Document zQBa28n5mDGv8yYKg6Ob1k5Da

1. A. P. Hottpot, DC CO r. C. Mfolt. Soft. It, 1975. 2. A. P. Mottatr, Itoefealcal Raport, Soft. 15, 1975. 3. M. M. O'Hara aa4 A. P. Matsgar, Toehaieal Raport, Auguot 18, 1975. 4. g. F. Smith, Toeholea1 Raport, Dacootar 10, 1975. 5. A. P. Katigar, 10C to J. t. Oeoto, 7ww>n 3, 1975. 6. A. P. Natspar, DC to J. K. Oooco, Oetohor 2, 1975. 7. A. P. Hocsgar, Toehaieal Raport, Poo--tor 3, 1974. 8. C. A. Doalola, DC to P. I. Kra--o, Poo--tar 4, 1975. 9. C. A. Doaialo, DC to J. R. Qgoto, --ator 14, 1975. 10. C. A. Ooololo, DC to A. P. Motagir, Po--tar 19, 1975. 11. C. A. Ooololo, DC to A. P. jjarojOt, Joo--ry H, 1974. 12. C. A. Ooololo, IOC to J. I. 8--CO, 3--17 24, 1974. 25. TOOS5ZZZ BFG19824 6FGTHCHWCAI document B. P. Goodrich Chamlcal Company A MVWON OT THS a F. OOOOWCH COMPAMY DEVELOPMENT CENTER MIGRATION Or RS5Z0GAL SOLVENTS IN CPVC POTABLE WATS* SYSTWS 40532 a. r. MnMO 17, U7t PISTBHOTM: CUrtlwd n. o. Scott R. A. Kruogor B. N. C. Xvlckor H. E. Roho K. CrNM E. C. Sehnm>tU J. L. Roloo* * R. V. ImlattM * E. J. Softs - j. C. HaslyF. J. Posit * R. B. Ooborso J. r. Noloso * G. S. --f . . TmmUU r. i. i. i. R. Afctoo J. Roaftas BbmII, Jr. E. t. BOOS j. A. ftiiRxm w C. Boehcol AL.TC J. t. Gooto * t. N. Rroofor - A. L. Schultz * R. t. Bloor>J. A. ToToo r. I. Rrasso - R. K. Schlttstr L. I. CrtAor I. C. DoCapitt c. A. Daslols R. J. Ifcyor Ctr/lafl. <4)^COWfOR-C: * R. J. Vwoott * G. |. lftospooo * C. I. Uftor * p. >. Ur A. t. Borosi T. W. Bopo* - B. A. CruUo MO HU (2) 41 1. R. RploaAo - R. S. Mother Loultrlllo F. R. Lootone* - S. S. Mlcholt ITC Tochsleol Pilot (2) Mrtiltwi A. I Webbe r K CS o o *trtdlctCo soAlfloA copy. Oooper - A. V. Otto BFG19825 1. Summary Potable water pipe made from Ceon CPVC compounds frequently eontelae sort than 0.5X of residual carbon tetrachloride. It has been shown that a portion of this "solvent" will Migrate Into water which remains in contact with the pipe. Attempt* to determine the migration rate of the "solvent" under specific con ditions have yielded Inconsistent reeults. for example, extractions carried out in an oven indicated that the amount of carbon tetrachlorlda which migrated into water did not depend on the amount la the pipe. The data also suggested chat reducing che temperature (from 160 to 140F) enhanced the migration. However, there was evidence that the pipe was somewhat heterogeneous and this probably Influenced che eigration. In ocher extraction work, e temperature gradient wee maintained across the pipe wall. Under thoee conditions, emeller amounts ef "emlvemt" eero axtractad and the concentration la the watar dotmaned with lasreaalag exposure time. Apparently soma of the carton tetrachloride mactad with other ingredients la che system. Obviously, it la not possible to predict the contoatratlon ef fugitive carbon tetrachloride. a typical erve toe water system wee Installed to provide "real-life" samples. Temperature measurements demonstrated that the CfTC pips was at as elevated temperature only a smell percentage of the elms. Aaalyela of the water saaplaa, collected over a 9 week period, thawed that sheet MX of them coatained seaaurable amounts of "solvsat". The cmmcsatretlos varied widely hut ia e random pattern. There wee evidence that water trapped la certain pertlomn of the ayacen contained high comceetrations ef carton tetrachloride. Ooeaslonalip, eoam of Chet water aimed into che sample and canoed the analyst* to he unexpectedly high O 100 ppb). Presumably, that would also occur la seat aatatiag systems. Consequently, if the risk of carton tatraehleride migration it to to reduced, the amount in the rests muac to graatly dacreaped. Although so acceptance criteria have brneo established, it aeeae likely that ths realdual carton tetrachloride should be radoced from the curreatC levels ef stout 5000 ppm to perhaps a maximum of 100 ppm. BFG19826 C006Szz?. introduction Thera has been concern about the poaalbla migration of%restdual CC14 from potable water pipe fabricated of Ceon CPVC. Last year, aanples of water were collected fro* operating CPVC llnea Installed at various locations at ALTC. Analysis of the water showed that CCI4 wee