Document V3Me69RwD0jNbnDmeNBKK355N

6'! W. 11. EDWARDS PROCESS ENGINEERING Philadelphia Plant July 22, 1965 {2JEKN& B0t H* ORCULATHW MMSANDUM PE-65-S:/ RC-230Q21 BP LATEX - SEMI-M3RKS MANUFACTURE This report covers procedures, observations and results of the BP latex production campaign carried out in the 100-gallon smi-works in Building $77. The campaign was a joint operation by Process Engineering and Research Engineering. The report was prepared by John S. Prendergast Experimental Operator who had the operating responsibility for Process Engineering in the campaign. This information has value as a consolidated record of operations for plant start-up and operating procedure preparation. JAB:IL Att. J. A. BURKE Fii4ic Ally :20s I96S PROCESS BMOiKHERiMS IQMgMffiUI REPOST pr o c es s mmMsmmi mm ?-n, m u:m sms-m nc-23002,. bp yem. __ During the period Mar. Up IS6S.c Z% 196.5 * Pisces 'Bitgzxm?i.mfl m ejmctic& with the fesliall Moratory 8mma & Sa w , ^ "' normal 90-gallon batches of SC-2SKSI2 (RC-B-24$.82) BP .Latex in' the 199-gsyllm S/S reactor in Buiidiag Nb, ??<, `Ms late was sta& fox tw reasons? a) To produce material for-the Marshall Pereli^ment 0r# Ssr use in their interior wall .paint field test program, - , b) To give Process Bigineering m opportmitf to familiarire themselves with the' equipment* process 'and to OTpare the late -produced with previous batches mads by the 'Research -Group* I, ISSdgPTlON QPJQIgffM A*. Reactor v.ys., I---.** The reactor .is & 169~gellon pressure vessel* rated fey. ISO psig md is constructed of. 516 stainless Steele It is supplied with a ..fell jacket, The inside diameter: of the reactor is .52 inches and has a straight wall length of 28-5/8 Inches, '.The agitate? is &>uHleded eacfc blade 'is M~!/4 inches long* 1/4 inch wide aid ,5=3/4 indies high, My are mounted perpendicular to-'tbe shaft and 'equidistant- from each other, ' Agitato? drive is va?iMe speed. The interior surface of the reactor was polished to a number 1 finish appnnsliwtely a year and a half .ag The i of 4b.# .reactor is supplied with a variant system* gtplml rent 'ttae, inert gas* nature vent line- containing. a r^ture disc rated for 13S 'psig .and various fLaoged fitting .fox the introduction'of materials, B: Reactor Jacket The reactor jacket is supplied with a circulating water system capable of .naintainirg-almost constant water -temperature* thTOi%tet -the system, Sot or cold water is introduced to the system through a cross \af the sitterof tJ jactet and disdiasies but the top and bottom., . Water leading -the top -section. is split s with-socae. being discharged at a.-regulated rate through a rotameter to the drain. M -remaining water joins the water leaving the bottom section -sod is pumped feudt to the" jadeet center, .. G, te?s*:r. Weigh Ttil. ,... .. > % wm .* -a. i.** ' ". -The' monomer weigh teak is m 100: gallon steritaless steel pressure vessel rated' for-125 psig mated on a portable seal. The dam of this -vessel is supplied with the fallowiagi1 2 1,: lightning iiiK? - electrically -sperated -{220j -'single threat bladed agitator, 2,' -Yaaiimi-line, 5,' Rupture rent line st^lied with - 100 ps% restore disc, 4o dw hsf:ssi$mm drs?gi$gs> % recifeisfistica ;/f wiumm* N 30695.01 DUP030002738 ~2- Hie center of the vessel is supplied with two inspection ports* located opposite each other for viewing aid fixing of the monomers .and inspection of the vessel, Hie bottom has .a single discharge line connected to a cmtrifegal pmpa Dp Surfactant. Weigh Tank Hi surfactant'weigh tank is an pea-head* SS-gallon* S/S drum exported by an angular stand whidi is rousted on a portable scale., A single .discharge Ite at the'bottom of tbs drum is connected to a centrifugal pimp. Bp.. Puaps lo Mancmer and Surfactant feed. Fiagis' Hie feed pisips used for both the wmcwer