Free Gravitons Break de Sitter Invariance ( arXiv:0907.4930 , - PowerPoint PPT Presentation

Free Gravitons Break de Sitter Invariance ( arXiv:0907.4930 , 1002.4037 ) S. P. Miao (Utrecht) N. C. Tsamis (Crete) R. P. Woodard (Florida)

Spacetime Exp. Strengthens QFT � Why? � Loops � classical physics of virtuals � Expansion � holds virtuals apart longer � Maximum Effect for � Inflation � M=0 � No conformal invariance (classically) � Two Particles � MMC scalars � gravitons

Primordial Inflation was nearly ² de Sitter with small GH � ds 2 = 9dt 2 + a ² (t) dx.dx � H(t) = á/a & ε(t) = 9Ḣ/H ² � For single9scalar inflation with k = H(t k )a(t k ) ² � @ ² R (k) ≈ GH ² (t k )/ π ε(t k ) & @ h (k) ≈ 16GH ² (t k )/ π � WMAP data for k = .002/Mpc � @ ² R = 2.441 x 10 99 & r = @ ²h /@ ² R < 0.22 � Hence � ε ≈ r/16 < 0.014 (even smaller before t k !) � GH ² ≈ π /16 x r x @ ² R < 10 910

MMC Scalar Models λφ 4 ( Brunier , Kahya , Onemli ) 1. M ² (x;x') � @u(t,k) & <T �ν > � Growing scalar mass & pos. vac. Energy � SQED ( Kahya , Prokopec , Tornkvist , Tsamis ) 2. M ² (x;x') � @u & [ � Π ν ](x;x') � @ε � � <φ*φ>, <(D � φ)*D ν φ>, <F �ν F ρσ > & <T �ν > � Growing photon mass & neg. vac. Energy � Yukawa ( Duffy , Prokopec , Miao ) 3. M ² (x;x') � @u, Σ(x;x') � @u & <φψ † γ 0 ψ> � Growing fermion mass & neg. vac. Energy �

Quantum Gravity Models QG + Dirac (Miao) 1. [ i Σ j ](x;x') � @u(t,k) � Growing fermion field strength � QG + MMC Scalar (Kahya,Park) 2. M ² (x;x') � @u(t,k) � [ �ν Σ ρσ ](x;x’) � @ε �ν & force of gravity � Possible tilt in Power Spectrum � QG (Tsamis,Mora) 3. [ �ν Σ ρσ ](x;x') & <h Zν > � Consistent with relaxation of Λ �

Enhanced QFT as IR Logs � What? � factors of ln(a) = Ht � Eg ρ = λ (H/2 π ) 4 x ⅛ ln ² (a) + O( λ  ) � Through propagators � i@(x;x’) = (dS inv) + H ² /8 π ² ln(aa’) � i[ ij @ kℓ ](x;x’) = [2δ i(k δ ℓ)j – 2δ ij δ kℓ ] x same � Also from vertex integrations � ∫ t dt’ 1 = t = ln(a)/H � NB occur even if no dS breaking in i@!

Math Guys ���� ���� IR logs � Reluctantly accept in i@(x;x’) � But struggle to avoid consequences � But deny in i[ ij @ kℓ ](x;x’) � NB vertex integrations still break dS ∫d 4 x’ √9g(x’) θ(x 0 9x’ 0 ) θ[9ℓ 2 (x;x’)] = ∫ t dt’ a 3 (t’) x 4 π /3H 3 (1/a’ – 1/a) 3 = 4 π /3H 4 [ln(a) + O(1)] � But simplest IR logs come from props

dS Inv Eqns Don’t Always Have Invariant Solutions � MMC φ: □i@(x;x’) = iδ 4 (x9x’)/√9g � Allen & Folacci, PRD35 (1987) 3371. � ds 2 = 9dt 2 + a 2 (t) dx—dx a(t) = e Ht � i@(x;x’) = ∫ d 3 k/(2 π ) 3 e k—(x9x’) x[θ(t9t’)u(t,k)u * (t’,k) + θ(t’9t)u * (t,k)u(t’,k)] � u(t,k) = H/(2 k 3 ) ½ [1 – ik/Ha] Exp[ik/Ha] � IR problem: uu * ~ H 2 /2k 3

What about i[ �ν @ ρσ ](x;x’)? Cosmologists: not invariant � Grishchuk (Sov. Phys. JETP 40 (1975) 409) � Gravitons have same u(t,k) as MMC φ � This IS observable! @ ²h = k 3 /2 π  ∫ d 3 x e ik—x <h ij (t,x) h ij (t,0)> = k 3 /2 π  × 32 π G × 2 × |u(t,k)| 2 = 16/ π GH 2 (a.k.a. SCALE INVARIANCE) � Kleepe (PLB 317B (1993) 305) � Comp. trans. does not restore invariance

What about i[ �ν @ ρσ ](x;x’)? Math Physicists: Yes it is! � Add α (D ν h ν� + β D � h νν ) 2 � Solve in Euclidean space & continue � Ok except few “singular” choices of α and β Burden of my Talk: Math Physicists are wrong � Obstacle to adding gauge fixing term � Obstacle to analytic continuation � Origin of “singular” gauges

“Exact” vs “Average” Gauges � Illustrate with EM in flat space � Exact: ∂ i A i = 0 (Coulomb) � Average: L � L 9 ½ ( ∂ � A � ) 2 (Feynman) � Derive Average from Exact � Start in canonical functional formalism ∫ [dE T ] [dA T ] e iSfixed � S. Coleman, Erice 1973

Coleman’s Seven Steps Integrate out E T 1. Use ∫ [dA T ] = ∫ [dA] δ[ ∂ i A i ] √det[ ∂ i ∂ i ] 2. Undo A 0 constraint 3. Write integrand as invariant 4. δ[ ∂ � A � ] w field dependent gauge trans 5. δ[ ∂ � A � –f(x)] w C9number gauge trans 6. Multiply by ∫[df] Exp[9½i∫f 2 ] 7.