present and that the amount Increased with longer residence tlee^^. Those results confined that Migration does occur. There Is no criteria for an acceptable level of CCI4 In water but tr is *e ie r 1! v agreed that It should be lees than 0.0) ppm. Extraction studies nave : -"ur.r-A that amounts far in excess of 0.0) ppm will Migrate under certain condit ior.s '> . Those InvestIgatlone also Indicated, as expected, that the amount of CCl4 which migrated Increased with higher tanperature, longer extraction time and larger surface-to-volume ratios. There was sows criticises of those studies because sections of pipe had been lustersad In water end In normal use extraction nuld occur only from the interior. Irrespective of that feet, there wer' a number of Inconsistencies In the dste end therefore it wee not possible to develop a mathematical relationship which would predict the migration of CCl^. * Subsequently, sections of pipe were filled with distilled water and scored in ovens. Periodic analysis of the seatplee stored at about 105F indicated that the CCI4 content Increased with time and the migration rate was similar to chat observed in the earlier extraction work with immersed samples. However, unalyses of the pipe samples stored at 180? wore confusing because the CCI4 concent reached a maximum end then decreased with continued storage^). There was speculation thee the long term decrease was evidence of e chemical reaction involving the CCI4. The disappearance of the CCI4 woa studied (by C. F. Smith) In a series of experiments carried out in gleeaware. Thera was evidence thee CCI4 had reacted with the brass fittings used to cap the pipe sections^*). Obviously, any further Investigation of the migration of OClg should not bo conducted In the presence of brass. The study indicated that CClg did not react with stainless steel. It wee ewident chat further axtractioo stadias wore neadod to define the migration problem sad clarify the contradictions in the previous work. This report describes the additional work. Moot of Chn previews extract Ion studios have Involved lamera Ion of pipe sec dona la vocnr gat tTirsf at ltOF. Tha algratloa of CClg from potable water pipe la our ante oemaorn amd charuforu it wee decided that any further investigat lone should morn lonely einnlata that application. Consequently, in thla work we will extract only from tho Interior pipe surface at somewhat lower temperatures. Of course, in actual ayscams tho water la la contact with tho pipe for va*ying time, dependent 00 specific 000 conditions. It la difficult to Interpret date from a flowing system amd chorefora both static aad dynamic ayeterns were planned. Each system is described below. As already mentioned, in tho Initial effort to extract from the pipe mtertor, analyses of tho water indleatad that tha CCI4 content decreased with increased storage elms, presumably because of reaetloa with the brass fittings. Never the; = ss BFG19827 3. that experimental arrangeMnt had Mr It and therefore It waa Modified to eUiii-.jtr contact between tha water and Mtal. Ttola waa accomplished by machining CPVC parta to fora cloauraa for tha pipe aactIona. Saptuaa wara attached to the closures ao that a hypodermic naadla could ba inaartad and water removed from the lnalde of tha plpa. Figure 1 ehowe a eketch of the arrangement. The pipe samp Lea aelectad for thia atudy (163-17-50-1 and 163-17-50-2) had been exrrided at ALTC In December, 1974. They had boon atorod at ambient temperature (Bldg. 427) prior to Initiation of tha extraction atudy (November, 1975). The pipe aectlona (8") were not cleaned before being filled with dletllled water and stored In a constant temperature oven. Periodically, samples of water ware removed and analysed for CHCl^ and CClg. "Ttt te^fBTr Tff Of course, in normal use tha pipe In a hot water system has a temperature gradient across the pipe wall. There waa acne speculation that thla gradient influenced migration of tha aolventa. Therefore, a static cent arrangement waa devised to simulate tha gradient. This experiment wee described earlier^), but basically It required the use of metal commommats to cap the pipe aectlona and