∝ r−2 :
ρc
ρ(r) = . (25) 11 The spherical approximation holds if rB is larger than the cir-
1 + (r/rc )2
cularization radius rc = j 2 /GM• , where j is the specific angular
The above density profile has been confirmed by simulations momentum of the gas.

c 2012 RAS, MNRAS 000, 1–??
12 Ferrara et al.
a steady adiabatic accretion flow, i.e. the classical Bondi the density increases to values that are large enough to effec-
problem, read tively trap photons; within this region the energy is convec-
tively rather than radiatively transported by diffusion. It is
1 dρ 2 1 dv
=− − , (31) easy to transform this condition (which is also equivalent to
ρ dr r v dr
the Schwarzschild criterion for stability against convection)
into one on τ . The radiative luminosity can be written as
dv 1 dp GM•
v + + = 0. (32) 16πr2 c 3 dT
dr ρ dr r2 Lr = − aT , (38)
3κρ dr
Taking c2s,∞
= γp∞ /ρ∞ , where γ is the adiabatic index, as
where a is the radiation density constant, and κ = σT /µmp .
the sound speed at large distances, and further assuming12
On the other hand, the convective luminosity can be written
v∞ = 0, we can integrate eq. 32 to get the Bernoulli equa-
as
tion,
Lc = 16πr2 pv; (39)
1 2 cs (r)2 GM• c2s,∞
v + − = , (33)
2 γ−1 r γ−1 if pressure inside the trapping region is dominated by radia-
tion, then p = aT 4 /c. By equating the two luminosities and
from which the classical Bondi accretion rate can be derived
recalling that dτ = κρdr we obtain the implicit definition
by evaluating the previous expression at the sonic radius
for the trapping radius, rtr :
rs = GM• /2c2s (rs ):
v(rtr )
ṀB = 4πrs2 ρ(rs )cs (rs ) = πqs rB
2
cs,∞ ρ∞ , (34) τ (rtr ) = 1. (40)
c
where Thus, for r < rtr radiation is convected inward faster than it
 (5−3γ)/(2γ−2) can diffuse out and therefore within this radius (convective
1 2 region) photons cannot escape; the flow is then almost per-
qs (γ) = . (35)
4 5 − 3γ fectly adiabatic. At larger radii, radiation can start diffuse
and transport energy outwards: we refer to this region as the
The numerical value of qs ranges from qs = 1/4 at γ = 5/3 to
radiative layer. It can be easily shown that Lc (rtr ) = βLE ,
qs = e3/2 /4 ≈ 1.12 when γ = 1 (isothermal);
√ in a radiation-
with β = O(1). Using eqs. 37 and 34 to express τ and re-
dominated fluid (γ = 4/3) then qs = 2/2.
calling that v = ṀB,γ=4/3 /4πρr2 , we finally obtain
Inside the sonic radius rs = 1/8rB for γ = 4/3, the
√
Bernoulli equation reduces to (1/2)v 2 ≈ GM• /r, which 2  cs,∞ 
yields v(r) = cs,∞ (r/rB )−1/2 . To conserve the Bondi rate rtr = τB rB  rB , (41)
4 c
then the radial density dependence can be easily shown to
and
satisfy  
cs,∞ c
 −3/2 Lc = 48β ṀB . (42)
r τB
ρ(r) = ρB , (36)
rB
The temperature at the trapping radius is then simply ob-
4 2
where ρB = 3
3Mg /8πrBis a normalization constant obtained tained from Ttr = L/πacrtr . We will show later that the
by requiring that at each time the mass contained within radiative region is very thin. The DCBH growth rate can
rout is equal to the current gas mass Mg (t). then be determined by equating its accretion luminosity
Two points are worth noting. First, the −3/2 depen- η Ṁ• c2 /(1 − η) to Lc to obtain
dence of density is independent of the value of γ. More- 
1−

