Document V4XqL8YEzzrJVxVOkMv0qGvo

BUSINESS CONFIDENTIAL PROJECT REPORT RECEIVED JAN 3 1968 R* N. Wheeler SUSPENSION VINYL RESINS: PROPOSED PROCESSES TO PRODUCE A POWDERBLEND COPOLYMER RESIN FOR THE SOUND RECORD MARKET author ; A. A. Peterson supervisor! Dean E, Richardson datei December 20, 1967 PROJECT MO.I 333 A 10 F|LE HOi 8645 SUMMARY The advent of powderblend feed for the manufacture of vinyl records has called for a vinyl chloride/vinyl acetate copolymer with good dryflov properties. To date, there is but one resin on the market with the excellent properties required, Borden VC 113 FSP. The Process-10 resins now made by UCC have poor flow properties due to static electricity and mediocre particle sphericity. During the past few months, three processes have been found to produce good particle sphericity: 1) Continuous addition of vinyl chloride monomer withheld from the initial charge; 2) Batch addition of vinyl chloride monomer withheld from the initial charge; 3) Use of Elvanol 50-^2 as suspending agent instead of polyvinyl pyrrolidone . All three methods have been successful in producing spherical particles. However, methods 1) and 2) still have the disadvantage of static electricity. Another disadvantage of method 1) is the fact that the north line autoclaves (those designated for VCl/VAc copolymer manufacture in 1968) are not equipped with meters for continuous vinyl chloride addition. Dis advantages of method 2) would be the additional attention required by opera tors, additional hold-up time in the charging of autoclaves, and possible failure of pumps during the critical addition period. As stated, the Process-10 resins made with methods 1) and/or 2) still have the static electricity problem. This report includes a compari son of three additives which would alleviate the static problem and permit free flow of the spherical resin particles. Efforts are also being made by Dr. J. E. Glass of the Tech Center to study the effects of silicones and caprolactone/ethylene oxide copolymer upon the static electricity problem with the polyvinylpyrrolidone. research and development department CHEMICALS AND PLASTICS UNION CARBIDE CORPORATION TEXAS CITY, TEXAS UCC 070280 2 As to the relative merits of the antistatic additives, the Magnesol and Cabosil were almost identical. On those resins which had spherical particles, but had the static problem, 0.1# of either the Magnesol or Cabosil greatly improved the flow properties. Two things are evident from these tests: l) excellent dry flow requires spherical particles whether or not the static electricity is neutralized, and 2) the Process-10 copolymer will not flow well without the antistatic additives in spite of the improved sphericity gained by batch or continuous vinyl chloride addition. " An additional interesting and promising fact was found during the writing of this report. The Magnesol can be added at concentrations of 0.3#-0.5# to antistatic (with Catenae SN) VYCR-10 (a sticky, poor flowing powder) and the resin flows better than the original VYCR-10 without the Catanac. This might prove valuable for RCA technical service or, for that matter, our own handling of the resin, if RCA is converted from VYAR-2 to the Catanac-modified VYCR-10. The Process-10 resin with continuous vinyl chloride addition (or with 65# of the VC1 added batchwise) with the aid of 0.1# Cabosil or Magnesol was equivalent in flow properties to the Borden resin. The VYCR made with Elvanol 50-1-2 was equivalent to the Borden resin without the aid of antistatic materials. During the next few weeks, efforts will be made to scale-up the Elvanol 50-1-2 recipe for a powderblend record resin from the 10-gallon to the 600-gallon autoclave. INTRODUCTION In July of 1966, pictures were made by the author of several blends of VYCR-10 resin. It was noticed at this time that a few blends of this resin had spherical particles. Most of the Process-10 copolymer (and homopolymer) observed had an irregular non-spherical character. At that time there seemed to be no particular desire for such a characteristic, since most resin was being compounded into pellets. Daring the pact year, interest grew toward increasing the dry flow properties of the record resins for powderblend formulations in which good dryflov properties are critical. After looking at several blends of VYCR-10, it was decided that an increase in particle size and a decrease in fines should help the dry flow proper ties. After extensive efforts, consistently satisfactory blends were made with a