The post-Minkowskian limit of f(R)-gravity Salvatore Capozziello⋄,Arturo Stabile♮, Antonio Troisi‡ ⋄ Dipartimento di Scienze Fisiche, Universita` di Napoli ”Federico II” and INFN sez. di Napoli Compl. Univ. di Monte S. Angelo, Edificio G, Via Cinthia, I-80126 - Napoli, Italy ♮ Dipartimento di Ingegneria, Universita’ del Sannio and INFN sez. di Napoli, gruppo collegato di Salerno, C-so Garibaldi, I - 80125 Benevento, Italy and ‡ Dipartimento di Ingegneria Meccanica, Universita’ di Salerno, Via Ponte Don Melillo, I - 84084 Fisciano (SA), Italy (Dated: January 7, 2010) Weformallydiscussthepost-Minkowskianlimitoff(R)-gravitywithoutadoptingconformaltrans- formations but developing all the calculations in the original Jordan frame. It is shown that such an approach gives rise, in general, together with the standard massless graviton, to massive scalar modeswhosemassesaredirectlyrelatedtotheanalyticparametersofthetheory. Inthissense,the 0 presence of massless gravitons only is a peculiar feature of General Relativity. This fact is never 1 stressed enough and could havedramatic consequences in detection of gravitational waves. Finally 0 theroleofcurvaturestress-energytensoroff(R)-gravityisdiscussedshowingthatitgeneralizesthe 2 so called Landau-Lifshitz tensor of General Relativity. The further degrees of freedom, giving rise tothe massive modes, are directly related to thestructureof such a tensor. n a J PACSnumbers: 04.50.Kd,04.25.Nx,04.30.-w Keywords: alternativetheoriesofgravity,weakfieldlimit,gravitationalwaves 6 ] c I. INTRODUCTION q - r Astrophysical observations of the last decade suggests the introduction of new ingredients in order to achieve a g self-consistentpicture of cosmos. In particular,the observationthat Hubble flow is currently experiencing a speeding [ up has completely changedthe approachto standardcosmologyinducing to take into accounttheoreticalapproaches 1 more general than the standard lore of General Relativity (GR). v The simplest explanation of such a cosmic acceleration requires to include the cosmological constant in the 7 Friedmann-Robertson-Walker cosmology (Concordance Model). This ingredient gives rise to a negative pressure 4 contribution needed to balance the standard matter attractive interaction. Although the Concordance Model repre- 8 sents the best fit model with respect to all samples of data coming from supernovae,large-scalestructure and cosmic 0 . microwave radiation [1], several conceptual problems come out to theoretically define and to give an explanation to 1 the observedvalue of the cosmologicalconstant . Furthermore, assuming the existence of both dark energy and dark 0 matter, weshould findout new fundamentalingredients capableofgiving accountto almost95%ofthe totalamount 0 of cosmic matter-energy. 1 : Due to these difficulties, people have considered alternative approaches to GR that could be able to frame the v observed late time acceleration and missing matter without introducing new ingredients. In this sense, higher order i X gravity [2] and, in particular, fourth order gravity represent an interesting scheme which could, potentially, address the problems. r a Up to now, these theories have been investigated both at cosmological scale and in the weak field limit with significant results [3–6]. It has been shown that an accelerating late time behaviour can be easily recovered [7] and, in