ΜATERIALS AND ΜETHODS

we quantified losses when the flow rates were

from the TCE-contaminated well and drew water

moderately fast (1 L/min) and slow (100 mL/

off at flow rates that varied from 100 mL/min to

min), determined if equilibration occurs and

3 L/min. The well water was diverted from the

when, and investigated whether organic constit-

outflow pipe to a water distribution system con-

uents or constituents that interfered with organic

sisting of a master flow valve; an electronic flow

analyses were leached. In our desorption study,

meter connected to a Linear 555 chart recorder; a

we determined whether TCE was released from

three-port valve, which allowed samples to be si-

contaminated sampling tubings when organic-

phoned off but still allowed flow to continue

free water was pumped through them at a slow

through the tubings; and a bank of five multi-

flow rate, and again looked for leaching of or-

turn needle valves, where up to five tubings

ganic constituents or compounds that interfered

could be connected (Fig. 4). The materials used in

with organic analyses.

the distribution system consisted primarily of

copper, brass, and glass. There was a small

amount of Teflon, perfluoroalkoxy, and Kel-F

ΜATERIALS AND ΜETHODS

that were used in the O-rings, three-port valve,

Μaterials

and needle valve body. To flush the system, the

The five tubing materials used in this study

well water was pumped through the system for

several hours prior to connecting the tubings.

were LDPE, one rigid fluoropolymer (PVDF), one

Strip chart recordings showed that the flow rate

flexible fluoropolymer [P(VDF-HFP)], and two

into the distribution system varied by approxi-

formulations of plasticized polypropylene (PP1,

mately 10%. The rate of flow through each tubing

PP2). We selected these tubings because, in our

was routinely monitored by collecting a volume

previous study (Parker and Ranney 1996), we

of water from the outflow of the tubing in a grad-

found that these tubings ranged from relatively

uated cylinder and measuring the time with a

highly sorptive (PP1 and PP2) to least sorptive

stopwatch. Flow through each tubing was regu-

(PVDF) and that they leached little or no organic

lated by turning the appropriate multi-turn nee-

constituents under long-term static storage (as

dle valve at the end of the distribution system. To

determined by HPLC analyses with a UV detec-

expel air from the tubings as they filled with wa-

tor). All the tubings had an inside diameter of

ter, the tubings were wound around a 55-gallon

0.25 in. and were not pre-cleaned prior to the

study.

metal drum in an upward spiral fashion.

Water contaminated with trichloroethylene

In our first study, all five tubings [PVDF,

(TCE) was obtained from an artesian well that

P(VDF-HFP), LDPE, PP1, and PP2] were tested.

can supply up to 380 gal/min, located at our fa-

TCE-contaminated well water was pumped

cility. The water from this well is used by our fa-

through three different lengths (10 ft, 50 ft, 100 ft)

cility for cooling after it is treated to remove the

of tubing at a slow flow rate (100 mL/min). In

TCE.

our second study, the TCE-contaminated well

water was pumped through 500 ft of LDPE at the

same flow rate (100 mL/min). In our third study,

Experimental design for the sorption studies

For these studies, we tapped the outflow pipe

the contaminated well water was pumped

through 100-ft sections of three tubings (PP1,

LDPE, and PVDF) at a much faster flow rate, 1 L/

Figure 4. Distribution system from water source to tubings.

Meter

6