Document 99R5R45BmYY7L7DZ9RBeKzvZ6

,<L /) A AMERICA!i VISIT 10th to 2kth May. 1966. RECEIVED APR lb to R. N. Wheeler INFORMATION GATHEiSD 1. Puraoso of Report A report has already been issued retailing information given to Union Carbide Corporation during tho above visit. The main purpose of this report is to describe Union Carbide's plant and methods, and critically compare them with our own. A subsidiary purpose is to report on ny brief visit to Bound Brook. At the request of Mr. J. M. Austin of the Union Carbide Corporation, information which could be of use to competitors of Union Carbide, for example F.W. Goodrich via British Geon, has been omitted from this report. It has been passed to those whom it concerns by private communication. 2. Contents of Report General description of Plant aid Equipment for P.V.C. manufacture Section 3 Pages 2-9 Comparison of Methods of making Suspension Resins Section 4 Pages 9 -16 Bulk Handling of P.V.C. Semi-scale Equipment Section 5 Pages 16-19 Section 6 Pages 19-21 Organisation Section 7 Pages 21 Possible Applications at Aycliffe Visit to Bound Brook Section 8 Pages 21-23 Section 9 Pages 23-25 Appendix i ir e 039827 -2- 3. c^ V 3.2 General Description of Plant and Equipment f_or JEV/jC^ tonmjicture The Suspension Resin Plant at Texas City has been installed alongside the already existing non solvent Resin Plant and shares mary facilities with it. The Suspension aixl Won-Solvent Plants are ad ministered as one unit and will therefore be described together. Reference should be made to the attached sketch No. 1. Non-solvent System VYNW and VYNA are still manufactured in the 6 northern autoclaves of the old Non-solvent Resin Plant. The 6 southern non-solvent autoclaves are converted, or being converted, to extra viryl chloride storage tanks. Union Carbide have made one major alteration to the method of manufacturing non-solvent resins, which in my opinion has made the whole operation very much more safe and controllable. They have changed their catalyst from D.A.P. to I.P.P. They manufacture this latter themselves in the old D.A.P. plant as a solution of approximately 20?o concentration in nonane (a naphtha fraction). They circulate this by pumping round a header, keeping it at between -5C and 0 C by brine refrigeration, and meter it automatically directly to the autoclaves. Peed mixes, containing catalyst, are no longer made up, uncatalysed monomer being fed directly to the autoclaves. As only half the autoclaves are being used, the monomer recovery system is shared with the Suspension Resin Plant. All recovered monomer from both suspen sion and non-solvent processes is used up in the non-solvent process. Non-solvent resin is recovered, dried, and handled in the same way as previously. Suspension Resin Plant The Suspension Resin Plant consists of twenty-one 5,700 U.S. gal. autoclaves (21.6 mr), all of which are glass lined and fitted with a Pfaudler-typo agitation system. They are driven at 120 r.p.m. by a 50 horse power motor. They are highly automated and the reaction system is run by two operators. -3- 3.2.1 The autoclaves are arranged as three arms of a cross, with the Control Room, which is pressurised, and suspending agent manufacturing facilities at the centre. Future expansion will be along the other arm. Five of the six east autoclaves were commissioned in the summer of 1962, together with the blowdown tank and the Bird centrifuge. One extra autoclave was com missioned in January, 19&3 5 the other fifteen were commissioned in January 1965 AH except the six south side autoclaves are rated at 220 p.s.i., but the six southern ones are rated at 330 p.s.i. to enable the special Type 11 copolymers to be made. Distillation is not carried out from autoclaves. At the end of the reaction, the autoclave contents are blown down to one of four blowdown tanks, from which monomer is recovered via the non-solvent and recovery system. These blowdown tanks are 10,100 U.S. gal. capacity (8,400 imperial gal.) and agitated by side entering agitators. After blowdown, slurry is trans ferred by pumping to one of three atmospheric blending tanks, each of 55,000 U.S. gal. capacity (46,000 imp. gal.). These are nitrogen purged. Dilution to 10-1596 solids content is carried out in these tanks. From the blending tanks, the 3lurxy is fed to the drying system. This consists of Bird centrifuges feeding 2 Louisville type rotary dryers. They are not always able to dry all the suspension resin in these dryers and have brought into use non-solvent plant flash dryers for drying the more porous resins. After drying, the resin is sieved through a Derrick sifter and blown to the bulk handling system. This will be described later. The individual items of plant will now be described in more detail. Autoclaves The original six autoclaves were made by Ffaudler of glass-coated steel of 5700 U.S. cal. capacity (4,600 Imp. gal.). They wanted 25mr autoclaves (6,600 U.S, gal.), but Ffaudler were not able to provide the requisite length of overhung shaft, so geometrically similar autoclaves with 1 ft. shorter shafts were bought. ucc 039829 ucc 039 S3 -*- One of these six is actually of 4,600 U.S. gal., having been transported from South Charleston, where it was 'used for experimental work. It has an infinitely variable spued equipment up to 125 r.p.m. The 15 new autoclaves were all bought from Glascote, who offered better delivery than Pfaudler. They standardised on 5,700 U.S. gal. All autoclaves are cooled by brine at -5 C. The 4,600 gal. autoclave has a falling-film jacket; all the others have totally filled jackets, with brine entry at the bottom of the jacket through nozzles set to give a swirling action. The autoclaves are heated up to reaction temperature by hot brine. This is achieved by diverting the brine feed through a heat exchanger on a recycle system, sec sketch No. 1. An interesting feature was the autoclave shaft seals. Seal pressure is balanced by pressurised glycerine, but no circulation is used. They were asked if these did not get very hot - they replied that they had checked and they hardly heated up at all. The drives of the autoclaves supplied by tfaudler are of Pfaudlor design. The drives of the Glascote auto claves are by Philadelphia Mixers. These are much more compact and are preferred. Autoclave Instrumentation For each autoclave there is: An autoclave temperature controller, namely a Taylor Cascade temperature and pressure controller. These are electrical instruments with pneumatically relayed signals. The brine outlet temperature is recorded and there is a high temperature alarm. Remote controlled slurry motor valve operator. Remote controlled agitator slow speed and fast speed operator. Remote controlled hot brine valve operator. These are located in the Control Room on a control panel. Twelve autoclaves are accommodated on one panel -5- and fifteen on the other. On each panel there are the following meters, which are connected to all the autoclaves on that panel:- Vinyl Chloride meters - these are Fisher Porter turbine meters and are accurate to *1$. They measure a set number of U.S. gallons into the autoclave. Graphs are provided to con vert weight into volume at the