Document bao2zw0QKRbmj5r2aKKmakGjD

> t v, a*-. ; jk:; re;v Batch operation of the reactor is the dominant characteristic of almost all commercial processes for making many different varieties of polyvinyl chloride because of the great ease and flexibility in changing recipes from one product to another. f ..MODERN CHEMICAL TECHfJOLOGf PART ID .-.a v . ( : *v r; c*. v* l.C; *^ : v r I * nji V TO! T-*! "? aL.-?' L*L J*. an rl? v>v *-* J* s U-a. v_/ rT* hy {; T n ** *: O, VT* -- 'oeess1. \o LYLE F. ALBRIGHT, Purdue University Emulsion Cor dispersion) polymerization is the oldest technique that is used on a large scale for the product:::, of polyvinyl chloride. It was the main pro'-c-ss uii it about ]9o0 v hen suspension methods were developed. iteu.-ntiy some of the products produced by emulsion processes have found expanded uses.31-2 Polyvinyl chloride- (PVC) resins produced hy emul sion polymerization general].'- retain all or most of the emulsifier added to the reaction system. About 2 to Cc of the emulsifier is generally present in the poly mers. Suspension polymers, however, contain only a relatively small amount of surfactant or stabilizing agent. Jn emulsion polymers, the emulsifier is concentrated at the Mirfacc of the granular particles that are bead like ur.d icla:ivcly smooth. When plasticizer is mixed with these particles at ambient conditions, it wets then', but the penetration of plasticizer in the PVC beads is slow. Such mixtures of PVC resin stud plas ticizer form a relatively tluid mixture, assuming that a rather large amount of plasticizer is present in the final matei ial. These arc called plaslisols. When an organic solvent is also included in the formulation, the litpiid mixture is termed an organosol Plasfiso's ami organosols are used in dip coatings, sb.,>h n'oUI-ne-\ rotational moldings and foam applica tions In the liaa! steps of the fabrication operation, tlie tcniper.it ire is lai.scd to 1G0 0. for a shout period of time. The heat causes tote! or partial solvation and unifeini ni:\i.ig of the PVC ami pL.stiir/a r. Upon ion]::.i" tin.-' )r.i\':n-c.s. the plasticizer iray be laOier uniCimly m,v..l throughout the plastic. In other va-cs, the Pua! pic -tie has a l'YCcni.timious pilose, and a plas'.iei. e.-ihsccntlmioiis phase. Such a plastic has rl t jj (t Hi*' :*Hi**i\ m c <'). i h /'<!., 1 t*. 1 3, ) !H*7, p. * t<4. O.ruicn) --July 3, 196T different physical and chemical properties as compared to that containing only a single phase. Factors Affecting Polymerization Many recipes are used commercially in emulsion polymerization. The following materials are always prc-eiit, and the approximate range9-13.ir- 31-3- at .which each is added are: Vinyl chloiide .................... Water .................................... Soap or srrfactarl ............ Water-soluble mi:i..tor .... 100 parts by weirht 150 to 250 parts by Weight 2.0 to 5.0 parts by wucht 0.1 to 0.4 parts by Wtig.it For an outline of the mechanism of emulsion poly merization. see the summary on the next page. Actdition.il details were given in an earlier article.1 In the emulsion polymerization process, only highly purified vinyl chloride .should be used. In addition, deionized and deaerated water is required. A wide variety of soaps or surfactants arc used in cluding sodium lauiyl sulfate, other stilfatcd c-.ster?. ammonium stearate, sulfonated castor oil, sulfonuted higher alkyl alcohols, alkyl naphthalene sulfonic acid--, oilier type? of synthetic detergents, and just plain soaps (salts of latiric. myristir. palmitic, and stearic ncieO. Jr. addition, nonionic emulsifiers are sometimes used, especially to make polymers for electrical ap plications. Hop.T and Fakla1'' investigated the relationship between the chemical nature of the emulsifiers ard rule of po'vinoi i/i'l ion, and the propel ties of the i esuiiiug l'VC. Sodium .-alts of the siilfuiie acid e.-S'i-of fatty alcohols, iiavine 10 to IS laibnn atoms, g.i\e l'Vd polyim-rs nitii the highest molecular woi--! ts. I'ar.d'in sulfonates, sulfates, and salts ef fatly a.c'ds CTL030843 rs ce pcrpisH-a 1 had s.'n'.i r. rrh.livc 1n puiy -'wr;. Vn- i! i )- .!(>...- lit* i.f r.i.'v-.n.Ti- 7:ti'hi. t mul-il'icrs were .v'lve `ii.ui the anionic vr.rictt. In ivner.il. the effect r>f emulsifiers on the y.r.z.a! i i;i reaction rca-cs a-. the molecular weight of the emulsifier increases I'VC re-ins f..r electrical applications such as wire (noting- r: : ] r,-1-:tro.' he a-big wi.'.r:1.' r:ni'!- s if: its or iIccenV'aMe emui-in- r~. Tne l.rlcr ,a't quite stable ciun'.ig polymerization. but they can be saponified later.--' Kc'IdiiCs ol' the emulsifier arc removed by washing the granular PVC product. Some emuisifi<rs in i. :y u.iude: If .\i.l:v. - .1; :c .. ud- fattv acid e-tors, sulfonic acid esters, and phosphonic acid fst-.'i". Initiators Are Wafer Soluble Some of the nacre important water-soluble initiators are: f.-nmr-r.iim persulfate, potassium persulfate, hy drogen pc; \;d.:a an;l varint:.- redox sy-toms such as chlorat-.-ljis.L'ite c .m.binatioi:.-. The initiators are soluble to -;,mo extent in the vinyl chloride phase. As ?. result, a small amount of an essentially suspensiontype polymerization may occur in that phase, l'eggion and others-' indicate that the polymerization rate is proportioned to the O.C to 0.7 power of the initiator concentrathn; whereas Smith and Ewart" predict that it should be proportion..! to the 0.4 power. Al- 'har.'li sever t1 i"\<- t; rs ;i ''h pre-cut vnluahle i :f ::i i i -r.i. i r,ir.>: 11i i!