Instability of scalar perturbation in a phantomic cosmological scenario
aa r X i v : . [ g r- q c ] D ec INSTABILITY OF SCALARPERTURBATIONS IN A PHANTOMICCOSMOLOGICAL SCENARIO
J.C. Fabris Institut d’Astrophysique de Paris98bis, Boulevard Arago, 75014 Paris, FranceD.F. Jardim and S.V.B. Gon¸calves Departamento de F´ısica, Universidade Federal do Esp´ırito Santo, 29060-900,Vit´oria, Esp´ırito Santo, Brazil
Abstract
Scalar perturbations can grow during a phantomic cosmological phaseas the big rip is approached, in spite of the high accelerated expansionregime, if the equation of state is such that pρ = α < − . It is shownthat such result is independent of the spatial curvature. The perturbedequations are exactly solved for any value of the curvature parameter k and of the equation of state parameter α . Growing modes are foundasymptotically under the condition α < − . Since the Hubble radiusdecreases in a phantom universe, such result indicates that a phantomscenario may not survive longtime due to gravitational instability. PACS: 98.80.-k, 95.36.+xIn the ordinary theory of cosmological perturbation [1, 2, 3] there are twofundamental regimes defined from the notion of
Jean’s length , λ J . Perturbationswhose scales are such that λ > λ J suffer gravitational instability since thegravitational attraction dominates over the pressure reaction to contraction; onthe other hand, perturbations whose scales satisfy the condition λ < λ J , tends tooscillates due to the effectiveness of the pressure opposition to the gravitationalcollapse. In the relativistic version of the theory of cosmological perturbation wemust add a new relevant scale, given by the Hubble radius. The Hubble radiusdefine, in some sense, the effective causal region, and consequently the regionwhere the effects of the microphysics phenomenon (responsible for the pressure)may play a relevant rˆole in the process of gravitational collapse. Perturbationswhose scales are much large than the Hubble radius tends to become frozen,while for those smaller than the Hubble radius, the pressure tends to producea damping effect. Moreover, if the spatial section of the metric representingthe universe is not flat a new scale appears which is connected to the curvatureparameter. e-mail: [email protected]. On leave of absence of Departamento de F´ısica, UniversidadeFederal do Esp´ırito Santo, 29060-900, Vit´oria, Esp´ırito Santo, Brazil e-mail: [email protected] e-mail: [email protected]
1n cosmology, concomitant with the interplay between the gravitational at-traction and the pressure resistance, there is a supplementary relevant effectdue to the expansion of the universe, which acts as a friction term in the funda-mental equations of the perturbed quantities: the expansion leads to a dampingin the evolution of perturbations. In general, for very large perturbations, thatare connected with the large structures existing in the universe, there are twomodes: a growing or a constant mode and a decreasing mode. However, if thepressure is negative, and the expansion becomes accelerated, gravity may be-come repulsive and we generally find only decreasing modes: the perturbationsare always damped and structures can not be formed.There are quite strong evidences that the universe today is in an acceler-ated expansion phase [4]. If this is true, the material content of the universemust be dominated by an exotic fluid whose pressure is negative. There aremany candidates for this exotic fluid, like cosmological constant, quintessence,K-essence, Chaplygin gas, etc., each of them having its advantages and disad-vantages from the theoretical point of view. We address the reader to recentreviews on the many dark energy models existing in the literature [5, 6]. Suchnegative pressure fluid may generate repulsive effects driving the acceleratedexpansion. There are some claims that the observational data favors a phan-tom fluid, that is, a fluid whose negative pressure violates the dominant energycondition p + ρ ≥ big rip . This is, of course, a very undesirable feature.But it has been shown [8] that in a single fluid approximation, for a spatially flatuniverse, the scalar perturbations can grow when the scales of the perturbationare greater than the Hubble radius. Since, the