We follow the evolution of helium stars of initial mass $(2.2 - 2.5) M_\odot$, and show that they undergo off-center carbon burning, which leaves behind ${\mathbf \sim 0.01 M_\odot}$ of unburnt carbon in the inner part of the core. When the carbon-oxygen core grows to Chandrasekhar mass, the amount of left-over carbon is sufficient to ignite thermonuclear runaway. At the moment of explosion, the star will possess an envelope of several $0.1 M_{\odot}$, consisting of He, C, and possibly some H, perhaps producing a kind of peculiar SN. Based on the results of Waldman and Barkat (2007) for accreting white dwarfs, we expect to get thermonuclear runaway at a broad range of $\rho_c \approx (1 - 6) \times 10^9 \mathrm{g cm^{-3}}$, depending on the amount of residual carbon. We verified the feasibility of this scenario by showing that in a close binary system with initial masses $(8.5 + 7.7) M_{\odot}$ and initial period of 150 day the primary produces a helium remnant of $2.3 M_{\odot}$ that evolves further like the model we considered.

Exploring stellar models which ignite carbon off-center (in the mass range of about 1.05 - 1.25 Msun, depending on the carbon mass fraction) we find that they may present an interesting SN I progenitor scenario, since whereas in the standard scenario runaway always takes place at the same density of about 2 X 10^9 gr/cm^3, in our case, due to the small amount of carbon ignited, we get a whole range of densities from 1 X 10^9 up to 6 X 10^9 gr/cm^3. These results could contribute in resolving the emerging recognition that at least some diversity among SNe I exists, since runaway at various central densities is expected to yield various outcomes in terms of the velocities and composition of the ejecta, which should be modeled and compared to observations.

We explore the hypothesis, that helium stars in a certain mass range can evolve to a carbon core explosion similar to what is widely accepted as an explanation for the SN I phenomenon. This should happen when their carbon-oxygen core grows thanks to the helium shell burning above the core. We found that in the mass range of about 1.7-2.2 Msun, indeed this can happen. The main new insight we believe we gained is the crucial importance of an "early" off-center ignition of carbon, which at a later stage prevents the carbon which forms below the helium burning shell and ignites, from burning the carbon all the way to the center. When helium is almost depleted in the convective envelope by the helium burning shell at its bottom, the now super-Chandrasekhar mass carbon-oxygen core contracts, and the residual degenerate carbon at the center is ignited, resulting in a runaway similar to the classical SN I scenario. Since the structure and behavior of the carbon-oxygen core of the helium stars of our interest is very similar to that of a mass accreting carbon-oxygen star, we also thoroughly examined the behavior of carbon-oxygen stars. We discovered that the models which ignite carbon off-center (in the mass range of about 1.05-1.18 Msun, depending on the carbon mass fraction) present an interesting SN I progenitor scenario of their own, since whereas in the standard scenario runaway always takes place at the same density of about 2E9 gr/cm3, in our case, due to the small amount of carbon ignited, we get a whole range of densities from 1E9 up to 6E9 gr/cm3.

The previous analysis of the convective Urca neutrino loss process in degenerate, convective, quasi-static, carbon-burning cores by Barkat and Wheeler omitted specific consideration of the role of the kinetic energy flux. The arguments of Barkat and Wheeler that steady-state composition gradients exist are correct, but chemical equilibrium does not result in net cooling. Barkat and Wheeler included a "work" term that effectively removed energy from the total energy budget that could only have come from the kinetic energy, which must remain positive. Consideration of the kinetic energy in the thermodynamics of the convective Urca process shows that the convective Urca neutrinos reduce the rate of increase of entropy that would otherwise be associated with the input of nuclear energy and slow down the convective current, but, unlike the "thermal" Urca process do not reduce the entropy or temperature.