Study of Foregrounds and Limitations to EoR Detection

1. Study of Foregrounds and Limitations to EoR Detection Nithyanandan Thyagarajan N. Udaya Shankar Ravi Subrahmanyan (Raman Research Institute)

2. Outline • Estimates of uncertainties in EoR power spectrum • Classical Radio Source Confusion • Sidelobesdue to Frequency dependent beams (mode-mixing) • Thermal Noise • Understanding effects of array configuration on EoR sensitivity • Understanding effects of bandwidth • Conclusions

3. Framework of our Study • Radio Source Distribution • 128-tile MWA Layout • Relations Hopkins et al. (2003)

4. Classical Source Confusion • Solid Angles in (l,m) domain

5. Classical Confusion in k-space • Consider zenith pixel • Smooth variation along frequency of residuals • Delta function at k|| = 0 • Array configuration determines variation along • Lies outside of EoR window • Bandpass spillover into EoR window k

6. Response in u-v Annulus • Consider as point source of strength σC • Uniform response across all baselines • Integrate coherently over bandwidth

7. Sidelobe Confusion • Classical confusion is the source of sidelobes • Weighted by primary beam • Sidelobe confusion at zenith comparable when Brms=Nbeams-1/2 • Sidelobes have frequency structure (results in mode-mixing)

8. Modeled Primary Beam

9. Mode-mixing Principle Bowman et al. (2009) ηcont = umax l / f Vedantham et al. (2011) Transverse structure of contamination translates to a line-of-sight structure due to mode-mixing l-f invariance

10. PSF Decomposition • Fourier Basis Coordinates (u,v) are linear & independent • Net PSF conveniently decomposed into component PSF sensitive to specific k

11. Sidelobe Confusion in k-space k|| m v IFT k l u FT

12. Sidelobe Confusion response in u-v annulus • Baselines inside independent cells add coherently • Incoherent across independent cells • Integrate coherently over bandwidth

13. Sidelobe Confusion in k-space Horizon (l2+m2=1) First null of Primary Beam Edge Brightening

14. Thermal Noise • Vrms(u,v,f)= 2kBTsys/Ae (Δfτ)1/2 • Snapshot mode: 8 sec integration time • Thermal noise uniform along k|| (No convolution) • Distribution along determined by Baseline distribution k

15. Properties of Thermal Noise • Incoherent integration along bandwidth • Incoherent across baselines Morales (2005)

16. Thermal Noise in k-space

17. Combined Uncertainty • Add in quadrature • Thermal noise dominates in snapshot mode

18. Modeled EoR Power Spectrum • P(k) from Lidz et al.(2008) for z=7.32, ionization fraction=0.54 • Peculiar Velocity corrections applied

19. Measured EoR Power Spectrum • Pass through bandpass • Observed through array configuration

20. Summary of Uncertainties and EoR Power Spectrum

21. Conclusions • Radio source statistics and 128T MWA layout • Comprehensive estimate in k-space of • Sidelobe confusion due to mode-mixing • Classical Source Confusion • Thermal Noise • Array configuration has different effects on each • Thermal Noise overwhelms in snapshot modeand requires longer integration times (~ few hours) • In thermal noise limited regime, lower effective bandwidth degrades sensitivity