Comments on Supernovae

• Riess 2004 sample of SNIa• Comments on SNIa systematics• Next SNIa surveys• Some Kosmoshow analysis of present SNIa data

Charling TAOApril 2004, Toulouse

SN Ia 2004 : Riess et al, astro-ph 0402512

Fits well the concordance model : 2= 178 /157 SNe Ia

183 SNIa selected Gold set of 157 SN Ia

SNIA 2004: Riess et al, astro-ph 0402512

* Low z : 0.01 < z < 0.15• Calan-Tololo (Hamuy et al., 1996) : 29 • CfA I (Riess et al. 1999): 22• CfA II (Jha et al, 2004b): 44 (not published yet)

16 new SNIA with HST (GOOD ACS Treasury program)

6 / 7 existing with z >1.25 + Compilation (Tonry et al. 2003): 172 with changes…

* Knop et al, 2003, SCP : 11 new 0.4 < z < 0.85

reanalysis of 1999 Perlmutter et al.

* 15 / original 42 excluded: inaccurate colour measurements and uncertain classification

* 6 /42 and 5/11: fail « strict  SNIA » sample cut

* Barris et al, 2003, HZT: 22 new varying degrees of completeness on photometry and spectroscopy records

* Blakesly et al, 2003 : 2 with ACS on HST

Determination of Cosmological parameters

Riess et al, astro-ph 0402512

w=p/

w= w0+w’ z

Determination of acceleration

Riess et al, astro-ph 0402512

New physics?

• Cosmological constant• Dark Energy: Dynamical scalar fields,

quintessence…. General equation of state p=w r r = R-3(1+w)

Perhaps a bit early !!!

« Experimentalist » point of view…

Constraints on cosmological parameters

m= 0.2 - 0.3 effect!

Systematic error on magnitude

3 fit with no prior

20% calibration error on intermediate fluxes gives no cosmological

constant

Use Kosmoshow:

an IDL program by A. Tilquin!

marwww.in2p3.fr/~renoir/kosmoshow.html

Riess gold set sensitivity

Kosmoshow, A. Tilquin

A m=0.27 shift of low z data

No need for But Universe is not flat!

Shift z <0.15 data by m= 0.27

m= 0.43 +/-0.2 and = 0 +/-0.34

Use Kosmoshow: an IDL program by A. Tilquin!

Systematic differences between methods

A 3 steps method: o Discovery: subtraction of an image

with a reference one.

o Supernova type identification and redshift measurement: spectrum.

o Photometric follow-up: light curve.

Final analysis: Hubble diagram.

The “classical” observation method

m(z) = M + 5 log (DL(z,M,L))-5log(H0)+25

The Hubble diagram

Absolute magnitude

Less luminous/z =>Accelerated expansion less matter or more dark energy

Too luminous/z =>Slowed down expansion => decelerationMore matter, less dark energy

– Light Curve in local reference frame – K correction– Galactic extinction correction

- Standardisation methods : stretch (SCP), MLC2k2 (HiZ), m15, ...

mag

Magnitude at maximum

light curve

Standardisation: stretch method

Before: mB After : mBcor = mB – (s-1)

Precision on the magnitude dominated by intrinsic dispersion:

mint 0.15

Stretch uncorrected

Stretch corrected

Precision on the magnitude at the maximum

Fit cosmological parameters

– From Hubble diagram, fit best cosmological model agreeing with observations.

– Determine dark energy parameters , ou (X, w, w’)

and matter density M

Spectroscopy needed

• SN Ia Identification– Spectrum structure

• Redshift z measurement– From position of identified

lines from spectra SN and/or underlying galaxy

data analysis physics

The « classical » method

Images

Spectra

+ identification.

Ia

magnitude z(redshift)

galaxy

Hubble

Systematic effects

Extragalacticenvironment

Supernovaenvironment

reduction/correlationsSNIa contamination

Selection bias Inter calibration filters

local

Normal Dust absorptionLensingGrey DustSN evolution

Systematic effects

• Observational problems– Standardisation method– Light curve fitting

– Subtractions– Calibrations– Atmospheric corrections– K-corrections– Selection bias

– Heterogeneity of SN data– SNIa identification

• SN evolution

• Internal extinction not negligible in spiral galaxies

• Corrections for peculiar velocity effects

• Grey dust• Lensing

Rowan-Robinson astro-ph/021034

Perlmutter & Schmidt 0303428

Knop et al (2003) light curves

Spectrum is dilated by (1+z) :The integrated flux in a filter is Shifted. Filters responses are not flatSometimes, need different filtersCorrect for differences systematic effects

Le flux est intégré sur unfiltre pour un point dephotométrie

Redshift calibration

SNIa sample contamination

Need strict selection criteria Gold sample is probably well selected

Supernovæ identification

With SpectraMain stamp of the SNe Ia: Si II at 6150 Å:

o Hardly observable beyond z > 0.4-0.5.

