Document EqagRaXjd75gGOKaBeG64Jggj

-diforflia baffles in icse tests it ions in 1 one t the e in pre fer. the on the l of a some of ix-stage rcuiting ijor obprocess e units tired, . lolyiner. lytliing except, : slides ! nmhnd the f units do not =quire- s - the t ever ze it least 1' i Oil rease : pri and laiity proosts. Co., been tons iem- I ited f tars i vitli ; in ade eai is >ni- 1 1 ;ra, ich \J ` `7 MONOMER RECOVERY IN GR-S MANUFACTURE C.` R. JOHNSON and W. M. OTTO Firestone Tire & Rubber Company, Akron, Ohio Tli National Defense Advisory styrene molecules (mole wt 130,000Gtmmittcc -first considered a syn- 250,000). Since the charged monomers, Development of Batch Recover/ Process tlietic rubber program in August, 1940b. utadiene and styrene, are reacted to In May, 1941, the President approved only 72% conversion, it is necessary for the program of constructing four GR-S economy and quality reasons to recover (butarfieue-styrene-type) plants provid the unreacted monomers before the latex ing a total annual capacity of 40,000 is coagulated to form GR-S rubber. long tons of synthetic rubber. - By late spring >n 1942 the program had been expanded to 805.000 tons, including four types of synthetic rubber. Construction of tlie $700,000,000 industry was nearly complied in 1943, and by the end of 1944, 949.305 long tons of synthetic rubl>cr had been produced. Agreements providing~~Tdr" the ex change of patent rights and technical information were executed in the var ious branches of the industry and the Rubber Reserve, rubber, chemical, pe troleum, and industrial companies com bined their technical and scientific in formation to design and successfully operate the plants. The 705,000 yearly long-ton capacity assigned to GR-S rublier production made up 87.5% of the program. An important part of this program was the development of the monomer recovery process. This process is the method of recovering unreacted butadiene and styrene from GR-S latex. This report will describe how the monomer recovery process was developed from the basic Composition of the latex before re covering unreacted butadiene and sty rene is presented in Table 1. It is seen that 15.9 lb. of butadiene and 12.11b. of styrene are available to 'be recovered/ 100 lb. of monomers charged. Impuri ties present in the butadiene and styrene charged to reactors are concentrated in .the recovered monomers in relation to their boiling points. Several properties of GR-S latex greatly affect the design and operation of recovery equipment Soap makes the latex foamy in nature, and it froths violently when it is degassed of buta diene or is suddenly subjected to reduced pressures. Scum or precoagulum (ag glomerated rubber particles) forms on the surface of latex when water evap orates from it due to a drying action on the particles unless agitation is used to keep them wet. Coagulation also occurs with addition of superheated steam' or when latex is subjected to abrasive action such as when it is rul>1>ed between a stator and rotating disc. Batch recovery equipment is u-xl in two of the first four plants built in the Government program. Knowledge gained in operating these plants aided in designing the continuous recovery process used in the Standard plants. The hatch process lias proven particu larly effective in the production of high solids GR-S latices. ---- Only a.few months' pilot plant rk had been completed when it was m sary to design "the equipment for ic first plants using hatch recovery systems in'1941. The facts known about the properties of GR-S latex and its raw . materials, together with information gained in rubber company pilot plants were used to design - the recovery sys tems. Most of the information presented in this section is based on information obtained at the Firestone operated syn thetic rubber plant Since butadiene and styrene at atmos pheric pressure boil at widely different temperatures, 23.7 and 293.4 F. re spectively, it was logical to 'try a t\s ostage distillation process for separa1 ig and recovering the monomers. Ber. se butadiene and styrene are only slit; ly soluble in water, the recovered n>< omers may be separated from water by iilea to the batch and continuous plant As would be expected, latex becomes decantation. scale level- Emphasis will be placed on more viscous when it. is concentrated. When samples of freshly reacted the practical engineering work per Heating latex to more than 150 F. be GR-S latex were exposed to the air, formed in setting up the new opera fore or while the monomers are re most of the butadiene volatilized rapidly tional procedures and in eliminating covered tends to polymerize residual with simultaneous foaming of the latex. equipment problems* styrene, and changes the properties of The size of compressors ami the rates TIk paper will first define GR-S latex the final coagulated rubber. for butadiene recovery were governed and wiH present important properties of To aid in designing recovery equip by the latex foaming at various temper latex and its raw materials used in de ment the monomer properties were in atures and pressures as well as by die signing recovery equipment. Then, the vestigated and are presented in Tabic 2. rate of diffusion of butadiene frmu hatch and continuous recovery systems It will be noted that there is a wide latex. Research and pilot plant iinv- ti- with related process and equipment de differential between the boiling points gations were limited due to lack of t: le. velopment work will be described. and vapor pressures of butadiene and Considerable 1>asic theoretical in for. na styrene, and the monomers are only tion was published later within the n- Properties of GR-S Latex, Buta diene, and Styrene slightly soluble in water. Due to the high heat of combustion and wide explosive limits of the mon dustry by Princeton University throe gh the Rubber Reserve Co. Since tlie butadiene and styrene were Standard GR-S latex is a colloidal