cs,∞

over, although it has been obtained under a steady-state as- Ṁ• = 48β ṀB . (43)
 cτB
sumption, it has been shown to hold also for time-dependent
(Sakashita 1974), and even optically thick (Tamazawa et al. To determine the thermal structure of the radiative region
1975) accretion flows. Thus we consider it as a robust fea- and compute the photospheric temperature of the accreting
ture of our model. In addition, as the dynamical time at DCBH we need to improve our treatment of the opacity. So
small radii is much shorter than at rout , we keep the latter far we have assumed a constant electron scattering opac-
fixed during the evolution and allow ρB to decrease as gas ity, κT = σT /µmp . In a metal-free gas, this is a good ap-
is incorporated into the DCBH. proximation as long as the temperature remains ∼ >
5 × 104
From the density we can compute the optical depth K. At lower temperatures, as the gas starts to recombine,
to Thomson scattering (we will discuss later on when this additional processes increase the gas opacity: (a) free-free
simple opacity prescription breaks down and adopt a more (κf f ∝ ρT −7/2 , known as the Kramers opacity); (b) bound-
precise formulation): free and free-bound; (c) H− (κH − ∝ ρ1/2 T 9 ) which is mostly
effective in the temperature range (0.3 − 1) × 104 K. A full
Z r
ρ(r)  r 1/2 r
τ (r) = − σT dr = 2τB
B  (37) calculation of the opacity is given in Mayer & Duschl (2005);
rout µmp r here we use a fit to their results suggested by Begelman et al.

rout
(2008):
where we have defined τB = nB σT rB . Towards the center
κT
κ(T ) = , (44)
1 + (T /T∗ )−s
12 This may hold only approximately if the gas accretion onto with T∗ = 8000 K and s = 13. The above expression
the halo from the intergalactic medium is still occurring for the Rosseland mean opacity is independent on density.

c 2012 RAS, MNRAS 000, 1–??
 The IMF of black holes seeds 13

Figure 8. Dependence of various characteristic scales of the prob- Figure 9. Dependence of several characteristic radii of the system
lem as a function of the core gas density, ρc for a DCBH of mass (see text for definitions) on the DCBH mass along with the tem-
M• = 105 M , located in a dark matter halo with Tvir = 104 perature at the trapping radius, Ttr , (blue) and photospheric tem-
K formed at z = 14. We have assumed  = 0.1, fb = 1.0, and a perature, Tph (red). The DCBH of initial mass M• = 104.5 M , is
radiation-dominated equation of state corresponding to γ = 4/3. located in a dark matter halo with Tvir = 104 K formed at z = 14.
In addition to the core, rc , Bondi, rB , the outer gas distribution, We have assumed  = 0.1, fb = 1.0, and a radiation-dominated
rout , and virial, rvir radii, also shown are the Bondi and Edding- equation of state corresponding to γ = 4/3.
ton density regimes corresponding to the above (Tvir , M• ) pair
and the electron scattering optical depth out to rB .

