USEPA - Technology Innovation and Field Services Division (TIFSD)

EPA studies identify techniques for critical leak testing prior to soil-vapor sampling

Researchers at EPA’s National Risk Management Research Laboratory (NRMRL) in Ada, OK, are developing quality assurance (QA) measures for soil-gas and sub-slab sampling methods that help differentiate contaminant vapors due to vapor intrusion from background sources. Recent research focused on measures to identify leakage of ambient air into conventional vapor probes, which can significantly impact sampling result.

During sub-slab or soil-gas sampling, ambient air may enter the sampling vessel (e.g., sampling bag or canister) through loose fittings connected to the probe or through openings or cracks in the concrete and bentonite seals used to isolate screened intervals. If leakage occurs and gas concentrations at the point of leakage are less than soil-gas concentrations, concentrations measured in a sampling vessel will be less than true concentrations in the subsurface. In the absence of leak testing, leakage is assumed to have occurred if anomalous results are observed; otherwise, measurements are assumed to be valid.

The QA measures for soil-gas leak detection involve use of a small, sealed chamber that can be integrated into an above-ground sampling train. Placement of the chamber directly on top of a sub-slab or soil-gas probe enables each component of the sampling train to be vacuum or pressure tested with a gas tracer such as helium (He). Valves or gas-tight, quick-connect fittings are used to isolate each component during testing and containerize the vapor.

To test integrity of fittings for the leak detection chamber, NRMRL used a peristaltic pump to create a vacuum of 97.3 kPa in the flowmeter and tubing of a laboratory-deployed sampling train. The vacuum was initially recorded every second and then relaxed to every 120 seconds as it gradually dissipated over 35 hours. The Ideal Gas Law was used to calculate flow rate into individual components of the sampling train. Results indicated nearly a complete vacuum, with leakage of only 0.72 standard cubic centimeters per minute. NRMRL considers leakage less than 1% of the flow rate to be insignificant and below the detection limit.

Similar QA measures can be used in the field to leak test boreholes. One method is to flood the detection chamber surrounding the top of a borehole with a gas mixture containing a tracer (usually He). Concentrations in the chamber and line or in the sampling vessel are then monitored. Typically, He is injected into the chamber as a pure gas, and a portable thermal conductivity detector (TCD) is used to monitor the sampling train. The density of dry, pure-phase He is only 0.16 g/L at 20°C, while the density of soil gas typically exceeds 1.2 g/L. As a result, pure-phase He is buoyant and will not be drawn down a compromised borehole without sufficient vacuum in a screened interval.

NRMRL developed a heuristic model (Figure 1) to provide a conceptual understanding of borehole leakage. Only vertical compressible gas flow is allowed down a compromised borehole having an integrated gas permeability of k1. The integrated permeability of the borehole incorporates the presence of cracks and openings in and around an essentially impermeable matrix of concrete and bentonite. Only radial compressible flow is allowed to a screened interval in a homogeneous isotropic medium having a gas permeability of k2.

These calculations indicate that leakage is primarily a function of the permeability contrast between the formation and borehole. As the ratio of formation to borehole permeability decreases, the potential for leakage increases. The potential for leakage then is greatest in soil having low gas permeability.

The leak detection chamber was integrated into a borehole sampling train used to evaluate past releases from underground storage tanks at an automotive station in Green River, UT. Onsite soil consists primarily of clay, and the soil gas consists of 79.8% N2 and 20.2% O2, with a calculated density of 1.19 g/L at 20°C and 100% relative humidity. To minimize the effect of buoyancy, a gas mixture of He and argon (Ar) was injected to achieve a near-constant gas mixture inside a chamber of 42% Ar, 21% He, 7.8% oxygen (O2), and 29.2% nitrogen (N2) with a calculated gas density of 1.15 g/L at ambient temperature. A portable TCD and landfill gas meter were used to measure He and O2, carbon dioxide and methane in the sampling train and specifically within the chamber, respectively. QA tests indicated nearly 100% leakage in the borehole, which in turn prompted evaluation of whether to redesign or abandon the borehole.

NRMRL recommends that leak testing always precede soil-gas sampling, especially in media of lower permeability, until the integrity of a borehole is well established. All components of the sampling train should be vacuum or pressure tested with quantified flow into or out of the system prior to sample collection. For leak testing of conventional probes, a chamber containing a gas mixture approaching the density of soil gas should be used, and tracer concentrations in the chamber should be held constant in order to quantify the leakage.

NRMRL has developed purge and transient gas permeability test measures that can be used simultaneously with the leak detection methods. The concurrent testing process relies on multiple tracers introduced into multiple intervals of soil-gas probe clusters.

Read the full article online at

Customer comments

No comments were found for EPA studies identify techniques for critical leak testing prior to soil-vapor sampling. Be the first to comment!