• Dissolved-phase contaminant signature may be a significant issue for hydrogeological investigations involving coalescing groundwater contaminant plumes from multiple sources, as well as for groundwater treatment system design; and

• Evaluation of natural attenuation of contaminant plumes may necessitate collection of indicator-level and inorganic groundwater quality data from monitoring wells in which LNAPLs are present.

This paper introduces an ice-coating method for collecting groundwater samples from beneath LNAPLs and presents results from a field trial in which this method was used at a site in central New Hampshire.

ICE-COATING METHOD

Ice is used as a temporary barrier to protect sampling implements from becoming product coated as they pass through LNAPLs within monitoring wells. Sampling implements are coated with an approximately 0.1 to 0.3-inch-thick layer of ice (laboratory-grade distilled water) using simple molds fabricated from polyvinyl chloride (PVC) pipe and end caps. Bench-scale testing of two different ice-coating procedures demonstrates that product initially coats the ice, but sloughs off within seconds as the ice begins to melt. The ice coating melts completely within a few minutes and the product-free implement is used to sample groundwater. Melting ice is expected to have a negligible effect on groundwater quality due to the minimal volume of ice relative to the storage capacity of most monitoring wells. If the impact of melting ice on groundwater quality is a concern, the standing water column could be purged or the well could be allowed to equilibrate prior to sampling. Two different ice-coating procedures are described below for sampling beneath LNAPLs.

CONDUIT PROCEDURE

This procedure involves placing a silicon stopper in one end of a Schedule 40 PVC pipe and ice coating the end of the PVC pipe containing the stopper. The ice-coated pipe is lowered through the LNAPL until the stoppered end of the PVC pipe extends at least 3 feet into groundwater (Figure 1A). Following melting of the ice coating, a messenger rod is used to push the stopper from the end of the PVC pipe, creating a portal in the LNAPL through which sampling may be performed. A monofilament line attached to the stopper allows retrieval of the stopper from the well bore at the time the conduit is retrieved.

DIRECT COATING PROCEDURE

This procedure involves coating a Waterra™ 1-inch outer diameter thermoplastic Standard D-25 inertia pump foot valve with ice (with the exception of the threaded end), connecting it to a 1/2-inch outer diameter high density polyethylene tubing, and then lowering the inertia pump through the LNAPL until the foot valve is about 4 feet into groundwater. As with the Conduit Procedure, ice serves as a temporary barrier that protects the inertia pump from product (Figure 1B). After complete melting of the ice coating, the product-free inertia pump is used to sample groundwater below the LNAPL.

FIELD TRIAL

BACKGROUND

The field trial was performed at a manufacturing facility in central New Hampshire. Chlorinated and non-chlorinated VOCs and PHCs have been detected in site groundwater at concentrations exceeding applicable regulatory standards. A hydrogeological investigation was conducted which identified the following three hydrogeologic units at the site: 1) an unconfined sand upper unit; 2) a saturated clay and silt unit; and 3) a confined lower sand unit.

Dissolved-phase VOC plumes were detected in both the upper and lower hydrogeologic units. In addition, up to 4-feet of LNAPL was measured within several monitoring wells installed in the upper unit. Product samples were fingerprinted as used cutting or grinding oil containing VOCs, and an oil fraction composed of moderately heavy aliphatic PHCs in the C₂₀ to C₃₀ range. Estimates of product volume were on the order of 10³ gallons.

Remedial activities underway include vacuum-enhanced groundwater extraction in combination with product recovery. Groundwater from the product recovery system requires treatment before being discharged to surface water. Characterization of groundwater quality beneath the LNAPL was critical for treatment system design. Two groundwater monitoring wells located proximal to areas of greatest apparent LNAPL thickness (i.e., wells MW-10 and GZ-205) were selected for water quality sampling.

METHODS

Monitoring wells MW-10 and GZ-205 were first gauged for LNAPL and groundwater level to size Direct Coating Procedure equipment. Inertia pumps (Waterra™ Standard Flow Systems) were then used to purge approximately three times the standing water column from each well. Though both monitoring wells were purged before sampling for this field trial, purging is likely unnecessary when using the Direct Coating Procedure because an LNAPL establishes a virtually impervious barrier that limits atmospheric exchange with the water column. However, because a portal is established through the LNAPL during use of the Conduit Procedure, well purging will likely remain necessary for subsequent sampling rounds if the conduit is dedicated to the well. After purging, groundwater samples were collected using the following methods:

• Standard inertia pump;
• Inertia pump via the Direct Coating Procedure; and
• Standard bailer (Voss™ 3/4-inch polyethylene weighted bailer) using a Voss™ bottom-emptying device (well MW-10 only).

