Environ. Sci. Technol. 2002, 36, 1592-1599
Diffusive Partitioning Tracer Test for Nonaqueous Phase Liquid (NAPL) Detection in the Vadose Zone DAVID WERNER AND PATRICK HO ¨ HENER* Swiss Federal Institute of Technology (EPFL), ENAC-ISTE-LPE, CH-1015 Lausanne, Switzerland
This paper proposes the theory and practical application of a new partitioning tracer test for nonaqueous phase liquid (NAPL) detection in the vadose zone, which is based on diffusion. A mixture of chlorofluorocarbons as gaseous tracers is injected into the vadose zone to form a point source at the injection point. While the tracers diffuse away, small volumes of gas are withdrawn from the injection point. The quantitative determination of the NAPL saturation is based on a comparison of the concentration decline of tracers with different air-NAPL partitioning coefficients. The test has been evaluated in laboratory sand columns contaminated with dodecane. NAPL in saturations of 0.84% of the total porosity have been quantified in a wide range of different water contents. Actual and measured NAPL saturations calculated as an average from four different tracer pairs agreed within (30%. The new method was successfully used for repeated NAPL quantification in a largescale field lysimeter contaminated with artificial kerosene. This rapid and inexpensive test is potentially of value for site investigations especially in combination with soil gas measurements, because it requires similar equipment. Possible applications are source delineation and repeated NAPL quantification in situ during a remediation.
Introduction Knowledge of the presence and quantity of nonaqueous phase liquids (NAPLs) is crucial for the management and remediation of contaminated sites (1). During the migration of NAPLs through soil, a certain amount of liquid will be retained in the soil by capillary forces. This fraction is known as residual saturation and may occupy ∼2-20% of the available pore space (2). The presence of NAPLs defines the so-called source zone (3), from which gaseous and aqueous contaminant plumes are formed. Source zone delineation is a key procedure at any NAPL contaminated site. Delineation of source zones with traditional sampling techniques (e.g., soil cores) presents considerable difficulties and costs (4). Innovative methods for NAPL detection based on partitioning tracers have been introduced and evaluated under laboratory and field conditions for the saturated and unsaturated zones (see ref 5 for a review). Partitioning tracers may be naturally occurring gases such as 222Rn (6) or a variety of harmless organic chemicals (5) with different affinities for NAPLs. The general principle of partitioning tracer tests is that of chromatography (7): a mixture of tracers is injected into a stream of gas or liquid (water) created in the subsurface. The * Corresponding author phone: 0041 21 693 57 50; fax: 0041 21 693 28 59; e-mail:
[email protected]. 1592
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 7, 2002
different tracers partition into the stationary NAPL phases or other phases to different degrees, and the resulting difference in the migration of the tracers is used for the location and quantification of the phases of interest. The field application of this concept is called a partitioning interwell tracer test (PITT; 8). An advective flow field in either the gas (vadose zone) or groundwater (saturated zone) is created between two or more wells, and the information obtained from the tracer data reflects the zone between those wells. With an advective flow field, subsurface environments of considerable spatial extensions can be assessed with a single tracer test. Practical problems associated with such a setup are the incorrect interpretation in heterogeneous subsurface environments, especially when NAPLs are located in zones with low permeabilities that are bypassed by the tracers (9), and short tracer residence times with respect to partitioning kinetics. For the vadose zone, the generation of flow fields creating long residence times is difficult and biased by air pressure variations. Field applications of long duration are costly. The measurement of breakthrough curves requires a high number of analyses and the use of sophisticated analytical equipment on site. We propose a new partitioning tracer test for the vadose zone based on diffusion rather than advection, which we will call a diffusive partitioning tracer test (DPTT). This test obtains local information on NAPLs in the vadose zone. The investigation can be performed with the equipment used for soil gas surveys. The tracer mixture is injected by a soil gas profiler and creates a point source. After the injection, the tracers diffuse into a larger spherical zone. Slower diffusion velocities of those tracers partitioning into NAPLs indicate the presence of an NAPL. Johnson and co-workers (10) reported a test for the in situ determination of effective vaporphase porous media diffusion coefficients, using SF6 as the sole tracer gas. In their study they consider partitioning of SF6 into the stationary phases as negligible. Our approach includes two or more gaseous tracers with known partitioning behavior into the stationary phases. The objective of this study is to develop the theory underlying the DPTT and to test the new method in laboratory experiments and in a largescale lysimeter study.
Theory Partitioning of Gaseous Tracers in Soil. The pore space of the soil is described as a three-phase system consisting of air, water, and NAPL. Assuming an instantaneous linear equilibrium between these three phases and the solid, the partitioning of a gaseous tracer can be described using the air-water partitioning coefficient or Henry coefficient H, the air-solid partitioning coefficient Ks, and the air-NAPL partitioning coefficient Kn (see Figure 1). Dimensions and definitions for the notations are given under Notation. The fraction of the gaseous tracer fa in the soil air can be calculated as
fa )
1 Fs(1 - θt) θw θn 1+ + + Ksθa Hθa Knθa
(1)
where θa, θw, θn, and θt denote the air-filled, water-filled, NAPL-filled, and total porosity and Fs denotes the density of the solids. The air-solid partitioning coefficient Ks can be interpreted as the ratio of the Henry constant H and the distribution coefficient Kd between the solid and the water phase. The inverse value of fa is equal to the retardation 10.1021/es010098x CCC: $22.00
2002 American Chemical Society Published on Web 02/22/2002
at distance x from the plane source at time t after the release is given by
Ca(x,t) )
m 0 fa 2AθaxfaDmτπt
exp[-x2/4faDmτt]
(5)
The concentration at x ) 0 is given by
FIGURE 1. Equilibrium consideration for soil gas tracers in an NAPLcontaminated soil. coefficient R ) fa-1, a quantity used for the interpretation of advective partitioning tracer tests (5). For the gaseous tracers used in this study with a Henry coefficient >5, the third term in the denominator is much smaller than 1 for a wide range of soil water contents. To simplify the expression further, also the second term is neglected, which accounts for sorption to the solids, a simplification that will be discussed in the following sections. For two gaseous tracers, termed tracers 1 and 2, eq 1 can then be written as follows:
fa,1 ≈ fa,2
θn Kn,2θa θn 1+ Kn,1θa
(2)
θn ≈ θt
[
[
]
fa,1 -1 fa,2
m0fa 2AθaxfaDmτπ
×
1 xt
(6)
For two different compounds one finds
[
]
fa,1 m0,2Ca,1(0,t) 2Dm,1 ) fa,2 m0,1Ca,2(0,t) Dm,2
(7)
Note that this ratio does not depend on soil physical parameters such as the tortuosity factor. The gas-phase fraction ratios fa,1/fa,2 are calculated according to eq 7 and used to derive the NAPL saturation according to eq 3. To compare the measured data with the mathematical model, a quantity Nplane(0,t) is defined as follows:
1+
Nplane(0,t) )
Ca(0,t)xDm f 0.5 a 1 ) × m0 2Aθ xτπ xt
(8)
a
By transforming eq 2, one obtains the NAPL saturation Sn. The residual NAPL saturation in soils is typically 98%. In a field lysimeter, artificial kerosene made up from 14 different compounds was used as NAPL (15). Sands. Two alluvial coarse sands were used in this study: In all laboratory experiments, a sand obtained from Masson SA, Renens, was used. It originated from a pit on the southern shore of Lake Geneva and had a total organic carbon content (foc) of 0.06% (weight percent of dry sand). Grain size distribution was as follows: 53 min (point source) fell on straight lines when plotted according to the theory and measured concentration ratios varied by
