Surfactant‐Induced Unsaturated Flow: Instrumented Horizontal Flow Experiment and Hysteretic Modeling

Abstract

Organic compounds with surface‐active properties reduce the surface tension of water and can change soil water contact angles depending on their aqueous concentration. The presence of such compounds in the unsaturated zone affects soil moisture characteristics and unsaturated hydraulic conductivity relations, and can cause flow as a result of induced soil water pressure gradients. Some studies have used horizontal column experiments to measure the surfactant‐driven movement of water in unsaturated media; however, they have all used destructive sampling methods to determine water contents and did not measure soil water pressures. We attempted to gain better insight into this unsaturated soil water flow process through continual monitoring of water contents with time domain reflectometry and soil water pressures with pressure transducer equipped tensiometers. One half of the horizontal one‐dimensional flow cell was prewetted (but unsaturated) with water while the other half contained the same fluid content of 7% butanol solution. The temporal and spatial water content and soil water pressure information improved our understanding of the surfactant‐induced flow, including perturbations associated with solute gradients. Backflow in the flow cell was observed at later times as the water‐content‐induced component of the hydraulic gradients became more important. Hysteretic numerical simulations of the one‐dimensional horizontal flow cell using a modified version of HYDRUS 5.0 including coupled flow and transport through concentration‐dependent surface tensions were performed for the case in which butanol is the surface‐active compound. The numerical simulations, which used independently measured flow and transport parameters, provided a good fit to the experimental data and provided further insight into the induced flow behaviors.

Chemically Induced Flow in Contaminated Unsaturated Soil

“…Henry et al (1999) used a horizontal column 13.7 cm long, half wetted by deionized water and half wetted by a 7% butanol–water mixture, both at 18% water content. Similar experiments were conducted in a horizontal column 40 cm long, 10.2 cm wide, and 4.5 cm high, with Ottawa graded sand and a 7% butanol–water mixture at 15% initial water content by Bashir et al (2008). In those works, water was repelled from the contaminated half because butanol is a highly soluble organic substance that reduces σ from 72 to 26 mN/m at 25°C.…”

“…Experiments by Henry et al (1999) and Bashir et al (2008) (Fig. 4 and 5, respectively) show that water moved away from the butanol‐contaminated side toward the water side.…”

“…Four numerical analyses (T1–T4) were implemented to account for increasing model complexity, model comparison, and the number of estimated parameters. Specifically, T1 accounted for only the solute effect on the physicochemical properties (ρ X , σ X , and μ X ) but excluded osmotic potential ψ o and adsorption F and was used only with data from Henry et al (1999) and Bashir et al (2008) to initially assess the model; T2 allowed the hydraulic parameters in Eq. [6] to be adjusted to improve fitting against the experiments with butanol– and glycerol–water mixtures but excluded osmotic potential and adsorption; T3 extended T2 by including osmotic potential ψ o as in Eq.…”

“…Hence, water flow in unsaturated soil where water also transports dissolved chemicals can be affected not only by the osmotic pressure Π but also by other induced effects on the capillary pressure and hence on the soil matric potential ψ m . Both experimental and numerical studies conducted in horizontal unsaturated sand columns half wetted by water and half wetted by water and organic surfactants have attributed the causes of water flow to the difference Δψ m at the interface of the two halves induced by surface tension effects (Tschapek and Boggio, 1981; Bashir et al, 2008; Henry et al, 1999) and contact angle changes (Demond and Roberts, 1991), showing that these could move up to 50% of the water (Karkare and Fort, 1993). However, effects caused by the osmotic potential were not investigated and were excluded from modeling.…”

“…Then, the Richards equation for unsaturated flow was modified to account for both processes and to analyze the condition that can induce backflow during the transient toward equilibrium. The model was first validated on experimental data from the literature (Bashir et al, 2008; Henry et al, 1999) to describe and assess the effect of osmotic flow induced by a butanol–water mixture. The model was next used with our experimental data related to a glycerol–water mixture: a flow analysis, a parametric sensitivity analysis, and a thermodynamics interpretation were conducted to complement our observations.…”