Document kY931E33bEBL72QwaGbQx6Ey

i Determination of Vinyl Chloride Migration From Polyvinyl Chloride Pipe Into Water Ronald C. Dressman and Earl F. McFarren ASI 000018596 3as-ehromatograph determinations of possible vinyl chloride leaching from poly vinyl chloride pipe into potable water systems alleviate fears of harmful concentra tions in the systems tested. Polyvinyl chloride (PVC) pipe has been in use to varying degrees in some water-distribution systems in the US for the past sixteen years. Recently concern has been raised that unpolymerizcd vinyl chloride (VC) in the pipe can migrate into the water flowing through it. To investigate this theory, existing meth odology1 2 for vinyl chloride determina tion was modified' to permit the detec tion and measurement of this material at very low concentrations. EPA's Water Supply Res. Div. then studied five water distribution systems that use PVC pipes. Determinations generally were made using gas chromatography with a microcoulometric titration detector (GC/MC)--the most specific detector available. The data are not unequiv ocal, and although results should be ac cepted with caution, the authors consid er them to have a high degree of validity. Experimental Sampling. Samples were collected in serum reaction bottles, which were then sealed with PTFE-lined rubber septa held in place by crimped aluminum caps. The bottles were filled to overflowing, and the septa were slid over the top to preclude any air space at the time of sealing. Five water-supply distribution sys tems (Table 1) incorporating PVC pipe were sampled. Samples were taken and analyzed in triplicate both before and after entering the pipe. Samples were shipped on ice and kept refrigerated in the laboratory until warmed to room temperature immediately before being analyzed. Sample concentration (purging). Samples were purged in the sampling container using the procedure and apparatus described by Dressman et aid Silica gel, JANUARY 1978 35/60 mesh, was used as the trap adsor bent. Gas chromatography (GC). Qualitative accuracy was achieved through the innovative application of two types of column material, with the express aim of eliminating interference from cyanogen chloride. A 180- x 4-mm-ID glass col umn was packed for its first third with Chromosorb 101 and the last two thirds with Chromosorb 102. The column was operated in a program mode. The initial temperature was set at 50 C and programmed to rise 10 C/min to 160 C. Nitrogen carrier flow was 80 ml/min, and the desorb (inlet) temperature was 150 C. A microcoulometric titration system* was used as the GC detection system. Operating parameters were: reaction fur nace, 850 C; outlet furnace, 850 C; and transfer line, 240 C. The humidified oxygen-reactant gas flow rate was 40 ml/ min. A range ohm setting of 600 was used to produce a 65 per cent of full-scale response for 32 ng of vinyl chloride. Standard solutions. The standard solu tions of vinyl chloride used for gas chro matographic analysis and dosing con trols were dilutions of a stock solution prepared by bubbling vinyl chloride into a fared amount of acetone as described by Bcllar et al.1 The actual concentration of vinyl chloride in the stock solution was determined by gas chromatographic analysis using a reaction coulometer4 as a quantitative GC detector.1 Identification and quantification. Identifi cations on each sample were ac complished by comparing the retention time of suspect peaks with the appro priate standard. `Model 5-400 equipped with T-300-S silver hjluir* titration cell and ('-200 microcoulomi'ter, Dohrmann/finviroloch. Santa Clara. Calif. 0003-150X/78/0100-0029 $01-00 r. 1978 American Water Works Association Quantification was accomplished by comparing the area of the sample peak with a measured amount of the vinyl chloride standard using height times width at half-height measurements. The standards were injected directly onto the GC column, which was then pro grammed as for sample analysis. Results obtained on the analysis of samples by the bottle-purging procedure were then divided by a factor of 0.52 to correct for incomplete recovery during purging.1 Five field sites were selected to deter mine the extent, if any, to which vinyl chloride might contaminate drinking wa ter supplies. The sites were selected as being representative of extremes in climatic conditions and because data were available concerning the age, length, and size of the pipes used. The results and other pertinent data are listed in Table 1. Very low concentra tions of vinyl chloride were detected in four of the five water supplies studied. In all but one instance (Georgetown, Tex.) values were too low to be verified by GC--mass spectrometry. In any event at the particular time of analysis the equip ment was not available. As with any gaschromatography data, caution must be exercised in interpreting the results. Without mass-spectrometric verification the results lose a degree of confidence. The method is designed, however, to overcome all known interferences (i.e., cyanogen chloride, acetaldehyde) and to retain a high degree of validity owing to the highly specific halogen detector used. The initial values obtained at Coolidge, Ariz., probably can be discounted because VC was detected both before and after the PVC pipe run. This suggests that an outside source had contaminated either the system or the sample itself. No VC could be detected a month later. By way of contrast, the concentration of VC (1.4 