LOW TEMPERATURE C RUBBER BY A CONTINU PIILQT P ANT PLtQ Robert W. Lanndrie and Roger F. MeCann G o v e r n m e n t Laboratories, L'nicersity of Akron, Akron, Ohio
T h e Office of Rubber Reserve recently authorized the conversion of approximately one half the capacity of the operating GR-S copolymer plants to the production of polymers made a t low temperatures. Because many of the plants are now producing standard GR-S a t 117" F. by a continuous process, investigation of continuous polymerizations a t low temperature in the pilot plant of the Government Laboratories has been spurred considerably, so t h a t the basic operating techniques and pertinent data for some of the suitable polymerization formulas may be available for the design of additional equipment necessary for the operation of a continuous unit. Three formulations developed for the batchwise preparation of chemical rubber were utilized in pilot plant runs conducted a t 41' F. in a continuous polymerization unit consisting of twelve 20-gallon reactors connected in series: (1) a for-
T
mula containing about 3% of commercial dextrose; (2) a formula with approximately 1 % of~commercial dextrose, and (3) a formula without dextrose. Cumene hydroperoxide as the oxidant and ferrous iron salts as the activating medium were utilized i n each formulation with a monomer ratio of about 7 2 to 28 butadiene-styrene, by weight. Experimental lots (5000 to 10,000 pounds of polymer in each) w-ere distributed to rubber companies for evaluation and factory processing tests and for comparison in tires with standard GR-S and low temperature polymers made b y a batchwise process. Compared to batchw-isepolymerization, the preparation of chemical rubber a t 41" F. by a continuous process in the pilot plant revealed ease of control of temperature; smaller variations in conversion levels attained and &looney viscosities of !final polymers; and similar physical and chemical properties of the products.
of activator make-up, charging procedures, and operating techHE superior tensile strength, elongation, flex life, and resilience exhibited by low temperature polymers (%@, as niques have had to be developed for each formulation. The generalized formulations are shown in Table I. well as improved tread m a r in the tire tests conducted on the government fleet and those operated by the private companies, have caused the Office of Rubber Reserve to authorize the conversion of nearly one half of the capacity of the operating GR-S Table I. Activated Cumene Hydroperoxide-Redox Formulas for Emulsion Polymerization by a Continuous plants to the production of chemical rubber a t low temperatures. Process a t Low Temperatures Production a t the converted plants is expected to yield approxiFormula A B C mately 200,000 long tons of low temperature polymer per year. Description High-sugar Low-sugar, Sugar-free Because of the advantages of continuous polymerization in CHP-redo; CHP-redox CHP-redox Reaction temp., F 41 41 41 the production of the standard polymers, GR-S and GR-S-10, Parts such as increased productivity, steadiness of the load on the Ingredient Butadiene 71.5 72 71.5 monomer rpcovery and cooling systems, and uniformity of the Styrene 28 28.5 28.5 Modifier (tertiary mercaptan) 0.21 0.18 0.20 products, the Office of Rubber Reseive requested the GovernEinulsifier ment Laboratories to prepare sufficient quantities of 41 ' F. Rosin acid type 4.6 3.76 4.7 F a t t y acid type ... 0.94 ... redox polymer in their twelve 20-gallon reactor continuous system Cumene hydroperoxide 0.1-0.2 0.1 0.20 hctivatora 3.73 1 . G 1 0 .50 t o obtain operating specifications and conditions as well as polyElectrolyte 0.5 0.5 .. Water mers for factory evduation and tire tests. 180 180 200 \Then the work was started in 1948, it was not known vhether a Complered iron compound. the redox formulaq, which had been developed for a batchwise process, would be satisfactory in a continuous process. The first continuous low temperature polymerizations were made in Equipment accordance n-ith adjustments of the baeic formula used by the The continuous unit consists of twelve glass-lined, corli-insuCopolynyr Corpox atioii in the batch process. Ementially, this lated, 20-gallon reactors connected in series. X feed-stock cooler conqisted of a redox formula ( 5 ) a t 41 ' F. in which soluble iron or heater, depending on the temperature of polymerization, is coxnpoiinds, cuinene hydroperoxide, and about 3% of dextrose manually controlled 50 that the feed in the first reactor wili be at the desired temperature. The temperature in each set of three were utilized. For the sccond set of polymerizations, a lowreactors is regulated by one temperature recorder-controller and sugar (lyo) redox formula ( S ) , suggested by the staff of the B. F. the agitators in each reactor of the set are operated by a common Goodrich Chemical Company ( 1 ), was employed. Currently, drive. As the unit has been operated a t pressures of 20 to 80 polymerisations according to a sugar-free formula (2, S , 6 ) , pounds per square inch, inhibition of polymerization due to air is eliminated. Unlike the equipment in the production plants, no adapted by the United States Rubber Company and the Governdisplacement tubes are provided in the pilot plant installation, ment Laboratories from a formulation suggested by the Phillips The monomers and modifier are weighed separately, fed to a Petroleum Company, are in process. weigh tank, and mixed prior to being proportioned into the preT h e v basic formulas for batchwise polymerizations have had cooler. The catalyst is mixed with the soap solution 01' fed separately through a proportioning pump from a weigh tank into the to be adjusted for continuous polymrrizations and the conditions 1568
