Partitioning Interwell Tracer Test (PITT) applies two kinds of tracers, one is the partitioning tracer which penetrates into Non Aqueous Phase Liquid (NAPL) nuggets contaminating the subsurface zones, the other is the non-partitioning tracer which evades penetrating into it. By using these two kinds of tracers, PITT is potentially capable of estimating the quantity of NAPL. Since this chromatographic separation of the tracers is based on the partition law, it is most important to determine the partitioning coefficients of the tracers. However, the accuracies of the PITT would be influenced not only by the partitioning coefficients, but also by the conditions of NAPL contamination and tracer velocities. The effects of these factors on the accuracies of PITT have not been confirmed sufficiently up to now, probably due to the lack of the basic research compared with the practical requirement. The questions are 1) what is the effect of the flow velocity of tracer materials on the partition phenomena, and 2) what is the effect of the NAPL nugget sizes on the partition phenomena. We used Toyoura-sand as porous medium packed in columns, Trichroloethylene (TCE) as a NAPL, Isopropyl alcohol (IPA) as a non-partitioning tracer, and 4-methyl-2-pentanol (4M2P) and 5-methyl-2-hexanol (5M2H) as partitioning tracers which have different partition coefficients. Five types of tracer flux were used to answer the above question 1), and three kinds of NAPL nugget sizes were used to answer the question 2). To change the NAPL nugget size, we filled the porous medium made of kaolinite that absorbed TCE into columns with Toyoura-sand. From the experiments where five tracer velocities were set up, the peaks of the break through curves (BTCs) of the partitioning and non-partitioning tracers were not separated when tracer flux was larger (0.065, 0.088, 0.091 cm min^-1). The tailing of the BTCs of partitioning tracers were separated into two stages. These separations were more clearly explained when the BTCs were plotted in semi-log scales. The first stage of the tailing may be attributed to the insufficient partition of the tracers and the second stage may be due to more sufficient partition implementation.

In the experiments of changing the NAPL nuggets sizes, the separation of BTCs was larger when the NAPL nuggets sizes were smaller. The tailings were longer when the NAPL nuggets sizes were larger. It is practically important that the degree of separations of the BTC-peaks and the lengths of the tailing are sensitive to the sizes of NAPL nuggets embedded in the subsurface zones.