Document 8EJvK71Vn8V8MQv3NQN0bKmZ

SAFETY & HEALTH WITH PLASTICS NATIONAL TECHNICAL CONFERENCE Society of Plastics Engineers November 8-10, 1977 Denver, Colorado BFG22241 SOCIETY OF PLASTICS ENGINEERS, INC. r; 2643001 4U>* I (9*r\ WATER EXTRACTION OF ADDITIVES FROM PVC PIPE n ig 04 tu^ CSaur I-? 04sue/ aj-*,/ ^**A G.R.Dicu* w **0- Aff/** *** ^ (jut{A f' t*9 . Diamond Shamrock Corporation Plucks Division Post Office Box 191 Painesvilk. Ohio 440T7 A)o hov^ , Oo 'N-*f` *G>Authm: J.D. Umio E.M. Miller c4j O Yo't oie-SU* WATER EXTRACTION OP ADDITIVES FROM PVC PIPE INTRODUCTION Although numerous data ha* boon dnvnlopad concerning the water extraction of VCH from PVC pipe, no date was found in the current open literature on the extraction of tin stabilizer* and other material* used to produce potable water pipe. A limited amount of information was found regarding the extraction of lead based stabilizers that are used for some European potable water pipe service. <1> 2'> (3) (4) (5) (6) Therefore, because of our extensive work on extraction of VCM,f?) we docided to initiate e study on other direct pipe compound additives to insure the total safety of PVC potable water pipe. EXPERIMENTAL Pipe for this study was extruded using e 2*1/2 inch single screw extruder with typical pip* die and vacuum sizing -eguipoant. Product mad* for testing was on* inch schedule 40 pip*. This typ* pip* could be used in e variety of applications, including cold water piping systems. The formulation used for producing th* pip* for this study is shown in Table 1. Pip* samples for this extraction testing were pre pared by cutting 14 cm. lengths from sach pip* com pound extruded. Eight lengths of each sample to be tested wpre placed in a glass jar whieh was fillad with 2.0 liters of deionized water. This surface to volume ratio (pipa surface to water amount! correspond* to 4 ml of water per square inch of pip* surface as has been recommended by the MSP. These jars were than sealed with aluminum lined ecrew tops. Duplicate jars were prepared for each pipe sample. One jar was stored at room temperature (2S*C) and the other at 50aC. Jars were stored for two week*. At the end of this initial two week period, a 100 ml sample was withdrawn for trace metal analysis. After sampling, 100 ml of fresh deionized water was added back to each jar and eh* jar returned to storage. Subsequent analyses at two week intervals were corrected for the dilution effect of the 100 ml of fresh water. A blank sample was run at aach temp erature. ANALYTICAL TESTING PROCEDURE The samples obtained for this tasting Were analyzed on a Perkin-Elmer Model 503 Atomic Absorption Spectro photometer operated in the absorbance mode. A single 100 ml aliquot was withdrawn from each sample and blank jar. Samplat were preserved between sampling and analysis with 0.1% nitric aeid (Baker (Jltrex) to prevent leas of metal from solution by precipitation or absorption. Tin and titanium were analyzed using the graphite furnace. Calcium and magnesium were analyzed using an air-aeetylcn* flame. For calcium analyses, lanthanum chloride was added to sample and standards at a 20% concentration to suppress calcium ionization. At least three replicate analyses of each sample were preformed. All results were corrected for blank value*. RESULTS AND DISCUSSION The first data presentad in this study are the tin extraction values. These results (Table 2) show the individual levels extracted on two butyl, two methyl, and on* octyl tin stabilized pip* compounds. Th* stabilizers used are cosmercially available products. The two butyl and two mathyl samplas are not dupli cates and were suds using different stabilizers. Th* initial test results are reported after two weeks extraction at 25* end 50*C and on through eight weeks extraction at two week intervals. The data shows that relatively low levels of tin ars extracted from these pip* samples with th* higher levels occuring after the initial two week testing period. After this point we did not get an increase in the tin extraction level and the values drop off to a relatively constant level and do not go up again. These tin extraction results were also plotted a* shown in Figures I and II. For these plots the extraction value* for eh* methyl and butyl tins were averaged and only one curve is shown for eaeh on both the 2S* and S0*C plot*. These data shew little difference in the methyl and butyl extraction at 2S*C, although both show a alight doereas*. The octyl is much lower at both extraction temperatures. Th* data in the 50*C plot (Figure II) shows th* same trend with overall extraction levels decreasing with increasing extraction time*. Th* unusual feature of this S0*C plot is the rather high initial extrac tion value for th* methyl tin, followed by a dramatic decrease with time to th* point where it is about equivalent to th* butyl sample. A possible explanation for this phenomenon is pre cipitation of insoluble tin compounds or absorption of axtraeted tin on th* walls of th* jar or th* pipe. In order to verify this, on* methyl tin sample was acidified with concentrated KOI to dissolve insoluble tin compounds. Th* tin level before acid addition was 0.04? ppm.- After addition of acid, the tin level in solution increased to 0.38S ppm, an eight fold increase. This confirms our hypothesis that extracted tin compound* are subsequently converted to an insoluble fora and are prtcipitatinq from solution or are being absorbed on th* walls. Another important point that can be made from these data is the fact that with increasing chain length, fro* butyl or octyl, there is a significant decrease ZM V9Z7. BFG22242