temperature of the vinyl chloride being charged. These are very reliable ^on the whole, but on the few occasions when they go ' haywire', it can take some time before this is found out, as no indication other than by the process behaviour is given. They intend to circulate vinyl chloride from a weigh tank and use this to check the meter before and after every charge. Water Meters These are Foxborough orifice-type totallising meters, and like the vinyl chloride meters, shut the valve to the autoclave when the correct amount of water has been charged. Process water is de-ionised and filtered steam conden sate and is usually at a temperature of about 50C. Suspending Agent Meters These are Vane type meters, called the Remote Repeat Batch Controllers, made by the Badger Metering Manufac turing Company, Milwaukee, Wisconsin. Vinvl Acetate Meters These are Fisher Porters like , ` the vinyl chloride meters. Continuous Vinvl Chloride Addition Meters The southern line high pressure autoclaves arc set up for continuous vinyl chloride addition, and this is metered in by Pottermeters. These only measure the amount added. Control ia on pressure by demand, similar to the Aycliffe system. NOTE: Ingredients charged in small quantities are added manually into a cat- pot in the process water line to each autoclave, mostly D.L.P. catalyst. 6- - 3.2.3 Blowdown Tanks There are four 10,100 U.S. gal. stainless steel blowdown tanks, each of which will withstand 50 p.s.i.g. working pressure or full vacuum, and are agitated by side entering 7.5 horsepower, 280 r.p.m., propeller agitators, . made by Lightnin Mixers. At the centre of the top head of each is fitted an entrainment separator and monomer is recovered from them via a 15 p.s.i.g. vent header. They are installing a 2 p.s.i.g. vent header, but it is not yet in service. They have a lot ox" trouble with carry-over of resin foam from these tanks into the recovery system, which from ny observations was similar to what we were experiencing when we were using this system. I consider that it is a major cause in keeping their vinyl chloride efficiencies down to the 80-85% level and so recommended that they install trap tanks like us. Slurry is continuously recirculated in these tanks and pumped as required into the blending tanks. 3*2.4 Blending Tanks These are 55,000 U.S. gal. (46,000 Imp. gal.) stainless steel, atmospheric tanks, in which the slurry is blended and diluted to the correct consistency for feeding to the Bird centrifuges. They are approximately 15 ft. high by 25 ft. diameter, as near as could be judged by eye. They were agitated by top-entering agitators with lifter type impellers, designed to give overall circulation. Two of these had conical bottoms and one a flat bottom. The conical bottoms wore specified after experience with the flat-bottomed one. These tanks are nitrogen purged and are fitted with pressure vacuum relief heads, set at 1 oz. per sq. in. pressure and 5 lbs. per sq. in. vacuum- These heads are made by Black-Sivails and Biyanson Inc., Oklahoma City. Slurry is recirculated in the slurry blending tanks and pumped to the Bird with Durco pumps. These are centri fugal pumps with a special claw-type impeller and have Teflon asbestos packing on the gland with blowback. 3.2.5 Bird Centrifuges All except one of these are the same as ours. The new one has what is known as a "contoured" bowl. This only means that it has both a parallel and a conical section for improved clarification. It gives an output of 6,000 lbs. per hour on a dry resin basis of VYCR-10, at a feed of 15% Vw. of solids, which is a 50% improvement on a non-contoured uco 039333 -7- bowl. They had similar trouble with theirs as we have experienced with ours and have overcome them by fitting deflector plates with close clearances to the beaters, 3*2.6 Rotary Dryers They have two Louisville-type rotary dryers, similar to A.N.I.C., made by the General American Transportation Corporation, One is 7 ft, x 50 ft, and the other 3 ft. x 4-0 ft., and they intend to go still bigger. borne details of the 8 ft. x 40 ft. are as follows:- C Speed 4.5 r.p.ra. Resin: 3>000 lbs. per hour Water; 1,300 - 1,600 lbs. per hour Total: 4,300 - 4,600 lbs. per hour Dwell Time: 30 - 4O minutes Heater Control: Cascade on steam. They dry all types of resin satisfactorily on the rotary dryer, but even with this dwell time cannot dry glassy polymers to less than 0.5% heating loss consistently. They are using the old non-solvent flash dryers success fully for drying porous suspension resins. 3.2.7 Sifters U.C.C. are using the Lferrick sifter almost exclusively for sifting suspension resins after flash drying- This is not a plan-sifter. They did use plan-sifters but had a lot of trouble with sealing between screens, as we do with our Allis Chalmers. They find the Derrick sifter much more satisfactory from this point of view. It is an angled vibratory sifter. For most of their products, they sift through stainless steel mesh. There is, however, an angle factor, i.e. larger holes are required than the particle size, for example, 50 mesh bolting cloth is used with approxi mately 500 microns opening for sifting 100-320 mesh material. They find this very good for coarser materials, but do not say whether it is as good on fine powders, i.e. in that it could let through oversize. -8- For our use, I would not be happy about using oversize screens, although this machine appears to be much simpler and more foolproof tlian our Allis Chalmer typo sifters. If a plant expansion at Aycliffe takes place, however, I feel that this sifter should be tested. Fines from the outlet of the Derrick sifter are blown to the bulk handling system. Coarse material is collected and reground. 3*2.8 Monomer Recovery System This is similar in effect to the Aycliffe system after the reconstruction in 195b, 57 and 58- Viryl chloride vapours from the suspension and non-solvent plants pass through the rectifying column, from which the bubble caps have been removed to prevent freeze--up. They then pass tlirough three low pressure condensers in series, the re fluxes from which return to the rectifying column. The vapours then pass to two compressors, which load and unload on demand. The compressed gas is condensed in the high pressure condenser, the condensate returning to the recti fying column via a condensate receiver and the gases being vented into the solvent recovery system of the Solvent iiesin Unit. Recovered moncmer is pumped from the base of the column to the recovered monomer tanks in the Tank Area, where water and resin (11) are drained off. It is re-used in the Non-solvent Plant only. 3*2.9 Suspending Agent Manufacturing Facilities They use several different suspending agents and, therefore, have relatively more suspending agent storage tanks. They have one dissolving tank, which is a 3>500 U.S. gal. fibre glass reinforced polyester tank, agitated by a side-entering propeller type mixer. Suspending agents, emulsifiers, and coagulants are all dissolved in this tank, but are stored in separate tanks. Fibre glass polyester was chosen to avoid iron contamination of calcium chloride solution. Heating and cooling i3 done by recirculating through external heat exchangers. Solutions arc centri fuged and pumped to storage tanks. Separate dual-compartment storage tanks are provided for 5 different types of solution. Suspending agent tanks are made of aluminium; calcium chloride tanks of polyester fibre glass, and the one for Polyox of stainless steel. A storage tank is also provided for chain transfer agent tri-chlorethylene in polyester fibre glass. 