swelling agent of the polymer, then diffusion in-
taminated by pumping deionized water through
creases several orders of magnitude and becomes
them.
first anomalous and then Case II (i.e., has a swell-
Schuh et al. (1993) compared the effectiveness
ing front) (Berens 1985, Jenkins et al. 1986, Holsen
of distilled water rinses for removing seven pesti-
1988). In instances where Case II diffusion has
cides from a PVC bailer. They selected a brief
occurred, the polymer will be seriously degraded
contact time (one minute) to represent the time
(swollen, softened, dissolved) and the sampling
required to take a bailed sample in a shallow well
device will no longer be useful.
or gravelly aquifer. With one exception (dimetho-
Parker (1995) proposed that decontaminating
ate), they found a reasonably good correlation
permeable materials should involve more than
between the effectiveness of this procedure and
removing surface contaminants if desorption of
the analyte's solubility in water or its Kow value.
absorbed contaminants is significant. Unfortu-
They found that the most hydrophilic contami-
nately, there have been only a few studies (Miller
nants (i.e., with a solubility greater than 500 mg/L
1982, Barcelona et al. 1985, Parker et al. 1990) that
or log Kow less than 2.3) were removed from the
have examined desorption of organic contami-
bailer with no residual carryover with just one
nants from polymeric materials. Barcelona et al.
rinse, while the most hydrophobic analytes (i.e.,
(1985) followed the kinetics of desorption of chlo-
with a solubility of less than 3 mg/L or log Kow
roform from five tubing materials (polypropy-
~4.64.7) had residual carryover after six rinses.
lene, polyethylene, PVC, silicone rubber, and
Fink and Boyajian (1989) decontaminated a
polytetrafluoroethylene [PTFE]) that had been pre-
stainless steel bladder pump (with a PTFE blad-
viously exposed to a 100-ppb solution of chloro-
der and PTFE-lined sample tubing) after it had
form for one hour. They found that 8090% of the
been used to sample VOCs. The pump was cleaned
desorption they observed occurred within the first
by steam cleaning the outside of the pump and
tubing, and then circulating hot (120F, 49C)
510 minutes and that the amount desorbed after
one hour constituted less than 10% of the total
aqueous detergent (1%) solution through the
chloroform sorbed. Those results indicate that con-
system followed by ambient temperature rinse
taminants at or near the surface desorb rapidly
water. In each test, the pump was contaminated
and that release of the remainder of the absorbed
by pumping at least five sampling pump/tubing
contaminants is slower, most likely because it is
assembly volumes of contaminated water. Detect-
able levels of trichloroethylene (TCE) (12 g/L)
controlled by diffusion. All three studies found
that the compounds that were present in the great-
were found in the final rinse water when the pump
est concentrations following desorption were not
was used to sample the two wells with the highest
TCE concentrations (1600 g/L and 2215 g/L, re-
the same compounds that had been sorbed the
spectively) but not when the pump was used to
most rapidly or to the greatest extent. Parker et al.
sample wells with lower concentrations.
(1990) also noted that the smaller molecules were
Because both Mickam et al. (1989) and Parker
more readily desorbed and attributed this to the
(1995) concluded that there have been almost no
fact that diffusion is more rapid for smaller mol-
systematic scientific studies on decontamination
ecules (Berens and Hopfenberg 1982, Reynolds et
of groundwater sampling devices, we initiated
al. 1990).
these studies to determine the effectiveness of
various decontamination protocols. These studies
Studies that demonstrate the factors
were designed to consider the type of contami-
that affect decontamination efficiency
nants, concentration of contaminants, materials
Although not many studies have examined the
being decontaminated, and contact times.
efficiency of the various decontamination proto-
cols, Parker (1995) found three studies demon-
strating the impact that the type of contaminants,
Findings from our initial studies
In the first phase of our studies, we (Parker and
the level of contamination, or the materials being
Ranney 1997b) exposed three materials commonly
decontaminated can have on decontamination ef-
used in sampling devices (rigid PVC, PTFE, and
fectiveness. These studies are summarized below.
stainless steel) to a test solution containing either
Devlin (1987) found that polyethylene tubing
VOCs or pesticides and then tested various de-
was harder to decontaminate than Teflon tubing
contamination protocols. These chemicals were
(the actual type was not specified). These tubings
selected because they ranged from being rela-
had been exposed to ppb levels of a suite of vola-
tively hydrophilic to relatively hydrophobic. In
tile organic compounds (VOCs) and were decon-
3
Previous Page
