Table 3. Continued - SR97

Table 3 (cont'd).

ΣTNT

Well no.

pH

Type

HMX

RDX

TNB

DNB

DNA

TNT

DNT

2ADNT

4ADNT

UMATILLA

BANGOR

(kit value - lab value)

The RDX kit produced very poor results for

100.

RPD =

(kit value + lab value)/2

these samples. It failed to detect RDX in eight of

the samples that contained RDX above the EPA

Health Advisory limit of 2 g/L (EPA 1988a). It

A 50% RPD is the commonly used control limit

also produced six false positives.

for field kits. The EnviroGard, Quantix, and Ohm-

The cost of the kits requiring daily calibration

icron kits met this criterion more than 85% of the

varies depending on the number of samples anal-

time. The DTECH kits failed this criterion over half

yzed per day. The more samples that can be anal-

the time. Positive results require confirmation by

yzed, the better. For this investigation of monitor-

laboratory analysis. In on-site assessment, false

ing wells at NSWC, 33 samples were generated in

positives add to the cost of laboratory assays. In

3 days. For cost comparison estimates, ten sample

remediation activities, false positives could be-

analyses per day was chosen. The strip-plate kits

come quite expensive if they triggered an addi-

have a unique problem that becomes apparent

tional treatment step (e.g., replacing a purification

because of their design. There are 96 assay wells

cartridge or extending a composting period or ex-

arranged in long strips containing either 12

cavating an additional lift of soil). For the discus-

(EnviroGard) or 8 (Quantix) wells each. All wells

sion of results that follows, a false positive is de-

in a strip must be used. Unused well-strips can-

fined as an erroneously high value for TNT con-

not be saved for future use. Thus 10 samples, one

tamination above the EPA 1989 Health Advisory of

2 g/L.

standard, and one blank run in duplicate (24

7