3.2.10 Catalyst Manufacture I.P.P. is manufactured as a 20$ solution in nonane. Details and procedures were obtained. 4. COMPARISON OP MPTHOES OP '..AKIhG SUSPENSION REBUS In this section, the operating methods, us ad. in the manufacture of suspension resins at Texas City will be compared with those in use at Aycliffe and at '..acker Chemie in Burghausen. The main difference between Texas Citymethods and ours is that theirs have been tailored to use the maximum amount of automation and the minimum amount of labour, whereas ours is virtually manual operation. 4.1 Main differences are as follows:- 4.1.1 U.C.C. wish to make the maximum number of runs between cleanings: they do not clean after every batch as we do, but after every ten batches approximately. 4.1.2 In order that this can be effected, they run their autoclaves at 75$ full instead of 90$ as we and the Germans do. In addition, they do not distil from the autoclaves but blow down into blowdown tanks. They have two Pfaudler rinse valves in each autoclave, and with these are able to keep the heads of the autoclaves clean and the nozzles do not get blocked. I inspected an autoclave after the sixth run, after it had been washed, and it appeared on the surface as clean, if not cleaner than ours do after they C have been cleaned. 4.1.3 Since their methods do not involve much manual work between batches, their philosophy is to decrease reaction times using relatively lower charge weights and obtain greater production by increasing the number of batches made. Our philosophy - and also the German philosophy - since a lot of manual labour is used between batches in the turn round and cleaning, is to increase autoclave loadings and produce the extra output by getting a greater yield per batch. 4.1.4 Union Carbide charge their D.L.P. catalyst into a cat. pot in the water charge line and wash it in with process water. Their process water is at about 50 0, being de-ionisod steam condensate, and they heat up by using hot brine in the jackets. This is nearer to what Wacker -10- Chemie do than us. Vacker charge the catalyst into the water through the open manway and heat up by hot water on the jacket. We suspend the catalyst in a polyethylene bag and charge hot water so that no heating up is required. 4.1*5 Apart from one autoclave, Union Carbide have adopted a variant of YYackers fully filled jacket, butthey still use , cold brine for cooling. They heat up autoclaves by passing the brine through a heat exchanger on a closed system, whereas Wacker heat their water by means of a steam and water mixing valve. Y/e con only cool with our jackets, which are of the falling film type, and save heating up time by charging hot process water at such a temperature c that when it is mixed with vinyl chloride, it is on reaction temperature. Both methods have their limitations and a combination would be ideal for maximum flexibility. 4*1*6 Union Carbide, like us, prefer I.P.P. to A.C.S.P. as a high activity initiator or catalyst. They have used it for a considerable period and treat it with much less concern than we do. They say that the stories about its shock sensitivity are not true, and that it is no more sensitive to shock than D.L.P. They make it themselves and pump it about and meter it in a very convenient manner. If it had not been for reports of the shock sensitivity of this material, we would have adopted a similar system, but have installed the present, more inconvenient method of handling because of our fears in that direction. 4.2 Comparison of Procedures 4.2.1 Charging U.C.C. With the agitator on slow speed, set the required amount of water on the water meter and start charging. Immediately afterwards, start charging suspending agent components and chain transfer reagents, if specified. These are injected into the water charging line. YYhon approximately half the water has been charged, stop the water flow, charge catalyst and solid component to the cat. pot. Purge the air in the cat. pot with water and continue to charge water through it. A hooter indicates when the water has been charged. Shut the water valve on the autoclave and open the viryl chloride valve. Set the correct amount of vinyl chloride on the vinyl chloride meter and charge the vinyl chloride. Ydien hooter indicates that vinyl chloride charge is complete, turn agitator to high speed. Shut the autoclave and manifold vinyl chloride valves. IJCQ 039837 -11- Ay cliffe Charge hot process water via the open nonhead to an external marker. Charge suspension agent components from a pump cart or paper bags, if solids. Agitate for seconds. Suspend D.L.P. Catalyst in a cat. bag in the top of the autoclave by the neans of hooks provided on shaft and manhead. Replace manhead, Pull a Nash vacuum to boil the water. Break the vacuum with vinyl ohlorido. Charge the remainder of the vinyl chloride with agitator stopped; by liaison between the charger on the weigh tank scales and the autoclave operator on the autoclave viryl chloride nozzle cock, shut nozzle cock and. header valves- Start agitator. V/acker Chemie Start agitator on slow speed. Charge cold process water through open manhead via an automatic water meter. Charge suspension system components from a pump c;jrt and paper bags, similar to Aycliffe. Charge catalyst into water. Replace manhead. Pull a vacuum similar to Aycliffe. Break vacuum with vinyl chloride (agitator stopped) and charge vinyl chloride through an automatic meter. Start agitator. 4*2*2 Heating up U.C.C. Set autoclave temperature control at reaction temperature and alarm 5 C below reaction tempera ture. Pull knob to start tho hot brine. When the auto clave temperature reaches the alarm set point, hot brine automatically shuts off. The alarm is reset 2W higher than reaction temperature. The controller takes over and no more attention is necessary from the operator. Aycliffe Hot process water having been used for the charge, the reaction is already at, or just above reaction temperature. Operator adjusts set point of instrument to bring onto reaction temperature. Wacker Chemie Essentially similar to U.C.C. except that the heating fluid is hot water not hot brine. 