Hubble radius decreases, fora phantom dominated universe, the isotropy and homogenous condition wouldnot be satisfied anymore as the big rip is approached, leading perhaps to theavoidance of this future singularity, leaving a very inhomogeneous universe. Thissituation may occur if the pressure is negative enough in order to satisfy thecondition pρ = α < − .In the present work we extend those result showing that phenomenon ofenhancing of the inhomogeneities in large scales is independent of the spatialcurvature, and that the critical point α = − is present in any class of homo-geneous and isotropic universe. In order to do so, we will solve the perturbedequations for scalar modes for any value of k and α . An asymptotic analysiswill reveal the existence of critical behavior associated to α = − .For a universe dominated by a fluid whose equation of state is given by2 = αρ , the relevant equation is a ′ a + k = 8 πG ρ a , ρ = ρ a − α ) . (1)The primes indicate derivations with respect to the conformal time η , and k isthe spatial curvature parameter, k = ± ,
0. The solution for this equation maybe written in a unified form as a ( η ) = a (cid:20) √ k sin (cid:18) α √ kη (cid:19)(cid:21) α . (2)For k = 1, this solution represents a universe that begins and end at a singularityat a = 0 for 1 ≤ α < − , while for α < − it represents a bouncing universe.On the other hand, for k = − ,
0, it represents an ever expanding universe forany value of α , with the following characteristics: for 1 ≥ α > − , the expansionimplies 0 ≤ η < ∞ corresponding, in terms of the cosmic time t , to 0 ≤ t < ∞ ;for − > α , the expansion implies −∞ < η ≤
0, corresponding to 0 ≤ t < ∞ if − > α ≥ − −∞ < t ≤ − > α . This last feature leads to the notionof big rip.To study the evolution of scalar perturbations, we use the gauge invariantformalism. For a perfect fluid the evolution of the scalar perturbation is givenby a single equation for the gravitational potential Φ [3, 10]:Φ ′′ + 3(1 + α ) H Φ ′ + (cid:26) α n + 2 H ′ + (1 + 3 α )( H − k ) (cid:27) Φ = 0 , (3)where H = a ′ a and n is the eigenvalue of the Laplacian operator ∇ Φ = − n Φ.The perturbed equation can be recast under the following form(1 − z ) z Φ ′′ + 7 + 9 α α ) (1 − z )Φ ′ + ˜ n Φ = 0 , (4)where z = 1 + cos( √ kθ )2 , ˜ n = 4 k (1 + 3 α ) (cid:20) αn − α ) k (cid:21) , θ = 1 + 3 α η . (5)The solution of (4) can be represented under the form of hypergeometricfunctions. It reads in general, for any value of k and α , as follows:Φ n ( η ) = c F (cid:20) A + , A − ; B ; z (cid:21) +¯ c z − B F (cid:20) A + − B + 1 , A − − B + 1; 2 − B ; z (cid:21) , (6)where A ± = 12 (cid:26) α α ± s
36 (1 + α ) (1 + 3 α ) + 4˜ n (cid:27) , B = 7 + 9 α α ) , (7)3nd c, ¯ c are constants.The asymptotic analysis for the flat case k = 0 was made in reference [8].In this case, the hypergeometric’s function reduces to Bessel’s functions. It hasbeen shown that one of the two modes remains constant, in the long wavelengthlimit, for any value of α . In the same limit, however, the other mode decreaseswhen α > − , but grows with time when α < − . For α = − , both modes areconstants. In the case k = 0, such analysis, in terms of large or small wavelengthlimit, is more involved since, contrarily to the flat case, we have a scale given bythe Hubble radius and another scale given by the curvature. It becomes easier,in this sense, and for our purpose more relevant, to consider the behavior in theextremes of the time interval. In order to perform this analysis, we must takeinto account some convenient transformation properties of the hypergeometricfunctions. In special, the following transformations will be useful (see reference[11]): F ( A, B ; C ; z ) = Γ( C )Γ( C − A − B )Γ( C − A )Γ( C − B ) F ( A, B ; A + B − C + 1; 1 − z ) +(1 − z ) C − A − B Γ( C )Γ( A + B − C )Γ( A )Γ( B ) F ( C − A, C − B ; C − A − B + 1; 1 − z ) ;(8) F ( A, B ; C ; z ) = Γ( C )Γ( B − A )Γ( B )Γ( C − A ) ( − A z − A F (cid:18) A, A + 1 − C ; A + 1 − B ; 1 z (cid:19) +Γ( C )Γ( A − B )Γ( A )Γ( C − B ) ( − B z − B F (cid:18) B, B + 1 − C ; B + 1 − A ; 1 z (cid:19) . (9)Hence, we have the following asymptotic behaviors: z → ⇒ F ( A, B ; C ; z ) ∼ z − C , (10) z → ⇒ F ( A, B ; C ; z ) ∼ Γ( C )Γ( C − A − B )Γ( C − A )Γ( C − B ) (1 − z ) C − A − B + Γ( C )Γ( A + B − C )Γ( A )Γ( B ) , (11) z → ∞ ⇒ F ( A, B ; C ; z ) ∼ Γ( C )Γ( B − A )Γ( B )Γ( C − B ) ( − A z − B + Γ( C )Γ( A − B )Γ( A )Γ( C − B ) ( − B z − A . (12)Using these expressions, we can determine the behavior of the perturbationsin the two different extremities of the time interval, for each value of k . • k = 1. In this case the conformal time interval is 0 ≤ η ≤ π α for α > − and π α ≤ η ≤ α < − . Using the asymptotic expressionswritten above, we find the following behaviors: for α > − , there isinitially two decreasing modes and, as the universe approaches the bigcrunch at η = π α , there are a constant mode and a growing mode; for − > α > − there is initially, during the contraction phase, a growing4ode and a constant mode, and as the scale factor diverges in the otherasymptotic, there is a constant mode and a decreasing mode; for α = − ,both modes are constant at the beginning of the contraction phase and atthe end of the expansion phase; for α < − , there is a constant mode anda decreasing mode at the universe begins to contract, and there are twoincreasing modes as the universe approaches the big rip. • k = −
1. The range of the conformal time is 0 ≤ η < ∞ for α < − and −∞ < η ≤ α < − . The open case is more involved because, inopposition to the closed universe, the asymptotic behavior of the modesdepends on the scale of the perturbation. Let us consider the situationwhere the eigenvalues of the Laplacian operator is null. Hence, when α > − there is initially and in future infinity two decreasing modes.For − < α < − there is initially two decreasing modes, but in thefuture infinity there is a constant mode besides a decreasing mode; thecase α = − differs from the preceding one by the fact that in futureinfinity both modes are constant. Finally, when α < − , both modesare initially decreasing but they become growing modes as the big rip isapproached.When α = − , in all cases, the same features observed in the flat universe arereproduced here, since for this particular equation of state the matter densityscales as the curvature parameter in the Friedmann’s equation.The main conclusion of the previous analysis is that a phantom cosmologi-cal scenario is highly unstable against scalar perturbations under the conditionof an isotropic and homogeneous background universe if the pressure is nega-tive enough, that is pρ < − : the scalar perturbations grows as the big rip isapproached. As in the flat case, the Hubble radius shrinks with time in thephantomic case. This means that the large scale approximation becomes es-sentially valid asymptotically for all perturbation scales in the phantomic case.The analysis was made using a perfect fluid material content. However, at largescales, it is expected that a more fundamental representation, using for example,scalar fields, must give the same results as the perfect fluid case [12].A curious feature of the above results concerns the critical point α = − .As far as we know, it does not correspond to any energy condition (contrarily to α = − and α = − α = − [14]: it is not a critical point for the background.Moreover, in the perturbed equation (3) there is no explicit special structurefor that particular value of the parameter α . The nature of this critical pointmust still be cleared up. Acknowledgements:
We thank J´erˆome Martin for his criticism and sugges-5ions. We thank also CNPq (Brazil) and the French-Brazilian scientific coopera-tion CAPES/COFECUB (project number 506/05) for partial financial support.
References [1] S. Weinberg,
Gravitation and cosmology , Wiley, New York (1972).[2] T. Padmanabhan,
Structure formation in the universe , CambridgeUniversity Press, Cambridge (1993).[3] R. Brandenberger, H. Feldman and V. Mukhanov, Phys. Rep. , 203(1992).[4] D.N. Spergel et al, Astrophys. J. Suppl. , 377 (2007).[5] V. Sahni, Lect. Notes Phys. , 141 (2004).[6] R. D¨urrer and R. Maartens,
Dark energy and dark gravity , arXiv:0711.0077.[7] R. Caldwell, Phys. Lett.
B545 , 23 (2002).[8] J.C. Fabris and S.V.B. Gon¸calves, Phys. Rev.
D74 , 027301 (2006).[9] F. Piazza and S. Tsujikawa, JCAP , 004 (2007).[10] V. Mukhanov,
Physical foundations of cosmology , Cambridge Univer-sity Press, Cambridge (2005).[11] I.S. Gradshteyn and I.M. Rhyzik,