and surfactant were a -three-stag centrifugal pimp manufactured by Eastern Pptsp Company (Model 3J34B) * 2, Emulsification Pumps. TV .types of Eastern centrifugal pmps were used during the series for 'the emulsification of monomer and surfsetmt. So Single pass (one-stage), type fell* Eastern centrifugal, b. Three-pass (three-stage),'type 3J-34D* Eastern centrifugal. So Initiator feed. Pump - Milton itoy Fo Emulsion System Connected-to the discharge side of both the irawr ad surfactant centrife? gal pumps.are a rotameters for controlling the individual feeds of-these two systems. Hie discharge sides-of both--rotameters enter a common line which 'feed* 'to the .inlet ids' of the emulsification pump, 'Two types of emulsion pmps were used dining the series' * single stage and. three stage centrifugal pumps, Once .mdsified* the mmi&im is discharged through a badc-pressro relief value to a' dip tuba. under the IfeuM surface ini'the reactor,. fi,... Heat awhangsrs-1, Appnadmateiy' midway in the series* a small water-cooled heat exchanger was placed' in the monomer* surfactant feed line on ih: inlet side of the emulsific&tion pusife "Hie'exchanger was maufactured by the Young' Radiator Qmpmy end-is a type containing 31 tsAes*- s' long* 3/16" 3uB, - Hie mslsi was intfdd to the shell, side of the - beat, exchanger with cooling water passing through the tiabeii- If, Pressure- Scabs, .Small pressure bombs wire used fer the .introduction' of initiator* xasonlw hydradd and colloids to the.reactor, " DUP030002739 <* 3 - II0 MANUFACTURING PROCEDURES w'rgUS'c--t ;i nvi lyii'.^urmtiiiijnn..................... . A, Preparation and Loading, la Reactor The reactor is rinsed with deionized water before materials are .diarged,. Triton 770 and deionized water are charged to the reactor followed by th addition o Sodiisa Meta Bisulfite in a water solutitm. It was found that addition of Sodium Meta SisiM^TaTTtfir^Iti5M teg>erMure. before th addition of the Muinitm' Persulfate solution had no effect orfeither initiation of the reaction or'product properties. * 20 hfaiomer.Weigh Tank- fgn.-nrwi*-rfrr-- i. ,i ......... r ^e=qg,>e^cJt*T^-^.a?>^c . id ^ . The methyl methacrylate' monomer and methactylic acid are pisped into th weigh tank using the monomer food pimp* The pressure in the tank is reduced to 20" Mg . Absolute by-use of the mosm jets and the waetmn source shut off* Butadiene is thin.'charged to the weigh tank, while the rammer composite is being redtreulatesL This is dm to avoid,-line freeze-up due to th vaporization and expansion of the butadiene0 A 101 excess erer formula weight of those rammers is charged for tie purpose" of priming th o t o t k ? pi3p. ihe wsnoraer composite has been charged to the wnosier weigh tank in the following ways: a? toder-vacuum. b.n In a nitrogen atmosphere orp c. ' In' ^ nomal method described fdve0 The composite has been -used immediately after charging md also been alli ed to sit - for several ;days before being used. Mm of these affected initiation or latex -properties. . ' ' ..3 Ammoniai..Persulfate Charging (Initiation Charge) - The Awmiisfi Persulfate is made up into a water'solution prior.to being used and. is charged by suction. to its evacuated, bomb. ' During th course'of th eis^aigaB it. was found that this solution:.was being mtarainated with copper fro a. copper line connected to a hood aspirator used for the eradiation. Th Awamim Bsrsulfate . .solution' -cleavage left in- the. bomb was sucked Into this line fowsirtg capper sulfate. Since foe removal and replacement of this line presented a problem^, -foe diarg line andiwnbs were Hushed with deionized water after its. use. .4o Surfactant Feed ' M excess of .101 over the formula weight of Triton 770 and deiteired water is charged to foe .surfactant drum and stirred with a S/S paddle. Single