Skip to main content

Full text of "Quantum cosmology of the brane universe"

See other formats


Quantum cosmology of the brane universe. 



A.Boyarsky^, A.Neronov^, I.Tkachev'^ 
^ Ecole Polytechnique Federate de Lausanne, BSP 720, CH-1015, Lausanne, Switzerland 
^ ISDC,CH-1290 Versoix Switzerland , Switzerland, 
^ Theory department, CERN, 1211 Geneve 23, Switzerland 

We canonically quantize the dynamics of the brane universe embedded into the five-dimensional 
Schwarzschild-anti-deSitter bulk space-time. We show that in the brane-world settings the formu- 
lation of the quantum cosmology, including the problem of initial conditions, is conceptually more 
simple than in the 3-|-l-dimensional case. The Wheeler-deWitt equation is a finite-difference equa- 
tion. It is exactly solvable in the case of a fiat universe and we find the ground state of the system. 
The closed brane universe can be created as a result of decay of the bulk black hole. 



Introduction. Quantum effects almost certainly played 
crucial role in the early universe evolution and in the 
process of universe creation. Understanding and study 
of quantum cosmology is important not only from the 
conceptual point of view, but, hopefully, may provide 
us with constraints on possible topology of the universe 
and initial conditions for the inflationary stage P, 0, • 
Appropriate theoretical frameworks which would incor- 
porate all quantum gravitational effects are yet to be con- 
structed, however. 

String theory, eventually, may provide the consistent 
approach to the quantum cosmology realm, but the for- 
mulation of the string theory on a non-trivial and signifi- 
cantly Lorentzian space time is very complicated and un- 
solved task ( see for example 0| and references therein). 
That is why the approaches based on canonical quanti- 
zation of the Einstein gravity still prove to be more 
successful in addressing the problems of quantum cos- 
mology. Here one has to adopt a modest approach and 
restrict consideration to quantum phenomena below the 
Plank energy scale. Quantizing the universe as a whole 
one has further resort to the "mini-superspace" model- 
ing 1, 2, 3,'6] in order to get to definite final results (for 
a recent interesting development see, however, Ref. Q 
where effective action for the scale factor was derived in- 
tegrating out other gravitational degrees of freedom using 
numerical simulations). 

Even then, within the "mini-superspace" approach, 
many conceptual and technical problems remain, such as 
the problem of ascribing physical meaning to the wave 
function of the universe [g . Other important issues are 
the choice of the boundary conditions which one imposes 
at the big-bang point (e.g. "no-boundary" jQj], "tunnel- 
ing" 2], etc.) and the problem of unboundedness of the 
gravitational action (see e.g. Q)- 

In the present paper we pursue the viewpoint that the 
presence of extra dimensions can resolve or relax some 
of these problems. Indeed, in the brane world scenario 

llOj l- the problem of quantum cosmology (i.e. quan- 
tization of gravitational degrees of freedom) is replaced 
by a much better defined problem of quantum mechanics 
of the brane (matter degrees of freedom) which moves 
in the bulk space-time. This has several important con- 
sequences. First, one may hope that probabilistic inter- 



pretation, initial and boundary conditions, "tunneling", 
" scattering" and "ground" states of the Universe become 
better defined. Second, one can escape, to some extent, 
solving the problems of quantum gravity. Indeed, the 
big bang point, i.e. the point of vanishing brane size, 
can be unreachable due to quantum uncertainty. Thus, 
quantization of matter in a self-consistently calculated 
"external" gravitational field can be sufficient. 

The conceptual simplicity of the brane quantum cos- 
mology does not imply its "technical" simplicity: one 
has to take into account self-consistently the interaction 
of the brane with the bulk both on classical and quan- 
tum levels. Here we can benefit capitalizing on the fact, 
that the dynamics of (3-|-l)-dimensional brane embed- 
ded in (4-|-l)-dimensional bulk, is very similar to the dy- 
namics of self- gravitating shells in conventional (3-1-1)- 
dimensional General Relativity, which was studied ex- 
tensivel y b oth at classical [ill and quantum levels 
[l^ 113 . TTsll. In the present paper we generahze the for- 
malism developed in j3| to the case of the (3-1-1) dimen- 
sional brane universe embedded into (4-f l)-dimensional 
bulk. 