omers in air and the volatile nature of to be recovered in two stages, it was dispersion of rubber particles in soap butadiene, a thorough study of safety necessary to determine the mole frac and water. The rubber particles are equipment needed in the recovery proc tions of butadiene, styrene, and water long chain copolymerized butadiene and ess was initiated. vapor above latex at various tempera- 49 Yol. 45. No. A (.Util) f CHEMICAL ENGINEERING. PROGRESS |JQQ 041337 Page 407 -/.V Tail* 1,--Composition of Latex Be hires and pressures as is shown in Table ' water/lb. of styrene would be distilled fore Monomer Recovery 3. It was evident that at 104 F. a pot when the residual styrene content ap 72% Conversion stripper (batch recovery vessel) con proached 0.1%. taining freshly reacted latex could be Although it was not known at that Parts/100 Parts Total Monomers Charged evacuated to at least 152 mm. Hg. ab solute pressure before much styrene would be distilled into a compressor. Recovery studies, soon revealed that time, analysis of experimental data indi cates that the controlling resistance to transfer coefficients is located in the liquid film between the phases. Labora Water-polymerization and short stop .......................................... 185 Butadiene (polymer) (76.5%) ... 55.1 Styrene (polymer) (23.5%) ....... 16.9 Butadiene (unreacted) .............. 15.9 Styrene (unreactcd) ................. 12.1 ' Butadiene (impurities)* ............ 14 Styrene (impurities)t ................ 1.4 Soap ........................................... 4.3 Plasticity modifier,catalyst and re action stopping agent t ............ OJ it was advantageous to steam distill the styrene at the lowest practical absolute pressure. At lower equilibrium pres sures and temperatures the latex foamed less, and polystyrene formation was minimized. A two-stage ejector system capable of obtaining a suction pressure of 50 mm. Hg. absolute and which would handle the required quantity of gas was tory data indicate the possibility of slightly lower steam consumption in batch operations when steam is added at high velocity through parallel orifices into the latex; however, in plant opera tions the orifices plug with coagulum too rapidly for this method of steam addition to be practical. When it was noted that some buta Total , ...................................... 294.9 specified so that it would be possible diene vented from the styrene system, to steam-distill styrene at temperatures . it was evident th^t the discharge from * Principal impurities are: 1-butene, 2butene, propylene, n-butane. t Principal impurities are: ethyl benzene, 4 vinyl-l^cydo-hexene, isopropyl benzene, 1-3 butadiene. t Parts given are the quantities charged. as low as 120 F. Experiments showed that 4 lb. of water would be distilled/Ib. of styrene recovered at the beginning of the proc ess, and several hundred pounds of die ejector system should be tied into the suction line of die butadiene com pressor. * In designing the recovery equipment, careful consideration had to be given to the properties of the latex. Foam space was needed in th? strippers, and liquid Table 2.--Physical Constants of Monomebs -- ' Butadiene Styrene lines could not be sized using standard fluid flow' calculations since allowances had to be made for gradual coagulum Formulae Molecular weight CH, = CH--CH.-- CHi 54.088 CJLCH *= CH, 104,14 build-up. Vapor lines had to be over sized for the same reason. Also, desu perheaters were specified to reduce pre Physical appearance dear, colorless liquid under Colorless liquid, sharp coagulation in the latex caused by pressure; gas at normal at- aromatic odor superheated steam._ mosphereic conditions Boiling point at 760 mm. Hg. kVapor pressure ` Melting point Density at 60* F. Solubility in water at 15* C. Heat of vaporization Specific heat of liquid Heat of combustion of liquid Explosive limits in air --4.6* C. (23.7* F.) 145.2* C (293.4* F.) 11-2 mm. Hg.--78.51* C. 760 mm. Hg.--4.6* C 2144 mm. Hg. 25* C. 3338 mm. Hg. 40* C. 13 mm. Hg 0* C. 6.9 mm. Hg 25* C. 20.7 mm. Hg 50* C. 166 mm. Hg 100* C --108.9* C -^30.60* C 5.229 lbygal. 7.582 Ib./gal. 0.13 g./100 g. water . 0.0063 g./lOO g. water 99.8 cal,/g. at --4.6* C 179.6 85.0 cal./gmole B.tu./Ih. 153.0* B.tu./lb. 146* C 0.517 cal./(g.) (* C) at 25* C 0.413 cal./(g.) (* C.) 11,158 cal./g. or 607.4 kg. 1,046 kg. cal./g.mole cal./gjnole Z0 to 11.5% by vol. ' 1.1 to 6.1% by vol. Table 3.--Vapor Composition Above Latex at Various Temperatures and Final Venting Pressures (3) 25* C (77* F.) 40* C (104* F.) 60* C. (140* F.) Total Pressure (abs.) in. Hg. - 1.8 3.4 77 19.1 3.8 6.0 12.0 30.5 8.1 12.5 22.0 53.5 mm. Hg. 44 85 182 485 97 152 303 772 204 316 566 1353 Butadiene 0.424 0.707 0.865 0.951 0.403 0.622 0.812 0.928 0736 0.510 0.728 0.887 Mole Fractions * Styrene 0.030 0.013 0.0050 0.0015 0.029 0.015 0.007 0.002 0.033 0.018 0.009 0.003 -- Water Vapor 0.546 0.280 0.130 0.048 0.568 0.363 0.183 0.070 0.730 0.472 0.263 0.110 Description of Process. Figure*1 is a flow sheet of the batch recovery process used at the Governmentowned synthetic plant operated by The Firestone Tire & Rubber Co. in Akron. A 3200-gal. batch of latex under 35 lb./ sq.in. gage butadiene pressure in the reactor containng 1575 lb. of unreacted butadiene and 1050 lb. of unreacted sty rene is pressured into a 7500-gal. strip per. During this process the butadiene flashes off, and the flow of latex to the stripper is controlled so that the recov ery compressor maintains a vacuum in the stripper. After all the latex is in the stripper, degassing continues at the maximum rate possible without excess foaming. The rate is controlled by throttling of the compressor suction valve. The 1800-gal. surge tank knocks .out entrained latex and foam. The butadiene vapors are compressed to 65 lb./sq.in. gage pressure and liquefied in the shell side of a two-stage .watercooled removable tube bundle-type con denser. The butadiene and nonconden sables flow by pressure differential to the tank farm blend tank and butadiene absorption system. Butadiene recovery operations are terminated at 152 mm. Hg. absolute pressure at 110 F. Sty rene and water drained from the com pressor are blown into the next batch of latex. Approximately 1 hr. are re quired to recover the butadiene. Page 408 CHEMICAL ENGINEERING PROGRESS ucc 041338 June, 1949 TO a*MC TOBU, i ctn pi Dti .