-<in polymeriza tion, inucli additional information i- needed. Relatively little quant it.e mfni mation is avail able in the h'oi.itare re/tardin/ temperature. Appar ently temperatures used commercially vary from about .40 to r.j C. II ich.er temperatures are not generally ti.-eil (a-. i:i th- -u-per.-m:, pruce-s) beenu-e the PVC would li.it e a te'.itr. e.y lov. n'olccul.ir ueignt and con-e- qnenlly would be relatively thermally unstable. Tem perature would affect not only the rate of decomposi tion of the initiator and other chemical reactions, but a!-'> the no "f the i;r.:.d-. m i trail-for of reac tant to the prowins polymer paiticic. Batch Processes [or Emulsion Polymerization In the I'.tv. batch pro-.-e---e-' are used prod 'mman'l.c if not exrlusi', ly a., comp., red to cnntinuo.i.--flow proc esses that are or hate been u-ed in Europe. Batch processes hate two imp irtant advantages: 1. Increased er.-e of producing a large number of distinct types of PVC rcsjns without producin': sig nificant quantities of off-specification product. In a continuous-flott process, there is always the possibility of producing relatively poor quality materials espe cially during start-tip or while switching from cr.e product to another. 2. Better control of operating variables during r"- INITIATION 1. Reaction of initiator in aqueous phase: Init.a'.or ---------- >- Radicals (l) 2. Transfei of roGical to r.:icel'e. 3. Ir.itiat.on in micelle: I 4- CH2 -- CHCI ------------>- l-CH2-CHCI Methor.lsm of emulfijn ;ror/merirc:-!on fins su-'-'.vy ''or emulsion po'>rieriration does not tc^e i-': s ts.'t t proooscl m ' c.u a'd .. ,i.rs-'1 v/l.o state t - 1 some poymer./.-.ycn occurs ir. ire water pliase in wh ch vinyl cliorid: is slightly sclut-le. I--(CHj--Ch'CI)n--CHj--CHCI + CMj-CHCI---------v- I--(CHj--CHCI)---5--CHj--CHCI 1 Part-cie containing CHj-CHCI *CH2kCHCI monomer plus polymer-----.. (^Jl0Vi ryi chloride) (CH, =CHC!) droplet --------------------- > k* CHj = CHCI CHj-CHCI (CH: --CHCl)cfrop!ct " -------- (CHj-CHCI) jqueous (CM;--Cf iCl)iiquc*ous ~ ^ (CM j -- CHCI) particle P'OpIrlH ^r.hdtiA'I s (*i'..'; '.> Bolh\ir,l c!!or*do and f . * M tc i' ;1 . C '-.:a n.:. j s .n, i , . *\_r__. ._ >\ ' C> \ * , i.r f y,.'*' ^ * , ,, ` ... TLR.V.IMAT10N AND REINITIATION Termination Reinitiation <S> CTL030844 I polymer:/.r.a.n ;i: I. nu.ie, belter (nntrol <'f (:i<11ly in the final polymer. Iidheulty is probably still rxpen- cnctd ;n <.ii:;r.i.t);e hichc-t <|ii111y PYC rc-ms in Lo:itin-(jU>-:'.(/.. prcie.-scs. Mn.h of i;,c ci; i.;.n.o:11 n.~fc! for emulsion polymeriza tion is the -sm" that ti-ed fur i^;mm>: -?i pop. : z.itici. In u c i- .i i -:.i-io pi.oil. the on'.;, major <. j...| meat ihuererii-c seems to be m tin* drying operation.-1 Since m'lch of tie litmrni is i.-oil interchanpcably, the operating `lops tire also very similar. I):.. . ! m'd.'w ' '"o autoclu'e are no: Uru'wi. Pi i-sumably there are several C"rn;r.-..\ .il v;.:.-. I'Ir~ 1. the cm i'-1vr. in:;!.it if. an'I j' - -! 1.1 y a. b..f!\r n>-o dissnkorl in < i i :i i .'f. i ..`er. Heiativiiy cm.! v.,ter i- normally used because the initial r begins : decompose a- -<as the Irw'iern- taie i' raised. T:.e water ion is added to tl.e autoclave, ar.d care is taken to exclude air for oxygen} from ;.-e sy-lem Tt.o dr.-trod .iir.i i.r.t of vinyl ch'or.do is added, and ayit.it'on is started to form a relatively stable emulsion. Wuen tiie emtilsion is heated, the reaction begins. Tcmpc-rntu:e o r.trol is often eriticti! during ti e most rapid phase of po'ymerization. Adequate temperature control is maintained by cold brine or water in the jacket. Eu:.mJ . ..rlltms s..;ne of the method for increas ing the over-;::; l.e.-.t-lransfer coefficients: a. Introducing the brine through nozzles to produce tangential, hieb.v it;e swirling bring or water. b. Baffled de-im for the jacket. Proper installation is relatively o-pi'.s;re, ar.d this method is apparently not too widely used. c. Rapid convection or pumping of brine through the jacket. When po-silde. cold \inyl (blonde is added to the poly met i z n t Ion ve-se! iu order to moderate the tem perature. Baum* indicates that a reflux condenser is sometimes used in some tyres of emulsion polymeri zation. As indie..ted earlier,-' such a technique is prob ably not applicable for producing PYC resins. Often BO to f'5r. of the vinyl chloride is polymer ized in each batch tun. The rates of polymerization decrease' rapidly at higher con\ersions, and reaction time is 12 to IS hr. Pressure decreases as the run nears completion. Recovery of PVC Particles Sepaiation of unrcnctcd vinyl chloride from the omulsion and its subsequent recovery, purification, and recycling involves the same1 equipment as used in the si'-pou-ion procc-s. The remaining emulsion or latex is stored in blend tanks until several batches are collected. The most popular method for recovering solid PYC particle's is by spray diving because it is rel.itixely diflicult to separate the emulsion by centri fuging or I'd'.'i'iu'' Tlvsavy1" has indicated some of the facloi- of in.poi t.ibcc i. ; 11 A .............. di". n *. and I .:ibilli '' i!c-c i ;be . b.- -; . 1 ! . i . i ' . Ib'iim dry ci< h.iv.' al-.i been u-ed. rsniiict imes. tiie cmul-l-i.'.c ' . i - 11.,;. '..! lo i'. 