CA1238145A - Modified lewis acid catalyzed polymerization - Google Patents

Abstract

ABSTRACT OF THE DISCLOSURE
Protection against molecular weight depres-sion in the polymerization reaction producing polyiso-butylene rubber or butyl rubber is achieved through ad-dition of minor amounts of certain metal alkyls, particu-larly the lower alkyl aluminum compounds or lower alkyl aluminum hydrides which function as buffering, neutral-izing or complexing agents against polymerization poisons and extraneous catalyst activators.

Description

3L23~

BACKGROUND OF TIE INVE~ITIO?~

2 l. Field of the Invention

3 This invention relates to the polymerization of

4 the elastomeric homopolymer polyisobutylene rubber and the elastomeric copolymer bottle rubber, especially the polymer-6 ization reaction required to produce the isobutylene-iso-7 prone form of bottle rubber. More particularly, the invent 8 lion relates to a method of protecting the polymerization g process used to produce polyisobutylene and bottle rubbers against molecular weight depression and other undesirable if effects caused by the presence of polymerization poisons 12 and/or activators, through addition of a buffering neutral-13 icing or completing agent to the polymerization zone.
14 2 Desert lion of the Prior Art p A significant problem encountered in the bottle 16 rubber production process is the sensitivity of the reaction 17 system to the presence of polymerization poisons and gala-18 lust activators, which, even in trace quantities can sub-lug staunchly adversely affect the polymerization process by reacting or completing with the catalyst species or prop-21 grating carbonium ions to profoundly affect the initiation, 22 propagation, transfer, and termination steps of the polyp 23 merization. The consequences are observed as severe mole-24 ular weight depression in the bottle rubber product, active-lion or deactivation of the catalyst system, and often times 26 rapid warmup and plugging of the polymerization reactors.
27 The polymerization poisons and activators include Oxygen-28 axed compounds such as alcohols, kittens and ethers; acidic 29 compounds such as hydrogen chloride, chlorine, chlorine I containing organic compounds, and organic acids. Many of 31 these polymerization poisons or activators are produced in 32 the various zones of the bottle rubber production system 33 where water is present and result from hydrolytic reactions 34 involving methyl chloride and/or ethylene chloride. Some are produced by chemical reactions which take place in the I alumina drying step of the process which is used to remove .

I

1 water in the delineate recycle section of a bottle plant. Still 2 others are introduced as impurities present in the monomer 3 feed streams. Specific examples of polymerization poisons 4 and/or activators include methanol, isopropanol, acetone, dimethylether, diethylether, dimethoxymethane, t-butyl at-6 cool, t-butyl chloride, bis-chloromethylether, gaseous 7 chlorine, hydrogen chloride, prop ionic acid, and the like.
8 Recognition of this problem in the prior art is g found in U.S. Patent 3,005,008, issued October I 1961, to Kelly et at wherein polymerization poisons are removed as 11 one aspect of a process for treating the recycled delineate of 12 the bottle rubber polymerization process. So far as the 13 inventors hereof are aware, no effective technique has been 14 disclosed in the art for adding a completing or buffering agent to the polymerization reaction mixture in order to 16 eliminate the molecular weight depression and other pro-17 bless associated with the presence of the polymerization 18 poisons or activators noted above.
19 The use of aluminum alkyds to purify hydrocarbon delineates used in the polymerization of alpha-olefins is 21 disclosed in British Patent 920,513 (1963) issued to Eastman 22 Kodak Company. Similarly, in British Patent 1,298,909 23 (1972) issued to Monsanto Company, is disclosed the use of 24 trialkyl aluminum compounds as a scavenger for water, oxygen and trace amounts of alcohol and other impurities in con-26 section with a process for the Ziegler polymerization of 27 olefins such as ethylene. While some of the trialkyl 28 aluminum compounds used in this Monsanto British Patent are 29 employed in the present invention, the manner of their use in the reference and the polymerization process of the 31 reference are distinct from the present invention. Sims 32 ilarly, in U.S. Patent 3,257,332 issued June 2 , 1966 to 33 Ziegler et at, excess quantities of aluminum trialkyl are 34 used to eliminate impurities in feed stock ethylene employed 35 in an ethylene polymerization process. Also in U.S. Patent 36 3,~42,878 issued May 6, 1969 to Gippin, dihydrocarbon alum 37 minus chlorides are used to destroy impurities present in S

1 isoprene feed stock and in the hydrocarbon solvent used for 2 isoprene polymerization.
3 3. Summary of the Invention 4 In accordance with the present invention, there has been discovered an improved method of preparing C4-C7 6 isoolefin homopolymers such as polyisobutylene rubber by 7 Lewis Acid catalyzed polymerization or bottle rubber by the 8 Lewis Acid catalyzed polymerization of a C4-C7 isoolefin 9 with a C4-C14 conjugated dine in an inert delineate, come prosing adding to the polymerization reaction zone up to 11 about 1.0 to about 1.25 mole of a completing or buffering 12 agent per mole of catalyst in an amount effective to protect 13 the rubber product against molecular weight depression due 14 to the presence of polymerization poisons and catalyst activators, the completing or buffering agent comprising 16 certain metal alkyds, particularly a lower trialkyl alum 17 minus, the alkyd being methyl or ethyl, or a lower dialkyl 18 aluminum hydrides the alkyd in said hydrides being a Cluck 19 alkyd group. Other metal alkyds such as magnesium din-hey' have also been found to be useful in the practice of 21 this invention but are generally less effective and more 22 expensive than the preferred aluminum alkyds.
23 The particularly preferred completing or buffer-I in agent for use in the present invention has been found to be triethylaluminum. Other preferred completing or 26 buffering agents found suitable are trimethylaluminum and 27 diisobutylaluminum hydrides In a preferred embodiment, the 28 alkyd buffering or completing agent is added to the cold 29 monomer feed stream just prior to its introduction into the polymerization reaction zone.
31 A variety of substantial improvements accrue from 32 employment of the process of the present invention. The 33 polymerization reaction is protected effectively against 34 the harmful effects caused by sudden or unexpected variation in the level of the trace quantities of the polymerization 36 poisons noted above. The alkyd acts as a buffering agent or 37 completing agent which maintains the free concentration of 1 these poisons at constant, low levels. The problem of 2 molecular weight depression is essentially eliminated, en-3 ambling the desired molecular weight polymer to be produced 4 at a relatively lower steady-state unrequited monomer con-cent ration, which is a substantial economic benefit to the 6 process. The buffering agent of the present invention 7 permits higher tolerable levels of these polymerization 8 poisons to be present. The need for extensive purification 9 of feed and recycle monomer and delineate streams is sub-staunchly reduced. Also, startup of the polymerization system is greatly facilitated wince the need for excessive 12 initial catalyst flow to overcome poisons present is at-13 levitated. Operation with the alkyd buffering agent present 14 also results in a reduced rate of reactor warm -up as it polymerization proceeds.
16 4. Detailed Description of the Invention 17 The term "bottle rubber" as used in the specific 18 cation and claims means copolymers of C4-C7 isoolefins 19 and C4-C14 conjugated dines which comprise about 0.5 to about 15 mole percent conjugated dine and about 85 to 21 99.5 mole percent isoolefin. Illustrative examples of 22 the isoolefins used in the preparation of bottle rubber 23 are isobutylene, 2-methyl-1-propene, 3-methyl-1-butene, 24 4-methyl-1-pentene and beta-pinene. Illustrative examples of conjugated dines which may be used in the preparation 26 of bottle rubber are isoprene, butadiene, 2,3-dimethyl 27 butadiene, piperylene, 2,5-dimethylhexa-2,4-diene, cycle-28 pentadiene, cyclohexadiene and methylcyclopentadiene.
29 Preparation of bottle rubber is described in U.S. Patent 2,356,128, and is further described in an article by RUM.
31 Thomas et at in Industrial and Engineering Chemistry, 32 Volume 32, pages 1283 et seq., October, 1940. Bottle 33 rubber generally has a viscosity average molecular weight 34 from about 100,00 to about 800,000, preferably about 250,000 to about 600,000, and a Wits Iodine No. of about -1 0.5 to 50, preferably 1 to 20. The term isoolefin home-2 polymers as used herein is meant to encompass those home-3 polymers of C4-C7 isoolefins, particularly polyisobutylene, 4 which have a small degree of terminal unsaturation and

5 certain elastomeric properties.

6 The principal commercial form of bottle rubber

7 is isobutylene-isoprene bottle rubber and both bottle rubber

8 and polyisobutylene are produced in a low temperature g cat ionic polymerization process using Lewis Acid type 10 catalysts, typically aluminum chloride (Alec); it is 11 recognized that BF3 is also useful. The process extensively 12 used in industry employs methyl chloride as the delineate 13 for the reaction mixture at very low temperatures, that 14 is less than about -90C. Methyl chloride is used for a 15 variety of reasons, including the fact that it is a solvent 16 for the monomers and aluminum chloride catalyst and a 17 non-solvent for the polymer product. Also, methyl chloride 18 has suitable freezing and boiling points to permit respect 19 lively low temperature polymerization and effective swooper-20 lion from the polymer and unrequited monomers. Other 21 suitable delineates for use in the bottle rubber polymerization 22 process are ethylene chloride, vinyl chloride, ethyl 23 chloride, propane, butane, pontoon and hexane, but methyl 24 chloride is the particularly preferred polymerization I delineate.
26 In ?reparj.ng bottle rubber about 70 to ~9.5 parts 27 by weight, preferably 85 to 99.5, of an isoolefin such as 28 isobutylene is reacted with about 30 to 0.5 parts ox a con-29 jugated dine, such as isoprene. Particularly referred is 30 the reaction product of 95 to 99.5 percent by weight of i50-31 battalion, with 0.5 to 5 percent by weight of isoprene.
32 These monomers with about 1 to 5 volumes of inert delineate, 33 such as methyl chloride, are cooled to a temperature between I about -60C to -130C. The cold mixture is polymerized in 35 a reaction zone by the addition of a Friedel-Crafts catalyst, 36 preferably aluminum chloride. Generally about 0.02 to about 37 0.5 percent by weight of catalyst, based on the weight of I

