Microscopic optical potentials have been successful in describing nucleon-nucleus and nucleus-nucleus scattering. Some essential ingredients of the framework, however, have not been examined in detail. Applicability of the microscopic folding model is systematically investigated. Effect of an antisymmetrization factor (ASF) appearing in multiple scattering theory, theoretical uncertainty regarding the local density approximation (LDA), and the validity of a prescription for nonlocality, the Brieva-Rook (BR) localization, of the microscopic potential, are quantitatively estimated for nucleon-nucleus scattering; investigation on the ASF is carried out for also deuteron-nucleus scattering. A single folding model with the Melbourne g-matrix interaction and the SLy4 Skyrme-type Hartree-Fock-Bogoliubiv (SLy4-HFB) density is employed for evaluating a nucleon-nucleus microscopic optical potential. Deuteron-nucleus scattering is described by the continuum-discretized coupled-channels method incorporating the microscopic proton-nucleus and neutron-nucleus potentials. The ASF is found to affect proton total reaction cross sections for a 12C target below 200 MeV by about 10%. Effect of the ASF on total reaction cross sections is negligibly small if a target nucleus is heavy or scattering energy is above 200 MeV; elastic cross sections are hardly affected by the ASF for all the reaction systems considered. Below 65 MeV, still the BR localization works quite well. However, at energies below about 50 MeV, the LDA becomes less accurate for evaluating elastic cross sections at backward angles. This is the case also for the total reaction cross sections of p-12C below about 200 MeV. The microscopic model is applicable to nucleon-nucleus scattering above 25 MeV for target nuclei in a wide range of mass numbers. Deviation of calculated results from experimental data is less than about 10%.