NEUTRINO COSMOLOGY

STEEN HANNESTAD UNIVERSITY OF AARHUS

LAUNCH WORKSHOP, 21 MARCH 2007

e

A BRIEF REVIEW OF PRESENT COSMOLOGICALDATA

BOUNDS ON THE NEUTRINO PROPERTIES

IF THERE IS TIME:BOUNDS ON NEW NEUTRINO INTERACTIONS

HOW DO NEUTRINOS INFLUENCE THE DETERMINATION OF OTHER PARAMETERS

OUTLINE

WHAT IS TO COME IN THE FUTURE?

WMAP-3 TEMPERATURE POWER SPECTRUM

THE COSMIC MICROWAVE BACKGROUND

SDSS SPECTRUMTEGMARK ET AL. 2006

Astro-ph/0608632

LARGE SCALE STRUCTURE

OTHER AVAILABLE STRUCTURE FORMATION DATA:

THE LYMAN ALPHA FOREST (THE FLUX POWER SPECTRUMON SMALL SCALES AT HIGH REDSHIFT)

McDonald et al. astro-ph/0407377 (SDSS)Viel et al.

THE BARYON ACOUSTIC PEAK (SDSS) – THE CMB OSCILLATIONPATTERN SEEN IN BARYONS

Eisenstein et al. 2005

WEAK GRAVITATIONAL LENSINGMassey et al. 2007

EISENSTEIN ET AL. 2005 (SDSS)

THE SDSS MEASUREMENT OF BARYON OSCILLATIONS IN THEPOWER SPECTRUM PROVIDES A FANTASTICALLY PRECISEMEASURE OF THE ANGULAR DISTANCE SCALE AND TURNS OUT TO BE EXTREMELY USEFUL FOR PROBING NEUTRINO PHYSICS

NEUTRINO MASSES ARE THE LARGEST SYSTEMATIC ERROR NOT ACCOUNTEDFOR IN THE ANALYSIS

017.0)94.01(98.0

469.035.0

fn

A

GOOBAR, HANNESTAD, MÖRTSELL, TU 2006

EFFECTIVE ANGULAR DIAMETER DISTANCE, A

THE LYMAN-ALPHA FOREST AS A TOOLFOR MEASURING THE MATTER POWERSPECTRUM

Ly- forest analysis

• Raw data: quasar spectra

remove data not tracing

(quasi-linear) Ly absorption

• Flux power spectrum PF(k)

hydrodynamical simulations + assumptions on thermodynamics of IGM

• Linear power spectrum P(k)

Power Spectrum of Cosmic Density Fluctuations

FROM MAX TEGMARK

SDSS BAO

NEUTRINO PHYSICS

LIGHTEST

INVERTED

NORMAL

HIERARCHICAL DEGENERATE

Tritium decay endpoint measurements have reached limitson the electron neutrino mass

This translates into a limit on the sum of the three mass eigenstates

(95%) eV 3.22/1

22 iei mU

eV 7im

Mainz experiment, final analysis (Kraus et al.)

em

THE ABSOLUTE VALUES OF NEUTRINO MASSESFROM COSMOLOGY

NEUTRINOS AFFECT STRUCTURE FORMATIONBECAUSE THEY ARE A SOURCE OF DARK MATTER

HOWEVER, eV NEUTRINOS ARE DIFFERENT FROM CDM BECAUSE THEY FREE STREAM

1eVFS Gpc 1~ md

SCALES SMALLER THAN dFS DAMPED AWAY, LEADS TOSUPPRESSION OF POWER ON SMALL SCALES

eV 932

mh FROM K2

11

43/1

TT

5/6

03 )()1(

1

)()(

f

nr

CDM

a

agf

kPkP

)(ag : DAMPING FACTOR (COMING FROM LAMBDA DOMINATION)

ITS POSSIBLE TO FIND AN ANALYTIC ESTIMATE OF THE SUPPRESSION (SEE E.G. LESGOURGUES & PASTOR 2006)

k < kfs

k >> kfs

THIS CAN BE APPROXIMATED WITH THE MORE WELL-KNOWN

)/81)((~)( MCDM kPkP k >> kfs

m = 0.3 eV

FINITE NEUTRINO MASSES SUPPRESS THE MATTER POWERSPECTRUM ON SCALES SMALLER THAN THE FREE-STREAMINGLENGTH

m = 1 eV

m = 0 eV

N-BODY SIMULATIONS OF CDM WITH AND WITHOUT NEUTRINO MASS (768 Mpc3) – GADGET 2

eV 9.6m 0m

T Haugboelle, University of Aarhus

256Mpc

COMBINED ANALYSIS OF WMAP AND LSSDATA (Spergel et al. 2006)

WMAP-3 ONLY ~ 2.0 eVWMAP + LSS 0.68 eV

COMPARE WITH WMAP-I:

WMAP-1 ONLY ~ 2.1 eVWMAP + LSS ~ 0.7 eV(without information on bias)

WHAT IS THE PRESENT BOUND ON THE NEUTRINO MASS?