111. or ee:t- i.. , .i. 1 i .o i.: 11 11 i' \ i jua i! is eo!b ctvd, ci.i.ssi :.c d, ami stored mittl p.us.igiJ m silos paving captsc iti-< of 100,<'00 to 200,000 lb. Modified Emulsion Polymerization In a cot", ftf ton :! emulsion polytrci i-ation. initia tion ...u;'. ,n l i. i n.neho-; but in some n.od.bcd methods, few if aiiv nucelies are present, llatlter the conventional initiation of polvmer particles is by passed. Instead, the reactor is seeded with tiny poly- tU' r t IV n . cc-'il- in t he c o n- veniicinal n.,;r,!'et, n.tnieiy polymci iz.ition occurs r.n or iti the- latex -if s-cel i particles. S`\ era! eu dllie.it ions of sc-cdine ate prol ;.Vy used, but th.e patented process liy Powers-'' is .in in.pi riant method. Powers cxp'.sms !'. :a lat :r n-hit' between PYC par- tick- s;/.e at..: the .ut of (b.-pe'- eg agent i su> h as soap) necessary to maintain a slab'.j emulsion, that is the prevent:, it of i".i!e-ci`nve of the PYC particle.-, or formation, of additional dt-persod particles in the emulsion. The area A covered to a thickness of one molecular layer by a mole of dispersing agent is: A " .v, .1. Il) where .Y,, is Avogndro's number, and A., is the crosssectional area for a .molecule of the dispersing agent. Such values are often. a\ailnh!e in handbooks. Since the individual TVC particles are essentially spherical in shape, the weight IP for the dispersing agent of mc.eculu: v.etgnt .if required to cover the surface of a panicle ol di..meter d L: 11' = x<n M/X. .1, (2) The weight IP,. ;: polymer with density p in each particle is as follows: H j * rtf' p 'G (3) Then tiie weigh: percentage P of the dispersing agent (based on the weight of the polymer) that is needed to c..ver the surface of a particle is obtained by dividing Eq. (2) by Eq. to), and then multiplying by 100: /' \r \i k x ion non u A .. A o p (-0 Of course, the particles in an emulsion have a range of diameters. The ".irfa.ee area is proportional to the diameter squared, '"herons the volume is proportional to the diameter cubed. In general, the diameters of the particles present in an emulsion follow a normal probability distribution. Corrections must then be made lo Eq. (2). < 3 > and < l'i when P is calculated. A semithceietica.! correction technique has been pre sented by 1'i.v ci's.-"' He reports that when the weight percentage P of 2is to u<V 1 tpreferably 25 to -10Y ) is used, sigmficai.'. conic 'conic and formation of par ticles do not oieur. Tiie actual fi.nlion of P that is required is un doubtedly to some exten* a filiation of toniporutiiic. atro i::;.- of reactants and other additives, and degree of . ilio:i liv implication. Powers apparently be- lic\c- liiai 25 to o1'.' of /' i- a e.md \auic lor the i .ere , f , , in emu! - ion : '. lie: i. -oi. pi lie- I . CTL030845 C K";, - ' - r p I 111 1 r ' J ! i' .<. ;:.ii \ .11,1 I l ill' '! 1 do and di.'pci MM" ,! are added tin:*! i'.j: the course of the bat. ii rw. ' ifj'i. 1 .be a.Mitim :.il v:nyi chloride p' n\ ides mo: e rr.nj> fur p -Iviner;izalion. ; uni the add it I011.ll di-pci ir.a.nt.im-. .'1 eiivd-inn as the PYC jvt r! LJS gi'iw. The \ i*tv1 i hi ' ci'! ' ui..in'.ihledtv wnuM i*K- p.M. . : i n il be< a.i 'i1 it ' ...a! ioi;::n: 11 ar- tor to -. limec, a 1 liglier r. ito of pulyn'.oriza- tion i.-; I'"' 'iblc In an example (>f the Power-.' process, the final PVC latex i ..a \ . r -t.f 2." ,i1 Sodinrn sic; rate- u; - ;iie di-pc.-mg event. Tlie follow ing C V'.'c was .-c 1 at tie.' 1 <,f a haul: iv.a in a 15-g.ii. reacth n \o~-el; Vinvl chloride.............................................................. 0.0 li>. W. ................................................................................... Pc', 7 lh. IV';.--Kim pc ; ............................................. n.27 1 i>. Amaioi.i'uu liy.mixide (t?>rr solution).............. 0.27 ]'t>. S,c ! lutes ............................................................ u.-a Hi. Sodium srenrat-.................................................. OOtP. )ii. The seed latix contained PYC particles with an average diameter of 342 A. Table 1 indicates how the \inyl r.p.d cli-;.orsiiter avmit were added during the ran, and the formation of polymer. The dispersing .-.cent was added in a 5r. water solu tion. The n ; ;.d te be added more rapidlv at the beginning of tlie r.n than ;.t tite end. Vinyl chloride was added at an approximate I'-slc of 2.1 ib. hr. for the tii-t 4ri hr. nf the rum and at ahnit 3.2 !:\ 'hr. for the next C hr. The rate of vinyl chloride addition was decreased f"r :h; next 1 hr., and finally stopped at the end of 20 hr. The latex particle produced at the end of the run 23.2 hr and 5 o'"', conversion) had an average si,;e p. . tide of 2.530 A. essentially that predicated it. the p..ient.:" Adding bo:}; '.! vinvl chloride and the dispersing agent at a variable rate dining the hatch run is a complicatin;: factor. Hence variable-speed or variahlc- flo pumps or feeding devices arc required. Several analyses of the emulsion during the run are probably done in order to check the course of the polymeriza tion. lliriih ti.lined oper.iLo] - aie needed to ensure high quality PYC. l'nlvnicr fo'n.ation laps behind the amount of vinyl chiiu ide added to the rc.icfor bv I to Id lb., as show;, in Table' I. The largest lag oceened at about if, la in the ranci fi-en 1 f. to 23 h , i.ite of polymevi-'al i. ;e mo-; rapid, except ;>1 for the lute during the- first hour. It would seem as if an auto-acceleration reaction is rucurring during tiie-o periods. Using the Powers' process. PVC particles carving i". .--.v fr 5'M to da.ooii (.,ii h-. prod a - i becaii'e the cmul-ions formed m this range ai-e hjg'ulv stab'.' U.itiei- i:i winch the PYC partules are as -m..'.'