what we have learnt from planck (or how i learned to love...
TRANSCRIPT
What we have learnt from Planck
Wayne HuIAP, June 2019
(organizer’s title!)I learned(or how I learned to love tension)strange anomalies
curiositiesΛCDM?
slim pickings
What we have learnt from Planck
Wayne HuIAP, June 2019
I learned
What we have learnt from Planck
Wayne HuIAP, June 2019
(or how I learned to love tension)strange anomalies
curiosities
What we have learnt from Planck
Wayne HuIAP, June 2019
(or how I learned to love tension)anomalies
curiositiesΛCDM?
slim pickings^stop worrying and
Paradigm of Precision Cosmology.
-160 160 µK0.41 µK
tem
pera
ture
pola
rizat
ion
lens
ing
Planck (2018) I
• Precision measurementsand maps
– temperature
– polarization
– lensing
– 9 frequencies
• Control over systematics
– most recently polarization
• Accurate and precisetheoretical predictions
– Gaussian, adiabatic
– ΛCDM
Near Perfection in 6 Numbers.
10−3
10−2
10−1
100
101
102
103
D[µK2]
TT
EE
BB
TE
lensing
PlanckWMAPACTSPTACTPolSPTpol
POLARBEARBICEP2/Keck
BICEP2/Keck/WMAP/Planck
−200
−100
0
100
DTE[µK2]
2 150 500 1000 2000 3000 4000
Multipole
−0.5
0.0
0.5
1.0
1.5
107[(+1)]2C
φφ/2π
Planck (2018) I
• All this precisiondata described by6 ΛCDM parameters
– Ωch2: CDM
– Ωbh2: baryons
– θs: sound scale
– As: amplitude
– ns: tilt
– τ : reionization
• Measuredto sub percentprecision (except τ )
Predictive Power.
−100
−60
−20
20
60
100
∆D
[µK
2]
−8
−4
0
4
8
∆10
5C
[µK
2]
0 500 1000 1500 2000 2500
−8
−4
0
4
8
∆D
[µK
2]
Planck (2018) I
TE glitch
• Small residuals fromΛCDM in variousspectra
• Temp↔ polresiduals in ΛCDMwith reduced samplevariance
• Largely consistent,but with highprecision, moderatelysignificant deviations
• ∼ 2σ outliers, expectedbut some alsodrive parameters
Predictive Power• Predicts all other observables, which direct measurements test
0.24
0.25
0.26
YBBN
P
Aver et al. (2015)
Standard BBN
0.018 0.020 0.022 0.024 0.026
ωb
2.2
2.6
3.0
3.4
y DP
Cooke et al. (2018)
Planck TT,TE,EE+lowE
Standard BBN:
(Adelberger et al. 2011)(Marcucci et al. 2016)
Planck (2018) I,VI
0.5 1.0 1.5 2.0 2.5
0.90
0.95
1.00
1.05
1.10
(D/r d
rag)/(D
/r d
rag) P
lanck
6DFGS
SDSSMGS
SDSS quasars
WiggleZ
BOSSDR12
BOSS Ly-α (DM)
DES (DM)
DR14 LRG
−0.8
−0.4
0.0
0.4
0.8
JLA and Pantheon
Pantheon only
0.01 0.1 1
z z
z
−0.10
−0.05
0.00
0.05
µ−
µPlanck
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
0.2
0.3
0.4
0.5
0.6
0.7
0.8
fσ8
6dFGS
SDSS MGS
SDSS LRG
6dFGS+SnIa
GAMA
VIPERSBOSSDR12
WiggleZDR14 quasars
FastSound
BBN SNIa BAO
P(k) growth rate
baryon density
• Good agreement, even weak lensing, clusters, and yes H0 (< 10%)
Anchors Sink ΛCDM?• When distance ladder calibrated by CMB sound horizon, H0
discrepant with local measurements at 4.4σ (Riess et al 2019)
• Relative distances forward/backwards by ladder: CMB to BAO toSN isolating discrepancy as anchors (e.g. Aylor et al. 2018)
• Relative distances ∼ ΛCDM: little room for any new physics atintermediate redshifts to resolve
0.5 1.0 1.5 2.0 2.5
0.90
0.95
1.00
1.05
1.10
(D/r d
rag)/(D
/r d
rag) P
lanck
6DFGS
SDSSMGS
SDSS quasars
WiggleZ
BOSSDR12
BOSS Ly-α (DM)
DES (DM)
DR14 LRG
−0.8
−0.4
0.0
0.4
0.8
JLA and Pantheon
Pantheon only
0.01 0.1 1
zz z
−0.10
−0.05
0.00
0.05
µ−µPlanck
SNIa BAO
Soun
d H
orzo
n
Loca
l Anc
hors
0.0 0.5 1.0 1.5 2.0 2.5
56
62
68
74
H(z)/(1
+z) Local Anchors
Sound Horizon anchor tension
Planck (2018) I,VI
Driving in the Anchor• CMB anchor is sound horizon, must calibrate propagation time