the system remains relatively low, reaching in this case only

This turns out to be a very good approximation as long as 15970 K. Because of this low temperature the ionizing rate
ρ∼<
10−10 g cm−3 . As the use of the correct opacity becomes from the accreting envelope has a relatively mild feedback
important outside the trapping radius where densities are effect onto the overlying atmosphere, making it difficult to
comparable or below the above validity threshold, we con- stop the accretion of the leftover gas onto the DCBH.
sider this approximation as a safe and handy one. Fig. 9 gives a full view of the evolution of the system as
Armed with these prescriptions, we can solve for the the DCBH mass increases due to accretion, self-consistently
temperature structure in the radiative region using the en- calculated using eq. 43. As the DCBH mass increases the
ergy transport equation in the diffusion approximation: convective region shrinks due to the decreasing density as
"  0 −s # matter is progressively swallowed by the DCBH. At the same
3κT ρ(r0 )L 0
Z T Z r
03 T time, such contraction induces a temperature increase at the
T 1+ dT 0 = − 02
dr , (45)
Ttr T∗ rtr 16πacr convective/radiative layer boundary, paralleled by a similar
increase in the photospheric temperature. In particular, in
whose solution can be written, using the expression for the this specific case of a DCBH growing inside a dark matter
density eq. 36 and the definition of τB (eq. 37) as halo with Tvir = 104 K formed at z = 14 and fb = 1, Tph
"  4−s T # "
 r 5/2  r 5/2
# initially increases slowly and remains below 5500 K up to
4T∗4 T 3τB L B B
T4 + = 2
− . the point at which the DCBH mass crosses the value M• =
4 − s T∗ 10πacrB r rtr 105.4 M . Beyond that point the photospheric temperature
Ttr
(46) increases more rapidly and reaches about 30,000 K once the
To get the photospheric radius, rph , we solve numerically the DCBH has grown to M• = 106 M .
above equation together with the additional constrain that Thus it is only in these more advanced evolutionary
τ (rph ) = 2/3 as canonically used in stellar atmospheres, e.g. phases that copious amount of ionizing photons start to be
see Schwarzschild (1958). The temperature at rph is defined produced. As a result of radiative energy deposition the gas
as the photospheric temperature of the system. can be heated to a temperature far exceeding the virial tem-
The resulting structural properties of an accreting flow perature of the halo and therefore be evacuated from the
onto a DCBH of mass M• = 105 M , located in a dark mat- halo, preventing the accretion of gas located beyond rph .
ter halo with Tvir = 104 K, fb = 1 formed at z = 14 are The gas within the photosphere is eventually accreted and
shown, as an example, in Fig. 8. We find that the radiative the final state of the system is a naked IMBH embedded in
region is extremely narrow, i.e. rph ≈ rtr = 2.52 × 10−5 pc. the parent dark matter halo. Thus we are left with the final
Both radii are considerably smaller than the Bondi (9.9 pc) question of establishing when accretion, and hence DCBH
and virial (559.5 pc) radii. Inside rph the optical depth raises growth, will come to a halt.
to very large values. However, the effective temperature of From the detailed properties computed above we derive

c 2012 RAS, MNRAS 000, 1–??
14 Ferrara et al.

Figure 10. Initial Mass Function of IMBH seeds (shaded gray Figure 11. Probability distribution functions of the accretion
histogram) averaged over 10 MW merger histories for the fertile phase duration for the fertile (upper panel) and the sterile (lower
(upper panel) and for the sterile (lower panel) minihalo cases. panel) minihalo cases. The results are averaged over 10 MW
The errorbars correspond to ±1σ errors. The birth mass function merger histories and the ±1σ errors are shown.
of DCBH and SMS (see Fig. 6) is also shown (dotted histogram).

halo mass remain constant during this phase (no merging

the ionizing rate,
processes). The final mass distribution of DCBH is then
πφ(Tph ) 2 4 summed with the birth mass function of SMS (as these
Q(M• ) = rph acTph (47)
hhνi objects also finally evolve into black holes, see Section 5),
where hhνi ≈ 1 Ryd is the mean ionizing photon energy and normalized. The results of this calculation are shown
and φ is the fraction of the bolometric energy emitted by in Fig. 10, where the final IMF of IMBH (gray shaded his-
the accreting DCBH, whose spectrum is assumed to be a tograms) is compared with the birth mass function of DCBH
black-body, Bν (Tph ): and SMS (empty histograms). The low-mass end of the IMF
R∞ is identical to the birth mass function, while the peak is
dνBν (Tph )
νL
φ(Tph ) = R ∞ , (48) shifted towards higher masses. This is simply a consequence
0
dνBν (Tph ) of the feedback-regulated growth, only affecting DCBH. As
where hνL = 1 Ryd. In order to ionize the entire atmosphere the growth of DCBH is fed by the available halo gas, the
(i.e. the gas outside rph ) and increase the gas temperature displacement is larger when minihalos are sterile because in
above Tvir ≈ 104 K, the ionization rate must exceed the this case DCBH hosts are more gas rich (Fig. 6 and Fig. 3,
recombination rate, R, of the gas within rout . The latter right panels).
can be written as As a consequence, the IMBH IMF is very different in the
Z rout  2    two scenarios: in the fertile case, in particular, it exhibits a
r ρ
R(M• ) = 4π dr (49) bimodal distribution with two separate peaks at M ≈ (0.7−
rph trec µmp 1.2) × 105 M and M ≈ (5 − 10) × 105 M . The distribution
where trec = (nα(2) )−1 is the recombination timescale and extends over a broad range of masses, from M ≈ (0.5−20)×
α(2) = 2.6×10−13 (T /104 K)−1/2 is the Case B recombination 105 M . If minihalos are sterile, the IMF spans the narrower
rate of hydrogen (Maselli et al. 2003). By substituting eq. 36 mass range, M ≈ (1 − 2.8) × 106 M , which contains > 90%
and performing simple algebra we obtain the final expression of the IMBH population.
for the recombination rate: These differences are also reflected in the duration of the
accretion phase. DCBH can continue to grow almost unim-
4πα(2) ρ2B 3
 
