ebook img

U(1) symmetry and elimination of spin-0 gravitons in Horava-Lifshitz gravity without the projectability condition PDF

0.15 MB·English
by  Tao Zhu
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview U(1) symmetry and elimination of spin-0 gravitons in Horava-Lifshitz gravity without the projectability condition

U(1) symmetry and elimination of spin-0 gravitons in Horava-Lifshitz gravity without the projectability condition Tao Zhu, Qiang Wu, and Anzhong Wang ∗ Institute for Advanced Physics & Mathematics, Zhejiang University of Technology, Hangzhou 310032, China Fu-Wen Shu College of Mathematics & Physics, Chongqing University of Posts and Telecommunications, Chongqing 400065, China (Dated: October21, 2011) In this paper, we show that the spin-0 gravitons appearing in Horava-Lifshitz gravity without the projectability condition can be eliminated by extending the gauge symmetries of the foliation- preservingdiffeomorphismstoincludealocalU(1)symmetry. Asaresult,theproblemsofstability, ghost,strongcoupling,anddifferentspeedsinthegravitationalsectorareautomaticallyresolved. In 1 addition, with the detailed balance condition softly breaking, the number of independent coupling 1 0 constantscanbesignificantlyreduced(from morethan70downto15),whilethetheoryisstillUV 2 completeand possesses ahealthyIR limit, wherebytheprediction powers of thetheory areconsid- erably improved. The strong coupling problem in the matter sector can be cured by introducing ct an energy scale M∗, so that M∗ < Λω, where M∗ denotes the suppression energy of high order O derivativeterms, and Λω thewould-be strong coupling energy scale. 0 PACSnumbers: 04.60.-m;04.50.Kd 2 ] I. INTRODUCTION z must be z d [2, 14]. The associatedgauge symmetry h ≥ nowis brokenfromthe generaldiffeomorphismsdownto t - Quantization of gravitational fields has been one of the foliation-preserving ones, p e the main driving forces in Physics in the past decades δt= f(t), δxi = ζi(t,x), (2) h by following several different paths [1]. Recently, Ho- − − [ ravaproposeda theoryof quantumgravityin the frame- to be denoted by Diff(M, ). 2 work of quantum field theory, with the spacetime metric AbandoningthegeneralFdiffeomorphisms,ontheother v as the elementary field in the language of the standard hand, gives rise to a proliferation of independently cou- 7 path-integral formulas [2]. Applied to cosmology, it re- pling constants [2, 15], which could potentially limit the 3 sults in various remarkable features and has attracted a prediction powers of the theory. In particular, only the 2 great deal of attention [3]. In particular, high order spa- sixth-orderspatialderivativetermsaremorethan60[16]. 1 tial derivative terms cangive rise to a bouncing universe To reduce the number of the independent coupling con- . 8 [4, 5]; the anisotropic scaling solves the horizon problem stants, two conditions were introduced, the projectibility 0 and leads to scale-invariant perturbations without infla- and detailed balance [2]. The former requires that the 1 tion [6, 7]; the lack of the local Hamiltonian constraint lapse function N be a function of t only, while the latter 1 : leadsto“darkmatterasanintegrationconstant”[8];the requires that the gravitational potential should be ob- v darksectorcanhaveits purelygeometricorigins[9]; and tained from a superpotential W , where W is given by Xi