Document 2jZKemKRJ59MgVjpQ1aoOkbQ7

^ete S W TY\Ob hili i\\ad u, 15 . 4 - i 6 n\e^tHll-c **June 3> 1951* MEMORANDUM OF DISCUSSIONS SHELL DEVELOPMENT CO REC'D ;n 1 > oia l54 Shell Chemical Corporation Denver, Colorado Denver Mr.'R. H. Beilin Mr. P. R. Bolenbaker Mr. J. E. Knaus Mr. E. M. Lando Mr. C. C. McKenna (part time) Mr. G. McMasters (part time) Mr. R. E. Scheldt Mr. R. M. Stager (part time) Mr. M. Smith Mr. S. Zevin Denver, Colorado May 10 and 11, 1954 Hew York Us* Mr. B. H. Cuu&ings Mr. W. C. Walker TO HAN0LE-. cc - iW/f tc - M( *1 ANS'O: BY: EPON ADHESIVES VI AND VIII A meeting was held between Denver and New York personnel to dis cuss the nature of the shortcomings of the Epon Adhesives VI and VIII man ufactured at Denver and to develop information to carry to Emeryville re garding procedures and equipment for discussion with Baeryville personnel. Excessively high viscosity for the Adhesives and the reported short "pot life" for one batch of Adhesive VIII were the main factors necessitating such a meeting. The viscosity of a number of batches of Adhesives from Denver has exceeded the sales requirements of 1 and 2 MM cp. for Adhesives VI and VIII, respectively. In addition, recent information was received , from Western Division Marketing of a complaint from North American Aviation ` that the "pot life" of a batch of Adhesive VIII was only 20 minutes com pared with the minimum of one hour which they consider satisfactory. Discussions were initiated with a review of the Denver and Emery ville procedures for preparation of Adhesives VI and VIII. A description of each of the manufacturing processes for both locations is appended. The principal difference between the Emeryville and Denver procedures is the much longer batch times at Denver. Preparation of the Adhesives is essentially a mixing operation, with all of the ingredients blending together relatively easily excepting the Vinylite AYAF which is added to the mixture as 1/8"-l/4" pellets. The Vinylite dissolves very slowly at temperatures less than 100C and must actually be melted to obtain ready solution. At Emeryville the Vinylite is predissolved in the Epon 828 for a batch of Adhesive at l60C in a stirred, heated kettle. At Denver no separate kettle is used to predi6solve the Vinylite - - all the ingredients for a batch being added to a Read mixer and mixed at 90C. At this lower temperature the time required to dissolve the Vinylite is not definitely known, but the mixture is not homogeneous after six hours mixing. For operating convenience the mixing is allowed to continue overnight at this temperature, with a total of twenty hours elapsing from the beginning of the batch until the Adhesive is ready for packaging. ABS-018254 -2 - June 3, 195^ A brief discussion of the conditions which might possibly lead to loss of allyl glycidyl ethyl (AGE) -- loss of which would increase the viscosity of the Adhesive -- indicated that the Denver apparatus is not completely vapor tight but that a metal cover on the mixer prevents stray air currents from contacting the batch. She Joints between the cover and the mixer appeared to be effectively sealed by the Adhesives. Preliminary laboratory tests have shown that substantially all the AGE Is retained In the batch based upon a volatiles recovery test. However, the volatiles removed were not analyzed and may contain materials other than AGE!. This volatiles test presently consists of measuring the weight loss from a uniform film of Adhesive l/l6" thick on a glass slide heated for one hour at 180C at 10 mm. mercury absolute pressure. Denver has found that a batch of Adhesive VIII prepared in the laboratory under plant conditions had a lower viscosity than the plant material, 1.3 MM cp. versus 2.0 MM cp. However, the Epon 828 used in the laboratory batch was not identical with the plant batch and, therefore, results are not strictly comparable. For instance, the viscosity of Epon 834 varies widely. Viscosity determinations of 1.1 MM cp. and 2.0W cp. have been made at Denver on separate batches of Epon 834. A discussion of the possible effect of raw material properties on viscosity of the finished products indicated that there is little like lihood that variations in the Vinylite AYAF have any significant effect. The proportion of Vinyllte to that of 828 is low (4.3$v), end the varia tions of viscosity of 828 are undoubtedly fully as large as those of Vinylite. ,, Pending discussions with Emeryville personnel, a further program 'Of process Improvement was discussed. It was generally agreed that all process improvement studies and the corresponding new Adhesive prepara tions would be made at Denver. In any event, Denver's program will In clude laboratory test runs to establish laboratory repeatability and re producibility of viscosity measurements using a new 1-30 MM cp.