. as 100 to 2n0 A. can be u-ed f..r seeding, but the-e cmiil'ions are i ;:to un-rablo and must tie u-ed \,itii:u a few ho;;; <. If the -i-ed !:.fe-'' l,a\e '"nevb..t b;g_r i partn be', they will lie stable for longer pci iocU, up to several days nr even wcel;s. Tbe volume peieent of PYC particles that cai. be stabilized in an emulsion v.:r;e- up to about G3'f. In most commercial applicatoms, the volume peieent is probably reasonably high in order to minimize the problem# of removing waiter from the polymer. This modified emulsion technique lies important advantages as compared to regular emulsion proc esses. Particle size can be better controlled. Further more. larger particle size can bo produced uniformly. In addition, the amount of dispersing agent can be minimized nccause it ;s r........1 omy a.- iieedecl to sta bilize the particles. In the above examole, the filial PVC polymer contains about I O', emul-ifying agent. This is considerably less than that for most conven tional emulsion polymers. An even smaller quantity would l-.e present if larger PYC particles were pi o- dticcd. The modified method of emulsion polyme: ization is related to suspension polymerization in that only polymer-monomer particles occur in the dispersed plia.-e of the water. The method of preparing the seed polymer is tint known. Presumably it would be prepared in a relatively small reactor. Exomple of a botch run for o seeded emulsion-polymerization process --Table 1 Time Hr. Total Vinyl Chloride Added "10 Reactor Lb. Total Dispersing Aejenl Added To Reactor Lb. 0 1 2 4 6 8 10 IG 20 23 2 O V- ' ' 9.0 14.4 14.4 18.9 23.2 27.8 33.3 G2.4 GO.O GOO * i. 0.013 0.1 G4 0.1 G4 0.20 0.21 0 37 0.3 7 0.53 0 GO 0.G0 ` i * '.it i.. ,i : i ! % Theoretical Amount Of Dispersing Agent In Reactor % 35 22 30 20 23 35 32 35 34 31 Amount of Polyvinyl Chloride In Reactor Lb. 0.15 4.2 0.7 14.1 13.0 22.0 26 4 39.G 49.1 57.0 CTL030846 f Continuous-Flow Emulsion Polymerization J)i -it:j. inns r.f )!.rivelv **!.! (leiman pmces-rs have been given liy 1Junlop and Kev-c,1-1 1 >e-1 <>i mr.1" JI'M'tf,' a*.I Kt.hlcr ml Witt".-" The ic.iU"!' ve--i!s ]>:ni' ;i r.-!>- ka-re height-l >-:iii:idrr rut:n. One ve--.-l i- (-;.'] ,lied being 2:', 11. tall, alma; a. 1 It. diameter, and having a capacity of :5.f>00 yah1-' It is glas.--lined and is j:.. hr ted for cn.-iT.g as desired vvith brine or cold water. A simple-blnde or paddle agitator is ]-i i ti Ti ;; w, specified . being -1 ft. diameter and I.a ft. high. A variable-'|-i: .1 itl *. ' .- ! tarn the agitator at spo. d- -in t-> ad r, Water eon tail::.ne :ac- dissolvu! emulsifier. initiator and h..:'.ci i f.,.i ( :: la-'y. :;.c :..n ti',- u,,. tar through one fee.I into. vinyl A.or.be i-- metered to the reactor through another. A mav.nvnm decree of shear i< ant.,.:early tte.ded in i.te i:r..,!-i`n near tlie feed point: Bat Dur! i ami Bee-c1-1 cliaraeterute it as "gentle swilling." wh,.:: irrnlic- relatively low speed. They further indicate tiiat the level of agitation is mo-t important. Of c.-.ir.-e, re..:i\ely little agitation is provided in the bottom of the reactor. As PVC particle? farm, they become heavier than the vinyl chloride droplet' in the trnu'sior. because of the greater density of I'VC as compared to vinyl chloride. It is doubtful that any pgnijic.int phase separation occurs in the of th.' rvo.oto!* t'o.TKiic the emu normally farmed are wry stable. Variations in the r.r.'.-.ngenur.: of reactors occurred in German plants. In one. multiple reactors were ar ranged in pare.l'el-dnv 17 In another., the relatively large reactor tpi evioaslv ues-nbed) was connected in series with a much smaller unit.1-' In the latter, about SST polymerization of the entering vinyl chloride oc curred in the first (and larger; reactor, and an addi tional IT reacted in the second. The theoretical residence time of vinyl chloride in the two reactors was 2S hr. Blue dye injected in the feed stream did not appear in the exit stream for 2*> hr. Hence a small amount of back mi.\.ng was occurring, but the flow in the bottom portions of the reactors must have been essentially plug ilow. The blue color persisted in the exit stream for 100 hr. Temperature was leported as the controlling factor relative to the capacity of the reaction system. De pending on the specific polymer desired, these tempera tures vary from ;>$ to jO C. Uniform temperatures are desired throughout the reactor, but with the laminar flows that undoiibtod'y occur in the bottom of the re actor. the over-ail he. Mraii'fcr mo:l;< ionts tend to be low. Some convection currents occur because of tem perature gradients in the omul-ion fiom the heat of polymeric. tion. Operators of the Get man plants indicated that a deli- ca:,' ,\i . *1. i e k eiiiui: 1I ea-cd initia tor ioiiccntrations v.u-e the leiutimi temperature and rate of re.i. * ion. The "per.il ai e of the l-i ini' or water ill the jacket is .l> ia * rta111 relative to teniper.itlire cent rol. : '( mi the 1 e.ietor reipnrcd scveial hours in order to minimize production of oir-grade p.,l;.mer. The reactor was filled with <ool water, cmul-nier solu tion. initiator a.nd vmvl ihlonde1-1 Warm water was ii.-ed in the iael.et of the realtor until the icaetion mixture was -ulliviently heated to s'.ait )">!v n'.-i i/almi:. Oi adnallv. co-din:' water wa.- -i.b-t itnted in the jacli". and the densitv of the emn'-am was me.i--.ied fre quently until'it readied the desired level. At this time, tiie continuous flows of vinyl chloride and water solu tion were started. Start-up was then considered to be com),:, ted. When multiple reactors arc arranged in series, the number of !\>-.l pump-, feed lines, etc. v. m.'