0% found this document useful (0 votes)
9 views5 pages

kolb1994

The paper investigates large-amplitude isothermal fluctuations in dark matter density using a spherical collapse model, finding that such fluctuations can lead to the formation of dense dark matter clumps. The final density of these clumps is expressed as a function of the fluctuation amplitude, with implications for dark matter detection and astrophysical phenomena such as Bose star formation. The study emphasizes the importance of understanding the nonlinear evolution of these fluctuations for both theoretical and observational cosmology.

Uploaded by

Chia-Feng Chang
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
9 views5 pages

kolb1994

The paper investigates large-amplitude isothermal fluctuations in dark matter density using a spherical collapse model, finding that such fluctuations can lead to the formation of dense dark matter clumps. The final density of these clumps is expressed as a function of the fluctuation amplitude, with implications for dark matter detection and astrophysical phenomena such as Bose star formation. The study emphasizes the importance of understanding the nonlinear evolution of these fluctuations for both theoretical and observational cosmology.

Uploaded by

Chia-Feng Chang
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 5

PHYSICAL REVIEW D VOLUME 50, NUMBER 2 15 JULY 1994

Large-amplitude isothermal fluctuations and high-density dark-matter clumps


Edward W. Kolb*
NASA/Fermilab Astrophysics Center, Fermi National Accelerator Laboratory, Batavia, Illinois 60510
and Department of Astronomy and Astrophysics, Enrico Fermi Institute, The University of Chicago, Chicago, Illinois 6063 7

Igor I. ~ k a c h e v ~
NASA/Fermilab Astrophysics Center, Fermi National Accelerator Laboratory, Batavia Illinois 60.510
and Institute for Nuclear Research of the Academy of Sciences of Russia, Moscow 11 7312, Russia
(Received 9 March 1994)
Large-amplitude isothermal fluctuations in the dark-matter energy density, parametrized by
@=8pDM/pDM, are studied within the framework of a spherical collapse model. For k 1, a fluctuation
collapses in the radiation-dominated epoch and produces a dense dark-matter object. The final density
of the virialized object is found to be pF= 140a3(@+l)peq,where p,, is the matter density at equal
matter and radiation energy density. This expression is valid for the entire range of possible values of a ,
both for @ >> 1 and << 1. Some astrophysical consequences of high-density dark-matter clumps are
discussed.
PACS nurnberk):98.80.Cq, 05.30.Jp, 95.35.+d, 98.70.- f

I. INTRODUCTION In this paper we are interested in isocurvature fluctua-


tions that enter the horizon before the temperature of
In almost all modern cosmological models, galaxies, equal energy densities of matter and radiation,
clusters, and all large-scale structures develop through T,,=s. 5n,h2 eV [2]. We will consider scales much
the gravitational instability of small-amplitude, seed den- smaller than the horizon, so the radiation energy density
sity fluctuations. In most of these models cold dark should be homogeneous.
matter is an important constituent of the total mass den- It is well known [3] that the growth of small-amplitude
sity of the Universe. There are two basic types of seed isothermal fluctuations is suppressed by the cosmological
density fluctuations, curvature and isocurvature,' and in expansion, the fluctuations do not grow until after the
general, both are expected to be produced in the early equality epoch. However, this is true only in linear
Universe. By definition, the total energy density in an theory. The self-gravity of large-amplitude, nonlinear
isocurvature fluctuation is constant; the fluctuation is in fluctuations may become important before Teq, and con-
the relative contribution to the total energy density of sequently they collapse earlier. Therefore they are cap-
different components in a multicomponent system. Im- able of producing very dense objects after they separate
portant examples of this type are the fluctuations induced out from the general expansion and virialize.
in the baryons by some dissipative process in a universe We refer to these isothermal fluctuations as "clumps."2
containing both baryons and dark matter, and topologi- Let us specify the density of a dark-matter clump as
cal or nontopological field configurations such as cosmic
strings or textures. While the amplitude of either type of
fluctuation on large scales is strongly restricted by mi-
crowave background anisotropy constraints, the ampli- where CJis not necessarily small. For example, "typical"
tude of small-scale fluctuations can be large, even non- axion miniclusters [4] will have @- 1. In Ref. [S] it was
linear, at the epoch of last scattering. The spectrum of found that accounting for nonlinear effects in the evolu-
small-scale fluctuations do not necessarily have to reflect tion of axions at the crucial epoch when the axion mass
the shape of the power spectrum of the primordial fluc- switches on can lead to considerably larger density in
tuations generated at the inflationary epoch, since the many miniclusters, with @ in the range 1 to lo4. Dark-
small-scale fluctuations may well be generated later, e.g., matter clumps seeded by wakes induced by cosmic strings
during various cosmological phase transitions. or by textures also will have @- 1 [6].Seeded clumps are
particularly interesting in the case of weakly interacting
massive particle (WIMP) dark matter.