1 idea offense, is used.
2 Roy -no sDecificat;on, eye no clair,ls 3 the term completing agent as used herewith is used inter-4 changeably with the term buffering agent. The inventors hereof do not mean to draw any distinction between these two 6 terms.
7 In practicing the present invention, a number of 8 conditions and limitations must be observed in order to g realize the benefits hereof. These relate to the quantity of completing or buffering agent used and the manner of its 11 addition to the reactor, that is, the control of injection 12 temperature and residence time prior to the entry of the I buffering event into the polymerization. reaction zoo-.
14 The quantity of buffering agent used is a function principally of the amount of catalyst employed. It has been 16 found that the amount of buffering agent should not exceed 17 about 1.0 to about 1.25 mole, preferably should not exceed 18 about 1.0 mole per mole of polymerization catalyst. If a 19 significant molar excess of buffering agent over catalyst is present, reaction of it with the aluminum chloride catalyst, 21 or the Lewis Acid catalyst, is likely to occur with very 22 sharp reductions in catalyst efficiency. Catalyst effi-23 Chinese is defined as the weight of polymer produced per 24 weight of catalyst. Through controlled experiments, it has been found that the minimum effective amount of buffering 26 agent is that which is in molar excess over the polymer-27 ization poisons present in the reaction mixture. Therefore, 28 in connection with the practice of the present invention, 29 the proper guideline is that the quantity of completing or buffering agent be not more than about 1.0 to about 1.25, 31 preferably not more than about 1.0 mole per mole of polymer-32 ization catalyst since the polymerization poisons when 33 present are always present in amounts which are signify-34 gently less than the amount of catalyst used. A substantial molar excess of buffering agent over polymerization poisons 36 is not harmful to the practice of the present invention.
37 Generally speaking, the amount of completing or buffering I

1 agent to be used it in the range of about 0.1 to 1 mole ratio 2 to 1 to 1 mole ratio; the preferred quantity is between 0.5 3 to 1 and 1 to 1 moles of buffering agent per mole of 4 polymerization catalyst such as aluminum chloride.
Location of the injection point of the buffering 6 agent and the injection temperature are also important 7 factors to the successful practice of the present invention.
8 If the buffering agent is added to the cold incoming monomer

9 feed stream containing monomers and inert delineate it must be

10 done so immediately prior to introduction of the feed stream

11 into the polymerization reaction mixture. When bottle

12 rubber is being produced, the monomer feed stream typically

13 contains a blend of the isoolefin and conjugated dolphin I monomers. Residence for any appreciable time in the feed 15 stream will result in undesirable side reactions with moo-16 men taking place, which will limit the effectiveness of the 17 buffering agent. A residence time in excess of several 18 minutes in the cold feed has been found to be undesirable in 19 many cases.
In one method of practicing the present invention 21 therefore, the buffering agent is injected into the polyp 22 merization zone such that the residence or contact time of 23 the buffering agent with the monomers prior to polymer-24 ization is generally less than about 60 seconds, preferably 25 less than about 30 seconds, most preferably less than about 26 10 seconds. Using this technique, one method of adding the 27 buffering agent to the polymerization zone would be to 28 introduce it into the reactor at the startup of the process, 29 prior to the addition of monomer(s) and catalyst to the 30 polymerization zone. The agent would then be continuously 31 introduced as polymerization proceeds. Alternatively, and 32 preferably, the agent is added directly to the cold monomer 33 feed stream immediately before the reactor, with due con-34 side ration of residence time. In another approach, the 35 buffering or complexir.g agent is present in the polymerize-36 lion zone prior to the introduction of monomer(s) and then 37 as monomers are introduced, the agent is added to the monomer 38 feed stream in an appropriate amount to achieve buffering or Lo i completing. Alternative equivalent methods will be Papa-2 rent to those skilled in the art based on a description of 3 the important variables as described herein, particularly 4 in the examples.
Preferred embodiments of bottle rubber to which the 6 invention is applicable are isobutylene-isoprene bottle rub-7 biers made using aluminum chloride catalyst and methyl Shelley-8 ride delineate, the product having a viscosity average mole-g cuter wt. of about 250,000 to 600,000 and a wits Iodine No 10 of about l to 20. The invention is further illustrated by 11 the following examples, which are not to be considered as 12 limitative of its scope.
I Procedure and Materials for Examples lull

14 A series of batch polymerizations were performed

15 in a nitrogen purged dry box to show the deleterious effect

16 of traces of an oxygenated base on bottle rubber molecular

17 weight and to show the effectiveness of the buffering agents

18 in accordance with the present invention. The entire series

19 of runs was made using allocates removed from a common feed

20 blend so that the feed was the same for each run in the

21 series; and the entire series was also made using allocates

22 removed from a common batch of catalyst solution so that the

23 catalyst was identical for each run in the series. The

24 various runs in the series differed then only in the van-

25 axles under investigation.

26 Polymerizations were run using standard Pyre

27 glass 500 ml 4-neck round bottom flasks with ball-jointed

28 necks. A standard bearing and ground glass stirrer shaft

29 were connected through the center neck and glass thermals

30 were connected through two of the side necks. One of the

31 thermals was for a thermometer to visually monitor react

32 lion temperature; the other for a thermocouple connected to

33 a recorder. A jacketed and cooled dropping funnel was

34 mounted over the third side neck. Measured amounts of

35 catalyst solution were allowed to drip slowly from the

36 dropping funnel into the reactor as desired. In making a

37 run, the following general procedure was used:

~3~5 g 1 1. An Alcott of the feed blend was weighed into 2 the reaction flask, which was then clamped into position, 3 immersed up to the neck in a cooling bath.
4 2. The stirrer shaft was connected to the stirrer 5 motor mounted above the bath and stirring was started. The 6 thermometer and thermocouple were placed into their rest 7 pective wells.
8 3. After reactor contents had reached desired S temperature, the desired poisons, activators or buffering I agents were introduced through a volumetric pipette be-11 fore the catalyst dropping funnel was mounted over the third 12 neck and clamped into position. The dropping funnel jacket 13 was filled with cold bath liquid and the the desired catalyst 14 solution was poured into the funnel.
4. After the desired residence time, catalyst 16 solution was allowed to drip slowly into the reactor until 17 the desired volume had been added. The catalyst addition 18 rate was controlled to maintain only a small temperature 19 difference between the bath and reactor so that polymer-20 ization temperature was carefully controlled.
21 S. When the desired amount of catalyst had been 22 added and the desired reaction time had elapsed, a measured 23 volume of cold methanol was added to the reactor as a quench.
24 The quenched reactor was then unclamped and removed from the 25 bath and placed into the dry box vestibule from which it was 26 removed and placed into a hood, where it was allowed to warm 27 and boil off unrequited monomers and methyl chloride. Is-28 propel alcohol containing 1 percent PX-441 as an inhibitor 29 was added to the reactor flask as the volatile boiled off 30 to keep the polymer from discoloring and degrading due to 31 catalyst residues. After it had warmed to room temperature, 32 the polymer was quantitatively recovered. It was kneaded in 33 isopropyl alcohol to remove catalyst residues and then 0.2 34 percent PX-441 was mixed into it before drying in a vacuum 35 oven at 70C. The dried polymer was weighed and subjected 36 to analyses.
37 Isobutylene employed was purified by scrubbing

38 through an hydrous calcium chloride and an hydrous barium I

1 oxide to remove any moisture and distilled just prior to use 2 through a small fractionating column located in the dry box.
3 The methyl chloride used was Annul high purity grade ox-4 twined from Air Products Speciality Gas Division. Isoprene 5 used was dried over white Drierite and freshly distilled 6 through a fractionating column before use. The catalyst 7 solution used in the batch polymerization runs was aluminum 8 chloride dissolved in methyl chloride. A saturated solution n containing 0.9 grams Alec per 100 grams of solution was I prepared by refluxing methyl chloride over Alec and this 11 was diluted to the desired concentration with freshly scrub-12 bed and filtered methyl chloride.
13 A batch of feed for a series of runs was prepared 14 by weighing the desired amount of cold isoprene isobutylene 15 and methyl chloride into a cold feed flask and then mixing 16 and storing in a stopper Ed flask in a cold bath. Allocates 17 of this feed blend were weighed into the individual reactors 18 for the series of runs. In each run of Example 1 set forth 19 below, and Examples thereafter, 230 grams of a feed blend 20 consisting of 0.75 grams isoprene, 24.25 grams isobutylene 21 and 205 grams methyl chloride were charged to the reactor.
22 The reactor was then stirred and cooled as described above 23 to -140F, the indicated amount of polymerization poison 24 and/or stabilizer or buffering agent was added and the 25 catalyst solution consisting of 0.18 grams Alec per 100 26 grams methyl chloride was allowed to drip in slowly to 27 initiate polymerization. Over the course of about 10 28 minutes enough catalyst was added to convert about 25 29 percent of the monomer to polymer while maintaining reaction 30 temperature at -140F. The reaction was quenched with moth-31 anon and the polymer recovered and dried as outlined above.
32 In the following Examples, "INOPO" is a measure of 33 the degree of unsaturation, also known as the Iodine-34 Mercuric Acetate Method as reported in In. Erg. Chum 40, 35 1277 (1948). The abbreviation TEAL is used for tri-ethyl 36 aluminum. MY is the viscosity average molecular weight of 37 the polymers, based on a dilute solution viscosity meat I surmount in diisobutylene at 68F.

3~5 2 Run Number ADDED TEAL ADDED BOTTLE ANALYSIS
3 ivy INOPO
4 1 None None 971,500 7.7 2 12.6 ppm None 486,500 6.7 6 3 12.5 ppm 58.1 ppm 941,000 7.3 7 4 None 58.1 ppm 962,10rl I
None None 945,00C 7.3 3 The results clearly show that addition of 12.6 ppm of methanol strongly depresses polymer molecular weight from 11 more than 900,000 to less than 500,000; whereas addition of 12 58.1 ppm of tri-ethyl aluminum (TEAL) as a buffering or 13 complexion agent has essentially no effect on molecular 14 weight by itself, but that its presence prevents the moth-anon from causing a molecular weight depression (i.e., the 16 molecular weight of run 3 of the series with both methanol 17 and TEAL present is the same as that of the controls, 1 or 18 5, with no methanol present).

