Document VJbwMpmQYOZBM571MXK3pjLoq

Adsorption ot Polychlorinated Biphenyls from Aqueous Solutions and Sewage John l*wf*ncs' and HH* M. Tosbta Water Chamltay Section. ProcMi ftatoarch Division, Canada Centra to Inland Water*. Surfing*on, Ont, Canada LTR 4AC The adKtrjHioii of PCH's from synthetic aqueou* soiuIlona and raw sewage ha* been studied on a variety of ad sorbents, including eetivated carbons, polymeric resins, polyvinyl chloride, and polyurethane foam*. The carbons, foam*, and XAD-2siwniiy adsorbed ihe PCH't from aque ous solution* but were much lest effective in raw irwage. The PVC, however, we* very effective in raw *rws|e. A mall-acale treatment unit waa designed to demonstrate the feasibility nf removing PCBs from sewage by this method. Polychlorinated biphenyls, PCHa (tradename "Aroclor'' in North America), are ubiquitous persistent pollutant* of the environment with adverse ecological and toxicological effect* (i>. Sinee first being introduced in 1929. they have had widespread application in the electrical, palm, paper, and plastics Industries (2) and have seen ettenslve use a* hydraulic fluid* and pesticide additives. Il ha* heen **i). mated that the total world production during the pa*t 45 year* has eseeeded one million tom, of which ehoul 400 000 tom have been released to the environment (31, Concerned for environmental quality. Ihe Monsanto Co., the only North American producer of PCBs, restricted sales In 1971 to closed-system applications such as elsrtrleal caparllor* and transformers; however, it Is not known how many VCH-twrrtaining Items ar imported. Even with these resirirttois, significant quanlilis* of PCBs are sill) released iniit the environment. For reasnn* that are not fully understood, raw municipal sewage often contain* appreciable amount* of PCH* For example, raw aewage entering the Hamilton Sewage Treat ment plant contains on the order of 10 ppb PCH [mainly Arorlor 1254 and 1250 If)). Considering Hamilton's flow rale of 230 X 10* I./day (32 mgd), this fnnueni concentra tion corresjuwtd* to approiimately one metric ton entering ihe plant annually. Concentrations of PCBs in aewage from other riliea in southern Ontario, while not as high a* that from Hamilton, are nevsrlheleas ailH significant. The physical chemical properties of PCH*- i.e. low vapor pressure, low aqueoas solubility, and chemical inert ness to water, arid, alkali, and heat make them extremely persmtent environmental contaminant*. There are 2)0 pos sible combination* of chlorine auhatitiHinn* on the biphen yl nucleus but only about half that numlier are aterically poeslblc. The numlier of PCHa occurring in the environ ment is of the order of 30-40 If I. The main coolrilrutun are the various isomer* of pentachlorobiphenyl (Arurior 1234), hexachtooWphenyl (Aroclor 1300) and, to a lesser extent, triehloeobiphenyl ('Aroclor 12491. However, aumc lower chlorinated isomer* ire present a* a mult of limited degradation. Degradation usually proceed* via lurresslve dechlorination prior to breaking of the liable aromatic bi phenyl structure (3). The low soluliility and hydrophobic nature of the chlori nated biphenyl* make them relatively easy to adsorb from aqueous solution (0-9). The purpose of ibis work wm to study Ike feasibility of extracting PCH* from raw aewage by adsorption onto various media Hom-ier. het niise of the varying nature of raw sewage, the adsorption wa* fir*i si lul led using synthetic aqueous solution* of Aroclor tj'.t ami 1242 and then this wa* followed by measurement* on raw aewage collected from the Hamilton Sewiigc Treatment plant. Rxperimrnlnl The two atoek solution* of Aroclor 1242 and r>*-4 , ii prepared by vigorously mixing an exie ol emh At.tier (Monsanto Co.) with waler for A h. allowing the solution* to stand nvttnight ami carefully decanting oft Ihe Iroe aqur out phase. The waler used wa* double distilled, ihe second distillation being from an all-glass system The rsimentra turn nf these tohition* determined by gaa chromatography wa* 43 * 10 ppb. which i* consistent with the pohlUhed oiulhlitv fur Arurior 1234 of JWtppbtfll All soivsnla used were glass distilled pesticide grade (Caledon l.abnrslurle* Inc ). The activated rarlauts em ployed were lignite baaed Hydrodarco 400 IAt la* Chemical Industries) and anthracite baaed Fihrmoeb 4(91 i('aigon Cuip.). These were prstreated by heating to Hit *(' for 12 h. cooling, and twice extracting each AMI g with 2 I. of het ane The extracted carbon was then filtered and air dried The pulyureihane foam* used were IliSI'o |dug if'antah Supplies l.td.l and Foams f I in and 2t2 if* P. Goodrich l.td). Thr first two digits relate to the density- I.e. 1.1 snd 2.3 Ik/ft5 and the second two lu the hardne-.