addition, it can be coherently related to an early time inflationary expansion [8]. Furthermore, such an approach seems to deserve attention even at smaller scales. In fact, modifying the gravity action in favor of a non-linear Lagrangian in the Ricci scalar implies, in the Newtonian limit, corrections to the gravitational potential which can induce an astrophysical phenomenology interesting at galactic scales. In particular one can fit the rotation curves of spiral galaxies and the haloes of galactic clusters without the introduction of dark matter [9]. Besides, several of these extended models evade Solar System tests so they are not in conflict with positive experimental results of GR [10, 11]. A relevant aspect of higher order gravity theories is that, in the post-Minkowskian limit (i.e. small fields and no prescriptions on the propagation velocity), the propagation of the gravitational fields turns out to be characterized by waves with both tensorial and scalar modes [12, 13]. This issue represents a striking difference between GR and extended gravity since, in the standard Einstein scheme, only tensorial degrees of freedom are allowed. As matter of facts, the gravitationalwaves can represent a fundamental tool to discriminate between GR and alternative gravities [15, 16]. In this paper, we want to develop, formally, the post-Minkowkian limit of analytic f(R)-gravity models which, in our opinion, has never been pursued with accuracy stressing enough some peculiar points. As shown by the same 2 authors for the Newtonian limit, we will show that it is different from the same limit of GR since massive modes naturally come out in the gravitational radiation. This occurrence has a deep meaning since points out that the presence of massless modes only is nothing else but the particular case of GR while massive and ghost modes are present in general [13]. The layout of the paper is the following. In Sec. II, we discuss the post-Minkowskian limit of f(R)-gravity. Considerations on gravitational wave massive modes are developed in Sec.III. Sec.IV is devoted to the discussion of the role of the stress-energy tensor in f(R)-gravity. Concluding remarks are drawn in Sec. V. II. THE POST-MINKOWSKIAN LIMIT OF f(R) - GRAVITY Any theory of gravity has to be discussed in the weak field limit approximation. This ”prescription” is needed to test if the given theory is consistent with the well-established Newtonian Theory and with the Special Relativity as soon as the the gravitationalfield is weak or is almost null. Both requirements are fulfilled by GR and then they can be considered two possible paradigms to confront a given theory, at least in the weak field limit, with GR itself. In [17, 18], the Newtonian limit of f(R)-gravity is investigated always remaining in the Jordan frame [14]. From our point ofview, this is importantsince, by perturbatively approximatinga field, some conformalfeatures could be lost. Here we want to derive, formally, the post-Minkowskian limit of f(R)-gravity. Thepost-Minkowskianlimitofanytheoryofgravityariseswhentheregimeofsmallfieldisconsideredwithoutany prescription on the propagation of the field. This case has to be clearly distinguished with respect to the Newtonian limitwhich,differently,requiresboththesmallvelocityandtheweakfieldapproximations. Often,inliterature,sucha