Simulations of polarised dust emission

Similar documents
Magnetic field structure from Planck polarization observations of the diffuse Galactic ISM

Primordial B-modes: Foreground modelling and constraints

Measurements of Degree-Scale B-mode Polarization with the BICEP/Keck Experiments at South Pole

C-BASS: The C-Band All-Sky Survey. Luke Jew

Cosmic Microwave Background

The microwave sky as seen by Planck

CMB Foreground and separation methods. Astro 448 Yuxi Zhao 12/5/2018

Galactic radio loops. Philipp Mertsch with Subir Sarkar. The Radio Synchrotron Background Workshop, University of Richmond 21 July 2017

Dust polarization observations towards interstellar filaments as seen by Planck: Signature of the magnetic field geometry

Physics of CMB Polarization and Its Measurement

Polarised foregrounds (synchrotron, dust and AME) and their effect on the detection of primordial CMB B-modes

First scientific results from QUIJOTE. constraints on radio foregrounds

Scientific results from QUIJOTE. constraints on CMB radio foregrounds

Polarized Galactic foregrounds: a review

X name "The talk" Infrared

The CMB sky observed at 43 and 95 GHz with QUIET

arxiv:astro-ph/ v1 5 Aug 2006

Polarized Foregrounds and the CMB Al Kogut/GSFC

Toward an Understanding of Foregrounds in the BICEP2 Region

DES Galaxy Clusters x Planck SZ Map. ASTR 448 Kuang Wei Nov 27

Dust Polarization. J.Ph. Bernard Institut de Recherche en Astrophysique et Planetologie (IRAP) Toulouse

BINGO simulations and updates on the performance of. the instrument

Primordial gravitational waves detected? Atsushi Taruya

The international scenario Balloons, LiteBIRD, PIXIE, Millimetron

Galactic dust in the Herschel and Planck era. François Boulanger Institut d Astrophysique Spatiale

Magnetic Fields in the Milky Way

CMB component separation intution and GreenPol

Observational Cosmology

Delta-map method to remove CMB foregrounds with spatially varying spectra

Really, really, what universe do we live in?

A5682: Introduction to Cosmology Course Notes. 11. CMB Anisotropy

PLANCK lately and beyond

Multi-frequency polarimetry of a complete sample of faint PACO sources. INAF-IRA (Bologna)

Polarised synchrotron simulations for EoR experiments

PILOT. A far-infrared balloon-borne polarization experiment. Jonathan Aumont IRAP Toulouse, France

Cosmology Large Angular Scale Surveyor. Wednesday, September 25, 13

Short course: 101 level introductory course to provide a foundation for everyone coming to the workshop to understand the field.

PICO - Probe of Inflation and Cosmic Origins

Microwave Background Polarization: Theoretical Perspectives

arxiv: v3 [astro-ph.ga] 13 Sep 2018

Gravitational Lensing of the CMB

Cosmic Microwave Background

Detection of hot gas in multi-wavelength datasets. Loïc Verdier DDAYS 2015

Consistency tests for Planck and WMAP in the low multipole domain

Preparation to the CMB Planck analysis: contamination due to the polarized galactic emission. L. Fauvet, J.F. Macías-Pérez

News from BICEP/Keck Array CMB telescopes

Inferring source polarisation properties from SKA observations. Joern Geisbuesch

The first light in the universe

Constraints on primordial abundances and neutron life-time from CMB

Detection of B-mode Polarization at Degree Scales using BICEP2. The BICEP2 Collaboration

The CO Mapping Array Pathfinder (COMAP)

CMB Polarization and Cosmology

Rayleigh scattering:

Examining the Effect of the Map-Making Algorithm on Observed Power Asymmetry in WMAP Data

AME characterisation in the Taurus Molecular Clouds with the QUIJOTE experiment

astro-ph/ Oct 94

Delensing CMB B-modes: results from SPT.