* The foams were shredded and suecessively waslu-d with n hr* ane (several lime*), acetone, and distilled water They were then air dried. The pretreatment was developed in remot e trace organic contaminant* from the surfwe f the foam*. The macrotetkuhtr polystyrene wins Amlwrilte XA|t i and XAD-4 (Rohm and Haas Co.t were peetrealrd lit am eesaively washing each 600 g of retin with I I. Imlt iu-* of water, melhatud. and water. Ilte cleaned resin* wi re sion-d in mated glass containers under methanol In prevent them from drying out. Polyvinyl chloride chip* t Monsanto Co l were washed several lime* with n -hexane and air drier! To determine (he adsorption characteristic*. IIM) ml ali quots of stock Arurktr solutions were stirred vigorously with 9-10 weighed amounts of adsorbent for ;M> mm lire quantities of adsorbent wer riuwen to rover lire range from I mg to approximately l I- After contacting. the ad sorbent wa* removed by filtration through yert-niirr pad (Millipore l.td I. Five milliliter* of n-hrxane were then tig ornusly stirred with the fihrste for 43 min ami the organic extract was withdrawn. These exlraci* were anahred with a ga* chromatograph iVariaw series) eqwwwH with an rim tron-eapture detector INI41). The gas column OH m x l A mm i.d.) was parked srilh 4% OV lOl and if- Ot .'Id on Chrumsorb W HI' 90/100 mesh. Nitrogen was u*ed a* a car rier gaa at .30 ml/min. The injection port, column, snd deterlor temperatures were 2.30, 200, and 100 *(' re*i*-rfi\*-lc Thr amount of PCH adsorbed per unit weight of ad*orla->it was then calculated and plotted as a function of the equi librium concentration of l>CH remaining in solution. Wastewater was collected fmm the Hamilton Scwag* Trratment plant at the raw sewage lTet pltre Sampling M0NS 036671 Votuma 19, Numbw 4. Apr* *9*9 991 wa* carried out lifting all-glass containers to ensure against adsorption onto container walls. The samples were stored at a constant temjierature or .1 *C and in all cases were treated and/or extracted within 24 h of collection. The con centration of PCBs in the raw sewage ranged from 1-20 pph hut was usually lietween 5 and 13 pph. In the evalua tion of PCM adsorption from sewage, the procedure de scribed previously for pure Aroclor solutions was followed except that 100-ml samples of raw sewage were stirred vig orously with the weighed adsorbent for 1 h and the adsor bent was separated by filtration through a 60 mesh, stain less steel screen. After washing, the screen did not retain any raw sewage and, with the exception of activated car bon, 100% separation of adsorbent was achieved. The sam ples were then twice extracted with 50 ml of n-hexane in 500-ml separator funnels. The aqueous portion was dis carded and the organic phase, after being dried through IS g of Na^SO*, was reduced to about 3 ml, using a rotary evaporator. Tbc sample was purified by liquid-solid chro matography on a llorisil support column using petroleum ether to elute the PCH fraction and (hen the eluate was evaporated to 3 ml (9). Prior to Injection of the sample into the gas chromatograph, it was shaken with 0.2 ml mercury to remove residual sulfur compounds. The PCBs in the samples were identified by comparison with chromatograms of standard Aroclors (Figure 1). Con centrations were determined from the peak height ratios of peaks A, H, C, and I) for Aroclor 1254 and peaks of R, F. and (I for Aroclor 1260. The minimum amount of PCH that could lie deleeted with confidence by the GC was 22 p/g in a 10-pi injection. When