4.2.3 Control during the reaction U.C.C. & Wacker Chemie No control is required during the reaction, since charge weights are adjusted to be well within autoclave jacket cooling capacity. Aycliffe The operator must watch the temperature of the reaction, since the instruments, being adjusted -12- for proportional control only, havo an off-set. It is necessary for the operator to manually adjust for the off-set. Touurds tho end of a reaction, when the temperature begins to rise, the operator must manually add cold high pressure process water to the reacting mixture to control the kick at the eid of the reaction. He may well have to add this in seven to nine shots of 50 gallons each. 4 2.4 B1q\ down and Monomer Recovery U.C.C. When the autoclave reaches the prescribed pressure drop, open the manual valve on the slurry header; adjust the slurry motor valve by renote control from the control panel to blowdown the batch to the blowdown tank, using residual viryl chloride vapour pressure in the autoclave. Blowdown at the maximum rate possible, maintaining the blowdown tank pressure below 50 p.s.i.g. Reduce agitator speed to slow speed half-way through the blowdown. Rinse tire autoclave with the spray rinse valves and blow into the blor.down tank with vinyl chloride pressure from the charging header. Close slurry motor valve and hand valve. Avcliffe When the termination criterion of the batch is achieved, either pressure drop or temperature rise, start the compressor. Open the distillation cock and adjust to distil at the maximum rate which the compressor will take, i.e. maintaining the pressure in the pre-trap tank below 30 p.s.i.g.. Pull the autoclave to 10" of vacuum. Shut the distillation cock and header valve. Break the vacuum in the autoclave with inert gas and pressure up to approximately 75 p.si. with the agitator still running. Blovdown to the blowdown tank rinsing through with the Pfaudler rinse valve. Wacker Chemie Distillation essentially similar to Aydiffe, except that the pressure in the autoclave is never allowed to go below atmospheric. When distillation is completed, the autoclaves are vented to atmosphere and opened up and the slurry fed out by gravity, with agitator at slow speed, to the slurry holding tanks. Autoclaves are hosed out manually after the slurry has all drained. 4*2*5 Cleaning U.C.C. Union Carbide clean after approximately 10 runs. Therefore, their normal turn-round contains no cleaning step and charging commences immediately ucc 039839 -13- blowdown is completed. When cleaning is done, great care is taken so as not to damage the glass lining. Plastic scrapers and 3 M's cleaning doth are used and 8 to 10 hours are taken. NOTE: They have had several "biffs" already and some of the autoclaves have small pieces of glass off - this in less than six months. It was noticed, however, that the characteristic liquid level build-up on the agitator shaft seen on stainless steel did not occur on the glass lined shafts of their autoclaves. Aycliffe Autoclaves are cleaned after every batciv. After testing for 'No explosive vapour', the manheads are removed and the autoclave jet washed for half an hour. They are then cooled, air-blown, and manually cleaned on a rotational basis. Wacker Chemie Autoclaves are cleaned manually, after airing. No jet washer is used. A-*2.6 Slurry Handling U.C.C. After stripping to 15 p.s.i. in the blowdown tanks, slurry is pumped to -the slurry blending tanks, where it is diluted to the correct consistency for feeding to the Bird cen+rifuges. Blending tanks are filled full and solids content checked by means of a measuring cylinder before feed to the centrifuges is commenced. No header tanks ore used on the centrifuges. Aycliffe Slurry from the atmospheric blowdown tanks is pumped to the dilution tanks, where it is continuously diluted before punping to the Bird centri fuges. No. 2 centrifuge has a header tank; one is being fitted to No. 1 centrifuge. With certain resins, steam sparge is used in the dilution tank to assist the drying by supplying the resin to the dryers at wet bulb tempera ture. Wacker Slurry is pumped from the large "suspension tanks" to the Wedag filters, filtered, reslurried up to the right 3olids content and fed to the horizontal basket pusher-type centrifuges via a header tank. Effluent from the centri fuges, which is not completely clarified, is returned to the Wedag filters for recovery. Li CO 03*3840  # -14- 4.2.7 U.C.C. Union Carbide have 2 rotarv drvers. which thev use for all grades of material. They also use flash dryers successfully for porous materials; even using the rotary dryer, they cannot get below 0.5% heating loss with "glassy" polymers. They sift through an inclined vibratory sifter, called a Derrick sifter, which is very robust and efficient, but since an angle factor lias to be taken into consideration, the mesh size must be wider than the particles being screened and if bouncing takes place, it is possible for oversize to get through. C A.ycliffe Drying is by flash dryer; sifting by multi-deck horizontal plan-sifter through nylon mesh. These sifters have sealing difficulties and require a lot of maintenance. Wacker have a combined flash drying and rotary drying system. Porous polymers are dried to about 5% moisture content in the flash dryer and finished off in the rotary dryers. Non-porous resins are dried solely in the rotary dryers. Sifting is in multi-deck, horizontal plan sifters of more robust design than Ayoliffe. They sift through silk bolting cloth. 4.3 Comparative Evaluation of the Different Methods 4.3.1 Experiments at Aycliffe have shown that when making porous resins some 5Q& of the recovered monomer comes off only when the pressure falls below 15 p.3.i.g,, and 33^S of it cooes off during the pulling of the 10" vacuum. However, C with non-porous resins, 95% of the recovered monomer comes away before a vacuum starts to be pulled. With a plant making a large amount of porous resins, the ability to pull a vacuum can obviously increase the vinyl chloride efficiency by a considerable amount. In this respect, Aycliffe is more efficient than U.C.C, or Wacker Chemie, although measures to be taken by U.C C. in the near future should improve matters. 4.3.2 Wacker Chemie, U.C.C. and Aycliffe all realise the importance of low oxygen contents and have tailored their procedures to effect this. U.C.C. never pull an autoclave below atmospheric pressure via the recovery system like Wacker, and do not use recovered monomer in their suspension ' `JCC 0393-f 15- procesaea. U.C.C. have probably got lower oxygen contents than Aycliffe . However, Wacker Cherniy who charge cold water, have no monomer purge, and rely on a liquid ring pump for vacuum pulling, must have high gaseous oxygen contents in their autoclaves. 4*3*3 I did not like Union Carbide methods of charging D.L.F. catalyst. While I .was present, the header from the cat,pot to the autoclaves had to be taken out and cleared of solid ivory. 