• H(z < 103): with radiation, baryons fixed only Ωch2 unknown
• Ωch2 controls matter-radiation ratio and radiation driving from
potential decay due to Jeans stability
10 02515
Ψ
ks/π
damping
drivingtemperature
decay
Glitch and Oscillatory Residuals• Shifts in CMB anchor between low & high ` (Addison et al 2015)
• Low multipole glitch – deficit of power – looks like peaks shouldbe higher, more driving, less matter, higher H0
• High multipole oscillatory residuals: smoother acoustic peaks, lessdriving, more matter, lower H0
– also drives “lensing tension”: Pavel Motloch’s talk
−1.5
−1.0
−0.5
0.0
0.5
1.0
1.5
∆C
TT/σ
101 102 103
−15
−10
−5
0
5
−2∆
lnL
TT
ΛCDM TT ( < 1000)
Obied et al (2017)
glitchoscillatory
Planck Intermediate (2016) LI
penaltyΛCDM TT (all)
H0~70
Glitch and Large Angle Features• Exclude low ` < 30 glitch (Planck Intermediate 2016 LI)
• H0 falls, consistent with high multipoles within errors
• Alternately marginalize over possible features during inflation,mild ∆χ2 ∼ 12 improvement, but 5 params
• Likewise H0 returns to low value (Obied et al 2017)
−1.5
−1.0
−0.5
0.0
0.5
1.0
1.5
∆C
TT/σ
101 102 103
−15
−10
−5
0
5
−2∆
lnL
TT
ΛCDM TT ( < 1000)
SB TT
Obied et al (2017)
glitch
feature marginalized
oscillatory
Planck Intermediate (2016) LI
penalty
Driving and Oscillatory Residuals• Oscillatory residuals indicate smoother peaks (and persist even in
best fit ΛCDM)
• Driving sharpens the peaks
• Residuals indicate less driving, higher matter-radiation ratio
• Higher Ωch2, lower H0
10 02515
Ψ
ks/π
damping
drivingtemperature
decay
Driving and CDM• Signatures of CDM Ωch
2: TT amplitude and oscillatory residuals
• Polarization sharper test: projection effects for TT (Galli et al. 2014)
Obied et al (2017)
(+1)C
/2π
TT TE EE
3
−
−
0.
0.
2
−0.1
0.0
0.1
0.2
0.3
Ωbh2/σ Ωch
2/σ θMC/σ 500τ500τ
σ−1
dC
/dp
101 102 103
101 102 103
101 102 103
Polarization Signatures• Polarization sensitivity provides independent calibration using` < 1000 TE Planck data
• TE glitch at ` ∼ 165 enhances sensitivity, since lowering Ωch2
(raising H0) further raises predictions
Obied et al (2017)
−2
−1
0
1
2
∆C
TE
/σ
101 102 103
−15
−10
−5
0
5
−2∆
lnLT
TEE
ΛCDM TT ( < 1000)
SB TT
2σ glitch
Dark Radiation• Extra dark species whose energy density redshifts faster than
matter change sound horizon calibration
• Raise H(z < 103), lower sound horizon, raise H0 at fixed θs
• Additional driving from Jeans stable ∆Neff radiation compensatedby raising matter Ωch
2
10 02515
Ψ
ks/π
damping
drivingtemperature
decay
Driving and Damping• But damping provides second standard ruler in diffusion scale
• Random walk distance of photon scales as harmonic meanbetween horizon and mean free path
• Consistency check that ΛCDM passes and constrains anyadditional radiation ∆Neff
10 02515
Ψ
ks/π
damping
drivingtemperature
decay
Dark Exotica• Decreasing additional dark components during radiation
domination changes damping vs sound horizon
• Reconcile ratio if timed exactly right
0.00
0.03
0.06
0.09
10−5 10−4 10−3 10−2
a
frac
tion
of to
tal d
ensi
ty equality recombination
Lin et al (2019)da
mpi
ng se
nsiti
vity
horiz
on se
nsiti
vity
• Poulin et al (2018): specific anharmonic, periodic scalar potential
General Mechanism• Must compensate effect of raising Ωch
2 which reduces decay ofpotential
0.5
1.0
-(Ψ
+Φ
) aeq
-0.03
-0.02
-0.01
0.00
0.01
∆(Ψ
+Φ
)
k = 0.04/Mpc
10−5 10−4 10−3 10−2
a
raise CDM
General Mechanism• Dark exotica with relativistic sound speed, acoustic oscillations,