we allowed for the increase in con centrations during extraction from the aqueous phase, the limit of detection in the aqueous samples was estimated as 0.5 pph with an accuracy of 0.1 pph. licsitlht and Dincusnion Adsorption data for Aroclor 1254 and 1242 on activated lignite carlwin, anthracite carbon, two polyurethane foams, Amlierlite XAD-2 and XAD-4 and PVC are shown in loglog form in Figure 2. The weight of PCH adsorbed per unit weight of adsorlient is expressed as a function of the equi librium concentration of PCH remaining in solution. The error bars have not been drawn on Figure 2 as they distract from the clarity of the data: In (he worst rases, the diame ter of the error circles corresponds to about two-and-onehalf times the diameter of the marked {mints. These are in sufficient to linearise the isotherms. The sets of data do not follow any of (he common isotherm expressions- c.g., ],angmuir, Prcundlich, BKT, and so forth, and consequent ly a theoretical interpretation of the results has not been attcmfMcd. It is evident that the two carbons and XAD-2 have the greatest adsorption capacit ies but a residual con centration of leas than 3 pph could not be obtained with lignite carbon. Both polyurethane foams appear to be good adsorbers with relatively high adsorption capacities and low residual levels. DifiPo polyurethane foam plugs were also evaluated hut these had identical adsorption proper ties to the Goodrich foam 1115. The lower efficiency of XAD-4 is surprising in view of the similarity between XAD-2 and XAD-4~ they differ only In pore diameter: 90 A for XAD-2 and 60 A for XAD-4. The lower efficiency of PVC can be explained by the lower surface area of this ad sorbent. The surface area per unit weight Is reported as 500- 2000 m? g'1 for carbon; 750 m? g_l for XAD-4 and 330 m* g"1 for XAD-210 but only 2 x 10"* m1 g"' for PCV chips, there being no maeroretirular structure in PVC. This gives hii area ratio X AD-2/PVC of about 10*. There are two complicating condition* associated with adsorbing I'CB* from raw sewage rather than synthetic aqueous solution: Sewage contains other hydrophobic or ganic matter that competes for the active sites on the ad sorbent and much of the PCH has already edmutn-d mtl<> the suspended solids by the time the sewage rearhe* the treatment plant. The second condition can easily hr dem onstrated by filtering raw sewage and monitoring (He change in PCH concentration. With typical raw sewage containing 10 ppb PCH, vacuum fill rut ion through a Millipore prefilter pad resulted in the removal of sImiuI 75% It is therefore necessary to find an adsorlient that is not only relatively specific to PCHs, Iml that also has auflicienl at'- figure 1. Qas ohrometogroms of Aroclor 1254, Aroclor 1360. and raw sewage figure t. Adsorption ol Aroclor 1254 on PVC lignite oerbon, enttwo cite carbon, polyurethane foams, and Amberftte XAO-2 and XAD-4 Asterisk Indicates edsorotlcn 0< Aroclor 1242 rather than 1254 962 Environmental Science 6 Technology SONS 036672 Table 1. Adsorption of PCBi from Row 8owogo %PCB AtfMrtiMt adsorbed* LIpnHa carbon 4613 Potyurethana loam 3$ i 3 AmbarHIa XAD-2 23 A 2 AmbarHO XAD-4 00 * 3 PVC 73 A 4 * mchidot both Aroclor* 1254 and 1260 Data ara averaged ovar tavaral Ueternttnattoni to minimize (ha variation* in raw sewage. AOSOWiio* VCSWL t t rLOVACUn ^Nttwf *IV I t. AOSCWMCH vessel mw atwsor ncwitvon coti tenon vessel Figure 3. Apparatus used lor tank adsorption experiments < SMOIE TANK Figure 4. Percentage of Aroclor removed as function of time for tank tests finlty for thr i*CBs such that the PCB-suspcnded solid equilibrium is reversed. Table I show* the percentage of PCB (including both Amrhrr I2ft4 and 1260) adsorbed from raw sewage by five different media. With the exception of the PVC, approxi mately I R of each media was stirred with 200 ml of raw sewage for 4ft min; approximately 10 g of PVC were used because of the lower surface area. To minimize the incon sistency of raw aewage, the data have been averaged over apvcral determinations on different days and with different samples of sewage, The data