- gc an 14 ; ess ei w ate denst grav knoc to a cove tem (Iron ove '.le - ape ^ysti !-ate -,eco char twoof resi j.lr A the 25 wit' late the Vc TO ATMOSPHERE *...,' ''^kv ` ,- - . -- v * * 4 ` `'*<){ -` ",^^1 ` ~ry^ss^f -rw TO BLEND TANK CENTRIFUGAL PUMP DEFOAM ER rUM,,CIfc- STEAM,. ---------FLOW METERC _____ ____ STM PPER ' _ LATEX .am un. DESUPERHEATED ' STEAM _^To LATEX LEND TANKS BUTADIENE COMPRESSOR Fig. 1. Batch Monomer Recovery Procesi. DRAIN TO 8LEHD TANK Desuperheated steam at 15 lb./sq.in. it may be evacuated by quickly venting gage is added at a controlled rate to the steam into the styrene condenser. steani-distill the styrene from the latex Originally, nitrogen was used, but time at 140 F. and 100 mm. Hg. absolute was lost evacuating the gas using a pressure. The styrene and water Vapors separate plant vacuum system. are condensed in the shell side of a- water-cooled drip-type vacuum con Operational and Equipment Problems. denser. The styrene and water flow by Solutions for many operational and gravity from the condenser through a equipment problems were needed before knock-out drum down the barometric leg the above recovery process was devel to a decanter where the styrene is re oped. A few of these problems will be covered. A two-stage steam ejector sys considered. tem tied into the top of the knockout Originally latex was dropped rapidly drum maintains the vacuum in the re- from the reactor creating butadiene -covery system. During the first hour of pressure in the stripper, but subsequent the styrene recovery cycle, butadiene degassing of the latex was accompanied vapors arc discharged from the ejector by violent foaming. It was soon discov system into the compressor suction line. ered that when the rate was slowed Later when all the butadiene has been down to maintain a vacuum in the recovered, noncondensables are dis- - stripper, the butadiene flashed off with charged into the atmosphere. Over a out excess foaming. two-hour period approximately 5000 lb. . At first the steam distillation of of steam are required to reduce the styrene was controlled by maintaining residual styrene content in the latex to a constant temperature in the latex 0.1%. using a temperature controller which After the styrene has been recovered, regulated the steam flow. Later, foamT the latex is pressured to storage using ing was reduced and more efficient re 25 lb./sq.in. gage steam above the latex covery was possible when a steam flow with the agitator turned off so that the rate controller was installed. In the latex will not be appreciably heated by former system coagulum formed on the the steam. When the stripper is emptied, thermometer well and caused poor tem- peraturc control; surges of steam caused excess foaming._______ An' important development in the batch recovery process was the discov ery that high molecular fatty acids, candelilla wax, and Dow anti foam A. could be effectively used to control foaming. Subsequently a chemical feeder was installed to add specified amounts of the defoamer into the top of the stripper to break the foam. Coagulum build-up in styrene con densers has been minimized by the con tinual spraying of liquid styrene into the top of die condensers to wash away and dissolve coagulum as it forms. Originally noncondensables and buta diene in the ratio of their partial pres sures at the operating temperatures were vented from the butadiene con densers through vent traps and dia phragm valves activated by pneumatic pressure controllers. Pumps were used to transfer the recycle butadiene from the accumulators to a tank farm blend ing vessel. Now the vent traps and pumps have been eliminated and the compressor discharge pressure is used to transfer the butadiene to the blend tank. The pressure in the blend tank is con trolled by venting the noncondensables Vol. 45, No. 6 CHEMICAI ENGINEERING PROGRESS ucc 041333 Page 409 and butadiene vapors through a dia latex. Tile technical committee first de pitch, uses adjustable weirs to main phragm valve activated by a pneumatic cided to use four vessels in series for a few inches of latex on the trays, ail pressure controller into a styrene ab tlie continuous styrene recovery system, operates at 60-100 mm. Hg. absolute sorption unit where the vented butadiene but to go forward with column develop pressure at the top of the column- More is recovered. ment By Sept, 17, 1942, sufficient pilot attention will be given to tlie alxivc- plant work had been performed so that listed design problems later in tile re- Efficiencies ami Advantages of Batch a decision was made to substitute a port when operating problems are dk. Process. At the present time 98 to 99% multiple plate column. cussed. butadiene and 96 to 99% styrene effi Preparatory work was performed in Time was not available to exjvcriment ciencies are obtained for GR-S produc the laboratory using an 8-in. glass col with various types of pumps, control in tion in plants using batch recovery oper umn, Then a steel 4-plate (bubble-cap struments, condensers, etc. Decisions ations. The butadiene efficiency ap type) styrene purification column was for designing the complete monomer re- proaches that of .the Standard plants converted into a recovery unit to obtain covery unit were based on fundamental which use the continuous process, but it preliminary operational data. Most knowledge of unit operations and avail A is difficult for the batch styrene recov pilot plant work was performed on a ability of equipment ery process to compete with the more flexible, multiple plate. 