We may hope that some results found in frameworks of 
brane quantum cosmology may hold even if the universe 
is (3+l)-dimensional. In particular, the distinctive fea- 
ture of quantum mechanics of branes is that the differen- 
tial Schrocdinger (or "Wheeler-deWitt" ) equation for the 
wave function is replaced by a finite-difference equation 
[TsL fisjl . This may be a general property of "true" quan- 
tum cosmology. Note in this respect that finite-difference 
equations for the wave function of the universe appear 
also in the frameworks of loop-quantum gravity [ifil . 

Hamiltonian description of the classical motion of a 
gravitating brane. We construct the Hamiltonian for- 
malism which describes the motion of a self-gravitating 
thin shell of matter starting from the action of (4-1-1)- 
dimensional Einstein gravity with bulk cosmological con- 
stant. The brane part of the action contains the term 
proportional to the brane tension A and the term which 
describes (in the simplest case) dust-like matter on the 
brane with the mass fi per unit co-moving volume. The 
total action of the system is 



S 



1 

4/3" 



bulk 



[a 7^ -t- (Tr/C)^ - Tr/C^ 



2 



-Svr/i 



dr-X 



branc 



—gdrdx 



(1) 



branc 



where Ip^ is the (4+l)-dimensional Planck mass, g is the 
induced metric on the brane, r is the proper time of co- 
moving observers in the brane universe, A is the bulk cos- 
mological constant and ^^^7?,, /Cyis are the 4-dimensional 
Ricci scalar and the external curvature of the spatial sec- 
tion of (4-|-l)-dimensional space-time. We restrict our- 
selves to the case of homogeneous and isotropic brane 
which may describe open, flat, or closed brane universe. 

For a generally-covariant systems the Hamiltonian dy- 
namics is encoded in a system of constraints 5]. For a 
spherically symmetric space without matter, and in any 
space-time dimensions, these constraints can be solved 
explicitly classically as well as quantum mechanically, see 
Ref. 01 . This result can be understood noticing that in 
this case gravity has only global degrees of freedom. The 
most convenient way to parameterize these global degrees 
of freedom is to use the Schwarzschild-like representation 
of the metric 



ds^ = -F{t,r)dT'^ 



dR^ 
F(t,r) 



R^dnl , 



(2) 



where T = T{t,r) and R{t,r) are arbitrary functions 
of time and radial coordinates {t,r), while the function 
F{t, r) has the form 



F{t,r) 



i?2 



KR' 



(3) 



where k = 0,±1 for the cases of flat, closed and open 
spatial sections, respectively. 

In the Hamiltonian formalism the canonical vari- 
ables describing the bulk gravitational field are 
{R,M]Pr,Pm). It turns out that T' = dT/dr is the 
momentum conjugate to M [T^ . The conventional con- 
straints of canonical formalism reduce to the set of equa- 
tions, Pr = and M' = dM/dr = 0. One can see that 
if M = const, the metric ^ coincides with the metric of 
five-dimensional Schwarzschild-ant-deSitter black hole of 
mass A/. 

The canonical constraint on the brane is 



F\ cosh ■ 



(M + Ai?=')=0, (4) 



where hat denotes the values of corresponding variables 
on the brane, e.g. R = R{t, r)!^^^^^ and a = ±1. For the 
geometrical meaning of the sign function a see Ref. [llj . 
At the classical level a is integral of the motion, but the 
change of sign is possible at the quantum level O . 
Note that the Hamiltonian constraint Eq. I^J does not 
describe the most general case (e.g. the Schwarzschild 
parameter M can be different on both sides of the brane 
in general situation); rather, the Z2 symmetry was as- 
sumed following Ref. To]. Positive (negative) sign of a 
corresponds to the positive (negative) brane tension in 



the case of classical regime of Randall-Sundrum cosmol- 
ogy. For the discussion of general brane Hamiltonian in 
the quantum case see Ref. [Tj] . 

The equation of motion for R found from the Hamil- 
tonian (gl) is dk/dr = o-i/lTT'sinh (^3ZpiPr/7?2^ ^ which, 
upon substitution into gives 



{dR/drf k _ Z|,i(^ + Ai?3)2 ;3^^^ 



i?2 



i?2 



i?4 



A 



(5) 



Being written in this form, the equation of motion of the 
brane resembles closely the Friedmann equation 0|, in 
which the density of matter on the brane pm = f^/R^ 
enters quadratically at small i?, the presence of non-zero 
bulk black hole mass M results in the effective "dark 
radiation" contribution p^r = M/R^ and the effective 
cosmological constant on the brane is a certain combi- 
nation of the bulk cosmological constant and the brane 
tension A(3+i) = l^^X'^/9 + A. Note however that Eq. © 
is a "square" of true dynamical equation and important 
information encoded in a is lost. Therefore its use can be 
inappropriate in some situations, especially in the quan- 
tum regime. 