-range Brookfield Viscometer Just received at Denver May 11. During the time remaining from the discussions at Denver, lab oratory and plant facilities were inspected keeping in mind the require ments which might be imposed should it be decided to provide product ap plications testing facilities at Denver. In general, Denver has adequate laboratory and plant space for any foreseeable developments in Adhesive manufacture. ABS-018255 Shell Development Company Emeryville, California Shell Development Mr. R. k. Cole Mr. H. DeJong Mr. D. V. Elam Mr. E. C. Klute Mr. F. M. McMillan Miss M. Hap8 Mr. C. Rapean -3 - June 3, 195* Emeryville, California May 12, 13, and 1*, 195* Denver Mr. P. R. Bolenbaker Mr. J. E. Knaus Hey York Mr. B. E. Cunnings Mr. W. C. Walker Western Division tar. A. IP. Jtowe Mr. J. Morrill Discussions vere held at Emeryville on May 12, 13, and 1*, 195*> to review the current situation on Adhesives and to determine upon a program for future operation in this field. The present problem of high viscosity in the Denver production of both Adhesives VI and VIII was reviewed from the standpoint of manufacturing technique, customer requirements, and im mediate supply. An exhaustive discussion of the process variableswhich might in fluence the properties of Adhesives VI and VIII indicated that the major point of departure of Denver from Emeryville's procedure was the long mix ing time at Denver for the complete formula at 90C. Miss Haps reported that the pH of asbestos floats is on the alkaline side (8-11 pH) and that there was the possibility of further hydrolysis at 90C to yield alkaline materials. Since Epons are known to react with alkali at high temperatures with consequent increases in melting point and weight per epoxide equiva lent, there is a strong possibility that excessive mixing of the Adhesives at elevated temperatures may lead to a viscosity increase. The possibility of reaction between AGE and Epon also exists. Denver's laboratory experi ment, reported in discussions at Denver, does not substantiate this theory, and further control experiments will be required to establish the point. Denver will proceed tinnedlately with the necessary experiments in the laboratory. Meanwhile, the plant will initiate equipment changes to permit following Emeryville's procedure in all details. It will be necessary that promising laboratory samples be evaluated for performance by both Union and Emeryville before Denver will proceed to make plantscale batches. Close visual inspection of sanples of Adhesive VIII from Emery ville and Denver reveals discrete particles of Titanox in the Denver sample. Here again, a difference in manufacturing procedure is probably responsible for the difference. At Emeryville a 6($w concentrate of Titanox in Epon 828 or Epon 83* is made by mixing the two materials and passing them four times through a Morehouse Ry-R-Speed mill. This procedure was adopted at ABS-018256 -k - June 3, 195^ Emeryville to eliminate large particles of Titanox which were otherwise obtained in the Adhesives. It is felt that a ecngmrable method of Titanox dispersal should be adopted for Denver. Since a grinding mill is not available at Denver, means will be Investigated for having this operation performed for Denver by seme firm specializing in this work. Some time was devoted to discussing Horth American's report of excessively short "pot life" for Denver Batch 11 which had shown a good viscosity (1.67 lftf cp.). Prom the scanty evidence at hand it would appear that "pot life", as Indicated by North American's test, is not necessarily measurable as viscosity, and further study of consistency will probably be required. Various types of apparatus for determining consistency were described by Hr. McMillan. Among those mentioned were the Gardner Mobilometer and the Stoxmer Viscosimeter as being perhaps