Correlation of the South Pole 94 data with 100 µm and 408 MHz maps

Soma De Post-doctoral Fellow, Arizona State University presentation at 27th Texas Symposium, Dec 11,2013

Polarization from Rayleigh scattering

Rotation Measure Synthesis of the Local Magnetized ISM

Planck 2014 The Microwave Sky in Temperature and Polarisation Ferrara, 1 5 December The Planck mission

arxiv:astro-ph/ v1 22 Feb 1999

The cosmic microwave background radiation

LFI frequency maps: data analysis, results and future challenges

SPATIAL UNIFORMITY OF THE GALACTIC GAMMA-RAY EXCESS. Manoj Kaplinghat, UC Irvine

Using training sets and SVD to separate global 21-cm signal from foreground and instrument systematics

Power spectrum exercise

BipoSH Representation: Modulation & Systematics

An Estimator for statistical anisotropy from the CMB. CMB bispectrum

The Extragalactic Radio Background

The ultimate measurement of the CMB temperature anisotropy field UNVEILING THE CMB SKY

FAST AND EFFICIENT TEMPLATE FITTING OF DETERMINISTIC ANISOTROPIC COSMOLOGICAL MODELS APPLIED TO WMAP DATA

Dark gas contribution to diffuse gamma-ray emission

arxiv: v1 [astro-ph.co] 3 Feb 2016

An Introduction to Radio Astronomy

Low-order multipole maps of cosmic microwave background anisotropy

A joint analysis of Planck and BICEP2 B modes including dust polarization uncertainty

THE WMAP HAZE: PARTICLE PHYSICS ASTROPHYSICS VERSUS. Greg Dobler. Harvard/CfA July 14 th, TeV09

Hemispherical CMB power asymmetry: observation vs. models

Data analysis of massive data sets a Planck example

Imprint of Scalar Dark Energy on CMB polarization

Study of the very high energy gamma-ray diffuse emission in the central 200 pc of our galaxy with H.E.S.S.

arxiv: v3 [astro-ph.co] 5 May 2016

The FIR-Radio Correlation & Implications for GLAST Observations of Starburst Galaxies Eliot Quataert (UC Berkeley)

Challenges of foreground subtraction for (primordial) CMB B-modes

Cosmic Microwave Background Polarization. Gil Holder

Polarization simulations of cloud cores

Where is the COBE maps non-gaussianity?

Spectral Line Intensity Mapping with SPHEREx

arxiv: v1 [astro-ph.im] 1 Oct 2012

Large Scale Polarization Explorer

SUNYAEV-ZEL'DOVICH EFFECT WITH PLANCK SURVEYOR. Institut d'astrophysique Spatiale, B^atiment 121, Universitçe Paris Sud, F Orsay Cedex, France

Anisotropy in the CMB

SUPPLEMENTARY INFORMATION

CMB polarization and cosmology

FOUR-YEAR COBE 1 DMR COSMIC MICROWAVE BACKGROUND OBSERVATIONS: MAPS AND BASIC RESULTS

A Fast Algorithm for Cosmic Rays Removal from Single Images

arxiv: v1 [astro-ph.co] 2 Sep 2014

A Far-ultraviolet Fluorescent Molecular Hydrogen Emission Map of the Milky Way Galaxy

Transcription:

Simulations of polarised dust emission Flavien Vansyngel - Institut d Astrophysique Spatiale Collaborators: François Boulanger, Tuhin Ghosh, Andrea Bracco, Jonathan Aumont CMB-S4 meeting 03/016

Dust contamination Toward North pole Toward South pole Planck collab. PIPXXX 014 current upper limit for r BICEP field

Fake CMB B-modes Figure 1. Recovered posterior distribution P (r) of tensor-to-scalar ratio and impact of incorrect dust modelling. The oretical input tensor-to-scalar value (vertical solid black line) is r = 0 in left-hand panels and r = 0.05 in right-hand panels. Top panels: no foregrounds (left: Model 0a, right: Model 0b). Middle panels: correct foreground modelling (left: Model a, right: Model 1a). Bottom panels: incorrect spectral modelling of rmal dust (left: Model b, right: Model 1b). Recovered tensor-to-scalar distributions: COrE (solid yellow), COrE+ Light (solid light-blue), COrE+ Extended (solid blue), LiteBIRD (dotted red), PIXIE (dashed green), EPIC-LCTES (long-dashed yellow), EPIC-CS (long-dashed purple), EPIC-IM-4K (long-dashed orange), PRISM (dash three-dot black). The top left panel compares overall sensitivity of di erent satellites in absence of foregrounds by showing for Model 0a (r = 0, no foregrounds) r.m.s of residual noise B-mode map after component separation. Remazeilles, Dickinson, Eriksen & Wehus, 015

Topical questions What experimental design for optimal cleaning? What level of B-modes can be actually reached? How to quantify confidence in cleaning?

Topical questions What experimental design for optimal cleaning? What level of B-modes can we actually reach? How to quantify confidence in cleaning? Need sky simulations to test component separation methods

Log[power] Usual ISM dust model At map level a D spatial template scaled through frequencies using a grey-body law At power spectrum level a power law freq. Log[angular scale] Effective model

Actual ISM dust...