August 1949
INDUSTRIAL AND ENGINEERING CHEMISTRY
feed-stock cooler or directly into the first reactor. The emulsifier solution, which includes the water, emulsifiers, and electrolyte, is made up in quantities sufficient for 24 hours of operation and is transferred, as needed, to a weigh tank from which it is proportioned into the precooler. The monomer and emulsifier streams are mixed in a turbine-type pump with a by-pass to the suction side. The activator is either mixed with the soap solution or proportioned separately into the feed-stock cooler or the first reactor. Both the activator and catalyst may also be added under pressure, intermittently from a glass tube into the feed line at several places or into any of the reactors. The shortstop emulsion or solution is added to the latex on the downstream side of the airoperated diaphragm valve used to maintain a constant pressure on the unit. All feeds are calibrated periodically to check the operations of the pumps. The effluent is collected for 4- to 6-hour periods in a combination blowdown tank and stripper and is subsequently freed of unreacted styrene by either a batchwise or continuous steamstripping process. After addition of an antioxidant and blending of the latices from several collection periods, coagulation is carried out batchwise by the usual salt and acid technique employed for GR-S. The crumb is placed on trays and dried a t 126 t o 140"F. in a forced-draft oven. Refrigeration is supplied by a Worthington Freon 22 unit with a capacity of approximately 50 tons at 0 F. Inhibited aqueous calcium chloride (specific gravity, 1.28) is used as the secondary refrigerant and is circulated through each bank of three reactors. The flow of brine through the jackets is relatively high to obtain satisfactory temperature control in each reactor. O
Polymerization Formula A (High-Sugar Type). The first redox polymerizations conducted in the continuous unit were made in accordance with a high-sugar (about 3y0 of commercial dextrose based on the monomers) cumene hydroperoxide formulation a t 41 ' F. similar to that given in Table I. Inasmuch as polymerizations in 5-gallon reactors indicated that polymers of 60% conversion could be obtained in 14 to 15 hours with this formula, the initial continuous run was started on a similar cycle. However, it was necessary to increase the holdup time to 21 to 22 hours to obtain polymer of 58 to 60% conversion. Considerable effort and time were spent in adjusting the amounts of the ingredients, revising the equipment, and standardizing the operating techniques, especially with respect to preparation of an activator of uniform reactivity. The activity of each activator was checked by polymerizations in bottles. The most satisfactory activators were obtained under the following conditions: Preparation and storage of the activator in a closed vessel under a blanket of nitrogen Use of a reducing stabilizer Solution of the activator salts in 20 parts of water Aging of the activator for 15 minutes a t 212' F.
For most of the polymerizations with Formula A, the activator was premixed with the emulsifier solution and charged into the feed-stock cooler, whereas the catalyst was charged into the first reactor. These procedures were somewhat different from the optimum procedures for batch polymerizations. Approximately 12,000 pounds of polymer of 60% conversion were made in continuous operations up to 1180 hours per run. Formula B (Low-Sugar Type). The next project undertaken in the continuous unit a t 41" F. was in accordance with a lowsugar (1%) formulation (Formula B). The emulsifier was composed of 80% rosin acid soap and 20% fatty acid soap of the type used for GR-S-IO and GR-S, respectively. Although batchwise polymerizations required 11 to 12 hours in 5-gallon reactors to attain 60% conversion, a holdup time of 12 hours in the continuous unit resulted in a conversion level of about 50%. An increase of 1hour in the holdup time, calculated from the previous rate curve to produce 60'% conversion, increased the final conversion only slightly. T o attain 60% conversion, an additional displacement time of 4 hours was necessary. With either cycle the increase in conversion in each
1569
reactor was linear with time. These differences in activity indicate that the redox balance is upset somehow in the change from batchwise to the continuous process. As with the previous run, the conditions for preparation of the activator had to be adjusted and standardized, The results of bottle polymerizations, as well as the results obtained in the continuous unit, indicated a much greater reproducibility of the activator for Formula B compared to that of Formula A. Again, blanketing of the activator with nitrogen was beneficial. Most of the polymer that resulted from the run, nearly 1500 hours in duration, was prepared under the following conditions: Activator, catalyst, monomer, and emulsifier phases charged separately into the feed-stock cooler. An activator charge of 1.61 parts of salt and 12 parts of water per 100 parts of monomer. A displacement time of about 16 hours. About 5000 pounds of polymer were prepared and shipped to the B. F. Goodrich Company for factory and tire evaluation. Formula C (Sugar-Free Type). An exploratory run in accordance with a sugar-free formulation is being conducted at. present in the continuous polymerization unit a t 41 O I?.
