REPORT DOCUMENTATION PAGE

Five relatively newly developed groundwater-sampling devices (the Kabis, HydraSleeve, Discrete Interval, PneumoBailer, and USGS Passive Diffusion Bag [PDB] sam-

plers) were tested to determine their ability to recover representative concentrations of a variety of analytes, including volatile organics, explosives, pesticides, and metals. The

first phase of the study included several standpipe experiments with known concentrations of analytes. In the second phase, the devices were used in the field to sample TCE

from a deep well and were compared with samples taken using low-flow sampling.

We found that the PDB sampler was the easiest device to use but should be used only for selected VOCs. The HydraSleeve and the Kabis Sampler are thief-type samplers

that were also relatively easy to use. Although these devices could produce representative concentrations of explosives, pesticides, and metals in the standpipe experiments,

they elevated the turbidity in our monitoring well. Therefore, we would recommend that their use be limited to wells where the turbidity is not affected by their use, especially

if sampling for metals or the more hydrophobic organic contaminants. In addition, there were small but statistically significant losses of some VOCs with the HydraSleeve in

the standpipe studies (<5%) and of TCE in the field study (11%). Concentrations of VOCs taken with the Kabis Sampler did not show a substantial and consistent bias in either

direction, except for the low-level study where loss of TCE was substantial, 18%. In the field study, loss of TCE was small (<8%) with this device and not statistically

significant.

The Discrete Interval Sampler and PneumoBailer are pressurized thief-type devices that are designed to collect a sample when activated. The PneumoBailer was heavy

and awkward to handle, required taking a nitrogen tank into the field, and was difficult to operate. The Discrete Interval Sampler required only a bicycle pump to pressurize the

chamber, was smaller and lighter in weight, and easier to handle and operate than its counterpart. Both devices generally delivered representative concentrations of all the

analytes tested in the standpipe experiments. Although loss of TCE was statistically significant for the Discrete Interval Sampler in the field study, loss was very small (<5%).