*Electronicaddress: rocky @ fnas0 1.fnal.gov


TElectronic address: tkachev @ fnas 13.fnal.gov
'The division of fluctuations into curvature and isocurvature is 2Since the clumps may be very dense compared to the back-
strictly true only outside the Hubble radius [I]. ground, we do not refer to them as "perturbations."

0556-282 1/94/50(2)/769(5)/$06.00 50
- 769 @ 1994 The American Physical Society
770 EDWARD W. KOLB A N D IGOR I. TKACHEV --50

It was pointed out in Ref. [ 5 ] that the final virialized nate, d q z d t / a ( t ) , and then to rewrite this equation of
density in a clump has to scale as p~ -@4peq. Because of motion in the comoving reference frame, r =a ( r] )R <( 71 )<,
the dependence upon the fourth power of (D, even a small where is the comoving label of a given shell and R t ( ~ l
increase in @ is very important. For the same reason, the measures the deviation of the shell motion from the uni-
final density can be sensitive to the details of the evolu- form Hubble flow of the background Friedmann
tion of the clump in the radiation-dominated era. To our universe In what follows we shall omit the subscript ;
knowledge, a detailed study of the nonlinear evolution of on R ( r ] ) ,but it should be understood that there is a
large-amplitude isothermal fluctuations has never been separate evolution for each shell.
performed. However, it is very important in various phe- We shall assume that the scale factor a ( r ] i satisfies the
nomenological implications including both direct and Einstein equations for an no= 1 universe filled with radi-
nondirect dark-matter searches. In this paper we consider ation and pressureless matter:
this problem.
The clumpiness of the dark matter has important Im-
plications for attempts to detect dark matter. Clearly the
signal in direct detection experiments for dark matter is where a prime denotes d / d q . We parametrize the radia-
proportional to the dark-matter density. For the rare tion and matter energy densities as p R =p,,(aeq/a l4 and
direct encounter with a clump, there could be a huge p,,, =peq(aeq/a 13. The solution to the background equa-
amplification of the signal. However, if the clumpiness is tions, Eqs. (2.21, is
too high, the flux of unclumped dark matter will be too
small for a reasonable detection rate. The rate of WIMP
annihilation contributing to the y-ray background [7,8] is
proportional to the density as well. In the case of
where r]* = 2vGpeqa /3. ,?,
The equation of motion [Eq. (2.111 in terms of confor-
clumped dark matter, there will be stronger constraints
mal time is
on dark matter from indirect searches. In very dense ax-
ion clumps, Bose star formation becomes possible [9]
(clumps with @ - 30 already satisfy the critical condition
for this [ 5 ] ) ,which in turn can lead to the formation of
radio sources [lo]. Another possible manifestation of The radiation energy density does not enter this equation
high-density clumps is the phenomenon of microlensing. explicitly, but its effect is encoded in the evolution of the
To study the structure and evolution of high-density scale factor. We also parametrize the total mass of
clumps, a full three-dimensional numerical simulation is matter inside the shell in terms of the excess over the
needed. However, for an isolated clump some relevant homogeneous background, denoted as @( 6) = 6pM/pM.
physical information can be extracted from a one- The total mass within the region is
dimensional spherical model. The spherical model
proved useful in studies of the gravitational nonlinear
evolution in the epoch of matter domination when it is
possible to find exact analytic solutions [ l l ] . In the Changing from 7 to x = a /ae, as the independent vari-
present paper we generalize this model to include radia- able, we finally obtain
tion. Although there are no analytic solutions, the result
turns out to be very simple: The final density in a virial-
ized clump is pF = 140@~( +(D 1)p,, in the whole range of
possible values of (D, both for (D >> 1 and for @ << 1.