distinctionisnotclearlyremarkedandseveralcasesofpathologicalanalysiscanbeaccounted. Thepost-Minkowskian limit ofGR givesrise to masslessgravitationalwaves. An analogousstudy canbe pursuedconsidering,insteadof the Hilbert-Einstein Lagrangian linear in the Ricci scalar R, a general function f(R). The only assumption that we are going to do is that f(R) is an analytic function. The gravitationalaction is then = d4x√ g f(R)+ , (1) m A Z − (cid:20) XL (cid:21) 16πG where = is the coupling, is the standardmatter Lagrangianandg is the determinantofthe metric. The X c4 Lm field equations, in metric formalism, read1 1 f′R fg f′ +g (cid:3) f′ = XT (2) µν − 2 µν − ;µν µν g 2 µν 3(cid:3)f′+f′R 2f = XT , (3) − 2 2 δ(√ g ) df(R) with T = − − Lm the energy momentum tensorof matter (T is the trace), f′ = and (cid:3) = ;σ. We µν √ g δgµν dR g ;σ adopt a (+, ,− , ) signature, while the conventions for Ricci’s tensor is R = Rσ and Rα = Γα +... for − − − µν µσν βµν βν,µ the Riemann tensor, where 1 Γµ = gµσ(g +g g ) (4) αβ 2 ασ,β βσ,α− αβ,σ aretheChristoffelsymbolsoftheg metric. Actually,inordertoperformapost-Minkowskianlimitoffieldequations, µν one has to perturb Eqs. (2) on the Minkowski backgroundη . In such a case the invariant metric element becomes µν ds2 =g dxσdxτ =(η +h )dxσdxτ (5) στ στ στ withh small(O(h)2 1). We assume thatthe f(R)-Lagrangianis analytic(i.e. Taylorexpandable)intermofthe µν ≪ Ricci scalar, which means that 1 Allconsiderations aredeveloped hereinmetricformalism. 3 fn(R ) 1 f(R)= 0 (R R )n f +f′R+ f′′R2+.... (6) n! − 0 ≃ 0 0 2 0 Xn The flat-Minkowski backgroundis recoveredfor R=R 0. 0 ≃ Field equations (2), at the first order of approximation in term of the perturbation [19], become: R(1) f′ R(1) η f′′ R(1) η (cid:3)R(1) = XT(0) (7) 0(cid:20) µν − 2 µν(cid:21)− 0(cid:20) ,µν − µν (cid:21) 2 µν df d2f where f′ = , f′′ = and (cid:3) = ,σ that is now the standard d’Alembert operator of flat space-time. 0 dR(cid:12)R=0 0 dR2(cid:12)R=0 ,σ Fromthe zero-o(cid:12)rderofEqs.(2), o(cid:12)negetsf(0)=0,while T isfixedatzero-orderinEq.(7)since,inthis perturbation (cid:12) (cid:12) µν scheme, the first order on Minkowski space has to be connected with the zero order of the standard matter energy momentumtensor2. TheexplicitexpressionsoftheRiccitensorandscalar,atthefirstorderinthemetricperturbation, read R(1) =hσ 1(cid:3)h 1h µν (µ,ν)σ− 2 µν − 2 ,µν  (8)  R(1) =h ,στ (cid:3)h στ −  f′′ with h=hσ . Eqs. (7) can be written in a more suitable form by introducing the constant ξ = 0 , that is σ −f′ 0 1 1 1 hσ (cid:3)h h (h ,στ (cid:3)h)η +ξ(∂2 η (cid:3))(h ,στ (cid:3)h)= X T(0). (9) (µ,ν)σ− 2 µν − 2 ,µν − 2 στ − µν µν − µν στ − 2f′ µν 0 Bychoosingthetransformationh˜ =h hη andthegaugeconditionh˜µν =0,oneobtainsthatfieldequations µν µν−2 µν ,µ and the trace equation, respectively, read 3 (cid:3)h˜ +ξ(η (cid:3) ∂2 )(cid:3)h˜ = XT(0)  µν µν − µν −f0′ µν . (10) (cid:3)h˜+3ξ(cid:3)2˜h= XT(0) −f0′  InordertoderivetheanalyticsolutionsofEqs. (10),wecanadoptamomentum-description. Thisapproachsimplifies the equations and allows to fix the physical properties of the problem. In such a scheme, we have: k2h˜ (k)+ξ(k k k2η )k2h˜(k)= XT(0)(k)  µν µ ν − µν f0′ µν (11) k2h˜(k)(1 3ξk2)= XT(0)(k) − f0′  where h˜ (k)= d4x h˜ (x) e−ikx µν (2π)2 µν  R (12) T(0)(k)= d4x T(0)(x) e−ikx µν (2π)2 µν  R 2 Thisformalismdescends fromthetheoretical settingofNewtonianmechanics whichrequirestheappropriateschemeofapproximation when obtained from a more general relativistic theory. This scheme coincides with a gravity theory analyzed at the first order of perturbationinthecurvedspacetimemetric. 