18 in. Lapp por The Office of Rubber Director pro efficient continuous column styrene re celain column. Sets of distillation data vided valuable assistance in expediting covery process to be described later. As were gathered to solve many design die delivery of critical materials and was mentioned previously, large quanti problems, some of which were the fol equipment so that many of the plants ties of stcani are required to reduce lowing: . ' - were completed and started operations residual styrene values below 0.1% in m 1943. the batch process. Although continuous 1. Foam space needed at top of col operations are preferred for straight umn and spacing of plates in rela Description of Process. The contin GR-S production, batch operations have tion to various latex and steam uous recovery of the monomers (Fig. 2) a definite place in the sydthetic rubber flow rates is accomplished by first passing about program. It is not feasible to recover styrene in 'perforated plate columns from high solids, viscous, andunstable latices which may be processed in batch cquipT 2. Number of perforated or bubblecap plates needed foy effective styrene recovery with a minimum steam consumption 40-60 gaL/min. of latex through two flash tanks where die butadiene is re moved by reducing' the pressure from 30 lb./sq.in. gage to 3 lb./sq.in. gage and then to 220 mm. Hg. ment. Batch design by nature Tends- iti_ -- 3. Size and number of bubble-caps absolute. From the second flash tank self better to plants producing a variety or perforations in plates required the latex is pumped to a perforated of products and experimental latices. for effective styrene recovery and plate-type column where the styrene -These facts are important since unco for minimum pressure drop across is removed by steam stripping at a pres agulated synthetic latices are in demand column sure of 60-100 mm. Hg. absolute. The in manufacturing present-day latex > butadiene vapors from the flash tanks products. 4., Depth of latex needed on plates to are compressed by the use of centrifugal maintain liquid seals and to give vacuum pumps and compressors and Development of Continuous Recovery Process In the spring of 1942 when construc tion of the first four individually de styrene time to diffuse from latex 5. Choice of materials for plates to minimize plugging and to facilitate cleaning discharged to a vertical tube condenser at 60 lb./sq.in. gage' pressure. The li quid butadiene then flows by gravity to a receiver from which it is pumped to feed tanks and is returned to the tank signed plants was near completion, and 6. Type of downcomers and method farm for partial reuse. Vapors from the operations started, the four major rub of sealing plates to minimize plug stripping column contain water, styrene, ber companies, Standard Oil Develop ment Co., Blaw-Knox Construction Co., and Rublier Reserve Company, worked together in Technical and Design Com ging small amounts of butadiene, and en trained particles of latex. Vapors pass 7. Most effective operating pressures through a foam trap where the entrained and temperatures . particles of latex are removed and then mittees to design a standard plant This design was used in constructing the rest of the plants in the industry. The sub 8. Method of adding steam to latex flow through a three-stage vacuum con- . denser where most of the imreacted 9. Method of obtaining desuper styrene and water are condensed. The committee on monomer recovery was heated steam to minimize preco liquid styrene and water, and styrene, formed in July, 1942, and in three agulation in column butadiene, and water vapors flow to. a months* time it obtained the necessary pilot plant data for setting up the con tinuous recovery process. Figure 2 is a flow sheet of this process slightly re vised to include a few later changes. Ex perience gained from batch operations 10. Effects of many latex variables in cluding per cent conversions of monomers, use of short-stopping agents, antioxidants, and varying quantities and types of soap booster suction drum where the vapors are compressed by a steam jet from ap proximately 40 mm. Hg. absolute to 220 mm. Hg. absolute and discharged into a booster condenser where most of the re maining styrene and water are con and knowledge of latex, butadiene, and 11. Effect of all variables on quality densed. Liquid styrene and water ami styrene properties made it possible for of rubber butadiene vapor then enter the vacuum the committees to agree quickly on a two- suction drum and the vapors arc scut stage butadiene recovery process. In After the committee evaluated all to the vacuum pump suction line where this process the butadiene is flashed available data, the standard styrene the butadiene is recovered. The liquid from 110 F. latex at 3 lb./sq.in. gage stripping column was designed. It is 9 outlets of both the booster and vacuum and 220 mm. Hg. absolute pressure. ft in diameter, 49 ft. 6 in. high, contains suction drums are connected to the In August. 1942, little design and twelve perforated porcelain-enameled styrene decanter. The recovered styrene operational data were available on the No. 11 gage Armco iron trays having overflows from the decanter into an ac process of steam distilling styrene from in. holes on a J^-iu- triangular cumulator and on to the feed tanks, and Pope 410 Chemical fngineering progress ucc 041340 June, 1949 'j, . *>' :v;- =t . s' -iZy> ' . i :> - . - - - - . VENT TO ATMOSPHERE | VIA OIL ABSORBER wV< aoiene CO NDENSER VENT REFRIG. UNIT SURGE TANK '-W STYRENE CONDENSER STRIPPING COLUMN BOOSTER VET BOOSTER SUCTION AND VACUUM SUCTION DRUM VAPOR TRAP LATEX PETO J ll !f>RES! OTC Oi^TaEnMdP INOIC ATORS "o^ butadiene 1 ACCUMULATOR (LATEX *LASH HANK LATEX VACUUM PLASH TANK OESUPER- HEATED STEAM 'CONTROL LEVEL^ CONTROO i STTRENE [DECANTER tl-HG ^C_ TO BUTADIENE PEED TANK vacuum COMPRESSOR PUMP STRAME I VENTURI CENTRIFUGAL PUMPS DEC. WATER