Background: The eikonal approximation is a high-energy reaction model which is very computationally efficient and provides a simple interpretation of the collision. Unfortunately, it is not valid at energies around 10 MeV/nucleon, the range of energy of HIE-ISOLDE at CERN and the future ReA12 at MSU. Fukui etal. [Phys. Rev. C 90, 034617 (2014)] have shown that a simple semiclassical correction of the projectile-target deflection could improve the description of breakup of halo nuclei on heavy targets down to 20 MeV/nucleon. Purpose: We study two similar corrections, which aim at improving the projectile-target relative motion within the eikonal approximation, with the goal to extend its range of validity down to 10 MeV/nucleon in nuclear-dominated collisions, viz. on light targets. The semiclassical correction substitutes the impact parameter by the distance of closest approach of the corresponding classical trajectory. The exact continued $S$-matrix correction replaces the eikonal phase by the exact phase shift. Both corrections successfully describe the elastic scattering of one-neutron halo nuclei. Method: We extend these corrections and study their efficiency in describing the breakup channel. We evaluate them in the case of $^{11}\mathrm{Be}$ impinging on $^{12}\mathrm{C}$ at 20 and 10 MeV/nucleon. Results: Albeit efficient to reproduce the elastic channel, these corrections do not improve the description of the breakup of halo nuclei within the eikonal approximation down to 20 MeV/nucleon. Conclusions: Our analysis of these corrections shows that improving the projectile-target relative motion is not the ultimate answer to extend the eikonal approximation down to low energies. We suggest another avenue to reach this goal.