supply the heat. Since the previous workW had shown that stainless steal did not react with CCL4, It was used for the Mtal parts. The end cape were drilled and tapped so that tubing (V) could be extruded through the plastic pipe. An external water bath was Mlntalned st an elevated temperature sod this water was pumped through the tubing to heat tha water la the annular space (between tha CPVC pipe and S.S. tubing) to the dMlred temperature. Thermocouples la tha static water and on the pipe surface were used to record tbs temperature gradient. A "saddle" around the plaeclc pipe held e eeptia which permitted sampling of the water In contact with the CPVC. Tha axperlMatal arrangement ie shown la Figure 2. The material used in the heat exchanger experiments wars the seme as those used la the oven tests. Semples, about 14" long, of 163-17-50-l and 143-17-50-2 were Installed and main tained at ta^erature for several daye. Veter saaplee were removed periodically and analysed for CSCl3 end CClg. Simulated hot Veter Ststin Although the static tents described shove should provide daea for predicting th amount of CCln mtgvmetmg wader a given eat of cooditlone, It was decided that water from a "real-life" system theald also be analysed. Consequently, a "typical" hot watar -yf>n tW tlWtlUH end operated on a schedule which presumably rapraaented the habits of a 4-Mmbar fasrtly. Although this system has been described pre viously!*^ tha essential details will be repeated, for convenience. Coeamrcial pipe (1/2" end 3/4" CTS) ends shout August, 1975, end fittings produced about Hey, 1975, were obtained from Geneve Products. The system, which wee fabricated from these eompomaata, com lated of epproxlMtely 32 feat of 3/4 Inch plpa end 46 fact of lech pipe (om Figure 3). SoleMld valves at 8 locations were actuated by tlMrs which allowed flow to the drain on s prearranged schedule, ort flce dowmetream from tha valves coatrolled the flaw rata. These facts are tabulated below. BFG19828 9QQ6GZZZ location 1 2 3 4 5 6 7 a Run mean tin. Rata (sal./draw) Bath #1 Lav. Bath #1 Show Bath #2 Lav. Bath #2 Tub Laundry Tub Cloches washer Kitchen sink Dish weaker 1* 10 2 15 20 5 19 10 manager 7 1 8 3 2 3 19 2 Total (sal./day) '-o* 10 16 45 40 15 28* 20 It should b noted that about. 185 gallons of hot water waa flushed through this svstea each day. An electrically heated hot water eaek supplied the syscea. Theraocouples were located at 6 points (Me Figure 4) ao chat Che teaperature of the water, pipe surface aad aableot air could be recorded continuously (3 a.note Intervals). The installation of this ayetea was conplsted la early October, 1975. During the following eenth, it rsaalasd filled with water (at aablent teaperature) while being checked for leaks aad adjustaaat of flow rates. The aystea was then given three 5*alnute flushes with hot water aad placed La service. At weekly intervals, wjeer saaplea were collected et eech Locatlea ead aaalysed for CHCI3 and CCl^. After 9 weeks (about 7260 gallons of water flow) the eaperlaaat waa tarainatad. In additloe to tha hot water ayetea deacrlbed shove, a eecoad loop vet installed for cold water. This consisted of sheet 15 Coat of 3/4 Inch pipe and 14 feet of * inch pipe. The solasmld value oa this loop epnretsd at the seas frequency aa location #7 00 the hot water eyetea. Than, flew through the cold water line was about 28* gal loos per day. This loop waa swpls* aad the water analysed on the saae schedule as the hot water system. 9006ZZZZ The pipe aaaples used in this esperlaeat contained different aaounts of residual 'solvents". The fMplm ld--tlflsd as 143*17-50-1 wee based on our accodard CPVC reals (603XS40) p|g|n a "stripped" reels wee weed la 143-17*30-2. The residual "solvent" UmnMp&ft catty (ppm) CCI4 (w) 143-17-30-1 143-17-50-2 tad plugs 172 130 253 5179 370 3443 it was expected that the CCI4 level la the water saaplea would reflect the nearly 10 fold difference la CClg seatsat of the pipe secttoes. Furthermore, since the sane water remained la oeatact with the pipe throughout the experlaeet, It would be expected that each aueessalus analysis (with Increasing extraction ties) would show