Laserfiche WebLink
It is important to select the appropriate materials, and type, diameter, and length of casing <br />and screen, as these can affect the quality of a ground-water sample. Biased water-quality data <br />can arise from chemical and physical interaction between ground water and materials used to <br />construct monitoring wells (tables 5 and 6). These biases can result from leaching, sorption/ <br />desorption, or volatilization. Leaching and sorption/desorption studies that examined casing <br />materials are described by Hewitt (1994a and b, 1992); Parker and Ranney (1994); Ranney and <br />Parker (1994); Parker, Hewitt, and Jenkins (1990); Parker and Jenkins (1986); Gillham and <br />O' Hannesin (1990); Reynolds and others (1990); Reynolds and Gillham (1986); Cowgill (1988); <br />Barcelona and others (1983); Sosebee and others (1983); and Curran and Tomson (1983). Parker <br />(1992) provides a recent summary of the findings of several of these and other studies. <br />The well screen potentially can alter water quality because of the large surface area ex- <br />posed to ground water. The screen is the part of the monitoring well most susceptible to corro- <br />sion and (or) chemical degradation, and provides the highest potential for sorption or leaching of <br />contaminants (Aller and others, 1989, p. 192). Thus, when selecting the screen materials, resis- <br />tance to leaching or sorption/desorption for the broad suite of NAWQA constituents is a major <br />consideration (table 5). Therefore, PVC is the material of choice for well casing screens installed <br />for NAWQA ground-water studies. In cases where the well will be used only for sampling one <br />class of chemical constituents, casing and screen materials can be selected to minimize bias <br />caused by that material (table 5). <br />The PVC casing selected should be National Sanitary Foundation-approved schedule 40 <br />(or 80) and flush jointed and threaded. In low-yielding materials, such as till and loess, leakage <br />of water through improperly sealed PVC joints can contribute a significant amount of water to a <br />well compared to the amount of water contributed through the well screen (van der Kamp and <br />Keller, 1993). Under such circumstances, 0-rings or Teflon tape on threaded joints below the <br />water table helps prevent this leakage, as does a properly installed annular seal. <br />Joints of PVC or other plastic casing used for NAWQA-installed wells must be threaded <br />and not glued. Organic compounds that leach from PVC primer or adhesive can compromise <br />sample integrity (Sosebee and others, 1983). Compounds listed as ingredients in one or more of <br />six PVC adhesives and one primer included THE (tetrahydrofuran), MEK (2-butanone or meth- <br />ylethylketone), MIBK (methylisobutylketone), cyclohexanone, and DMF (N, N-dimethylforma- <br />mide). In addition to sample contamination, such compounds can mask or made the identi- <br />fication of other VOCs difficult by co-eluting with other VOCs during sample analysis. <br />The length of a well screen is determined on the basis of the scale and objectives of the <br />investigation. The length of a well screen is important in relation to the vertical interval of in- <br />vestigation. In terms of water-level and water-quality measurements, a short screen generally <br />provides measurements of hydraulic head and ground-water quality that more closely represent <br />point measurements in the aquifer than measurements provided by a long screen. Ground-water- <br />quality samples also reflect an integrated measurement of water quality vertically throughout the <br />screened (or open) interval. Pumping a well with a long screen is more likely to induce mixing <br />of waters of different chemistry than pumping the same well with a short screen. Thus, concen- <br />trations of constituents in samples obtained from wells with long screens are less likely to reflect <br />the maximum concentrations of those constituents at any point within the screened interval than <br />samples obtained from wells with short screens. <br />28