good improvement in flow properties due to the larger particle size gained through lower agitator speed in the pilot plant and production auto claves . In the past few months, Bound Brook personnel have expressed a desire for an even greater improvement as the flow properties were still not as good as the best competitive resin, Borden VC 115 FSP. It was noted that this resin had regular, very spherical particles which accounted for the excellent flow properties. Thus, attempts were made to improve the sphericity of the UCC record copolymer. UCC 070281 5 DISCUSSION Three successful methods have been discovered to produce the desired particle geometry. The first came from pictures of old VYCR-10 blends made in late 1965. A report by H. H. Savage concerning the reduction of free vinyl acetate content of copolymer by continuous vinyl chloride addition mentioned that "the appearance of the resin made at 75^> and 73^ initial vinyl chloride was appreciably different from the others--the particles were very spherical and had a wider size distribution." From this observation a pilot plant run was made using the continuous vinyl chloride .addition (controlled by feeding the remaining vinyl chloride to hold the reaction pressure constant). The initial vinyl chloride/vinyl acetate ratio was 75/25 and the final ratio was 87.6/12.4 for a VXKF-10 copolymer. The result was a resin with greatly improved sphericity and higher bulk density (also noted in H. H. Savage's report). The 1968 schedule for production of VCl/VAc copolymer calls for the north line autoclaves to be used. These autoclaves are not equipped for continuous vinyl chloride addition. Thus, several pilot plant runs were made by withholding from the initial charge a fraction of the vinyl chloride, which was then batch-charged after a few hours of reaction had occurred. It was found that the larger the fraction of vinyl chloride withheld, the more spherical the particle. It is known that the suspen sion polymerization of vinyl acetate produces very spherical particles. Perhaps the lower vinyl chloride/vinyl acetate ratio at the beginning of the experimental copolymer runs is what causes the improved sphericity. A third method for producing improved sphericity of copolymer was found: the use of Elvanol 50-42 polyvinyl alcohol to replace polyvinyl pyrrolidone (FVP K-90) as the suspending agent. The Borden Chemical Company has been producing a polyvinyl alcohol in the molecular weight range which should theoretically make a good suspending agent, but for some reason they stopped marketing it. We suspect that they are still using it internally as a suspending agent for the Borden vinyl chloride/vinyl acetate copolymer series. A duFont product, Elvanol 50-42, is the most likely polyvinyl alcohol on the market with the molecular weight suitable for a good sus pending agent. Elvanol 50-42 has been used in the past in pilot plant work, but the evaluation was not a high priority item. A statistical series of polymerizations was made in the 10-gallon stainless steel autoclave using Elvanol 50-42 with the new Pfaudler agita tion system. After defining the limits of the experiments with several ag glomerated runs due to excessive agitation, eight successful runs were made studying four operation variables. The results will be reported separately from this report. All of the runs produced very spherical particles. Due to the study of the operating variables, all runs did not produce the de sired particle size, but the sphericity attained was equivalent to the Borden VC 113 FSP. This can be seen In the accompanying photomicrographs. Another feature of the resin produced by the Elvanol 50-42 system was the lack of electrostatic charge, which is a great shortcoming of the Process-10 resin. The study with the antistatic additives showed no im provement of funnel flow upon mixing the resin with them. It was quite evident from the drying, screening, and microscope visual observation that the static charge is immeasurable by sight or from resulting flow proper ties. After discussion with Dr. J. E. Glass, who theorized that the poly vinyl alcohol should be more difficult to remove from the resin particle ucc 070282 4 than the polyvinylpyrrolidone, it is thought that the polyvinyl alcohol may have some antistatic properties. Dr. Glass will be studying the effect of adding silicones or caprolactone/ethylene oxide copolymer to the polyvinylpyrrolidone recipe as a solution to the static problem with Process-10 resins. Work is now being