Increased amounts of residual solvent. BFG19829 t i f (n the first trial, tha samples vara to rad In a 160F oven for 10 jjvs. :>ie control water contalnad no CClg, but aftar contact with the pipe section 1. r ,>:i. .'4 hours analysis showed that water contalnad 800 ppb (Table l). i ere were further Increases In CClg with longer extraction time but the mUrui n ra>e was much slower. It required 9 days for an additional 800 ppb of CClg t> njrate into the water. It should be noted that In some cases, water samples miiyred after longer exposure time appeared to have lost CClg. Subsequently, the con centration of CClg would then Increase again. It Is doubtful that there would be Intermittent leakage ftoai the system and therefore It can be (peculated that the random dlacrepancles are e reflection of overall sampling and analytical errors. A plot of the date (Figure 5) suggests Chet after the Initial su.ge,* the migration rate of CClg Is about 3 or 4 ppb per hour. It was surprising to ftr.d that the amount of residual CCI4 In the pipe seemed to have little effect on the amount of CCI4 which migrated Into the water. At leeet In this experiment, re ducing the CCI4 content of ths pips did not decrease the smount which migrated Into the water. In the second trial, new sections of ths pips (163-17-50-1 and 163-17-50-2) were filled with distilled water amd storsd la a 140F oven. Water samples were re moved after very short exposure times to determine the initial migration rate. The extraction was continued for s total of 15 days. Analyses of ths water samples ere shown In Tsbls II. Omce agala, a few of the samples appeared to loss CCI4 but in general the amount of CClg which migrated lacressec continuously with claw. The Initial rate (first 24 hours) at 140? appears to be about one half that at 160F (see Plgura 5). however, the migratloe rets at 140? during the remaining extraction period la actually greatar than that at 160F. It should also be noted that the pipe section containing loo residual CClg (163-17-50-2) lost as much "*oIventM as the sampla which contained nearly 10 times as much CClg (163-17-50-1). The observations that: (a) the migration of CClg at 160F is amt linear in short tins intervals (b) tha nlgr'tloe^rate at 1M*F is greatar than that at 160F and (c) the anenat ef CClg which migrates la not dependent on the residual quantity la the sample are Inconsistent with supertetlose from diffusion theory. There was speculation thee the interior surface of the pipe was not uniform and that this led to the inomelome raseIts. It was postulated that tha lubricant In the compound tended to concerncrate oe the Interior surface during extrusion of ths pipe end that CCI4 wes trapped In this layer. As a result tha extraction of CClg from both pipe (actions would be relatively rapid sad not influenced greatly by differences in the pipe wall. If that rationalisation were valid, than rsmoval of ths surface layer before extracting should raseIt la (a) decreased migration of the CClg and (b) differences la tha amount of CClg which migrated from pipe sections containing "high" end ''low" Isvels of residual solvest. ^ In the third trlel, approximately 5 alls of tha Interior surface was mactunsU from (6") lengths of 163-17-60-1 amd 163-17-50-2. The machined surface appeared Ol ^ BFG19830 .40539 co he <> saooth as eh original surfaca. Thasa sections ware capped, filled with Jisclllad water and atored In a 140F oven. Water saaplaa were removed after 24, -.8 and 72 hours and analysed for CHCI4 and CCI4. Comparison of chose results, TabLe III, with the corresponding values for the pipe "as extruded" (Table II) demonstrates that there la Increased algratlon of residual solvents from the -nachtried surface. The Increase Is greater for the pipe section containing tr.e larger amount of residual solvent. As already pointed out, when the "as ex truded" pipe sections were extracted nearly the seam amount of CCI4 migrated from both (Table II). However, when new samples were machined and then extracted, the amount of CCI4 which ailgratad depended on the aaunt that was in the plp>-. Obviously, the pipe walls are not uniform In tens of CCI4 content