In figures 1 and 2 in this paper the spectral index of the magnetic field spectra are wrong  due to a CAMB bug I set it to exactly scale invariant rather than n=2.9 as stated in the text. This only has a slightly visible effect on the power spectrum result and does not affect any of the conclusions.
The CAMB bug will be fixed in the next update.
[astroph/0406096] CMB anisotropies from primordial inhomogeneous magnetic fields
Authors:  Antony Lewis 
Abstract:  Primordial inhomogeneous magnetic fields of the right strength can leave a signature on the CMB temperature anisotropy and polarization. Potentially observable contributions to polarization Bmodes are generated by vorticity and gravitational waves sourced by the magnetic anisotropic stress. We compute the corresponding CMB transfer functions in detail including the effect of neutrinos. The shear rapidly causes the neutrino anisotropic stress to cancel the stress from the magnetic field, suppressing the production of gravitational waves and vorticity on superhorizon scales after neutrino decoupling. A significant large scale signal from tensor modes can only be produced before neutrino decoupling, and the actual amplitude is somewhat uncertain. Plausible values suggest primordial nearly scale invariant fields of ~ 10^(10)G today may be observable from their large scale tensor anisotropy. They can be distinguished from primordial gravitational waves by their nonGaussianity. Vector mode vorticity sources Bmode power on much smaller scales with a power spectrum somewhat similar to that expected from weak lensing, suggesting amplitudes ~ 10^(9)G may be observable on small scales for a spectral index of n ~ 2.9. In the appendix we review the covariant equations for computing the vector and tensor CMB power spectra that we implement numerically. 
[PDF] [PS] [BibTex] [Bookmark] 

 Posts: 1618
 Joined: September 23 2004
 Affiliation: University of Sussex
 Contact: