Analysis of PCB sample

The first step in any success analysis is the collection of a sample. The sample collected must meet the objectives of the sample design. e.g., representative of the whole. The sample collection, handling, and storage techniques should eliminate or at least minimize losses or contamination which can affect the accuracy of the reported results.

PCBs are an inert, nonpolar class of semivolatile organic compounds. As such, they are reasonably well behaved during sample collection and storage. Consequently, with the exceptions of the air and water sampling techniques which concentrate the PCBs from the matrix onto an adsorbent, most PCB sample collections have utilized the standard or customary methods for semivolatile organics in the subject matrix. For example, the American society for testing and materials(ASTM, 1991c)procedure for gas chromatography/electron capture detector(GC/ECD)determination of PCBs in environmental PCB Assembly matrices references general methods for water sampling, sediment and soid sampling, and air sampling. The reader is referred to general sampling and storage procedures in the absence of a PCB-specific method. Sampling for chemical analysis has been reviewed(Kratochvil et al., Keith 1991). Numbers(1994)reviewed PCB sampling, including setting objectives, approach and design,layout,and protocol.

This chapter discusses PCB-specific sampling design condiderations, sample collection techniques, and sample storage.

SAMPLING DESIGN

The selection of sampling sites, frequency of sampling, number of samples, measurement of physical and chemical parameters of the sample, and the overall statistical design of sampling methods are critial to a realistic assessment of the PCB concentration of the whole.The sampling design is directly related to the objectives of the specific study, research program, or regulatory action(see data quality objectives discussion in Chapters 9 and 11), the sampling design often strives to achieve representative samples to determine the mean and other statistical values for the whole(e.g., finding the mean concentration of PCBs in adipose of the U.S. population). Another common design maps the extent of contamination over an area or in a volume. For certain applications, the sampling design may strive to collect samples which represent the highest concentrations or hot spots. Other biased sampling designs also may be appropriate for certain applications. Two types of error traceable to the sampling design are possible. the first is a false positive, i.e., concluding that PCBs are present above electronic assembly an action limit when, in fact, they are not. The second is a false negative, i.e., failure to detect the presence of PCBs above an action limit. False nagetives can readily occur with hererogeneous contamination. For example, if a small area is contaminated within a much larger area being sampled, the probability of finding that area can be quite low without the aid of phycical clues such as staining of the surface. General considerations on sampling design have been provided in extensive detail(see, for example, Moser and Huibergtse, 1976; Mason, 1982; Boomer et al.. 1985; Kelso etal.. 1986 Gilbert, 1987).

PCB sampling desiges have been developed for assessment of the extent and levels of contamination of spills, landfills, ponds, and other sites. Two general designs are possible; random and grid. Grid designs are generally favored because they are easier to implement in the field via a rota protocol and are statistically more efficient than random design. A grid design is certain to detect a sufficiently large contaminated area, while a random design could miss detection of that same area. The classic square grid, widely used in environmental sampling environmental sampling, has even been applied to PCB sampling in a pond(Prohammer et al.,1985). Number(1994) provides an overview of grid sampling.