A further study of $t_{\rm burst}$ of GRBs: rest frame properties, external plateau contributions and multiple parameter analysis

[Zhang 2014] propose to redefine the true GRB central engine activity duration, $t_{\rm burst}$, by considering the contributions from the prompt $\gamma$-ray emission, X-ray flare and internal plateau features. With a comprehensive study on a large sample of Swift GRBs, it is shown that the $t_{\rm burst}$ distribution in the observer frame consists of a bimodal feature, suggesting the existence of a new population of ultra-long GRBs. In this work, we make a series of further studies on $t_{\rm burst}$: we update the Swift GRB sample up to June 2016; we investigate the properties of $t_{\rm burst}$ distribution in the rest frame; we redefine $t_{\rm burst}$ by involving external plateau contributions; we make a multiple parameter analysis to investigate whether the bursts within the ultra-long population being statistically different in sense of other features besides the duration distribution. We find that for all situations, the distribution of $t_{\rm burst}$ requires two normal distributions in logarithmic space to provide a good fit, both in observer frame and rest frame. Considering the observational gap effect would not completely erase the bimodal distribution feature. However the bursts within the ultra-long population may have no statistically different in sense of other features besides the duration term. We thus suggest that if the ultra-long population of GRBs indeed exists, their central engine mechanism and radiation mechanism should be similar to the normal population, but they have longer central engine activity timescale.