and It seems inevitable that extraction rasulta will very widely. Thus, It would be very risky to use such data to predict the migration of the residual solvents In any use application. fltis Itiii Since the pipe sections selected for this experiment were the seme materials as used In the oven teats, It was expected that the extraction results would also be similar. In the first trlsl, tbs pipe sections were filled with cop water and extracted over a 10 day period. The water temperature wet maintained at 160F and the resulting surface temperature wee 12S7 (when the ambient air cemperacure was 797). Thus, there wee a 35F gradient across the pipe well (80 mils). Analysis of the tap water (control) indicated Chet it contained 79 ppb of CHClj but no CCl^. Periodic analysis of the water samples showed that the CCl^ content was at a maximum after 1 or 2 days end then decreased with longer exposure time (Table IV). As noted earlier, e decrease In CCI4 with Increased exposure time had also been observed at 180F'3'. Since there was no brass necal In contact with che water in che current study, Che deerneee la CCI4 can not be attributed to a reaction with braas(*). Apparently, other chanleal reactions are occurring. This heat exchanger experiment and dm previously described oven teat have common characteristics. The asms materials were used la both and the water temperature m contact with the CPVC surface was the asms. However, comparison of che data in Tab la I and Table IV show dramatically different extractions. For example, In che oven teat after 24 hours the water contained about 800 to 900 ppb of CCl^ irraapectlvn of ths residual solvent level In the pips. However, in the heat excharder experiment, after 24 hours water from 163-17-50-1 contained only 117 ppb of CCl^ and there wee only 22 ppb la the water from 163-17-SO-2. It was speculated that either the use of tap water or the presence of a taaperature gradient across the pipe wall wee responsible for the decreased extraction In the heat exchaapsr test. in che second trial, new sections of pipe (163-17-50-1 sad 163-17-50-2) were installed and filled with distilled water. The taaperature of the water (in che annular apace) wee 13tF and the surface taaperature of the pipe was 114F with an amblaat air taaperstnre of 79?. Thus, there was a 24F taaperature gradient across the pipe well. Hater samples were removed after short cine intervals and analysed to determine how rapidly e aaxiwue extraction wea reached. The experiment wea continued over e 19 day period. The analyses (Table V) show chat che aaaount of CCI4 in the water reached a eexlmum In 12 to 24 hours and then gradually decreased with seat1sued exposure. Thee* results, like those obtained c 160F, are In sharp contrast to cowperable data from the oven cast. 8 0 0 GGZZZ BFG19831 For example: Time (hrs.) WU (Fpfr) iR Kmc heluaf r 30-1 50-2 QVtn 30-2 24 313 3 314 385 US 309 37 601 595 72 335 35 566 664 192 283 11 1401 1519 364 114 0 1697 2002 IC * clear that different factors euat be controlling Cb migration in the two experiments. The most obvious difference Is the feet that there Is a cempcrature gradient across the pipe wall tn the heat exchanger experiment. However, it is difficult to explain why this gradient didn't affect both pipe staples the saae way. The fact that the seas general trends were obeerved at both 160F end 140F suggests chat it was not an experimental fluke. Nevertheless, the Inconsistencies in the results of these two experiments demonstrates that the algratlon of CCl^ from CPVC can not be predicted. Simulated Hot Water System The CPVC materials used to fabricate this hot weter syetea contained very high levels of residual "solvents": 3/4 lech pipe 1/2 Imch pipe 1/2 Inch fitting CMClj (ppm) 229 240 240 CCU (W) 6900 6938 6592 Although those valuee are enmeshat higher then the "average" (77 ppm CHC13 end <*677 ppm CClg) found In previous analyses''). It is llkaly that a portion of the CPVC compounds do contain similar high eoacsntratIons'. Thus, this hot water ayscae ta probably typical of moms existing lastsHat loos. As pointed one kfon, the water system was flushed three tlaaa before