3 Thegauge transformationish′µν =hµν−ζµ,ν−ζν,µ whenweperformacoordinate transformationasx′µ =xµ+ζµ withO(ζ2)≪1. To obtain the gauge and the validity of the field equations for both perturbation hµν and ˜hµν, the ζµ have to satisfy the harmonic condition(cid:3)ζµ=0. 4 are the Fourier transforms of the perturbation h˜ (x) and of the matter tensor T(0). We have defined, as usual, µν µν kx = ωt k x and k2 = ω2 k2; h˜(k) and T(0)(k) are the traces of ˜h (k) and T(0)(k). In the momentum space, µν µν − · − one can easily recognize the solutions of Eqs.(11); h˜ (k) turns out to be µν T(0)(k) ξ k2η k k h˜ (k)= X µν + X µν − µ νT(0)(k), (13) µν f′ k2 f′ k2(1 3ξk2) 0 0 − which fulfils the condition h˜µν = 0 (that is h˜µν(k) k = 0). The perturbation variable h (k) can be obtained by ,µ µ µν inverting the relation with the tilded variables. In particular, by inserting the new stress-energy tensor S(0)(k) = µν T(0)(k) 1η T(0)(k) with the trace S(0)(k)=ηµνS(0)(k), one obtains: µν − 2 µν µν S(0)(k) ξ k2η +2k k h (k)= X µν + X µν µ νS(0)(k), (14) µν f′ k2 2f′ k2(1 3ξk2) 0 0 − which represents a wave-like solution, in the momentum space, with a massless and a massive contributions. The massive term is due to the pole in the denominator of the second term: the mass is directly related with the physical properties of the pole itself and, thanks to the parameter ξ, depends on the analytic form of the model (i.e. f′ and 0 f′′). The wavelike solution in the configuration space is obtained by the inverse Fourier transform of h (k). 0 µν III. MASSIVE MODES IN GRAVITATIONAL WAVES Thepresenceofthemassivetermisafeatureemergingfromtheintrinsicnon-linearityoff(R)-gravity. Specifically, itisrelatedtothefactthatf′′ =0,whichiszeroinGRwheref(R)=R. Thismeansthatmasslessstatesarenothing 0 6 else but a particular case among the gravitational theories. A similar situation emerges also in the Newtonian limit: the Newton potential is recovered only as the weak field limit of GR. In general, Yukawa-like corrections, and then characteristic interaction lengths, are present [17] Some considerations are in order at this point. It is worth noticing that field Eqs. (2) can be written putting in evidence the Einstein tensor in the l.h.s. [3]. In such a case, higher than second order differential contributions can be considered as a sources in the r.h.s. as well as the energy-momentum tensor of standard matter: 1 G =R Rg =T(curv)+T(m), (15) µν µν − 2 µν µν µν where T(m) = Tµν µν f′(R)  (16)  T(curv) = 1g f(R)−f′(R)R + f′(R);µν−gµν(cid:3)gf′(R). µν 2 µν f′(R) f′(R)  Actually,ifweconsidertheperturbedmetric(5)anddeveloptheEinsteintensoruptothefirstorderinperturbations, we have 1 1 1 G G(1) =hσ (cid:3)h h (h ,στ (cid:3)h)η (17) µν ∼ µν (µ,ν)σ− 2 µν − 2 ,µν − 2 στ − µν while the curvature stress-energy tensor gives the contributions T(curv) ξ(∂2 η (cid:3))(h ,στ (cid:3)h). (18) µν ∼ µν − µν στ − This expressionallowsto recognizethat, in the space ofmomenta, such a quantity will be responsible of the pole-like term which implies the introduction of a massive degree of freedom into the particle spectrum of gravity. In fact, inserting these two expressions into the the field Eqs. (15) and considering Eqs.