started to reproduce the Elvanol 50-42 recipe results in the 600-gallon autoclave and for possible scaleup efforts to production autoclaves. * From a suggestion by Bound Brook personnel, comparative funnel flow tests were made with resins produced by the continuous and batch vinyl chloride addition process, resin produced by the Elvanol 50-42 system, the Borden VC 113 FSP, VSKF-10 (B-6) made with the lower agitator speed of 104 rpm, and a recent production sample of VYCR-10 produced at the normal high agitator speed of 120 rpm. The greatest improvement in flow properties upon the addition of an antistatic material was realized by the resin produced by the con tinuous vinyl chloride addition. The resulting flow properties were comparable to the Borden VC 113 FSP (without antistatic material). The next best improvement was the resin made with the largest amount of vinyl chloride withheld and charged by batch addition. Comparing these improve ments with the sphericity of the resins, it would seem that the static of the Process-10 copolymer can indeed be overcome with antistatic materials, but to give flow properties comparable to the Borden resin, the sphericity must also be present. This can be seen in the minor improvement in those resins tested with poor sphericity. The Magnesol, Cabosil, and carbon black gave similar improvement with few exceptions. Carbon black would not be a good additive as it has been found to segregate. For that matter, segregation may occur with the other materials also, but this has not been detected. All three materials are very dusty and would be difficult to handle, unless added while the resin was still wet (for example, between the Strong-Scott Solidalre and fluid-bed drier in the Texas City Number 7 Drying System), CONCLUSIONS Through the examination of the flow properties and particle geometry of the resins and the effect ofthe three antistatic materials, the following conclusions have been made: 1) At present, the Elvanol 50-42 suspending system for copolymers gives the best candidate for a powderblend record resin. The scale-up of this recipe in the next month will determine the success of this approach. The Elvanol 50-42 process produces a resin particle unlike the Process-10 copolymer in that is possesses a great degree of sphericity, virtually no electrostatic charge, an 8-to 10-pound higher bulk density, and excellent flow properties. 2) The continuous vinyl chloride addition method with Process-10 gives improved sphericity over the normal Process-10 procedure, but the resin still has an electrostatic problem. This can be overcome with the ad dition of an antistatic material at low concentrations. With the anti static material and improved sphericity the flow properties are equivalent to those of the Borden VC 113 FSP, and with the resins produced in the 10gallon autoclave with Elvanol 50-42. ucc 070283 5 5) Batch addition of the majority of the vinyl chloride after two or three hours of reaction produces resin similar to that produced by the continuous addition method. When more than 50$ of the vinyl chloride is charged in the initial charge, the resin produced begins to look more like the normal Process-10 resin, possessing mediocre sphericity. 4) Both Magnesol and Cabosil inhibit the electrostatic charge of the co polymer at very low concentrations. It was also found that the Magnesol will aid^he flow properties of copolymer resin containing Catenae SN, without which the resin will not flow at all, as was experienced by a recent production run of Catanac-modifled VYCR-10 (QEX-1325). The only recommendation concerning future work is a study to determine the feasibility of adding Magnesol to the Catanac-modified VYCR-10 for RCA. A sample has been sent to Steve Krumm at Bound Brook for evaluation. Scale-up studies of the Elvanol 50-42 recipe are being made at this time. No future work is planned with the continuous or batch addition of vinyl chloride at this time. Dr. J. E. Glass is ex perimenting with caprolactone/ethylene oxide copolymer and silicones for the inhibition of the electrostatic charge accompanying the Process-10 copolymer. EXPERIMENTAL The procedure for determining the flow properties of the resins were as follows. 1) Weigh the resin in a tared dry jar. 2) Add the antistatic material to the concentration desired. 3) Record total weight. This should be about 5 grams. 4) Mix the resin and additive vigorously by shaking with the jar-lid closed. 5) Transfer all the material to the steel funnel (see accompanying drawing) holding a finger at the bottom of the funnel to prevent leakage. 