it was placed la eatwtem* Samples of wstsr were collected at location #7 before and after each flMp these wee e tins delay (about one hour) between each flush so chat the wntht. in the task could recover temperature. Consequently, the water sample VMawmi before the flush bed been in cootace with tha CPVC longei chan the voter sample takas Immediately after the flush. Analyses showed chat this raaldeswa time Influences the amount of solvent which had algrated. OKly (ppb) CCU (PP<>> 600G SZZ7. first Plush Before After Becemd Flush Before After Third Plush Before After 47 52 71 <1 11 4 99 4 301 <1 179 4 BFG19832 8. It wu surprising Co find that significant quantities of CCl^ migrated into the water In such abort periods of time. In hli thr cases, the water which had been In contact with tha pipe for about ona hour contained acre CCI4 than rhe water removed immediately (after flushing), There had baan conjecture chat during extrusion a surface layer would accumulate on the interior surface of the pipe. Conceivably such a layer eight be removed during use. The analysis cited above indicate that, if such a layer exists, It la not readily removed. Although thla system delivers hot water, the pipe la at elevated temperature* for relatively short periods of time. Obviously that exposure to elevated teepeiacute depends on tha amount of hot water flowing through tha system. For exaarple, at location #1 (sea Figure 4) the thermocouple on the pipe surface indicated that the maximum temperature was about 110V. glace that location was scheduled for 7 uses par day, tha temperature of the pipe was above 100F for about 51 of the time (as estimated from the recorded temperatures). On thu other hand, at location #7 there were 19 uees per day and the thermocouples indicated that .ue water temperature reached 125*7 while the pipe surface wee about 115F. The temperature charts showed that tha pipe was above 100*7 about 201 of tha time. Of course, the eoctioa of pipe batesaa the hot wetar beater and tha first tea (see Figure A) hod the moat exposure to high taaperature. 411 tha water flowing through thu system (it5 galloon par day) passed through this section. According to the thermocouples, the water temperature la this section wee usually about 1407 (during each of tha 454*ooo) oed the pipe surface was about 125F. Tha recorded temperatures indicated that the pipe wee above 100*7 about 70i of tha in tha typical home, there are frequent drama of water so thee the contact time with the CPVC pipe Is relatively short, la this hot water system, there were 4 5 draws per day for each wash day. lamever, at tee locations, representing the laundry, no draws were made on the weahami. Ooneeqwamtly, water removed on Monday morning from those locations had heme in the pipe for about 61 hours. Analyses of those samples showed that the water contained relatively high values of CCI4. Location Cl3 (pph) PCI*JPP> #S 110 #4 73 55 20 Presumably meet ef the samples removsit from this system will contain less CCI4 because they will keen shorter residence times. since this experiment wee deeigpnd to rapreseat the migration encountered in a typical hot water system, rhere were ee special precautions uaad whan the water samples were removed. A valve was opened end the flowing water wee directed into j glass vial which was sapped after filling. At the time of analysis, ths cap was removed end a pertlee ef the water wee ramoved with a hypodermic syringe and injected into the GC ualt. Of teeres, there were opportunities for volatiles to ?*cap during these transfers. Iks errors inherent in the sampling and analytical procedures were examined by tabiag neltiple samples at two locations. The analyses wra fairly coneIstest tmi therefere It tee be semsluded that the sampling errors were small. j,* 0T06Sz z BGH19833 t t / V. CHC13 (ppb) CCI4, (Ppb) Location #2 2U hour contact t lme Location #6 pH hour contact time A B C D A a c D 25 34 25 27 21 61 83 70 <l <I <l < 4 4 9 <i .unplea with the longer residence tlaw do have some scatte jc 31L me vii_ n the low range where analytical probleaa are greatest. !However, it was czling to recall (eee above) chat che pravloua