Metals study

Groundwater Sampling Devices

29

Vroblesky and Campbell (2001) for g/L and low mg/L concentrations. There-

ore, we believe that the reason the more volatile compounds were slightly higher

in the PDB samplers than in the standpipe was because they had volatilized from

the test solution. (The standpipe in these experiments was covered but not sealed,

and there were several inches of headspace.) Thus, the standpipe had lower

concentrations of these analytes at the time of sampling (i.e., the controls) while

the concentrations in the samplers reflected the previously higher concentrations

in the standpipe. The more volatile compounds had the largest differences.

Metals study

The variability between replicate samples was very low (<5% RSDs), except

for chromium concentrations in the PDB samplers where concentrations were

very low (Table 8).

Concentrations of metals in samples taken with the Discrete Interval

Sampler, HydraSleeve, Kabis Sampler, and PneumoBailer were generally not

significantly different from control values (Table 4). In the instances where there

were significant differences, the differences were less than 6%.

However, as with the pesticides, there were dramatic differences between the

metal concentrations in the PDB samplers and the controls. After 21 days' con-

act, three of the metals (As, Cd, Pb) were not detected in these samplers, and

concentrations of Cr were approximately 95% lower. Again, these results support

USGS guidance (Vroblesky 2001) that does not recommend using these devices

for inorganic ions. However, we should mention that other membranes, such as

the cellulose acetate (dialysis) membrane of the Multi-Level Sampler (available

from Margan Ltd.), can be used to sample metals and anions (Ronen et al. 1986,

1987a, 1987b; Magaritz et al. 1989, 1990; Ronen and Magaritz 1989; Kaplan et

al. 1991) or the Versapor membrane (acrylic copolymer coating over nylon

fabric) can be used for sampling colloids (Weisbrod et al. 1996).

Analytical results from the field study

Again, even though these samples had to be brought over 130 feet to the

surface and transferred to a sample vial, the variability was very low with the

Discrete Interval Sampler, PDB samplers, and HydraSleeve, with RSDs less than

4% (Table 9). Slightly greater variability was observed with the PneumoBailer

and the Kabis Sampler, with 15% and 8% RSD, respectively. The RSD for the

pumped control samples (using low-flow sampling) was less than 3%. With all

these devices, mean concentrations were 514% lower than the mean values for

the control samples.