Consistent analysis of the ν_μ to ν_e sterile neutrinos searches of ICARUS and OPERA

The two long-baseline experiments ICARUS and OPERA have recently provided bounds on light ($\sim$ eV) sterile neutrinos exploiting the negative results of the $\nu_\mu \to \nu_e$ appearance searches. Both collaborations have performed the data analysis using an effective 2-flavor description. We show that such a simplified treatment neglects sizable genuine 4-flavor effects, which are induced by the interference of the new large squared-mass splitting $\Delta m^2 _{14}$ with the atmospheric one. The inclusion in the data analysis of such effects weakens the upper bounds on the effective appearance amplitude $\sin^2 2\theta_{\mu e}$ approximately by a factor of two. In addition, we evidence that, in a 4-flavor scheme, the flavor oscillations involve also the $\nu_e$ component of the CNGS beam and can suppress the theoretical expectation of the background in a substantial way. The inclusion in the data analysis of the $\nu_e$ disappearance effects leads to a further weakening of the upper bounds on $\sin^2 2\theta_{\mu e}$, which overall are relaxed by a factor of three with respect to those obtained in the effective 2-flavor description.