Nuclear electromagnetic breakup processes at low energy are particularly relevant in the astrophysical context. In this Thesis we analyse the Beryllium-9 photodisintegration reaction, whose inverse process, under certain astrophysical conditions, is related to the Carbon-12 formation. A preliminary study of the Carbon-12 photodisintegration is also carried out. The interaction of these nuclei with a low-energy photon induces a transition to a state consisting of cluster sub-units, the alpha-particles, and possibly a neutron, n. The theoretical study of the cross section in the low-energy regime is conducted by using a three-body ab initio approach. Beryllium-9 exhibits a clear separation of energy scales, since its alpha-alpha-n three-body binding energy is shallow compared to the binding of the alpha-particle. Within this framework a halo/cluster Effective Field Theory (EFT) can be developed. The alpha-alpha and alpha-n effective interactions are defined in momentum space as a series of contact terms, regularized by a momentum-regulator function. The Low Energy Constants are expressed in terms of scattering observables, i.e. scattering length and effective range. A three-body potential is also introduced in the model. Carbon-12 is studied on the same footing. By means of an integral transform approach, the problem of the transition to a state in the continuum can be advantageously reformulated in terms of a bound-state problem: in the calculations we use the Lorentz Integral Transform method, in conjunction with the Non-Symmetrized Hyperspherical Harmonics method. In determining the low-energy photodisintegration cross section, the nuclear current matrix element is evaluated through the electric dipole, or quadrupole, transition operator (Siegert theorem). Since the continuity equation is used explicitly, the contribution of the one-body and the many-body current operators is implicitly included in the calculation. By comparing the results with those obtained by using a one-body convection current, the effect of the many-body terms can be quantified. The dependence of the results on different EFT parameters is discussed, always in connection with the experimental data available in the literature. By following the power counting dictated by the EFT approach for Beryllium-9, the inclusion of different partial waves in the potential model is explored. In addition to a alpha-alpha S-wave, a alpha-n P-wave and a three-body effective interaction, a alpha-n S-wave term is also required to obtain results more consistent with the experimental data. The contribution of the many-body currents to the cross section is found to be non-negligible. Although at an early stage, Carbon-12 results show interesting features. The formalism presented in this Thesis can be extended to study the photodisintegration of Oxygen-16 within a fully four-body ab initio approach.