Dust SEDs in Milky Way-like galaxies in the IllustrisTNG simulations based on the evolution of grain size distribution
To understand how the evolution of grain size distribution in galaxies affects observed dust properties, we apply a post-processing dust evolution model to galaxy merger trees from the IllustrisTNG cosmological hydrodynamical simulation. Our dust model includes stellar dust production, sputtering in hot gas, dust growth by accretion and coagulation in the dense interstellar medium (ISM), and shattering in the diffuse ISM. We decompose the grain size distribution into different dust species depending on the elemental abundances and the dense ISM fraction given by the simulation. In our previous work, we focused on Milky Way (MW) analogs and reproduced the observed MW extinction curve. In this study, we compute dust spectral energy distributions (SEDs) for the MW analogues. Our simulated SEDs broadly reproduce the observed MW SED within their dispersion and so does the observational data of nearby galaxies, although they tend to underpredict the MW SED at short wavelengths where emission is dominated by polycyclic aromatic hydrocarbons (PAHs). We find that metallicity and dense gas fraction are the most critical factors for the SED shape, through their influence on coagulation and shattering.The overall success of our models in reproducing the MW SED further justifies the dust evolution processes included in the model and predicts the dispersion in the SEDs caused by the variety in the assembly history. We also show that the most significant increase in the dust SED occurs between redshifts $z\sim 3$ and 2 in the progenitors of the simulated MW-like galaxies.