.i he mini mized. Furthe. more, temperature control would le easier. Although the reason? for using reactor- ar ranged with pai a'lel-fced .-v-Irms are s.-vj . sudi a ilow arrangement would allow muili greater flexibility in producing various types of PVC lesirs. Pmees.-ing of the emulsified product from the re actor was similar to that in batch proct.-.-o.- except a continuous-flow system cmild be used. Process improvements have no doubt been made in the continuou=-flovv processes described here. Unfor tunately details are not available. Bulk Polymerization Bulk polymerization of vinyl chloride has established itself as an important process within the !r.-t few years.r' K' Announced world capacity fog the FochineySt. Gobair. two-step process is G"h-mi!li m 71 yr. of PVC. Two American companies are presumably plan ning to use this process. Important advantages of this process are: 1. The final polymer contains no impurities such as emulsifying agents or surfactants. 2. The polymerization process is likely or potentially cheaper than other processes because there are no inert materials such as water or solvents that must be sepa rated from the final polymer. Quality of the polymer obtained by this process is claimed to be at least as good as that of the best sus pension-type PVC. Two recipes-'1 for bulk polymerization are: Ileripe 1 Hccipc 2 Baits by Weight Vinyl chloride ............ '......................... 100 100 A7o-/'/.-'-isobutyionitrile .................... 0.01G 0.02 Diisopropyl percurbonate.................. -- 0.00005 Two-Step Bulk Process In the l'cchiiicy-St. Goboin process, two reactors operated bateliwise and in scries are used commer cially.-* In the first reactor, approximately 10', of the vinyl chloride is polymerized. Polymerization loiuinues in the soioinl reactor until a tal of 7"> P* So i- p -!v - meri/ed. The remaining 2n to -j.V; of the vmvl liilotule is v ,ipori/ed, i ccoverod and rccvcled. Tin* first reactor is a vertical one : I i- pevidod with vei v f.i-; .ipii.itii.il.' -* It * no --- nated as a iirepidymerizer. Details on t1..- i iMt \11r CTL030847 i Cfc REFRESHER ... I ;uul the phenomena occurring in il arc not known. The autoclaves are jacketed, and water or brine is However, several investigators:- who liavc con used as coolant r' Part of the heat of icaction is removed ducted mass polymerisations of \inyl chlurid" have re- by the vaporization of 20 to 23rt of the original vmvl l>ortcd considerable informal ion about this phase of chloride feed. A reflux condenser can possibly he the polymerization. A precipitate of l'V(, particles is ucd to control the reaction temperature, assuming that first noted when about ! to H',.` of th.e vinyl ihl->ride excessive polymerization does not occur on the con has polymerized. The exact renditions at wliah precipi denser surfaces. tation starts depends on various operating effects in Materials of construction for the commercial auto cluding temperature. Rather surprisingly coagulation clave are not .specified, but a stainless steel one was of the PVC particles is shifted to higher conversions used in the examples givc-n in the patent.By impli as the temperature is decreased. ''' but perhaps viscosity cation, perhaps the problems of polymer formation plays a role in the rate of coagulation. on the reactor walls are quite diiferci.t than tho.-e in The precipitate of l'YC solvates considerable suspension polymei iz.aticn where glass-lined reactors amounts of vinyl chloride--perhaps 3 to G parts of are required. Additional inform;.'ion is nec-sary vinyl chloride per part of PVC pulymei. At about 15 relative to what surfaces and opera:mg conditions to 20T polymerization, the liquid phase of vinyl chlo promote `'polymer skin-." in bulk polvmei ization. The ride essentially disappears. Hence when a IOC. poly temperature used commercially in this reactor is not merization has occun ed, probably over half of the reported, but is likely in the range from 30 to GO C. original liquid phase lias disappeared, i.e., il has either Presumably the autoclave is operated only partly polymerized or been solvated. Temperature control up filled in order to obtain adequate blending. Perhaps to lO'T reaction is possible by heat transfer through it would he 50rk filled, which would mean 210 cu. ft. the reactor walls. Sufiicienl vinyl chloride liquid is of material. Bulk densities of PVC product are about available to provide relatively high heat-transfer co 0.3 g./cc., or higher. Hence, PVC product per batch efficients. will be 4,000 lb., or greater, and 5.000 lb./batch might High rates of agitation arc needed for two reasons: be an average value. A total of 500 to GOO batchcs/yr. (1) formation of particles with ratner uniform sizes. is pio'oable because time must he allowed for clean Sizes from 100 to 200 microns, or even larger,11' are ing and maintenance. Average annual capacity of each possible. Willi proper operating conditions, the parti autoclave is tiicn estimated to be 2.5 to 3.0-million lb. cles can be quite porous anu ha\e low bulk densities. The prepolymerizer vessel, however, can be essen The agitation, of course, affects the complicated phe nomena of dispersion followed by agglomeration of ihe particles; (2) promotes good heat transfer with the cooled walls of the prepolymerizer. About 10Tr polymerization of vinyl chloride is achieved in the prepolymerizer after about o hr.