Actual ISM dust dust grains properties turbulence 3D magnetic field bright molecular clouds...

Toward realistic dust simulations Why is dust emission polarised? Dust grains are heated by star light They radiate rmal emission (microwave domain) grains are not spherical and aligned with ambient galactic magnetic field One dust grain produced in laboratory

STEP 1: Galactic magnetic field ~B gal = ~ B 0 ˆB0 + f m ˆBturb with random Model for mean magnetic field Relative strength of turbulent field 3D power index of turbulent field

STEP : Stokes parameters Q/I One random realisation of magnetic field U/I

STEP 3: depolarisation Q/I U/I + + Stack N layers of Q/I,U/I B1 B B3

STEP 4: dust Stokes parameters Q/I U/I I multiply by an external intensity map I Q U

TE correlation TE power Two-point statistics of model TE correlation BB/EE ratio Data PIPXXX 014 ratio 0.5 multipoles fsky Simulations N laye 7 0.8 -.6 ratio 1 multipoles multipoles

STEP 5: introduce correct statistics (I,Q,U)A such that: (I,Q,U)B such that: linear transform at alm level 8 >< b T`m = tat`m b >: È m = p 0(a È m + at`m ) b B`m = p 0fa B`m In particular: include E-B asymmetry and TE correlation

Properties of simulations BB power spectrum E-B asymmetry power data sims ratio 60 00 60 00 multipoles multipoles The simulations are able to reproduce data at power spectrum level Vansyngel, Boulanger & Ghosh (in prep.)

PDF of power spectra PDF of E power at multipole ~110 Residuals 0.006 0.01 0.018 very close to Gaussian Vansyngel, Boulanger & Ghosh (in prep.)

EE power at ell=80 EE power at ell=80 Properties of simulations A EE /hii.1 Empirical law A EE /hii 1.9 PIPXXX 014 mean intensity mean intensity Vansyngel, Boulanger & Ghosh (in prep.)

Case of synchrotron Suppose that we can apply same procedure with differences: 1. Magnetic field Same statistical model, but a different realisation from that of dust. Intensity map Haslam map with small scales added

Dust-Synchrotron correlation dust E x synch E 30% 50% 70% dust B x synch B 30% 50% 70% Common mean field not sufficient

Dust-Synchrotron correlation independent realisations: (Q/I,U/I)1 and (Q/I,U/I) dust[qu] = dustint (κ ([QU]/I)1 + (1-κ) ([QU]/I) ) sync[qu] = syncint ( (1-κ) ([QU]/I)1 + κ ([QU]/I) ) κ: ad-hoc input

Furr work is needed Full sky modeling Physical modeling of matter-polarisation correlation filament align with magnetic field Need high resolution noiseless intensity template Extrapolation through frequencies beta spatial variation angle decorrelation