Quantum dynamics of the brane universe. In canoni- 
cally quantized theory the Hamiltonian constraint is 
replaced by an operator equation on the wave function of 
the universe, H^! = 0. However, the quantization proce- 
dure in the coordinate representation would result in the 
differential equation of infinite order. In addition, the 
definition of operator 



cosh 



ll^pR/m^ =cosh -i{llJ?,R^)d/dR) 



suffers from ambiguity related to the operator ordering. 

These problems can be solved if one makes canonical 
transformation v = R^; Py = P^/(3i?2) ^ which brings 
the Hamiltonian H into the form 



H 



ay \F\ cosh {lliPy} - i^i + Xv) 



(6) 



In the new variables, after quantization P-^ —id/dv, 
the hyperbolic cosine which enters H becomes an 
operator of finite shift along the imaginary axis, 
exp (^ll^Py) {v) = *(ti - i^li). Substituting this into 
H'i> = we find the following finite-difference equa- 
tion which determines the quantum dynamics of a self- 
gravitating brane universe 



v^/'F'^^^{v + iPp^) + ^iv-iPp^)} 
2 



0. 



(7) 



Since the shift of the argument of the wave function is 
along imaginary axis, one has to consider the above equa- 
tion in the complex plane, or, more precisely, on the 
corresponding Riemanian surface. Indeed, the function 
_F^/2 is a branching fimction on the complex plane. The 



3 



two branches, F^/"^ = ±^/F correspond to the two pos- 
sible choices of sigma. Therefore, if one finds the solu- 
tions of the above equation on the Riemann surface, the 
wave function 4" is defined simultaneously in = +1 and 
(T = — 1 domains. 

In order to understand qualitatively the behavior of 
solutions of Eq. (0 we start with an analysis of the 
distances much larger than Ipi. In this limit we can 
expand ± ilp^) in powers of the shift parameter, 
^{v±ill,i) ~ ^{v)±ill,i'i/'{v)-^ll.i^"{v) + . . . In the first 
non-trivial order Eq. Q reduces to (we restrict ourselves 
to the case a = 1 here) 



Eq. |7J takes the form 



(8) 



which is a Schroedinger-like equation for particle motion 
in a potential 



U = 1 



+ Xv) 



3 (fcw4/3 _ 2G'Afw2/3 + |A|w2 



,1/2 



(9) 



For large v — W the potential approaches a constant, 
U l-Z|,jA/(3v/|A|). If ;|iA > SyiAl the wave function 
behaves in the limit of large as a flat wave which 
describes an expanding or contracting universe. 

Exactly solvable case of the flat universe. To make 
more detailed analysis of the quantum mechanics of the 
brane, e.g. to study its spectrum, one needs to impose 
boundary conditions at the origin. At first sight the issue 
of boundary conditions at the Big Bang point v = looks 
conceptually more simple for the brane universe. Indeed, 
since the scale factor of the universe is now just a po- 
sition of the brane moving in the external space (rather 
than purely gravitational degree of freedom), this is just 
the question of boundary conditions on the wave func- 
tion at the origin of spherical coordinates . However, in 
the region v ~ Zpj one can not expand Eq. ((T)) in powers 
of ^pi and the intuition based on Eq. ((HJ is not appli- 
cable anymore. Instead, one has to deal with the exact 
finite-difference equation Q . (We assume that the mini- 
superspace model based on the thin-wall approximation 
is still valid in the limit of small v.) 

The finite-difference equations, and in particular the 
Eq. Q, possess a number of interesting general prop- 
erties. Being understood as an infinite-order differential 
equations, they have to be supplemented with an infinite 
set of boundary conditions. At the same time, starting 
from a single particular solution 'So one can generate an 
infinite set of solutions simply by multiplying (z) by a 
function C{z) which is periodic with respect to the finite 
shift parameter (i.e. C{z + ilp^) = C{z) in the case of 
Eq. O). The appropriate methods of anaj^is of finite 
difference equations are discussed in Refs. El, 11101 ■ 

In order to illustrate these methods it is convenient to 
consider the special case when Eq. Q is exactly solvable, 
namely the case of the flat universe k — and zero bulk 
Schwarzschild mass M = 0. For this choise of parameters 



*(^;+4i)+*(^;-4,)~-^ (a +9 *H-0 (10) 



which coincides with the finite-difference analog of 
quantum-mechanical problem of motion in Coulomb po- 
tential . A general solution of Eq. (fTUIl is given by 
(up to multiplication by an arbitrary ilp^ - periodic func- 
tion) ^(5) = ve-'^^'J^il - w,l - /3 : 2 : 1 - e^^^"), 
where is the hypergeometric function. Parameters a 
and P are defined by relations cos a — Zp[A/(3y^|A|) and 
/3sina = Z3^/z/(3v1^). 