-6 The exact time depends on the temperature (probably 30 to GO C.), and the amount of initiator. tially filled for each batch. It should he huge enough to handle a batch for the autoclave, wl.icli might be as high as G.OOO lb. Since some vinyl chloride is vapor ized during polymerization, up to 8,000 lb. of vinyl chloride may be added to the prepolymerizer. This vessel should then be capable of holding about -140 cu. ft. of liquid; am! for the vessel described, a capacity of 1G0 to ISO cu. ft. would presumably be adequate Operation of Second Reactor for this vessel. Attempts have been made to perform both pre- polymerizalion and final polymerization in the same The second reactor, or autoclave, is stirred with rib bon blenders. The jo horizomal autoclaves are reported to have volumes of approximately -12-1 cu. ft.*'-111 Rased vessel. In one case,-'1 the initial prepo'ymoriz.ation was conducted by using a very high rate of .gitation. Less vigorous agitation and a different agitator were used on published pictures, each autoclave appears to be for the later stages of polymerization. Probably tnis about S-ft.-dia. and 20 ft. long. Details of tlie ribbon method is not used commercially, and it would have the blenders arc not known, but a patented version lias major disadvantage that a large vessel would be needed two or three ribbons wound on whorls of different di for both stages; whereas a smaller vessel is adequate ameters.'1 One of the ribbons turns in an opposite, for prepolymcrizalion in the two-stage process. In the direction to the others. The autoclave blades are con conventional process, a prepolymerizer vessel can prob tinuously denned by the stirrer blades, and good heat ably service at least two and possibly thicc autoclaves. exchange is maintained.-7 Operation of the second reactor is the critical por Cost Dnta for Process tion of this process. Ihulesircd agglomeration of the l'YC particles must be prevented, and adequate tempeiaturo control must lie maintained as the reaction mixture is transformed from a slurry, tlnougli a sticky solid stage, to C'-untialiy dry particles. The speed of the ribbon blenders is decreased as polymerization Tlie flowsheet for the entire process lias been re ported by Krause.,!> Recent patents'--s probably explain some of the specific details. Recovery, purifica tion and recycle of mireacted vinyl chloride appeals much simpler than in suspension-typo processes. The progresses ftom perhaps 20 to 30 rpm. to 3 to 10 rpm.1'' The second portion of the i caction pinh.ihly re quites 10 to Jo hr. in order to obtain the desired polymerization.1 PYC pi oducl requires little additional ps ocossing. I reli ably the cost of polymerization by this method would be considerably less than by suspension niolhod'. per haps about 2y/lb. of PYC formed as compared In '>? CTL030848 I f by the ss: jXTisior. prove---. In nddii inn. ir.vi tvnent costs 2. Possibility of air leaking into the svstein because I arc churm-d to he A -JS-xnii; 1 > n Iii./yr. plant is tiip filter prcs-cs arc opened and exposed to air during reported in cost $1.7 million, excluding storage tallies the cleaning cycle. and packaging. Either centrifugal separators or continuous-flow filters arc probably uced in modern commercial proc Solution Polymerization esses. Lead-lined autoclaves were specified in the patents, Solution polymerisation, as practiced in most com but glass-lined reactors are likely used now-. The lead mercial processes, is not a true solution polvmerizution was probably used as a corrosion-resistant material since polymer precipitates from the solution The term and not as a catalyst. ``precipitation" polymerization is sometimes applied Benzoyl peroxide was specified as the initiator at to this type. In several respects, this polymerization concentrations of about 0.5'f by weight. Popular is related to bulk polymerization because the polymers initiators such as lauroyl peroxide and isopropyl are quite insoluble in tli? iiquid phase. peroxide would likely be satisfactory in piesent com In bull: polymerization, the polymer begins to mercial processes. Azo-his-isobutyronitrile was used precipitate from the liquid phase when about 8 to 10ra in the recent experimental program of the precipita of the vinyl chloride has pol.vrr.cn/.cd. In solution tion polymerization of vinyl chloride.-'1 polymerization, the solubility of the polymers in the The separated solid polymer or copolymer particles mixture of solvent and vinyl chloride depends on contain considerable amounts of dissolved \inyi chlo the solvent, tne concentration of vinyl chloride in the ride and some solvent.2-'* Polymerization reactions will solution, and composition and molecular weight of continue until the vinyl chloride is separated from the the polymer and copolymer.'-'1 When cyclohexane is the solid particles. Temperature and pressure to which the solvent, a granular precipitate occurs from the "start" separated particles are subjected is most important of the run. Tolrahydrofuran is a "good" solvent for relative to flushing of the vinyl chloride from the PVC polymer, and precipitation doe* not immediately particles of solid product. occur. Temperature control is easier to obtain for Solvents, other than n-butane, have been used in solution (or precipitate) polymerization as compared experimental investigations,- and include the fcl- to bulk polymerization because the large amount of lowing: benzene, chlorinated alipnatics, cyclohexane solvent acts as a heat sink. and tetraliydrofuran. Normal paraffins such as /(-butane When polymer begins to precipitate from solution, contain only primary and secondary carbon atoms, and auto-acceleration is frequently noted because the they are less active as chain-transfer agents than iso precipitate contains consjdei able amounts of solvated paraffins such as isobutane that contain a tertiary vinyl chloride (ana po=sibiy comonomer) in addition carbon atom. to polymer or copolymer. The polymerization reaction An important advantage of solvent polymerization is continue'- in `lie pre-eitmate, and the polymer so pro that water is not used. Hence, recovery of the final duced ha-- a much higher molecular weight than product is simplified. The unrcacted vinyl chloride and that produced in solution. This latter polymer tends solvent arc generally easy to remove, especially if a to June relatively low molecular weight bc-cair-e of relatively volatile solvent such as 7!