Desorption experiments were carried out in flow through columns following long-term sorption batch experiments (up to 1010 days at 20 degreesC; Rugner, H.; Kleineidam, S.; Grathwohl, P. Long-term sorption kinetics of phenanthrene in aquifer materials. Environ. Sci. Technol. 1999, 33, 1645-1651) to elucidate sorption/desorption hysteresis phenomena of phenanthrene in aquifer materials. Most of the sorbents employed in this study (homogeneous lithocomponents separated from aquifer sediments or fresh rock fragments) showed highly nonlinear sorption isotherms because of coal particles embedded inside the grains. Because sorption capacities were high, sorption equilibrium was not reached in most of the sorbents during the initial sorptive uptake experiments lasting up to 1010 days. Desorption was studied up to 90 days at 20 degreesC. The temperature was raised after that stepwise from originally 20 to 30, 40, 50, and finally to 70 degreesC for selected samples to estimate activation energies of desorption. A numerical intraparticle pore diffusion model was used to fit sorptive uptake data and subsequently for pure forward prediction of the release rates in the desorption column experiments. Desorption was initially fast followed by extended tailing which in other studies is fitted by using multirate first-order models. Our results demonstrate that the retarded intraparticle pore diffusion model can predict the desorption rates with a single diffusion rate constant obtained independently from the long-term batch sorption experiment. No evidence for hysteresis was found, suggesting that many hysteresis phenomena reported earlier are experimental artifacts resulting from nonequilibrium effects and "nonphysical" models. The different temperature steps allowed one to additionally calculate activation energies of desorption (45-59 W mol(-1)), which were in reasonably good agreement with results from earlier studies for a retarded pore diffusion process. In addition, equilibrium sorption isotherms were determined at 20 and 40 degreesC to compare sorption and desorption enthalpies. Both were in good agreement, confirming that desorption was not significantly different from sorption.