BENCH-SCALE MODELS OF DYE BREAKTHROUGH CURVES

Fluorescent dye breakthrough curves obtained from quantitative traces in karst flow systems record multiple processes, including advection, dispersion, diffusion, mixing, adsorption, and chemical reaction. Recently developed laboratory techniques are measuring breakthrough curves of small, bench-scale physical models in an attempt to isolate, understand and quantify some of these processes.

Our initial apparatus is a closed Pyrex glass mixing chamber containing a magnetic stirrer with Pyrex glass inlet and outlet tubes. The internal volume of the glass network is 165 mL and is connected to a constant flow apparatus. Dye is injected with a syringe, flows through and mixes in the chamber. The mixture is collected in vials and measured in a scanning spectrofluorophotometer. The purpose of the apparatus is to eliminate or at least minimize any adsorption or chemical reaction during tracer breakthrough, allowing for characterization of hydrodynamic processes occurring along the flow path.

Independent variables in the model are flow rate, stirring rate, and dye concentration. Our initial runs have varied flow rates of 1 to 4 ml/s with constant stirring rates. Complete breakthrough curves are obtained in 5 to 10 minutes. The initial runs yield realistic-looking breakthrough curves with steeply rising leading edges, a peak and a single, asymmetric, exponential tail. The slope of the exponential tail is inversely proportional to the flow rate.

If the tails of quantitative dye traces in karst systems are monitored long enough, they often display, in semi-log space, two or more progressively less negative slopes. The limited data we currently have shows single slopes on the tails of the curves from the bench-scale models. If this observation is correct, the initial steepest negative slope of the breakthrough curve tail is a significant function of the flow rate through the system. The subsequent shallower slopes must be due to other processes.

A small volume fluorometric detector that would allow higher time-resolution breakthrough curves to be obtained would be significant improvement to the bench-scale apparatus.