rout peded for several tens of Myr before gas accretion is shut
R(M• ) = r B ln . (50)
(µmp )2 rph down by feedback, as illustrated by Fig. 11. From there we
Once the condition Q > R is satisfied, we assume that the see that in the fertile case there is a spread in the duration
remaining gas has been heated and ejected by the accreting of the accretion phase from 10 to 100 Myr, which arises from
DCBH radiative feedback and its growth is quenched. This a combination of differences in the initial DCBH mass and,
sets the final mass of the DCBH, or the mass of the resulting more importantly, in the amount of gas available set by the
IMBH. past history of the host halo. The accretion phase duration
In order to determine the IMF of IMBH we follow distribution peaks at around 40-60 Myr. In the sterile case,
the feedback-regulated growth of DCBH present in 10 re- durations are both longer (70 − 120 Myr, due to the larger
alizations of the merger tree by assuming that the hosting reservoir of gas available for accretion in the halo) and more

c 2012 RAS, MNRAS 000, 1–??
 The IMF of black holes seeds 15
7 SUMMARY AND DISCUSSION
In this paper we have derived for the first time the Initial
Mass Function of Intermediate Mass Black Holes (104−6 M )
formed inside metal-free, UV illuminated atomic cooling
(virial temperature Tvir > 104 K) halos either via direct
collapse followed by GR instability or via an intermedi-
ate Super Massive Star (SMS) stage. These objects have
been recently advocated as the seeds of the supermassive
black holes observed at z ≈ 6. Assembling the SMBH mass
(M• = 2 × 109 M ) deduced for the most distant quasar
ULAS J1120+0641 at z = 7.085 (Mortlock et al. 2011) when
t(z) = 0.77 Gyr, requires a seed mass > 400M . Such value
is uncomfortably large when compared to the most recent
estimates of the mass of first stars, which now converge to-
wards values  100M (Greif et al. 2011; Hosokawa et al.
2012b; Hirano et al. 2014). This is why IMBH seeds, with
their larger masses, are now strongly preferred as the most
promising seeds.
We have obtained the IMBH IMF with a three-step
strategy, as described below.

