On the longtime behavior of solutions to a model for epitaxial growth