Optimised detection of SZ-clusters with PLANCK workshop “SZ-effect and ALMA“ Institut d’Astrophysique Spatiale, Orsay

Bj¨orn Malte Sch¨afer (MPA) [email protected]

Optimised detection of SZ-clusters with PLANCK – p.1/14 outline: SZ-clusters with PLANCK • try to cover every aspect of SZ-observations • SZ-physics: hydrodynamical simulations of clusters deviation from scaling-laws, asymmetry, non-isothermality • include (Galactic) foregrounds: synchrotron, dust, free-free, carbon monoxide, infra-red emission of planets and asteroids • SZ-detection with PLANCK non-uniform detector noise, scanning, beams • amplification and extraction of the weak SZ signal matched and scale-adaptive filtering, multifrequency observations, spherical geometry

thanks to: M. Bartelmann, C. Pfrommer, M. Reinecke, R. Hell

Optimised detection of SZ-clusters with PLANCK – p.2/14 thermal and kinetic SZ-maps

−51 −51 4 6 3

5 2 −50.5 −50.5 g] g] 1 [de 4 [de β β 0

−50 −50 3 −1 latitude latitude

−2 2 ecliptic ecliptic −49.5 −49.5 −3

1 −4

PSfrag replacements PSfrag replacements −5 −49 0 −49 99 99.5 100 100.5 101 99 99.5 100 100.5 101 ecliptic longitude λ [deg] ecliptic longitude λ [deg] thermal SZ kinetic SZ

• complex substructure, orbiting clusters • 50 clusters per square degree • halo-halo correlation, (evolving) mass distribution • velocities correspond to densities • deviation from idealised scaling relations

Optimised detection of SZ-clusters with PLANCK – p.3/14 Galactic foregrounds

free-free carbon monoxide

• Galactic foregrounds: dust and synchrotron • new foregrounds: Galactic free-free and carbon monoxide • infra-red emission from (moving) planets and asteroids • assumption: isotropic spectral properties (likely violated)

Optimised detection of SZ-clusters with PLANCK – p.4/14 sky at sub-mm frequencies 44 GHz 100 GHz 353 GHz 143 GHz

German virtual observatory (try yourself!): http://gavows2.xray.mpe.mpg.de:8080/planck/ Optimised detection of SZ-clusters with PLANCK – p.5/14 optimised filtering: basics • convolve sky map s(α) with filter kernel ψ[R](α)

∞ +` w(α, R) = dΩ s(θ)ψ(θ − α) = s`mψ`0[R] · Y`m(α) Z `X=0 mX=−`

• filters should meet 3 (2) requirements 2 2 1. variance σw(R) = 4π ` ψ`0[R] · C` is minimal at scale R0 P 2. hw(α, R0)i = 4π ` ψ`0[R] · s(`) is an unbiased estimator for the peak height sP(α)

3. there exists a scale R0, such that hw(α, R0)i = max • functional minimisation for obtaining scale-adaptive filter ψ(α) (1-3) and matched filter φ(α) (1+2) • pioneered for flat geometry and analytical profiles by Sanz et al.: ApJ 552:484 (2001), and Herranz et al.: MNRAS 336, 1057 (2002) (great read!) Optimised detection of SZ-clusters with PLANCK – p.6/14 matched and scale-adaptive filtering

0.5 1000

800 0.4

600

0.3 400 ) ,ν 0 θ ` ( ν ψ 0.2 ψ 200 filter

filter 0

0.1 −200 matched 0 matched −400

−600 −0.1 −800 PSfrag replacements PSfrag replacements −0.2 0 1 2 3 4 −1000 10 10 10 10 10 0 5 10 15 20 25 30 35 40 multipole order ` angle θ [arcmin]

Y`m-space real space

• extended to spherical geometry • gives filter kernel shape as function of • model profile (β-profile works reasonably well) • spectral dependence of signal (SZ spectral dependence) • power spectra Cν1ν2 (`) (45=36+9 spectra) Optimised detection of SZ-clusters with PLANCK – p.7/14 filtered maps: visual impression

