Soil Processes and Chemical Transport A Flow-Through Reactor with an Infinite Sink for Monitoring Desorption Processes Dirk Freese,* Peter G. Weidler, Daniel Grolimund, and Hans Sticher ABSTRACT The assessment of elemental release processes in soils are of large interest in soil research. Atechnique is presented where the advan- tages of a flow-through reactor arecombined with a specific infinite sink. Between the two reactors an electrolyte solution (pH 6) acting as a transport carrier was circulated continuously in a closed-loop set- up. To demonstrate the capability of the new technique, phosphate desorpfion kineticsfrom differentsynthetic Fe and Mn hydroxides coated on silica sand were studied. The relative amount of P desorbed increased in the order ferrihydrite (10% of the initially sorbed P) goethite (20%) << hematite (65%) < Mn dioxide(80%). We suggest that this technique can also beadapted to other issues,for example, desorption of other anions and cations from solid phases (soils and sediments) because the composition of the infinite sink is variable. D ESORPTION and diffusion are major processes con- trolling long-term release of elements and sub- stances from solid soil phases into the soil solution. The study of desorption processes is often limited by low concentrations of the released element of interest and rapid achievement of equilibrium between solid and liquid phases. Different methods of assessing release processes in soils have been developed in the past that can be summa- rized in four main categories: simple dilution methods in batch vessels, resin-exchange methods, flow-through methods, and infinite sink methods (Amacher, 1991; Freese et al., 1995; Chardon et al., 1996). Infinite sink methods are based on the idea that the desorbing species is continuously removed from solu- tion providing a high concentration gradient between the solid and liquid phase as a driving force. The infinite sink technique has been successfully adapted to phos- phate studies in soils. Iron-oxide impregnated filter pa- per (Pi-method) were used to assess the amount of re- versibly adsorbed P in soils (Van der Zee et al., 1987; Menonet al., 1988). The method has two major draw- backs: (i) the paper strips are mechanically unstable and (ii) part of fine soil materials adhere. An advance was madeby Freese et al. (1995), who used a dialysis mem- brane tube (DMT-method) filled with a ferrihydrite sus- pension as an infinite sink to monitor long-term desorp- tion kinetics of P in soils. Short-term kinetics could not be measured because of the dialysis membrane have D. Freese, Humboldt-Univ. to Berlin, Dep. of Soil Science, D-10115 Berlin Invalidenstrasse 42, Germany; and P.G. Weidler, D. Groli- mund, and H.Sticher, Swiss Federal Institute of Technology Zurich, Institute of TerrestricalEcology, CH-8952 Schlieren, Grabenstrasse 3, Switzerland. Received 19 Nov. 1997. *Corresponding author (dirk. freese@agrar.hu-berlin.de). Published in J. Environ. Qual.28:537-543 (1999). acted as an diffusion barrier for P at the beginning of the desorption process. Recently, Grolimund et al. (1995) introduced a simple flow-through reactor for the measurementof adsorption isotherms on soils. This technique was described as a hybrid between batch and column experiments that can be a useful tool in different ways, such as precondi- tioning of samples with background electrolyte solution or removal of competitive ions before adsorption exper- iments. In general, flow-through designs are steady-state systems that have the advantage that the experimental parameters in the reaction chamber can be kept con- stant, for example, concentration of the effluent solu- tion, pH, flow rate (Van Riemsdijk and Vander Linden, 1984; Glasauer et al., 1995). The chief advantage of flow- through methods is the continual removal of reaction products and therefore readsorption or competition ef- fects of ions are minimized (Miller et al., 1989; Amacher, 1991; Backes et al., 1995). Van Grinsven and VanRiems- dijk (1992) gave an example of measuring long-term weathering rates in soils by flow-through techniques and pointed out that these techniques were much more suitable than any batch technique. Nevertheless, there was some evidence that the percolation rate may affect the weathering rate, but this phenomenoncould not be explained. Thin-disk methods pertain also to the group of flow- through methods with different applications. In princi- ple a solution is continously pumped through the thin- disk (a dispersed solid phase, <1 mm thickness, placed in a filter holder) and effluent aliquots are collected by a fraction collector. Miller et al. (1989) compared batch- and flow-generated (miscible displacement) adsorption isotherms of anions on mineral phases. Thabet and Selim (1996) used a thin-disk methodfor the determina- tion of cation exchange isotherms in binary systems and the exchangekinetics. The kinetics of the exchangereac- tions (equilibrium was reached within 25 min) could only be assessed if parameters concerning to mass trans- fer (dispersion/diffusion) were considered. Thin-disk methods were also applied to study desorption kinetics of trace metals from minerals (Backes et al., 1995; McLaren et al., 1998). The authors point out that two main problems occur using thin-disk methods for de- sorption studies: (i) preferential flow through the mate- rial on the filter and the occurence of diffusion, which can introduce artifacts and (ii) the residual solution the solid after sorption may influence the amount of desorbed species of the first desorption step. For further Abbreviations: DMT, dialysis membrane tube; XRD, x-raydiffraction; FTRD method, flow-through reactor desorption. 537 Published March, 1999