Top: IR echo model for a spherical shell of silicate dust. The dust shell has a radius of 0.65 pc, to match the observed photometry between 200 and 400 d. The contribution of the hot TDE blackbody, and any hot dust emission at early times, has been subtracted from the data. The shell does not reproduce the observed drop in the MIR light curves, and under-predicts the potential cool dust component at early times. Bottom: IR echo models for different silicate dust clumps. The dust is at 0.65 pc from the TDE and has θ₀ = 30°, 40°, 50°; θ_clump = 60°, 80°, 100°. The models with larger values of θ_clump decline at later times.