ID STYRENE PEED TANK he ks Us 2. Standard Plant Monomer Recovery Process. is then pumped to the tank farm for trolling the latex level and allowing the reuse. latex to enter the tank with a minimum Butadiene efficiency is kept -high by of turbulence would lessen the.foaming passing the vapors above the butadiene in the flash tanks and permit good strip receiver through a refrigerated vent ping at increased flow rates. To attain trap and then through an absorption these objectives, a large basket-type column where 90 to 95% of the remain strainer was installed in the latex line ing butadiene vapors are removed be to the first flash tank, and the throttling fore the noncondensables are vented to valves were replaced with diaphragm the atmosphere. The water which con motor controlled Saunders-type valves. denses with the styrene is decanted, but Tlie ball float mechanism was replaced still contains an appreciable (0104 to by a bubbler-type differential pressure 0.1%) quantity of styrene, The alyienc controller using butadiene vapor to h recovered by steam stripping the de transmit the static bead pressure of the canter water in a packed column which liquid level in the flash tank. In addi is tied into the vacunm system. tion, the sparge lines were removed, and a piece of pipe with a spoon-like end was Flash Tank Problems. The flash attached to the latex inlet so that the tanks are horizontal cylindrical vessels incoming latex would hit the side wall of 10,000-gal. capacity with inlets at of the vessel and run down into the both the bottom and the center of one liquid in the flash tank. As a result of cud. They were originally equipped with these changes foaming was reduced. steam sparge lines and a ball float me Thus, it was possible to increase the chanism that operated a throttling valve latex flow rate from 40 gal./min. to 60 on the inlet at the end of the Link. Diffi gaL/min. with automatic level control culties arose from coagulutn build-up on and to decrease the cleaning schedule the ball float, around the sparge lines, from every six weeks to approximately and in the throat of the inlet valve. It every four months. soon became apparent that the addition ' of steam was not necessary to obtain Calnmn Problems. Problems en good stripping, but that a means of con countered in the styrene stripping col- uinRS were usually of three basic types: (1) foam control, which directly af fected tlie rate at which latex could be stripped, (2) coagulation or prefloc build-up, which determined the length of time the column could be operated at 'economical stripping rates before it was taken out for service for cleaning, and (3) stripping efficiency, which was im portant both from the standpoint of cost and quality.. Considerable thought was given to the use of various materials as defoamers since they were successfully used in batch operations; however, sub sequent development work revealed that it is possible to minimize tlie foam prob lem without using them. Although the soap used in the preparation of CPS latex is subject to very rigid tests, it was noted that foaming varied to a greater or lesser extent when using soaps from various sources. Foaming was partially controlled by blending the soaps as they were used. It was also found that if the pH of the latex was maintained at 8.8 to 9.0, less foaming occurred than when latex of a higher pH was stripped. The rate of steam addition was found to have a direct bearing on the amount of foaming and considerable work was done to find an optimum rate that would Vol. 45, No. 6 CHEMICAL ENGINEERING PROGRESS ucc 041341 Page 411 give good stripping with an amount of . operation each of the plants experienced Prior to the discovery of nitrogen foam that could be tolerated at the de plugging of the vacuum and booster dioxide and sodium nitrite as agents for sired rate. This rate was found to be condensers, vapor lines, flash tanks, and the deactivation and inhibition of "pop 5.5 lb. of steam/Ib. of styrene removed. other equipment with a white "popcorn corn" polymer, considerable effort had The hop tray was removed from some like'' polymer. Once this polymer had been made to revise the vacuum system cohmons to allow more vapor space. In formed, some vessels and lines required so that the formation of "popcorn" other columns a solid top plate was in cleaning every six weeks to three polymer would be less critical in the stalled for weathering the latex at low months, thus reducing the operating ca operation of the recovery equipment. pressure to remove the remaining buta pacity of the plants, increasing main Since it was recognized that "popcorn" diene before the latex passed over the tenance costs, and reducing monomer formed more readily at high tempera perforations on a second tray where it efficiency. Considerable effort has been tures, the booster jet discharging at was raixed with steam. made to determine the nature of this 180-200 F. was suspected of being the PrdSoc formation which plugged the "popcorn" polymer and to prevent its source of "popcorn" formation. An in holes in the plates and limited the life formation. stallation was made whereby the booster of at -column between cleanings was It is formed first as small "seeds" jet was replaced by a centrifugal va greater on the top three plates and the which then greatly accelerate the growth cuum pump. This pump proved capable bottom two plates than throughout the of additional "popcorn." Growth activ of handling the vapor load and reduced rest of the column. It thus appeared ity is ascribed to two properties of the the temperature of the vapors entering that Ae flashing of large quantities of seed (1). (1) "The seed is a material the booster condenser; however, since styrone from the incoming latex and the which is insoluble in the monomer, but the seed was still present in the booster effect of superheated steam were con which swells therein, and (2) the active condenser, "popcorn" continued