1. Cambridge CMB meeting 20 th July 2009 CMB B-modes: Foregrounds Paddy Leahy, Clive Dickinson, Mike Preece, Mike Peel (Manchester)

2. Cambridge CMB meeting 20 th July 2009 Polarized Foregrounds rms Q,U @ 1° E B, r = 0.1 3.4% anomalous dust 10% thermal dust QUIJOTE

3. Cambridge CMB meeting 20 th July 2009 FG separation strategy Adjacent bands give little leverage on spectral parameters –No point in having bands too close together Widely separated bands susceptible to subtle departures from simple spectral models (power law) –No point in having bands too far from CMB minimum If only the CMB had a spectral “feature”!

4. Cambridge CMB meeting 20 th July 2009 Synchrotron spectral are smooth! Power law is just an approximation… …but a good one The best-measured synchrotron sources are well fit by a 2 nd - order log-log polynomial over 2 decades of frequency

5. Cambridge CMB meeting 20 th July 2009 Cosmic ray spectrum CR energy spectra well known to be smooth over many orders of magnitude… … but dominated by baryons. What about the electrons & positrons that produce the synchrotron radiation? Simpson (ARNPS 1983)

6. Cambridge CMB meeting 20 th July 2009 Cosmic ray primary electrons Barwick et al (1998) –Electron energy spectral index 3.09±0.08 –Corresponds to β = −3.045±0.04 Moskolenko & Strong (1998) –Primary & secondary –Solar modulation at low energy –Model includes steepening of initial spectrum at 10 GeV, from 2.1 to 2.4, followed by energy loss.

7. Cambridge CMB meeting 20 th July 2009 Cosmic ray primary electrons Alpha Magnetic Spectrometer results –1998 Space shuttle flight (AMS collaboration, 2002) –Smooth! –Dominated by electrons, especially above 1 GeV

8. Cambridge CMB meeting 20 th July 2009 Fermi electron spectrum 5 20 100 350 GHz (B sinθ = 2.5 µG)

9. Cambridge CMB meeting 20 th July 2009 Fermi e − /e + results Apparent curvature in spectrum suggests new feature @ E > 100 GeV, perhaps related to increasing positron fraction in PAMELA data But with current calibration, data consistent with pure power law, p = −3.04 (i.e. β = −3.02) Synchrotron emitted in CMB band (< 300GHz) dominated by E < 100 GHz. TBD: assess impact of apparent curvature.

10. Cambridge CMB meeting 20 th July 2009 Synchrotron Polarization Synchrotron polarization varies with frequency for curved spectra (as expected in the Galaxy). Detail of variation depends on B-field geometry, dependence of electron energy on pitch angle. –Diagnostic of scattering efficiency. Degree of polarization vs. scaled frequency for “single burst” spectral ageing model (Leahy, Black & Chan in prep.)

11. Cambridge CMB meeting 20 th July 2009 Spectral Index 21:1.3 cm

12. Cambridge CMB meeting 20 th July 2009 WMAP I spectral index Gold et al (2009)

13. Cambridge CMB meeting 20 th July 2009 Spectral Index: 13:7 mm Low sensitivity in WMAP data at λ < 1.3 cm gives limited sky coverage Note flat spectrum for Crab nebula Mean β P ≈ −3.0 –Slightly flatter than at lower frequencies. (−3.1 in same regions) Kogut et al (2007) claim detection of flattening from β P ≈ −3.2 to −3.0 from WMAP data alone… –Use smoothing from 7° to 18° –No allowance for pol. bias at 23 GHz: artifact?

14. Cambridge CMB meeting 20 th July 2009 Spectral index distribution Kogut et al. (2007) Gold et al. (2009)

15. Cambridge CMB meeting 20 th July 2009 Fractional Polarization Minimum polarized intensity coincides with minimum 408 MHz intensity Typical fractional polarization at high latitude outside loop ≈ 10% = 75%/√N –Unless strongly contaminated…free-free? anomalous dust? N ~ 50: much line-of-sight structure in field direction, even straight up out of plane. If angle can vary on LOS, so can spectral index. Kogut et al (2007)

16. Cambridge CMB meeting 20 th July 2009 Loop I / North Polar Spur WMAP Haslam map 2° smoothing

17. Cambridge CMB meeting 20 th July 2009 Local & Distant B-fields North polar spur supposed to be at ~140 pc. Synchrotron scale height ~ 1 kpc Projected B-field angle the same in spur and in “diffuse” emission outside it!!?

18. Cambridge CMB meeting 20 th July 2009 Finkbeiner Davis & Schlegel (1999) #7: “Physical” model: –silicate grains, emissivity α = 1.5,  T  = 9.6 K –carbon (?) grains, α = 2.6,  T  = 16.4 K –α from lab measurements, T from fit to FIRAS data. #8: “Free-fit” model: –Emissivity indices allowed to float: α = 1.67, 2.70;  T  = 9.4, 16.2 K –Reduced χ 2 : 2.03  1.85 Fits exclude |b| < 7° Good evidence that cold component more dominant in HI vs H 2 clouds (NB composition not T dust !) –Composition or emission/abs properties –15% effect; not included in released FDS models. FDS #7 & #8: good fits to WMAP 94 GHz dust –outside mask –Model underpredicts by 26% (Gold et al) or 15% (my analysis). Grotesquely over-simplified?

19. Cambridge CMB meeting 20 th July 2009 Point sources 30 GHz 97 GHz 150 GHz 220 GHz r = 0.1 r = 0.01 r = 0.001 All > 1 Jy > 0.1 Jy > 0.01 Jy Toffolatti et al. (1998), Scuba IR counts (1% pol)

20. Cambridge CMB meeting 20 th July 2009 State of the art CMBPol foreground subtraction report –(Dunkley et al. 2008. arXiv:0811.3915): –We got away with it for the FIR Background –We have codes ready to run to do B-mode foreground separation. e.g: Codes that assume spectrum of each component is uniform over the sky (ILC, ICA) Codes that assume each foreground component has simple spectrum (e.g. power law) (FGFit/Miramare) –These assumptions known to be wrong: How wrong? How much difference does it make? Simulations in progress…

21. Cambridge CMB meeting 20 th July 2009 PSM Models Update to CMBPol report Now with latest dust polarization level: –Gives 5% at high latitude after geometric effects.

22. Cambridge CMB meeting 20 th July 2009 Summary Next 3 years should define problem –Planck HFI on thermal dust spectrum & polarization (not to mention BICEP) –Fermi + AMS on cosmic ray electron/positron primaries. What to minimise… –Sensitivity? Observe 60-150 GHz –Foreground uncertainty? Observe 200-350 GHz

23. Cambridge CMB meeting 20 th July 2009 Faraday Rotation

24. Cambridge CMB meeting 20 th July 2009 Faraday Rotation Away from Galactic plane, RMS Faraday rotation between λ1.3 cm and λ21 cm is 33° –< 3° at 6 cm –< 0.2° at 1.3 cm Significantly less than Faraday rotation of extragalactic sources –Diffuse synchrotron emission is mixed with ionized layer. PA differences between WMAP bands (22.5 – 33 GHz) suggest large Faraday rotation near Galactic Centre: –3°-4° at 1.3 cm, –RM ≈−700 rad m -2 A few pixels show up to 22° rotation between 22.5-33 GHz –Random errors (~ 5σ, but non- Gaussian) –Change of emission mechanism (dust polarization?) –Very large RM??