20 Run Number ACETONE ADDED TEAL ADDED BOTTLE ANALYSIS

22 1 None None 705,000 7.5 23 2 24.5 ppm None 334,000 8.1 24 3 None 81.3 ppm 835,000 7.8 4 24.5 ppm 81.3 ppm 765,000 7.1 26 5 None None 725,000 7.6 27 The results show that addition of 24.5 ppm acetone 28 strongly depresses molecular weight from more than 700,000 29 to 334,000; addition of TEAL alone at 81.3 ppm causes a small 30 molecular weight increase, and that addition of acetone to 31 the buffered system containing TEAL has little effect on 32 molecular weight. TEAL is able to counteract or prevent the 33 molecular weight depression caused by acetone just as it was 34 effective with methanol in example 1. TEAL is an effective buffer with acetone as it is with methanol.

iota .

2 Run No. DIM ETHYL TORY ADDED TEAL ADDED BOTTLE ANALYSIS
_ _ 4 1 None None 705,000 7.5 2 10.7 ppm None 457,900 8.3 6 3 None 81.3 835,000 7.8 7 4 10.7 ppm 40.7 ppm 705,100 7.8 8 5 None None 725,000 7.6 9 These results show that dim ethyl ether is a severe molecular weight poison when TEAL is not present but that its 11 effect is nullified in the buffered system containing TEAL.

_ 13 A series of batch polymerization runs was made 14 following the procedures outlined above using tertiary bottle alcohol as the poison with the following results:
16 Run No. t-BuOH ADDED TEAL _ DYED BOTTLE ANALYSES
17 ~vINOPO
18 1 None None 830,000 7.9 19 2 25.1 ppm None 650,000 8.3 3 None 65.0 ppm 7.8 21 4 25.1 ppm 65.0 ppm 7.6 22 5 None None 800,000 8.0 23 The results show that t-butanol is a strong mole-24 cuter weight depressant with no TEAL present but that it has little effect in a buffered system containing TEAL.
26 Examples 1 to 4 show that when an effective buff 27 firing agent like TEAL is used at proper concentrations 28 relative to catalyst and poisons, and is added properly, it 29 is able to prevent all classes of oxygenated bases from adversely affecting the bottle polymerization. In these 31 examples the TEAL was added at a molar excess to the poison 32 but at less than 1 mole per mole of Alec catalyst. The TEAL
33 was added to the cold feed blend at -140F and allowed to 34 complex with the poison before catalyst was added to in-shut polymerization. Only a short residence time (several 36 minutes) was allowed between complex formation and in-Al shoeshine of polymerization.

~3~5 2 This example is provided to demonstrate that the 3 buffering agent is not effective against molecular weight 4 depressants unless it is present in a molar excess in relation to the amount of polymerization poison in the 6 system.
7 RUN No. Mesh ADDED TEAL ADDED MOLE TEAL/ BOTTLE ANALYSES
8 My ye TV INOPO
9 1 None None - 848,500 7.8 2 12.6 ppm None - 469,400 7.6 11 3 12.6 ppm 8.2 ppm 0.18 381,700 7.3 12 4 12.6 ppm 16.5 ppm 0.37 442,000 7.6 13 5 12.6 ppm 41.0 ppm 0.87 863,600 7.3 14 6 12.6 ppm 57.4 ppm 1.28 883,800 7.8 7 12.6 ppm 121.7 ppm 2.75 924,000 7.7 16 8 None 123.5 ppm - 843,500 7.8 17 9 None None - 810,000 7.4 18 The data show that TEAL is ineffective in prevent-19 in the molecular weight depression caused by methanol unless there are about as many moles of TEAL present as there 21 are moles of methanol. Excess TEAL is not harmful provided 22 (as shown in later Example 7) its molar ratio to Alec is not 23 too high. Apparently TEAL is effective at a level of about 24 1/1 molar relative to the polymerization poison.

26 The data in these experiments show that the buff 27 firing or completing agent, in this case TEAL, is not 28 effective if the residence or storage time prior to in-29 tinting polymerization in the polymerization reactor is too long or if the residence time is at too high a temperature.
31 RUN No eye ADDED TEAL ADDED STORAGE TIME MY INOPO
33 1 None None None 877,000 7.8 34 2 12.5 ppm None None 576,000 8.4 3 12.5 ppm None 3 1/2 his. 612,000 8.7 36 4 12.5 ppm 58 ppm None 864,000 9.0 37 5 12.5 ppm 58 ppm 3 1/2 his. 414,000 8.4 I

1 These data show that methanol is a strong Milwaukee-2 far weight depressant without TEAL present and that storing 3 the feed with only methanol present has no further effect.
4 The data from run 4 show that methanol has no adverse effect in the buffered system containing TEAL, but the 6 data from run 5 show that TEAL completely loses its buffering 7 ability when the feed containing methanol and TEAL is 8 stored for 3 1/2 hours at -70C before adding catalyst to initiate polymerization.

RUN No. Lowe ADDED TEAL ADDED STORAGE TIME MY INOPo 11 6 None None None 803,000 8.1 12 7 12.7 ppm None None 555,000 8.0 13 8 12.7 ppm 58 ppm None 925,000 7.8 14 9 12.7 ppm 58 ppm 1/2hr @20C 459,000 8.1 15 10 12.7 ppm 58 ppm his @-70C 395,000 7.8 16 11 12.7 ppm 58 ppm ho @ -95C 361,000 8.5 17 These data again show that methanol is a strong 18 molecular weight depressant but that its adverse effects can 19 be completely prevented in the buffered system containing TEAL. however, molecular weight is severely depressed if 21 the buffered system is stored at 20C, -70C or even -95C
22 before catalyst is added to cause polymerization. These data 23 show that TEAL forms an effective buffered system only when 24 it is added to the cold feed immediately before polymerize-lion is started.

27 13 None None None 700,000 7.8 28 14 24.5 ppm None None 334,000 8.3 29 15 None 81.3 ppm None 836,000 7.8 30 16 24.5 ppm 81.3 ppm None 756,000 7.8 31 17 24.5 ppm 81.3 ppm 1 ho ~-100C 255,000 8.4 32 These data show that acetone is a strong molecular 33 weight depressant but that it has no adverse effect in a 34 buffered system containing TEAL. However, molecular weight in the buffered system is severely depressed when 36 the feed blend containing acetone plus TEAL is stored 1 37 hour at -100C before polymerization.

I

1 RUN NO. ACETONE ADDED TEAL ADDED STORAGE TIME MY INOPo 2 18 None None None ~30,000 7.6 3 19 24.9 Pam None None 453,000 8.3 4 20 24.9 ppm 81.3 ppm None 302,000 7.6 S 21 24.9 ppm 81.3 ppm 5hrs @20C 673,000 7.5 6 22 24.9 ppm 81.3 ppm lhr @ -99C 680,000 US
7 These data again show that acetone is a severe 8 molecular wright depressant but that its adverse effect is 9 completely negated in the buffered system containing TEAL.
However, if the buffered system is stored as in runs 21 and 11 22 prior to polymerization then TEAL Loses effectiveness and 12 molecular weight is depressed.
13 RUN NO. POISON ADDED* TEAL ADDED** STORAGE TIME MY INOPO
14 23 0 None None 662,000 7.0 24 12 ppm None None 439,000 7.1 16 25 12 ppm 1.2 None 640,000 6.8 17 26 12 ppm 1.21 ho @ ~97 405,000 6.7 18 27 12 ppm 1.2 None*** 646,000 7.4 19 * Poison was 4 ppm each of methanol, acetone, and dim ethyl ether.
21 ** TEAL was added at 1.2 moles per mole of total poisons 22 added.
23 *** No storage time prior to polymerization but 15 min.
24 residence time @-95C after initiation and before completion of polymerization.
26 These data show that molecular weight was severely 27 depressed by adding the mixed poisons to the unbuffered 28 feed (Run 24) but unaffected in the buffered feed (Run 25).
29 Storage of the buffered feed prior to initiating polymerize-lion caused the TEAL to lose its effectiveness (Run 26); but 31 allowing residence time in the reactor after polymerization 32 had been started (as in run 27) had no adverse effect on 33 buffering ability. These data show that TEAL is an effective 34 buffering agent when properly used.

The data from all these experiments show that 36 TEAL must be used properly in order to be an effective 37 buffering agent. It is effective only when added to cold 38 feed immediately before polymerization. It loses effective-

39 news if the feed is warm or is stored with TEAL present I prior to polymerization.

I

1 Example 7 2 This example is provided to demonstrate that 3 the molar ratio of buffering agent to catalyst affects 4 both polymer molecular weight and catalyst efficiency at a level significantly greater than about 1:1. Here TEAL
6 and Alec were used.

8 No. TEAL/Alcl~ WITH TEAL vs. CONTROL LUST EFFICIENCY
9 WITH NO TEAL WITH TEAL vs.
CONTROL WITH NO

12 1 0.16 1.04 1.0 13 2 0.20 1.11 0.9 14 3 0.42 1.01 0.8 4 0.62 1.08 0.8 16 5 1.05 1.13 0.7 17 6 2.01 0 09 0.15 18 The data show that addition of TEAL to pure 19 feed (as free of oxygenated impurities as possible) has only a mild beneficial effect on molecular weight and a 21 mild negative effect on catalyst efficiency until sign-22 ficantly more than one mole of TEAL per mole of Alec is 23 added, then molecular weight and catalyst efficiency 24 fall off rapidly. In this example a 5% molar excess had not yet reduced molecular weight, but caused a 30% reduce 26 lion in catalyst efficiency. At a 2/1 molar ratio of 27 TEAL/AlC13 molecular weight is less than 10% of the control 28 molecular weight and catalyst efficiency has been reduced 29 by nearly a factor of six. The operating window over which TEAL is effective as a buffer is defined by the 31 amount of poisons present and the amount of Alec catalyst 32 used. The buffering agent becomes effective at some 33 lower level where the amount added on a molar basis just 34 exceeds the total amount of poisons present and it ceases to be effective at some upper level where the amount 36 added on a molar basis significantly exceeds the amount 37 of Alec catalyst used.