As it is usual in quantum mechanics, the single solution 
can be selected only when the proper set of boundary con- 
ditions is chosen. The correct boundary conditions can 
be determined from the requirement of vanishing of the 
probability fiow J = i(\I'tif* - -^H^i'f) at w = 0. In the 
case of Eq. H10|l this reduces to the set of conditions |23| 
^f(2")(0) = 0, n = 0, 1,.... 

Similarly to the conventional quantum mechanics with 
the Coulomb potential, there are bound states and con- 
tinuous spectrum. Using the above boundary conditions 
as well as appropriate conditions at infinity, one can see 
that bound states exist when the quantization condition 



A 



(3+1) 



9 



is satisfied. It relates the effective brane cosmological 
constant A(3+2) to the matter density on the brane. In 
particular, the ground state of the universe corresponds 
to n = 1. The wave functions of continuous spectrum 
(ZpjA > 3\/]A|, which corresponds to the positive effec- 
tive cosmological constant on the brane) contain both 
the collapsing (in-going wave) and expanding (out-going) 
branes. Thus, in the case of continuous spectrum, the 
wave function of the universe corresponds to the so-called 
"big bounce" situation. One can consider also transitions 
between the bound states and the states from contin- 
uous spectrum (e.g. an expanding brane universe can 
result from the excitation of the ground state). How- 
ever, the analysis of perturbations of the spherically sym- 
metric system considered above goes beyond the mini- 
superspace approximation. 

Tunneling from the bound states. In order to study 
qualitatively the more general cases when the bulk 
Schwarzschild mass in not zero let us come back to the 
analysis of the truncated equation ((SJ. The behavior of 
the potential U for the cases fc 7^ and/or M 7^ is 
shown in Fig. One can see that if ZpjM > (A(3_|_i))~i 
there is a potential barrier, which separates the regions 
of bound and unbound motion of the brane. This means 
that the spectrum of quantum states of the brane can 
contain, apart from the discrete and continuous part, also 
"resonances" . In this case the expanding brane universe 
is the result of decay (or "tunneling") of an almost sta- 
ble state localized near the origin. The main difference 
of the tunneling states considered here from the (3-1-1)- 
dimensional ones is that in the (3 -I- l)-dimensional case 



4 



0.4 











k=l,M=0 






k=i,]vi;=o 















0.5 1 1.5 2 2.5 3 



R 

FIG. 1: The potential U lO for difFerent choices of parame- 
ters. The potential is singular at the gravitational radius of 
the bulk Schawrzscild-anti-deSitter black hole when M 7^ 0. 

the choice of the boundary conditions at i? = is am- 
biguous and the existence of the "tunneUng" state is, in 
fact, just postulated 0. 

Discussion. In this paper we have constructed quan- 
tum cosmology of the brane universe and have shown 
that it has several distinctive features. In particular, one 
can avoid the conceptual problems related to the inter- 
pretation of the wave function of the universe. Indeed, in 
the brane-world setup one does not quantize pure grav- 
ity, but rather deals with quantum mechanics of a matter 



source (brane) moving through higher-dimensional space- 
time. The problem of the choice of boundary conditions 
on the wave function of the universe is also free from am- 
biguities: one simply has to impose the usual quantum 
mechanical conditions on the wave function at the origin 
of coordinates. This allows for the detailed analysis of 
bound states, continuous spectrum and tunneling states, 
where creation of the universe from "nothing" can be 
interpreted as a decay of a bound state resonance. 

When gravitational self-interaction of the brane uni- 
verse is important, as, for example in the setup of 
Randall-Sundrum cosmology studied here, one has to cor- 
rectly account for the bulk-brane interaction not only 
classically, but also on quantum level. As a result, the 
classical brane Hamiltonian constraint becomes after 
quantization a finite-difference equation Q. 

Although the appearance of finite-difference equations 
is a novel feature of the quantum brane cosmology, the 
analysis of the boundary conditions and of the wave func- 
tions of discrete and continuous spectra can be carried in 
a way similar to the one used in conventional quantum 
mechanics. From the point of view of quantization of 
gravitating systems the appearance of a non-local equa- 
tion (with non-locality at the Plank scale) is natural to 
expect. Such equations appear in several other models 
(see e.g. HBIS^IMii)- It implies a deformation of the 
Lorentz symmetry and generalized uncertainty principle 

mill. 