md sulti' charges of surfactant' hare been 'made which, had no effect on initiation or product properties. ' So Pipp. Priming Prior to priming the monomer foed'presv 3S psig Hg pressure is applied to the.roswer weigh tank for maintaining a positive, neau. in mmomers are saiwd and circulated for -5. minutes with th .lightning mime?. This-miking cycle km bmm DUP030002740 r- 4 w .varied from 1-30 minutes with no changes in initiation or product properties, l!ie monomer pimp is first 'primed, with the emulsion pump operating in order to adjust the badc-pressure relief valve to stop the vaporization of the butadiene which would affect the rotameter control- of this, feed, When this has been com pleted, the surfactant-feed is started and both rotameters adjusted to'their proper settings.., This material' is taken off into a container which is vented to the atmosphere and the emulsion being formed visually inspected, 6,, Ammonium Persulfate (cOTtanuous foed) Charge ' . - Double the formula amount of Pmonim Persulfate and deionised water is .mixed prior to use .mid the same method being, employed as used in charging the initial Persulfate .solution, Nitrogen Purging and Evacuation . #ih the-preliminary, operations completed, the reactor is purged with nitrogen for 10 minutes with the agitator off. However, the' agitator has been run during this period to insure good-mixing of the Bisulfite, surfactant and water, Some ..foam-is created but this was.overcome by shutting off the agitator, closing the vent and allowing the pressure to rise to. 3-S psig before the evacuation ofthe'reactor. the reactor, is evacuated to 20-25..in Hg aid then'broken with % to atmospheric pressure. The reactor is again evacuated to 2S-28. in, Hg and held. Foam is created during the evacuation step but the level is low, ,C, jjaat-up.Initiation and Ctotinuous, Feed The agitator., is'turned on and the speed adjusted. In this program, the agitation has'.been , varied-from 80 rpa to llS .rpm with no product dissimilarity or- ' control problems being evident. The reactor is heated with 90C, water.until the batch', temperature reaches 7&C, An additional rise in temperature of 2-3*C, is experienced during initial monomer feed. At this temperature, the Sodium Mete Bisuliite solution was added cm several occasions without any change in .initiation or final product properties. The reactor pressure increased from 2S288J Hg to t4~ir Mg', Six percent-'of the total formula weight-of monomer tod ssirfactant Is fedio the reactor Containing Sodiiai Meta Bisulfite,' Triton 770 and deionised water over a ID-minute period,' Foam immediately appears and the level initially rises, .several feet but begins" to decrease to .6'* to 8s9 above to .liquid surface'as the pressure: increases. In the manufacture of these batches, this foed was also Introduced-ftt a lower temperature, 76C,, requiting "an adjustment ..in batch, 'teapesature'before adding the initiators This has no effect to filial predict.. At the. completion of toe' 61 emulsion feed period, the reactor tempereture reduced to ?6e, rwi'.78C, due to'the mtroduction of cool mdsim and the pressure 'increased to 3035 pslg, '.0, 'Initiation . The initial to&rge of Potassium Peretslfate solution is added to initiate the reaction with no" adjustments in batch tmperature being required. The Imtcfi temperature begins to rise-'with -a decrease in pressure .at-7Pa, s the continuous feed Of emulsion and: initiator'..is'begun and fooling water is intreduced to the jacket circulating system at a. regulated: rate to maintain' a desired batch tewpetoture, -Batches haw been msipfectuted .at lewis over the. temperature range 79-83C, with DUP030002741 T* S " no effect on latex properties* Normal continuous feed periods were ,M0 minutes.. 