store suited to Epon Adhesives than the Brookfield Viscometer. A group of tests made with Adhesive VIII uncatalyzed, using a Brookfield Viscometer operated at different rpm. at roam temperature, showed that apparent viscosity drops with rate of shear on Adhesive VIII but did not explain the short "pot life". RPM. Shell Dev. VIII* Denver VIII Batch 11** 10 800.000 CD. 5 960.000 2 1,5^0,000 1 2,170,000 805.000 cp. 930.000 1,120,000 1,280,000 * Old batch supposedly OK. ** Insufficient sample. Brookfield spindle not at prescribed depth. The effect of the presence of catalyst on stiffening was examined for three batches of Adhesive VIII: Batch 11, Batch 8, and a sample of Emeryville production. The Adhesive was catalyzed with 6 parts by weight of Catalyst A per 100 parts of VIII and spread with glue spreaders on sheets of 20 in. x 20 in. masonite panel. Immediately after adding cata lyst and 20 minutes later there was little noticeable difference in spreadability. However, to some observers the 20-minute spread for Batch 11 seemed slightly more stiff and tacky. All three batches were still spread able after one hour. After standing four hours the catalyzed batches had sill stiffened noticeably, but Batch 11 was much stiffer and more rubbery than either Batch 8 or Emeryville material. From this it would appear that the cure may have been advanced in Batch 11 before the catalyst was added but that viscosity of the uncatalyzed resin is not necessarily an index of cure. On the basis of these findings. It was decided that an accurate evaluation of North American Aviation's problem could be achieved only by first-hand inspection of their operations. Accordingly, arrangements were made to visit their assembly plant on Monday, May 17. On reaching this ABS-018257 -5- June 3, 1954 decision, it was agreed that Emeryville would not be asked to make any Adhesive VIII until after a visit to North American.* Stocks of Adhesive VIII presently at Denver are as follows: Pounds Viscosity Batch Lot on Hand MM cp. Tensile Shear 77(2500) 180(1250) -70(2500) 2 A 495# 2.43 3110 3144 2834 4 A 360# 2.56 3183 2966 2962 5 A 720# 2.11 3453 3266 2792 6 1350# 2.43 3013 3242 - 10 A 720# 1.67 3150 3200 2840 B C ft ft 3126 3090 3125 3212 2610 2470 12 A 1395# 1*73 3330 3412 2470 B C It n 2920 2838 3055 3139 2570 2440 13 1300# 1.92 2820 3000 2550 14 1400# 2.05 2910 3120 2360 Batches 16, 17, and 16 (not shown above) were made by blending existing stocks to reduce the Vinylite AYAF to 6 phr. These are under going evaluation tests. These three batches amount to 4,200 lb. The situation, therefore, is one where about 2,000 lb. of marginally accept able material is on hand at Denver, with the possibility that more may pass specifications but have yet to be evaluated. Whether it will be necessary to have Emeryville make more Adhesive VUI will depend upon the outcome of discussions to be held at the North American Aviation assembly 'plant in Torrance on May 17 and upon Denver's ability to develop an ac ceptable Adhesive on short notice. Seme time was taken to discuss testing equipment for possible installation at Denver in order to expedite application evaluation of Adhesive products. Mr. H. C. Klute of Emeryville advised that the Tinius- Olson Tensile Tester, costing in the neighborhood of $6,000, was the cheap est apparatus that would be capable of giving satisfactory results. To this would need to be added a high temperature and a low temperature en closure for samples at about $1,000 and $1,600, respectively. With all auxiliary equipment he "guesstimated" that the complete testing apparatus might cost about $10,000. Other manufacturers of such equipment were listed and will be investigated. There was considerable discussion on coordination and close liaison between Shell Development, Marketing, and Manufacturing on both new processes and on any desired changes in specifications for products * Emeryville ha6 subsequently been requested to make 3,000 lb. of Adhesive VUI and to maintain minimum stocks of 2,000 lb. each of Adhesives VI and VIII until further notice. ABS-018258 -6 - June 3, 195^ being manufactured. A case in point vith regard to changes in specifica tions on an existing product was the recent change In the formula for Ad hesive VIII, reducing the Vinyllte content from 6 to 6 phr. Epon 628. Denver vas not advised of this change, and several batches of Adhesive VIII were made to the old recipe before the change vas discovered. For the Information of the group, the representative from Shell Chemical Manufacturing Operations described the procedure for establish ing a specification change In a commercial product as follows: 1. Marketing may Initiate a request for a change In specifications by letter to Manager, Manufacturing Operations. 2. Upon written notice of the acceptability and estab lishment of the revised specifications by Manufac turing Operations, Marketing may place orders for product to meet the new specifications. 3. Conversely, Manufacturing may desire to change a product specification and will write a request for approval to Marketing. Upon written approval by Marketing, the change will be established in the Product Specifications Manual. 4. Ho shipping orders for products outside authorized specifications will be honored by the manufacturing plants without authorization from Manufacturing Operations. With respect to new processes. Shell Development brought out the benefits to be gained by a Shell Development representative(s) most familiar with the process following the process to Shell Chemical for initial operation. This applies for small projects as well as major ones since frequently the effects of all process variables have not been com pletely worked out for a small project where exhaustive research is not Justified. In some cases, the feel, smell, or looks of a product may be extremely significant but are often difficult to convey by the written word. For example, manufacturing of Adhesives might be compared with hnn) baking where, as is well understood, two cooks working from the same rec ipe can turn out entirely different types of cake. The visit to North American Aviation on May 17, 1954, is covered in a separate memorandum of discussion. t w e. - KJ i')pn t\ cl) B. H. Cummmiinnggs J BHC/WCW:mev W. C. Walker Si // ABS-018259 -7Distribution: Julius Hyman & Co. - Mr. J. G. Bejarano (4) Shell Development Co. - Messrs. B. S. Greensfelder*THis-coPY-FOR F. M. McMillan D. V. Elam R. M. Cole Shell Chemical Co. - Messrs. F. S. Svackhamer M. R. Sprinkle R. L. Kittle W. F. Schoenthaler ABS-018260 -8 - DENVER PLANT EPON ADHESIVE VI MANUFACTURE OPERATING PROCEDURE At least 24 hours before a batch of Adhesive VI is to be nixed, tvo drums of Epon 834 should be placed In the hot-box to liquefy the resin. The adhesive is prepared by mixing by weight in the blender as follows: 1. Allyl glycidyl ether (AGE), 177 lb. 2. Acetamlne orange, 147 grams (in cabinet, M. Smith's office). 3. Vinylite AYAF, 94 lb. 4. Tltanox, 8 lb. 2 ox. Mix ten minutes, turn on steam, and add 5. Epon 834, 500 lb. (1 drum). 6. Asbestos floats, 353 lb. This should be added in 100 lb. increments, seeing that each addition is mixed in before adding more. 7. When asbestos has almost all been pulled into the mix, stop the blender, scrape down the walls to the point where the blades reach, mix again for a few minutes to be sure all the asbestos has been worked in, then add 8. Epon 834, 500 lb. Allov to mix overnite, reversing the direction of ' agitation occasionally. Inspect to see that no unmlxed resin remains on top. If there is such, work it in with the scraper by hand. 9. When batch is ready, bolt down the blender lid, set up the scale and pour the adhesive into cans as specified. Containers will be 1-gallon, holding 8 pounds, or 5-gallon, bolding 40 pounds net. Label each can with designation "VI" and batch number. Each batch is split into lots "A", "B", and "C"; lots to be 500 lb. About at the middle of each lot, take two 1-pint samples properly labeled, which are delivered to R. Beilin in the Laboratory. 10. Toward the end of the packaging, it will be necessary to run the blender in the proper direction to push out the bulk of the adhesive. Hie last of the contents will have to be scraped out. If another batch of the same adhesive is to be made, the blender need not be scraped clean. If another type of adhesive will be made next, scrape out as much of this batch as possible. Elapsed time/batch is 20 hours at the beginning of packaging. abs-oi8261 -9 - EEKVER PLANT EPOH ADHESIVE VIII MAKUFACTOBE OPERATOR PROCEDURE At least 24 hours before a batch of Adhesive VIII is to be mixed, check to see whether the Epon 828 to be used Is In bung or open-head, drums. Open-head drums can be used cold; bung type drums must be heated In the hot box before they will run well. The adhesive is prepared by mixing by weight in the blender as follows: 1. Allyl glycidyl ether (AGE), 20 lb. 2. Vinylite AYAF, 60 lb. (recently reduced from 80 lb.). 