Lo 2 This example is included for the purposes of 3 comparison and demonstrates the criticality of the buffering 4 agents claimed in accordance with this invention. In this example, tri-isobutyl aluminum (TUBAL) was found to 6 be ineffective as a buffering agent. The poisons added 7 here were t-butanol (TBOH) and methanol.
RATIO

10 No. MOLE TBOH
11 1 None None - 487,000 8.3 12 2 26 ppm None - 249,000 8.4 13 3 52 ppm None - 226,000 8.4 14 4 26 ppm ll0 ppm 1.6 240,000 8.7 52 ppm 220 ppm 1.6 244,000 9.0 16 These data show that t-butanol causes a severe molecular 17 weight depression whether or not TUBAL is present.
18 Results of a series of runs with methanol as 19 the poison and TUBAL evaluated as a buffering agent are shown below:

22 RUN Mesh ADDED TUBAL ADDED MOLE TUBAL/ MY INOPO
23 No. MOLE Mesh 24 6 None None - 908,000 7.9 7 None 111 ppm - 619,000 8.4 26 8 9 ppm None - 522,000 8.2 27 9 9 ppm ill ppm 1.0 547,000 8.3 28 Methanol causes a severe molecular weight depress 29 soon whether or not TUBAL is present; and in fact the to 30 data from run show that addition of TUBAL by itself 31 causes a molecular weight depression. The effect of 32 TUBAL alone is also quite different from the effect of 33 TEAL alone as shown in the data of example 7. TEAL alone 34 tends to raise molecular weight whereas TUBAL alone depress-en it.

1 Example 9 2 This example is also included for the purpose 3 of comparison and further demonstrates the criticality of 4 the buffering agents claimed in accordance with this invention. In this example, deathly aluminum chloride 6 (DEAR) was found ineffective as a buffering agent despite 7 its similarity to the buffering agents claimed in accord 8 dance with the present invention. TBOH is t-butanol; DYE
9 is dim ethyl ether. RATIO
11 HUN POISON (BASE) BUFFERING MOLES DEAR/ MY
12 No. AGENT MOLE BASE
13 1 None None Control Run 836,000 14 2 None 82 ppm DEAR Buffer only 944,000 15 3 25 ppm TBOH None Poison only 481,000 16 4 25 ppm TBOH DEAR 1.5 Moles/Mole 436,000 17 5 11 ppm DYE None Poison only 593,000 18 6 11 ppm DYE DEAR 1.4 Moles/Mole 339,000 19 7 25 ppm Acetone None Poison only 334,000 20 8 25 ppm Acetone DEAR 1.1 Moles/Mole 408,000 21 These data show that DEAR by itself does behave 22 similarly to TEAL in that it has a mild beneficial effect 23 on polymer molecular weight but that DEAR is completely 24 ineffective as a buffering agent with various oxygenated bases. Acetone, dim ethyl ether, and t-butanol all severely 26 depress bottle molecular weight whether or not DEAR is 27 present as a buffering agent. Thus DEAR, like TUBAL, is 28 not suitable as a buffering agent during bottle polymerize-29 lion.

31 This example demonstrates the utility of trim ethyl 32 aluminum (TEAL) as a buffering agent in accordance with 33 the present invention.

I

3 No. MOLE POISON
4 1 None None Control 891,000 8.1 2 None 41.6 ppm TEAL only 969,000 7.8 6 3 12.8 ppm Mesh None Mesh only 763,000 8.1 7 4 12.8 ppm Mesh 58 ppm 2.03 967,000 7.7 8 5 24.9 ppm None Acetone 509,000 7.6 g Acetone Only 10 6 24.9 ppm 56 ppm 1.84 790,000 7-4 11 Acetone 12 These data show that methanol or acetone severely 13 depress polymer molecular weight when added to the control 14 polymerization but that the molecular weight depression is largely prevented when these same poisons are added to 16 buffered feed containing TEAL. TEAL is an effective 17 buffering agent when added to cold bottle feed immediately 18 before polymerization.
19 Another series of batch polymerizations similar to those of example 1 were run with dim ethyl ether as the 21 poison and TEAL as the buffering agent. Some of the 22 buffered feeds were also stored prior to initiating polyp 23 merization as in example 6 to determine whether TEAL
24 loses effectiveness if feeds containing it are stored prior to polymerization.

28 No. MOLE DYE TIME
29 7 None None Control None 589,000 8.2 30 8 34.5 ppm None DYE only None 315,000 7.9 31 9 34.5 ppm 78.7 ppm 1.45 None 621,000 7.9 32 10 34.5 ppm 78.7 ppm 1.45 4 ho @ 424,000 8.3 34 These data show that addition of dim ethyl ether to the control polymerization severely depresses polymer 36 molecular weight but that molecular weight depression is 37 prevented when dim ethyl ether is added to buffered feed 38 containing TEAL. However, if the buffered feed is stored 39 for significant periods prior to polymerization, as in run 8, then TEAL loses effectiveness as a buffering agent 41 and is no longer able to prevent dim ethyl ether from 42 depressing bottle molecular weight. These data show that 1 TEAL is an effective buffering agent for bottle polymerize-2 lion, but that like TEAL, it must be used in the proper 3 way in order to be effective.

This example demonstrates the utility of dyes-6 bottle aluminum hydrides (DIABLO) as a buffering agent in 7 accordance with the present invention. RATIO
g RUN POISON ADDED BUFFERING AGENT MOLE BUFFER/ MY INOPO
10 No. ADDED MOLE POISON
11 none None Control 635,000 8.6 12 none 87.4 ppm DIAL H DIAL H only 8.6 13 324.2 ppm None Acetone 424,000 8.2 14 Acetone Only 15 424.2 ppm 87.4 ppm DIAL H 1.48 671,000 8.6 16 Acetone 17 524.2 ppm 63.4 ppm TEAL 1.33 668,000 7.8 18 Acetone 19 612.3 ppm None Methanol 487, boo 8.5 methanol Only 21 712.3 ppm 87.4 ppm DIAL H 1.60 860,000 8.1 methanol 23 These data show that methanol or acetone severely 24 depress bottle molecular weight when added to the control 25 polymerization but that the molecular weight depression 26 is prevented when these same poisons are added to buffered 27 feed containing a molar excess of DIAL H or TEAL. DIAL
28 H in run 4 is as effective in preventing molecular weight 29 loss caused by acetone as is TEAL in run 5. The data 30 also show that DIAL H alone (run 2) tends to raise bottle 31 molecular weight as compared to the control rather than 32 depress it. Both in terms of buffering effectiveness and 33 molecular weight effect, ~IBAL H behaves much more like 34 TEAL than tri-isobutyl aluminum.
I RATIO
36 RUN t-BUTANOL DIAL H MOLES DIAL H/ TV INOPO
37 No. ADDED ADDED MOLES t-BUTANOL
38 8 None None Control 745,000 8.5 39 9 23 ppm None t-butanol only 237,000 9.0

40 10 23 ppm 66.2 ppm 1.50 75S,000 8.1 ~3~5 1 These data show that t-butanol severely depresses 2 molecular weight when added to the control but has no effect 3 when added to buffered feed containing a molar excess of DIAL H. DIAL H is indicated to be an effective buffering agent to prevent the harmful effects of t-butanol on bottle 6 polymerization.
8 RUN DIM ETHYL DIAL H Malaysia B / MY INOPO
g No. ETHER ADDED MOLE ETHER
ADDED
11 11 None None Control 589,000 8.2 12 12 34.5 ppm None Ether only 315,000 7.9 13 13 34.5 ppm 15.2 ppm 1.43 678,000 7.1 14 These data show that DIAL H is an effective buffering agent to prevent the harmful effects of dim ethyl 16 ether on bottle polymerization.
17 All o f these data show that DIAL H, like TEAL
18 and TEAL, is an effective buffering agent for bottle polyp 19 merization when used in a proper manner. These buffering agents counteract or prevent the adverse effects which 21 "poisons" have on bottle polymerization. Their use results 22 in an improved polymerization process.

24 In order to confirm the advantages of this invent lion under actual plant conditions, a series of tests were 26 run in a large scale bottle reactor to show that the dole-27 tedious effects of traces of poisons and/or activators in 28 the bottle feed could be negated by the appropriate use of 29 completing or buffering agents. These tests were run using a conventional large scale bottle reactor which had been 31 equipped with auxiliary equipment to permit injecting con-32 trolled and measured amounts of poisons, activators, and 33 buffering agents into the reactor feed. The reactor was of 34 tune conventional draft tube design. Reactor contents were circulated by a bottom entering propeller pump centered in 36 the draft tube. The reactor was cooled by circulating 37 liquid ethylene to the jackets where it vaporized to 38 provide refrigeration. Reactor temperature was maintained Lo 1 at the lowest level the facilities would permit and warmed 2 slowly during a run as the reactor fouled. Warmup rate 3 is a measure of the fouling rate. All tests were made at 4 a constant feed rate with a residence time of about 40 minutes. The reactor effluent overflowed into a flash 6 tank to which steam and hot water were added for recovery 7 of the bottle rubber as a slurry in water. Unrequited moo-8 mews were flashed overhead for recycle.
9 The tests were carried out by establishing steady state conditions and then beginning continuous injection 11 of the desired poison/activator and/or buffering agent at a 12 controlled and measured rate. After a new "test" steady 13 state was established, the effects of the injected agents 14 were determined by comparing the steady state reactor con-dictions before, during, and after the test injection.
16 Tests were usually run at constant monomer conversion by 17 adjusting catalyst rate to correct for the effect of the 18 injected agent on catalyst activity, but all other pane-19 meters were maintained as constant as possible during a series of tests so that the effects of the injected agent 21 could be clearly ascertained. Control reactors were being 22 operated simultaneously with the test reactor so that 23 perturbations caused by uncontrolled variations in the 24 plant could be observed and not erroneously ascribed to the injected agent.
26 The feed chilling facilities included two stages 27 of ethylene refrigerated heat exchangers. The auxiliary 28 facilities for injection of agents were built so that the 29 poison or activator was injected into the reactor feed at the inlet of the first ethylene feed chiller whereas 31 the buffering agent was injected downstream of this point 32 at the inlet of the second ethylene feed chiller. This was 33 done in order to conform to conditions which the laboratory 34 glassware studies had shown were necessary in order to achieve effective buffering. In this way, the buffering 36 agent was injected into the cold "poisoned" feed with a 37 contact time of only several seconds before entering the I

1 reactor where polymerization was being continuously carried 2 on.
3 Data from a test series which show the adverse 4 effects of injecting methanol into the reactor feed and then the negating of these effects by injecting TEAL downstream 6 of the methanol are summarized below. Feed rate to the 7 reactor was 6655 Kg/hr. of a feed containing 31.5~ by weight 8 isobutylene and 0.82% by weight isoprene in methyl chloride.
9 During the test period methanol was injected into the feed at 0.129 kg/hr. = 4.03 moles/hr. = 19 ppm by weight on feed.
11 During the TEAL injection period, TEAL was injected down-12 stream of the methanol at 0.854 Kg/hr. = 7.5 moles/hr. =
13 128 ppm by weight on feed. The molar ratio of TEAL to 14 methanol during the test was thus 1.86. Data from the various steady-state periods of the test series are summarized below.
16 Reactor Index is defined as Bottle product Mooney viscosity 17 divided by the isobutylene content of the reactor liquid 18 phase.

