Nonlocal Cosmology
arXiv:0705.0153 & 1307.6693
S. Deser & R. P. Woodard
Problem: what is making the
universe accelerate?
FRW: ds2 = -dt2 + a2(t) dx.dx
H0 ~ 67 km/(s Mpc)
q0 ~ -.54
General Relativity with 1+z = a0/a
2
H(t) = dln(a)/dt
q(t) = -1 - Ḣ/H2
2
3H = 3H0 [Ωr(1+z)4 + Ωm(1+z)3 + Ωvac]
2
2
-2Ḣ-3H = 3H0 [1/3 Ωr(1+z)4 + 0 - Ωvac]
ΛCDM works
Ωr ~ 8.5 x 10-5, Ωm ~ .306, Ωvac~.692
But why is GΛ ~ 10-122 & why NOW?
Quintessence works
L = -½∂µφ∂νφgµν√-g – V(φ)√-g
3H2 = 8πG [½ḟ2 + V(f)]
-2Ḣ-3H2 = 8πG [½ḟ2 - V(f)]
Given a(t) Reconstruct V(f)
t
-2Ḣ = 8πG ḟ2
f(t) = f0 ± ∫0 dt’ [-Ḣ(t’)/4πG]½
Monotonic t(f)
Ḣ+3H2 = 8πG V V(f) = [Ḣ(t[f]) + 3H2(t[f])]/8πG
But who ordered that!
Why homogeneous?
Why so small?
Why no new scalar force?
f(R) models do not really work
L = f(R)√-g/16πG
Unique solution for ΛCDM is . . .
f(R) = R - 2Λ
Dunsby et al, arXiv:1005.2205
Hence deviations occur at zeroth order!
Other problems
Why now? new scales
New scalar DoF need screening
Modifications of Gravity
Only local, stable, metric-based is f(R)
Nonlocal modifications proposed for
Summing quantum IR effects from inflation
Explaining late time acceleration w/o DE
Explaining galaxies & clusters w/o DM
Isaac Newton in 1692/3
“that one body may act upon another at a
distance thro' a Vacuum, without the
Mediation of any thing else, by and
through which their Action and Force
may be conveyed from one to another,
is to me so great an Absurdity that I
believe no Man who has in philosophical
Matters a competent Faculty of thinking
can ever fall into it.”
Was Newton wrong about
action-at-a-distance?
We don’t think so
Primordial inflation IR gravitons
Fundamental theory is local
But quantum effective field eqns are not
M=0 loops could give big IR corrections
N(t,k) ~ [Ha(t)/2kc]2 for every k
Perhaps their attraction stops inflation
Late modifications from vacuum polarization
Would affect large scales most
But for now, just model-building
Late-Time Acceleration
(arXiv:0705.0153 with Deser)
Nonlocality via
-½
for □ = (-g) ∂µ(√-g gµν∂ν)
Retarded BC {□-1 & ∂t□-1} = 0 at t=0
Act it on R □-1R dimensionless
L = √-g R[1 + f(□-1R)]/16πG
□-1
f(X) the “Nonlocal distortion function”
Gµν + ∆Gµν = 8πGTµν
∆Gµν = [Gµν + gµν□ - DµDν] (f + □-1[Rf’])
(ρ
σ)
ρσ
+ (δµ δν - ½gµνg ) ∂ρ(□-1R) ∂σ(□-1[R f’])
Specialization to FRW
ds2 = -dt2 + a2(t) dx.dx
R = 6Ḣ + 12 H2
t
t’
-1
3
□ = -∫0 dt’/a ∫0 dt’’a3
Two Built-In Delays
R=0 during Radiation Dom. (H = 1/2t)
5
No modification until t ~ 10 years
□-1R ~ -4/3 ln(t/teq) during Matter Dom.
10
□-1R ~ -15 at t = 10
years
Reconstructing f(X) for ΛCDM
(arXiv:0904.0961 with Deffayet)
How It Works
for slowly varying a(t)
Gµν + ∆Gµν = 8πGTµν ,∆Gµν ~ Gµν(f + □-1[Rf’])
Just rescale G! Geff = G/(1 + f + □-1[Rf’])
Friedman Eqn: 3H2 ~ 8πGeff ρ0/a3(t)
But Geff strengthens gravity!
