Planck Data and Ultralight Axions

We examine the effects of photon-axion mixing on the CMB. We show that if there are very underdense regions between us and the last scattering surface which contain coherent magnetic fields (whose strength can be orders of magnitude weaker than the current bounds), then photon-axion mixing can induce observable deviations in the CMB spectrum. Specifically, we show that the mixing can give rise to non-thermal spots on the CMB sky. As an example we consider the well known CMB cold spot, which according to the Planck data has a weak distortion from a black body spectrum, that can be fit by our model. While this explanation of the non-thermality in the region of the cold spot is quite intriguing, photon-axion oscillation do not explain the temperature of the cold spot itself. Nevertheless we demonstrate the possible sensitivity of the CMB to ultralight axions which could be exploited by observers. ar X iv :1 40 5. 17 25 v1 [ as tr oph .C O ] 7 M ay 2 01 4 The QCD axion remains the best solution of the strong CP problem [1–3]. The most successful UV completions [4] which do not suffer from fine tuning [5,6] contain many more ultralight axion-like fields. Only one of their linear combinations will couple to the gluons of the Standard Model, becoming the QCD axion. Another linear combination might couple to the photon, and thus remain extremely light. Such an ultralight axion could have cosmological consequences, since it can affect photons coming from distant cosmological sources, including the Cosmic Microwave Background (CMB). Previously we considered [7] the possibility that photon-axion oscillations [8] in intergalactic magnetic fields could contribute to the dimming of distant supernovae (SNe) [9]. This mechanism requires the photon-axion coupling scale to be M ∼ 4 · 10 GeV, which is just above the bounds from SN1987A [10], assumes an axion mass m < 10−15 eV, and an intergalactic magnetic field B ∼ 5 · 10−9 G, with a domain size of order a Mpc, in agreement with observational bounds [11, 12]. Further properties of the model were investigated in [13–22]. The current status of this scenario is that while it cannot account for the entire SNe dimming, photon-axion oscillation can still yield a sizeable contribution, contaminating the direct determination of the dark energy equation of state from the supernova Hubble diagram by as much as several tens of percents [15,21], and making the observed parameter w = p/ρ look more negative than it really is. Further, the most precise CMB measurements to date made by the Planck collaboration [12] have pointed to curious small discrepancies between CMB and SNe data. Specifically the Planck collaboration determined the fraction of the dark energy to be ΩΛ = 0.69 ± 0.02 which is smaller than previously obtained by fitting to the dimming [9] of distant SNe. For example the SNLS combined sample [12,23] gives ΩΛ = 0.77± 0.05. The tension between the different measurement techniques suggests that there may be some as yet unidentified systematic effect involved. If this survives further (ongoing) scrutiny, it might imply that the dimming of SNe is not only due to the geometry of the universe (via cosmic acceleration), but also some additional non-geometric effect, such as photon-axion mixing. Thus it is interesting to look for other signatures of photon-axion mixing to either constrain the couplings or to identify possible glimpses of new axion physics in the sky. One possibility is to search for signatures of photon-axion mixing in the CMB. This has already been considered in [21,22], where the bounds on photon-axion mixing from non-thermal distortions of the CMB spectrum [24] averaged over the whole sky (i.e., the ‘monopole’ in the expansion) have been