theinclusionofaU(1)symmetry(withtheprojectability an integral of the gravitational Cghern-Simogns term over condition) [10, 11] not only eliminates the spin-0 gravi- a 3-dimensional space, W ω (Γ). With these two ar tons but also leads to a flat universe [12]. conditions, the general acgtio∼n nΣow3contains only five in- With the perspective that Lorentz symmetry may ap- dependent coupling constantsR[2]. The detailed balance pearonlyasanemergentoneatlowenergies,butcanbe condition has several remarkable features [2, 17]. For fundamentallyabsentathighenergies,Horavaconsidered example, it is in the same spirit of the AdS/CFT cor- a gravitational system whose scaling at short distances respondence [18], where a string theory and gravity de- exhibits a strong anisotropy between space and time, fined on one space is equivalent to a quantum field the- ory without gravity defined on the conformal boundary x b 1x, t b zt. (1) → − → − of this space, which has one or more lower dimension(s). Yet, in the non-equilibrium thermodynamics, the coun- This is quite similar to Lifsitz’s scalar field [13], so the terpart of the superpotential W plays the role of en- theory is often referred to as the Horava-Lifshitz (HL) g tropy, while δW /δg represents the corresponding en- gravity. In (d+1)-dimensions, in order for the theory to g ij tropic force [19]. This might shed further lights on the be power-counting renormalizable, the critical exponent nature of gravitational forces, as proposed recently by Verlinde [20]. Despite of all the above remarkable features, the the- ∗On Leave from GCAP-CASPER, Physics Department, Baylor ory (with these two conditions) is plagued with several University,Waco, TX76798-7316, USA problems [3], including the instability, ghost [2, 21], and 2 strongcoupling[16,22]. However,allofthem areclosely +2K ˆijlka α+2(1 λ)K 2α+ak α ,(6) ij (l k) k related to the existence of the spin-0 gravitons [3]. Be- G ∇ − ∇ ∇ cause of their presence, another important question also where ˆijkl = ijkl , ijkl(λ) (cid:0)(gikgjk+gilgjk(cid:1))o/2 raises: Their speeds are generically different from those G G λ=1 G ≡ − λgijglk, G R Rg /2 and f = (f +f )/2. of gravitational waves, as they are not related by any ij ≡ ij(cid:12)− ij (ij) ij ji R (R) is the Ric(cid:12)ci tensor (scalar) of g . To have symmetry. This poses a great challenge for any attempt ij ij the U(1) symmetry, we introduce the gauge field A [10], to restore Lorentz symmetry at low energies where it which transforms as has been well tested experimentally. In particular, one needs a mechanism to ensure that in those energy scales δ A=α˙ Ni α. (7) α i all species, including gravitons, have the same effective − ∇ speed and light cones. With these in mind, Horava and Then, adding 1 Melby-Thompson (HMT) [10] extended the symmetry (2) to include a local U(1), S = ζ2 dtd3x√gA(R 2Λ ), (8) A g − − U(1)⋉Diff(M, ). (3) Z F to Sˆ , one finds that its variation (with Λ = 0) with g g With this enlarged symmetry, the spin-0 gravitons are respect to α exactly cancels the first term in Eq.(6). To eliminated[10,23,24],andthetheoryhasthesamelong- repairthe rest, weintroduce the Newtonianprepotential distance limit as that of general relativity. This was ini- ϕ [10], which transforms as tially done in the special case λ = 1, and soon general- ized to the case with any λ [11], where λ is a coupling δ ϕ= α. (9) α − constant defined below. Although the spin-0 gravitons Under Eq.