EUPHORIA TEST + Mesh Moe AFTER
21 & TEAL TEST
22 Mooney Viscosity 60 49 60 60 23 of Bottle 24 Reactor Index 11.7 9.8 10.7 10.5 25 unrequited Isobutylene, 5.1 5.0 5.6 5.7 26 Isobutylene Conversion 87.3 87.3 86.0 85.8 27 Alec Rate, Kg/hr. 1.374 1.642 1.79 1.25 28 Mole/hr. 10.3 12.3 13.4 9.36 29 Alec Efficiency, 1330 1115 1003 1439 31 Warm-up Rate,C/hr. 0.12 0.085 0.006 0.14 32 In this test series the reactor was brought to 33 a steady-state operation and operating conditions were meat 34 surged and recorded; then methanol injection was started 35 and catalyst rate was adjusted to try to maintain conversion 36 of isobutylene approximately constant. The reactor was 3 Lo I
1 equilibrated at a new steady-state which was measured and 2 recorded; then the TEAL injection was begun downstream of 3 the methanol. Catalyst rate was again adjusted to keep 4 conversion about constant and the reactor was equilibrated at a third steady-state which was again measured and record-6 Ed finally both the methanol and TEAL injections were stop-7 pod and catalyst rate was again adjusted to keep conversion 8 constant and the reactor was equilibrated at a fourth steady-9 state. The first and fourth steady-states were both with no materials being purposely injected into the feed and 11 represent the base case with which the test periods can 12 be compared. Although both the initial and final steady-13 states represent the base case with no injections into the 14 feed, they are not identical because the reactor had been 15 on-stream for a long period by the time the fourth steady-16 state was reached. In the interim it had suffered con-17 siderable fouling and had warmed-up several degrees genii-18 grade. Because it takes at least six hours to line a rev 19 actor out at a new steady-state after a change is made, the reactor had been in production at least 24 hours by 21 the time the fourth steady-state was reached. In general, 22 as a reactor fouls and warms, certain trends can be expected:
23 1.) Catalyst efficiency tends to increase slowly 24 (note the initial catalyst efficiency was 1330 and the final 1440).
26 2.) The warm-up rate tends to rise at an ever 27 accelerating rate (note the initial warm-up rate 28 is 0.12C per hour and the final 0.14C per hour).
29 3.) Reactor index which is simply the ratio of Mooney to unrequited isobutylene tends to decrease 31 slightly (note the initial index is 11.7 and the 32 final 10.5). Reactor index is one measure of the 33 performance of a reaction system in terms of the 34 absence of poisoning effects with a high index being preferred.

Lo 1 It is important to be aware of these trends as the reactor ages in interpreting the data.
3 A comparison of the before test period with the period during methanol injection shows that methanol severe-lye depresses Mooney viscosity (a measure of polymer Milwaukee-6 far weight) and reactor index as well as decreasing catalyst 7 activity. These changes are undesirable and if they occur-8 Ed due to random fluctuations in the feed methanol level 9 would make it very difficult to control the reactor and produce specification bottle rubber. Data from the period Jo of simultaneous injection of methanol and TEAL show that 2 injection of the TEAL has counteracted the adverse effect 13 of methanol on Mooney viscosity and reactor index. A come 14 prison of this steady-state with the final steady-state after stopping the injection of both agents shows that the 16 TEAL has completely prevented the methanol from depressing 17 Mooney viscosity or reactor index. Clearly injection of 18 methanol has severely affected the reactor but it has little 19 effect in the buffered reactor containing TEAL.
TEAL does not counteract the adverse effect of 21 methanol on catalyst efficiency because as is shown later 22 TEAL itself depresses catalyst efficiency, but this is not 23 a significant practical problem. It is also apparent though 24 that introduction of methanol and TEAL has a favorable effect in reducing the rate of reactor warm-up which is highly 26 desirable. This favorable effect of TEAL is demonstrated 27 further in later examples.
28 These data show then that TEAL is an effective 29 buffering agent under large scale bottle production conditions and that it is able to counteract the adverse effects of 31 methanol on polymer molecular weight and reactor index.

32 Furthermore, it is shown to have a beneficial effect on 33 reactor warm-up rate.

I Another series of plant tests were run in a manner 36 similar to that described in example 12 to confirm the of-I

festiveness of TEAL as a buffering agent against the adverse 2 effects of methanol in a bottle polymerization reactor.
3 In this series of tests though, the TEAL was injected into 4 the reactor first and then the methanol was injected to evaluate the effects of methanol in the buffered reactor 6 containing TEAL. The injection points were as in example 7 12 with the TEAL being injected into the cold feed down-8 stream of the methanol and immediately before entering the g reactor. For this series of tests, feed rate into the no-actor was 6655 Kg/hr. of a feed containing 32.4% by weight 11 isobutylene and 0.84% by weight isoprene in methyl chloride.
12 As in example 12, TEAL was injected into the feed at 0.8i4 '3 Kg/hr. and methanol at 0.129 Kg/hr. during the injection 14 periods. The molar ratio TEAL/methanol was again 1.86.
Data from the various steady-state portions of this test 16 series are summarized below:

18 BEFORE TEST + TEAL + TEAL & AFTER
19 Mesh TEST
20 Mooney Viscosity 60 62 59 58 21 Reactor Index 10.5 10.9 10.5 10.0 22 tlnreacted Isobutylene, 5.7 5.7 5.6 5.8 23 %
24 Isobutylene Convert 86.4 86.0 86.3 86.1 25 soon, %
26 Alec Rate, Kg/hr. 1.13 1.68 1.70 0.96 27 Mole/hr. 8.4412.59 12.73 7.18 28 Alec Efficiency 16551103 1095 1930 29 Warm-up Rate,C/hr. 0.14 0.23 0.17 0.68 30 In this test series the reactor was again equal-31 brazed at four separate steady-state conditions as in exam-32 pie 12 with catalyst rate adjustments being made after each 33 change to keep conversion approximately constant. The no-34 actor was first equilibrated in normal operation, then with TEAL being injected, then with TEAL and methanol being in-36 jetted, and finally again with the injection of both agents 37 stopped.
3~3 The data show that injection of TEAL alone has 39 a mild beneficial effect on Mooned viscosity and reactor I

1 index as in the laboratory tests and causes a significant 2 (although unimportant) drop in catalyst efficiency. Then 3 injection of methanol into the reactor containing TEAL was 4 begun and had alr,lost no effect at all. The buffered reactor into which TEAL was being injected was virtually unaffected 6 by the methanol. This is in sharp contrast to the effect 7 of methanol injection into an unbuffered reactor as in ox-8 ample 12. Simultaneous removal of the TEAL and methanol 9 again had no appreciable effect on Mooney viscosity or no-actor index, although catalyst efficiency of course rose 11 sharply. reactor warm-up rate again rose sharply also when 12 the TEAL and methanol were removed, showing their beneficial 13 effect on this important parameter. Clearly these data I show that TEAL is an effective buffering agent under large scale polymerization conditions to prevent methanol from 16 adversely affecting reactor performance. Variations in 17 the methanol level of a feed to a buffered reactor contain-18 in TEAL would not upset the reactor in any way whereas 19 similar variations in the methanol level of the feed to an unbuffered bottle reaction system would greatly upset 21 the reactor and cause off-specification rubber to be pro-22 duped and reactor control to become difficult.

24 A series of large scale tests similar to those described in examples 12 and 13 were run to confirm the 26 effectiveness of TEAL as a buffering agent to counteract 27 the adverse effects of acetone in a bottle polymerization 28 reactor. In this series of tests, acetone was injected 29 into the equilibrated reactor to measure its adverse effect.
This was followed by a period in which TEAL and acetone 31 were injected simultaneously to determine whether TEAL could 32 negate the adverse effects of acetone. Feed rate into the 33 reactor for this series of tests was 6542 Kg/hr. of a feed 34 containing 32.0% isobutyler.e and 0.83% isoprene by weight in methyl chloride. Acetone was injected at 0.22 Kg/hr.
'6 = 3.77 moles/hr. = 33 ppm on feed and TEAL was injected I I

1 at 0.854 Kg/hr. = 7.5 moles~hr. = 130 ppm on feed during 2 the injection periods. The molar ratio of TEAL/acetone 3 was 1.99. Data from the various steady-state portions of 4 this test series are summarized below:
SUMMARY RESULTS-ACETONE FOLLOWED BY TEAL TEST SERIES
6 BEFORE + ACETONE + ACETONE AFTER TEST

8 Mooney Viscosity 68 59 70 64 9 Reactor Index 9.4 10.5 10.5 10 Unrequited Issue 6.3 6.7 6.1 11 battalion, %
12 Isobutylene Con- 84.0 84.3 82.7 85.2 13 version, 14 Alec Rate, Kg/hr. 1.14 1.40 1.74 0.95 Molejhr.
16 Alec Efficiency 990 1870 18 Warm-up Rate,C/hr. 0.085 0.10 0.14 0.27 9 These data show that injection of acetone into the equilibrated reactor severely depressed Mooney viscosity, 21 reactor index, and catalyst efficiency. A random variation 22 in the acetone level of the feed to a bottle reactor would 23 severely upset the reactor and result in off-specification 24 production. Injection of TEAL downstream of the acetone completely counteracted its deleterious effects on Mooney 26 viscosity and reactor index (but not on catalyst efficiency 27 as already explained). Simultaneous removal of the TEAL
28 and acetone had no effect on reactor index (and would have 29 had no effect on Mooney viscosity if isobutylene conversion had been held constant). Again it is clear that TEAL is 31 an effective buffering agent to counteract the adverse of-32 feats of acetone in a bottle reaction system. Introduction 33 of acetone into a buffered reactor containing TEAL would 34 have little effect on the reactor performance or product.