Otherwise, search for features in the range 3500-5500 Å (supernova rest frame):

o Ca H&K, SiII at 4100 Å, SII, …

Ca H&K SiII 4100

Simulation of a SN Ia spectrum at

z0,5

Spectroscopy : a Supernovae

Atmospheric transmission (ground)

Reduction of transmission in visible Absorption water & O2 reduce visibility in IR .

Reduced efficiencyNot homogeneous filtersRedshift dependent !!!

Seeing+ weather+ moon+ field not always visibleabsorption

mmsz

arcmms

arcmms

///1050. :light maximumat SNIa 7.1

sec////1050. : Zodiacal

sec////10602 :Continuum

22

222

222

Les Atmospheric emission

Spectroscopy : Subtract galaxy

Dependence on SN Environment

Blue have a lower metallicity Can be seen further

Supernovae evolution Peak magnitude can change

– Explosion changes with environment– Difference of chemical elements around SN– Depends on galaxy morphology, age, type,…

Sullivan et al (2002) SCP

SNIa host galaxy morphological classification

Not a large effect, but statistics are low

Extinction and Dust

• Extinction by dust from Our or SN galaxy

Rv=3.1 +/- 0.3 for OUR galaxy Very large correction

• Grey dust: not well known, intergalactic,?

Before extinction

After correction

Correction factor to the magnitude

A = R* E(B-V) Measurements in many filters Select minimal dust regions ?

A strong limit on grey dust?

• A 24.7 hr Chandra exposure of QSO 1508-5714 z=4.3 shows no dust scattering halo

• Upper limit on mass density of large grained (>1m) intergalactic dust: dust < 2 10-6

Peerels, Tells, Petric, Helfand (2003)

Dust and evolution ?

Evolution: shift due to progenitor

• mass?• metallicity?• Ni distribution?• Other effects?

Dust :Homogeneous gray intergalactic dust?Galactic dust responsible for extinction?

Sensitivity to dark energy decrease for z > 0.6

Is there a region of deceleration? Needs to go to z> 1

Gravitational Lensing

Some estimates of Systematics

Effect of de/amplification

Systematics

Understand environmentTo classify and correct

Need precise measurementswith statistics

Perlmutter

SN demographics studies

Summary

Ideally• Many SN for a negligible statistical error and study of systematic conditions. wide field

• Detect deceleration zone (z>1) measure in IR (or have large local UV sample for SNIa identification) • Control the correction precision for SNIA standardisation (environment and measurement corrections)

• Control non corrected systematic effects to the same level Very precise light curves and spectra to determine the explosion parameters, at all distances.

space

Ground limitation at z around 1 due to atmosphere

ground simulation

Hubble diagrams: Space vs ground

Advantage of space

•More galaxy surface density•Less impact from a more constant PSF• More information on shape

same observation in space and from

ground

Optimisation of mission

SNAP a dedicated satellite

Large statistics: 2000 Sne Ia/yr redshift to z<1.7, Minimal selection Ia identification

2m wide field telescope

Science

• Measure M and • Measure w and w (z)

Data Set Requirements

• Discoveries 3.8 mag before max• Spectroscopy with S/N=10 at 15 Å bins• Near-IR spectroscopy to 1.7 m

Statistical Requirements• Sufficient (~2000) numbers of SNe Ia• …distributed in redshift• …out to z < 1.7

Systematics Requirements

Identified and proposed systematics:• Measurements to eliminate / bound each one to +/–0.02 mag

Satellite / Instrumentation Requirements

• ~2-meter mirror Derived requirements:• 1-square degree imager • High Earth orbit• Spectrograph • ~50 Mb/sec bandwidth (0.35 m to 1.7 m)

•••

•••

Mission : % level

Need same precision on extracted magnitude Fit the magnitude on light curve after corrections of stretch, galactic extinction, K-corrections, everything that modifies luminosity Study models and parameter extraction Determine camera properties

reach 1 to 2 % on cosmological parameters

SNAP goals

data analysis physics

SNAP: Observation method

Images

Spectra

+ same spectra, allows identification.