enhances decay
0.5
1.0
-(Ψ
+Φ
) aeq
-0.03
-0.02
-0.01
0.00
0.01
∆(Ψ
+Φ
)
k = 0.04/Mpc
10−5 10−4 10−3 10−2
a
ADE
4 Parameter Tuning• Tune the impact with sound speed (first dark acoustic peak) and
equation of state (redshifts away faster)
0.5
1.0
-(Ψ
+Φ
) aeq
-0.03
-0.02
-0.01
0.00
0.01
∆(Ψ
+Φ
)
k = 0.04/Mpc
10−5 10−4 10−3 10−2
a
sound speed
w
Reducing Tuning• Poulin et al (2018) 4 parameters, amplitude, time scale, c2
s and wmust be carefully tuned (top of oscillatory potential c2
s ↓)• Data favor something both more specific and generic: c2
s = w,transition to kinetic energy domination (Lin et al 2019)
~0.08 ρeq
PE: Hubble drag
KE: w=cs2
V,V
0.00
0.03
0.06
0.09acoustic DE
m φ2 2
10−5 10−4 10−3 10−2
a
locally
frac
tion
of to
tal d
ensi
ty
equality
Lin et al (2019)
Potential to Kinetic• With c2
s = w = 1 leaves 2 parameters: amplitude and slope ofpotential and requires kinetic energy redshift away (not oscillate)
• Amplitude ∼ 0.08ρeq and slope must be large enough to releasefrom Hubble drag
~0.08 ρeq
PE: Hubble drag
KE: w=cs2
V,V
0.00
0.03
0.06
0.09acoustic DE
m φ2 2
10−5 10−4 10−3 10−2
a
locally
frac
tion
of to
tal d
ensi
ty
equality
Lin et al (2019)
Fit and Predictive Signatures.
-1
-0.5
0
0.5
1
∆C
TT
/σ
CV
-1
-0.5
0
0.5
1
∆C
EE
/σ
CV
30 500 1000 1500 2000
-1
-0.5
0
0.5
1
∆C
TE
/σ
CV
Lin et al (2019)
• Fits joint data betterby ∆χ2 ∼ 12− 14
for 2-3 parameters
• Fits CMB itselfbetter, largely TE
• TE glitch ` ∼ 165
highly sensitive
• Dark componentredshifts away byrecombinationleaving nearlybare Ωch
2 signature
Potential Conversion of H0 Tension• Raises H0 to bring CMB and local anchors into better agreement
• Minor further improvements with additional parameters
• Posterior depends on parameter volume near ΛCDM, maximumlikelihood (ML) more reflective
• But mainly converts H0 question to “why this, why then”!
67.5 70.0 72.5 75.0H0
67.5 70.0 72.5 75.0H0
3 param2 param
4 param(oscil.)
ΛCDM
ML
Lin et al (2019)
Potential Conversion of H0 Tension• Raises H0 to bring CMB and local anchors into better agreement
• Minor further improvements with additional parameters
• Posterior depends on parameter volume near ΛCDM, maximumlikelihood (ML) more reflective
• Already limited by Planck TE polarization, distinguishing details
67.5 70.0 72.5 75.0H0
67.5 70.0 72.5 75.0H0
+POL
3 param2 param
4 param(oscil.)
ΛCDM
ML
Lin et al (2019)
Potential Conversion of H0 Tension• Raises H0 to bring CMB and local anchors into better agreement
• Minor further improvements with additional parameters
• Posterior depends on parameter volume near ΛCDM, maximumlikelihood (ML) more reflective
• EE residuals are ∼ 0.3 vs cosmic variance per multipole
-1
-0.5
0
0.5
1
∆C
EE
/σ
CV
30 500 1000 1500 2000
Lin et al (2019)
• Opportunity for testing ideas based on changing CMB anchor!
Summary• Planck and other precision CMB experiments have firmly
established ΛCDM as the standard model
• ΛCDM unreasonably effective and efficient in describing suite ofcosmological observables
• 6 numbers, mostly measured to sub percent precision, mostlyconsistent at this level with everything
• Tensions, anomalies and curiosities: imperfection is moreinteresting than perfection
• H0 at 4.4σ, can only be explained by changing one of the anchors
• CMB anchor is sound horizon and cross checked by damping scale
• Potential conversion illustrates designer difficulties, one or moreparameters per effect:
– look for predictive power of any explanation