New Supernova Constraints on Sterile-Neutrino Production.

We consider the possibility that a light, sterile-neutrino species νS can be produced by νe scattering during the cooling of a proto-neutron star. If we parameterize the sterile-neutrino production cross section by a parameter A as σ(νeX −→ νSX) = Aσ(νeX −→ νeX), where X is an electron, neutron, or proton, we show that A is constrained by limits to the conversion of νe to νS in the region between the sterile-neutrino trapping region and the electronneutrino trapping region. This consideration excludes values of A in the range 10−4 <∼ A <∼ 10 −1. The possibility that the solar-neutrino problem [1] may be solved via the oscillation of electron neutrinos to “sterile” neutrinos (so named because they have superweak interaction with the W and the Z and thus avoid the LEP bound on the number of neutrinos) has been widely discussed in literature. The most compelling case for sterile neutrinos [2] arises when one tries to solve simultaneously the solar-neutrino problem and the atmospheric-neutrino deficit, as well as accommodating either (or both) the reported ν̄μ −→ ν̄e oscillations at LSND [3] or the idea that neutrinos constitute about 20% of the total matter content of the universe [4]. In the sterile-neutrino hypothesis the solar-neutrino deficit is resolved via MSW oscillations between νe and νS. In constructing realistic gauge models [2, 5, 6] that lead to the mixings between the electron neutrino and the sterile neutrino one generally introduces various new interactions which can lead to the desired mixing without conflicting with known low-energy constraints as well as cosmological and astrophysical ones [7]. A well known astrophysical phenomenon that leads to strong constraints on the static properties of the neutrino is the dynamics of supernovae inferred from the neutrino signal from supernova 1987A (SN1987A) observed by the underground detectors of the IMB and Kamiokande collaborations [8]. Two classes of restrictions on sterile neutrino properties may be obtained. One is on νe oscillating into νS, thereby depleting the νe signal and contradicting observations. This possibility has been analysed by Kainulainen et al. [9], who showed that the high density in the supernova suppresses such oscillations for the range of masses and mixing angles needed to solve the solar neutrino problem. The other class of constraints may arise in models where there exist direct interactions of electron neutrinos with visible particles such as e, p, n, νe,μ,τ that can convert a νe into a νS. This can also potentially deplete the νe luminosity. It is this class of effects that we discuss in this letter. We will also