STH, ASTRO-PH/0505551 (PRL)

THERE IS A VERY STRONG DEGENERACY BETWEEN NEUTRINOMASS AND THE DARK ENERGY EQUATION OF STATE WHEN CMB, LSS AND SNI-A DATA IS USED.THIS SIGNIFICANTLY RELAXES THE COSMOLOGICAL BOUND ONNEUTRINO MASS

HOW CAN THE BOUND BE AVOIDED?

IF A LARGE NEUTRINO MASS ISMEASURED EXPERIMENTALLY THISSEEMS TO POINT TO w < -1

DE LA MACORRA ET AL. ASTRO-PH/0608351

MAKING THE BOUND SIGNIFICANTLY STRONGER REQUIRES THE USE OF OTHER DATA:

EITHER ADDITIONAL DATA TO FIX THE m – w DEGENERACYTHE BARYON ACOUSTIC PEAK

OR

FIXING THE SMALL SCALE AMPLITUDELYMAN – ALPHA DATA

HOW CAN THE BOUND BE STRENGTHENED?

GOOBAR, HANNESTAD, MÖRTSELL, TU (astro-ph/0602155, JCAP)

NmbAnHwBM ,,,,,,,,, 010 FREE PARAMETERS

WMAP, BOOMERANG, CBISDSS, 2dFSNLS SNI-A

NmbAnHwBM ,,,,,,,,, 010 FREE PARAMETERS

WMAP, BOOMERANG, CBISDSS, 2dFSNLS SNI-A, SDSS BARYONS

sBM QNmbAnwH ,,,,,,,,,,, 012 FREE PARAMETERS

WMAP-3, BOOMERANG, CBISDSS, 2dFSNLS SNI-A, SDSS BAO

95% @ eV 3.2 m 95% @ eV 48.0 m 95% @ eV 62.0 m

WITH THE INCLUSION OF LYMAN-ALPHA DATA THE BOUND STRENGTHENSTO

IN A SIMPLIFIED MODEL WITH 8 PARAMETERS

No BAO

BAO

LY-

95% @ eV 45.02.0 m

BAO+LY-

USING THE BAO DATA THE BOUNDIS STRENGTHENED, EVEN FORVERY GENERAL MODELS

VERY SIMILAR RESULTS HAVE SUBSEQUENTLY BEEN OBTAINED BY CIRELLI & STRUMIA AND FOGLI ET AL.

SELJAK, SLOSAR & MCDONALD (ASTRO-PH/0604335) FIND

95% @ eV 17.0 m

IN THE SIMPLEST 8-PARAMETER MODEL FRAMEWORK WITH A NEW SDSSLYMAN-ALPHA ANALYSIS.NOTE, HOWEVER, THAT THIS DATA IS (EVEN MORE) INCOMPATIBLE WITHTHE WMAP NORMALIZATION.VIEL ET AL. FIND A DIFFERENT NORMALIZATION BASED ON A DIFFERENT ANALYSIS OF THE SAME DATA.

SELJAK, SLOSAR & MCDONALD

SDSS LYMAN-ALPHA

WMAP-3 NORMALIZATION

SELJAK, SLOSAR & MCDONALD (ASTRO-PH/0604335) FIND

95% @ eV 17.0 m

IN THE SIMPLEST 8-PARAMETER MODEL FRAMEWORK WITH NEW SDSSLYMAN-ALPHA ANALYSIS.NOTE, HOWEVER, THAT THIS DATA IS (EVEN MORE) INCOMPATIBLE WITHTHE WMAP NORMALIZATION.VIEL ET AL. FIND DIFFERENT NORMALIZATION BASED ON DIFFERENT ANALYSIS OF THE SAME DATA.