3. Titanox, 7 lh. Mix a few minutes and add 4. Epon 828, 500 lb. (1 drtxn). Turn on steam, mix 10 minutes, and add 5. Asbestos floats, 300 lb. This should be added in 100 lb. increments, seeing that each addition is mixed in before more is added. 6. When asbestos has almost all been pulled into the mix, stop the blender, scrape down the walls to the point where the blades reach, mix again for a few minutes to be sure all the asbestos has been worked in, then add 7. Epon 828, 500 lb. Allow to mix ovemite, reversing the direction of agitation occasionally. Inspect to see that no unmixed resin remains on top. If there is such, work it in with the scraper by hand. 8. When batch is ready, bolt down the blender lid, set up the scale and pour the adhesive into cans as specified. Containers will be 1-gallon, holding 8 pounds, or 5-gallon, holding 40 pounds net. Label each can with designation "VIII" and batch number. Each batch is split into lots "A", "B", and "C"; lots to be 500 lb. About at the middle of each lot, take two 1-pint samples properly labeled, which are delivered to R. Beilin in the Laboratory. 9- Toward the end of the packaging, it will be necessary to run the blender in the proper direction to push out the bulk of the adhesive. The last of the contents will have to be scraped out. If another batch of the same adhesive is to be made, the blender need not be scraped clean. If another type of adhesive will be made next, scrape out as much of this batch as possible. Elapsed time/batch is 20 hours at the beginning of packaging. ABS-018262 - 10 SHELL DEVELOPMENT COMPAJff EPOH ALEBSIVE VI MANUFACTURE 1. Dissolve Vinylite AXAF in AGE at 50C 4 to 10 hours in a stirred, heated kettle. Fifteen parts AGE and eight parte Vinylite are used. 2. Add 0.03 parts acetamine orange. 3. To a Read mixer charge 85 parts Epon 83^ Heat to 70C. 3. Add 30 parts asbestos floats and 0.7 parts Titanox as a preground 60f> concentrate in Epon 83^* 6. Mix for two hours at 70 to 75 C. 7. Add the 23 parts AGE-Vinylite mixture from the stirred kettle. 8. Blend for 30 minutes. 9- Drum at 50-6oC. AES-018263 - 11 SHELL DEVELOPMENT COMPANY ETON ADHESIVE VIII MANUFACTURE As a preliminary step in the formulation, a 60Jtw concentrate of titanium dioxide (Titanox) in Epon 628 Is made by premixing the two mate rials and then passing the mixture through a Norehouse Hy-R-Speed mill using GO grit carborundum stones* Four passes at about 80-100C are nor mally required to obtain a smooth mix and sufficient material is prepared for several adhesive runs. In a 100-gallon tank, 500 pounds of Epon 828 is heated to l60C and 30 pounds of Vlnyllte AXAF is added. The mixture is then stirred at this temperature for one hour to obtain complete solution of the solids in the liquid. As the Epon received at Emeryville occasionally contains foreign particles, the solution obtained after dissolving the Vinylite is filtered hot through an 18-mesh screen. Approximately one-half of the filtered Epon-AYAF solution is transferred to the 150-gallon Read mixer and to this is added 5-8 pounds of Epon-Tltaaox concentrate. The tempera ture at this point is about 100-110C, and after mixing for a few minutes, 150 pounds of asbestos floats is added in portions over a period of 15-30 minutes. Mixing 16 continued for 2 hours at 90-100C, and'during this time the direction of the blades is reversed at 15-minute intervals to insure thorough blending. With completion of the above period, the re mainder of the Epon-AXAF is added to the mix and blending continued for one hour. The charge is completed by the addition of 10 pounds of allyl glycldyl ether, and after a final mixing period of 30 minutes at 80-90C, the finished adhesive is discharged to suitable containers. ' As no corrections are made for the amount of Epon 828 introduced with the addition of the Titanox concentrate, the fined formulation has the following composition: Epon 826 Vinylite AYAF Asbestos Floats Allyl Glycidyl Ether Titanox 100.5 parts by weight 6 30 2 0.7 ABS-018264