-butane is present. chair,-tran-l'cr reactions with the solvent. A thorough The recovered vinyl chloride and solvent arc prohably diseu-'ion of the complex mechanism of solution (and recirculated as a mixture. The final polymer product precipitation) polymerization is given by Mickley, contains few- impurities since presumably no emulsi Michaels and Moore.--1 fiers or surfactants were present. However, the prod The basic process that Union Carbide Corp. is thought ucts probably tend to have relnti\ely low molecular to be using is probably described in two patents.1'-0 weights because all solvents act to some extent as chrin Presumably the process is used to make copolymers transfer agents. containing 75 to vinyl chloride, with the re mainder being \inyl acetate. The mechanically-stirred Low-Temperature Solvent Polymerization autoclave maintained at approximately 10 C. contains a slurry consisting of about SO'T solvent, designated Solvent polymerization of vinyl chloride has recently as n-biilane. The remaining 20'r consists of the comon been publicized as at least one of the techniques for omers dissolved in the solvent, and solid copolymer producing an improved polyvinyl chloride that is being present as a precipitate. The slurry is pumped to a used in making lOO'd- PVC fibeis.:ili Several American filter press to remove the precipitate, anil filtrate is companies arc thought to be interested in those newer returned to the autoclave. Make-up eomoaomers and types of J'YC fibers. In France. PVC fibers have cap initiator aie continuously added to the autoclave. tured the following shares of the market: flags, Ofl'7 ; The pi ocess as described in the 10.12 ami 10-'17 patents curtains. underwear, 30rI ; ski and mountain has undo 'I'tedl.v been modil'.ed as rower equipment has clothing, 2a'I ; sweaters and blankets. 2^'. ; caul auto become a\eilahk\ Two batch Idler presses were orig and airplane interiors, yarns, and artificial furs, lo'I. inally specified. One tiller pros was emptied and These PVC libers arc significantly cheapei than nylon, cleaned while the other v as being used. Such filters polyesters and acrylic fibers. would have two disadvantages: New jo messes have been developed to obtain PVC I. Co i-Mcrablo hand labor. fibers with improved properties. Low-temperature Cli/S.icul --July 3, )Vi7 CTL030849 VI I CC REFRESHER . . f j>o!_vr'.H-n;-;iti<>:] icchmivj"' nrc generally u?c i v. >h a References I Miila;/" catalyst and solvent. S.^ciete Jim" yli'1 is I AP-tirM. I, V , 1t,yruoriratJon of Vinxl Chloride. Chrm. though! to be iiolymerizii'.g vinvl chloride at ti:i;irr- Hi.'/ . .v.i\ ' l ".7 p I ` 1 - AI Pri-hi I. I*. Ymvi Chloride Pol vmcrlTntlan b> Sn^ps-n- atures from --JO to --20 C.. and Apjdicazinni Chirmche Societa per Azioni10 polymerizes at --20 to --00 C. .s,4n I'luMss. . chthi /.n*/. .)iue i t:*.7. p 14*. 5 A'trn.t. .. .1 :i r: i \\ M. f'orne I'.sperun* nt* on l?tj!k pm ri^.tt iot of Vinvl Chh*rMo. ./ Vohnmr Net . '''I by using mass-polymerization technique. The n..mU- (I ' 3 ) < I'. I'm S t I>*;i r.*TSf tr Ap***M of UmulMon Pol vmriz'\- injt polymers arc more highly svndiot..ctic because of f ( /*; , t ,, . 4*.. 17 7 (l '*..7> "* f. up'' W ,i t.l V..rri-h H C \V . Moihani'ni ted the low polymerization temperature^, and they are Kit - : . .>r ` i - t. r- .t.v |\* \ ri7 oi Vi:\ 1 r 11 < r j r. \ 1 | --i *.tr i A /.!1 \.| \ m-'i 7.11 i*n of Yirvl Ci.Mnd**, quite resistant to boiling water and to dry-cleaning solvents. l'vn l.'t.wil '->> ( ) \imi. ,10 1 b il("i<K<*r i< t'' I,.it. vt t. l`.vy I'erhmex -SL Gobi^n Pnlx uieri- zation Pt"* ? . ( ?. a i'u .hm<' \3. '***'.. j> sj Sociele Rhovyl indicates that fibers can be produced 7 t 'ntiip ue 11*- iitio't " Frwn-h Patent 1. J . I * t'onl 1 miiawm Uulk PoJymcriza( \pr 17. V^J) with break strengths of greater than 2g./denier. S Ciunp.u-d- S.nrtt G..h.t'r, Hulk-1':\ iip-nzatton Appara tus" 1: r 111 *.' i I* its-rit ( 11- - *3'< VC2) shrinkage in boiling water of less than 10%. and P C'/rso <. .ir.i! Krrr.tr1 K I'mul -mn }'i>!\ m'Tiz.tli.in of Vinvl Ch'-fle. l/,.ri.- y-MsfitAr. \n JO ( 1 -h.j ) -.*7. \. S fl`.*:i> shrinkage in trichloroethylene at 40 C. of less than 10%.15 Actual shrinkage values are probably signifi 1- Ivt.'rr J. 1V1\\in\l Chloride, ft. <75 ur*'"i >. Mnltrrr.* riatujurs. II 6 H < to Vnion Carbide atni ^'nr?on C.i^p >. "Pr--t'-s u.r l>roKHin4 \'ir,\.' KoNin^," L'S I'.iUnl 2.0 7. cantly less than JO'f . Adhesive fibers can be prepared <>f..r 1'-37) if a copolymer of vinyl chloride and vinyl acetate is i - ImijIo'i. 1* p -riI Kci'O. F H. Continuous r'ol\tncrj7att>n in G.'1 ttuiit. h >1 hn <1 Chi'". 4 o, *1 * 4 tll'h) produced or dissolved in various water-soluble solvents before spinning.'-14 Suitable solvents include acetone, 12 h*lti )i. K M. Tl.t* T! of Cmuision Pnl\miri7.T;'."n. J(K-!.'<7;>) 1 7Vc/. 