[1] Y. B. Zeldovich and A. A. Starobinsky, Sov. Astron. Lett. 
10, 135 (1984); A. D. Linde, Sov. Phys. JETP 60, 211 
(1984) [Zh. Eksp. Teor. Fiz. 87, 369 (1984)]; A. Vilenkin, 
Phys. Rev. D 30, 509 (1984). V. A. Rubakov, Phys. Lett. 
B 148, 280 (1984). 

[2] A. Vilenkin, Phys. Lett. B 117, 25 (1982) and Phys. Rev. 
D 27, 2848 (1983). 

[3] A. Linde, JCAP 0410, 004 (2004). 

[4] A. Giveon, E. Rabinovici and A. Sever, Fortsch. Phys. 

51, 805 (2003). 
[5] B. S. Dewitt, Phys. Rev. 160, 1113 (1967). 
[6] J. B. Hartle and S. W. Hawking, Phys. Rev. D 28, 2960 

(1983). 

[7] J. Ambjorn, J. Jurkiewicz and R. LoU, hep-th/0411152| 
[8] A. Vilenkin, gr-qc/020461. 

[9] V. A. Rubakov and M. E. Shaposhnikov, Phys. Lett. B 
125, 136 (1983). 
[10] L. RandaU and R. Sundrum, Phys. Rev. Lett. 83, 4690 
(1999). 

[11] V. A. Berezin, V. A. Kuzmin and I. I. Tkachev, Phys. 
Lett. B 120, 91 (1983) and Phys. Rev. D 36, 2919 (1987). 

[12] A. Vilenkin, Phys. Lett. B 133, 177 (1983); J. Ipser and 
P. Sikivie, Phys. Rev. D 30, 712 (1984); S. K. Blau, 
E. I. Guendelman and A. H. Guth, Phys. Rev. D 35, 
1747 (1987); A. Aurilia, R. S. Kissack, R. Mann and 
E. Spallucci, Phys. Rev. D 35, 2961 (1987). 

[13] V. A. Berezin, N. G. Kozimirov, V. A. Kuzmin and 
I. I. Tkachev, Phys. Lett. B 212, 415 (1988). 



[14] W. Fischler, D. Morgan and J. Polchinski, Phys. Rev. D 
42, 4042 (1990). 

[15] V. A. Berezin, A. M. Boyarsky and A. Y. Neronov, Phys. 
Rev. D 57, 1118 (1998) and Phys. Lett. B 455, 109 
(1999); A. Y. Neronov, Phys. Rev. D 59, 044023 (1999). 

[16] M. Bojowald, Phys. Rev. Lett. 86, 5227 (2001); Phys. 
Rev. Lett. 87, 121301 (2001) and Phys. Rev. Lett. 89, 
261301 (2002). 

[17] K. V. Kuchar, Phys. Rev. D 50, 3961 (1994). 

[18] J. M. Cline, C. Grojean and G. Servant, Phys. Rev. Lett. 
83, 4245 (1999); P. Binetruy, C. Deffayet and D. Lan- 
glois, Nucl. Phys. B 565, 269 (2000); P. Brax, C. van 
de Bruck and A. C. Davis, Rept. Prog. Phys. 67, 2183 
(2004); R. Maartens, Phys. Rev. D 62, 084023 (2000); 
A. Neronov, gr-qc/0101060 A. Neronov, JHEP 0111, 
007 (2001). 

[19] V. G. Kadyshevsky, R. M. Mir-Kasimov and M. Freeman, 
Yad. Fiz. 9, 646 (1969); V. A. Berezin, Phys. Rev. D 55, 
2139 (1997). 

[20] P. Hajicek, Commun. Math. Phys. 150, 545 (1992). 
[21] S. N. M. Ruijsenaars and H. Schneider, Annals Phys. 

170, 370 (1986). 
[22] H. S. Snyder, Phys. Rev. 71, 38 (1947). 
[23] A. Gorsky and N. Nekrasov, Nucl. Phys. B 436, 582 

(1995). 

[24] G. 't Hooft, Class. Quant. Grav. 13, 1023 (1996). 
[25] T. Jacobson, S. Liberati and D. Mattingly, 
gr-qc/0404067 



5 

[26] G. Amelino-Camelia, gr-qc/0309054j