A minimal extension of the standard model to naturally generate small neutrino masses and provide a dark matter candidate is proposed. The dark matter particle is part of a new scalar doublet field that plays a crucial role in radiatively generating neutrino masses. The symmetry that stabilizes the dark matter also suppresses neutrino masses to appear first at three-loop level. Without the need of right-handed neutrinos or other very heavy new fields, this offers an attractive explanation of the hierarchy between the electroweak and neutrino mass scales. The model has distinct verifiable predictions for the neutrino masses, flavor mixing angles, colliders and dark matter signals.

We consider a reconstructing scheme using observational data from SNIa, BAO and CMB, based on a model of dark unification using a single non-minimally coupled scalar field. We investigate through a reconstruction program, the main features the current observational data imposes to the scalar field potential. We found that the form suggested by observations implies a step feature in the potential, where the kinetic and potential energy becomes of the same order of magnitude.

We introduce the curvature $\Omega_k$ as a new free parameter in the Bayesian analysis using SNIa, BAO and CMB data, in a model with variable equation of state parameter $w(z)$. We compare the results using both the Constitution and Union 2 data sets, and also study possible low redshift transitions in the deceleration parameter $q(z)$. We found that, incorporating $\Omega_k$ in the analysis, it is possible to make all the three observational probes consistent using both SNIa data sets. Our results support dark energy evolution at small redshift, and show that the tension between small and large redshift probes is ameliorated. However, although the tension decreases, it is still not possible to find a consensus set of parameters that fit all the three data set using the Chevalier-Polarski-Linder CPL parametrization.