36 A series of large scale tests similar to those 37 already described was run to confirm the effectiveness of I TEAL as a buffering agent against the adverse effects of 39 deathly ether in a butvl polymerization reactor. In this to 1 series of tests, TEAL was injected into the reactor airs.
2 and then the deathly ether was injected to evaluate the 3 effects of the ether in the buffered reactor containing 4 TEAL. The injection points were as previously described with the TEAL being injected into the cold feed downstream 6 of the ether injection point and immediately before entering 7 the reactor. The feed rate for this series of tests was 8 6542 Kg/hr . of a feed containing 30.4~ isobutylene and 0.79 9 isoprene by weight in methyl chloride. TEAL was injected into the feed at a rate of 0.69 Kg/hr. = 6.1 moles/hr. =
11 105 ppm on feed and deathly ether was injected at 0.288 12 Kg/hr. = 3.89 moles/hr = 44 ppm on feed during the injection 13 periods. The molar ratio of TEAL/ether was 1.57. As in 14 the previous tests, the reactor was again equilibrated at ' four separate steady-state conditions with catalyst rate ; adjustments being made after each change to keep isobutylene ' conversion approximately constant. The reactor was first 18 equilibrated in normal operation, then with TEAL being in-19 jetted, then with TEAL and deathly ether being injected simultaneously, and finally with the injection of both 21 agents stopped. Data from the four steady-state portions 22 of this test series are summarized below:

24 BEFORE + TEAL + TEAL & AFTER TEST
TEST DEATHLY

27 Mooney Viscosity 64 68 64 65 28 Reactor Index 11.7 11.2 11.4 29 Unrequited Is- 5.5 5.8 5.7 5.7 battalion, %
31 Isobutylene Con- 85.5 84.3 84.5 85.1 32 version, 33 Alec Rate, Kg/hr 1.20 1.84 1.84 1.18 34 mole/hr.
35 Alec Efficiency, 1410 912 915 1440 37 Warm-Up Rate, 0.09 0.05 0.07 0.13 38 cry.
39 The data show that introduction of TEAL alone had a very mild beneficial effect on Mooney viscosity and no-

41 actor index and caused a substantial drop in catalyst effi-I

1 Chinese. Then introduction of the deathly ether into the 2 reactor buffered by NEAL had almost no effect at all. TEAL
3 was very effectively buffering the reactor against the ad-4 verse effects of deathly ether so its introduction did not upset the reactor in any way. In a comparison test, the 6 results of which are shown below, introduction of deathly 7 ether into a reactor not protected by the buffering action 8 of TEAL sharply depressed Mooney viscosity and reactor in-9 dew, and resulted in poor quality bottle rubber. Simultan-eons removal of the TEAL and ether again had no effect on 11 Mooney viscosity or reactor index showing that the TEAL
12 was preventing the ether from depressing these parameters.
13 Catalyst efficiency of course did jump when the buffering 14 agent and poison were removed. These data again show that TEAL is an effective buffering agent under large scale polyp merization conditions to prevent deathly ether from adversely -7 affecting reactor performance. The buffered reactor contain-18 in TEAL had been rendered insensitive to variations in 19 the ether level in the feed and would not be upset by the natural feed quality variations which occur from time to 21 time.
22 Data from the comparison test showing the effects 23 of deathly ether injection into a normal, unbuffered reactor 24 are shown below. In this test feed rate was 6542 Xg/hr~
of feed containing 32.5% isobutylene and 0.84% isoprene 26 by weight in methyl chloride. The deathly ether was in-27 jetted at 0.25 Kg/hr. = 3.4 mole/hr. = 38.5 ppm on feed.
I SUGARY RESULTS-DIETHYL ETHER EFFECTS
29 BEFORE TEST+ DEATHLY ETHER
30 Mooney Viscosity 55 46 31 Reactor Index 10 8.5 32 Unrequited Isobutylene, % ,.5 5.4 33 Isobutylene Conversion, % 87.1 87.2 34 Alec Rate, grow. 1.04 1.30 Mole/hr.
JO Alec Efficiency, W/W 1780 1420 3, Warm-up Rate, cry. 0.17 0.14 3~LL7~5 i The date show that injection of ether into the 2 normal, unprotected reactor caused a severe depression in 3 Mooney viscosity, reactor index, and catalyst efficiency and resulted in the production of off-specification bottle rubber.
6 Results of these plant tests 12 through 15 have 7 shown that TEAL is an effective buffering agent for all 8 classes of oxygenated poisons (alcohols, kittens, and ethers).
9 It effectively prevents these poisons from affecting a bottle lo reactor so that stable on-specification production can be if maintained despite fluctuations in the levels of these pot-12 sons in the feed. In addition, these data show that TEAL
13 injection into the reactor has a beneficial effect in no-lo during the reactor warm-up rate, which has important equine-mix consequences because a reduced warm-up rate permits extending run length and/or increasing polymer production 17 rate without reducing achievable run length to an unaccep-I tally low value.
lo EXPEL 16 I Buffering against the effect of Hal in a bottle 21 polymerization reaction was the purpose of this example.
Hal is commonly referred to as a catalyst activator in the I bottle process. It is known to depress Mooney viscosity and I reactor index and to also dramatically increase catalyst efficiency and reactor warm-up rate when present in reactor feed. Increased catalyst efficiency is often desirable, ? but faster reactor warm-up is always undesirable because it , L limits run length and bottle production.
I A series of large-scale tests was carried out 3G hollowing the procedures previously described. However, be-31 cause of the very powerful effect Hal has on the reactor I warm-up rate it was not possible to conduct a prolonged --; test to achieve a number of steady-state conditions in the I same reactor as in the previous tests. Instead, the Hal 3J tests were run one at a time and in some instances it is I necessary to compare the reactor containing HC1 wit a ~L23~

1 side-by-side control reactor instead of compa-irg two dip-2 fervent steady-states in the same reactor as in the previous 3 examples.
4 The extremely deleterious effects that HC1 has on bottle polymerization is shown by the results of the lot-lowing experiment.
7 An hydrous Hal gas was injected into the feed of 8 an equilibrated reactor at 7.5 ppm by weight on the feed.
9 The injection point was the same as that used for inject lion of the oxygenated poisons, at the inlet of the first 11 ethylene feed chiller. As soon as the Hal gas injection 12 was started, the reactor temperature began rising rapidly.
13 It continued to rise rapidly accompanied by a sharp drop 14 in unrequited isobutylene level despite all efforts to no-dupe catalyst rate into the reactor. It proved to be imp 16 possible to stabilize the reactor to achieve a new steady-17 state. Within less than one hour, the reactor temperature 18 had risen by more than 5C and the reactor plugged so that :' it had to be taken out of production for cleaning. Clearly 'O a random fluctuation in the Hal level in the feed of this _ magnitude would be severely deleterious to polymerization 22 operations and could result in temperature run-away of all 2. operating reactors.
2 Because the effects of HAL were so powerful that I, it was not possible to inject 7.5 ppm into the feed of a I running reactor without causing it to "burn-up" and plus.
27 An attempt was therefore made to initiate polymerization 28 with Hal already present in the feed from the start. This I technique permitted operation of a reactor counterweighing Hal long enough to achieve a steady-state. However, it was not 31 possible to remove the Hal and recalibrate the reactor 32 without Hal to compare the two steady-states in the same 33 reactor because of the rapid warmup rate. Therefore, the I reactor containing XCl must be compared with another react I ion running at the same time as a control reactor. Both I reactor were operated in the same manner and at the same I time except that an hydrous Hal was being injected into the feed of the test reactor at 7.5 ppm on feed. The feed rate to both reactors was the same at 6542 Kg~hr. of a feed con-3 tunneling 33% isobutylene and 0.86% isoprene by weight in I} methyl chloride. Steady-state conditions in the two react ions are compared below:
o CONTROL REACT

8 WITH 7.5 p m Hal __ Hal _ 9 Mooney Viscosity 44 70 lo Reactor Index 9.4 12.5 11 Unrequited Isobutylene, 4.7 5.6 12 Isobutylene Conversion. 89.8 86.8 13 Alkali Rate, grow. 0.22 1.31 14 Mole/hr. 1.636 9.82 15 Alec Efficiency, WOW 1430 lo Warm-up Nate, cry. 1.2 0.1 '7 The data show that the Hal was severely depress-18 in Mooney viscosity and reactor index and causing a six-l? fold increase in catalyst efficiency. it the same time, reactor warm-up rate increased by a factor of 12 to 1.2C/
21 hr. In the five hour period necessary to achieve a steady-22 state, the test reactor had warmed more than 6C. It plugged 23 and had to be taken out of production for cleaning shortly I after the Hal injection was stopped to try to reestablish 25 an equilibrium without Hal. Clearly Hal in the feed has a I disastrous effect upon reactor performance.
27 In another test, Hal and TEAL were injected Somali-28 tonsil into an equilibrated reactor to confirm the of-29 festiveness of TEAL as a buffering agent to prevent the 30 harmful effects of Hal on the bottle polymerization. In this I test, the reactor was equilibrated at normal conditions and 32 then injection of Hal and TEAL was begun simultaneously into 33 the feed to establish a new steady-state. Jo third steady-34 state was reached by discontinuing the injection of both 35 agents. Feed rate to the reactor was 6542 kg/hr~ of a feed 36 containing 33.2% isobutylene and 0.85% isoprene by weight in 37 methyl chloride. During the injection period, an hydrous Hal 38 gas was injected at the first ethylene chiller at a rate of ' 0.042 Kg/hr. = 1.15 moles/hr. = 6.4 ppm on feed. TEAL was 2 injected downstream of the Hal at the inlet of the second 3 feed ethylene chiller into cold feed at a rate of 0.992 4 Kg/hr. = 8.7 moles/hr. = 152 ppm on feed. The molar ratio of TEAL/HCl was 7.57. Data from the various steady-state 6 periods of this test series are summarized below:

7 SUMMARY RESULTS - HCl/TEAL SIMULTANEOUS INJECTION
O BEFORE Hal AND TEAL AFTER

10 Mooney Viscosity 57 65 59 11 Reactor Index 10.2 10.8 9.8 it reacted Isobutylene, % 5.6 6.0 6.0 13 Isobutylene Conversion, %86.9 85.9 85.9 14 Alec Rate, Kg/hr.1.23 1.23 1.23 5 Mole/hr. 9.19 9.19 9.19 6 Alec Efficiency, W/W 1541 1524 1524 17 Warm-up Rate, cry 0.2 9.1 18 Methanol, ppm on feed 11 11 11 :.~ These data show that TEAL is an effective buffer to prevent the harmful effects of Hal on the bottle polymer-21 ization reaction. The TEAL completely prevented the Hal 22 from depressing Mooney viscosity and reactor index, and also 23 prevented the Hal from raising catalyst efficiency and warm-24 up rate. In fact, the simultaneous introduction of Hal and TEAL had very little effect on the reactor at all as 26 contrasted to the disastrous effects Hal had in the absence 27 of the protective buffering action of TEAL. The slight 28 increase in Mooney viscosity and reactor index which ox-29 cuffed in this test series when Hal and TEAL were introduced simultaneously was due to the fact that a small amount of 31 methanol was present in the feed and the TEAL was counter-32 acting its effects as well as preventing the HC1 from 33 affecting the reactor.
Clearly, TEAL is an effective buffering agent 3' against the harmful effects of HC1 gas on bottle polymer-l ization. Very small variations in the HC1 level of bottle 2 feeds do occur in normal practice and drastically affect 3 reactor performance in the absence of TEAL but would have 4 little effect on a buffered reactor containing TEAL.

6 An hydrous chlorine gas is another powerful gala-7 lust activator like Hal gas. At very low levels in the 8 reactor feed, it severely depresses Mooney viscosity and g reactor index and strongly increases catalyst efficiency and reactor warm-up rate. Injection of 7.5 ppm of chlorine if gas into the feed of an equilibrated reactor causes effects 12 very similar to those reported in example 16 for injection 13 of 7.5 ppm of Hal gas. The reactor warms at a rapid rate, 14 cannot be stabilized and plugs before a steady-state can be achieved. Since a test with injection of chlorine alone 16 could not be run, a test with simultaneous injection of 17 chlorine and TEAL was made to confirm that TEAL is an 18 effective buffering agent against the harmful effects of lo chlorine on bottle polymerization. In this test the reactor was equilibrated normally and then injection of chlorine and 2' TEAL into the feed was begun simultaneously to establish a -I new steady-state. The injection points were as in example 23 16. Feed rate into the reactor was 6542 Kg/hr. of a feed 24 containing 33.5~ inbutylene and 0.87~ isoprene by weight in methyl chloride. The feed analyses showed it was con-26 staminate with about 8.2 ppm methanol. Chorine was injected 27 at 0.049 Kg/hr. = 0.7 moles/hr. = 7.6 ppm on feed, and TEAL
28 was injected at 0.74 Kg/hr. = 6.5 moles/hrs. = 113 ppm on feed. The molar ration of TEAL/Cl2 was 9.28. Data from the JO steady-state portions of this run are summarized below:

~23~

I

4 Mooney Viscosity 51 53 5 Reactor Index 9.1 9.5 6 Unrequited Isobutylene, 5.6 5.6 7 Isobutylene Conversion, 87.1 87.4 8 Alec Rate, Kg/hr. 1.39 0.91 g Moles/hr 10.46 6.8 10 Alec Efficiency, W/W 1367 2106 11 Warm-up Rate, cry. 0.1 0.2 12 These data show that TEAL is an effective buffer-13 in agent to prevent the disastrous effects of chlorine upon I bottle polymerization. Simultaneous introduction of Shelley-fine and TEAL caused a mild improvement in Mooney viscosity, 16 reactor index and catalyst efficiency and had little effect 17 on warm-up rate. Some of the mild improvement was due to 13 TEAL canceling out the negative effects of the small 19 amounts of methanol also present in the feed.
In the absence of TEAL, injection of a similar 2 amount of chlorine would have caused the reactor to warm-up rapidly and plug, and would have severely depressed Mooney viscosity and reactor index as well as greatly increasing Jo catalyst efficiency.

Tertiary bottle chloride is another very powerful 27 activator in bottle polymerization. Its adverse effects are 2& similar to those of Hal and chlorine but not quite as 29 pronounced so that it is possible to run a sequential test and obtain several steady-states in the same reactor run. A
I series of large scale tests similar to those described in ,2 examples 12-18 was run to confirm the effectiveness of DIAL
H as a buffering agent against the adverse effects of t-butyl I, chloride in a bottle polymerization reactor. In this series or of tests, t-butyl chloride was injected into an equilibrated I reactor to establish a new steady-state. Then DIAL was ~23~ 5 injected into the cold feed downstream of the t-butyl 2 chloride and immediately before entering the reactor to 3 establish a third steady-state with DIAL H counteracting 4 the adverse effects of t-butyl chloride. Catalyst rate adjustments were made after each change to keep conversion 6 as constant as possible. Feed rate into the reactor for this 7 series of tests was 6542 Kg/hr. of a feed containing 33 & isobutylene and 0.86~ isoprene by weight in methyl chloride.
9 Tertiary bottle chloride was injected at a rate of 0.046 o Kg/hr. = 0.5 moles/hr. = 7 ppm on feed. DIAL H was injected 11 at 0.32 Kg/hr. = 2.24 moles/hr. = 48 ppm on feed. The molar 12 ratio of DIAL H/t-butyl chloride was 4.48. Data from the 13 various steady-state portions of the run are summarized 14 below:

15 SUMMARY RESULTS - t-BUTYL CHLORIDE FOLLOWED BY DIAL H SERIES
16 BEFORE + T-BuCl 17 TEST + t-BuCl DIAL H _ 18 Mooney 59 62 19 Reactor Index 10.5 8.710.3 Unrequited Isobutylene, 5.6 5.6 6.0 21 Isobutylene Conversion, 87.0 87.6 85.8 22 Alec Rate, Kg/hr.1.23 0.375 1.43 3 Moles/hr. 9.22 2.8 10.7 I Alec Efficiency, W/W 15205045 1295 These data snow that introduction of tertiary I bottle chloride into the equilibrated reactor severely de-27 presses Lyon and index and causes more than a three-fold 28 increase in catalyst efficiency. Introduction of the DIP
29 BAY H completely cancels out all the adverse effects of the t-butyl chloride. Inn viscosity and reactor index 31 are restored to the before-test levels and the catalyst 32 activation effect is more than completely canceled.
33 These data show that DIAL H is a very effective 34 buffering agent against the harmful effects of t-butyl chloride in a bottle reactor. An increase in the t-butyl 36 chloride level of the bottle feed would cause production of 37 off-specification product and very seriously affect per-I

1 pheromones of an unbuffered reactor, whereas the same in-2 crease in feed t-butyl chloride level would have almost no 3 effect in a reactor protected by effective buffering agents 4 such as DIAL or TEAL. This anility of these agents to buffer the reactor against the serious consequences of in-6 curs ions of powerful activators into the feed is of great 7 economic benefit.
8 E~PLE 19 g As still further confirmation of the effective-10 news of TEAL as a buffering agent to protect operating bottle 11 reactors against the adverse consequences of variations in I the levels of trace quantities of poisons and activators inn the feed streams, a third type of large scale test was made I in addition to those already described. In normal operation I of large scale bottle reactors, the feed components (methyl chloride, isobutylene, isoprene and isobutylene diver) are continuously metered at the desired rates and Russ into a 13 feed blend tank where they are mixed and blended to prepare 13 the feed blend, which is metered to separate reactors which may be in operation. A convenient way of demonstrating the I effectiveness of a buffering agent against a poison or ,2 activator is to have two reactors in production being fed I from the same feed blend tank with only one of the reactors I being protected by injection of the buffering agent into its reed stream. Then a quantity of poison or activator is 26 injected into the feed blend tank so the poisoned feed is fed 27 to both reactors. The difference in the response of the Jo protected reactor containing the buffering agent and the I normal reactor can then be used to evaluate the effective-3C news of the buffering agent against that poison. This type 31 of test was run with prop ionic acid as the activator and TEAL
I as the buffering agent.
3- This is not a steady-state test as in the previous 34 examples, but a test which measures the transient response I of the reactors to a fluctuation in the feed poison level.
I In this test with prop ionic acid, a quantity of prop ionic Al acid was introduced into the feed blend to yield 24.2 ppm by I

1 weight of prop ionic acid in the feed at the start of the 2 test. This level then decreased exponentially in the feed 3 blend tank as more clean feed was continuously added.
4 Computer calculations were made of the decay rate and showed that the prop ionic acid level had fallen to 1 ppm about 107 6 minutes after the start of the test. The prop ionic acid 7 level in the reactors started at I, rose for a time as the 8 poisoned feed was fed to them and then slowly decayed again as the prop ionic acid was flushed from the system. Computer calculations showed the maximum prop ionic acid level in the 11 reactors reached 7.4 ppm about 39 minutes after the start of 12 the test and had fallen to 1 ppm after 170 minutes. Reactor 13 conditions were continually changing during this time, but 14 no changes in catalyst rate were made. The reactors were allowed to respond naturally to the changing prop ionic acid 16 level. Prop ionic acid is a fairly potent activator for bottle 17 polymerization, but much weaker than Hal, t-butyl chloride, 18 or an hydrous chlorine gas discussed in earlier examples.
19 Feed rate to both reactors during this test was 20 7870 Kg/hr. of a feed containing 33.1~ isobutylene and 0.87%
21 isoprene by weight in methyl chloride. TEAL was being 22 injected into the protected reactor at 0.724 Kg/hr. = 6.35 23 moles/hr. = 92 ppm on feed. Introduction of the spike of 24 prop ionic acid into the feed blend had no noticeable effects 25 on the reactor into which the TEAL was being injected as a 26 buffering agent. This protected reactor continued to prod-27 use on-specification rubber with no sudden warming or change 28 in unrequited isobutylene level. On the other hand, the 29 control reactor containing no buffering agent suffered a 30 severe upset due to the prop ionic acid contaminated feed.

