Objective

The objectives of this project were to determine if the GORE® Module, a passive groundwater sampler, can provide (1) technically defensible analytical data for volatile and semi-volatile organic compounds (VOCs and SVOCs) and (2) substantial cost savings when compared with conventional low-flow purging and sampling methodology. To achieve these objectives, the Gore technology was compared with conventional low-flow purging and sampling at two test sites: the Southern Bush River section of Aberdeen Proving Ground (APG), Maryland, and the former Pease Air Force Base in Portsmouth, New Hampshire.

Technology Description

At both test sites, the GORE Modules were placed at the same depth in the well as the pump or tubing used to collect the low-flow samples. Additional Modules were placed in most wells to allow the project team to profile analyte concentrations in the wells with depth. The wells were initially sampled with the GORE Modules (i.e., the pre-purge samples), then low-flow samples were collected, and finally a second set of Modules were collected (i.e., the post-purge samples). Analytes at the APG site included several chlorinated VOCs: tetrachloroethylene (PCE), cis-1,2-dichloroethylene (cDCE), trichloroethylene (TCE), 1,1,2,2-tetrachloroethane (TetCA), and chloroform. VOCs and SVOCs at the Former Pease site included: benzene, toluene, ethylbenzene, the xylenes (i.e., BTEX compounds), 1,2,4-trimethylbenzene (124TMB), 1,3,5-trimethylbenzene (135TMB), naphthalene, isopropylbenzene, and 2-methylnaphathalene.

Demonstration Results

The analyses of field duplicate Modules revealed that this method provided good reproducibility in most instances. For three of the analytes (TCE, TetCA, and Benzene) at APG, 90% of the replicate samples had a relative standard deviation (RSD) that was 20% or less. For the remaining analytes (PCE, cDCE, and chloroform), at least 70% of the duplicate pairs had a similar RSD. In instances when there was poor reproducibility, it was observed that this primarily occurred in a few wells after purging and where the upper portion of the screen was near the water table.

At the Former Pease site, reproducibility was very good for almost all of the analytes and 80% of the duplicate pairs had a RSD that was 20% or less. For three other analytes (benzene, ethylbenzene, and the xylenes), at least 60% of the sample pairs had a similar RSD. The poorest reproducibility was with toluene. It was observed that in instances where the reproducibility was poor this could be attributed to sampling three wells where the samplers had been left in the well for more than 2 hours and the depth below the water table for the samplers was at least 40 ft. It may be that leaving the samplers for more than 2 hours is too long a contact time especially given the sampling depth.

At APG, 10% of the samplers were also analyzed by an independent contract laboratory using the same analytical method used by the Gore Laboratory. There was excellent agreement between the analyte concentrations of the replicate samples analyzed by the two different laboratories for all the analytes that were compared.

With respect to the sensitivity of the sampling method, at the APG site the GORE Modules provided data that was below the action level, i.e., EPA’s maximum contaminant level (MCL) for drinking water. However, the detection capability of the low-flow method was one twentieth of that for the GORE Modules. Because some agencies require or recommend lower quantitation limits, it was recommended to the manufacturer that they continue work to develop a lower detection capability (i.e., prior to the next field trial).

Subsequently, at the Former Pease site the detection capability for the Gore method was comparable to that for the low-flow samples for most of the analytes (e.g., benzene, toluene, ethylbenzene, the xylenes, 124TMB, 135TMB, and naphthalene). For the remaining analytes (n-butylbenzene, n-propylbenzene, isopropyltoluene, isopropylbenzene, TCE, and 1,2-dibromoethane), the method detection limits (MDLs) were higher for the GORE Modules than they were for the low-flow samples. However, in all cases the MDLs were below one tenth of EPA’s MCLs.

Also at the Former Pease site, in many instances low concentrations of contaminants were detected when using the GORE Modules but not when low-flow sampling was used, even though these concentrations were well above the detection capability of the analytical method used for the low-flow samples. This data likely reflects enhanced sensitivity of the GORE method, and the project team recommends that this difference should be examined further.

The data for the mid-level samplers and the data for the mean concentrations for the three samplers (at three depths) were compared with the data for the low-flow samples for each of the analytes. At both sites and in all cases but one, there was a statistically significant linear relationship between the Gore data and the low-flow data. This relationship was typically one to one (i.e., the slope of the line was not significantly different from 1.0). The exceptions to a one-to-one relationship were TetCA and chloroform (mid-level data only) at APG and benzene and toluene at the Former Pease site.

Although there was generally good agreement between the Gore data and the low-flow data, plots of the GORE Modules with depth showed that there was substantial stratification of some contaminants with depth in some of the wells at both sites. This was especially true for the wells near a contaminant source. At APG, pronounced stratification of VOCs in a shallow well with a short (5-ft) screen was observed; analyte concentrations were up to 50 times higher in the upper Module than in the lowest.

Implementation Issues

With respect to where to place passive samplers within the well screen, there was good agreement between the mid-level sampler and the low-flow concentrations for some wells and thus placement of the sampler at the mid-point of the well screen would be advisable. However, in other instances, purging brought water into the well from a more permeable upper or lower zone and thus low-flow analyte concentrations agreed best with the upper or bottom sampler. While the mid-level sampler did not always best represent analyte concentrations obtained by low-flow sampling, the opposite is also true where low-flow sampling did not always collect the highest concentrations of contaminants in the wells and this is important to regulators.

The field crew found that the Modules were easy to use and did not require any special training. It was found that this sampling method was not time consuming, required very little auxiliary equipment or clean up, and there were less concerns with sample handling and safety. No scale-up constraints that would prevent wide-scale use of this technology are foreseen. These samplers can be used in any well or piezometer that is larger than 0.25 inches in diameter. 

The long-term monitoring (LTM) costs were determined for each sampling method for 10 years using the initial start-up costs, annual field sampling costs, annual sample processing and analyses costs, and the estimated operations and maintenance costs over the 10-year period. For the GORE Modules, it was determined that 99.75% of the total 10-year LTM cost is associated with sample collection, and 85% of that cost is the price of the samplers (labor is the other 15%). For low-flow sampling, sample collection accounts for 45% of the total LTM costs and of that amount, 93% is labor. Laboratory analyses account for another approximately 25% of the total LTM costs, and the start-up costs (including dedicated pumps and the equipment for measuring purge parameters) account for less than 10% of the total LTM costs. Although estimated cost was determined to be lower with the GORE Modules, the degree of the cost savings depends heavily on the price of the samplers. As an example, cost savings of approximately 30 to 45% can be achieved if the price of the Modules is around $190 per sampler.