to form tributing to prefloc formation. It was seed contains peroxide groups. When at this point. Further experiments - found that by installing a water jacket these peroxides are decomposed, insol proved that both the vacuum and booster on the low pressure (15 lb./sq.in, gage) uble free radicals are formed. These [condensers were considerably overde steam fine feeding the'column, and also add to the monomer units and thus lead signed, thus permitting the removal of by laProducing a fine spray of water to the formation of `popcorn' polymer." le tube bundle from the vacuum con- into Ac steam, essentially desuper High temperature, the presence of fer lenser and eliminating the need for heated steam would enter the bottom of rous metals, water vapor, and oxygen [about one-third of the tubes in the the cwturon at approximately 150 F. are factors influencing the growth of booster condenser. Although these and ISO mm. Hg. absolute' pressure. this polymer. changes reduced maintenance and clean Further improvements were made by in' Strong caustic solutions and sodium' ing, they did not affect the growth of trodomng the steam at the bottom of the bisulfite have been used to decompose "popcorn." It is interesting tq nfl'" ihni column through a sparge ring which the active butadiene peroxides present in in one of the nlants. wunloyin^ a driw peiimttcd intimate mixing of the latex "popcorn," but their effectiveness can cdUdeHf^r*^tgr_st^tone in place of the with Ae hottest steam, thus tending to not be depended upon in plant equip stdllB5fd"Taccunico!uIenser. no jpoo- reduKE -prefloc. ment since theyarejjoteffectiveinthe 'Coril ^pOlymer was flound in this con rw--T** d*3* were made to control vapor phase anSfiCTnoralway^cnetratc denser arrer several years ot operation foaming and prefloc also had a direct sand pieces o{1lDopcorn." At tliT pres- altlllraj;ll II WAS present at otliet-ooints effect in improving the styrene efficiency. OITlime there are two methods of at in thS 'leUH'lll'y system. It has been sug- It was found that the best stripping was tacking the problem. The first method gestW-rtHT^TTFTeason for this is that obtained when the absolute pressure at is seed deactivation in which the growth this condenser has no stagnant areas tlie top of the column was 80 mm. Hg. activity of seed already present in plant such as are present in the horizontal . or less. - Removing foam baffles from equipment is destroyed or greatly re baffled tube bundles in the vacuum and ' the trays, increasing the range tubes in duced. The equipment Is first cleaned, booster condensers. the hoex flow controllers, and revising and then warm air is circulated to re tiie instrument leads between the latex move moisture. Nitrogen dioxide gas Seal Water System Problems. Since ventral tube and the flow controller (2) diluted with 20 or 30 volumes of air is ` the maintenance of low pressure is die .aided in more accurately controlling the circulated for several hours. In a single basis of all operations in the recovery operation of the column. Under normal treatment "popcorn" seed has been de system, it is natural that considerable operations, it is now possible to strip activated to such an extent that it takes thought be given to the operation and latex to an average styrene content of six months or more for the seed to form maintenance of the equipment involved, 0.04% compared with 02% when the again. This operation involves consid namely, the centrifugal vacuum pumps, colmrats were first put in operation. erable effort and expense and for these compressors, and their related water seal reasons it is not being used extensively system. The vacuum pump operates at Pmcm System Problems. Polymer- at the present time. The second method a suction pressure of 220 mm. Hg., re izatitei of styrene was recognized as- a is called seed inhibition. It is accom- ceiving butadiene vapors from the va factor in fouling of condensers and plishedjby continugu^ly^ajdinga small cuum flash tank and vacuum suction Other equipment at the outset of the! amount_oTsosd_li'um nitrite `to the n -j drum and discharging them at 3 Ib./sq. GRS program. Para tertiary butyl] fflasewherever "popcTMorn'1l1 -is lllikkd`1lyv~T in. gage and W F. to the compressor catechol was used as a polymerization occur. iFT!5STJ???rToundthat 0.2% t< suction. The compressor removes the inhibitor in fresh styrene, and in a short 0.5% sodium nitrite in the water set butadiene vapors from the pressure time, most plants were adding it to the system and 0.02% in the latex and i flash tank at 3 lb./sq.in. gage and com vapors entering the vacuum condensef the water condensing in the vacuum anc bines these vapors with those of the and booster condenser. Thus it wa booster condensers will effectively in-, vacuum pump discharge. Vapor at 60 present when the styrene was condense hibit the growth of polymer throughout lb./sq.in. gage and 130 F. is discharged and it remained in the recycle styrene! the major portion of the recovery sys from the compressor and sent to the during recovery and storage operations.] tem. This latter method has been used butadiene condenser for liquefaction. Formation of polystyrene in recovery' in a number of plants with considerable Operation of both the vacuum pump and equ^mient never reached the critical success either alone or following seed compressor is dependent on the use of stage, but after six to nine months of deactivation.. water to compress the vapors within the Page 412 CHEMICAL ENGINEERING PROGRESS ucc 041342 June, 1949 *11 lor >Pfld t nt. n" ra nt the in ter.. ta ble :ed ng ice ter rm nts ter ieof m--- :or :he tte mof tat r'P :he >P->n- * lAt tas tat ml ICC :hc ;ry Me nd m3, ps. sat at ,*e'aon 5q. or he ire nhc 60 ed he m. nd 4 49 pump body. Water used for this pur WKyvmAvniM pose is called seal water. A brief de scription of the original cyclic water rmt*c iicMfiR seal system (Sev Fig. 3) is as follows: Steam