CAUP Researchers: Carlos J. A. P. Martins
Team at CAUP: Galaxies and Observational Cosmology
Other Researchers: Eloisa Menegoni (IT), Silvia Galli (IT/FR), Gianpiero Mangano (IT), Alessandro Melchiorri (IT)

Varying couplings in the early universe: correlated variations of α and G,
Physical Review D, Volume 82, pp. 023532 (2010)

The anisotropies of the Cosmic Microwave Backround provide a unique opportunity to constrain parameters of the standard model while also searching for new physics. This paper is the latest of the efforts of the CAUP Galaxies & Observational Cosmology group along these lines. In collaboration with groups in Roma and Paris we provided the first joint constraints on variations of the fine-structure constant α and Newton's gravitational constant G. Assuming that the QCD sector is unchanged (and thus particle masses are constant), we have shown that current data gives no clear indication for such variations, but indicates that any relative variations in both must have the same sign for the maximally allowed variation at recombination of ∼1%. We also show how big bang nucleosynthesis data further constraints these models, and discuss the future bounds achievable by the Planck satellite mission.


Figure 1 | 68% and 95% C.L. constraints on the α/α0 vs λG for different data sets.The contours regions come from CMB data (blue), CMB data and SN-Ia (red), and CMB + HST (green).

Interest in astrophysical searches for varying fundamental constants as evidence of new physics beyond the standard model has been rekindled by a number of recent advances and results, some of which were discussed at a JENAM 2010 symposium organized by CAUP.

The CAUP's Galaxies and Observational Cosmology team, in collaboration with groups in France and Italy, has been particularly active in the field of fundamental constants variations and had recently, using Cosmic Microwave Background (CMB) observations, constrained variations of α at the epoch of recombination to be below 1% [1].

In these analyses, the standard approach is to phenomenologically assume that only one of the constants may vary, with all the others retaining their standard values. This stems in part from the need of simplicity in the analysis and in part from the fact that the available data is not precise enough to constrain joint variations. Nevertheless, from the theoretical point of view it's clear that if one of the dimensionless couplings varies then so should (at some level) the others, and observational studies should allow for this possibility.

Given its exquisite quality, cosmic microwave background data is ideal for such an analysis, and in this paper we show that indeed the data is now becoming good enough to accommodate such an analysis. Specifically, we studied scenarios where both the fine-structure constant α and Newton's constant, G, can vary. For simplicity we assumed that the QCD sector of the theory is unaffected, and therefore particle masses remain the same (this assumption is relaxed in ongoing work).

CAUP's contribution to this project has been the detailed study of these theoretical scenarios, leading to a simple but sufficiently generic parametrization:

with Q being a free parameter which we conservatively assumed to be in the range -10<Q<10. Different fundamental physics models will predict different values of Q, which may be positive or negative, and whose absolute value may be of order unity or much smaller. A measurement of Q can therefore be used as evidence for or against specific models. G was expressed as the product of its current and local value by a dimensionless parameter, λG2, such that variations in Newton's constant are translated into variations of λG.

For our analysis, done using a full Markov chain Monte Carlo package, we allowed α and G to vary during recombination and sampled a set of 10 cosmological parameters, adopting flat priors on each of them. For the CMB data, we used the 5-years WMAP observations, as well as the ACBAR, QUAD, BICEP, BOOMERanG and CBI data sets. To further constrain our results, the UNION catalog of type Ia supernovas and the recent constrain on the Hubble constant of h=0.747 ± 0.036 at 68% C.L. were also used. Consistently will other datasets, we considered purely adiabatic initial conditions and imposed spatial flatness.

The results obtained from CMB and supernova data were reproduced in Figure 1 and, although they don't allow us to conclude in favor of variations in α and/or G, they clearly show a correlation between them: the Q parameter is significantly greater than 0, with 68% C.L. ranges of about 0.7 - 0.9. Thus, if any substantial variations do exist, the variations of α are expected to have the same sign as those of G.

Assuming that neither α nor G evolve during the radiation era (a good assumption for dilaton-type models) one can extend the above analysis by using of numerical codes for big bang nucleosynthesis (BBN), allowing us to tighten the bounds on α and G up to a redshift of z~1010, as shown in Figure 2. The improvement is particularly dramatic for the case of G, since BBN is very sensitive to the expansion rate of the universe.


Figure 2 | 68% and 95% C.L. constraints on the λG vs Δα/α0 obtained from current CMB data (blue) compared with the constraints obtained from current CMB data plus BBN bounds (red).A cosmic complementarity between those two data sets helps in breaking the degeneracy and a much stronger bound is achievable on λG.

Our results have shown that astrophysical observables can be extremely useful in placing constraints on high-energy physical models that, with the current technology, would be difficult, if not even impossible, to test in laboratory. This potential is expected to improve greatly with the next generation of observatories and experiments, such as the Planck Surveyor.

[1] E. Menegoni, S. Galli, J. G. Bartlett, Carlos J. A. P. Martins, A. Melchiorri, New constraints on variations of the fine structure constant from CMB anisotropies, Physical Review D, Volume 80, pp. 087302 (2009)