SiII Ia

magnitude

M ,

z(redshift)

galaxy

Hubble

The same !! But optimised for systematics

Discovery maximum 2 days (RF) after explosion ( max + 3.8 magnitude),

Ligth curve: At least 10 points in photometry until plateau (+2.5 m)

Spectrum very precise at maximum (identification, systematics, calibration)

SNAP SNIa strategy

Hubble Deep Field

Weak Lensing Survey

Supernova Survey

Surveys:• Supernova Survey:

• 2X7,5 sq. deg.• 2X16 months • R<30.4 (9 bands)

• Weak Lensing Survey• 300 sq. deg.• 0.5-1 year• R<28.8 (9 bands)

Each field is est observed ~4 daysAll images are cumulated

Observe repeatedly same

sky area

SNAP survey

Wide field !!

SNAP: control evolution systematics

Light curves

Multi band PhotometryPeak measurement 2 %K correctionSelection biasVery precise measurement of beginning and end of light curve

Simulated SNAP Light Curvesz=0.8 z=1.0

Rest R-band

Rest B-band

Rest V-band

z=1.2 z=1.4 z=1.6

Rest B-band

Rest V-band

SNIa Spectra

Wide lines!SII 5350Å, w = 200Å

SII “W”, w = 75Å

SiII 6150Å, w= 200Å

Study of spectra and correlation of line variations with explosion parameters and luminosity

Need MODELS

Quantification of systematics

Metallicity effect

Velocity differences

DataModels

Modelisation of explosion (T, v, M)

Control of evolution

Present errors on :(flat universe case)

statistics 0.085systematics (determination SCP)

Malmquist bias 0.04K correction/Calibration 0.025

Extinction by ordinary dust 0.03Extinction (galactic) 0.04Non SNIa 0.05

Gravit. Lensing <0.06

not determinedgrey dust ?SNIa evolution ?

SNAP

2000 SN

Detection at explosion

Adjust filters in B+ intercalibration

spectra colours SDSS/SIRF Spectra Id

Average on many SN

spectra NIR + z>1spectra z>1

A method for each systematics

Résultats-diagramme de Hubble SNAP Expectations

SNAP expected results

Weak Lensing + CMB

How to constraint systematic effects and get precise measurements?

• Ideally in space: SNAP/JDEM

Problem: > 2014

• In the meantime: More statistics from as homogeneous samples as possible

CFHTLS and ESSENCE + Nearby

Low z activities

•Nearby SuperNova Factory

–300 SNIa (2004-…)

–http://snfactory.lbl.gov/

•Physics of SNIa explosions

•Supernovae at CfA (ongoing…)

–Expect ~ 100

–http://cfa-www.harvard.edu/cfa/oir/Research/supernova.html

Low z: Nearby Supernova Factory (2004-…)

• Goals– ~100/yr 0.03<z<0.08– 10 spectro-photometric

between –14D and +40D – Spectra: 320-1000 nm

• Tools– Discovery: Two cameras (one

wide field) 1.2 m ground based telescopes: NEAT

– Lightcurve follow-up with YALO – Photo-spectro follow-up with Field

Integral spectrometre (SNIFS) for ground based 2.2m

telescope (Hawaii)

• Collaboration– France: CRAL,IPNL, LPNHE– US: LBNL, U.Chicago

Intermediate z (2003-2014)

• ESSENCE at CTIO– ~50 SN Ia/year– http://www.ctio.noao.edu/wproject/sne/

• SNLS with MEGACAM of CFHT Legacy Survey– MEGACAM working since march 2003– http://snls.in2p3.fr/– Foreseen : 700 SNIa z < 1.

The CFHT Legacy Survey Supernovæ Program

                            

SNLS : the instruments

• A wide field camera (1 square degree, MEGACAM 0.35 Giga pixels) on 3.6 m CFHT (Hawaii) telescope

The Deep Survey of the CFHTLSCharacteristics:

o 4 fields of 1°x1° (RA=2h, 8h, 14h and 22h).

o Each field observed every 2-3 dark/grey nights (6 months/year).

o Each field observed with different filters:

[u’ (15 min)], g’ (15 min), r’ (30 min), i’ (60 min) and z’ (30 min).

o ~200 dark/grey nights exclusively dedicated to the survey (5 years).

o Seeing < 0.9 arcsec.

Pre-survey started in: March 2003.

