Internal Linear Combination of simulated CMB maps and foregrounds in pixel space Linear combination of component maps:
map, cmb, foregrounds, noise respectively
Assumes:
- CMB spectral energy distribution is a perfect blackbody
- CMB is uncorrelated to other sky components
Affected by errors dramatically. The ILC of the CMB is obtained by linearly combining input maps with frequency dependent weights.
The weights:
- preserve CMB blackbody spectrum
- minimise variance
with
- Compute the covariance
$C_{i i^{\prime}} = \langle d_i (p)d_{i^{\prime}}(p)\rangle$ - Compute the inverse of the covariance
- The CMB spectral energy distribution
- ILC weights can be computed as:
- Combine the input maps with the weights:
Internal Linear Combination of simulated CMB maps and foregrounds in harmonic space Performing ILC in harmonic space (as opposed to pixel space) allows for better control over spatial scales using spherical harmonic decomposition.
- Transform Frequency Maps to Harmonic Space:
Each map
where
- Compute Covariance Matrix
covariance matrix of the harmonic coefficients:
where
- Compute weights that minimize variance of the combined map while preserving the CMB signal:
where
- Combine maps
Construct the harmonic coefficients of the cleaned map.
- Reconstruct the map
- Compute the power spectrum