We study the quench dynamics of a topological
p-wave superfluid with two competing order parameters, ${\mathrm{\ensuremath{\Delta}}}_{\ifmmode\pm\else\textpm\fi{}}(t)$. When the system is prepared in the
p+ip ground state and the interaction strength is quenched, only ${\mathrm{\ensuremath{\Delta}}}_{+}(t)$ is nonzero. However, we show that fluctuations in the initial conditions result in the growth of ${\mathrm{\ensuremath{\Delta}}}_{\ensuremath{-}}(t)$ and chaotic oscillations of both order parameters. We term this behavior phase III'. In addition, there are two other types of late-time dynamics---phase I where both order parameters decay to zero and phase II where ${\mathrm{\ensuremath{\Delta}}}_{+}(t)$ asymptotes to a nonzero constant while ${\mathrm{\ensuremath{\Delta}}}_{\ensuremath{-}}(t)$ oscillates near zero. Although the model is nonintegrable, we are able to map out the exact phase boundaries in parameter space. Interestingly, we find phase III' is unstable with respect to breaking the time-reversal symmetry of the interaction. When one of the order parameters is favored in the Hamiltonian, the other one rapidly vanishes and the previously chaotic phase III' is replaced by the Floquet topological phase III that is seen in the integrable chiral
p-wave model.
The generic Hamiltonian describing the zero temperature transition between the insulating Bose glass phase and the superfluid phase lacks particle-hole symmetry, but a statistical version of this symmetry is believed to be restored at the critical point. We show that the renormalization group relevance of particle-hole asymmetry may be explored in a controlled fashion only for small time dimensions, ${\ensuremath{\epsilon}}_{\ensuremath{\tau}}\ensuremath{\ll}1$, where we find a stable particle-hole asymmetric and an unstable particle-hole symmetric fixed point, but we provide evidence that the two merge for some finite ${\ensuremath{\epsilon}}_{\ensuremath{\tau}}\ensuremath{\approx}\frac{2}{3}$, which tends to confirm symmetry restoration at the physical ${\ensuremath{\epsilon}}_{\ensuremath{\tau}}\phantom{\rule{0ex}{0ex}}=\phantom{\rule{0ex}{0ex}}1$.