saeiple collected under similar mentions concerned aoaevhac aore CCI4. The hot water ayetea waa placed la operetloa on 11/10/73 and che Initial samples were taken after 42 hours. Saapllng was coatlowed over a 9 week Interval. Since the saaples were taken between 8 aad 9 In the aorains, the water fro 4 locations had resided in the pipe for only 1 or 2 hours while the contact tlae at the other locations was about 20 hours. However, as pointed out above, the water remained 31 an elevated tanperature for a relatively short tlae. The analyses of these water saaples are shown In Tables VI end VEL It should be noted (Table VI) chat the control water contained significant quantities of chlorofora. The concen tration ranged froa 3 to 102 ppb with so everape of about 43 ppb (during the 9 week test period). The control saaple wee tehee froa the inlet to the water neater and therefore did not nocooonrily represent the water that was in the CPVC pipe. However, due to che varying aaount of CBClj io the control water, it was not possible to corroct the analysis of the water saaples. Nevertheless, since most of the analyses (701) shew that the water which has been in contact with CPVC contains aore CBCI3 (than the control water), there oust have been some migration. f course, Che algration of CClg le of grant concern. During the 9 week Interval, onjlyses showed that about half the tlae the control wetnr contained a snail, out measurable, aaount (epproaiaetely 3 ppb) of OCI4. The eaelyeee of the water samples which were rirrsfl froa the CVfC eyecan were not corrected (for che aaount in the control watot) hot tha dote (Table VII) indicate that, at class, there was considerable algMtloa of OCI4. Xt waa surprising to find that, at each location, ch aaount of G&g fa tho satsr varied widely ovoa though there wee no known change In reeidasMO tie* or teopereture. About 101 of tho seaplee contained non* detectable queatidoo <<l Pk) of CCI4 (see Table V11X) but the other saaples contained relatively largo aaouats. Tho higheat concentratlone of CCI4 were detected in wetnr whleh had regained in contact with the CPVC pipe for che longer periods. Thus, water withdrawn froa tho ayetea after 1 or 2 hours residence tine contained froa <1 ppb up to perhaps 20 ppb of CCI4 but if the water regained Ln the pipe for 20 hours, it eeaetiaoe contained novo then 100 ppb (Table VII). The fluctuation# la CCI4 content appeared to be raadoa with no evidence that the pipe surface wee changing. It wee difficult to retlonelIts the fact tha' large laounts of CCI4 appeared to nlgrate ln ehort periods of tins at ralatlvely low comparators (Table VTI). At other tinea, seaplee collected froa tha bssm location after identical exposure contsteed aore detectable aaounte of CCI4. The CPVC N> Is) 01 iO |ojs4 BFG19834 10. system contained "atand pipes" upstream from aach outlet (Figure 3). It is common practice to lnatall theae aectlona of pipe ao that air will be trapped in the line. The air acta aa a "cuahlon" when the water preeaure fluctuate* and chu* prevents "water hammer". It waa speculated that soma of the water trapped in the "standpipes" regained In contact with the CPVC pipe for long periods of cine and therefore contained high concentrations of CCI4. Subsequently, during sampling (and depending on flow rate and preaaura) sows of that water elxed with the sample. This uncontrolled mixing could result In the random fluctuation*. It was decided that theao conjacturoa should bm axamlnad bp coilacting special samples. The automatic timers on the diacherfa valves were dlacoonectad ao that water couia be removed from the system only by manual operation. The CPVC lines. Including the "standpipes'* were then emptied of aeaantlally ell weter. The ayatea was refilled with hot water aad allowed to flush briefly ao that all lines wars filled with hot water. After closing the vsIves, the weter remained in the lines for 6 hours and then wea sampled ead analysed. Subsequently, the entire procedure was repeated three tines except that the meter remained in the lines for ^erloda of 12, IS aad 24 hours before sampling. Aa shown in Tablo IX , only 2 of the 36 samples contained