SELJAK, SLOSAR & MCDONALD

SDSS LYMAN-ALPHA

WMAP-3 NORMALIZATION

GOOBAR, HANNESTAD, MORTSELL, TU (ASTRO-PH/0602155)

OLD SDSS LYMAN-ALPHA

NEW SDSS LYMAN-ALPHA

VIEL ET AL. LYMAN-ALPHA

ADDITIONAL LIGHT DEGREES OFFREEDOM (STERILE NEUTRINOS)

STH & RAFFELT (ASTRO-PH/0607101)ANALYSIS WITHOUT LYMAN-ALPHA

LSND 3+1 UPPER LIMIT ON HEAVY EIGENSTATE OF ~ 0.6 eV AT 99% c.l.(0.9 eV AT 99.99%)

COMPARISON WITHSH & RAFFELT ’04

LSND 3+1EXCLUDED ATMORE THAN 99.9%

THE NUMBER OF NEUTRINO SPECIES – AN EXAMPLE OF WHY PRIORS AND THE PARAMETER EXTRACTION TECHNIQUE ARE IMPORTANT(USING VANILLA CDM + N+m) FROM STH, RAFFELT & WONG (TO APPEAR)

DATA SET MARGINALISATION MAXIMISATION

WMAP ONLY 12 (2.2 – 26) 3.7 (0.5 – 27)

WMAP+SDSS-DR3 6.2 (1.5 -11.7) 3.9 (0.8 – 18)

WMAP+2dF 3.1 (0.3 – 8.3) 1.2 (0 – 5)

WMAP+SDSS-LRG 4.4 (1.5 – 9.5) 2.5 (0.9 – 6.5)

WMAP+SDSS-BAO 4.1 (1.6 – 8.5) 3.0 (1.0 – 6.0)

WMAP+ALL LSS+

SNIa

3.6 (1.4 – 6.9) 3.0 (1.0 – 5.4)

WHAT IS THE DIFFERENCE BETWEEN MARGINALISATION ANDMAXIMISATION?

MARGINALISATIONFACTORS IN VOLUME IN PARAMETER SPACE

MAXIMISATION LOOKSONLY AT HOW WELLA MODEL FITS THE DATA

FINALLY: UNDERSTANDING NEUTRINO MASSES IS ALSOIMPORTANT FOR CONSTRAINING OTHER COSMOLOGICALPARAMETERS! (REVERSING THE PREVIOUS ARGUMENT)

EXAMPLE: ALLOWING FOR A NON-ZERO NEUTRINO MASS MAKESTHE SIMPLEST 4 INFLATIONARY MODEL COMPATIBLE WITH DATA! (HAMANN, STH, SLOTH & WONG ASTRO-PH/0611582, TO APPEARIN PRD)

m

WHAT IS IN STORE FOR THE FUTURE?

BETTER CMB TEMPERATURE AND POLARIZATIONMEASUREMENTS (PLANCK)

LARGE SCALE STRUCTURE SURVEYS AT HIGH REDSHIFT

NEW SUPERNOVA SURVEYS

MEASUREMENTS OF WEAK GRAVITATIONAL LENSINGON LARGE SCALES

Distortion of background images by foreground matter

Unlensed Lensed

WEAK LENSING – A POWERFUL PROBE FOR THE FUTURE

H

drPa

gHC m

0

2

240 ),/(

)(

16

9

H

dng

0

''

)'()'(2)(

FROM A WEAK LENSING SURVEY THE ANGULAR POWER SPECTRUMCAN BE CONSTRUCTED, JUST LIKE IN THE CASE OF CMB

),/( rP MATTER POWER SPECTRUM (NON-LINEAR)

WEIGHT FUNCTION DESCRIBING LENSINGPROBABILITY

(SEE FOR INSTANCE JAIN & SELJAK ’96, ABAZAJIAN & DODELSON ’03,SIMPSON & BRIDLE ’04)

WEAK LENSING HAS THE ADDED ADVANTAGE COMPARED WITHCMB THAT IT IS POSSIBLE TO DO TOMOGRAPHY BY MEASURINGTHE REDSHIFT OF SOURCE GALAXIES

HIGH REDSHIFT

LOW REDSHIFT

STH, TU & WONG 2006 (ASTRO-PH/0603019, JCAP)

THE SENSITIVITY TO NEUTRINO MASS WILL IMPROVE TO < 0.1 eVAT 95% C.L. USING WEAK LENSINGCOULD POSSIBLY BE IMPROVED EVEN FURTHER USING FUTURELARGE SCALE STRUCTURE SURVEYS

IF MEASURED AT ONLY ONE REDSHIFT THE NEUTRINOSIGNAL IS DEGENERATE WITH OTHER PARAMETERS

CHANGING DARK ENERGY EQUATION OF STATE INITIAL CONDITIONS WITH BROKEN SCALE INVARIANCE

HOWEVER, BY MEASURING AT DIFFERENT REDSHIFTSTHIS DEGENERACY CAN BE BROKEN

WHY IS WEAK LENSING TOMOGRAPHY SO GOOD?