11. A SPHERICAL MODEL


This equation reduces to the Meszaros equation [3] in
Let us consider a spherical region of radius r contain- the limit of small deviation of the shell motion from the
ing an overdensity of pressureless matter in an expanding general cosmological expansion, R 7= 1 - S and 6 << 1, if
universe. In a spatially flat universe, every overdense re- we assume no excess in total mass of the matter, i.e..
gion eventually reaches some maximum size and recol- @=O:
lapses. The total mass of matter in the region inside r,
M,,,, is an integral of the motion so long as the region ex-
pands. Since we will consider scales much smaller than
the Hubble radius, we can consider the radiation to be
The latter is hypergeometric equation, and its growing
homogeneous, with its time evolution determined by the
general expansion of the universe, and not by the local
+
mode is 6 =So( 1 3x /2 ), implying the well known result
that the growth of small fluctuations is significant only
conditions.
after the Universe becomes matter dominated.
The equation of motion for the radius of the region is
We have solved Eq. (2.6) numerically, assuming
R (x, ) = 1 at some early time, x o << 1. Note that at small
x, the second derivative in Eq. (2.6)can be neglected, and
the solution with initial conditions fixed at x,=O is
It is convenient to change to the conformal time coordi- R = ( 1 - 3(Dx /2 = 1 - (Dx /2, where the expansion is
LARGEAMPLITUDE ISOTHERMAL FLUCTUATIONS AND ... 771

FIG. 1. Numerical solutions to Eq.(2.6)for several different FIG. 3. The coefficient C, in Eq. (2.9) as a function of cP.
values of cP. The second-order fit is shown by the dotted line. The dashed line is the prediction of the standard spherical mod-
el (matter without radiation).

justified since the solution is valid only at small x. Actu- defined by i. =O, in coordinates of Eq. (2.6) is the solution
ally, this is the separatrix, i.e., independently of the initial to the equation R +xdR /dx =O. Up to order 0 ( x 2 )the
value of R', all solutions tend to it (provided xo << 1). function R only depends upon the product a x , so to this
The results of numerical integration of Eq. (2.6) proved order in x the scale factor at turnaround will be given by
to be insensitive to R 1 ( x 0 )already at xo@< For xta=const/@ and to the same order in x, R,,=const.
several different values of @ they are shown in Fig. 1. The matter density of a fluctuation at turnaround is
It is possible to find an analytic approximation to R ( x ) '
pta= ( 1/4vr ) d/dr, ~ with r =R ,,xta,$, so it is appropri-
as a power series in x to any given order. To third order ate to represent the parameters at turnaround as
it is

We expect only a weak dependence of Cx and Cpupon @,


since higher order corrections are small. Results of a nu-
The first three terms in this decomposition provide a merical integration for Cx and Cp are shown in Fig. 2
good practical fit to the solution. This fit is shown in Fig. and Fig. 3, respectively. These figures demonstrate that
1 by the dotted line. The last term in Eq. (2.8) shows that for practical applications we can consider C, and Cp to
the solution is not simply a function of the product @x. be constants in entire range of possible values of @, both
Our main goal is to find the parameters of the fluctua- @>>I and@<<l.
tion, i.e., its radius and density, at the moment when the Let us compare our results to predictions of a standard
fluctuation turns around. For later times the assumption spherical model, which is valid for the matter-dominated
of spherical symmetry breaks down; however, we can as- epoch, i.e., at small @. The spherical model predicts
sume that the radius of a virialized gravitationally bound pta/p, =9.rr2/16 for the density contrast at turnaround,
object will be one-half of the turnaround radius, and the where p, is the background density [ l l ] . Using Fig. 2 we
density inside the object will be eight times larger than can extract the corresponding values of p, and then cal-
the density at turnaround [ll]. The turnaround time, culate the standard spherical model prediction for C,.
This is presented in Fig. 3 by the dashed line. Both
dashed and solid lines coincide in the limit of small @, as
they must.