4*3*4 There are two main parameters which influence output in a vinyl resin plant. These are overall cycle time per batch and the yield per batch. The main constraints on the system ore the facility to remove heat of reaction and the characteristics of the catalyst used. Cycle times can be decreased by reducing the reaction or the outcycle time. There is a minimum outcycle time below which it is not possible to go; therefore, the main attack must be made on the reaction time. Prom the point of view of heat removal, high reaction rates and high yield per batch are mutually antagonistic and a balance must be struck between the two to obtain maximum output rate. The point at which this balance will lie depends greatly on the methods of operating. In conditions of manual opera tion and manual cleaning between batches, where there is a large amount of labour, used, low cycle times are not achievable without increasing the labour force. Under these circumstances, which apply at Aycliffe, the balance lies towards increasing the yield per batch and accepting longer cycle times. In the case of Texas City, where there is a very high degree of automatic control and automation, the balance lies towards reducing the cycle times- A series of calculations have been carried out, which are summarised in Appendix I, which show clearly that Union Carbide's philosophy for their plant is the best. (These were done for this report after return to the U.K.) They also show that great gains could be made by Union Carbide adopting Aycliffe's water addition techniques, especially if the water addition were made to increase the volume at the end of the reaction from 75% full to 90% full. Over 5C9S increase in output rate has been calculated; however, maiy assumptions have been made in these calculations and a figure of 2C% increase would be more conservative. Since UnionQCarbide's process water is de-ionised steam condensate at 50 C, and they do not circulate it at high pressures, a iJCC 039840 -16- cooler and high pressure circulation system would have to be installed. They would be wise also to set themselves up to measure heat release, aixL therefore reaction rates, so that they could optimise with maximum knowledge. It is possible that this technique could be applied at Aycliffe given the necessary degree of automation. The - balance, however, will be different due to the higher area to volume ratio of the smaller autoclaves. C c 5. BULK flANDLDiG OF P.V.C. It will be remembered from the description of the drying plant, that the screened resin is blown from the outlet of the Derrick sifter to the bulk storage bins. Some of the old concrete resin bins are used for suspension resins, but in addition, a new bin structure and handling facility has been constructed. 5.1 Containers Resin is despatched in Texas City by two different methods:- 5.1.1 In aluminium boxes, each holding 50,000 lbs. of 22 lbs. per cubic foot bulk density resin. These are 8 ft. x 11 ft. x 40 ft. and fit onto a detachable road chassis. These are harnessed to a prime move) which conveys them to the docks. The box is detached from the chassis and packed into the hold of the ship. (U.C.C. own two ships on this route). They are shipped to an East Coast port, transferred to another chassis and hauled to the East Coast depot, where they are unloaded and the resin bagged and despatched to customers. 5.1.2 In bags, in rail cars to Bound Brook. 5.2 New Bulk Handline System This system was designed to handle the output of six 5,700 U.S. gal. autoclaves. It consists of 9 aluminium bins, each of 1,700 cu. ft. capacity, designed to hold 50,000 lbs. of 32 lbs. per cu. ft. bulk density resin, mounted on a 75 f high construction (see sketch No. 2). Each bin is gupportoci at the start of the conical bottom, which is of a 45 angle. There is a remote controlled, pneumatic cylinder operated De-zurik slide valve on the outlet from each big, and the outlets from all the bins are connected by a 45 angled duct of 6" diameter to a small, fully-enclosed collecting tundish. + 11c.r -17- Undemeath the structure is a platform, onto which the box and chassis are run and disconnected from the prime mover. One end of the platform is raised by means of pneumatic cyli:iders, until the box is tipped at on angle of 45 to the horizontal. The flexible pipe from the bottom of the tundish is connected to tho inlet to the box; the pneumatic valve under the bin is opened and it taxes approximately 15 minutes for 50,000 lbs. of resin to be loaded into the box. It should bo noted that the bin3 are mounted on load cells, with automatic cut-off and remote controlled switching. V/ater sprays are connected to the tops of the bins for washing out when changing grade. An air blower and air heater are supplied for drying the bins out after washing. The bins have a vacuum pressure relief cover on the manhead, set at \ cz. pressure and -g- oz. vacuum. This is a similar device to those fitted on the sluny blending tanks and made by the same firm. Black, Sivals and Bryanson Limited, Oklahoma City. Access to the top platform, 70 ft. above ground level, is by means of a stairway aixl man-lift. This latter seemed a rather dangerous piece of equipment, consisting of an endless canvas belt, on which, at regular intervals, a flap-down platform and handholds were fitted. An interesting feature of these bins was the receiving station on top. Resin, blown from the sifter outlet by a 10" x 15" R.C.D. -type Roots-Connersville blower, travels approximately 500 ft. through a 4" diameter duct of x" wall thickness into the collector (shown in sketch No. 3). This collector is a high efficiency cyclone, only 3ft. high and 10" diameter, which is 1OC05 efficient down to 10 micron particle size, and no dust filter is fitted. It is made by the Dustex Corporation, Buffalo 25, New Jersey, the speci fication being as follows:- Cyclone: 8" W/G D.P. U.S. Pat. 2643737 10" s.s. D. 1600 This system conveys 75 lbs. of resin per minute, using an air volume of 350 ou. ft./minute (air velocity 4,400 ft. per minute). ucc 039846 039847 C C -18- 5.3 In 'the old system, resin is blow into one of ton intermediate bins, each of 1,500 cu. ft. capacity and 10 ft. x 10. ft. section- They are made of concrete, painted with a vinyl preparation, with 60 conical stainless steel bottoms. Resin is transferred from these to the main storage bins, 24 of which are used for suspension resins. These are each 6,000 cu. t. capacity of 12 ft. x 12 ft. section and again have 60 stainless steel bottoms. They have more bins than they require. Resin is sucked from the main storage bins by several transfers systems, which convey it to the bagging building and loading bridge, a distance of 300 to 400 yards. 5.3.1 Loading of Boxes The old method of loading boxes is from the bridge. Under the bridge are a number of hydraulic platforms for tilting the boxes. Over each platform is a receiving station arri 3ar-Nun sifter feeding via an air slide conveyor into the box. Loading of boxes by this method is very much slower than by the other method and has to be done by volume. 5.3.2 Bagging Resin is received into the cyclone from the transfer system, as shown in sketch No. 4, and is fed via a rotary valve into a Bar-Nun'sifter, thence through a magnetic sepa rator into the surge bin above the bagging machines. The cyclones are part of the transfer system and contain a bag filter, wdeh is necessary to protect the Roots blowers. The system is a standard Fuller-Lehige one. Each surge bin feeds a pair of fluopackers, both of wnich are operated by one man. There is a Bindicator level con troller in the surge bin, which controls the rate of feed into the conveying system and maintains the correct head of resin above the fluopackers to enable them to weigh accurately arkl consistently. The fluopacker spouts are equipped with a blowing nozzle, and when the correct weight is reached, the sack is blown off the filler spout, pivots on its bottom support, and falls against a spring loaded, pivoted plate alongside the bagging operator, in a convenient place for him to tuck in the valve. Pressure on a foot pedal releases the plate and drops the bag onto a conveyor. This enables a very high rate of bagging to be attained by the operator. They have a battery of six fluopackers `operated by three men. `JL-C 339848 -19- Tho layout of the conveyors is shown in sketch No. 5An interesting feature was that of the design of one of Lie two bag flattenera. Thio had been made up by themselves and relied on the weight of the rollers on the pivoted arm for tlie flattoning action. Bags wore automatically marked on their side as they passed along the oonveyor and loaded manually from the end of tho conveyor onto pallets or into roil cars for despatch to Bound Brook. 