Ease of Control of Temperature The limiting factor with regard to the capacity of most emulsion polymerization processes is the rate a t which the heat of polymerization can be removed from the equipment. The problem is not solely one of poor heat transfer coefficient or need for a sufficient temperature differential. The heat of polymerization must be removed and the temperature of the emulsion must be maintained as uniform as possible. Because there are changes in the viscosity of the latex during the polymerization process, the agitation and coolant requirements for good control of temperature will vary. As it is not feasible to change either the coolant or the agitation during the high viscosity periods of the batchwise reaction process, the capacity of this process is limited. I n q continuous polymerization in a series of reactors, considerable flexibility with regard to both the flow of coolant and the agitation is possible. The reactors in which the viscosity of the latex is high can be provided with more agitation and separate cooling systems. When the cooling requirements for several reactors are the same, they may be controlled by one instrument and one circulating cooling system. For these reasons, then, higher rates of polymerization can be controlled more readily in a continuous process than in a batch process.
Control of Final Viscosity and Hydrocarbon Conversien Because the catalyst in most low temperature polymerizations is not water-soluble, oil-soluble shortstopping agents, as well as water-soluble agents similar to those used for GR-S, have been employed in the pilot plant to ensure complete cessation of polymerization and to prevent a rise in the viscosity of the polymer, For any one set of operating conditions in the continuous polymerization unit, the variation in the average h a 1 conversion for extended operations has been well within ~ 3 and % the final raw Mooney viscosity of salt and acid coagulated polymer within *5 ML 4 points. This control of both viscosity and conversion is generally better (but within the same limits) than that obtained for low temperature polymerizations by a batchwise process in the pilot plant.
Chemical and Physical Tests The results of chemical tests conducted on the low temperature polymers were normal with respect to the ingredients charged and the conversion level attained. Because the conversion of the low temperature polymers is only 60%, the total soap in the polymer is higher than that in G R S l O (72 % conversion). Laboratory tests of mill-mixed stocks indicate t h a t the low temperature polymers are rated poorer in extrudibility by the Garvey
INDUSTRIAL AND ENGINEERING CHEMISTRY
1570 Table 11.
Laboratory Physical Data' C o m p o u n d i n g Recipe
Parts Compound
h
Ingredient Polymer EPC black Zinc oxide Sulfur Altax Stearic acid
100 40 5
2 3 1.5
Conipoun4 B 100 40 5 2 3
..
Test Conditions
Mooney viscosity, 4-minute reading large rotor, 212O F. Stress-strain tests, O R R specificatibn procedure, J a n u a r y 1, 1940, a t 77' a n d 212' F. Temperature rise, Goodrich flexometer; 143 lb./sq. inch load; 0.175-inch stroke; 1800r.p.m.: 30-minute test a t 212' F. Flex life DeRlattia flexometer; pierced specimen; d a t a reported as number of flex'uros t o produce 0.8-inch crack growth a t 312 cycles per minute a n d 212" F. Resilience Goodyear-Healy pendulum method; angle of inertia, 16'; Shore hArdness, 5-second reading with Shore T y p e A durometer. Quality index, ratio of flexures of test sample t o flexures of GR-S a t equal hysterimeter values. Formula Reaction temp., ' F. T y p e of polymerization Raw Xooney viscosity
66 1 I L 4
300% modulus, Ib./ sq. inch 1,200 Tensile strength, lb./ sq. inch 4,240 Elongation, % a t break 640 Set (10 min. after 11 break), % Flex life, flexures 10,000 Temperature rise, F. 54 4.3 Quality index Resilience, 7c 63 64 Shoi e hardness a
GR-S-10
Batch
Low-Sugar 41 ' Continuous Batch
fiO h l L 4
59 h l L 4 60 AIL 4
87 M L 4
High-Sugar 41 Continuous
-.