111. APPLICATIONS

With the function @(,$) in the general form we have as-


sumed, the effective radius after virialization for each
shell of a given label { will be half the turnaround radius
of the shell.3 Using Eq. (2.9) we obtain for the final densi-
ty profile in a virialized object

3~trictlyspeaking this is not always true, but for the practical


applications we will consider below the approximation should
FIG. 2. The coefficient C, in Eq. (2.9)as a function of cP. be reasonable.
772 EDWARD W. KOLB A N D IGOR I. TKACHEV 50
--

would be small, this would give a reasonable encounter


rate, and the question of formation and survival of small-
scale clumps within the galaxy is worth further study.
-
where we have set C, 17. For the core density, this for- Another astrophysical outcome of very dense axion
+
mula gives p F = 1 4 0 @ ~ ( @ 1~)peq,where Qo=@(0 ). The clumps can be the possibility of "Bose star" formation in
numerical value of the density at equality is axion miniclusters. The Bose-Einstein relaxation time in
pe,=3 x 10-16(flh 2 ) 4 cm-'. NOWlet us turn to a few the minicluster due to axion self-interaction is smaller
specific examples. than the present age of the Universe with @ 2 30 [5,9].
A. Axion miniclusters B. Accretion by a point mass
Fluctuations in the density of axions can be very high, The density profile in the halo accreted by a previously
possibly spanning the range 1 5 @ 5 lo4 [5,12]. Even with formed clump can be calculated in the approximation of
@ as small as 1, the density in miniclusters which form secondary infall onto an excess point mass of mass m. In
out of these fluctuations can be as much as 10'' times this case @( 6)= m /M, where M =473-peqa 2,g3 is the mass
larger than the local galactic halo density [4]. Using Eq. of the background dark matter within the shell with the
(3.1), we obtain pF-9X 10-I4n4h8 g c m - 3 for @ = I . label c . Substituting this into Eq. (3.1) we find
With @ >> 1 this result must be multiplied by ~ ~ / 2 .
The typical mass of an axion minicluster corresponds
to the total mass in axions within the horizon at-^- 1
GeV when the inverse axion mass is equal to the Hubble This can be translated into p F as a function of r since M
length: M,, - ~ O - ~ M ~ .present probability of a
The has to be understood as the mass of dark matter residing
within r. The result is pF a r -9/4 , the same power law
direct encounter with a minicluster is small; the en-
counter rate is 1 per lo7 years with @ = I . Although the one usually obtains for secondary infall in the matter-
signal in an axion detector [13] from a close encounter dominated era [14]. However, Eq. (3.4) is valid regardless
would be enormous, it might of the time when collapse actually occurs.
- be a long- wait with a weak
signal between encounters if a major fraction of the ax-
C. Cosmic strings
ions are part of miniclusters.
There should be some miniclusters with @ in the range We can apply Eq. (3.4) for clumps of dark matter seed-
1 05@51 These collapse during the matter- ed by loops of cosmic strings so long as the peculiar ve-
dominated epoch and have larger radii than those with locity of a loop is sufficiently small. Since the string loop