6. SEMI-SCALE BQUIRaENT 6.1 Description of Plant o The semi-scale equipment is situated alongside the production plant, and consists of one 600 U.S. gal. autoclave geometrically similar to the production autoclaves, with infinitely variable speed agitation, and two 10 U.S. gal. autoclaves. Too blowdown tanks are provided, which between them have sufficient capacity to take a batch of resin from one of the production 5,700 U.S. gal. autoclaves. This equipment, together with pumps and connections to the recovery system, stands on a platform alongside a sheeted structure, similar in size to the suspending agent tower at Aycliffe. Acconsnodated in this structure are a semi-scale Bird centri fuge and & fluidised "bed dryer, together with suspending agent dissolving facilities, oatalyst, resin bagging and storage facilities, and the control room for the autoclaves, blov.down tank and drying system. This system seemed to be ideal in that it allowed the running of works trials on the production scale, without the necessity for special preparation of blowdown tanks, drying lines, etc. It allowed the production resin handling and drying systems to be designed for mmrimum output rate without \he necessity of being constrained to give flexibility for running small works trials. 6.2 Fluidised Bed Dryer The most interesting item of the semi-scale equipment was the fluidised bed dryer. This had been made to their own design. It had been designed as the result of experience with a "lash-up" spouted bed-type dryer in use in the solvent plant. This had been installed because the existing three-stage flash dryers had not been getting the. moisture content low enough* ^od was now successfully removing 2? of moisture with a feed rate'of 4,000 lbs. per hour of resin. When first installed, the "lash-up" had no diffusion plate 039850 -20- and diffuolty ms experienced with accumulation in the bottom of the cone. A stainless steel screen mesh plate was fitted) but the metal overheated due to fluctuation in load and resin adhered to it, plugging it, Finally, a perforated plate was fitted with a ply wood underneath it to reduce heat transfer to the metal. This, together with a re-design of the outlet and fitting of special nozzles to the boles, has made it work satisfactorily. Using this experience, the semi-scale dryer was designed It consists of a stainless steel cylinder 6 ft. in diameter and approximately 8 ft, high, with a slightly conical bottom. On top of this cone is fitted an air distributor plate, consisting of a 16 gauge stainless steel plate fixed tc a ^r" ply wood board, drilled and fitted with special distri buting nozzles on a 2^r" square, plate spacing. A sample of these nozzles was obtained. They are made by General American Transportation Corporation. The conical bottom seotion acts as a Plenum chamber. On the 3ide of this plate, as shown in the sketch, is fitted the exit nozzle on the same level as the plate to prevent hold up, and is closed by an automatically operated De Zurik slide valve. Bed depth is controlled at 12 to 18" bp means of a D.p. cell with air blow-back operating the automatic De Zurik slide valve. Feed is by means of a screw conveyor, which can be adjusted to feed either in the middle of the bed for batch operation, or at the edge for continuous operation. It is installed on the opposite side of the dryer to the outlet. They have found that the temperature control of the bed is excellent, but the efficiency and ease of working depend to a great extent on the instrumentation. They find that control for hompolymers can be carried out on outlet temperatures, but for copolymers, it is essential to use the bed temperature. The transmitters control in cascade, as shown in the accompanying sketch No. 6. For facility of control on the air heater, a by-pass duct with slide valve control has been fitted. Normal air velocity is 1 ft. per second, but this can be varied 50 per cent each way. They find that with P.V.C. they can use 150 C. inlet Jemperature, but with copolymers this must be reduced to 110 C. This is to keep bed temperatures below 60C., at which temperature copolymers stick to the metal surface. A cyclone i3 supplied to colluct cany-over. They do not advise feed out by weir because of the risk of some material staying in the bed for a very long time. Con veyance of rosin from the discharge to the sifter inlet was by a venturl-type, pneumatic conveyor, similar to -21- what was seen at Wacker Cheraie, made by 'Whitlock Associates Inc., 2021 Coolidge Highway, Hope Park, Michigan. 7. ORGANISATION The organisation of the Texas City plant is shown on the attached charts. There is also a product-wise croas* organisation and in this Mr. Prank Dexter is the General Manager for P.V.C. His responsibilities include the manu facture of monomer, P.V.C. resins by n processes, and compounding; he is normally located in New York. He has an assistant, Mr. Max Sutherland, who normally operates at South Charleston. The same thing applies with the research and development side. Although the Plant Development Section is responsible through the Laboratory Manager to the Plant Manager, they are also responsible to the Central Development Department at South Charleston; thus Mr. Dean Richardson is responsible both to Mr. 0. T. Carlyle at Texas City, and to Mr. Fred Bailey at South Charleston. 8. POSSIBLE APPLICATIONS AT AYCLTFFB 8.1 Application to the present plant Prom 4.3*4 and the appendix, it seems probable that increased output rates could be obtained from the Aycliffe plant, over and above what we have heretofore thought to be the absolute maximum, providing that the manual work content of the operation can be reduced to bring it into line v.ith Texas City, It seems that an output rate increase of 10 to 2Q& might well be possible. Experiments will be carried out to check this. If they confirm that an increase can be obtained, the following plant additions and modifications will be required:- 8.1.1 The purchase of automatic charging meters for hot water, vinyl chloride and suspension components. These could be similar to U.C.C., or be of the weighing type by fitting load cells to existing storage tanks. 8.1.2 Either the modification of the existing autoclave temperature control instruments, or the purchase of new ones, which will control the resin batch without need for frequent manual adjustment. 