Scientific objectives of the Deep Survey:

o Confirm acceleration with statistical significant study of systematics

o Characterization of the equation of state of the Dark Energy .

o Evolution of galaxies and quasars.

o Detection of transient phenomenæ.

o…

SNLS: Detections

Simulation of a SNIa light curve at

z=0.49

Multiplexing: detection and follow-up on the same image.

Light curve:

o Usable between 0.3 < z < 0.9: about 700 SNe Ia in 5 years.

o Between [-10,+15] days in the SN rest frame.

o Multi-wavelength: [g’], r’, i’ et z’.

Spectroscopic trigger: estimate of the magnitude and the date of the maximum.

Photometric follow-up

The spectroscopic program

Supernova type identification and redshift measurement.

Require 8-10 meters class telescope.About 12 SNe/field/lunation to be identified

Telescope allocation: Large Program on the ESO FORS/VLT:

240 hours spread over 4 semesters

Service mode.

Gemini:

3 Canada, 2 UK and 1 US nights/year requested.

Service mode.

Keck:

Visitor mode.

4 nights/year requested

Beginning of the pre-survey:o March 2003.

As expected, data not optimal at the beginning (engineering time):

o New optics (Megaprime), new camera (Megacam), new softwares, …

Dedicated spectroscopy program started July 2003 ~ 50 well measured SNIa todate!

The CFHT LS pre-survey and survey

Reference

Image

Subtraction

Sn2003fh: SN Ia at z=0.25

R6D4-9

Candidate

Ia:

z = 0.94

Age = -1 days

Preliminary

The CFHT Legacy Survey Supernovæ

Program

Canada and France

Extra collaborators for the spectroscopy:

VLT: ESO, Portugal, Sweden, UK.

Keck: US.

Gemini: UK, US.

Simulation after a 5 years survey

=0.72 and M=0.28.

SNFactory (300 SNe)

CFHTLS (700 SNe)

A new generation of Hubble diagram

SNLS : expected results

contraint

contraint

SN only : ~0.1 and w~0.2

limited to z<0.95 (atmosphere)

Flat

Only statistical errors

68 %

Comparison with present measurements

SNLS present conclusions

The CFHT LS /SNLS , a high redshift supernovæ factory:

o Survey started officially in August 2003.

o Sample increased by ~10: 700 Sne between 0.3<z<0.9.

o Very good quality and homogeneity of the data.

o Systematic errors at high redshift better controlled.

o Measurement of the w parameter at w 0.1.

o…

First results soon! (Already > 50 well measured SNIa)

Present situation:183 SN from Riess 2004

Astro-ph0402512

Curve for Gold sample

Fit, for a full sample, no prior

•Simulation and analysis tool:

Kosmoshow developed in IDL by André Tilquin (CPPM)

Kosmoshow analysis

marwww.in2p3.fr/~renoir/Kosmoshow.html

Riess 2004, gold sample

m=0.445 0.105

X=0.972 0.190

k=-0.418 0.283

Large presentr ?

Blanchard et al., (+ others) large r (0.1-0.2) for CMB possible

Riess gold sample radiative component

Simulation and analysis tool:

Kosmoshow (A.Tilquin)

Add r in DL equation: (1+z)4

component

Strong correlation r with M

Positive r component not excluded!

BUT always NEED large

Full fit on gold sample

T=1±0.1

Flat Universe

m=0.27±0.04

How can present r0 be large?

• Expected r0 = (1 + f N) = 5.06 10-5 h2

70

N = number of relativistic species Trec~ 0.26 eV

f = numerical factor = 0.227 for neutrinos

Expected (low z) r0 ~< 10-4 !!!With (1+z)4 evolution r,z=1100 is 103 higher than m with (1+z)3

evolution

Bias from the time evolution of the equation of state

astro-ph/0403285 Virey et al.

•Quantitative analysis of the bias on the cosmological parameters from the fitting procedure, ie, assuming a constant w, when it is not!

•With present statistics, can be ignored

Not the case with larger samples!

Example of bias: large w1

• w0F=-0.7

• w1F= 0.8

Suggestion Maor et al...

4-fit

Ms, M, w0 , w1

3-fit,

Ms, M, w0

Comments on Supernovae: a summary

• Riess 2004 sample of SNIa: a strict selection• Comments on SNIa systematics. Not all understood!• Next SNIa surveys• Some Kosmoshow analysis of present SNIa data

Charling TAOApril 2004, Toulouse