condensate from the styrene stripping column and booster ejector flows from the styrene decanter to the water overflow tank, and is then pumped through strainers to the vacuum pumps and compressors, A float in the water separator on the vacuum pomp controls the water level, and a similar float in the compressor water separator controls water discharge to the compressor water tank."The water is then pumped from the compressor water tank through a cooler and back to die styrene de dro P'-pQ ffi ffl 00: canter. A float in the compressor water tank controls the rate at which water is pumped through the cooler to the sty rene decanter. Excess wafer in the sys wwnmuiwninneun-wmw mwniamramve Fig. 3. Original Water Seal System. tem causes a. high level in the water overflow tank where a ball float actuates a valve in the discharge fine from the compressor water pump! sending the water tank. Seal water is used on the die styrene content of decanter water - water to the $ewer. Tempoatone con{rollers m"the water discharge lines of the-vacuum pump and conpreaor cir culate the water through the pumps until the proper temperatures are reached. / packing glands to prevent dilution of has been reduced from 0.15%-0.20% t< the sodium nitrite solutioS. As. a result ^ average of_0.04%. of these changes, and the use of sodium ^J nitrite, polymer formation has been ef Pump Problems. It had originally fectively inhibited so that the compres been felt that mechanical agitation of sor water tank, the fin tube cooler, and latex during pumping would cause pre- Operational difficulties ^countered m., the float valves in the pumps are free of floc to form and as a result an extended the water seal system were both numer polymer and require little maintenance search was made to find a pump that ous and varied. For simplicity, the more and cleaning. Loss of seal water from would produce a minimum amount of important of these problens will be dis the system is just about balanced by the cussed and an account given of the - condensation of styrene in the vacuum agitation and yet have sufficient ca pacity to handle the large volumes of methods now used to overcome diem. pump and compressor. The styrene col latex flowing through.the recovery sys Once seed had developed in the com pressor water tank, it became <nc of lects in the compressor water tank where it is periodically skimmed off and tem. Diaphragm pumps controlled by variable speed drives were used for this tiie largest sources of `'popcorn'' pdy- replaced with fresh sodium nitrite solu service. These pumps performed satis ~~ aier, requiring cleaning every two or ' tion. Since this system does not recir three weeks. Polymer also tended to culate seal water through the styrene form and collect in the fin tube water . decanter, the flow through this vessel cooler and in the water separators, has been reduced from 25-30 gaL/min. making the operation of the float con to less than 12 gal./min. The increased trolled water discharge valve erratic. time for decantation has improved the Difficulty was experienced with temper separation of styrene from water so that ature control on the vacuum pumps and factorily except that diaphragm break age was excessive, and the capacity of one pump was limited to approximately 35 gal./min. so that two pumps were used in parallel when flow rates greater than this were required. Also, with the introduction of nondiscoloring . compressors and in maintaining the pro per water pressure. .After one to two years' operation, wear and erosion of die rotors in the vacuum pomp and com pressor became apparent and gradually STTMM KCAKTM decreased the efficiency of butadiene re moval. A further source of monomer loss was due to the rdafively large amounts (0.1%-0.Z%) of styrene con tained in the water being discharged to wattA eoou the sewer. Replacement of eroded cast iron rotors wtih bronze rotors and the introduction of seal water to both sides of the compressor body without recir /C OMmwo* watia tamk 9N,4r%Tfln4rOt 0tIoUuIIfnmI culation (thus reducing heat build-up) have greatly improved tbc efficiency of TO MI.WRI (TRIfOtO compressor operation. The revised water seal system (Fig. 4) consists of a closed system whereby a solution of 0.5% sodium nitrite b pumped from the compressor water tank, through the cooler to the vacuum pump and com pressor and back to the compressor UNmnm ___ MATTtv*n FUMF FwtfgBTCII FltF Fig. 4, Revised Water Seal Syitem. Vol. 45, No. 6 CHEMICAL ENGINEERING PROGRESS UCC 041343 Page 413  * , - yspla.. ' larices, theomtain illation of latex with a foam trap, vacuum condenser, booster sity, and Chicago University. Data were toil which resulted from breaking c a- suction drum, booster ejector, booster obtained from minutes of the subcom phraguis became undesirable. Use M condenser, vaoauni suction drum, and mittee on monomer recovery, technical open impeller-type centrifugal pL line to vacuum pump suction header. committee, design committee, and re found to be an improvement in * The vacuum condenser is one section of ports prepared by the Rublier Reserve ^^ne pump could deliver 70 gal./nnt:, or the original three stage condenser. The Co,, Office of Rubber Reserve, and the ^^Iiore of latex without forming an ip- booster suction drum is a 24-in. piece of Office of the Rubber Director, preciablc quantity of prefloc. Centrifu 18-in. pipe mooted on the bottom of the Acknowledgment is made to D. G. gal pumps have required less mainten vacuum condenser shell. The booster Jamieson and G. R. Dance of the Can ance and the problem of contaminating ejector is motmted on the side of the adian Synthetic Rubber, Ltd., for sup nondiscoloring laticcs has l>ccn elim suction drum and discharges vertically plying details concerning revisions inated. down into tlie booster condenser. The made at that plant. Similarly, it was originally felt that booster condctncr consists of two 6-ft. Acknowledgment is also made to \V, positive