Document 4O63ND1qMKE1rOoDdBvO2qoQ

COMPANY CONFIDENTIAL RESEARCH AND DEVELOPMENT REPORT ANNISTON PLANT OTHER LOCATIONS GENERAL OFFICE P. B. Hodges F. J. Holsapfel W, B. Papageorge W. R, Richard J, R. Savage Rvport No._______1_________ Job No. A46-S1A___________ Polo. Jinmry IS, 1971 ANNISTON PLANT J. T. Bell J. L. Etheredge V. R. Haupt J . C. Landwehr G. W. Miller W. F. Taffee E. G. Wright Title* Fertemel* Preblewi Aroclor Pollution Study 1. Ranaaw & A. Nelson (J. L. Brown) Removal of Aroclor from the waste water by filtration-adsorption. Jeetlllcetleei Pollution Control haeayi Aroclor waa removed from the plant affluent by filtration of the neutralisa tion pit solids on which Aroclor was adaorbad and by adsorption on the filtration media. Aroclor concentrations (100-8,000) In the plant affluent were conelatently reduced to 10 ppb upon passing through a four (4) inch diameter tube (packed to bed depth of S.5 inch with activated carbon) at flow rates up to 350 cc/min. In the sand filtration study conducted under similar operating conditions, the Aroclor levels in the effluent were not consistently reduced to the target level (< 10 ppb). The pretence of 200-1000 ppm of solid particles (size range 1.5-25Q40 tended to blind the filter, thus setting a maximum limit on the pressure end flow at which the unit had to be operated. lab Ranaaw PR&D Chemist HONS 042646 Aroclor Pollution Study PR&D Report Pago 2 BACKGROUND! Initial Rasaarch afforta for removing Aroclor from tha waste atraaai were cantarad primarily around thraa conventional approachaa: multiple distillation, adsorption and extraction. After considering the economic and technical factors involved in each approach, it vaa fait tha approachaa should be evaluated in aaquanca Hated above. Tha following paragrapha are presented to up-data previously research efforts and to show tha basis for the work being reported. Cost evaluations were made by Dr. Wise, CED, for both a diatlllation and extraction system for removing Aroelor from tha waste water. Dr. Vise concluded that the coat for either of these systems was likely to be considerable greater than that for an adsorption system. A decision was made at this point to abandon these approaches and pursue the adsorption approach in depth. Initial screening of various adsorbents (i.e, carbon types, porocel, attapulgus clay, powdered polyethylene) by adsorption Isotherm tests was very Inconclusive. The Aroclor levels were reduced to < 10 ppb in all the isotherm tests regardless of the adsorbent or the amount of adsorbent used. The reason for the erroneous results was later found to lie in the fact that Aroclor was being adsorbed on the reactor walls, particulates in the solution, as well as the adsorbent being tested. Prom the isotherm studies, the following conclusions wars drawn: A) Aroclor levels in plant effluent can be reduced to an acceptable level (< 10 ppb) by adsorption, b) Further evaluation should be done on an adsorption column utilising a feed of constant Aroelor concentration. The studies reported herein were conducted at the original neutralisation pit (hareafter designated as pit #1) and newly constructed neutralisation (hereafter designated pit #2). Pit el was distinguished from pit #2, during the course of study, by both its consistently higher suspended solids and Aroelor concentrations. Pit #1 has been operational since 1961 while pit #2 has been operational since October, 1970. In a 13 day pilot run at pit #1, initial concentrations of l^)00-8,000 ppb Aroclor in the waste were reduced to a range of 0-15 ppb upon passing through a series of packed carbon beds at a average flow of 23 cc/min. Over 99% of Aroclor removal occurred in the prefilter. This unusual reduction was found to be due, in part, to the filtration of solids on which Aroclor was adsorbed. Since the average flow per square foot was quite low for this particular study, it was felt that the prefilter observation should be evaluated at higher flow rates. This report contains the result of the studies carried out using both carbon and sand filtering media for treating waste water from pit #1 and #2. 1. The adsorber consisted of a prefilter (3.5 inch diameter packed to e 5.5 inch bed depth and a series of one (1) inch diameter columns packed to a total bed depth of 13.5 ft. MQNS QHZbbl Aroclor Pollution Study PR&D Report EXPERIMENTAL Page 3 2. Semolina A 24 hour composite temple (300 ml) wit token by catching a 100 ml. Maple every 8 houre. Occasionally, smaller samples were caught at shorter intervale and composited. 3. flow Rate flow to the filter mas controlled by adjusting two needle valves (diagram above). A calibrated rotameter was used to measure flow. A given flow rata was difficult to maintain due to the back pressure created by the build up of fine solids in the filter. The unit was back washed when hecessary to maintain a relative constant flow. In the studies conducted at pit #1, flows wars started at the lowest rata (100 cc/mln.) and than increased to maximum 330 cc/min. This approach was reversed In the sand study at the pit #2. The maximum flow (350 cc/aln.) could not be meintained over an eight (8) hour period due to the high back pressure. * AtalYtU.1 AialviU The Aroclor was extracted from the aqueous simple into a hexana solvent. The hexane solution wss then dried end analysed on a BC/GC. The concentration of Aroclor 1221, 1242, 1254 waa determined by comparing the