Document Xv6jrX7Dkg4ojD7k6Z3Gb8DJ

Cy*aui>4(OMdt E. I. du Po n t de N e m o u r s St C o m p a n y IMCQHPOflATED . P. O. Box 1217 Pa r k er sbu r g . W . Va. zeioi POLYMER PRODUCTS DEPARTMENT AR226-2380 CC; AE.. Wegchelar D.Champney -DoDr1d1r0e7c0ht R. A.Morgan . R. E.Putnam T D.Gierke R. G. Brown D. A. Erdman 0. T.Garza H. M.Cooper File: 8-7, 8-10, 8-42 November 30, 1981 TO: J. H. EANGENDAM - PD II WILMINGTON FROM: T. J. LECK ^ WASHINGTON LAB0K ANALYSIS OF EFFLUENT STREAMS FROM FEP Because of the need for environmental emissions data for Pro ject 8533, "Dordrecht Copolymer Finishing Facilities," a program was initiated to collect and analyze samples from all of the FEP plant effluent streams. Several additional samples were taken and analyzed so that material balances could be determined. Also, since Project 8533 includes a scrubber for the exhaust from the humid heat treatment oven it was necessary to assemble a small scrubbing apparatus so that pilot scale scrubbing of HHT oven exhaust could be done. Scrubber water from this unit was analyzed for chemical composition and was tested for aquatic toxicity at Haskell Laboratory. Harry Cooper of FEP Technical assisted with sample collection in this program. The aqueous samples tested are listed in Table TJL-1. These samples were analyzed for residual C-8 surfactant, organic and inorganic carbon, total fluoride ion content, and pH. Samples .were collected during production of T-100 FEP at normal production rates. The fluoride ion figures for the first two samples listed (coagulator effluents) are a factor of 10 higher than is expected for this stream. Repeat samples of coagulator effluent and raw dispersion supernate, which should have the same fluoride content, agree with historical figures of 300 mg/l fluoride or less. This is pointed out since this data was transmitted to you earlier. The two high fluoride levels should probably be dis carded. The other analyses on these two samples are consistent with expected values. Vent samples were collected from the following locations: torus disc dryer exhaust (ahead of the exhaust scrubber), torus disc JUZ003294 There's a world of things we're doing something about EID 931230 J . H. LANGENDAM - 2 - NOVEMBER 3 0 , 1981 dryer exhaust (after the exhaust scrubber), #1 humid heat treatment cooling tower exhaust, and the #1 humid heat treatment oven exhaust. Samples from each location were analyzed directly by electron capture gas chromatography and then by scrubbing the gas samples through water and analyzing the water for extracted compounds. The humid heat treatment oven cooling tower exhaust was analyzed only by electron capture GC. The only compound which could be detected was the C-8 decomposition product C y F ^ H at 3 ffiole3 per million. At the torus disc dryer exhaust we expected to see primarily C-8 surfactant and perhaps some oligomeric dicarboxylic acids which are expected to be formed in the polymerization. -As expected, GC analysis showed no volatile fluorocarbons. Analysis of water through which exhaust samples were scrubbed gave the results shown in Table TJL-2. The major component was C-8 surfactant. The HF and C0,, were just at the detection limits of the methods employed for their analysis - the figures given are upper limits to the actual levels. No oligomers could be detected in these samples. Results of the analysis of the humid heat treatment oven exhaust are shown in Table TJL-3. The COy, C-8, HF, and oxalic acid were measured by analyzing water collected from the pilot scale exhaust scrubber. The ocher compounds were detected directly by GC. This data was taken with the oven operating under normal con ditions as listed here: Product Oven Feed Oven Temperature Air Make Up Rate Humidity FBF T-100 500 lb/hr 36Q*C 500 cfm 60" to 65"C, wet bulb (about 5% HyO) Historical data on the composition of the humid heat treatment oven exhaust is sketchy at best so comparisons with historical data are difficult. The best historical data are those of J. F- Doughty, collected in 1978. Doughty detected TFE, HFP, F-23, K-125, and PF1B in the oven exhaust at levels which are comparable to my measurements as listed in Table TJL-3. This supports the sampling and measurement techniques used in this project. By summing all of the components found in the oven exhaust to get the total exhaust loading of vented volatiles and multiplying this by the temperature corrected oven exhaust rate one can calculate the rate at which these materials are vented from the system. This exercise was performed and summarized in Table TJL--3. The resulting material loss rate was ratioed to the oven feed rate, and it was found that about 0.95% of the oven feed is vented as volatiles. This figure is In very good agreement with measurements of polymer weight los3 performed in 1979 by R, A. Morgan. Morgan's data shows an average weight loss of 1.13% for dried FEP samples in the humid heat treatment process. RJZ003295 EID 931231 J . H. LANGENDAM - 3- NOVEMBER 3 0 , 1981 The pilot scale scrubbing of humid heat treatment oven exhaust was run at a low rate to be su^a that a high scrubbing efficiency was attained. Altogether 127.4 ft of exhaust air was scrubbed through 60.5 1 of water. This scrub ratio resulted in a fluoride level of 16 tng/1 in the scrubber water. Scrubbing was stopped at this point to keep the fluoride below its aquatic toxic level of 60 mg/l for fish. The analysis of .the water is shown in Table TJL-1 and in the attached Haskell Labora tory report. The 60 liters of scrubber collected was shipped to Haskell Laboratory for aquatic bioassay with fathead minnows and water fleas (Daphnia Magna). These tests show that the full strength scrub water was toxic, but when diluted about 1:1 with pure water the scrub water was rendered non toxic. The toxicity appears to be a result of the combination of the oxalic acid, hydrogen fluoride and perhaps the low levels of iron and nickel leeched from the sampling line during scrub bing. From these results it is clear that aqueous scrubbing and dis charge of a diluted effluent stream to the river can be managed in an acceptable fashion. The Haskell Laboratory reports are attached. TJLECK:ssb Attachments RJZ0Q3296 EID 931232