Table A1. For each analyte and time, analysis of
show the adjusted mean normalized concentra-
variance (ANOVA) tests were conducted to deter-
tions for the three materials that had different
mine if the tubings had any significant effect (at
surface-area-to-volume ratios (PTFE, ETFE, and
the 95% confidence level) on the analyte concen-
the polyamide).
trations. When significant differences were found,
As we have found in previous studies that test-
Fisher 's Protected Least Significant Difference
ed PVC and PTFE well casings (Parker et al. 1990,
Test was performed to determine which tubing
Parker and Ranney 1994), the more hydrophobic
materials were significantly different from the
analytes--ODCB, PDCB, and PCE--were the
controls and each other. The results of these analy-
most readily sorbed, and NB and MNT were the
ses can be seen in Table A2. Because three of the
least sorbed analytes. All the tubings sorbed at
tubing materials (PTFE, ETFE, and the polyamide
least some of the more sorption-prone analytes.
[nylon]) had different material surface-area-to-
For some tubings, sorption of some organic sol-
solution-volume ratios than the other tubing ma-
utes was very rapid; i.e., losses equalled or ex-
terials, the data for these tubings were marked
ceeded 95% after only a one-hour contact time.
with an asterisk.
For polyurethane this was true for seven ana-
Mean normalized concentration values were
lytes. For FPVC, the silicone-modified thermo-
derived (Table A3) by taking the mean concentra-
plastic elastomer (TPE), and one of the plasticized
tion of a given analyte exposed to a given tubing
PP tubings, it was true for six analytes. For the co-
at a given sampling time and dividing it by the
extruded polyester lining in a PVC shell and the
mean concentration for the control samples for the
other plasticized PP, it was true for five analytes.
same analyte and time. Thus a mean normalized
For LDPE, cross-linked high-density polyethyl-
value of 1.00 represents no loss of analyte for a
ene (XLPE), and PE cross-linked to an ethyl vinyl
given tubing and time.
acetate (EVA) shell, it was true for three analytes.
A second study was conducted so that sorption
For the PE liner in an EVA shell, it was true for
of the analytes by the three tubings with different
two analytes. These ten materials generally were
surface-area-to-solution-volume ratios could be
the most sorptive, and among this group the flex-
compared with the other seventeen tubings. This
ible tubings were the most sorptive. Among all
study compared sorption of the same analytes by
the flexible materials tested, only the two fluori-
the three tubings with three different surface-area-
nated products, the fluoroelastomer and P(VDF-
to-solution-volume ratios. The results from this
HFP), were not highly sorptive.
study can be found in Table A4. These data show
Using Figures 1a-h and the results from the
that the material surface-area-to-solution volume
ANOVAs and the least significant difference tests
does affect sorption and that, as this ratio increas-
(Table A2), we compiled a summary table show-
es, sorption of the more hydrophobic analytes also
ing the least sorptive materials (Table 3). We see
increases. Normalized values were derived as de-
that generally the rigid fluoropolymers (FEP,
scribed previously and are given in Table A5.
FEP-lined PE, PVDF, PTFE, perfluoroalkoxy
(Also included in Table A5 are the normalized val-
(PFA), and ETFE) were the least sorptive. Specifi-
ues from the previous study for these three mate-
cally, PVDF was the least sorptive material for
rials. This gave us a fourth surface-area-to-
TDCE, TCE, and PCE. FEP and FEP-lined PE
solution-volume ratio for comparison.) The best
were the least sorptive materials for the other five
fit equation was obtained for these data for each
analytes (NB, MNT, CLB, ODCB, and PDCB).
material, analyte, and time using Cricket Graph
However, even though FEP, FEP-lined PE, and
software. These equations were then used to
PVDF were generally the least sorptive materials
determine what the adjusted normalized values
tested, they were still highly sorptive of the more
would have been for these three materials if the
hydrophobic analytes (Fig. 1g and 1h). For exam-
surface-area-to-solution volume had been the
ple, after 24 hours, losses of PCE and PDCB by
same as the other seventeen tubings. This meant
these three materials ranged from approximately
that we had to extrapolate values for PTFE and
60 to 80%. Clearly, long-term storage of aqueous
the polyamide and interpolate values for ETFE.
solutions of organic compounds in fluoropoly-
These adjusted normalized values can be found in
mer containers can be problematic.
Table A3, in parentheses next to the original mean
With respect to shorter contact times, we see
normalized values. For each analyte, the mean
that in the first hour, sorption of PCE and PDCB
normalized concentrations have been plotted for
by FEP, FEP-lined PE, and PVDF tubings ranged
all 20 materials in Figures 1a-h. These figures
from approximately 25 to 40%. However, the con-
8
Previous Page