1m batches wests manufactured using a 135 minute continuous feed period, 'both had high pressures 5Q6Q psig vs* the more normal 25-35 psigo On, batch was manu factured using a continuous 17Sminute feed with no waiting period between initia tion and continuous feeds Hie initial initiator system was added 8 minutes after the start of the emulsion surfactant feed- and the continuous initiator feed started t minutes later. No problems were encountered .with operational control or in the properties of this latex* All batches in the series initiated but several batches experienced sluggish initiation due to. low- temperature initiation* E* Oontinteus Emulsion Initiator:feed (160 mins.) ar^%i^^tfaca<rffflrfugi)rjrig,Tr-TTrTtf.sa^ *w High and low.pressures during'.the first 90 minutes of continuous feed were experienced throughout the series* High pressure is attributed to the initial initiation, temperature which determined the. reaction rate. The average pressure during this period was 25 psig but begins to increase in the remaining 70 minutes 'to 35=40 psig- (Av*), The;temperature .differential between the batch and jacket water temperature is approximately 11C* for the first 120 minutes of continuous feed', mad.-increases to 20*(* In .the remaining'40 minutes.. .'NO .-trouble was ever-en countered in maintaining the desired batch temperature. ..However,. during this period several problems did-, develop* . . U Trouble on several occasions, was encountered with a foreign matter holding the diaphragm of the.tack-pressure -relief .Valve-in the emulsion feed line open,..causing the emulsion to' break and loss"df cmtTOi in the feed system:. .In the m^SSBfectere of one batch, toe-emulsion feed was stuped, the. backpressure relief valve removed fron the line and feed resumed, the oomtiaums. initiator feed was not in'tem?ted, ,. "- '2* In another instance, the nitrogen pressure on the continuous initiator feed .bomb exceeded .the .pressure .In the reactor, thereby causingthe -badc-pressure relief valve ,to open, discharging approximately 471- of the total continuous initiator feed'-in a. 10-mihute-Interval* 3c Erratic'.feed .control-Was also experienced due to high reactor and feed tank pressures and blocked.emulsion dip -tube * However, '.there.'-was no definite change, in latex properties dte to these difficulties. These, problems wefe-eliminated by til-inspecting and cleaning f the dip. tube and back-pressure relief valves after every third, balds* 'The nitrogenpressure on the 'cmtinuoto -initiator feed- bomb was reduced to 10 psig throughout the. rsmain^T-'of the-series*.. During the manufacture of on batch, high pressure, sluggish, rate @f ,. . reaction, low temperature differential and loss of wammr md Surfactant feed teir troi. were experienced* After the' templetitp of 90 minutes, of .continuous feed, . hydroqinnon was added, inhibiting the:reaction* .-The batch was cooled and .discarded* The only- TOriation:xmtributing to the cause,'Was that this was the thirteenth mto factored after the reactor whs cleaned and the. reactor eoagulum may have contained .some toreij^matteriiMbiting reaction* F* Initiator ^yiek1VMditim: ' M the .awpletion of the i60-minuts. continuous -feed, tee water the jacket circulation, system is immediately turned off md a small quantity of Per sulfate solution is pressure-fed to the "reactor -to .complete the poiymsjeisatiMic The formula hold period after this addition was extended. fjtm S' minutes to 15 minutes or until the pressure decreases to 28 psig whichever- is Imp* 'Ms h@M period'has been varied fro S to 45 minutes at- various pressures with no change in latex properties- DUP030002742 .6 C. Anaaoniup Hydroxide - Stabilization Step The amonluia hydroxide has been added in several ways with m change in the final product pH, ' i, Added to the reactor on and under the liquid surface, 