8.1.3 Automatic high pressure process water addition, controlled by rise in autoclave temperature with a volumetric top limit over-ride. -22- 3.1.4 The provision of hot; high pressure process water pump:, of considerably higher capacity to ensure a supply of water at the increased rate which will then be required. 8.1.5 Provision of remote controlled distillation and blowdown facilities. 8.1.6, Modification of autoclave cat. pots and provision of Hindle line blinds to facilitate and speed up taking out of service and putting into service of vessels, and inspection of top nozzles. 8.1.7 The bringing back into service of the old recovered monomer tanks, to allow virgin/recovered monomer mixes to be made up more quickly to keep up -with the more rapid charging. One of the present weigh tanks is wasted by being used as a recovered monomer storage tank. 8.2 Applications of Knowledze gained in Design of Future Plant a.2.i Plant Layout The cruciform arrangement of the Texas City autoclaves, with the Control Room at the centre of the cross, is ideal for control. It also cuts the amount of waling about of operators to a minimum- In our preliminary ideas for plant layout, our autoclaves were spread out in a row, ton autoclaves long. This would require a considerable amount of walking and would not bo so easily controlled. On measuring up the area between the existing autoclave struc ture and the edge of the neutralising pit, it has been found ^v - that part of a cruciform similar to Union Carbide's can be accommodated, including eighteen 5,500 Imp- Gal. autoclaves in all. This is shown roughly to scale on the attached sketch, No. 7. This sketch also shows five more 3,000 gal. autoclaves installed in the existing area, together with extra blowdown tanks, trap tanks, etc. 8.2.2 Both Wacker and Union Carbide have gone in for very large slurry holding tanks, which are in effect what our blowdown tanks are. It is considered that we should follow suit, and if the layout sketched is adopted, a convenient 3ize would be 16 ft. diameter x 20 ft. high, giving a capacity of 25,000 Iag>. Gals. 8.2.3 Both Union Carbide and Wacker are using rotaxy dryers successfully for drying suspension polymers. Fluidised bed -23' dryers are used for finishing off the drying of resin , .. has already been dried to about 1% moisture content means. It would seem that we ought to follow the host method seems to be like wacker's, ..capacity one-stage flash drying plant feeding Ui a rotary dryer second 3tage. Two rotary dryers arc shown rn thu sketch ol' the extension to the resin building. Those would not be sufficiont without a largo, first-stage flash dryer each, for the size of plant shown. Also, if a bulk iurrlling and bagging plant were built outside the present perimeter fence, a further rotary dryer could be accommodated on the ground floor of the resin building and the existing flash drying plant could be upgraded by replacement of the first stages with higher capacity ones. 8.2.A The new plant should be fully automated similar to Texas City, to enable maximum output rate and, therefore, maximum return of investment to be achieved. 8.2.5 3ecause of lindJations in space, once the tank area has been filled with normal size tanks, further storage should be provided in large spherical tanks (hortonspheres). A AO ft. diameter sphere would store 1,500,000 lbs. of viryl chloride and would have to be constructed in 1^" plate. Two of these are shown in the space between the Tank Area and the Refrigeration Area in the sketch. 8.2.6 Careful consideration has been given to the pros and cons of glass lined autoclaves versus stainless steel ones. In view of the fact that Union Carbide's glass lined auto claves were already damaged after a short period of service and chat the new catalysis leave the autoclaves much cleaner, stainless steel autoclaves are still to be preferred. 9. VISIT TO BOUND BROOK This visit was of one day's duration and consisted of a discussion of our range of resins and their properties with their Development Department in the morning, and a quick tour of the Calendering and Dress Polishing Plants in the afternoon. The morning's discussion has already been reported. The tour of the Plant in the afternoon was toe quick to be comprehensive, but various items of information were noted and are as follows;- 9.1 The main calendering lines were fully automated and the raw materials handled in bulk. There were four ribbon blenders manufacturing pre-blend. These were mounted on UCC 039858 -24- lo-nI cells and ingredients were charged into the..s by remote control from a p^nol. This was completely automatic and in corporated a punch card for_ feeding to the compator. The load cells were all by Richardsons, and were very good and reliable;. These blenders supplied three calendering lines aid required one operator. 9.2 Of the three calendering lines, two had conventional 3A Bardcries and tne other had a continuous Banbury. A fourth line was being installed with a continuous Banbury. The crews for the lines were one man upstairs and four men downstairs par line, with a cleaning crew of four men. These three lines were in charge of one foreman. There wurc no two-roll mills - these having been replaced with continuous extruders (made by Royal). The top operator could control the Banburics and extruder from a control p-r.l, on which were mounted monitoring television screens, one shewing the extruder hopper and the other showing the bank on the top nip of the calender. Each line was fitted with a Beta-ray gauge to control the sheet thickness. These gave continuous profiling and fed back and adjusted the rolls gaps. These are Aceu ray gauges and are made by the Industrial Neutronics Corporation, Columbus, Ohio. These were fitted after the third cooling rolls. Each line also had an automatic take-off and cutter made by the Hobbs Manufacturing Compary, Worcester, Massachusetts. Trim was chopped very efficiently and continuously by Cumberland 2-stage trim choppers, mounted in noise-proof boxes. These trimmed flexible as well as rigid. 9.3 Their experience with the continuous Banbury was as follows:- 9.3*1 They had had trouble with colour dispersion. This could be eliminated by blending with pigment in the ribbon blender, but required a long shutdown for cleaning after each colour. They had now installed a zig-zag conveyor/ blander, which was successful. It is made by the Patterson-Kelly Compaiy Inc., Process Equipment Division, East Stroudbury, Pennsylvania. Even so, it is not easy to start up. 9*3.2 The zig-zag conveyor i3 fed from a hopper containing pre-blond, together with the correct proportions of trim and maSurbatch. The pigment mnsterbatches are made up in a Pupenmeyer with pre-blend. The feeding of pigment pastes is not satisfactory. A Nouta mixer is used to add colour in unplasticioed formulations. IJCC -25- 9.1 Other snippets of information noted were: 9.4.1 Pressed metal are cleaned by immersing in boiling caustic soda and rinsing in detergent and finally rinsing with boiling demineralised water, after which they dry off by flash evaporation. This is carried out semi-automatically by hanging the plates from a conveyor, which passes through the different stages in turn. 9.4*2 To calender a s ecial grade of white rigid materials, which needs to be kept very hot ^11 the time, feed to the r.ip is in small lumps cut off from the extruder by a pneumatically operated knife. An operator keeps the bank in the nip at the lot level concerned by operating this knife by remote control, on observation of a television monitor picture of the nip on his control panel. 