Figure 12. Probability distribution function of the formation • We have first derived the condition for a proto-SMS to
redshifts DCBH (dotted histogram) and SMS (yellow shaded his- undergo GR instability and directly collapse into a DCBH
togram) for the fertile (upper panel) and sterile (lower) minihalo depending on the gas accretion rate; we found that, for a
cases are shown. The results are averaged over 10 MW merger non-rotating SMS, GR instability kick in when the stellar
histories. The errorbars correspond to ±1σ errors. mass reaches
 2/3
Ṁ?
M? = 8.48 × 105 M . (51)
M yr−1
concentrated (atomic halos have similar gas content, fb ≈ 1, Thus, for an accretion rate of 0.15 M yr−1 (typical of
due to the lack of star formation and supernova feedback in atomic cooling halos) a proto-SMS will collapse into a DCBH
the progenitor minihalos). We have also derived the distri- when its mass reaches ≈ 2.4 × 105 M . A similar expression
bution of the formation epoch of DCBH and SMS, shown has been obtained for rotating SMS, and given by eq. 22.
in Fig. 12. Remarkably, the conditions for the formation of However, the SMS growth can come to an end before the
these black holes seeds last relatively shortly during cosmic star crosses the above critical mass. This occurs if the host
history. In the fiducial fertile minihalos case, the first DCBH halo accretes polluted gas, either brought by minor merg-
and SMS (showing a more gradual abundance rise) appear ers or smooth accretion from the IGM, or suffers a major
in non-negligible numbers at z = 17; however by z = 8 their merger that generates vigorous turbulence, again disrupting
formation is already quenched as a result of the accretion of the smooth and quiet accretion flow onto the central proto-
polluted gas and/or a major merging disrupting the quiet SMS star.
accretion flow and inducing gas fragmentation. If minihalos • We followed these processes in a cosmological context
are sterile, then the termination epoch is delayed by about using the merger tree code GAMETE, which allows us to
2 redshift units, and DCBH become the dominant source of spot metal-free atomic cooling halos in which either a DCBH
production for IMBH seeds. or SMS can form and grow, accounting for their metal en-
We finally comment on the relation between the IMBH richment and major mergers that halt the growth of the
and their host halo mass. This relation is often necessary to proto-SMS by gas fragmentation. We derive the mass dis-
formulate physical seeding prescriptions, e.g. in studies of tribution of black holes at this stage, and dub it the “Birth
SMBH formation based on merger trees or numerical sim- Mass Function” (BMF). Most DCBH host halos (> 80% of
ulations. Our results show that a very reasonable prescrip- the total) have Mh ≈ (2.5 − 6.3) × 107 M . As a result of ac-
tion is to populate a given fraction of halos (a) of mass cretion physics, DCBHs span a very narrow range of masses,
7.5 < log Mh < 8, (b) in the redshift range 8 < z < 17, 2.5 × 105 M ∼ <
M• ∼ <
4.5 × 105 M . We find that the metal
(c) with IMBH in the mass range 4.75 < log M• < 6.25. pollution is by far the dominant process stopping the proto-
This prescription assumes our fiducial case of fertile mini- SMS growth. The resulting SMS are smaller than DCBH, al-
halos. If instead one wishes to consider sterile minihalos, though they span a larger range, MSM S ≈ (3−45)×104 M ,
then the previous prescription turns in the following set: (a) due to the stochastic nature of the merging/accretion pro-
7.25 < log Mh < 7.75, (b) in the redshift range 6 < z < 14, cesses. We can also note that the formation epoch of SMS
(c) with IMBH in the mass range 6 < log M• < 6.25. We is shifted towards lower redshifts with respect to DCBH,
recall once again that the above fraction of such halos to be 8 < z < 14 instead of 8 < z < 17. The previous results refer
populated cannot be obtained from our method as it would to the fiducial (fertile) case in which minihalos (Tvir < 104
require a detailed knowledge of the LW UV background field. K) can form stars and pollute their gas. Results are also
Therefore such information must be fixed from other physi- given and discussed for the sterile case in Sec. 5.
cal considerations or left as a free parameter. • As a third and final step towards the IMBH IMF we

