Cosmology with the EFTofLSS and BOSS: dark energy constraints and a note on priors

We analyse the BOSS DR12 multipoles of the galaxy power spectrum jointly with measurements of the BAO scale for three different models of dark energy. We use recent measurements performed with a windowless estimator, and an independent and fast pipeline based on EFTofLSS modelling implemented via the FAST-PT algorithm to compute the integrals of the redshift-space loop corrections. We accelerate our analysis further by using the bacco linear power spectrum emulator instead of a Boltzmann solver. We perform two sets of analyses: one including 3σ Planck priors on As and ns, and another that is fully CMB-free, i.e., letting the primordial parameters vary freely. The first model we study is ΛCDM, within which we reproduce previous results obtained with the same estimator. We find a low value of the scalar amplitude in the CMB-free case, in agreement with many previous EFT-based full-shape analyses of the BOSS data. We then study wCDM, finding a lower value of the amplitude in the CMB-free run, coupled with a preference for phantom dark energy with w = -1.17+0.12-0.11, again in broad agreement with previous results. Finally, we investigate the dark scattering model of interacting dark energy, which we label wACDM. In the CMB-free analysis, we find a large degeneracy between the interaction strength A and the amplitude As, hampering measurements of those parameters. On the contrary, in our run with a CMB prior, we are able to constrain the dark energy parameters to be w = -0.972+0.036-0.029 and A = 3.9+3.2-3.7, which show a 1σ hint of interacting dark energy. This is the first measurement of this parameter and demonstrates the ability of this model to alleviate the σ8 tension. Our analysis can be used as a guide for the analysis of any model with scale-independent growth. Finally, we study the dependence of the results on the priors imposed on the nuisance parameters and find these priors to be informative, with their broadening generating shifts in the contours. We argue for an in depth study of this issue, which can affect current and forthcoming analyses of LSS data.