117
Continuous
Vol. 41, No. 8
die test thanis GR-$10. The stress-strain properties of low teniperature rubbers made in accordance with Formulas A and B by batchwise and continuous processes are summarized and compared with GR-S-10 in Table 11. The :tocks were compounded according to the standard recipe, specified by the Rubber Reserve for rosin rubbers. S o stearic acid was used in compounding the polymers prepared mith the mixed emulsifier. Except for the resilience values, which were determined on stocks cured a t 292" F. for 10 minutes more than the optimum, values are presented for stocks a t their optimum cures. The tensile strengths of bot'h the batchwise and continuous low- temperature polymers are similar; they are about 30% superior to that for GR-SI0 and are generally equivalent t o that, of nat,ural rubber. For the most part, the moduli of the low ternperature rubbers, polymerized either batchwise or continuously, were equal and only slightly higher than that exhibited by GR-S-10. The over-all flex life-heat rise balance, as expressed by the quality indexes, and t,he resilience a t 77 ' F. of the batchwise and continuously polymerized low temperature polymers are about equal but considerably superior to corresponding values for GR-S-10. The resilience at 212" F. was similar to that of GR-S-10. The low temperature stocks exhibited slightly higher Shore durometer hardness values than did GR-S-10.
1,200
1480
1290
1140
Literatnre Cited
4,010
4190
4150
3100
670
530
600
580
10 12,000
9 8000
10 9000
10 4000
(1) Goodrich Chemical Co., B. F., grivate communication. (2) I n d i a RubberWorld, 119 ( 2 ) ,228 (1948). (3) Office of Rubber Reserve, unpublished work. (4) Schulze, W. A . , Reynolds, W. B., Fryling, C. F., Sperberg, L. I ? . , arid Troyan, J. E., I n d i a Rubber W o r l d , 117 (6), 739 (1948). ( 5 ) Shearon, W. H.. Jr.,McKenzie, J. P., and Samuels, M . E., 1x1). ENG.CHEM.,40, 769-77 (1948), (6) Troyan, J. E., Rubber Age, 63 ( 5 ) , 585 (1948).
55
4.9 61 66
54 3 4 63 64
59 3.2 61 63
__
60 1.4 JD
5g
RECEIVED M a y 26. 1940. Investigations carried o u t under t h e s1)onsorsliig of the Reconstruction Finance Corporation, Office of Rubber Reserve, in connection with t h e government synthetic rubber program.
Values presented for stress-strain tests conducted a t 77' F.
EFFECTS OF POL IZ TEMPEIEtATIJRE ON STRUC A. W.Meyer United States Rubber C o m p a n y , Pussaic, A-. J .
T
HE polymerization of butadiene may be carried out by various mechanisms under a variety of conditions. I n this paper, primary attention is directed toward the emulsion polymerization of butadiene and of butadiene with styrene using free radical catalysts. The effect of temperature of polymerization on the structures of such polymers and copolymers is reviewed. Various investigators have established that emulsion polymerization takes place by a free radical mechanism (9, 28, %), which includes initiation, propagation, chain transfer, and termination steps. Polymerization ternperature influences these reactions and therefore the structure of the polymers. Butadiene polymerization leads to various structural configurations. I n addition to the investigation of configurations for polybutadiene shown in Figure 1, this paper describes the occurrence of branches or side chains as well as cross links which lead to gel. Furthermore, differences in average molecular weight and molecular weight distribution are considered. I n the case of copolymers of butadiene and styrene, consideration is given to differences in the ratios of butadiene and styrene in various samples of copolymer and in the distribution of butadiene and styrene in various molecules in a given sample of copolymer, as well as differences in the airangement of butadiene
and styrene unit,s with relation t o one another in a given polymer molecule.
E-nsaturation I n all the structures indicated in Figure 1, there is one carbonto-carbon double bond in each butadiene residue. Various investigations have been made of the unsaturation of butadiene polymers and copolymers. The reaction of iodine monochloride 15-ithpolymers waj used by Lee, Kolthoff, and hIairs (PS), whose evperiments were based on the procedure of Kemp and Peters (PO). Rehner (50)sho\ved that difficulties such as substitution or "splitting out" of acid may occur, but KoIthoff and his con-orlrers modified the method to give accurate deterininations oi' unsaturation of polybutadiene, polyisoprene, and their copolymers Jvith styrene. They found that emulsion polybutadiene made a t 50" C. had 97 to 98% of the theoretical unsaturation; emulsion polyisoprene 977,. The unsaturation of GR-Sa t various conversions corresponded to the spectrophotometrically determined styrene content; thus, it was concluded that GR-S a t various conversions had the theoretical unsaturation. Change in the temperature of polymerization from 10" t o