c 2012 RAS, MNRAS 000, 1–??
16 Ferrara et al.
have followed the accretion of the halo gas leftover after the tion are located inside dark matter halos that have lost all of
formation of the DCBH onto the DCBH itself. This is neces- their gas. Some of these systems will be able to re-accrete gas
sary because, contrary to the case of the SMS in which ioniz- and turn it into stars (the raining gas is progressively more
ing radiation from the exposed hot photosphere ionized and likely to be polluted); others might be included in larger ha-
disperses the surrounding gas, the General Relativity (GR) los and their IMBH merge with other black holes. Finally,
instability induces a rapid, direct collapse into a DCBH, i.e. some of them could remain isolated and dead, thus becoming
without passing through a genuine stellar phase. The two virtually undetectable.
cases differ dramatically, as virtually no ionizing photons As a last remark, we stress that during the feedback-
are produced if a DCBH forms. Therefore the newly formed regulated growth we have assumed spherically symmetric
DCBH will find itself embedded in the gas reservoir of the accretion. Although we have given arguments in support of
halo and start accrete again. This accretion phase, similar to this assumption, it is unclear if accretion might go through
the quasi-stellar phase advocated by Begelman et al. (2008) a disk that could become thermally unstable (e.g. because of
(see also Ball et al. (2012)), remains highly obscured and it H2 formation), form stars and SNe, thus stopping the IMBH
is only in the latest phases (several tens of Myr after the growth. We plan to address these issues, which require dedi-
DCBH formation) that the DCBH will be able to clear the cated high.resolution numerical simulation, in a forthcoming
remaining gas when the photospheric temperature starts to study.
climbs from about 5000 K when DCBH mass crosses the On the observational side, our scenario can have impor-
value M• = 105.4 M . Beyond that point the photospheric tant implications. If a prolonged, obscured phase of DCBH
temperature increases rapidly and reaches about 30,000 K growth exists, this might explain the puzzling near-infrared
once the DCBH has grown to M• = 106 M , thus allow- cosmic background fluctuation excess and its recently de-
ing radiative feedback to clear the gas, stop accretion, and tected cross-correlation with the X-ray background (Cap-
determine the final IMBH mass. pelluti et al. 2013), which might imply that an unknown
faint population of high-z black holes could exist (Yue et al.
The IMBH IMF is different in the two scenarios consid- 2013a;Yue et al. 2013b). In addition, hints of a pervasive
ered: in the (fiducial) fertile case it is bimodal with two broad presence of IMBH in the center of nearby dwarf galaxies
peaks at M ≈ (0.7−1.2)×105 M and M ≈ (5−10)×105 M . have been convincingly collected by Reines et al. (2013).
The distribution extends over a wide range of masses, from Thus, our results might be a solid starting point to make
M ≈ (0.5 − 20) × 105 M and the DCBH accretion phase more detailed predictions on these and other related issues,
lasts from 10 to 100 Myr. If minihalos are sterile, the IMF including of course the puzzling presence of supermassive
spans the narrower mass range M ≈ (1 − 2.8) × 106 M con- black hole in the first billion year after the Big Bang.
taining > 90% of the IMBH population (the remaining part
being represented by the SMS low mass tail, see Fig. 10).
We conclude that a good seeding prescription is to populate
halos (a) of mass 7.5 < log(Mh /M ) < 8, (b) in the red- ACKNOWLEDGMENTS
shift range 8 < z < 17, (c) with IMBH in the mass range
AF acknowledges financial support from PRIN MIUR 2010-
4.75 < (log M• /M ) < 6.25.
2011 project, prot. 2010LY5N2T. SS acknowledges sup-
Although the present study combines the physics of
port from Netherlands Organization for Scientific Research
SMS evolution and DCBH formation/growth with a well-
(NWO), VENI grant 639.041.233. DRGS thanks the Ger-
tested cosmological scenario to derive the mass function of
man Science Foundation (DFG) for financial support via the
IMBH seed for the first time, it needs to be improved and
Collaborative Research Center (CRC) 963 on ”Astrophysi-
complemented under many aspects.
cal Flow Instabilities and Turbulence” (project A12) and via
First, we have not attempted to constrain the formation
the Priority Program SPP 1573 ”Physics of the Interstellar
efficiency of IMBH inside putative host halos. This would
Medium” (grant SCHL 1964/1-1).
require the knowledge of the LW radiation field and a solid
•
determination of Jν,c during their formation epoch. Fortu-
nately, given the very narrow mass range of the IMBH host
halos (7.5 < log Mh < 8) the LW intensity can be factorized
REFERENCES
safely.
Second, to follow the feedback regulated growth of Agarwal, B., Khochfar, S., Johnson, J. L., Neistein, E.,
DCBH we have assumed that the total mass of their hosting Dalla Vecchia, C., & Livio, M. 2012, MNRAS, 425, 2854
haloes remain constant during this short accretion phase. Ahn, K., Shapiro, P. R., Iliev, I. T., Mellema, G., & Pen,
Using our merger tree model we found that the average U.-L. 2009, ApJ, 695, 1430
time after which DCBH hosts experience a minor or ma- Ball, W. H., Tout, C. A., & Żytkow, A. N. 2012, MNRAS,
jor merger is respectively equal to ≈ 60 Myr and 90 Myr. 421, 2713
Hence the approximation is good for our fiducial fertile Baumgarte, T. W., & Shapiro, S. L. 1999, ApJ, 526, 941
minihalo case, as on average the accretion phase lasts for Begelman, M. C. 2010, MNRAS, 402, 673
h∆Taccr i ≈ 50 Myr. However, this assumption may affect Begelman, M. C., Rossi, E. M., & Armitage, P. J. 2008,
the IMBH IMF obtained for the sterile minihaloes as in this MNRAS, 387, 1649
case ∆Taccr ≈ 100 Myr. Begelman, M. C., Volonteri, M., & Rees, M. J. 2006, MN-
Third, a very interesting remaining question is the fi- RAS, 370, 289
nal fate of the population of IMBH that do not merge into Bovino, S., Latif, M. A., Grassi, T., & Schleicher, D. R. G.
SMBHs. As we have discussed, the IMBH seeds at forma- 2014, ArXiv e-prints