yg/1) detected in the first sampling of the Jonah water supply in Georgetown, Tex., was essentially the same (1.2 gg/1) on resample eighteen days later and R.C. DRESSMAN & E.F. MCFARREN 29 r TABLE 1 PVC Pipe Systems Studied Parameter Year pipe manufactured Coolidge. Anz. '1964 Georgetown. Tex. 1975 Pipe Length-km Length--mi Size--m Total wall area--m' Total wall area-sq ft Water temperature--C, Sampling point Sovirce Well head Treatment plant End of pipe Sampling date VC concentration*--fig/l Sampling date VC concentration,-/ii>/l l,b 1 l 1.0 0,69 32 50 6 000 -- 20 X X R/22/75 0.3T 0,3-f 9/22/73 tt 4.6 IS 30 95 666 465 6 200 66 000 22-25 X X 10/10/7S t 1.4 10/26/75 t 1.2 'Average of triplicate analyses, corrected for recovery efficiency 'Probably spurious JNune detected at a detection limit of 0 03 jig/i Water Utility Pioneer. Calif. Rosehurg, Ore '1966 19bh-b7 {1 39 km 1971) 7.6 3 7 4 7 23 64 4 800 52 000 22-24 4 44 0 6H 3,07 0 42 32 1 300 14 000 44 X X X 9/16 75 t 0.06 - - X 10/17/75 l 0 03 - - Salados, Tex. '1968 018 0 46 0 11 0 30 86 370 4000 24-23 0.18 0.11 2 X X 40/10/70 -t t 10/28/75 -tt As i 0001S597 \ occurred only in water sampled at iic end of the pipe. VC was detected m three of the five systems investigated. Although all of the concentrations were far below the mini mum levels that have been associated with a carcinogenic response in experi mental animals,* studies are still ongo ing. Because of the latency of the carcin ogenic response, the results from these investigations not be available for some time. Interpret,i.*,on of the data is difficult because the residual VC content of the pipes is not known. Within this limita tion, however, the newest, longest sys tem (Georgetown) hud the highest VC concentration, and the next longest (Pioneer, Calif.) hod the next highest concentration. Water temperature did not seem to influence the VC concentra tion. Finally, traces of VC at the micro- gram per liter level were still present in the Pioneer and Rosenburg, Ore., sys tems approximately nine years after they were installed. The concentration of VC in the newest, longest system tested (Georgetown) was low compared with the level deemed hazardous to animals. Manufacturers claim that pipe made for water use within the past year has a lower free *T< date, experiments that have been carried out demon strate that the minimum carcinogenic dose by chronic inhala tion is 50 mg/1 ynd by ingestion is 1Rb5 mg/kg'day. ' Additional investigations are being conducted to determine the incidence and type of liver cancer at lower dose levels by ingestion, including 3,33, 1.0. 0.3, and 0 03 mg/kg/day. The carcinogenic risk to man has been estimated by the Natl Academy of Sciences'as 5 1 x it) at a concentration of 1/tg/ 1, considering a consumption of 2 l,'day of water vinyl chloride content than pipe manu factured earlier. Large factors of safety therefore should be inherent in the use of the pipe. Summary A field study of the VC concentration in the water of five distribution systems utilizing PVC pipe showed concentra tions from 0.03 fig/1 to 1.4 ptg/l with the lowest concentration persisting in a system nine years old and the highest concentration present in the newest system. These levels are well below those that thus far have been associated with adverse effects in experimental animals, and these levels should decline as lower residual monomer pipe is put into use. Acknowledgment The assistance of Robert G. Tardiff in the evaluation of the toxicological data on vinyl chloride gratefully is acknowl edged. References 1. Bellar, T.A.; Lichtenberg, & Eiciiel- berger. J.W. The Determination of Vinyl Chloride at the jig/l level in Water by Gas Chromatography. Envir. Sc.-i. Tech., 10:927 (1970). 2. Beu.ar, T.A. & Lichtenberg, J.J. Deter mining Volatile Organics at Microgram- per-Liter Levels by Gas Chromatography, four AWWA. 66:12:739 (Dec. 1974). 3. Drf.ssman, R.C. & McFarren, E.F, A Sample Bottle Purging Method for the Determination of Vinyl Chloride in Water at Sub-microgram Per Liter Lev els. four Chromatog. Sci., 15:69 (1977). 4. Littlf.vvood. A.B. & Wiseman, W A. The Reaction Goniometer as an Absolute Standard in Calibration of Gas-Chro matographic Detectors. Presented at th& Pittsburgh Conference on Analytical! Chemistry and Applied Spectroscopy, Pittsburgh, Pa. (1970). 5. Kimm, Victor J. Letter from the dep. asst. admin, wtr. sply., EPA, to hearing clerk. Food and Drug Admin, (Dec. 12. 1975). 6. Maltoni, C. & Lefemine, G. Carcinogenic ity Assays of Vinyl Chloride. Current Results. Ann. N.Y, Acad. Sci; 246:195 (1975). 7. Maltoni, C. & Lefemine, G. Le Potenizialita dci saggi Sperimentali nella Prodi__ zione dei Rischi Oncogoni Ambientali. Un Esempio: il Closuro di Vinite. Accademia Nazionule dei Lined Rendicunti Delia Clause Di Scienze Fisiche, Moth matiche e Noturafe Estratto did Fasc.. Seri VIII, LVI (Mar. 1974). 8. Maltoni, C. et al. Gli Effecti Oncogen! Del Cloruro Di Vinile Somministratu Per Via Orale Nel Ratto. Estraftc dallti flevista-Gli Ospedali della Vila, 2:6 (Jul.-Dec. 1975). 9. Report of the Safe Drinking Water Committee. Natl. Acad. Sci., Drinking Water and Health, Washington, D.C. (1977). A paper contributed to and selected by the JOURNAL, authored by Ronald C. Dressman (Active Member, AWWA), res. chemist, Mumc. Envir. Res. Lab., and Earl F. McFarren (Active Member, AWWA), ch., distr. quality sect., wtr. supply res, div., both of the E11" ,c,nna Ohio 64464 4130, 4300, 4430 30 WATER TECHNOLOGY/QUALITY JOURNAL AWWA