@ 2 1 which collapsed in the radiation-dominated epoch, is not a point object, this formula breaks shown in the re-
so the probability of an encounter with a clump with gion of small M. Namely, when the given shell turns
@ << 1 can be larger. From the point of view of direct around at x,, =O. 7M/m, the loop size I, has to be small-
searches, even miniclusters with density contrast of order er than the physical radius of the shell, rta=xtaR,,{aeq,
two times the average with respect to the galactic halo for Eq. (3.4) to be valid. This gives the restriction
density are important. Such miniclusters form just prior M 2 3m 3/2,,,-3/4 1 / 4 , ~
to the moment of galaxy formation and started with peq - C, where m =PI,. We can con-
sider Mc as the mass of the core region. The correspond-
@- lo-'. For @ < 1 the expected time between en-
ing maximumdensity which can be achieved in the core
counters is given in terms of the number density of
clumps, n, the geometric cross section of the clump is p c - 1 5 2 / p 3 P e q / ~ C .This value of the core density
-
u R2 and the virial velocity v as could be many orders of magnitude larger than the densi-
ty at equality. However, as we see from Eq. (3.41, p~ is
much greater than pea only in the case when the mass of
the string loop is larger than the mass of the accreted
dark matter, Consequently, with gradual loop decay due
where pH is the halo density, and Rc,um,/u is the time the to emission of gravitational radiation, the dark-matter
Earth spends inside the minicluster. The factor of @ - ' clump will adiabatically expand and diminish in density.
appears because the number density of miniclusters with This process of clump expansion will continue until
@ << 1 is suppressed in our model [12]. A minicluster m 5 M. Since in the gravitational field the product rm is
with @ << 1 would require the initial misalignment angle an adiabatic invariant for each dark-matter particle,
8 (which is uniformally distributed in the range 0 to 2 7 ~ ) where r is the effective radius of the orbit, we conclude
to be finely tuned to the mean misalignment angle to an that in any clump for which M < m initially, the present
accuracy 68/8--@/2. Using Eq. (2.9),we finally obtain -
density will have the same order of magnitude, p 1 0 ~ ~ , , .
Dark-matter clumps seeded by wakes induced by long
segments of moving cosmic strings or by textures also
Note that the miniclusters discussed so far appear if will have @ - 1 (see [6]), and correspondingly the same
the axion field is uncorrelated on scales larger than the virialized density, p-102peq. While this density is
-
Hubble radius at T 1 GeV. However, miniclusters with sufficiently high to be significant in applications such as
@ << 1 can appear from primordial density fluctuations annihilation of dark-matter particles, it is too small to
generated by inflation without the suppression factor of cause microlensing, as we show below.
@-I. If this is the case, then T - @ ~ x lo7 yr. Since @ Cold dark-matter accretion onto string loops both in
50
- LARGE-AMPLITUDE ISOTHERMAL FLUCTUATIONS AND .. . 773