c 2012 RAS, MNRAS 000, 1–??
 The IMF of black holes seeds 17
Bromm, V., Ferrara, A., Coppi, P. S., & Larson, R. B. Montero, P. J., Janka, H.-T., & Müller, E. 2012, ApJ, 749,
2001a, MNRAS, 328, 969 37
Bromm, V., Kudritzki, R. P., & Loeb, A. 2001b, ApJ, 552, Mortlock, D. J., et al. 2011, Nature, 474, 616
464 Naoz, S., Noter, S., & Barkana, R. 2006, MNRAS, 373, L98
Bromm, V., & Yoshida, N. 2011, ARA&A, 49, 373 Omukai, K., Tsuribe, T., Schneider, R., & Ferrara, A. 2005,
Cappelluti, N., et al. 2013, ApJ, 769, 68 ApJ, 626, 627
Chandrasekhar, S. 1964, ApJ, 140, 417 Pallottini, A., Ferrara, A., Gallerani, S., Salvadori, S., &
Dijkstra, M., Haiman, Z., Mesinger, A., & Wyithe, J. S. B. D’Odorico, V. 2014, ArXiv e-prints
2008, MNRAS, 391, 1961 Park, K., & Ricotti, M. 2012, ApJ, 747, 9
Dopcke, G., Glover, S. C. O., Clark, P. C., & Klessen, R. S. Peimbert, M., Luridiana, V., & Peimbert, A. 2007, ApJ,
2011, ApJ, 729, L3+ 666, 636
Eisenstein, D. J., & Loeb, A. 1995, ApJ, 443, 11 Petri, A., Ferrara, A., & Salvaterra, R. 2012, MNRAS, 422,
Ferrara, A., Haardt, F., & Salvaterra, R. 2013, MNRAS, 1690
434, 2600 Planck Collaboration et al. 2013, ArXiv e-prints
Fialkov, A., Barkana, R., Visbal, E., Tseliakhovich, D., & Prieto, J., Jimenez, R., & Verde, L. 2014, MNRAS, 437,
Hirata, C. M. 2012, ArXiv e-prints 2320
Greif, T. H., Springel, V., White, S. D. M., Glover, S. C. O., Regan, J. A., & Haehnelt, M. G. 2009, MNRAS, 396, 343
Clark, P. C., Smith, R. J., Klessen, R. S., & Bromm, V. Reines, A. E., Greene, J. E., & Geha, M. 2013, ApJ, 775,
2011, ApJ, 737, 75 116
Hirano, S., Hosokawa, T., Yoshida, N., Umeda, H., Omukai, Reisswig, C., Ott, C. D., Abdikamalov, E., Haas, R., Mösta,
K., Chiaki, G., & Yorke, H. W. 2014, ApJ, 781, 60 P., & Schnetter, E. 2013, Physical Review Letters, 111,
Hosokawa, T., Omukai, K., & Yorke, H. W. 2012a, ApJ, 151101
756, 93 Sakashita, S. 1974, Ap&SS, 26, 183
Hosokawa, T., Yorke, H. W., Inayoshi, K., Omukai, K., & Salvadori, S., & Ferrara, A. 2009, MNRAS, 395, L6
Yoshida, N. 2013, ApJ, 778, 178 —. 2012, MNRAS, 421, L29
Hosokawa, T., Yoshida, N., Omukai, K., & Yorke, H. W. Salvadori, S., Schneider, R., & Ferrara, A. 2007, MNRAS,
2012b, ApJ, 760, L37 381, 647
Iben, Jr., I. 1963, ApJ, 138, 1090 Salvadori, S., Tolstoy, E., Ferrara, A., & Zaroubi, S. 2014,
Janka, H.-T. 2002, in Lighthouses of the Universe: The MNRAS, 437, L26
Most Luminous Celestial Objects and Their Use for Cos- Schleicher, D. R. G., Palla, F., Ferrara, A., Galli, D., &