−λ • filters optimised for various (θc, λ)-pairs (y(θ) ∝ (1 + θ/θc) ) • significance varies much! (try to get best lock on cluster) Optimised detection of SZ-clusters with PLANCK – p.8/14 number of recoveries

4 4 10 10

3 n 10 n

3 10

2 10 detections detections

1 2 10 10 1.4

1.2 32 1.2 32 16 16 1 1 PSfrag replacements 8 PSfrag replacements 8 0.8 4 0.8 4 2 2 slope λ 0.6 1 slope λ 0.6 1 core radius θc core radius θc matched filtering scale-adaptive filtering

• total number: 6000 to 8000 (above 3σ) • 25% lost when including Galactic foregrounds • watch out: figures not corrected for multiple detections!

Optimised detection of SZ-clusters with PLANCK – p.9/14 significances and position accuracy

4 10 180

160

3 10 140 σ )d σ ∆

( 120 n )d ∆ (

n 100 2 10 clusters

ution 80 of

60 distrib

1

number 10 40

20 PSfrag replacements PSfrag replacements 0 10 0 3 10 30 0 100 200 300 400 500 600 700 800 900 detection significance σ ∆ = θ2 2 squared distance arc [arcmin ] significances position accuracy

• scale-adaptive filter: many detections at threshold (3σ) • matched filter: highly significant detections • position accuracy: most of clusters within 100

Optimised detection of SZ-clusters with PLANCK – p.10/14 mass-redshift plane

0.8

0.7

0.6

0.5 z

0.4 redshift 0.3

0.2

0.1

PSfrag replacements 0 13.6 13.8 14 14.2 14.4 14.6 14.8 15 15.2 15.4 15.6 cluster mass log(M/(M /h))

• fairly well-defined region in M-z plane • population of low-M, z clusters not included in analysis

Optimised detection of SZ-clusters with PLANCK – p.11/14 distributions in mass and redshift

1000 2000

900 1800

800 1600 z m )d 700 )d 1400 z ( m ( n n 600 1200

500 1000 clusters clusters of

400 of 800

300 600 number number 200 400

100 200 PSfrag replacements PSfrag replacements 0 0 −0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 13.6 13.8 14 14.2 14.4 14.6 14.8 15 15.2 15.4 15.6 redshift z cluster mass m = log(M/(M /h)) redshift distribution mass distribution

• high-mass end fairly well sampled • no detection beyond z = 0.8 • large differences between filtering schemes

Optimised detection of SZ-clusters with PLANCK – p.12/14 SZ-observations: PLANCK and ALMA

• ALMA configuration for SZ-observations: compact, µK-sensitivity, 10 arcsec resolution, 1 arcmin fov • resolution: improvement compared to PLANCK by 30! • tiny fov → position of PLANCK clusters too coarse for follow-up • interesting topics: • cluster substructure, non-pressure equilibrium features • internal dynamics and turbulence via kinetic SZ

Optimised detection of SZ-clusters with PLANCK – p.13/14 summary and conclusion • PLANCK will yield a unique cluster sample with ∼ 8000 entries (more than Abell’s catalogue, larger than all X-ray catalogues) • very realistic simulation, observational + instrumental issues included • all-sky Sunyaev-Zel’dovich maps will be publicly available soon • 3 new foregrounds: carbon monoxide, free-free and planets/asteroids will be made public in the near future • fancy filtering: scale-adaptive and matched multifilters • possible to perform calculations with up to 5 × 107 pixels and ` ' 4096 with PLANCK simulation tools • paper submitted to MNRAS (see astro-ph/0407089 and astro-ph/0407090), 3rd paper to follow

SZ SKY IS GOING TO BRIGHTEN UP!

Optimised detection of SZ-clusters with PLANCK – p.14/14