main peak height of the unknown with height of the tame peek in the chromatograms of known Aroclor solutions of 1224 1242, and 1254. HONS 042648 Aroclor Polluclon Study PR&D Report Peg* 4 DI3CUSSI0M OF RESULTS For clarity of tha following discussion, it ia neceaaary bare to explain tha difference in tha quality of the effluent from pita #1 and #2. During the couree of thle study, the effluent from pit #1 generally contained about 1,000 ppm of eollde and Aroclor concentrations in the range (1000-80000 ppb) Table 1 & 2. A prolonged study of the effluent quality (i.e. solids and Aroclor levels) indicates that high Aroclor concentrations generally correspond to high solids. Presently, the solids level in pit #2 is kept at minimum. The effluent from pit #2 normally contains about 200 ppb of solids and 50-100 ppb Aroclor except during periods in which large quenities of acid (HC1) ere being sewered. This accounts for the higher values seen in Tables 3 A 4. Carbon consistently reduced the solid and Aroclor concentrations in the effluents from pits #1 A 2 to an acceptable level (< 10 ppb) at flow ratae up to 350 ee/min. (Tabla 1 A 3). Tha machaniam by which Aroclor la bting removed cannot totally be explained* Exlating evidence indicates adsorption on the pit solids* However, attempts to determine the adsorptive capacity of the eollde were unsuccessful. The lower quantity of solids (200 ppm) in the effluent from the pit #2 had seemingly little effect on the quality of effluent produced. As It can be teen from Tables 2 A 4, the quality of effluent realised in the send filtration was not as dramatic at in the carbon study. Aroclor lsvsls in pit #1 affluant (Table 2) were significantly end rather consistently reduced without producing an acceptable effluent (< 10 ppb) quality. Comparatively* the effluent quality produced at the pic #2 wee very inconsistent (Table 4). The effluent quality varied between high end low values without drastic changes in Influent Aroclor levels or operating conditions. The following factors may explain the results obtained: a. The eollde level of influent varied considerably during the test period. b. Silica type adsorbent requires a considerably longer contact time to adsorb Aroclor then carbon. The teat data for both the send end carbon studies suggest that the Aroclor removal ia taking place by filtration of pit solids on which Aroclor has been adsorbed end by adsorption on the filtering media Itself* Further, it also suggests that the adsorption eons with carbon is quite short end quite long with tend* The eollde in the wests stream caused considerable operational difficulties. These eollde, of which 42% of lte particles were in the 1.5 - 2.3^11 else, tended to blend the filter, eepiclelly St flow rates > 200 ce/mln. Backwashing the filter did not adequately remove the finer particles. Hence,thsss Aroclor con taining particles were eventually washed through the filter. This point was considered a breakthrough. 042649 HONS Aroclor Pollution Study PR&D Report Past 5 CONCLUSION 1. Aroclor la removed from the plant affluent by the filtration of the eolide on which Aroclor wee adsorbed. 2. Aroclor ia adaorbed on activated carbon within a very short adsorption tone. 3* The effactlveneaa of sand in removing Aroclor la not aufflcient at the experimental bed depth. 4. Solids removal la a primary concern in any final clean up method. mm wpm 1. A pilot adsorption study has been conducted in a series of 5 inch diameter columns. These result are being compiled for an evaluation and reconaendatlon. A report will be issued by 1/28/71. 2. A pilot clarification (flocculation) unit is scheduled to be put in operation by 1/29/71. This unit will also be uaed to carry out aemlcontlnuoua adaorption studies with other finely divided solids. Attachments *QNS CARBON FILTRATION STUDY AT OLD PIT TABLE 1 Page 6 Volina Treated (Gal.l 38 99 160 *175 198 236 Aroe lor Coricontr tlon oi Influnt (pp b) 1221 12&J[1334 84 106 NDA 1730 188 NDA MCA 618 NCA 6051 1638 MM 559 147 NDA 2106 441 NDA Avaraga Dally Flow (GPU) Aroe lor Concon era don 0 f Sfflusnt 1221 (ppb) 1242 1254 0.0263 0.042 0.042 0.0263 0.0263 0.0263 NDA NDA NDA 21 14 MCA NDA NDA MM ? 5 NDA NDA NDA NDA 4.0 1.2 NDA Unit plugged. Solids wort oread Chr >Vgh aa tht praaau rs on flltsr 1 created. 8A1IP FILTRATION STUDY AT OLD MCUMALIZATIOH PIT TABLE 2 Voliaea Traatad Aroe Lor ConeonfcirAtlon oi[ lafltimt <F1>b) ______mi -USi 38 4100 4000 MM 76 8000 2750 MM 114 152 48000 365CX NDA 190 2394 844 NDA 228 1B00 400 NOA Avaraga Dally Flo* (CPU) 0.0263 0.0263 0.0263 0.0263 0.0263 Aroe Lor Concan trotion ?f Bffluan : 1221 1242 1254 . 138 5 -157 12 178 12 MM . MM NDA MQNS 042651 Voluaa Traatad (Cpl.) 38 76 114 132 CARBON FILTRATION STUDY AT HEW FIT Fat* 7 TABLE 3 Aroel or C01icaneirot Lon of Infliionfc (Pi>b) 1221 \m 1234 Avtrigt Dolly Plow (GPH) Arodor Coneone ration of fflaont ppb) 1221 1242 1234 31 109 NDA 0.0263 NDA 2 NDA 4 40 HDA 210 32 HDA 0.0263 0.0263 NDA NDA NDA NDA NDA NDA 340 4 USA 0.0263 NBA 4 NDA Voluaa Traatad (Cal.) 36 76 163 277 364 446 316 BAUD FILTRATION STUDY AT SEW FIT TABLE 4 Aroel or Coneontr at ion of Influ nt (pn b) 1221 1242 1234 no 72 NDA Avaras* Dally Plow (CPU) 0.0263 Aroelor Coneantiration of ffluant 1221 1242 1254 48 3 NDA 1200 620 NDA 0.0263 0.0605 - -HDA NDA NDA 260 60 NDA 0.079 NDA NDA NDA 840 130 NDA 0.0605 7 NDA NDA 610 180 NDA 0.057 60 NDA NDA NDA 560 NDA 0.0482 55 NDA NDA HONS 042652