(4), the variation of the term, are also eliminated in the general case, the strong cou- plingproblemraisesagain[12]. However,itcanbesolved by the Blas-Pujolas-Sibiryakov (BPS) mechanism [15], Sϕ =ζ2 dtd3x√gN ϕ ij 2Kij + i jϕ+ai jϕ G ∇ ∇ ∇ iMn w<hicΛh a, wnhewereenMergydesncoatleesMth∗eissuipnptrreosdsuiocnede,nseorgtyhaotf +(1 Zλ) 2ϕ+aniϕ 2(cid:0)+2 2ϕ+a iϕ K (cid:1) ω i i hig∗h order derivativ∗e terms of the theory, and Λ the − ∇ ∇ ∇ ∇ would-be strong coupling energy scale [25]. ω + 1 ˆijlk h4(cid:0)( ϕ)a (cid:1) ϕ+(cid:0)5 a ϕ a (cid:1) iϕ 3G ∇i∇j (k∇l) (i∇j) (k∇l) Note that in [10, 11] the projectability condition was h (cid:0) (cid:1) assumed. In this paper, we shall study the case without +2 ϕ a ϕ+6K a ϕ , (10) (i j)(k l) ij (l k) ∇ ∇ ∇ it, and our goals are twofold: (i) Extend the symmetry (cid:0) (cid:1) io (3) to the case without the projectability condition, so will exactly cancel the rest in Eq.(6) as well as the one thatthe spin-0 gravitonsareeliminated. (ii) Reduce sig- from the term Λ A in Eq.(8), where G +Λ gij g ij ij g G ≡ nificantly the number of the coupling constants by the and a a . Hence, the total action ij i j ≡∇ detailed balance condition, whereby the prediction pow- ers of the theory can be improved considerably. To have Sg =Sˆg+SA+Sϕ, (11) a healthy IR, we allow it to be broken softly by adding is invariant under (3). allthe lowdimensionalrelevantterms [17]. We alsonote Equation (3) does not uniquely fix . To our sec- that in [26] a U(1) extension of F(R) HL gravity was LV ondgoal,weintroducethe“generalized”detailedbalance recently studied. condition, ˆ = E ijklE g AiAj, (12) II. THE MODEL LV ijG kl− ij where UndertheU(1)transformations,themetriccoefficients 1 δW 1 δW transform as [10], Eij = g, Ai = a, √g δgij √g δai δαN =0, δαNi =N iα, δαgij =0, (4) 1 2 ∇ W = Tr Γ dΓ+ Γ Γ Γ , where α is the U(1) generator, Ni the shift vector, and g w2 ZΣ (cid:16) ∧ 3 ∧ ∧ (cid:17) i the covariant derivative of gij. Since [15] 1 ∇ W = d3x√gai b ∆na , (13) a n i Sˆg =ζ2 dtdx3N√g(LK −LV(gij,ak)), (5) Z nX=0 Z where a (lnN), , K Kij λK2 and K = i i K ij ij ≡ L ≡ − ( g˙ + N + N )/(2N),thepotential isinvari- a−ntiujnde∇ri(4)j, w∇hijle tihe kinetic part transfoLrmVs as 1 Notethedifferencebetweenthenotationsusedhereandtheones used in [10, 11]. In particular, we have Kij =−KiHjMT, Λg = δS =ζ2 dtd3xN√g α˙ Ni α R +2αG Kij ΩHMT, ϕ=−νHMT,Gij =ΘHijMT, where quantities with the K − ∇i N ij superindices“HMT”wereusedin[10,11]. Z n(cid:0) (cid:1) 3 with ∆ gij and b being arbitrary constants. λ 1 ∂2B = 1 3λ ψ˙, (19) i j n ≡ ∇ ∇ − − Now, some comments are in order. First, the term ðφ=2℘ψ, (20) a ∆1/2ai inprinciplecanbeincludedintoW ,whichwill (cid:0) (cid:1) (cid:0) (cid:1) i a give rise to fifth order derivative terms. In this paper we where ℘ 1 ζ 2β ∂2. Eq.(17) shows that ψ is not − 7 shall discard them by parity. Second, it is well-known propagati≡ng, a−nd with proper boundary conditions, one that with the detailed balance condition a scalar field is can set ψ = 0. Then, Eqs.