Close binary systems of compact stars, due to the emission of gravitational radiation, may evolve into a phase in which the less massive star transfers mass to its companion. We describe mass transfer by using the model of Roche lobe overflow, in which mass is transferred through the first, or innermost, Lagrange point. Under conditions in which gravity is strong, the shapes of the equipotential surfaces and the Roche lobes are modified compared to the Newtonian case. We present calculations of the Roche lobe utilizing the second order post-Newtonian (2PN) approximation in the Arnowitt-Deser-Misner gauge. Heretofore, calculations of the Roche lobe geometry beyond the Newtonian case have not been available. Beginning from the general N-body Lagrangian derived by Damour and Schaffer, we develop the Lagrangian for a test particle in the vicinity of two massive compact objects. As an exact result for the transverse-traceless part of the Lagrangian is not available, we devise an approximation that is valid for regions close to the less massive star. We calculate the Roche lobe volumes, and provide a simple fitting formula for the effective Roche lobe radius analogous to that for the Newtonian case furnished by Eggleton. In contrast to the Newtonian case, in which the effective Roche radius depends only upon the mass ratio $q=m_1/m_2$, in the 2PN case the effective Roche lobe radius also depends on the ratio $z=2 (m_1+m_2)/a$ of the total mass and the orbital separation.

Binary mergers involving black holes and neutron stars have been proposed as major sources of gravitational waves, r--process nucleosynthesis, and gamma ray bursters. In addition, they represent an important, and possibly unique, observable that could distinguish between normal and self--bound neutron stars. These two families of stars have distinctly different mass--radius relationships resulting from their equations of state which can be revealed during their mergers if stable mass transfer ensues. We consider two cases of gravitational-radiation induced binary mergers: (i) a black hole and a normal neutron star, and (ii) a black hole and a self-bound strange quark matter star. We extend previous Newtonian analyses to incorporate the pseudo-general relativistic Paczy\'nski-Wiita potential or a potential correct to second--order post-Newtonian order in Arnowitt--Deser--Misner coordinates. These potentials are employed to study both the orbital evolution of the binary and the Roche lobe geometry that determines when and if stable mass transfer between the components is possible. The Roche lobe geometry with pseudo-general relativistic or post-Newtonian potentials has not heretofore been considered. Our analysis shows that differences in the evolution of normal neutron stars and strange quark matter stars are significant and could be detected in gravity waves. Both the amplitude and frequencies of the wave pattern are affected. In addition, details of the equation of state for either normal neutron stars or strange quark stars may be learned. A single merger could reveal one or two individual points of the mass-radius relation, and observations of several mergers could map a significant portion of this relation.

The mergers of black hole-neutron star binaries are calcuated using a pseudo-general relativistic potential that incorporates ${\mathcal O}(v^2/c^2)^3$ post-Newtonian corrections. Both normal matter neutron stars and self-bound strange quark matter stars are considered as black hole partners. As long as the neutron stars are not too massive relative to the black hole mass, orbital decay terminates in stable mass transfer rather than an actual merger. For a normal neutron star, mass transfer results in a widening of the orbit but the stable transfer ends before the minimum neutron star mass is reached. For a strange star, mass transfer does not result in an appreciable enlargement of the orbital separation, and the stable transfer continues until the strange star essentially disappears. These differences might be observable through their respective gravitational wave signatures.

Neutrino processes in semi-transparent supernova matter, opaque to semi-transparent protoneutron star matter, and catalyzed neutron stars are discussed.

Explicit expressions for the differential and total rates and emissivities of neutrino pairs from the photo-neutrino process $e^\pm + \gamma \to e^\pm + \nu + \bar\nu$ in hot and dense matter are derived. Full information about the emitted neutrinos is retained by evaluating the squared matrix elements for this process which was hitherto bypassed through the use of Lenard's identity in obtaining the total neutrino emissivities. Accurate numerical results are presented for widely varying conditions of temperature and density. Analytical results helpful in understanding the qualitative behaviors of the rates and emissivities in limiting situations are derived. The corresponding production and absorption kernels in the source term of the Boltzmann equation for neutrino transport are developed. The appropriate Legendre coefficients of these kernels, in forms suitable for multigroup flux-limited diffusion schemes are also provided.

The differential rates and emissivities of neutrino pairs from an equilibrium plasma are calculated for the wide range of density and temperature encountered in astrophysical systems. New analytical expressions are derived for the differential emissivities which yield total emissivities in full agreement with those previously calculated. The photon and plasmon pair production and absorption kernels in the source term of the Boltzmann equation for neutrino transport are provided. The appropriate Legendre coefficients of these kernels, in forms suitable for multi-group flux-limited diffusion schemes are also computed.