Growth of Geff balances 1/a3(t)
Not relevant for solar system
But should increase structure formation
Dodelson & Park (arXiv:1209.0836)
Not purely Geff(t) when space dependent
Delayed so late that only ~10-30% effect
Screening
f(R) models have a problem
R > 0 for cosmology AND solar system
Need “screening mechanism” to suppress
response inside solar system
f(□-1R) models avoid this problem
2
□ ~ -∂t + ∂x
2
□-1R < 0 for cosmology
□-1R > 0 for gravitationally bound systems
f(X) = 0 for X > 0 means NO solar system change
Local Version Is Haunted
(Nojiri & Odintsov, arXiv:0708.0924)
R[1+f(□-1R)] R[1+f(Φ)] + Ψ[□Φ–R]
Varying wrt Ψ enforces □Φ = R
NB both scalars have 2 pieces of IVD
Ψ□Φ -∂µΨ∂νΦgµν
-½∂µ(Ψ+Φ)∂ν(Ψ+Φ)gµν
+ ½∂µ(Ψ-Φ)∂ν(Ψ-Φ)gµν
Ψ-Φ has negative KE
No new initial value data for
the original nonlocal version
Synch. gauge: ds2 = -dt2 + hij(t,x) dxidxj
IVD for GR: hij(0,x) & ḣij(0,x) = 6 + 6
4+4 for constrained fields
2+2 for dynamical gravitons
IVD in nonlocal cosmo count the ∂t‘s
2
-2
0
R ~ ∂t & □-1 ~ ∂t □-1R ~ ∂t
2
∆Gµν has up to ∂t □-1
Hence hij(0,x) & ḣij(0,x), but what are they?
t=0 Constraints same as GR
Recall Gµν + ∆Gµν = 8πGTµν
∆Gµν = [Gµν + gµν□ - DµDν] (f + □-1[Rf’])
(ρ
σ)
ρσ
+ (δµ δν - ½gµνg ) ∂ρ(□-1R) ∂σ(□-1[Rf’])
Retarded BC [□-1 & ∂t□-1] = 0 at t=0
f(X) also vanishes at X=0
Only [gµν□ - DµDν] {f(□-1R) + □-1[Rf’(□-1R)]} ≠ 0
Synchronous constraints ∆G00 & ∆G0i
g00□ - D0D0 = ½hijḣij∂t - ∆ 0 at t=0
g0i□ - D0Di = -∂0∂i + ½hjkḣki∂j 0 at t=0
No Ghosts check ∂t2 terms
Recall Gµν + ∆Gµν = 8πGTµν
∆Gµν = [Gµν + gµν□ - DµDν] (f + □-1[Rf’])
ρσ
(ρ
σ)
+ (δµ δν - ½gµνg ) ∂ρ(□-1R) ∂σ(□-1[Rf’])
Dynamical eqns Gij + ∆Gij = 8πGTij
2
gij□ - DiDj -hij∂t + O(∂t) -1
∆Gij = 2 hij R f’(□ R) + O(∂t)
kl
Rij = ½ḧij + O(∂t) & R = h ḧkl
+hij□ + O(∂t)
+ O(∂t)
kl
Gij + ∆Gij ½ḧij – [½-2f’(X)]hijh ḧkl + O(∂t)
No graviton becomes a ghost!
Still might have a potential energy instability
Conclusions about nonlocal
modifications of gravity
Last chance for gravity based on gµν
Not fundamental (we think)
Simplest models based on f(□-1R)
From QG corrections during primordial inflation
Purely phenomenological for now
Built-in delays explain cosmic coincidence
Simple f(X) reproduces ΛCDM without Λ
Model seems ok
No problem for gravitationally bound systems
Localized version has scalar ghost
But no new DoF’s in nonlocal version
Initial value constraints identical to GR
No graviton DoF’s become ghosts