mology, ed. M. Gilfanov, R. Sunyeav, & E. Churazov, 357 Latif, M. 2013, ArXiv e-prints
Jeon, M., Pawlik, A. H., Greif, T. H., Glover, S. C. O., Schneider, R., Ferrara, A., Natarajan, P., & Omukai, K.
Bromm, V., Milosavljević, M., & Klessen, R. S. 2012, ApJ, 2002, ApJ, 571, 30
754, 34 Schneider, R., Omukai, K., Inoue, A. K., & Ferrara, A.
Johnson, J. L., Dalla, V. C., & Khochfar, S. 2013a, MN- 2006, MNRAS, 369, 1437
RAS, 428, 1857 Schwarzschild, M. 1958, Structure and Evolution of the
Johnson, J. L., Whalen, D. J., Li, H., & Holz, D. E. 2012, Stars (Dover Pubs.)
ArXiv e-prints Shang, C., Bryan, G. L., & Haiman, Z. 2010, MNRAS, 402,
—. 2013b, ApJ, 771, 116 1249
Latif, M. A., Schleicher, D. R. G., & Schmidt, W. 2014, Shapiro, S. L., & Shibata, M. 2002, ApJ, 577, 904
MNRAS, 440, 1551 Shchekinov, Y. A., & Vasiliev, E. O. 2006, MNRAS, 368,
Latif, M. A., Schleicher, D. R. G., Schmidt, W., & 454
Niemeyer, J. 2013a, MNRAS, 433, 1607 Tamazawa, S., Toyama, K., Kaneko, N., & Ono, Y. 1975,
Latif, M. A., Schleicher, D. R. G., Schmidt, W., & Ap&SS, 32, 403
Niemeyer, J. C. 2013b, MNRAS, 436, 2989 Tanaka, T., Perna, R., & Haiman, Z. 2012, MNRAS, 425,
Lodato, G., & Natarajan, P. 2006, MNRAS, 371, 1813 2974
Loeb, A., Ferrara, A., & Ellis, R. S. 2008, Cosmological Tornatore, L., Ferrara, A., & Schneider, R. 2007, MNRAS,
Feedbacks from the First Stars, 161–258 382, 945
Loeb, A., & Rasio, F. A. 1994, ApJ, 432, 52 Trenti, M., & Stiavelli, M. 2009, ApJ, 694, 879
Machacek, M. E., Bryan, G. L., & Abel, T. 2001, ApJ, 548, Valiante, R., Schneider, R., Maiolino, R., Salvadori, S., &
509 Bianchi, S. 2012, MNRAS, 427, L60
Madau, P., Ferrara, A., & Rees, M. J. 2001, ApJ, 555, 92 Volonteri, M., & Bellovary, J. 2012, Reports on Progress in
Maiolino, R., et al. 2012, MNRAS, 425, L66 Physics, 75, 124901
Maselli, A., Ferrara, A., & Ciardi, B. 2003, MNRAS, 345, Xu, H., Wise, J. H., & Norman, M. L. 2013, ApJ, 773, 83
379 Yue, B., Ferrara, A., Salvaterra, R., & Chen, X. 2013a,
Mayer, L., Kazantzidis, S., Escala, A., & Callegari, S. 2010, MNRAS, 431, 383
Nature, 466, 1082 Yue, B., Ferrara, A., Salvaterra, R., Xu, Y., & Chen, X.
Mayer, M., & Duschl, W. J. 2005, MNRAS, 358, 614 2013b, MNRAS
Miralda-Escudé, J., Haehnelt, M., & Rees, M. J. 2000, ApJ,
530, 1
Mo, H., van den Bosch, F., & White, S. 2010, Galaxy For-
mation and Evolution (Cambridge)

c 2012 RAS, MNRAS 000, 1–??

View publication stats