fs

f

fs

kkg

zzg

kkg

zzg

kPzkP 5/62

2

)0(

)1/()(

)0(

)1/()(

)0,(),(

THIS REDSHIFT DEPENDENCE COMES FROM THE INTERPLAYBETWEEN THE POTENTIAL DECAY DURING -DOMINATIONAND NEUTRINO MASSES

IN PRINCIPLE IT PROVIDES A UNIQUE WAY OF IDENTIFYINGTHE PRESENCE OF MASSIVE NEUTRINOS IN STRUCTURE FORMATION

BY MAKING DIRECT MEASUREMENTS OF THE MATTER POWER SPECTRUMAT DIFFERENT REDSHIFTS IT IS POSSIBLE TO EXTRACT THE SAMEINFORMATION AS WITH WEAK LENSING TOMOGRAPHY(STH & WONG ASTRO-PH/0703031)

Power spectrum suppression for a neutrino mass of 0.08 eV

SENSITIVITY VS. DETECTION THRESHOLD FOR G1+G2+SG(ROUGHLY REPRESENTATIVE OF WHAT WILL BE AVAILABLE IN 10 YEARS)

ASTRO-PH/0703031

NOTE THE HIGHLY NON-GAUSSIAN NATURE OF THE LIKELIHOODCONTOURS

COULD NEUTRINOS BE STRONGLY INTERACTING?

BEACOM, BELL & DODELSON (2004) SUGGESTED A WAY TO EVADETHE COSMOLOGICAL NEUTRINO MASS BOUND:

IF NEUTRINOS COUPLE STRONGLY ENOUGH WITH A MASSLESSSCALAR OR PSEUDO-SCALAR THEY CAN BE VERY MASSIVE, BUTHAVE NO EFFECT ON THE MATTER POWER SPECTRUM,EXCEPT FOR A SLIGHT SUPPRESSION DUE TO MORE RELATIVISTICENERGY DENSITY

WHY? BECAUSE NEUTRINOS WOULD ANNIHILATE AND DISAPPEARAS SOON AS THEY BECOME NON-RELATIVISTIC

BEACOM, BELL & DODELSON 2004

HOWEVER, NEUTRINOS WHICH ARE STRONGLY INTERACTING DURINGRECOMBINATION ARE NOT AFFECTED BY SHEAR (EFFECTIVELY THEYBEHAVE LIKE A FLUID).THIS HAS IMPLICATIONS FOR CMB BECAUSE THE NEUTRINO POTENTIAL FLUCTUATIONS SOURCING THE CMB FLUCTUATIONS DO NOT DECAY

THIS INCREASES THE CMB AMPLITUDE ON ALL SCALES SMALLER THANTHE HORIZON AT RECOMBINATION

Sth, astro-ph/0411475 (JCAP)

Strongly interacting neutrinos

Standard model neutrinos

SUCH STRONGLY INTERACTING NEUTRINOS ARE HIGHLYDISFAVOURED BY DATA

(STH 04, BELL, PIERPAOLI & SIGURDSON 05 CIRELLI & STRUMIA 06)

ALTHOUGH THE EXACT VALUE OF THE DISCREPANCYIS NOT FULLY SETTLED (but at least by 2 > 20 even with more d.o.f.)

THIS CAN BE USED TO PUT THE STRONGEST KNOWN CONSTRAINTSON NEUTRINO COUPLINGS TO MASSLESS SCALARS OR PSEUDOSCALARS (STH & RAFFELT HEP-PH/0509278 (PRD))

WE FIND

211

7

)eV/ 05.0(101

101

mg

g

ij

ii

Diagonal elements

Off-diagonal elements

FOR DERIVATIVE COUPLINGS THE BOUND WOULD BECOME EVENSTRONGER

THE BOUND ON g CAN BE TRANSLATED INTO A BOUND ONTHE NEUTRINO LIFETIME

310 )eV 05.0/( s 102 m

THIS BOUND FOR INSTANCE EXCLUDES THAT THERE SHOULD BE SIGNIFICANT NEUTRINO DECAY IN BEAMS FROM HIGH-ENERGYASTROPHYSICAL SOURCES (STH & RAFFELT 2005)

CONCLUSIONS

NEUTRINO PHYSICS IS PERHAPS THE PRIME EXAMPLE OF HOW TO USE COSMOLOGY TO DO PARTICLE PHYSICS

THE BOUND ON NEUTRINO MASSES IS ALREADY AN ORDER OF MAGNITUDE STRONGER THAN THAT FROM DIRECT EXPERIMENTS, ALBEIT MORE MODEL DEPENDENT

WITHIN THE NEXT 5-10 YEARS THE MASS BOUND WILLREACH THE LEVEL NEEDED TO DETECT HIERARCHICALNEUTRINO MASSES

NOT ACCOUNTING FOR A NON-ZERO NEUTRINO MASS CANBIAS OTHER COSMOLOGICAL PARAMETER ESTIMATES