(8), we obtain the solution: 5 (cid:3)h (x)= X S(0)(x)+Σξ (x) (19) µν −f′(cid:20) µν µν (cid:21) 0 where Σξ (x) is related to the curvature stress-energy tensor and is defined as µν ξ d4k k2η +2k k Σξ (x)= µν µ νS(0)(k) eikx. (20) µν 2Z (2π)2 1 3ξk2 − The general solution for the metric perturbation h (x), when the field equations are (15), can be rewritten as µν d4k S(0)(k) ξ d4k k2η +2k k h (x)= X µν eikx+ X µν µ νS(0)(k) eikx, (21) µν f′ Z (2π)2 k2 2f′ Z (2π)2 k2(1 3ξk2) 0 0 − where the second pole-like term is present. In vacuum (i.e. T(m) =0), Eqs. (10) become µν k2[h˜ (k)+ξ(k k k2η )h˜(k)]=0 µν µ ν µν −  (22)  k2h˜(k)(1 3ξk2)=0 −  showing that allowed solutions are of two types, i.e.: ω = k ±| |  , (23)  h (x)= d4k h (k) eikx with h(k)=0 µν (2π)2 µν  R and ω = k2+ 1  ±q 3ξ  h (x)= d4k ηµν+6ξkµkν h(k) eikx with h(k)=0 . (24) µν − (2π)2(cid:20) 6 (cid:21) 6  R Itisevident,thatthefirstsolutionrepresentsamasslessgravitonaccordingtothestandardprescriptionsofGRwhile thesecondonegivesamassivedegreeoffreedomwithm2 =(3ξ)−1 =−3ff0′0′′. Thankstothispropertiy,wecanrewrite Eqs.(10) introducing a scalar field φ=(cid:3)h˜ so that the general system can be rearrangedin the following way (cid:3)h˜ = XT(0)+ ∂µ2ν−ηµν(cid:3) φ  µν −f0′ µν (cid:20) 3m2 (cid:21) (25)  ((cid:3)+m2)φ= Xm2T(0)  −f0′ whichsuggeststhatthe higher ordercontributionsact, inthe post-Minkowskianlimit, asa massivescalarfieldwhose mass depends on the derivatives f′(R) and f′′(R), calculated on the unperturbed background metric. The massive mode is directly related to the coefficients of the Taylor expansion and it is interesting to note that they dermine alsothe value ofthe Yukawacorrectioninthe Newtonianapproximation[17,18]. Onthe otherhand, it is straightforwardto see that massive modes are directly related to the non-trivial structure of the trace equation as itiseasytosee fromEq.(3). InGR,the Ricciscalaris univocallyfixedbeing R=0invacuumandR ρinpresence ∝ of matter, where ρ is the matter-energy density. 6 IV. THE STRESS-ENERGY TENSOR IN f(R)-GRAVITY AND THE GRAVITATIONAL RADIATION As we have seen, higher order theories of gravity introduce further degrees of fredom which can be taken into account by defining an additional ”curvature source term” in the r.h.s. of field equations. This quantity behaves as an effective stress-energy tensor that can characterize the energy loss due to the gravitational radiation. Although the procedure to calculate the stress-energy tensor of the gravitational field in GR is often debated, one can extend the formalismto more generaltheories andobtainthis quantity by varyingthe gravitationalLagrangian. InGR, this quantity is a pseudo-tensor and is tipically referred to as the Landau-Lifshitz energy-momentum tensor [20]. The calculations of GR need to be extended when dealing with higher order gravity. In the case of f(R)-gravity, we have ∂ ∂ ∂ δ d4x√ gf(R)=δ d4x (g ,g ,g ) d4x L ∂ L +∂2 L δg = (26) Z − Z L µν µν,ρ µν,ρσ ≈Z (cid:18)∂gρσ − λ∂gρσ,λ λξ∂gρσ,λξ(cid:19) ρσ . = d4x√ gHρσδg = 0. ρσ Z − The Euler-Lagrangeequations are then ∂ ∂ ∂ L ∂ L +∂2 L =0, (27) ∂g − λ∂g λξ∂g ρσ ρσ,λ ρσ,λξ which coincide with the field Eqs. (2) in vacuum. Actually, even in the case of more generaltheories, it is possible to define an energy-momentum tensor that turns out to be defined as follows: 1 ∂ ∂ ∂ tλ = L ∂ L g + L g δλ . (28) α √ g(cid:20)(cid:18)∂g − ξ∂g (cid:19) ρσ,α ∂g ρσ,ξα− αL(cid:21) ρσ,λ ρσ,λξ ρσ,λξ − This quantity, together with the energy-momentum tensor of matter T , satisfies a conservation law as required by µν χ the Bianchi identities. In fact, in presence of matter, one has H = T , and then µν µν 2 (√ gtλ) = √ gHρσg = X√ gTρσg = (√ gTλ) , (29) − α ,λ − − ρσ,α − 2 − ρσ,α −X − α ,λ and, as a consequence, [√ g(tλ + Tλ)] =0 (30) − α X α ,λ thatistheconservationlawgivenbytheBianchiidentities. Wecannowwritetheexpressionoftheenergy-momentum tensor tλ in term of the gravity action f(R) and its derivatives: α ∂R 1 ∂R ∂R ∂R tλ =f′ ∂ √ g g + g f′′R g δλ f, (31) α (cid:26)(cid:20)∂g − √ g ξ(cid:18) − ∂g (cid:19)(cid:21) ρσ,α ∂g ρσ,ξα(cid:27)− ,ξ∂g ρσ,α− α ρσ,λ ρσ,λξ ρσ,λξ ρσ,λξ − It is worth noticing that tλ is a non-covariant quantity in GR while its generalization, in fourth order gravity, turns α out to satisfy the covariance prescription of standard tensors (see also [2]). On the other hand, such an expression reduces to the Landau-Lifshitz energy-momentum tensor of GR as soon as f(R) = R, that is 1 ∂ tλ = LGRg δλ (32) α|GR √ g(cid:18)∂gρσ,λ ρσ,α− αLGR(cid:19) − where the GRLagrangianhas been consideredinits effective form, i.e. the symmetric partofthe Riccitensor,which effectively leads to the equations of motion, that is =√ ggµν(Γρ Γσ Γσ Γρ ). (33) LGR − µσ ρν − µν σρ 7 It is important to stress that the definition of the energy-momentum tensor in GR and in f(R)-gravity are different. This discrepancyis due to the presence,inthe secondcase,ofhigherthansecondorderdifferentialterms thatcannot be discarded by means of a boundary integration as it is done in GR. We have noticed that the effective Lagrangian of GR turns out to be the symmetric part of the Ricci scalar since the second order terms, present in the definition of R , can be removed by means of integration by parts. On the other hand, an analytic f(R)-Lagrangian can be recast, at linear order, as f f′R+ (R), where the function satisfies the condition: lim R2. As a consequence, one can rewrite th∼e ex0plicitFexpression of tλ F R→0F → α as: ∂R 1 ∂R ∂R ∂R tλ =f′tλ + ′ ∂ √ g g + g ′′R g δλ . (34) α 0 α|GR F (cid:26)(cid:20)∂gρσ,λ − √ g ξ(cid:18) − ∂gρσ,λξ(cid:19)(cid:21) ρσ,α ∂gρσ,λξ ρσ,ξα(cid:27)−F ,ξ∂gρσ,λξ ρσ,α− αF − The general expression of the Ricci scalar, obtained by splitting its linear (R∗) and quadratic (R¯) parts once a perturbed metric (5) is considered, is R=gµν(Γρ Γρ )+gµν(Γρ Γσ Γσ Γρ )=R∗+R¯, (35) µν,ρ− µρ,ν σρ µν − ρµ νσ (noticethat = √ gR¯). InthecaseofGRtλ ,the Landau-Lifshitztensorpresentsafirstnon-vanishingterm LGR − − α|GR at order h2. A similar result can be obtained in the case of f(R)-gravity. In fact, taking into account Eq.(34), one obtains that, at the lower order, tλ reads: α ∂R∗ ∂R∗ ∂R∗ 1 tλ tλ = f′tλ +f′′R∗ ∂ g + g f′′R∗ g f′′δλR∗2 = α ∼ α|h2 0 α|GR 0 (cid:20)(cid:18)− ξ∂gρσ,λξ(cid:19) ρσ,α ∂gρσ,λξ ρσ,ξα(cid:21)− 0 ,ξ∂gρσ,λξ ρσ,α− 2 0 α ∂R∗ 1 ∂R∗ = f′tλ +f′′ R∗ g R∗δλ ∂ R∗ g . (36) 0 α|GR 0(cid:20) (cid:18)∂gρσ,λξ ρσ,ξα− 2 α(cid:19)− ξ(cid:18) ∂gρσ,λξ(cid:19) ρσ,α(cid:21) Consideringthe perturbedmetric (5), we haveR∗ R(1), whereR(1) is definedas in(8). Interms ofh andη, weget ∼ ∂R∗ ∂R(1) =ηρλησξ ηλξηρσ  ∂gρσ,λξ ∼ ∂hρσ,λξ − . (37)  ∂R∗ g hλξ h,λ ∂gρσ,λξ ρσ,ξα ∼ ,ξα− α  Clearly, the first significant term in Eq. (36) is of second order in the perturbation expansion. We can now write the expression of the energy-momentum tensor explicitly