indicate that PVC and XAD-4 arc more effective than carbon or polyurethane foams in terms of jwreentage of PCB removed from raw aewage. Thin is somewhat surprising since the graphs for PCB ad sorption for pure Aroclor solutions (Figure 2) would predict the opposite to be true. The likely reasons for this apparent anomaly arc; Thr active sites on carbon are preferentially occupied by hydrophobic s|M>cies in sewage other than PCBs and suspended solids adhere lo thr surface of curium and foam acting as a barrier to further adsorption. The above results indicate that PVC is superior to the other media for removing PCBs from sewage- To demon strate the feasibility of a sewage treatment method based on PVC adsorption, a small-scale flow-through unit was de signed, capable of treating about 3 I. of sewage per hour. The unit is shown in Figure 3. Each tank had a capacity of approximately 4.5 1. and was equipped with a paddle and baffle lo ensure thorough agitation of the chips. Seven hun dred grams of washed PVC were placed in each tank. The flow through the entire system was governed by tlie feed rale and this could be maintained at 50-60 ml/min (3 l./hl by adjustment of the needle valve on the inlet line Effluent samples from both tanks were collected for analysis every 30 min. The 50-60 ml/min flow rale was calculated on the Iwtsis of a 1-h retention time in each tank. Karlier batch equilib rium experiments with PVC and sewage had indicated that the adsorption equilibrium was established within the fir-t 15 min of contact If sufficient agitation was employed. An arbitrary factor of four was then selected as a margin in the transfer from batch to flow through condition-. Data from the flow through system are shown in graphi cal form In Figure 4. The efficiency nf thr single tank unit dropped from 75 to 47% during the first 8 h of o|>cration while that of the twin tanks dropped from 81 In 53"., over the same period. Although the second tank resulted in a 4-9% belter removal of the IH^Bs, the extra cost involved with the two'Stngc process would proltaldy not l>e justified in a large-scale application. Regeneration of the PVC was easily accomplished by successively flushing the < hi|*. with n-hexane, acetone, and water. The acetone and water rinses were to remove all traces nf hexane from the surface of the PVC. The PVC was then ready for reuse. In rottriunhm, PVC is very effective for adsorbing I'CH* from raw sewage. A small-scale unit has hern used todeni- onstrnlr the feasibility of the method but more engineering design research is required to optimize the technique. A system incorporating continual addition of fresh chip- and removal of spent adsorbent would probably yield more uni form results than those illustrated in Figure 4. Further re search into the regeneration of spent PVC and degradation of the recovered PCBs is currently under way in this labo ratory. The PVC adsorption process should be equally as effective for treating concentrated industrial effluent* as municipal sewage. Where a point source of I'CH-conlnining effluent can be located, there would be obvious economic advantages to locating it prior to discharge into the sewer system. Rpfcrcnte* tit l'eakail, I) B.. I.incer. -1.1... flower. 2tt, Me fl.|il97oi. (2) Itroadhurst. M Km iron. Health I'emp. 81 102. 1972. (3) Imerdeparlmeniat Task Force on PCBs, Puli. No COM 104)9, Washington, DC.. Mac 1972. (41 lawrence, J., unpublished data. (5) Hulr.inger. 0.. Safe. S., Zilko, V.. Kniirun. Henhh /Vr*p |S 20, April 1972. (G) Harjno. H.. ftchmedding, I). W., Freed, V. II , 5'm in,i S, i Techrwl. 8 (2). 139-42 (1974). (7) Gower, H. D., Chow, A.. Davis, K. C.. Uthc. -I F . Iteinke, I. Anal. I.ett-. 4,112). 883-9119711. (8) Musty, P- R.. Nickless,J ('hritmatog . M, IH5 90(197-1). (9) Chau, A S. V . "Analysis of Chlorinated Piwiii-idc* in Wnti-ph and Wastewaters". Kmironincni Canada Publication. pngi-r 40-43,1972. (10) "Amlrcr-hi-liics". No. 12:(, Rohm and Hass Co. Puhlimti.... . July 197). Rnvivvd for rrcien Seplrntlu-r 25, IV7.4 Attepi, it ll> > nilx r .'> 1975. M0NS 036673 Volume 10, Number 4. April 1976 313