31 reactor temperature increased by 2.5C, unrequited isobutyl-ennui level fell by an increment of 2.5%, and Mooney viscosity of the bottle rubber being produced fell by lo points. This reactor produced off-specification rubber for more than 4 hours 35 before it finally recovered from the upset. There was thus I a very dramatic difference in the response of the protected 37 reactor and that of the control reactor to the addition of I

1 prop ionic acid to the feed. The TEAL acted effectively as 2 a buffering agent to prevent the reactor into which it was 3 being injected from being affected in any way by the incur-4 soon of the activator into the feed, whereas the control reactor was badly upset and produced a large amount of 6 unacceptable rubber. The economic benefit of using the 7 buffering agent is dramatically shown by this test.
8 Clearly TEAL is a very effective buffering agent , against the adverse effects of prop ionic acid on bottle I polymerization. Although this was an artificial test in 11 that prop ionic acid was purposely injected into the feed, it 12 closely simulates an actual upset situation caused by pot-13 sons or activators in the feed. These materials normally 14 enter the feed when there is some upset in the recycle system (i.e., fractionating towers or dryers) or when a batch of 16 contaminated raw materials (i.e., monomers or Dylan 17 enter. The poison or activator is thus introduced naturally I in much the same way as simulated in this test. The very 19 excellent stabilizing effect of the TEAL is readily apparent and would be very valuable.

22 Another test similar to that described in example 23 19 was run to compare the transient response of a buffered 24 reactor and control reactor to an incursion of isopropyl alcohol into the feed. Isopropyl alcohol is a fairly strong 26 activator in bottle feed. In this test a large slug of 27 alcohol was introduced into the feed blend tank to determine 28 the response of the reactors to a large incursion of anti-29 valor. Enough isopropyl alcohol was injected at once into the feed blend tank to raise the alcohol level to 120 ppm and 31 this was allowed to decrease exponentially as new feed was 32 continually added and the contaminated feed was fed to the 33 reactors. Computer calculations showed the alcohol level in 34 the feed blend tank decayed to 1 ppm after 124 minutes. The isopropanol level in the reactors rose slowly from G to 31.7 36 ppm in 33 minutes and then fell back down slowly to 1 ppm 37 after 218 minutes. As in the test of example 19, it was 33 planned to allow the TEAL protected and the control reactor aye 1 to respond normally to the incursion of the alcohol-con-laminated feed, but the alcohol contaminated feed had such 3 a drastic effect on the control reactor that it became 4 necessary to stop catalyst flow to that reactor and take it out of production; the TEAL-protected reactor was not 6 significantly affected.
7 In this test, TEAL was injected into the feed of the buffered reactor at 0.724 Kg/hr. = 6.35 moles/hr. = 92 g ppm on feed. The feed rate to both producing reactors was 19 7870 Kg/hr. of a feed containing 33.1% isobutylene and 0.87%
11 isoprene by weight in methyl chloride.
12 There was no significant change at all in the 13 buffered reactor which was protected by having TEAL injected 14 into its feed. Temperature rose by a few tenths of a degree, but this is hardly more than the normal fluctuations a no-16 actor experiences; and there was no change in unrequited 17 isobutylene level or Mooney viscosity of the rubber being 18 produced. The TEAL was obviously acting as an effective 19 buffer.
On the other hand, the control, unprotected rev 21 actor was so severely upset by the isopronanol-contaminated 22 feed that it has to be taken out of production and cleaned.
23 Reactor temperature started rising rapidly about 8-10 mint I vies after the isopropanol was introduced into the feed 25 blend, and the temperature continued to rise rapidly. After 26 20 minutes, catalyst flow was stopped to this reactor in an 27 effort to keep it from overheating, but temperature con-28 tinted to rise. By 25 minutes the reactor temperature was 29 nearly 5C warmer than at the start and was continuing to 30 rise. The slurry was so warm at that point that the reactor 31 was beginning to plug and it had to be taken out of pro-32 diction and put through a wash cycle for cleaning. Unrequited 33 isobutylene concentration in the reactor had fallen by more 34 than an increment of owe and the Mooney viscosity of the polyp 35 men being produced had also started to plummet. Clearly the isopropanol-contaminated feed had drastically affected 3, this normal reactor causing the test run to be terminated 38 prematurely.

I
. , , ., , ., Jo

42 -1 These data again show the effectiveness of TEAL as it 2 a buffering agent to protect an operating bottle reactor 3 against upsets caused by variations in the levels of poisons and activators in the reactor feed streams. A TEAL buffered reactor is able to continue producing acceptable bottle 6 rubber during periods of feed quality fluctuation which 7 would cause a normal unprotected reactor to produce unyoke-suitably rubber or to become so upset that rapid fouling and 7 agglomeration of slurry particles force a termination of the I cycle for cleaning. The significant economic advantages 11 which accrue from this ability to maintain reactors in stable 12 operation despite variations in feed quality are obvious.

Claims (19)

1. THE EMBODIMENTS OF THE INVENTION IN WHICH AN EXCLUSIVE PROPERTY OR
   PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:
   l. In the soluble Lewis Acid catalyzed polymerization method of preparing C4-C7 isoolefin homopolymers or copolymers of a C4-C7 isoolefin and a C4-Cl4 conjugated diene in an inert diluent, the improvement which comprises adding to the polymerization reaction zone about 0.l to 1.0 mole of a buffering or complexing agent per mole of Lewis Acid catalyst in an amount at least effective to protect the said rubber product against molecular weight depression due to the presence of polymerization activators or poisons, the buffering or complexing agent being (1) a lower trialkyl aluminum, the alkyl being methyl or ethyl or (2) a lower dialkyl aluminum hydride, the alkyl of said hydride being a Cl_C4 alkyl group.
2. 2. In the soluble Lewis Acid catalyzed polymerization method of preparing C4-C7 isoolefin homopolymers or copolymers of a C4-C7 isoolefin and a C4-C14 conjugated diene in an inert diluent, the improvement which comprises adding to the polymerization reaction zone a buffering or complexing agent in an amount at least effective to protect the said rubber product against molecular weight depression due to the presence of polymerization activators or poisons, the buffering or complexing agent being (1) a lower trialkyl aluminum, the alkyl being methyl or ethyl or (2) a lower dialkyl aluminum hydride, the alxyl of said hydride being a Cl-C4 alkyl group.
3. 3. The method of claim 2 wherein the buffering or complexing agent is triethyl aluminum.
4. 4. The method of claim 2 wherein the buffering or complexing agent is trimethyl aluminum.
5. 5. The method of claim 2 wherein the buffering or complexing agent is diisobutyl aluminum hydride.
6. 6. The method of claim 2 wherein the amount of buffering or complexing agent is up to about 1.0 mole of said agent per mole of said catalyst.
7. 7. The method of claim 2 wherein the amount of buffering or complexing agent is at least in molar excess of the amount of polymerization poisons present in the polymerization reaction mixture.
8. 8. The method of claim 2 wherein said homopolymer is polyisobutylene and said copolymer is isobutylene-isoprene butyl rubber.
9. 9. The method of claim 2 wherein the catalyst is aluminum chloride.
10. 10. The method of claim 2 wherein the inert diluent is selected from the group consisting of methyl chloride, methylene chloride, vinyl chloride, ethyl chloride, propane, butane, pentanes or hexanes.
11. 11. The method of claim 10 wherein the inert diluent is methyl chloride.
12. 12. The method of claim 8 wherein said butyl rubber has a viscosity average molecular weight of 250,000 to 600,000, the diluent is methyl chloride, the catalyst is aluminum chloride and the buffering or complexing agent is selected from the group consisting of triethyl aluminum, trimethyl aluminum, or diisobutyl aluminum hydride.
13. 13. The method of claim 2 wherein said buffering or complexing agent is added to said polymerization reaction zone by adding said agent to the cold, incoming monomer feedstream at a point immediately prior to the introduction of said feedstream into said polymerization reaction zone.
14. 14. The method of claim 2 wherein said buffering or complexing agent is added to said polymerization zone prior to the time of introduction of monomers and catalyst.
15. 15. The method of claim 14 wherein said buffering or complexing agent is also added to the cold, incoming monomer feedstream at a point immediately prior to the introduction of said feedstream into said polymerization reaction zone.
16. 16. The method of claim 2 wherein said buffering or complexing agent is in contact with said isobutylene monomer or isoolefin and conjugated diene prior to polymerization for a period of time of less than about 60 seconds.
17. 17. The method of claim 13 wherein said isoolefin homopolymer is polyisobutylene.
18. 18. The method of claim 13 wherein said copolymer is isobutylene-isoprene butyl rubber.
19. 19. In the soluble Lewis Acid catalyzed polymerization method of preparing C4-C7 isoolefin homopolymers or copolymers of a C4-C7 isoolefin and a C4-C14 conjugated diene in an inert diluent, the improvement which comprises adding to the polymerization reaction zone about 0.1 to about 1.25 mole of a buffering or complexing agent per mole of Lewis Acid catalyst in an amount at least effective to protect the said rubber product against molecular weight depression due to the presence of polymerization activators or poisons, the buffering or complexing agent being (1) a lower trialkyl aluminum, the alkyl being methyl or ethyl or (2) a lower dialkyl aluminum hydride, the alkyl of said bydride being, a Cl-C4 alkyl group.