11 r, lt,ij / uttil I.rtjutccnno, -1^9, Part il, r - - 4 S 1-5 c7.'ilh-n, ,T. P nnil S'U'in. \V. M. Copolvni< r4Atu>ns Ktv- tetrahydrofuran. dimethyl forinamide (J)MF), and dimethyl acetamide (DMA). pk">,t*^ \*it 1 i C h.rulo or \*11 \ lulene Chhiri'le ns I'nwup.i) C..nw (''I p X tu "Cup..:> tnert/anon " l-AineJ by O. 17 ll.irn. lfiivi'w lentr. N*\s York. 4 Copolymerization by a sovent-polyir.crizalion tech nique must lead to interesting mass-transfer features. 15 Gorri. I. J (t SHiete Ithovyl). *'Po!vvinyl Chloride Pth-'rs and Pn-Ccsst"* fop F-'roiiuoiu? in me,'* U.S Patent 22. I < FVl>. 1 < M.irklf'*; W. in.. General Theory of the MechanJsm of The relative rates of transfer of the comonomers from Emulsion Po'wra-rizutson. J .-VC.''. Bft. H2S (11*47) IT Ilc-rte. . Horstollung \un Holy\invlchlorid, Chem. Tech., the solvent phase to the solid copolymer particle are 4, 327 Uf*22 ) ly lloprf If a*ul Pakla, I.. The Influence >f the Comn^sifim likely quite different. If so. the composition of the co polymer would presumably change as the extent of of the I7tnu.'.r.''r on the F'^lvmcrt/ation of Vinyl Chloride, Mul r-')itob Chru1 . <8. >4-74 (If"'*.*) IS. Krause A.. Ma^s PoV%menzation for PVC Hein. polymerization varies. >}'/, 1 *er I 'V5. pp. 7_'-74. 2^ Krt-hV r. }\. and \YicI:. G. "Kun-'tstotf-Handhuch : Hand II Polyvinv ;..Vond," Part 1. pp. 41-72, Carl Hai.aer Vcrlagr, Munich. r'3 Chlorination of PVC Resins 21 Lain*-.'-* U. A.. r>r\ir.? Tricks Tailor Hesin Properties. Chcii Far/. Vox 1C. It ;;> 1CC-1C0. 22 Mrs<xC> . ''j-chk*'. H and Scih'.d, E. Application of High-temperature stability of PVC polymers regard H> i.. olx/abb: 71. " 4 ( 1 ) in the Field of Pla&lks, .4/iycic. less of the method used for polymerization can be 23 aliikir. . H Miehac!*. A S and c>re. A. L. Knf.M^ of 1'(I'Cipifat !<> l-Mlx uten7.,iL:on of Yin\l Chloride. J. /*o/yacr improved by chlorination. Chlorine adds to the unsatu Set , on. \-j\ 24. Piusr"- i K T--t.i. F and Talamini. G.. A Kinetic ?:vdy rated and relatively unstable end-groups of the polymer of lh* ll:iiul*'n,i i'nlx r' rmatmt. of Vinyl Chloride, Mnk ro//>/ron- Ini r C f < u, . 11. 177 (!.' . molecules. As a result, zipper dehydrochlorination is significantly suppressed. 25 I'o..!. J H. to P. p CMVirtri*-?*. Co. >. ` Polx ui^n/ati-TP of Yi.:>! ' 1'S 3.32''*.250 (?t;jt. lti'.O). 2'- I'rml'::'.' Cfimuo.u'" IV* : lucv-Saini Gobatn. "rrepar.iti*! Chlorination of PVC resins is used to a limited Pro 0-5 no i'.-U ivort'.-.:- .p of and Cop *.x iu.*rs ttf V; >\I Ch)">* <1-" l'ren. h lMi-nt 1.T" T.7ft*7 plus 1st. 2n1 ami extent, but the commercial method for this chlorina 3rd .ni l.lion-* tr> p.Uc n; (Apr 10. l'"Hl. 27 Pro(i'<;,v* Ct.inv.*iues l\vKt"v*,'Uni O^'h^in 0v E. p.en-it tion is not known. The TVC resin could be dissolved at.*1. . x*. H\ . .*.) A.(*> htxe ,,*r Hulk P*!yn\vri7.i\io:\." Pr**' ^ Pat*'! 1 ! (M.ix *s. 10'U). also I'rciKh Athhtion <7.527 in a suitable solvent, and the the solution could be contacted, at the proper conditions, with chlorine. This (Ju,- 31 l,*`.4 ) L's Piodi.t!- Chinn ins P**.-hinev-S.unt Gohaln <b> J. C Thc'it*). ' i- ilk Poix im riz\:ion of Vinvl Chloride." French technique would appear to be expensive since PVC l\u.it 1 :>3 < 73 M'vt- i". 1 **;4> 2s* 'Uul i: \Y (to t'.tihi !, and Carbon C'hetnicaN Corp). resins are only slightly soluble in most common sol "Yrtv] r<`' l ..S i'*i-i:* 1..*15.577 (Vox* 14. l'57) V* . Prx i'r.* i. ) rf I'o'.vvtnvl ChU'riO..' T'.<('- vents. Large quantities of solvent would appear to lie T> p s. ui Sp'ax 1 a*-i aha' 1 u-. }\ mi stzl r. .i. No 7. lftS 51 Siiiiiu, \V . *\*n:;.l IVmm*." Kftnhfld. Nexv York. 1 `4. necessary, and the cost of evaporation and crystalliza tion would likely be high. 32 \Y M.. "M mulaet 1:1 e if IM.i-ties." Yul. I. Ilcmhoid, Ncxv Yo: k. f*:i TC Smith NV. V i.d !Txx `rt, 1*. H.. Kimdirs of Kmulfion Another possible technique would be to use a Pol' r 1 n Mtion I. Cb> ri. /'/;<"*. If.. 5*.`3 24 >>. t te llhoxxl (h\ 1. ! Gold ami 1\ Miele|), "Ai!h* si\C fluidized-bed in which a gas, containing chlorine, was Fii<r' I'r-il-h Ikurnt 1.414 v l < tcet. 23. 1 .V TakiinMn <7. and V'idattn. G. l\dx . of Ytuyl contacted with granular PVC resins. Chlorination Chi.-- id* lr'li its il bx* Yn--tu:\ll*oran*-t\x ceii S> steins. .Vul.ru- in 1W* < kii ;m. 13' (1 !`4*. 1). would initially occur on the surface of the solid par ticles. Chlorine molecules would presiiin; blv slowly i:.u-..m 1,, Ills tl.e \Va> m 100c PVC Fibers, Chcm. Fnp. .Yti.*. Mar. 15. I '57. pp. 55-3T. penetrate to the. interior of cadi particle. Possibly some substitutive chlorination would occur especially on tlie surface of the particles. Acknowledgement Several ropresenativos of indiistiial companies that produce PVC resins have leviewoil the articles in this series dealing with PVC polymers. The author is ex tremely grateful for their comments and suggestions. Key Concepts for This Article A ctn r (M Matil' . 1 aiuri:. :* i'r r ft ) 1 .:. X fix 1 1 hlorlde* t'* }*' A met ' PI IV ! 1 s 1 kiix nn 1 > Mmim/MrOmiU (10) Pr-i* e s p.itvimu ' it Inti* 1 Imulsti.ns* SmIiii Min*< ( pit \ t u **) P.ulk* < W * . iu ImM t re r.i* iiidlanP-rv . nttnih' T- m rf-pM*i'l Kir V 1 C1.K s ,.J. 1 1 , M apt f. 1 1J *le inD'l !l 11 ! tl A*l" t !* in it . \ .'.ii. . s 'i -a d ;<. tm!i mi, 1; < M !h r v . 1! add. i 11:1,4. >\ c 1 .11'. '11* .* all . Jt a. I Ilill'-Xlll. * n; 1. ; s |..-d Ml r 1 f/ . \ 1 11 t: .... p 1 N 7 <>r x mi tn IX Mflt.-I Key C"h" 1 x pi It l, .IK, tiMMJ: Pc i< U r Setr* iu 1`uste ai d.) r 92 CTL030850 ity 3, 1967--Cltimicnl ro^!-