Abstract
X-ray free-electron lasers (XFEL) are revolutionary photon sources, whose ultrashort, brilliant pulses are expected to allow single-molecule diffraction experiments providing structural information on the atomic length scale of nonperiodic objects. This ultimate goal, however, is currently hampered by several challenging questions basically concerning sample damage, Coulomb explosion, and the role of nonlinearity. By employing an original ab initio approach, we address these issues showing that XFEL-based single-molecule imaging will be only possible with a few-hundred long attosecond pulses, due to significant radiation damage and the formation of preferred multisoliton clusters which reshape the overall electronic density of the molecular system at the femtosecond scale.