se se a a Galactic Galactic magnetic magnetic field field model model Planck Collaboration: The local structure of Galactic magnetic field One-point statistics of model STEP 1 with 1 STEP with se a Galactic magnetic field model se a GalacticPolarisation magneticfraction field model Polarisation angle (direction of pola)fig. 10. spherical harmonic spherical harmonic decomposition Planck Collaboration: The local structure of Galactic magnetic field decomposition (Q+U )/I Nlayers Planck Collaboration: The local structure of Galactic magnetic field Fig. 10. Cartoon showing integration along line of sight of a modelled qc with four distinct polarization layers having a same value of fm and a same ordered-field direction. mean magnetic field mean magnetic field with 1 in which i = 1, and p = 11.89%. Thus, modelled p with results from Q Q +U U p =. (14) (D ) relative strength of turbulent field In bottom panel of Fig. 9 we show comparison between relative strength of turbulent field 0.4 histograms of p of data (black dots) and of model. In e M1 M M 353 M1 M of a mod same val particular, we present average over 0 realizations of model 3D power index of -1.5 3D power index of Nlayers spherical harmonic B (blue line) and corresponding 1 (bright blue shade) and (dark blue shade) variations. The dashed vertical line spherical harmonic decomposition refers to value of p = 11.89%. We notice that modelling p turbulent field decomposition allows us to nicely control level of noise in data. Indeed, turbulent field e mean magnetic field mean magnetic field Fig. 9. Top: histogram of about south Galactic pole 7 ROT (black dots), polarization angle inferred from Stokes parameters rotated with respect to best-fit uniform direction R of magnetic field (QR353 and U353 ). The error bars represent Poisson noise within each bin of histogram. The green line represents mean model B for fm = 0.4 over 0 different realizations. The green shades correspond to 1 (light green) and (dark green) variations of model. The dashed vertical line indicates no dispersion about uniform direction. Bottom: histogram of p obtained combining Yearmaps (black dots) as described in text for same pixels of top panel. The error bars represent Poisson noise within each bin of histogram. Model B is now in blue and it has same characteristics described in top panel. The dashed vertical line corresponds to an e ective polarization fraction of pe = 11.89%. in which results fr Planck Collaboration: we manage to recover all negative p values, which are only The local structure of Galactic magnetic field caused by noise in combination of Year-maps. Fig. 10. Cartoon showing integration From figure it is clear that our description of of a modelled qc with four distinct pola magnetic-field structure (using A and B) Fig. 1. Modelled histograms of models p (normalized to does unity not withsupp0 ) same value of fm and a same ordered-fie ply a satisfying of. The obtained around characterization south-galactic poledistribution from modelofc,p where data strong depolarization toward low p 1values, fm =show 0.9, aand value of N varies as follows: (dark which blue), is not seen in model, where distribution tends to peak at in which i = 1, and pe = 11.89%. (light blue), 7 (turquoise), 30 (yellow), 60 (orange), 100 (dark value of p. Moreover, large variance in data, also e results from red). In se models noise is not added. Q M1 Q M + U M1 found by Planck Collaboration Int. XIX (015) at intermediate pm = Galactic latitudes, produces a significant tail in distribution (D353 ) toward high values of p that is not reproduced by our model. In bottom panel of Fig. 9 we show t histograms of p of data (black do Making use of qc and uc, we replace qb and ub in Eq. (13), and 4.3. C: importance of modelled turbulencedistributions along line of particular, we present average over Q353 Model and U353 in Eq. (10), to get around sight B (blue line) and corresponding 1 south-galactic pole of p and ROT, given a magnetic-field and (dark blue shade) variations. T structure composed of an ordered field and LOS and POS turat this stage, it is important to keep in mind that observed refers to value of pe = 11.89%. We n bulent components. We stress that, thanks to model C, p in depolarization also depends on several factors that we havee not allows us to nicely control level of n Eq. (13) is now replaced by p. The modelled distributions deconsidered in modelling 0yet (see Eq. ()), such as variwe manage to recover all negative p pend on three parameters: p, f, and N. We fit data with 0 M ations of dust properties, encoded in p0, and fluctuations of caused by noise in combination of th model C structure exploringalong parameter of p0 between 15% and field LOS space and within Planck beam, From figure it is clear that o 40%, of fm between 1 and 17. parametrized by F. 0. and 1.8, and of N between magnetic-field structure (using models A For Being each triad of parameters we perform amagnetic-field -minimization of this work a specific study of ROT strucfig. 9. Top: about south Galactic pole ply Fig.a 1. Modelled histograms of of p (no histogram satisfying characterization combined reduced distributions of p -fit and -fit, ROT ture in Planck data at high-galactic latitudes, (black polarization dots), polarization angle inferred from Stokes pa- data obtained south-galactic po show around a strong depolarization towar as follows here we propose a phenomenological model (hereafter rameters rotated with respect to best-fitmodel uniform direction isfmnot=seen 0.9, and value of N varies in model, where distra R for variations R and C), which does of pu error southof not account magnetic field (Q ). The bars represent 0 across = +. (16) (light blue), 30 large (yellow 353 353 value of pe 7. (turquoise), Moreover, va tot p ROT Galactic pole, and which only LOS e ects. The green found Poisson noisefocuses within on each bingeometric of histogram. red). In models noise is not byse Planck Collaboration Int.added XIX relative strength of turbulent field relative strength of turbulent field 0.9 3D power index of turbulent field -1.5 3D power index of turbulent field Fig. 11. Same as in Fig. 9 but with modelled histograms now fm = 0.4. In to doing so Cwewith nowfmproduce two in corresponding model = 0.9, N = 7, andvariables p0 = 6% Eq. (6) where turbulent component is considered (hereafter (dashed-vertical line) (see text for definition of paramese two variables are qb and ub ). We n make four realizaters). tions of Planck statistical noise (nqi and nui, with i = 1, ), and, as in Eq. (7), we produce two pairs of independent samples data sims In bo histo figures from A. Bracco Ph.D sis

PDF of power spectra r cosmic variance total variance 30% 1%