displacement piston-type pumps vertical tube bodies eadi composed of H, Cain, W. E. Jones, W. E. Kaswell, would be required to handle recovered 1-in. tubes and is constructed so that the ' and J. E- Coincon of The Firestone Tire butadiene and styrene. It was found that vapors will be condensed inside the and Rubber Co. for their assistance and the small amount of water entrained in tubes. The conqwessor suction drum is suggestions. the recovered monomers tended to cause simply a 4-ft vertical section of 24-in. the valves to stick and that it was diffi pipe. To equafine the load on the va cult to-ferevont monomer leakage aroom! cuum pumps and compressors, the suc Literature Cited . die pistons. The capacity of these pump*.,gradually decreased as they be came .older 1 and maintenance became quite high. - It was found that centri tion and discharge lines each have been manifolded across the wings of the building so that all pumps and compres sors operate in parallel. 1. Kharasch, M. S., et al. The Labora tory Control of Popcorn Polymer-- ' CR950, report submitted to Office of Rubber Reserve Co. (1946). fugal pumps equipped with metallic sdals Changes deserted achieved the de could handle.both butadiene and styrene sired results in short stripping efficiency and (bos.eliminate the aforementioned was improved, production, was increased, - problems. Also, since centrifugal pumps cleaning and maintenance were reduced, discharge at a constant pressure, the and operation during the winter months breakage of rupture discs due to'-the caused less diffihdky. Front the changes flexing action of piston-type pumps-Has' fhathave^been made, it can be seen that 2. Pahl, W. H,, Chem. Eng. Progress, 43, 507 (1947). 3. Whitwell. T. C- and Wilhelm, R. H_ et al. Monomer Recovery Project, Princeton University. CD-309, re port submitted to Office of Rubber Reserve Co. (1944). been eliminated. considerable sua^ifkation is possible Comedian Plant Revisions. Due to its even in the plants where the effect of location geographically, the Canadian cold weather is leas critical. Discussion Standard plant located at Sarnia (OnL), lias some inherent peculiarities that sake recovery operations different in ^Bbie respects from the operations of ^Bnilar plants in the United States. Cooling water below 70 F. is available so that die need for refrigerated wafer is eliminated, and the operation of con densing equipment is improved. How ever, the low temperatures experienced .during tile winter greatly increase the difficulties due to frozen lines and gen erally retard all outdoor operations. The basic reasons for the changes made to the recovery system were (1) to reduce the hazard of peroxide accumulation, Conclusion Development wt the GR~S monomer recovery process was one small example of how Americas industry worked to gether in the war effort to develop, de sign, build, and ptace an entire industry in full production in less than two years. Of necessity, time-consuming re search and pilot plant work were held at a minimum; fundamental engineer ing, chemical knowledge and experience applied with spend and common sense succeeded in bolding the synthetic R. R. Bowles (California Research Corp., El Seguudo, Calif.): The buta diene producing plants in the Rubber Reserve program also have had difficulty with popcorn-polymer formation in dis tillation columns. In these plants sty rene probably is not present to the ex tent it is present in the copolymer plants. Has it been definitely established that popcorn-polymer contains styrene or that styrene must be present for its for mation ? C. R. Johnson: Varying amounts of styrene have been found in different samples of popcorn. (2) to improve latex stripping. (3) to minimize problems of winter operation, and (4) to increase production capacity approximately 40%. The recycle butadiene and styrene tanks, butadiene receivers and pumps, rubber industry. Changes described in this report have been adopted by most Standard plants although few plants have made all the specific changes. R. R. Bowles: The presence of sty rene then may not be absolutely neces sary? G. R. Johnson: I think you can get either polystyrene, polybutadiene, or co and refrigeration units were eliminated polymers of butadiene and styrene. to reduce "popcorn" and peroxide ac cumulation points. The butadiene is routed directly from the condenser sys tem to the tank farm recycle vessel. In addition, the Sarnia plant sulisti- tuted three styrene vacuum systems for the tiro included in the Standard plant design for three columns thus making it possible to shut down a vacuum system with a column for cleaning. When these vacuum system changes were made, Acknowledgment In preparing Ibis report information gathered by the subcommittee on mon omer recovery was used as well as that obtained from reports distributed within the industry by The B. F. Goodrich Co., The Goodyear Tire and Rubber Co., J. T. Cox: The question of popcoruing concerns the styrene producer, the butadiene producer and the copolymer producer. One butadiene plant has at tempted to cut down popcorn formation in its vessels by the use of sodium ni trite. It seems to be successful. I think the petro'eum companies producing butadiene can tell you on application of their experience in that particular line. large vapor line valves were excluded, revised vacuum system consists of The U. S. Rubber Co., The Firestone Tire and Rubber Co,, Princeton Univer (Presented at Eighth Regional Meeting, Cleveland, Ohio.) < I % r c he Tlong search, engineer a field 1 mgio ktt. wfcd fro i lb things > bvliave Hn-v as oi wer, ' ngir ill ut l tl, -c |HLI th ill -ft tr l< or tin d A Rcynd* of aqu< laws at Niindi method . of sett, tlncken The 11 * nt o ].< ndat tin: oaa <>f equ a 11 ribu iilsorp-Must i i tkav oi imp (i ere i I*. iria f.ctors ntearc The mindc as it c science its nat engine funtil engine ftjfl 414 ' CHEMICAL ENGINEERING PROGRESS UCC / 041344 June, 1949 ii Vol.