Blind duplicate samples were collected from monitoring well MW-10 using standard inertia pump and bailer techniques for quality assurance/quality control (QA/QC).

Samples were placed in a chilled cooler and submitted for laboratory analysis to GZA GeoEnvironmental, Inc.'s Environmental Chemistry Laboratory of Upper Newton Falls, Massachusetts in accordance with standard chain of custody protocol. Laboratory analysis for VOCs was performed in accordance with United States Environmental Protection Agency Method 8260.

RESULTS

Results of the laboratory analyses include:

• Groundwater samples collected from monitoring well MW-10 using the Direct Coating Procedure contained lower total VOC concentrations (i.e., 146 ug/l, Figure 2) than samples collected using standard bailer or inertia pump technique (i.e., 303 and 409 ug/l, respectively, Figure 2), as well as fewer detected analytes than standard techniques (i.e., bailer: 10; inertia pump: 8; Direct Coating Procedure: 3); and

• Groundwater samples collected from well GZ-205 using the Direct Coating Procedure similarly contained lower total VOC concentrations (i.e., 19 ug/l, Figure 2) than samples collected using standard inertia pump technique (i.e., 281 ug/l, Figure 2), but contained a greater number of detected analytes than the sample collected using standard technique (i.e., Direct Coating Procedure: 5; inertia pump: 2). We attribute the greater number of analytes detected in the sample collected using the Direct Coating Procedure to analytical method detection limits that were an order of magnitude lower than were used for the sample collected using standard inertia pump technique.

Water quality data are consistent with field observations that groundwater samples collected using standard techniques had entrained product emulsions and an iridescent sheen, whereas samples collected using the Direct Coating Procedure were visibly free of product. Collectively, water quality data supported by field observations suggest the Direct Coating Procedure was superior to standard techniques for collecting groundwater samples from beneath the LNAPL. Results of the laboratory analyses are summarized in Table 1.

Though not related to the ice-coating method, there was better agreement between analytical results for the sample collected from monitoring well MW-10 using standard bailer technique than standard inertia pump technique based on QA/QC blind duplicate results (e.g., bailer: 303 ug/l versus duplicate: 290 ug/l; and inertia pump: 409 ug/l versus duplicate: 558 ug/l). These results likely reflect sample collection method. Samples collected using the bailer were obtained from the bottom of the bailer using a bottom-emptying device whereas samples collected using the inertia pump were collected from the top of the water column where product emulsions would likely accumulate.

TABLE 1: SUMMARY OF VOC DATA, ug/l
VOC	WELL MW-10					WELL GZ-205
	Bailer		Inertia Pump		Direct Coating Method	Inertia Pump	Direct Coating Method
	Sample	Duplicate	Sample	Duplicate	Inertia Pump		Inertia Pump
1,1-Dichloroethane	31	32	40	37
cis-1,2-Dichloroethene	82	84	110	140	26	270	6
1,1,1-Trichloroethane	40	41	50	54	32	11	2
Trichloroethene	5	6
Tetrachloroethene	14	14	21	28			7
1,2,4-Trimethylbenzene	6		11	22
1,3-Dichlorobenzene	6	5
1,4-Dichlorobenzene	9	8	15	23
1,2-Dichlorobenzene	90	80	130	200	88
Naphthalene	20	20	32	54
Toluene							2
Xylenes (total)							2
Total VOCs	303	290	409	558	146	281	19

NOTES:

1. No entry indicates not detected above method detection limit.
2. "ug/l" indicates micrograms per liter, which is equivalent to parts per billion.

CONCLUSION

Though groundwater samples generally need not be collected from beneath LNAPLs, there are cases involving LNAPL mixtures, coalescing contaminant plumes from multiple sources, groundwater treatment system design, and collection of natural attenuation parameters that may necessitate sample collection beneath LNAPLs. In these cases, standard groundwater sampling techniques are not preferred because they entrain product emulsions in samples. The ice-coating method is an excellent alternative method for collecting groundwater samples from beneath LNAPLs because the method is simple, inexpensive, and limits the entrainment of product in samples thereby providing more representative groundwater quality data than could be obtained using standard sampling methods.