We present a first attempt to determine nucleon-nucleon potentials in the parity-odd sector, which appear in 1P1, 3P0, 3P1, 3P2-3F2 channels, in Nf=2 lattice QCD simulations. These potentials are constructed from the Nambu-Bethe-Salpeter wave functio
ns for J^P=0^-, 1^- and 2^-, which correspond to A1^-, T1^- and T2^- + E^- representation of the cubic group, respectively. We have found a large and attractive spin-orbit potential VLS(r) in the isospin-triplet channel, which is qualitatively consistent with the phenomenological determination from the experimental scattering phase shifts. The potentials obtained from lattice QCD are used to calculate the scattering phase shifts in 1P1, 3P0, 3P1 and 3P2-3F2 channels. The strong attractive spin-orbit force and a weak repulsive central force in spin-triplet P-wave channels lead to an attraction in the 3P2 channel, which is related to the P-wave neutron paring in neutron stars.
Recently a new approach to calculate the nuclear potential from lattice QCD has been proposed. In the approach the nuclear potential is constructed from Bethe-Salpeter (BS) wave functons through the Schroedinger equation. The procedure leads to non-l
ocal but energy independent potential, which can be expanded in terms of local functions. In several recent applications of this method, local potentials, which correspond to the leading order (LO) terms of the expansion, are calculated from the BS wave function at E~0 MeV, where E is the center of mass energy. It is therefore important to check the validity of the LO approximation obtained at E~0. In this report, in order to check how well the LO approximation for the NN potentials works, we compare the LO potentials determined from the BS wave function at E~45 MeV with those at E~0 MeV in quenched QCD. We find that the difference of the LO potentials between two energies are not found wihin the statistical errors. This shows that the LO approximation for the potential is valid at low energies to describe the NN interactions.