the matter- and radiation-dominated eras was also con- the observed upper limit to the y-ray flux in the direction
sidered in Ref. [15], however, only in the linearized limit of the galactic center [7], I y = 4 X ~m-~sec-'.
of Eq. (2.6),i.e., Eq. (2.7).
E. Gravitational microlensing
D. WIMP annihilation Two conditions must be satisfied for the clump to cause
gravitational microlensing [17]. First, the mass of the
In the case where the dark-matter particle species is a
clump has to be in a range near 0. lMo. Second, the
stable weakly interacting massive particle (WIMP) such
as a very massive neutrino or a supersymmetric particle physical radius of the clump has to be smaller than the
Einstein ring radius, R E =2- where d is the

-
(photino, Higgsino, or scalar neutrino), the WIMP's can
annihilate, contributing to the y-ray flux. This places effective distance to the lens (typically d -20 kpc). The
second condition r e s . s the density of the minicluster
severe constraints on the dark-matter density near the
center of the galaxy [7,8]. Clumped dark-matter annihi- ,,
to be p k l ~ ~ p , , / l / ~ -where M - , M/O. lMo. If
lation is even more efficient, and places a very strong lim- the lensing object is a clump of noninteracting cold dark
it on the clumpiness as a function of the WIMP proper- matter, it has to be formed from a density fluctuation
ties [6]. with @ !:20.
For the y-ray flux on Earth from WIMP annihilation Dark-matter clumps seeded by string loops or textures,
which were considered in Ref. [6], are in the appropriate
in the clump, we can write
-
mass range; however, they have @ 1. Axion miniclus-
ters can have @ k 20; however, they are too light. While
it is possible to invent models where both conditions are
met for some of the clumps (one example could be an ax-
where r g is the distance from the Earth to the clump ion model with an extremely small, but nonzero, value for
( r o = 8.5 kpc is the distance to the center of the Galaxy), the u-quark mass), it is hardly likely that a substantial
m, is the particle mass, and M is the mass of the clump. amount of the dark matter has evolved into clumps cap-
Since the particles are nonrelativistic, both in the clumps able of lensing. On the other hand, anticipating
and at the epoch of cosmological freeze-out of the significant numbers of microlensing events (for the first
WIMP's, the thermal average of the cross section in Eq. positive reports see Ref. [18]) in the future, it is not ex-
(3.5) is directly related to the cosmological abundance cluded that some of them could be caused by the clumps
[16]: in such classes of models (especially if collisional relaxa-
tion is significant). The corresponding light curve will be
different from the massive compact halo object
(MACHO)event since clumps are extended objects.
If we consider a large region of (possibly) clumpy dark When our paper was almost completed we became
matter, such as the galactic core or spheroid, we must aware of the paper Ref. [19] where Eq. (2.6) was studied
sum up the fluxes from each individual clump. As a re- in detail, however, again for the case which corresponds
sult, instead of M in Eq. (3.5) we have to substitute lMtot, to @ < < I .
where 6 is the mass fraction of all clumps to the total
mass Mtot in the region. For example, using p- 1 0 ~ ~ , , ACKNOWLEDGMENTS
we obtain, for the central spheroid (M,,, = 1 0 8 ~), o It is a pleasure to thank S. Colombi, A. Stebbins, and
R. Caldwell for useful discussions. This work was sup-
ported in part by the DOE and NASA Grant No.
where m2, m x /20 GeV. This has to be compared to NAGW-238 1 at Fermilab.

[I] W. H. Press and E. T. Vishniac, Astrophys. J. 239, 1 41 (1987).


(1980). [ l l ] P. J. E. Peebles, The Large-Scale Structure of the Universe
[2] E. W. Kolb and M. S. Turner, The Early Universe (Princeton University Press, Princeton, NJ, 1980).
(Addison-Wesley, Redwood City, CA, 1990). [12] E. Kolb and I. I. Tkachev, Phys. Rev. D 49, 5040 (1994).
[3] P. Meszaros, Astron. Astrophys. 37,225 (1974). [13] P. Sikivie, Phys. Rev. Lett. 15, 1415 (1983).
[4] C. J. Hogan and M. J. Rees, Phys. Lett. B 205, 228 (1988). [14] J. E. Gunn, Astrophys. J. 218, 592 (1977).
[5] E. Kolb and I. I. Tkachev, Phys. Rev. Lett. 71, 3051 [15] A. Stebbins, Astrophys. J. Lett. 303, L21 (1986).
(1993). [16] B. Lee and S. Weinberg, Phys. Rev. Lett. 39, 165 (1977).
[6] J. Silk and A. Stebbins, Astrophys. J. 411,439 (1993). [17] B. Paczynski, Astrophys. J. 304, 1 (1986); K. Griest, ibid.
[7] J. Silk and H. Bloemen, Astrophys. J. Lett. 313, L47 366,412 (1991).
(1987). [18] EROS Collaboration, E. Aubourg et al., Nature (London)
[8] J. R. Ipser and P. Sikivie, Phys. Rev. D 35, 3695 (1987). 365, 623 (1993);MACHO Collaboration, C. Alcock et al.,
[9] I. I. Tkachev, Phys. Lett. B 261, 289 (1991). ibid. 365, 621 (1993).
[lo] I. I. Tkachev, Pis'ma Astron. Zh. 12, 726 (1986) [Sov. As- [19] T. Padmanaban and K. Subramanian, Astrophys. J. 417,3
tron. Lett. 12, 305 (1986)l;I. I. Tkachev, Phys. Lett. B 191, (1993).

You might also like