In order to satisfy limits on the allowed variation of particle masses from big bang nucleosynthesis (BBN) until today, the chameleon scalar field is required to reach its attractor solution early on in its cosmological evolution. Brax et al. (2004) have shown this to be possible for certain specific initial conditions on the chameleon field at the end of inflation. However the extreme fine-tuning necessary to achieve this, poses a problem if the chameleon is to be viewed a natural candidate for dark energy. In this article we revisit the behaviour of the chameleon in the early Universe, including the additional coupling to electromagnetism proposed by Brax et al. (2011). Solving the chameleon evolution equations in the presence of a primordial magnetic field, we find that the strict initial conditions on the chameleon field at the end of inflation can be relaxed, and we determine the associated lower bound on the strength of the primordial magnetic field.

We consider the non-resonant mixing between photons and scalar ALPs with masses much less than the plasma frequency along the path, with specific reference to the chameleon scalar field model. The mixing would alter the intensity and polarization state of the cosmic microwave background (CMB) radiation. We find that the average modification to the CMB polarization modes is negligible. However the average modification to the CMB intensity spectrum is more significant and we compare this to high precision measurements of the CMB monopole made by the far infrared absolute spectrophotometer (FIRAS) on board the COBE satellite. The resulting 95\% confidence limit on the scalar-photon conversion probability in the primordial field (at 100 GHz) is P < 2.6x10^{-2}. This corresponds to a degenerate constraint on the photon-scalar coupling strength, g, and the magnitude of the primordial magnetic field. Taking the upper bound on the strength of the primordial magnetic field derived from the CMB power spectra, B < 5.0x10^{-9}G, this would imply an upper bound on the photon-scalar coupling strength in the range g < 7.14x10^{-13}GeV^{-1} to g < 9.20x10^{-14}GeV^{-1}, depending on the power spectrum of the primordial magnetic field.

We constrain the chameleonic Sunyaev--Zel'dovich (CSZ) effect in the Coma cluster from measurements of the Coma radial profile presented in the WMAP 7-year results. The CSZ effect arises from the interaction of a scalar (or pseudoscalar) particle with the cosmic microwave background in the magnetic field of galaxy clusters. We combine this radial profile data with SZ measurements towards the centre of the Coma cluster in different frequency bands, to find Delta T_{SZ,RJ}(0)=-400+/-40 microKelvin and Delta T_{CSZ}^{204 GHz}(0)=-20+/-15 microKelvin (68% CL) for the thermal SZ and CSZ effects in the cluster respectively. The central value leads to an estimate of the photon to scalar (or pseudoscalar) coupling strength of g = (5.2 - 23.8) x 10^{-10} GeV^{-1}, while the 95% confidence bound is estimated to be g < (8.7 - 39.4) x 10^{-10} GeV^{-1}.

We show that a direct coupling between a chameleon-like scalar field and photons can give rise to a modified Sunyaev-Zel'dovich (SZ) effect in the Cosmic Microwave Background (CMB). The coupling induces a mixing between chameleon particles and the CMB photons when they pass through the magnetic field of a galaxy cluster. Both the intensity and the polarization of the radiation are modified. The degree of modification depends strongly on the properties of the galaxy cluster such as magnetic field strength and electron number density. Existing SZ measurements of the Coma cluster enable us to place constraints on the photon-chameleon coupling. The constrained conversion probability in the cluster is P_Coma(204 GHz) < 6.2 * 10^{-5} at 95% confidence, corresponding to an upper bound on the coupling strength of g_eff < 2.2 * 10^{-8} GeV^-1 or g_eff < 7.2-32.5 * 10^{-10} GeV^-1, depending on the model that is assumed for the cluster magnetic field structure. We predict the radial profile of the chameleonic CMB intensity decrement. We find that the chameleon effect extends further towards the edges of the cluster than the thermal SZ effect. Thus we might see a discrepancy between the X-ray emission data and the observed SZ intensity decrement. We further predict the expected change to the CMB polarization arising from the existence of a chameleon-like scalar field. These predictions could be verified or constrained by future CMB experiments.