Multitension strings in high-resolution U (1 )×U (1 ) simulations

Topological defects are a fossil relic of early Universe phase transitions, with cosmic strings being the best motivated example. While in most cases one studies Nambu-Goto or Abelian-Higgs strings, one also expects that cosmologically realistic strings should have additional degrees of freedom in their world sheets, one specific example being superstrings from type IIB superstring theory. Here we continue the scientific exploitation of our recently developed multi-graphics processing unit field-theory cosmic strings code to study the evolution of U (1 )×U (1 ) multitension networks, which are a numerically convenient proxy: these contain two lowest-tension strings networks able to interact and form bound states, providing a convenient first approximation to the behavior expected from cosmic superstrings. We start with a discussion of our code validation, including a comparison of the evolution of these networks under three different assumptions: physical evolution (using the true equations of motion), the constant comoving width assumption (frequently used in the literature) and also the numerically convenient core growth case. We rely on the largest field-theory simulations of this model so far, specifically 4096³, Δ x =0.5 boxes. We present robust evidence of scaling for the lightest strings, measured through a complete and self-consistent set of correlation length and velocity diagnostics. We also find a linearly growing average length of the bound state segments, consistent with a scaling behavior. (In previously reported lower-resolution simulations, such behavior had only been identified with carefully engineered initial conditions, rich in those segments.) Finally, while we see no evidence of a large population of bound states forming at early stages of the network evolution, we do present tentative evidence for an asymptotic constant value of the fraction of bound states, with this value being different in the radiation and the matter eras. Our work demonstrates that our graphics processing unit-accelerated field-theory code can by successfully extended beyond the simple Abelian-Higgs approximation, and enables future detailed studies of realistic string networks and of their observational signatures.