SIMBAD references

Observations of AT2017gfo, the counterpart of the double neutron star merger, confirmed the existence of kilonovae that are suggested to originate from the decay of the very heavy elements. However, the different simplified assumptions used in analytical and semi-analytical models can lead to a variety of fitting results. Previous semi-analytical models have assumed a grey body because of the challenge in calculating the detailed opacity of ejecta and to save the calculation time. However, the variation of opacity with wavelength cannot be ignored when calculating the multicolour flux. Though numerical simulations can calculate kilonova details, they require an excessive amount of time for data modelling. In this work, we propose an extension of previous semi-analytical models by not assuming the grey-body opacity and constant thermalization efficiency assumptions. We demonstrate that the impact of opacity changing with wavelength is non-negligible and then introduce an empirical analytical opacity in optical and near-infrared bands to calculate multicolour light curves of kilonovae with a three-dimensional, three-component model. Additionally, we consider the evolution of thermalization efficiency at later times, enabling application of our model to both the early and later phases of kilonovae. To enable efficient fitting of models with many parameters (23 parameters in the case of double neutron star merger), a quick method based on the greedy algorithm is developed. We can well reproduce the infrared/optical/ultraviolet observations of AT2017gfo. This model can also be extended to black-hole and neutron star mergers as well.