Fitting dust/CIB amplitudes
Table of Contents
We use Planck 353 GHz power spectra reconstructed within ACT DR6 survey and for ℓ-multipoles [300; 2000]. To cancel the CMB contribution, we compute spectra difference between 353 GHz and 143 GHz frequencies i.e
\begin{align*} \mathcal{D}_ℓ^{353-143}&=\mathcal{D}_ℓ^{353×353}+\mathcal{D}_ℓ^{143×143}-2\mathcal{D}_ℓ^{353×143} \end{align*}
When fitting TT power spectra, we also add pa4_f220 ACT array for ℓ-multipole greater than 3500 and
jointly fit dust and CIB contributions to break degeneracy between these spectra. We also use a mask
with larger point source holes (21 radians) due to larger Planck beam.
1. Temperature results
When jointly fitting Planck 353 GHz spectra and pa4_f220, the fit does not give a good \(χ²\) value
due to tension between CIB and dust contributions. Whereas the pa4_f220 error bars might be
underestimated by ~20%, this partially alleviates the fit tension.
When only fitting Planck spectra, the goodness of the fit is restored, giving a lesser contribution to dust and a higher one to CiB.
If we now allow for a shift in frequency in ACT 220 GHz channel, recording that the central value is 226.7 ± 3.6 GHz (see wiki), we get a pretty good agreement with both measurements
2. Polarisation results
2.1. TE
2.2. EE
2.3. TB
2.4. BB