6) With a stopwatch, measure the time required for the resin to flow through the l/8" aperature at the bottom of the funnel. Record this time. 7) Repeat test once. 8) Divide average time in seconds by resin weight in grams. Report this number. uce 070284 6 ACKNOWLEDGEMENTS The author's thanks go to D. E. Richardson for his suggestions and help. The pilot plant work was co ordinated by R. C. Robinson. Appreciation is due technicians E. W. Dennard, T. W. Oliphint, and J. R. Lytle for their work and suggestions at the pilot plant. Technicians W. J. Tumlinson and J. T. Canon assisted with the resin flow testing. Special thanks go to Monty Montague for the reproduction of the photomicrographs. BIBLIOGRAPHY 1) Peterson, A. A., "Suspension Vinyl Resins: Pilot Plant Studies of Process-10 Copolymer Process Variables and Subsequent Scaleup Efforts", Project Report, July 19, 1967, file number 3^7B10. 2) Peterson, A. A., Letter to S. Krumm, October 16, 1967. 3) Savage, H. H., "VYCR-10 Copolymer Production: Reduction of Free Vinyl Acetate Content by Continuous Vinyl Chloride Addition", Status Report, December 29, 1965. NOTEBOOK REFERENCE Vinyl Chloride-Vinyl Acetate Copolymers, Number 9107 1 AAP, pages 1 - 23* Attachments A. A. Peterson ucc 070285 APPENDIX Table I: Funnel Flow Test Results Figure I: Drawing of Resin Flow Funnel Figure II: Photomicrographs of Resin Examined Figure III: Photomicrographs of Resin Examined UOCnc 070286 TABtE r FUNNEL FLOW TEST RESULTS Sample VYCR-10 (B-208) VXKF-10 (QEX1315) Borden VC 113 FSP 11.67-7-10S 11-67-B-lOS 11-67-5 FED 11-67-fl FED 11-67-9 FBD 11-67-10 FBD VXKF-10 (B-6) Remarks Normal Process-10 recipe at 120 rpn, particle sanevbat larger than normal Pilot Plant run with con tinuous VC1 addition Best competitive record resin with respect to flow properties Particle Size, Microns 132 150 151 Elvanol 50/42 recipe in 10-gal. autoclave from statistical series Elvanol 50/42 recipe in 10-gal. autoclave froo statistical series of VC1 charged at end of third hour (600-gallon auto clave) 32 of VC1 charged at end of fourth hour (600-gallon auto clave) 6j% of VC1 charged at end of second hour (600-gallon auto clave) 325 of VC1 charged at end of 2nd hour (600-gallon auto clave ) Normal Process-10 recipe at 104 rpn 175 107 nit 161 162 181 172 Size Distribution 0.59 1.18 0.17 0.61) 1 -31 1.16 0,90 1,12 0.88 o.i)9 Appearance Better than most Process-10 resins, but no sphericity Good sphericity but poor distribution Very smooth spherical-toegg-shaped particles, few agglomerates Very smooth, spherical-toegg-shaped particles, few agglomerates Very smooth, spherical particles, very few ag glomerates Fair sphericity, poor distribution Slightly more spherical than normal Process-10 resin Good sphericity, but poor distribution Just slightly smoother than normal Process-10 resin No sphericity 12 3Flow Time in Seconds/Gram for 50 Grams Resin Control Carbon Black Cabos11 Magnesol 2-31) 0.5* 2.23 0.1* 2.22 S-ff 2.24 24 m27DT i 2.20 1.65 1.56 1.57 1.51 1.65 1.57 1.49 1.66 1.97 1,41 1.44 1.63 1.61 1.64 1.76 1.58 j .79 1.93 2.46 4.44 1.98 2.06 2.18 1.92 2.05 2.00 2.26 1.98 2.06 2.09 2.19 2.22 2.24 2.26 2.07 2.06 2.46 1.66 .08 1.70 1.65 2.04 1.71 2.12 2.14 2.03 i.94 2.20 2.12 2.28 2.23 2.20 2.02 1) Distribution is (P - F ^5^)/ frco sieve analysis 50 grams of resin was tested in duplicate on 1/8" aperture funnel as described on accomii&nylng drawing 3) Industrial Grade magnesium silicate from Waverly Chemical Comijany, Inc. ucc 070287 rest:; flow fuituzl FIGURE I a-5/a:: ucc 070288 4 7~7~/0s nade JO ai/tochut, U - 67-6-/es made in /O *a/. a\t4e>d?vc 'i+h E/ySnd So/fZ *'W ENiftt) S0/4Z. fill pic4*re$ 40X fr>3^ni'f>`Cerhbt\ Border) VC.H3 Fsp -- --------- ---- fr* ^ <juc4in audcc/aves d4 /ZO rpm yi KF-/o (8-0 made i*> prcdt<4)6r} jv+edaves 94 fc4 rf> rr, ucc 070289 fh 6 7mwJ* i* too ?/ ivttcfave w#VX Si % 0-r vc/ ay 3V hr. ucc 070290 DISTRIBUTION Texas City Mr. J. H. Barrett Mr. 0. T. Carlisle Mr. ft. J. DeLuca/Mr. L. G. Peyton Mr. J. H. Field Mr. N. A. Gimber/Mr. W. D. Bush Mr. R. J. Greer Mr. J. L. Hockersmith/Mr. R. M. Arnold Mr. K. L. Meisner/Mr, R. C. Robinson Mr. D. E. Richardson Technical Center Dr. F. E. Bailey, Jr. Dr. J. E. Glass Dr. C. W. McGary Tarrytown Mr, W, H. Bauer/Mr. P. T, McCoy Librarian New York Office Mr. L. D. Harris Mr. J. R. Wilkinson Bound Brook Plant Mr. A. J. Costantin Mr. S. Krumin Mr. R. J. Stockman South Charleston Plant Mr. R. N. Wheeler Information Retrieval UCC 070291