Obstacle on T 3 × R � Invariant: ∂ i F i0 = J 0 � Q = 0 � Feynman: [9 ∂ t2 + ∂ i ∂ i ]A 0 = J 0 � Q ≠ 0 ok � Problems at Coleman’s steps 2 & 3 � No 09modes for δ[ ∂ i A i ] and A 0 � Hence no 09mode for gauge fixing term � Same Obstacle on de Sitter � IR ∞ of φφ * self9energy (gr9qc/0508015)

Analytic Continuation Sees Only Logarithmic IR Divergences � [□9M 2 ]i@(x;x’) = iδ 4 (x9x’)/√9g � i@(x;x’) = ∫ d 3 k/(2 π ) 3 e ik—(x9x’) x[θ(t9t’)u(t,k)u * (t’,k)+θ(t’9t)u * (t,k)u(t’,k)] � u(t,k) = [ π /(4Ha 3 )] ½ H ν (1) [k/Ha] ν = [9/4 – M 2 /H 2 ] ½ � � uu * ~ k 92 ν [1 + O(k 2 )] � IR ∞’s for 2 ν ≥ 3 � M 2 ≤ 0 � But only logarithmic for M 2 = 9N(3+N) H 2

IR ∞’s Signal Wrong Physics � DON’T subtract them, fix the physics � Exclusive � Inclusive in flat QED, QCD & QG � Physical Problem: � Can’t enforce Bunch9Davies for k < Ha initial � Standard Fixes � Vilenkin (NPB:226,527,1983) � Change Bunch9Davies for k < Ha initial � NCT and RPW (CQG:11,2969,1994) � Keep Bunch9Davies on T 3 xR with no k < Ha initial

How It Works In Practice � i@ naïve (x;x’) = ∫ d 3 k/(2 π ) 3 e k—(x9x’) x[θ(t9t’)u(t,k)u * (t’,k) + θ(t’9t)u * (t,k)u(t’,k)] � Just cut off IR � ∫ d 3 k/(2 π ) 3 e k—(x9x’) θ(k9k 0 ) x (Same) � Resolves old problem of Ford & Parker (1977) � Scalar9driven FRW Iliopoulos,Tomaras,NCT,RPW (NPB:534,419,1998) � MMC scalars on FRW with constant ε Janssen,SPM,Prokopec,RPW (CQG:25,245013,2008) ² and M ² ² ² � de Sitter with M S = 9N(N+3)H V = 9(N+1)(N+3)H SPM,NCT,PRW (JMP:51,072503,2010)

Why Not Use the Subtracted Solutions for Power Law IR ∞’s? � [□9M ² ] i@(x;x’) = iδ 4 (x9x’)/√9g but i@(x;x’) ≠ <ψ|T[φ(x)φ(x’) ]|ψ> � Eg i@(x;x’) � i G ret (x;x’) � Vanishes for x’=x vs <ψ|φ ² |ψ> ≠ 0 � SHO: 9m[(d/dt) ² +ω ² ] i@(t;t’) = iδ(t9t’) 9i sin[ω|t9t’|]/2mω + α cos(ωt)cos(ωt’) + β sin[ω(t+t’)] + γ sin(ωt)sin(ωt’) � Solves for any α , β & γ, but QM requires α + γ ≥ 1/2mω and α γ ≥ ¼ β ² Math: Reflection Positivity fails �

Exact de Donder Gauge � mtrc �ν = g �ν + h �ν with D ν h ν� = ½D � h ν ν αβ i[ αβ @ ρσ ](x;x’) ≠ g � ( ρ g σ ) ν iδ 4 (x9x’) � D �ν � Not consistent with gauge condition � rhs = [g � ( ρ g σ ) ν 9 ½ g �ν g ρσ ] iδ 4 (x9x’) + 2 √9g Sym{D � D ρ ’ i[ ν @ σ ](x;x’)} � [□+3H 2 ] i[ � @ ν ](x;x’) = g �ν iδ 4 (x9x’)/√9g � Corresponds to M V2 = 96H 2 � IR ∞’s M V2 ≤ 0, Log M V2 = 9(N+2)(N+3)H 2

Scalar Structure Functions in i[ �ν @ ρσ ](x;x’) � Spin 0 Part: P �ν (x) P ρσ (x’) F 0( x;x’) P �ν = D � D ν + ½ [□+6H ² ] g �ν � � ¾[□+4H ² ][□+6H ² ] ² F 0 (x;x’) = iδ 4 (x9x’)/√9g � M ² = 94H ² is Log ∞, M ² = 96H ² is power ∞ � Spin 2 Part: P �νβ δ (x) P ρσκθ (x’) [ T βκ T δ θ F 2 ] � C αβ γδ = P �ν αβ γδ h �ν + O(h 2 ) β δ = 91/2H ² P �ν αβ γδ D α D γ � P �ν � T βκ = 91/2H ² ∂ ² y/∂x β ∂x’ κ y= aa’H ² (x9x’) ² � □ ³ [□92H ² ] ² F 2 = 64 H 4 iδ 4 (x9x’)

Conclusion: Graviton Propagator Is NOT de Sitter Invariant � Plausible arguments each way � Pro: Inv. solns w some gauge fixing terms � Con: Dynamically same as MMC scalars + IR divergences in some gauges � Long controversy resolved � Obstacle to adding gauge fixing terms � Obstacle to Euclidean continuation � De Donder projection operator not invariant