Residual fluid dissolution fingering in porous media:

Experiment in a water-air system

 

FRITZ STAUFFER, OSCAR OSORIO and WOLFGANG KINZELBACH

 

Institut für Hydromechanik und Wasserwirtschaft

Eidgenössische Technische Hochschule Zürich, CH-8093 Zürich, Switzerland

Ph.: +41-1-633 30 79; Fax: +41-1-633 10 61; email: stauffer@ihw.baum.ethz.ch

 

 

ABSTRACT

The dissolution of immiscible fluid entrapped at residual saturation in porous media by flushing is investigated for the case of a water-air system, which represents an analogy to the dissolution of residual NAPLs in groundwater. A quasi-saturated sand packing at residual air saturation was horizontally flushed with deaerated water. Visible dissolution fronts could be discerned at the tank walls and their evolution was recorded. The dissolution fronts soon showed distinct fingers that increased in length from zero up to 19 cm over a total horizontal flow distance of 70 cm. These fingers were similar to the dissolution fingers described by Imhoff et al. (1996) for residual NAPLs. Tracer tests without dissolution effects were performed on the sand packing at residual air saturation. They showed that the packing was slightly horizontally layered and that the positions of the dissolution fingers in the sand packing coincided with zones of higher velocity. However the length of the dissolution fingers was more than two times larger than the length of the corresponding tracer fingers. Although similar results can be expected for the dissolution of residual NAPLs in groundwater, their applicability has to be further verified.

 

Keywords: Groundwater, residual air saturation, dissolution, hydraulic conductivity, inhomogeneity, fingering, analogy to NAPL-water systems

 

INTRODUCTION

Many groundwater contamination problems are caused by a release of immiscible liquids (Non Aqueous Phase Liquids, or NAPLs) into the subsurface, e. g., by a discharge of chlorinated hydrocarbons. The NAPL first migrates through the unsaturated zone until it reaches the saturated zone. After the bulk of the NAPL has migrated through the subsurface, a substantial fraction will remain trapped in pores. In saturated groundwater systems, these liquids are subject to a series of physical mechanisms. An important part of them remains immobile in the porous medium at a residual saturation, forming a source of long-term contamination of the groundwater caused by the low dissolution rate. The rate and manner in which a NAPL dissolves in groundwater will determine the magnitude and duration of contamination. A good understanding of the dissolution process contributes to the evaluation and optimisation of remediation measures.

Several theoretical and experimental studies have been performed on the dissolution of trapped fluids in porous media (e. g., Powers et al., 1994; Imhoff et al., 1994). In previous experiments (Imhoff et al., 1996), it was observed that the dissolution of residual NAPLs leads to instabilities causing fingers in the moving dissolution front.

This study is conceived as an experimental investigation of the dissolution of residual air in a quasi-saturated sand packing by flushing. The main focus is on the detection of the temporal evolution of the visible dissolution front. The experiment is an analogy to the dissolution of residual NAPLs in groundwater and is much more conveniently performed than a corresponding experiment with NAPLs.

 

EXPERIMENTAL SET-UP

The experiment was conducted in a Plexiglas tank with inside dimensions of 78 cm in horizontal length, 30 cm in vertical height and 8 cm in width. The tank was subdivided into three domains, an inlet reservoir, a 69.4 cm long porous packing, and an outlet reservoir. The packing was confined by two perforated vertical aluminium walls with a sandwiched polyamid web. A pump was placed between a storage tank for deaerated water and the inflow reservoir. The outlet reservoir was connected to a constant head vessel. Therefore a horizontal flow could be established in the porous packing. The sand used for the packing was a medium uniform quartz sand of the grain size 0.5-0.75 mm.

 

EXPERIMENTAL METHOD

The sand was packed by using a falling strands method. The method consisted of letting dry sand fall through a 4 cm diameter tube with a length of 51 cm, equipped with three sieves at the bottom. The instrument was slowly moved horizontally during the packing procedure. The method may cause a slight horizontal layering of the sand packing. The sand packing had an average height of 27.1 cm.

The dry sand packing was slowly flushed with tap water at room temperature until quasi-saturation was obtained, the trapped air remaining in the medium. The water weight in the tank was measured and the residual air saturation determined. Water was deaerated during 10 min in a 2.5 l volume container using a vacuum pump and then pumped to the storage tank. At the beginning of the experiment the volume of water in the space between the inlet and the sand packing was replaced by deaerated water. The sand packing was horizontally flushed with deaerated water at a rate of 87 ml/min. Periodic measurements of the inlet and outlet water levels as well as of the outflow were performed in order to determine the temporal development of the hydraulic conductivity. The advance of the dissolution front was optically observed at the tank walls by means of an illumination that facilitated the visualisation of the air bubbles zone.

Two tracer tests without air dissolution were performed on the quasi-saturated sand packing. A solution of 100 ppm Eosin dye solution without deaeration was used as tracer and introduced into the system as a long pulse input (tracer breakthrough {test I} and subsequent flushing with tap water {test II}) at the same flow rate as the deaerated water in the dissolution experiment. The visible tracer fronts were periodically observed and registered on both sides of the tank.

 

RESULTS AND DISCUSSION

The measured and calculated hydraulic parameters of the experiments are listed in table 1. The sand packing was flushed with 39 pore volumes (PV) of deaerated water. It was observed that air bubbles continuously shrank due to dissolution, until they disappeared within a zone of less than 1 cm of length. Figure 1 shows the temporal development of the dissolution front, with the inlet at x = 0, and the outlet at x = 69.4 cm. After the packing was flushed with about 2 PV of water (t = 135 min), a clear dissolution front with a rather uniform pattern was visible. However, at the points y = -10 cm, +2 cm and +8 cm the dissolution was more rapid than in the rest of the front. After 4 PV of water (t = 270 min), small fingers clearly developed with lengths between 1 and 3 cm. The finger length increased rapidly as the flux of water continued. After 39 PV of water (t = 2485 min), the length of the fingers ranged between 9 and 19 cm. The comparison between the left and right tank sides shows that the general structure was rather similar. Five main peaks were discerned on both sides. A measure of the finger length l_f can be defined as half the distance between the peak x-position minus the x-position of the next depression above. The development of l_f over time is shown in figure 2. The growth of most fingers was approximately proportional to time. The mean thickness of all fingers remained almost constant during the experiment (between 1 and 4 cm). The thickness to length ratio of the fingers remained at about 1 to 5. The mean front velocity was found to be 3.3×10^-4 cm/sec and was almost constant. The front on the right side of the tank was slightly slower than on the left side. The final mean hydraulic conductivity was nearly doubled due to the dissolution of the entrapped air.