Upper limits on neutrino masses from cosmology: new results Øystein Elgarøy (Institute of theoretical astrophysics, University of Oslo) Collaborator: Ofer.

Upper limits on neutrino masses from cosmology:

new resultsØystein Elgarøy

(Institute of theoretical astrophysics, University of Oslo)

Collaborator: Ofer Lahav (UCL, London)

SNOW06

NASA/WMAP Science Team

Room for neutrinos?

Absolute Masses of Neutrinos

E & Lahav, NJP 05

What do we mean by ‘systematic uncertainties’?

• Cosmological (parameters and priors)

• Astrophysical (e.g. Galaxy biasing)

• Instrumental (e.g. ‘seeing’)

Neutrinos decoupled when they were still relativistic, hence they wiped out structure on small scales

k > knr = 0.026 (m/1 eV)1/2 m1/2 h/Mpc

Colombi, Dodelson, & Widrow 1995

WDMCDM+HDM

CDM

Massive neutrinos mimica smaller source term

P(k)=A kn T2(k)

Neutrino Free StreamingNeutrino Free Streaming

P(k)/P(k) = -8 P(k)/P(k) = -8 m (Hu et al. 1998)m (Hu et al. 1998)

Neutrino mass from Cosmology

Data Authors Mmi

2dF (P01) Elgaroy et al. 02 < 1.8 eV

WMAP+2dF+… Spergel et al. 03 < 0.7 eV

2dF (C05)+CMB Sanchez et al. 05 < 1.2 eV

BAO+CMB+LSS+SNIa

Goobar et al. 06 < 0.5 eV

Ly- + SDSS+ WMAP (3 year)

Seljak et al. 06 < 0.17eV

WMAP (1 year) alone

Ichikawa et al. 04 < 2.0 eV

All upper limits 95% CL, but different assumed priors !

Example of “model” systematics: Dark energy

• Most cosmological neutrino mass limits have assumed that the dark energy is a cosmological constant

• There are (too!) many alternatives

• Common parameterization:

p = w where w is a constant (can be < -1)

Hannestad, PRL95 (2005) 221301

Why this degeneracy?

• P(k) sensitive to the combination fm

• But mh2 eV• If one allows for w <-1, SNIa data allow

large values of m

• The degeneracy is indirect, the effect of varying w on P(k) corresponds roughly to varying the amplitude

Modified gravity

TeVeS vs LCDM (D. F. Mota et al., in preparation)

Primordial power spectrum

The 2dF Galaxy Redshift Survey

• APM selected • Magnitude bJ < 19.45• Median z 0.1• 230 K measured• All public

• Cosmology• Galaxy properties

The 2dFGRS Team MembersI.J. Baldry, C.M. Baugh, J. Bland-Hawthorn, T.J. Bridges, R.D. Cannon, S. Cole, C.A. Collins,M. Colless (PI),W.J. Couch, N.G.J. Cross, G.B. Dalton, R. DePropris, S.P. Driver, G. Efstathiou, R.S. Ellis, C.S. Frenk, K. Glazebrook, E. Hawkins, C.A. Jackson, O. Lahav, I.J. Lewis, S.L. Lumsden, S. Maddox (PI),D.S. Madgwick, S. Moody, P. Norberg, J.A. Peacock (PI),B.A. Peterson, W. Sutherland, K. Taylor

http://www.mso.anu.edu.au/2dFGRS/

The 2dFGRS Power Spectrum

• 160 K (Percival et al.)

• Redshift space• Convolved• Good fit to CDM• Wiggles ?

l=200 l=1500

50 Mpc/h 7 Mpc/h

Empirical test of bias

• Red galaxies are more common in the centres of rich clusters than blue galaxies

• The opposite is the case in the rest of the universe

• Can get a feeling of the significance of bias for m limits by splitting the 2dF sample into red and blue galaxies

From Cole et al. 2005

Preliminary results

• Fix mh=0.18, fb=0.17, n=1

• Vary f, marginalize over amplitude and non-linear correction parameter Q (see Cole 2005)

• Look at various cuts in k

• Gives an idea of the importance of bias

ØE,OL,Percival, Cole & Peacock (in preparation)

ØE, OL, Percival, Cole & Peacock (in preparation)

Summary

• Cosmological neutrino mass limits start to probe the sub-eV range

• Need to focus on systematics• No data set can do the job alone• Dark energy: degeneracy with w understood

and can be dealt with, but not much has been done on “weird” models

• Bias: red and blue galaxies cluster differently, is it taken properly into account?