2, Added before and after venting. Adding the aasnonroa hydroxide under the surface before or after venting caused foam midway in the addition. Addition made before venting caused an extreme foaming condition due to vaporization during the venting procedure whether charged below or on .top of the'surface. - It was found that addition on the surface after venting during the cooldown period was the most practical. H, Venting The time cycle required to vent the reactor varied rm 2^45 minutes depending oh the anoint of .foam created during depressurization. It was learned that foam created in .this '.cycle can be'eliminated by the addition .of .part of. the colloids charge. At. atmospheric pressure,, the circulating system is set for full cooling md NHgCH addition is begun. I, Colloids Charge The remainder of the colloids is. charged on the liquid surface after the addition jof-t&rlBivCH solution by means of a . pressure both during the eoolds cycle. J, Post Stabilization A-post stabilization solution of triton X-100, Etylene Glycol and deionized water is added oh. the liquid surface over a 30-minut period while maintaining a batch temperature' of 35*C. At the completion of'this addition, the batch is agitated - for an additional 15; minutes and then, filtered. \ K, Sables Sables of the latex were' taken after the addition of the colloids and during the filtration operation. Evaluation' of the latex was, perfosd by the Marshall Laboratory IfeveXopaent Group. " h. Hitmtion 'The latex is gravity fed to a Yorti-Siv containing' a- lOO-mesh screen. The'time required, for the filtration of a drum varied from 5 to 12 Mantes depending on the viscosity' of the latex* Foam created by the aeration of the. latex Is the Yorti-Siv is almost eliminated by tbs insertion of a dip-tub to the bottom of the dross from the outlet side of the Vorti-Siv. . No prsblffl with filtration mm m- cooatered throughout the iiamisjEactwsing series, M, ComeuiM II III..................Ill II r IIHI.~i inn mi The eoagulw remaining after filtration and any tmccouatabj. losses have been equal to or less than the 11 formsla allowance, Th^cbagtftuai obtained throughout the series was fibrous ia structure. . "X DUP030002743 7 N. Equipment Cleaning After Kltr tion 1. Reactor - Use reactor has been cleaned between batches satisfactorily by using water and high-pressure rotating jets located on the center line, at the front and rear of the reactor or by the use of a high-pressure hose line. This operation is perforated iimsediately after the completion of the filtration. 2. Vbrti-Siv and Screen - the Vorti-Siv and Screen are hwediately cleaned-using city water at the cospletien of the filtration. Time required for the normal between-bateh cleaning was about one hour. ' O. Reactor and Agitator Condition , 1. Reactor - Twelve batches were the most mmufactured between reactorsolvent cleanings. 1h@ thirteenth was stopped 90 minutes after initiation because of slow and erratic reaction. Us entire. Interior surface of the. reactor begins to show a transparent coating after the manufacture of the first three batches. This transparency is still present on the dished bottom and extends up the straight wall :- IS indies after the- sixth batch. The remaining straight wall area. ' and dome - section become coated with varying amounts of co&gulssa .. - .buildup depending on the foam level of the previously juniac&irsd batches. .After the tenth batch: the dished bottom is coated with - a layer of coagulm which is no .longer transparent, the.straight' .'wali area extending 18 indies inward-from the dished bottle still . retains- its transparent coating, the remaining straight wall .aw. is heavily coated with coagultaa, in w areas this soagulw was- beginning to pull , exposing, the bare wall. This condition was more pronounced at the end of the thirteenth batch.. ' : Agitator Shaft - The agitator shaft after' the third batch has a . transparent- coating extending from the agitator coaling -downward to the-blades. This coating gradually built up to a..thickness of two -inches at the coupling, tapering-to die indi at the bottom at the eapleiion of. the series. 