in term of the perturbation h; it is 1 tλ f′tλ +f′′ (hρσ (cid:3)h)[hλξ h,λ δλ(hρσ (cid:3)h)] α ∼ 0 α|GR 0{ ,ρσ− ,ξα− α− 2 α ,ρσ− hρσ hλξ +hρσ λh +hλξ (cid:3)h (cid:3)h,λh . (38) − ,ρσξ ,α ,ρσ ,α ,α ,ξ− ,α} Considering the tilded perturbation metric h˜ , the more compact form µν 1 1 1 1 tλ = ˜h,λ (cid:3)h˜ h˜ (cid:3)h˜,λ h˜λ (cid:3)h˜,σ ((cid:3)h˜)2δλ , (39) α|f 2(cid:20)2 α − 2 ,α − σ,α − 4 α(cid:21) is achieved. As matter of facts, the energy-momentum tensor of the gravitational field, which expresses the energy transportduring the propagation,has a naturalgeneralizationin the case of f(R)-gravity. We have adoptedhere the Landau-Lifshitzdefinitionbutotherapproachescanbe takenintoaccount[21]. Thegeneraldefinitionoftλ,obtained α above, consists of a sum of a GR contribution plus a term coming from f(R)-gravity: tλ =f′tλ +f′′tλ . (40) α 0 α|GR 0 α|f However, as soon as f(R)=R, we obtains tλ =tλ . As a final remark, it is worth noticing that massive modes of α α|GR gravitational field come out from tλ since (cid:3)h˜ can be considered an effective scalar field moving in a potential: tλ, α|f α in this case, represents a transport tensor. 8 V. CONCLUDING REMARKS In this paper, we have formally studied the post-Minkowskianlimit of f(R)-gravity developing all the calculations in the Jordan frame. The main result is that, beside standard massless modes of GR, further massive modes emerge and they are directly determined by the analytic parameters of f(R)-gravity,that is the coefficients f′ and f′′ of the 0 0 Taylor expansion. This fact is extremely relevant since it does not depend on the considered f(R)-model but it is a general feature that can be enounciated in the following way: Massless gravitons are a peculiar characteristic of GR while extended or alternative theories have, in general, further massive or ghost states [13]. It is worth noticing that several indications in this sense are present in literature [12, 22, 23] but their relevance, from an experimental viewpoint, has never been stressed enough. On the other hand, a similar result comes out also in the Newtonian limit of the same theories: Yukawa-like correctionstothegravitationalpotentialemergeingeneralandtheyareabsentonlyinthecaseofGR.Itisinteresting to note that also the characteristic lengths of such corrections are related to f′ and f′′ as shown in [17]. Also in this 0 0 case, the Newtonian potential, coming from the weak field limit of GR, is only a particular case. TheseresultsposeinterestingproblemsrelatedtothevalidityofGRatallscales. Itseemsthatitworksverywellat localscales(SolarSystem)whereeffectsoffurthergravitationaldegreesoffreedomcannotbedetected. Assoonasone is investigating larger scales, as those of galaxies, clusters of galaxies, etc., further corrections have to be introduced in order to explain both astrophysical large-scale dynamics [7, 9] and cosmic evolution [3, 10]. Alternatively, huge amounts of dark matter and dark energy have to be invoked to explain the phenomenology, but, up today there are no final answerfor these new constituents atfundamental level. Furthermore,the fact that, upto now,only massless gravitationalwaveshavebeeninvestigatedcouldbeashortcomingpreventingthepossibilitytofindoutotherformsof gravitational radiation. Tests in this sense could come, for example, from the stochastic background of gravitational waves where massive modes could