Explosion rate of type Ia supernovae as function of redshift from SuperNova Legacy Survey

1. Explosion rate of type Ia supernovae as function of redshift fromSuperNovaLegacy Survey Tuesday lone supernovae presentation Moriond 2008 – La Thuile - Italy Pascal Ripoche – LPNHE - Paris Moriond 2008 - Pascal Ripoche

2. Outline • The SNLS • Measuring Rate in the SNLS • Photometric selection of type Ia supernovae • Type Ia Detection/Identification efficiency • The rate measurement Moriond 2008 - Pascal Ripoche

3. SuperNova Legacy Survey Moriond 2008 - Pascal Ripoche

4. The SuperNova Legacy Survey Main goal : measurement of cosmological parameters using type Ia supernovae. Method : • 5 years on 4 one square degree fields at CFHT (megacam + 4 filters) • Spectroscopic follow-up (VLT, Keck, Gemini) Results : about 300 type Ia supernovae up to z = 1.0 with spectroscopic identification during the 3 first years. Opportunity to accurately measure SN Ia rate vs z Moriond 2008 - Pascal Ripoche

5. Real time detection Pipeline ref detection new sub Increasing Sn Ia color Off-centered Good weather Spec time available  spectroscopy Spectroscopy: z = 0.81 Type : Sn Ia Day Moriond 2008 - Pascal Ripoche • 4 fields observed every 3-4 days in 4 bands (“rolling search”) • Variable object detection method : PSF matched image subtraction

6. Measurement of the Rate Moriond 2008 - Pascal Ripoche

7. How to measure supernovae explosion rate in the SNLS ? but Real Time spectroscopic identification efficiency is almost impossible to model due to spectroscopic selection (human selection, weather, available time, host luminosity,…).  pure photometric detection/identification A simple division : Observed type Ia supernovae sample Detection/Identification efficiency Moriond 2008 - Pascal Ripoche

8. Variable object detection Dataset : 2.5 years of SNLS images reprocessed First level selection : Variables objects • Images subtraction • Each detection is scored with a neural network and shapelet to remove artifacts • At least 5 good detections with S/N greater than 5. • Cut on field area and date to avoid edge effects. Selected objects are mostly physical objects : AGNs, variable stars, supernovae (all types). Spurious subtraction residuals Moriond 2008 - Pascal Ripoche

9. Selected objects 3051 variable objects Host galaxy photometric redshift : Ilbert et al. 2006 289 objects with spectroscopic datafrom real-time operations: redshift and type (not always)Used as control sample Redshift accuracy : 0.001 Moriond 2008 - Pascal Ripoche

10. SN Iaphotometric selection Level II Selection : Type Ia supernovae • Cut I : multi-color Lightcurve fit (SALT2) • Redshift fixed (spectroscopic or photometric) • Fitted parameters : stretch, color and brightness • Χ2 selection • Cut II: loose cut on brightness • gets rid of catastrophic redshift • minimize contamination Moriond 2008 - Pascal Ripoche

11. Test on control sample The control sample is used to test cuts (with host photometric redshift) • Blue : cut with chi2 • Red : cut on brightness • Selected objects Moriond 2008 - Pascal Ripoche

12. Results : Snia sample SNIa specSN specSN no spec • 435/3051 selected objects • No contamination with spectroscopic redshift • 2% Contamination with photometric redshift • Photometric redshift accuracy improved Moriond 2008 - Pascal Ripoche

13. Detection and selection efficiency Moriond 2008 - Pascal Ripoche

14. On image SN Ia simulation Main parameters: • Ra/Dec: coordinates • z : Redshift (from photoZ catalogue) • D0: Date of maximum • s : Stretch • c : Color (extinction) • disp : intrinsic dispersion • (α, β, Rv,…) Filter g’Filter r’Filter i’Filter z’ Lightcurves simulated with SALT2 Simulated supernovae added on all images (before subtraction)  Simulated images processed with the photometric detection pipeline Moriond 2008 - Pascal Ripoche

15. Detection/Selection Efficency 2.5 years of data processed (15 To) 720000 simulated supernovae • MC distributions tuned to match observed distributions: • Color • Stretch • Intrinsic dispersion • S/N • position on host I’ mag vega Moriond 2008 - Pascal Ripoche

16. SN Ia Rate measurement Moriond 2008 - Pascal Ripoche

17. Rate calculation Time dilatation correction Comoving volume Moriond 2008 - Pascal Ripoche

18. Corrections and Systematic Errors Corrections • Photometric Redshift dispersion • Photometric Redshift inefficiency • Contamination Systematic uncertainties • Efficiency calculation : distribution (stretch/dispersion/color) • Photometric Redshift dispersion • Contamination Moriond 2008 - Pascal Ripoche

19. Type Ia SN rate Photometric selectionSpectroscopic selection Rest frame Rate (z) assumed cosmology : LCDM (0.3,0.7) Moriond 2008 - Pascal Ripoche

20. Comparison with previous measurements Moriond 2008 - Pascal Ripoche

21. Rate and model • Rate = star formation rate (SFR)(Hopkins et Beacom 2006])∗ delay time function Φ(t) • σ/τ=0.2 τ=3.7±0.25 Gyrχ2=3.75 Slope measurement α=2.14±0.51 10-5 Sn/Mpc3/yr R0=2.0±0.54 χ2=4.77 Moriond 2008 - Pascal Ripoche

22. Conclusion • We have derived a photometric SN Ia rate vs z up to redshift 1.2 • Tested against a spectroscopic sample (40%) • Rate increase with redshift. Possible flattening around z = 1 • Good agreement with previously measured low and higher z rate (except with Barris and Tonry 2003) Moriond 2008 - Pascal Ripoche

23. The End Moriond 2008 - Pascal Ripoche

24. Taux spectroscopique (1) • Nouvelles coupures : sélection spectroscopique SNIa specSN specSN no spec Moriond 2008 - Pascal Ripoche

25. Taux spectroscopique (2) On suppose que l’absence de spectroscopie est aléatoire (temps disponible, méteo)  on estime l’efficacité spectroscopique : 69% ±10% On n’utilise que les supernovae Ia identifiées • Pas d’erreur sur les redshifts • Pas de contamination • Nouvelle systématique d’efficacité spectro Pas de spectroscopie pour z >1 Moriond 2008 - Pascal Ripoche

26. Moriond 2008 - Pascal Ripoche

27. (Z >1) Moriond 2008 - Pascal Ripoche 18 octobre 2007 - CPPM - Marseille 27