Non-aqueous Phase Liquid Spills in Freezing and Thawing Soils: Critical Analysis of Pore-Scale Processes

Abstract

The frequent use of non-aqueous phase liquids (NAPLs) in cold regions creates serious risks of soil and groundwater contamination. NAPL contaminants can stay in soil for long times due to their entrapment by strong interfacial forces, resulting in a source of pollution caused by their slow dissolution in groundwater over decades. The presence of these contaminants in ice-containing soils creates a four-phase problem, which must be fully understood to design/develop or improve remediation techniques in cold climate regions. In this review, the fate and transport of NAPL contaminants in periodically freezing-thawing and frozen soils is discussed, with emphasis on pore-scale processes. Three topics are identified for future research focus: (a) study of the dynamics of NAPLs during freezing and thawing of soils using non-destructive imaging techniques, and the effect of various factors, including wettability, pore size, consolidation of porous matrix, and fluid properties; (b) investigations of the fate and transport of NAPL contaminants in frozen soils with different wettabilities, and the effect of the spatial distribution of ice clusters on NAPL retention and movement; and (c) pore-scale modeling of the fate and transport of NAPL spills in freezing-thawing and frozen soils. This will lead us towards a complete pore-scale understanding of NAPL spills in cold climate soils, and their fate and transport over time.

KEY WORDS:

• cold regions
• freeze-thaw
• ganglia
• NAPL
• non-aqueous phase liquid
• porous media
• residual saturation
• soil contamination

ACKNOWLEDGMENTS

This work was supported by two Faculty Research Support Program grants, and Kamaljit Singh was supported by two scholarships and a travel grant, from the University of New South Wales. The authors also acknowledge the research group of Prof. Mark Knackstedt, Department of Applied Mathematics, Australian National University, for use of their X-ray tomography facility. They thank David Sharp, Jim Baxter, Wayne Jealous, and Vince Craig for experimental support; Adrian Sheppard, Andrew Kingston, and Ajay Limaye for software support; and the Supercomputing Facility at Australian National University and the Australian Partnership for Advanced Computing for computing time.

Notes

1. Residual saturation of NAPL is expressed as the ratio of NAPL volume in pores to the total pore volume.

2. There has been extensive research on the simulation of NAPL spills in the laboratory, especially to establish NAPL residual saturation in saturated and unsaturated porous media, the results of which vary with the porous medium and NAPL properties. To obtain the residual non-wetting fluid saturation, the wetting fluid is first displaced by non-wetting fluid until the effective minimum wetting-fluid saturation (irreducible saturation) is obtained, at which no more wetting fluid can be displaced. This process is called drainage. This is followed by the displacement of non-wetting fluid with wetting fluid at a particular capillary number (refer to section 2.2.1.4), in a process called imbibition, until a minimum non-wetting fluid saturation (residual saturation) is reached, where no more non-wetting fluid can be displaced.

3. Displacement of a non-wetting fluid by a wetting fluid is called imbibition (Bear, 1972).

4. Other variants of the Bond number based on intrinsic permeability (k) are also used in the literature (Chevalier and Fonte, 2000; Morrow and Songkran, 1984; Wilson and Conrad, 1984): . Some researchers have included a porosity term in the capillary and Bond numbers (Dawson and Roberts, 1997). In this case, the additional dimensionless term with a value less than one would change these numbers significantly (Chevalier and Fonte, 2000). The previous formulations are based on a zero contact angle (θ = 0°) at the solid-water-NAPL interface; nonzero contact angle relationships are also included by some authors (Dawson and Roberts, 1997; Li et al., 2001; Pennell et al., 1993; Pennell et al., 1996): & .

5. Each three-dimensional volume element in tomography is called a voxel.