Testing Dark Energy with Supernova (and other cosmological probes)

1. Testing Dark Energy with Supernova(and other cosmological probes) Marek Kowalski Physikalisches Institut Universität Bonn16.9.2009, Szczecin

2. Content • Introduction to supernova cosmology • SNe observations & cosmological parameters • Constraints on selected models

3.  M 1998: Discovery of dark energy

4. Flat universe L + M = 1.01+/-0.02  • Weak lensing mass census • Large scale structure • Baryon Accoustic Oscillations M= 0.3 M 1998: Discovery of dark energy

5. WMAP 2006

6. Baryon Acoustic Oscillation (BAO) SDSS, Eisenstein et al. (2005)

8. Supernova Type Ia • Type Ia supernovae (SNe Ia) provide bright “standard candle” that can be used to construct a Hubble diagram. • Accretion sends mass of white dwarf star to Chandrasekhar limit leading to gravitational collapse and a thermo-nuclear explosion of its outer layers. • Each one is a strikingly similar explosion event with nearly the same peak intensity. 1

9. Astronomers think in… Magnitudes: m = -2.5 log(Flux) + constant Filter: B,V,R,I,Z (400-900 nm)

10. Stretching the timescale: Correcting the Brightness: “Standard” Candles • Nearby supernovae used to study SNe light curve (z<0.1) • Brightness not quite standard Intrinsically brighter SNe have wider lightcurves.

11. Spectra used for Identification & Redshift determination

12. Reference new picture difference Searching for Supernovae • One example - the SNLS: • Canadian-French-Hawaii Telescope: 3.6 m • MegaCam Camera: 3.6 108 pixels • 5-year program finding about 500 SNe Ia SN-Candidate

13. SNLS-Lightcurves

14. Sloan Digitial Sky Survey (SDSS) First year papers: arXiv:0908.4274, arXiv:0908.4276

15. Sloan Digital Sky Survey

16. SNe at large Redshifts (z>1) Observations from Space with the Hubble Space Telescopes: 15

17. fainter then expected  slightly brighter M Normalization SNe Type Ia & Acceleration of the Universe Supernova Cosmology Project Kowalski et al., Ap.J. (2008) •  M • 1 0 0.73 0.27 0 1 fainter

18. Analysis aspects • Consistent Lightcurve Fits • SALT fitter used for all SNe using mostly • original band-passes (Guy et al 2005/2007). • Blind analysis • The analysis (i.g. cuts) were developed on a • blinded data set, all luminosities were offset • by a hidden redshift-dependent amount. • Robust analysis • Initial cosmological fit using median statistics. • 3-sigma outliers removed. • Assignment of sample dependent dispersion. Large sample of SNe allows new studies of systematic errors Stretch and color corrected luminosity:

19. A heterogenous data sample

20. A heterogenous data sample • Study of • mean deviation • residual slope

21. Test for Tension high-z low-z mean deviation: OK

22. Test for Tension high-z low-z residual slope: (OK)

23. Evolution test: Redshift Evolution test: Population Testing SN evolution by subdividing the sample No significant evidence for evolution!

24. Nuisance parameters for systematic errors: sample dependent common for all z>0.2 SNe Systematic errors

25.  M M Results: Cosmological fit parameters Perlmutter et al., 1999 Combination of SNe with: BAO (Eisenstein et. al., 2005) CMB (WMAP-5 year data, 2008) For a flat Universe: … and with curvature:

26. With systematics Equation of state: w=p/ SNe + BAO + CMB ... and allowing for curvature:

27. w0+ wa= 0 Redshift dependent w w=w0 + (1-a) wa

28. A floating non-SNe bin to decouple low from high-redshift constraints Redshift dependent w Assuming step-wise constant w:

29. Constraining Dark Energy models

30. mn - ln(1+z) Constraining Models (I):Growing Neutrinos Scalar-field couples to massive neutrinos. Once neutrinos become sub-relativistic, one obtains -like behavior. Today: massive neutrinos and small offset of w from -1: C. Wetterich (2007), L. Amendola et al. (2007), Early dark energy (e) is second parameter and LSS and WMAP constrain e to be less then a few % (Doran et al. 2007). We assume a 10% linear growth prior. D.Rubin, E. Linder, MK et al., ApJ (2009)

31. Early dark energy e Constraining Models (I):Growing Neutrinos with sys error Lab constraints: mn2 eV Katrin sensitivity: mn 0.2 eV n-oszillations: mn0.05 eV 3s 2s 1s mn<1.2 (h/0.7)2 eV @ 95 CLstat error only mn<2.1 (h/0.7)2 eV @ 95 CL with sys error

32. Constraining Models (II):Braneworld Gravity Gravity leaking into extra dimensions on the scales of the Hubble radius can mimic cosmic acceleration. Dvali, Gabadadze, Porrati (2000) - DGP We chose dark matter and curvature as DGP parameters to obtain an effective Dark Energy equation of state:

33. Constraining Models (II):Braneworld Gravity DGP-Model versus LCDM: 2stat = 15.0 (stat error only) 2sys = 2.7 (with sys. error) D.Rubin, E. Linder, MK et al., ApJ (2009)

34. Constraining Models (III):Backreaction From a solution to the Buchert Eq: Perturbation expansion (if appliciable): n= -1 Boljakov et al. 2008 Larena et al. 2008 n =1.6±0.9 @ 68% stat error only Ωm=0.42±0.04 @ 68% stat error only

35. SNAP Future projectsfor Dark Energy Project z-range # SNe Pan-STARRS 0.1-0.5 ~104 LSST (2015) 0.1-0.9 ~106 SNAP (2018) 0.2-3.0 >3000 (JDEM/Euclid) • Other important future methods: • Weak lensing • Cluster rates • Baryon acoustic osciallation

36. Summary SNe as standard candles provide a direct measurement of the Acceleration history Combining SNe, BAO, and CMB we (slowly) become able to test individual models for dark energy Today, systematic uncertainties are of similar size as statistical uncertainties (some even say they are larger). Next generation surveys offer the chance to improve on both.

37. without the 8 new nearby SNe 184 SNe 192 SNe 307 SNe Union w/o New SCP SNe Riess Gold Davis 2007 Union Union Union Comparison with previous compilations