Results from temperature stepped desorption (TSD) experiments are presented and compared with simulations from the TSD model presented in the first of this two-paper series. TSD columns were filled with a sand, a sediment, a soil, or a silica gel, all at 100% relative humidity. Next, TSD columns were equilibrated with trichloroethene (TCE), initially purged at 30 degrees C, and then heated to 60 degrees C after 100, 1000, or 10 000 min of slow desorption. One gamma distribution of diffusion rate constants at 30 degrees C and one gamma distribution Df diffusion rate constants at 60 degrees C were used to simulate column results at all three heating times for a single solid. At each heating time, diffusion rate constants of the gamma distributions at 30 degrees C and 60 degrees C were used to calculated an effective activation energy, E-act,E-eff. Values of E-act,E-eff for all solids were between 47 and 94 kJ/mol, on the order of activation energy values found for diffusion in microporous solids. Between 100 and 10 000 min heating times, the value of E-act,E-eff increased by a factor of 1.7 for the sand and by a factor of similar to 1.1 for the sediment and the soil. This suggests that diffusion occurs from micropores with a wider distribution of widths in the sand than in the other solids and that with decreasing mass remaining diffusion occurs from successively smaller width micropores. For the sediment, values of E-act,E-eff and (D//(2)(m)) were lower than those in the other solids. For a given sorbate, larger width micropores are associated with smaller values of E-act,E-eff and larger values of D. Hence, it is likely that micropores in the sediment are both wider and longer (i.e. larger value of /(2)(m)) than those in the other solids. These results suggest that micropore geometry varies between natural solids, and it is an important parameter that must be quantified to predict rates of slow desorption.