3. Blades - The- agitator blades throughout the. series did not show any apparent coagulwa buildup. Gaagulua was present m the trailing edges of the blade but was removed with a stream of hi^-pressme water. P.. Reactor Cleaning A 5/1 toiuem/isoptopanol solvent system developed by Researchhas worked very well for reactor cleaning. The reactor is filled to edacity,with the solution and heated to 8$rC with maximum agitation, and held at this temperature for four'hours.. It is then copied and .removed. ' The reactor is flushed with city water -and that' rinsed with deionised water. The interior surfaces of the reactor are perfectly clean except for small gmomts of coagulu around the seam section. This is removed with ease by. screwing. DUP030002744 8 III. SUBWKY , One IQO-gallon and few 90-gallen batches o BP Latex were made initially as part of the 34-hatch cooperative effort with Research. These batches did net produce specification latex since the viscosity varied from 670 cp to 4,820 cp without apparent relation to aeasuronent of other properties. Saaples for latex valuation by the Marshall Laboratory Development Group (T. G. Brown) were tabs before tbs addition of the post stabiliser and during the filtration step (after post stabilisation}. The results of .these.analyses are contained at the end of this .report'. It should be noted that the viscosities obtained by the Research Gswp in their progrm fer establishing an operating latex formula (RC-B-24852) differed widely from those obtained by the-Process Bagineering Grc^. A mber of. changes were Mete in the process to return to the process- actually used by feseard*. These changes .lowered, the reaction pressure but the variability of the latex viscosity still existed. Since the p@rfemnce.of the latex in the paint being evaluated was satisfactory, this formula was retained. In the remaining batches .manufactured in the campaign, no specific correlation of viscosity change with- another measured property could be noted within the formula. Raw tutorials were individually changed,'but had no effect on the viscosity. The introduction of the Sodium Meta Bisulfite at the initiation temperature or the addition of ferrous Ai Sulfate in the initial' reactor charge had no effect on the final latex viscosity. - It seem likely that viscosity, variation is caused by some unknown variable. . Hie principal effect introduced by this viscosity variability is on handling and not paint quality. Venting before and after sswmia and/or defesmr addition, hot .and cold stabilisation of-the latex'and lower polarisation were tried during the assign. One batch was' destroyed due to poor initiation and high pressure earned by m unknown inhibitatiem of the reaction. - The latex manufactured in this program except 'for- the two batches discarded was blended into two 1,400-gallm latex batches fer use by Sales; feveiopasmt in Conssmer Panel tests. XV, aMSJt^I@S 1. Variability in .reactivity remains a problem; slew reactivity causes . M#' reactor pressure and'seem to promote foaaing problem whm wroacted mmmv is vented at the end of the reaction. 2. Reactivity seem to be affected by feed sates, timing, quality of the monomer emulsion, temperature, or rm materials. 3. The present BP formula needs changes in operating procedures and product specification. 