(18)-(20) show that B, A, not UV complete [4], and the Newtonian limit does not and φ are also not propagating and can be set to zero. exist [27]. Following [17], we break it softly by adding Hence, we obtain ψ = B = A = φ = 0, that is, the allthe lowdimensionalrelevanttermsto ˆ ,sothatthe scalar perturbations vanish identically in the Minkowski V L potential finally takes the form, background, similar to that in general relativity. Thus, withtheenlargedsymmetry(3),the spin-0gravitonsare V = γ0ζ2 β0aiai γ1R +ζ−2 γ2R2+γ3RijRij indeed eliminated even without the projectability condi- L − − tion. +ζ−2 β(cid:16)1 aiai 2+β2(cid:17)aii 2+(cid:16)β3 aiai ajj (cid:17) +β4aihjaij(cid:0)+β5(cid:1) aiai R(cid:0) +(cid:1)β6aiajR(cid:0)ij +(cid:1)β7Raii IV. STRONG COUPLING AND THE BLAS-PUJOLAS-SIBIRYAKOV MECHANISM +ζ 4 γ C Cij(cid:0)+β (cid:1)∆ai 2 , (1i4) − 5 ij 8 where β bh2, γ ζ4/w4(cid:0), β (cid:1) i ζ4b2, and C de- Since the spin-0 gravitons are eliminated, the ghost, 0 ≡ 0 5 ≡ 8 ≡ − 1 ij instability, strong coupling and different speed problems notes the Cotton tensor, defined as Cij ≡ e√ikgl∇k Rlj − in the gravitational sector do not exist any longer. But, 1Rδj . Allthecoefficients,β andγ ,aredimensi(cid:16)onless the self-interaction of matter fields and the interaction 4 l n n between a matter field and a gravitational field can still and a(cid:17)rbitrary, except β0 0,γ5 0 and β8 0, as indi- lead to strong coupling, as shown recently in [25] for the ≥ ≥ ≤ cated by their definitions. γ0 is related to the cosmolog- theorywiththeprojectabilitycondition. Inthefollowing ical constant by Λ = ζ2γ0/2, while the IR limit requires we shall show that this is also the case here. Since the γ1 = 1, ζ2 =1/(16πG),whereGistheNewtoniancon- proofisquitesimilartothatgivenin[25],inthefollowing − stant. From the above, we can see that with the “gen- we just summarize our main results, and for detail, we eralized” detailed balance condition softly breaking, the refer readers to [25]. A scalar field χ with detailed bal- numberofindependentcouplingconstantsissignificantly ance condition softly breaking is described by Eqs.(3.11) reduced from more 70 to 15: G, Λ, λ, β , γ , (n = n s and(3.12)of[17]. IntheMinkowskibackground,wehave 0,...,8; s = 2,3,5). In the following, we shall show that χ¯ = 0 = V(0) = V (0). Considering the linear pertur- ′ such a setup is UV complete and IR healthy. bations χ = δχ, we find that the quadratic part of the scalar field action reads, III. ELIMINATION OF THE SPIN-0 1 1 1 GRAVITONS S(2) = dtd3x fχ˙2 V χ2 (1+2V )(∂χ)2 χ "2 − 2 ′′ − 2 1 Z To show this, it is sufficient to consider linear pertur- bations in the Minkowski background, given by +c1A∂2χ−V2(∂2χ2)2−V4′χ∂4χ+σ32∂2χ∂4χ#, (21) N = 1+φ, N =∂ B, g =(1 2ψ)δ +2E , i i ij ij ,ij − wheref isaconstant,usuallychosentobeone[17]. Since A = δA, ϕ=δϕ. (15) S(2) doesnotdependonψ, B andφexplicitly,thevaria- χ Choosing the gauge, E =0=δϕ, we find that tionsofthetotalactionS(2) =S(2)+S(2) withrespectto χ g them will give the same Eqs.(18) - (20), while Eq.(17) is Sg(2) = ζ2 dtd3x (1−3λ)(3ψ˙2+2ψ˙∂2B) replaced by ψ =c1χ/(4ζ2). Integrating out φ, ψ, B, A, Z n we obtain +(1 λ)(∂2B)2 φð+4β7ζ−2∂2ψ ∂2φ − − −2(ψ−2φ+2A+(cid:0)α1ψ∂2)∂2ψo, (cid:1) (16) S(2) =M2Z dtd3x"χ˙2−α0 ∂χ 2−m2χχ2 whereα ζ 2(8γ +3γ )andð β +ζ 2(β +β )∂2 (cid:0) (cid:1) 1 ≡ − 2 3 ≡ 0 − 2 4 − ℘2χ φζ−y4iβe8ld∂,4r.eVspaerciattivioenlys,ofSg(2) withrespecttoA, ψ, B,and −MA−2χ∂4χ+MB−4χ∂6χ+γχ∂2(cid:18) ð (cid:19)#, (22) ∂2ψ =0, (17) where M2 2πGc2/c 2 +f/2, γ 4πGc2/M2, α ψ¨+ 1∂2B˙ + 2∂2(A+ψ+α1∂2ψ) = 2℘∂2φ , (18) 1+2V1−≡4πGc21 /1(2|Mψ|2), MA2 ≡M≡2/ 2πG1α1c21+V02≡+ 3 3(1−3λ) 3(1−3λ) (cid:0)V4′ , MB4 ≡ M2/σ(cid:1)32, m2χ ≡ V′′/(2M2)(cid:0), and c2ψ = (1− (cid:1) 4 λ)/(3λ 1). Clearly, the scalar field is ghost-free for M &M ,thesixth-orderderivativetermsbecomedom- A B − f > 0, and stable in the UV and IR. In fact, it can be inant for processes with E & M , and the quadratic madestableinallenergyscalesbyproperlychoosingthe action (22) now is invariant only u∗nder the anisotropic coupling coefficients V . rescaling, n To consider the strong coupling problem, one needs to calculate the cubic part of the total action, which takes t b 3t, xi b 1xi, χ χ. (26) − − → → → the form, Then,onefindsthatunderthenewrescalingalltheterms 1 1 of λ in Eq.(23) become scaling-invariant, while the rest S(3) = dtd3x λ χ¨ χ∂2χ+λ χ¨ χ χ,i s 1 ∂2 2 ∂2 ,i scale as b δ with δ > 0. The former are strictly renor- Z ( (cid:18) (cid:19) (cid:18) (cid:19) − malizable, while the latter are superrenormalizable [28]. 2℘ 2℘ χ˙ +λ χ˙2 1 χ+λ χ˙∂i 1 χ∂ ThereforeEq.(25)makesthestrongcouplingproblemdis- 3 (cid:18) ð − (cid:19) 4 (cid:18) ð − (cid:19) i(cid:18)∂2(cid:19) appeared. ∂i∂j ∂ 2℘ V. CONCLUSIONS i +λ χ˙ χ˙ ∂ +3 χ 5 ∂2 ∂2 j ð (cid:18) (cid:19)(cid:18) (cid:19) (cid:18) (cid:19) Inthis paper,wehaveshownthatthe spin-0gravitons ∂ +λ χ˙χ,i iχ˙ +... , (23) in the HL theory even without the projectability condi- 6 ∂2 (cid:18) (cid:19) ) tioncanbeeliminatedbytheenlargedsymmetry(3). An immediateresultisthatalltheproblemsrelatedtothem where “...” represents the terms that are independent of disappear, including the ghost, instability, strong cou- λ, so they are irrelevant to the strong coupling prob- pling and different speeds in the gravitationalsector. In lem. λ are functions of λ,f,ζ and c . In particular, s i addition,therequirements[15], λ 1 β , 0<β <2, λ =c3/(64ζ4 c 4), whichwill yieldthe strongcoupling | − |≃ 0 0 5 1 | ψ| now become unnecessary, which might help to relax the energy Λ given below. The exact expressions of other ω observationalconstraints. Moreover,itisexactlybecause coefficients are not relevant to Λ , so will not be given ω of this elimination that softly breaking detailed balance here. For a process with energy E M ,M , the first two terms in Eq.(22) are dominant,≪and SA(2) iBs invariant condition can be imposed, whereby the number of inde- pendent coupling constants is significantly reduced from undertherelativisticrescalingt b 1t,xi b 1xi,χ bχ. Then, all the terms of λ →in S−(3) sca→le a−s b. As→a more than70 to 15. This considerablyimprovesthe pre- s diction powers of the theory. Note that, without the result, when the energy of a process is greater than a elimination of the spin-0 gravitons, the strong coupling certain value, say, Λ , the amplitudes of these terms are ω problem will appear in the gravitational sector and can- greater than one, and the theory becomes nonrenormal- notbesolvedbytheBPSmechanism,becausethiscondi- izable [28]. In the present case, it can be shown that tionpreventstheexistenceofsixth-orderderivativeterms (3) M 3/2 in Sg . Λ pl M c 5/2. (24) TheseresultsputtheHLtheorywith/withoutthepro- ω pl ψ ≃ c | | (cid:18) 1 (cid:19) jectability condition in the same footing, and provide a very promising direction to build a viable theory of However, if quantum gravity, put forwards recently by HMT [10]. M <Λ , (25) Certainly, many challenging questions [3] need to be an- ω ∗ swered before such a goal is finally reached. where M = Min.