a detectable amount of CClg aad aven thorn the concentration was very low. Those date, although not conclusive, support the postulation chat the random seep lee containing high concentrations are from "stagnant" weter. A somewhat more fundamental study of the migration of residual "advents" from CPVC compounds has beam undertaken by Staff Technical Services'*). That work Indicates that the dlffnslvtty of CCI4 decreasea as tha residual concentration is raduced(*>. There la also evidence showing that tha diffusion coefficient of CC14 is Influenced by the Ingredients mead la tha CPVC compound<L). Thus, any haterogmnolty in tha pip* could rooelt in variability la tha migration of rasldual CClf Tha ms asmred diffusion coafficlaat aad its oatlaatod activation energy were used to calculate the awpswtmd CClg contest of motor after contact with CPVC pipe under proscribed conditions, these calculations provide values which are in reasonable agreement with the "everages" detected la the meter samples takan from che hat water systems P*). This comlatency suggesta that tha CC14 contant of watar cam be predicted ter any desired use conditions. For example, If tha acceptance criteria ter migration of CCI4 la a maximum of 2 ppb in a typical horns, than tha calculations Indicate that tha CPVC pip* should contain* maximum of 100 ppm of ramideal CClg'1*'. `*1 BFG19835 JN) is) O] o 986IOdg 58S0* 091 611 SOU 8101 ou 68 m 818 Tf"T noo Ul in 81 611 18 U *9 69 jm M3D 9S91 m Ittl m on 691 ICI m 1001 m ltd til mi 9 4M 6S TTDSD TWT ClDB WTT-Tf!----- 4N1 aw HllTinv - 1W1MJUI 098 918 861 91 96 IL 89 98 TERtf U Mti 5AJB ai"ttilTHi IHH1 i* wmiiwi rnx Ia*la II HUrit Ian of ImUmI "ioWinf" fw CTK PIpi Ow T--t east - 4ttllUd NtM MN 1404V Tlaa (tab). 2 1*3-17-*M CK1 ^Cl* -- 21 42 9* 6-- tt 24 1 314 44 3 441 72 ll 3*4 94 3 m im 44 1343 m . m . . 21* ^iSfcf|t.|.;a./ 3* M ' 43 1441 171* 1*97 M 44 2334 444 It 2*44 1*3-17-30-2 CSKi1? CCl* (--41 2 SI 54 4 107 * 3*3 4 593 3 644 11 830 21 1300 23 1*19 23 1444 31 2002 34 2093 >2 1M0 BFG19837 U\ CD O Tim 1*11il 24 4S 72 xakiuai fnmcrrc fis. btraetMt - ttitllM tracer Tfant>n IMP? mefct--< lMriit Mrfee* of plpo l&fcUafcl aa, CCI4 rad lab 12 117* 94 4*43 3* 3*11 IteUzXbl OCI4 ecu Sam Until 23 328 2* 776 11 939 .. -V- '* x* 40532 BFG19838 CJ\ > o k" ? tf IrtU Vi Migration of 1--1 <m1 "BoUattta" Txom CFVC P1p Bmt ttrtwniL.lt tatxmtmt - cf aur - 14F TLm toihl 24 a 72 94 128 168 192 214 ^ *4.Jfe 143-17-30-1 csci3 CCI4 (skL !*! 38 117 42 149 n 125 86 45 98 32 114 32 115 54 Uft 19 105 17 143-17-50-2 OKI) CCI4 xml 42 22 52 24 53 24 52 22 83 2 44 2 72 2 48 1 27 2 2259016 BFG19839 15. UUUL. Miration of 1--1*1*1 "Solmnta" Trou CTYC PIpa Hoot Uchonoor Twt Extractant - distillad watar Taaporatur* 13tT Tlaa Qtflil 2 163-17-30-1 caci3 CC14 XBfel Xnfcl <i 36 6 <i 101 64 US 12 4 311 24 6 313 44 4 309 72 5 333 M-- 173 IM 13 331 191 222 *|L t u 203 232 300 444 <1 1U 32 163-17-30-2 cacis CCI4 f--41 iegfei, <i <1 <1 <1 <i <1 <1 20 3 u 37 2 33 <1 17 7 21 4 11 08 <1 l BFG19840 t\: v\ CoD li u 8 i 1 u2 N40^f<P w- NNN3N^2I*2K 40 PN0N^S ^ A fwSO 888888^s;8 88S88888888 8 888888Sn888 S m 'I*^hAnmIihmng * ^An n.* o 8NNNNSanq3nNNN nM 16. 888S88m838 2 J- Is IP *S5ISISiS BFG19841 ?o N u\ CD O I-- GO 1 b 32 35 N VV V V ** V V v v V V - VVVVVVV^V 3 aaj-^gsss- V VvV V V V V 4 N <\MO *>4IA<N4>< a-4 35v v v - v -< v v v v V VV V V 1i jji **smmn i iiffsiill^ 40539 BFG19842 6T0S9ZZ T is. Slt",rv nf XlUUttU |4--lli Of Watt c SiMlei Location Control 1 2 3 4 5 6 7 8 CoM Rosldooco TtM -tettii. -- 2 24 1 2 20 21 l 14 . ftimfrir nf ITmbIm Cniifimno Low IMariti Largo <1 p*4 >1 <30 p?4 >30 ppb Kongo (ppb) 5 5 i 3 1 2 - 4 4 4 <1 1 4 <1 15 4 2 <1 181 3 <1 12 4 <1 21 4 4 <1 443 4 3 <1 148 3 <1 21 3 <1 15 7 <1 21 Meatu Aaoun (PP*>) 3 6 44 5 9 75 46 13 6 7 2253020 BFG19843 Tabln n Migration of Kosltual "Solaanta" fna CTVC Mpa 2k caci3 CCI4 23 <1 5 <1 <1 <1 3 <1 41 <1 7 <1 1 <1 1 <1 1* 8 BFG19844 22259021 a 3r**s /** Z9f> $1 C/^/ry/^ /w^ CP*C *** ii *vW*e t*vt /*?* BFG19845 X & C 5 i)ijZ 3 9^861099 t----------- :------------1------------:------------ 1------------:------------ r &Z069ZZZ BFG19847 -rt*4 ' ' T2S ife-... 24. /r **/ /nr V-" i BFG19849 2259026