play a crucial role in the cosmic background spectrum [24, 25]. [1] T.Padmanabhan,Phys.Rept.380,235(2003); P.J.E.Peebles,B.Ratra,Rev.Mod.Phys.75,559(2003);V.Sahni,Lect. Notes Phys.653, 141 (2004). [2] S. Nojiri and S.D. Odintsov, Int. J. Geom. Meth. Mod. Phys. 4, 115 (2007); S. Capozziello, M. Francaviglia, Gen. Rel. Grav.40, 357 (2008); T. P. Sotiriou, V. Faraoni arXiv: 0805.1726 [gr-qc] (2008). [3] S. Capozziello, Int. J. Mod. Phys. D 11, 483, (2002); S. Capozziello, S. Carloni, A. Troisi, Rec. Res. Develop. Astron. Astrophys.1,625(2003),arXiv:astro-ph/0303041; S.Capozziello, V.F.Cardone,S.Carloni, A.Troisi, Int.J.Mod.Phys. D,12, 1969 (2003). [4] S.M.Carroll, V.Duvvuri,M.TroddenandM.S.Turner,Phys.Rev.D70,043528 (2004);S.Nojiri,S.D.Odintsov,Phys. Lett. B 576, 5, (2003); S. Nojiri, S.D. Odintsov, Phys. Rev. bf D 68, 12352, (2003); S. Capozziello, V. F. Cardone and A. Troisi, Phys. Rev. D 71, 043503 (2005); S. Carloni, P.K.S. Dunsby,S. Capozziello, A. Troisi, Class. Quant. Grav. 22, 4839 (2005). [5] G.Cognola, E. Elizalde, S.Nojiri, S.D.Odintsov,L.Sebastiani, S.Zerbini, Phys.Rev.D 77, 046009 (2008); G. Cognola, S.Zerbini, Int.J.Theor.Phys. 47, 3186 (2008). [6] G.J.Olmo,Phys.Rev.Lett.95,261102(2005);S.Capozziello, A.Troisi,Phys.Rev.D 72,044022, (2005);S.Capozziello, A. Stabile, A. Troisi, Mod. Phys. Lett. A 21, 2291 (2006); T. P. Sotiriou, Gen. Rel. Grav. 38; T. Chiba, Phys. Lett. B 575, 1 (2005). [7] R.Saffari, S.Rahvar,Phys. Rev.D 77, 104028 (2008). [8] S.NojiriandS.D.Odintsov,Phys.Rev.D77,026007(2008); S.Nojiri,S.D.OdintsovandP.V.Tretyakov,Prog.Theor. Phys.Suppl.172 81 (2008). [9] S. Capozziello, V.F. Cardone, A. Troisi, JCAP 0608, 001 (2006); S. Capozziello, V.F. Cardone, A. Troisi, Mon. Not. Roy.Astron.Soc.375, 1423 (2007); Y.Sobouti,Astron.Astroph.464,921 (2007); S.MendozaandY.M.Rosas-Guevara, Astron.Astroph472,367(2007);Ch.G.Boehmer,T.Harko,F.S.N.Lobo,Astropart.Phys.29386(2008);S.Capozziello, E. DeFilippis, V. Salzano, Mon. Not. Roy.Astron.Soc. 394, 947 (2009). [10] L. Pogosian, A.Silvestri, Phys. Rev.D 77 , 023503 (2008). [11] S. Capozziello, S. Tsujikawa Phys. Rev. D 77, 10750 (2008); S. Capozziello, M. De Laurentis, S. Nojiri, S.D. Odintsov, Gen. Rel. Grav. 41, 2313 (2009). [12] K.S.Stelle, Gen. Rev.Grav. 9, 343, (1978). [13] A. Nishizawa, A. Taruya, K. Hayama, S. Kawamura, M. Sakagami, Phys. Rev. D 79, 082002 (2009); C. Bogdanos, S. Capozziello, M. DeLaurentis, S.Nesseris, arXiv:0911.3094 [gr-qc] (2009). [14] G. Magnano, L.M. Sokolowski, Phys.Rev.D 50, 5039 (1994); V. Faraoni, E. Gunzig, Int.J.Theor.Phys. 38, 217 (1999). [15] M. Maggiore, Phys. Rep.331, 283 (2000). [16] C. M. Will, Living Rev.in Relativity 9, 3 (2006) (http://www.livingreviews.org/lrr-2006-3 or gr-qc/0510072) [17] S.Capozziello, A. Stabile, A.Troisi, Phys.Rev. D 76, 104019 (2007). [18] S.Capozziello, A. Stabile, Class. Quant.Grav. 26, 085019 (2009). 9 [19] S.Capozziello, A. Stabile, A.Troisi, Class. Quant.Grav. 25, 085004 (2008). [20] L. Landau and E.M. Lifsits, Field Theory, Pergamon Press, (1973). [21] T. Multamaki, A.Putaja, E. C. Vagenas and I.Vilja, Class. Quant.Grav. 25, 075017 (2008). [22] A.A.Starobinsky,Phys. Lett.B 91, 99 (1980). [23] R.Kerner, Gen. Rel. Grav. 14, 453 (1982). [24] S.Bellucci, S.Capozziello, M. De Laurentis , V.Faraoni, Phys. Rev.D 79, 104004 (2009). [25] S.Capozziello, M. De Laurentis, M. Francaviglia, Astropart. Phys.29, 125 (2008).