(M ,M ), one can see that before Acknowledgements: This work was supported A B the stron∗g coupling energy Λ is reached, the high or- in part by DOE Grant, DE-FG02-10ER41692 (AW); ω der derivative terms in Eq.(22) become large, and their NNSFC No. 11005165 (FWS); and NNSFC No. effects must be taken into account. In particular, for 11047008(QW, TZ). [1] S. Weinberg, in General Relativity, An Einstein Cen- 084008 (2009). tenary Survey, edited by S.W. Hawking and W. Is- [3] D. Blas, O. Pujolas, and S. Sibiryakov, JHEP, 1104, rael (Cambridge University Press, Cambridge, 1980); C. 018 (2011); S. Mukohyama, Class. Quantum Grav. 27, Kiefer, Quantum Gravity (Oxford Science Publications, 223101 (2010); P. Horava, ibid., 28, 114012 (2011); A. Oxford University Press, 2007); K. Becker, M. Becker, Padilla, J. Phys. Conf. Ser. 259, 012033 (2010); T.P. and J.H. Schwarz, String Theory and M-Theory (Cam- Sotiriou, ibid., 283, 012034 (2011); T. Clifton, et al, bridge University Press, Cambridge, 2007); C. Rovelli, arXiv:1106.2476. Quantum gravity (Cambridge University Press, Cam- [4] G. Calcagni, JHEP, 09, 112 (2009). bridge, 2008). [5] R. Brandenberger, Phys. Rev. D80, 043516 (2009); A. [2] P. Horava, JHEP, 0903, 020 (2009); Phys. Rev. D79, Wang and Y. Wu,JCAP, 07, 012 (2009). 5 [6] S.Mukohyama, JCAP, 06, 001 (2009). [21] T. Sotiriou, M. Visser, and S. Weinfurtner, Phys. Rev. [7] E.KiritsisandG.Kofinas,Nucl.Phys.B821,467(2009). Lett. 102, 251601 (2009); JHEP, 10, 033 (2009); C. [8] S. Mukohyama, Phys. Rev. D80, 064005 (2009); JCAP, Bogdanos, and E. N. Saridakis, Class. Quant. Grav. 27, 09, 005 (2009). 075005 (2010); A. Wang and R. Maartens, Phys. Rev. [9] A.Wang, Mod. Phys.Lett. A26, 387 (2011). D81, 024009 (2010). [10] P.Horava and C.M. Melby-Thompson, Phys. Rev.D82, [22] C. Charmousis, G. Niz, A. Padilla, and P.M. Saffin, 064027 (2010). JHEP, 08, 070 (2009); D. Blas, O. Pujolas, and S. [11] A.M.daSilva,Class.QuantumGrav.28,055011(2011). Sibiryakov, ibid., 03, 061 (2009); K. Koyama and F. [12] Y.-Q. Huang and A. Wang, Phys. Rev. D83, 104012 Arroja, ibid., 03, 061 (2010); A. Papazoglou and T.P. (2011). Sotiriou, Phys.Lett.B685, 197(2010); A.WangandQ. [13] E.M. Lifshitz, Zh.Eksp.Toer. Fiz. 11, 255 (1941); ibid., Wu, Phys.Rev.D83, 044025 (2011). 11, 269 (1941). [23] A. Wang and Y.Wu,Phys. Rev.D83, 044031 (2011). [14] M. Visser, Phys.Rev.D80, 025011 (2009). [24] J. Kluson, Phys.Rev. D83, 044049 (2011). [15] D. Blas, O. Pujolas, and S. Sibiryakov, Phys. Rev.Lett. [25] K. Lin, A. Wang, Q. Wu, and T. Zhu, Phys. Rev. D84, 104, 181302 (2010); Phys. Lett. B688, 350 (2010). 044051 (2011). [16] I.Kimpton and A.Padilla, JHEP, 07, 014 (2010). [26] J. Kluson, et al, Eur. Phys.J. C71, 1690 (2011). [17] A.Borzou,K.Lin,andA.Wang,JCAP,05,006 (2011). [27] H. Lu¨, J. Mei, and C.N. Pope, Phys. Rev. Lett. 103, [18] J.M.Maldacena,Adv.Theor.Math.Phys.2,231(1998); 091301(2009);R.G.Cai,L.M.Cao,andN.Ohta,Phys. O. Aharony, S.S. Gubser, J. Maldacena, H. Ooguri, and Rev. D80, 024003 (2009); A. Kehagias and K. Sfetsos, Y.Oz, Phys. Rept.323, 183 (2000). Phys.Lett.B678,123(2009);M.-i.Park,JHEP,09,123 [19] L.OnsagerandS.Machlup,Phys.Rev.91,1505(1953); (2009). S.Machlup and L. Onsager, ibid.,91, 1512 (1953). [28] J. Polchinski, arXiv:hep-th/9210046. [20] E.P. Verlinde, JHEP, 04, 029 (2011).

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.