On the different Floquet Hamiltonians in a periodic-driven Bose-Josephson junction

The bosonic Josephson junction, one of the maximally simple models for periodic-driven many-body systems, has been intensively studied in the past two decades. Here, we revisit this problem with five different methods, all of which have solid theoretical reasoning. We find that to the order of $\omega^{-2}$ ($\omega$ is the modulating frequency), these approaches will yield slightly different Floquet Hamiltonians. In particular, the parameters in the Floquet Hamiltonians may be unchanged, increased, or decreased, depending on the approximations used. Especially, some of the methods generate new interactions, which still preserve the total number of particles; and the others do not. The validity of these five effective models is verified using dynamics of population imbalance and self-trapping phase transition. In all results, we find the method by first performing a unitary rotation to the Hamiltonian will have the highest accuracy. The difference between them will become significate when the modulating frequency is comparable with the driving amplitude. The results presented in this work indicate that the analysis of the Floquet Hamiltonian has some kind of subjectivity, which will become an important issue in future experiments with the increasing of precision. We demonstrate this physics using a Bose-Josephson junction, and it is to be hoped that the validity of these methods and their tiny differences put forward in this work can be verified in realistic experiments in future using quantum simulating platforms, including but not limited to ultracold atoms.