4. The reactor should be solvent cleaned before thick eoaguta layers build up. So Blending latex of varying Viscosities may dampen the batch*to-batdh effect particularly since seme aging will be entailed in filling the blend tank. PREPARED BY: a > -JJ yL A, t/Usyu REVIEWED BY: APPROVED BY: /IL DUP030002745 Sp|mi- U i I a <t co <t <a a 03 sc 0 ^^^c^tfoaeQttooiscaj^cQtcflQaiOa - - . fi 6 t^ *o 0* 0* 0 6 N NSl O <5 0 6o oo <$ 6 oo 6 <3 AA N ?ivi s 5O n$s s^ k ff si 5 6 Isf N 5% 1*> k "J k * * 0 4<} oo oH 6 oo g O Hj k ) p I s. '-is ~- o rl V> ^ sYl o .y*1, o Hr O VJ 0 O ,K o' ^ S < % Is Vs ' N ^ OJ Vs 5 -S. o.?r o < 0* 6 v tso -- o m a- ? n o o M. rs \cMo frj os 0s r~C -- -- -> s%lO<< a, lB -s'*>l<>fK v S N O s .m OO Q <D O O O -* O *': * ' ' OO00 Oo "*"1*' <3 ^ O 0 ^ o njM \| r\j lo-w? AJ (\t l\l *S S. s. vs N S S *s S S S S "Ss s, Ss s*. o'<5 6 o' o' 6 o <5 o' ' NJ> n in is, s S rs q ^n s j w "' ... CDtsiy* <VtV)nsW)NV'>0>fsloV,Jo n co ro fs r* V> V) <t ** sS vj ^ *> to {Oivjfvj isiN o) IVjOu 'Vi^ < ^""ssC'is < n; < S! s.-J V s`<<<N'C< N. o rt * > vs 03 co wj g. 1 t o ^ <js Ns.ru 1 ^ Vo nj* V) <4. Jn H* *0 0 <M >v0 v*U> s\>S S\>3 sn3j s>sj V'>j Vwoi s>sj s3g |1 \\Sj s\SSONSs\OQ 'OHJ j sS s|> \Ji sfli \j> sjj sc nQ s0 \0 sS 00 00 <50 03 <0 03 TO CO Oo <30 4 Wffy5 <$0 00 00$ \ 6i OO 06 03 <6 05 aS <6 06 06 0$ W> V1 to 0 Cfry-O^ o <0 m H) V| frO K3 <$ Nl <M to No fo ^ K1 to Ny 0s 0^ O' 0* 0" Hf ' O'. 0* O' 6* O' O' O' 0? 04 0> 0"' gi <Js Os 0s` <s <S Oi O' 0= <. o V 00 > ts K ss n ^ -s. Cj > t*) <6 fs<V|M|'0 fs. s) f'ssOjs.Sj.vO 0 < 0 w>s9 o Vj Vo < V) O O' S3 'O O 0*0 s& N so CO O' O VO O' C0 V) *s. -s <J <J0s* ^ ^ ^ << \j > . ' co O'* -s. s. fq sj' rofsfOVo vosj-nr*5" ts,. rr<wi <^3 N><T tr1n<' CvtaMs $>ap U^ng ov ms rh l^NMT"ft*I<sjsl^M;V*o)j O'X 'Ni 50Ns3j> s jrss ^fs f^s O^ <, s: sTsSs' fs ?) ^ > Ssf., K^ 5k* Pj o vS k . VvXo s^3 Ns 0 ^ 05 k K cd rs K r* n in A) to <0 <o r? n to s| V) sS <0 ty> . nt- ? Co i3 sS n >0 10 -sS 0* S3 rf <V> IS sS vJ * si T rn co. sj 4 si tr CO Vo rs 1 <r H- O' sS sS sS St CO Ch 1 nr <a ts sS "O . HM 1 nf Si*' Fs sS <3N3 omr*-* <3r ri r* sR *4f* V *SS <> V p r~ s> Sj sj sS' .S*j s3 nr sr fro Ifo m S? O' Ol *> nr Yf* sj- sj> ^2 rs s3 s3 o Sss-S3 S|- VY> fro N3 to I. *n VO DUP030002746 tit 1*1p.4 8$ > in* K 8I ;?i* z c s a gas i\ s % 8II 3e * 5 :t V ; 1 % I *_l_ 1 ? ? ?. $ 5 5<4 4 Mi *I *li R".. at St {$2s < N tc Xo CD fNt*' in *8 Q0 Nft> s o fi 1 rto* O' ft* ft to SS |Vs|Xf : Of >4 ^ 5*5 *c ft 2! VI 9 *} ft to S 2? ^ <5 r ? 5 5?V * XX '* m~4 a. * Vft) 5e *?5f Q fl M0 ft kft 5. KQ * 6t * . <3 . $* Cft U4 3 * ft 0 ft ^ <5 53 3 3 5 O d o O 6 <? $ .6*o QI a w o -- P3 o,j k > cu /tj w 2 ? ~ A f- * tS 5 to -- 5? -- 5;. a <Q0 O I isn' wV <V<Sn VJoj 'wi' !*!* J^sS J*Vft <Vvj V<vi s<ov sw* t5oJ V^) to &N < < < < < ^** "** ^ Qc MN%i - ft ftv ft o> ft .15 ft ft ft ft : ft ft d I IlfN<10 l*^"* 03 o ft H ri ft q; 4 a ^8 <9 <1- Q JS || Ss y v> \ $^ . V| it <t iHU K O rt ^ Hi ft <0 $ K K K ft ft o ^Vo I ft ft ft ft o ft ft ^ ft ft O ^ 5 ^ 5..5?-5 a--= S 2 odd o .N3 <5 d d o o d 6 d N *0 ft- o y> ft* ft* Jh irt a ft ^ N) }fi ^<<<< SSUS s -j it K t' tr ^ V'J`'o s. Cv JA Vj vj* tj. V<j M is n# \# M) vs >5 > 4 4- 4 4 4 4 4 4 vj oo q j d>o6 Oo b 09 00 BO 0a G4 t)i Vj OS t* (b o5 % o d4 ftDd^>0i> feUj vi ^ Si ^ <s II t a ft V) to N *&M ft K > 2vt 0 *4 (\> *N> or <t| U a:. <o ifC Vft <5 tl g N> N) 4- M ^ <V| Ki M ^ ^oj's. ^ <$;tS <> (Stji^ovQcc1 > .c <t] ^-If Sis $ Sl V> (S frt <D rk \#^s INS` "cwo s. \# * *tVK'"I*s*'W') K' '1.s " <M ft *Nftl IriM s^hilft N)ftjrnMY O5 tKO*- Osftifti Qsfti Q^ftiCS ftsftJ Qi !t$^SS%!?|5%jS >i8!3^S:{J!25^^i55!o S ? N . *fl ^ ftZ c CQ K ft\ t^ OI-NQs OOOVfsv .4(-<BN.* ;'?!aS!S3teS'5|fifS! ' ft od > ft* N > 4s q N 4 4 > H| KS Hi M a 'io 0 8O?j s s sj N >0 *% tfi *o tN ^ 44 9 0-J0 > NJOO Jn 5 44 I5ftos to ,fts ^ V\ $s Qs ^ fts tjs CN Os K) NS K> H> N> fOr fO NS SD ro s !N *s 4 fts Nk 4 NT 49 N? NV jy> ft St Al Ol 4 so I** NO 4 NO $ NS Ni y4 CNJ cn <4 0 sD s9 DUP030002747 ^ &* * /r*> v ' ' fo ' os