9.1.3 Colour matching is done by machine. This is made by the General hlectric Carpany and called a spectro photometer. They claim that this is more sensitive than the human eye. 9.4*4 Compound is packed in polyethylene valve sacks by means of auger type packers. To prevent bags sliding on the pallet, g lue is spread on each layer of bags. This is adhesive 7139, made by Swifts of Chicago. B. S. Shatwell AarShEIX I Aycliffe versus U.C.C. Output Rate Philosophy Those calculations have been carried out assuming that (a) * (b) The overall heat transfer coefficient U = 40 3TU/hr/ft^/F. (Measured in 1959 at Aycliffe using brine) Reaction temperature is 54C. (c) Cooling is by brine at -5C. 2 (d) Effective jacketed area of autoclave = 300ft. (e) I.P.P. initiator is used and gives reaction times and rates as measured at Aydiffe. (f) Temp, rise of brine at maximum flow = 5C. (g) Reaction taken to 40 p.s.i. pressure drop i.e. 90^ conversion based on Grade 1 yield. 92?Z overall conversion. 1 Calculation of Maximum Heat Reinoval Rate q * UA4Tm (1 ti = 54 -(-5) * 59C A to = 54 - 5 = 49C A ti Aito 2 108 So Atm = 54C ^ * 40 x 300 x [54 x 1.8 = 1,170,000 KCU/hr Reaction rate tolerable 1,170,000 700 = 1,67C lb./hr. uoc 039359 Using U.C.C, charge weight 11.500 lb. Max. x'ate of reaction = 1>67 x 100 = 14.5^hr. 11,500 I.r.F. cone, to give this peak keac-cion time at this cone. U.C.C. outcycle time = O.Q125phm = 12.5 hrs. = 3*5 hrs. Total cycle time 16.0 hrs Average output rate = 11,500 x 0.9 16 = 650 It./hr 3. Check on U.C.C.'s rate if they used I.P,?. They run to 81$ conversion with a charge weight of 11,500 lb. They must blowdown on temp, rise not pressure drop. From our curves catalyst quantity to give 81^5 conversion at 1i.S^hr. = O.Oipphna Tine to reaoh this would be 8.5 hrs. Total cycle time = 3.5 + 3.5 = 12.0 hrs Output rate = 11,500 x .81 12 = 775 lb ./hr. V. If U.C.C. ran at 90% full same V/W ratio (a) Going to full pressure drop Charge wt. 11.500 x 90 = 13,800 lb. 75 kax. tolerable reaction rate = 1.670 x 100 = 12.2&^hr. 13,800 I.P.P. cone, to give this Reaction Tine Cycle time 0.0112 phm 14 hrs. 17.5 hrs. Output rate 15.800 x 0.9 17.5 712 lb./hr. (b) Going to 8l$ conversion I.P.P. to give reaction rate at 81$ con. = 0.0125phm Time to get to this = 11.5 hrs. Cycle time - 15 hrs. Output rate = 18.800 x 0.81 15 a 746 lb./hr. Run at full v,ith increase M/ft ratio (Aycliffe) (a) Full pressure drop Aycliffe 10* ratio = 1 : 1.75 Charge wt. * 9,140 x 4.700 5,000 a 14.500 lb. Max. output rate = 1.670 x 100 - 11.7%/hr. 14.500 I.P.P. to give this Reaction time Cycle time = 0.011 phm = 14.5 hrs. = 18.0 hrs. Output rate = 14.500 x .9 18.0 716 lb,/hr. U cc 039861 (b) 61$ Conversion. I.P.P. to give 11.7$ at 81$ conv. = 0.0112 phn Reaction time Cycle time = 13 hrs. = 16.5 hrs. Output = 14,300 x 0.81 "^5--------- = 700 lb./hr. o. Run at 76>-' r.0.1 ; ith Aycliffe I.'/V/ ratio Charge wt. = 14,300 x 75, 50 = 11,500 lb. Very little difference free; (3) say 11,900 x .81 = 772 Ib/hr. 12.5 i.e. no significant gain or loss. 7. With Water Addition 7.1 At 75$ full level fa) Making up contraction in volume only. (b) Making up contraction and filling to 9C$ full- N.S. U.S. process water is at anprox. 50C. It would have to be chilled. So 10 C. used. 7*1.1 U.O.C.'s present method usinr: I.P.P. Contraction in vol. at 75$ conversion = 11,500 x .75 x x lap* gal* e 11,500 X .75 X 0.043 360 Ian, gal. (430 U.S. pal.) If this is added over last lhg. of reaction and water temperature a 10 w. Heat absorbed * 3,600 x (54-10) x 1.8 3TU/hr. Reaction rate = 3.000 x la x 1.8 lb ./hr. 700 * 3.600 x 44 x 1.8 x 100 %/hr. JOO x 11,500 = 1,.5%/hr. liwc. reaotion rate = 14.5 + 3.5 = 18%/br. At 81$ conversion and 18%/hr. cat. = 0.C25phm and reaction time =5.5 hrs. cycle time = 9 hrs. Output rate = 11.500 x 0,81 9 - 1.035 lb./hr. 7.1.2 If made up to 90$ full: Vol. water = 4,700 (.9 - .75) Imp. gal. Wt. water = 4,700 x .15 x 10 lb. 7,000 lb. (840 U.S. gal.) Reaction rate if added over 1 hr. * 7.COO x 44 x 1.8 x 100 700 x 11,500 = 6.8%/hr. .*. Total reaction rate controllable *18+6.8 = 2Lt.jW.hr. This will give a reaction time to 81$ conversion of approximately 4 hrs. Total cycle time Output Sate * 7*5 hrs. = 11.500 x 0.81 7.5 = 1.250 lb ./hr, 7.1.3 Running to 90$ Conversion Contraction in vol. at 85$ conversion * 11,500 x .85 x 0.043 * 410 Imp, Gal. (490 U.S. Gal.) 039863 Heat absorbed a 4,100 x (54 - 10) x 1.8 BTU/hr. Reaction rate = 4.ICO x 41 x 1.8 x 100 700 x 11,500 = 4.0:Vhr. Reaction rate peak a 14.5 + 4-0 = 18.5/Vhr. Requires O.Q225phm I.P.P. to give 7*0 hr. reaction Total cycle = 7.0 + 5.5 = 10.5 Output rate = 11.500 x _Px2 10.5 * 990 lb./hr. 7.1.4 With make up to 90$> full iiax. reaction rate = 18.5 + 6.8 * 25.3?yhr. Reaction time = 6.0 hrs* Cycle time Output Rate = 9.5 hrs. = 11.500 x .9 = 1.090 lb./hr. 9.5 7.2 At 90%> Full 7.2.1 At 81^o conversion Contraction in vol. a 11,500 x 360 = 430 gal. (515 U.S. gal). Reaction rate = 430 x 3.5 360 = 4.2^hr. Max. reaction rate = 12.2 + 4.2 = 16.4^hr. Reaction time = 7.0 hrs. Cycle time - 10.5 hrs. Output rate = 13.800 x .81 1,080 lb./nr UCO 039864 7*2.2 At C?/'j Conversion Mux. W/A a 1? .600 x 410 11,500 = 495 gal. (595 U.S. gal.) Reaction rate = ^25 x 4.0 410 = 4.E^hr. . Max. Reaction rate = 12.2 + 4.8 = 17.0^hr. Reaction time = 8 hrs. Cycle time = 11.5 hrs. Output rate = 15.800 x ,9 11.5 = 1.080 lb./hr. UCC 039865 J. H, Field, Asst. Plant Manager 0. T. Carlisle Supt. Plant Process Dev, Dept, and Plant Labs. -V' Plant Process Development Dept, - Bldg. 133 Chesteals Group F.S. Provenzano - GL Ga3 Group f. Partridge - GL I Resins Group D.E. Richardson - GL 1 Operations Analysis Croup S.S. Aldredge - GL G.G.Herkreader Ch.E. H.H. Hack Chem. B.C. Robinson Chens. -------------------- Corrosion and "iiterials Group W.G. Ashbough - GL | Resins Lab* Polyy thylone R.C* Hurtcoolc - GL Plant laboratories iiicaJs Chemical's - Gas Anal. < Spec. Lab* J.E. Pinch - GL Resins Lab. Vir\/ls J.L, Hockersnith - GL ZL. Chemicals - Gas Special Problems J.S. Burch - GL GL = Group Leader 039866 A. R. Anderson Plant llanngcr (Vorks linage r) Frank Doctor (New York- Central Ungr. P.V.C. S: G< ' : 1' rg). Assistant: '.'uxx Sutherland J. H. Field Ass. Plant ling. KLastics !Shift Organisation Laboratories Process Development Industrial HeJ_'.tions fPolyethylene Area (Viryl Area (incl, monomer. i O.T, Carlisle Director of Labs. & Development IsE. Eisenhower Solvent Resins j J. Erdmann Non-solvent 4 Suspension Resins I Asst. P.V.C. Phil. Ki.lly Asu. PI, ] r^r. Chord cals (OXO aloohol.s) etc. lid, Uurndrett A:.s, PI, Kn^r, Cas Reparation (Oli; fines) R.M.a Ch.E. Ch.E. Ch.E, Ch.E. E.T. John Schlichter - Autoclaves ),, . Spec. Rrobleins)SusP- * f/b* H. H. Savage W. D. Bush l!. 11. Kuet.uan H.C. Felter - Resin Recovery ) Vononmr, Suap. Expansion) - Quality Co-ordinator, Production Scheduler. - Project Engineering - Day Production. Supervisor - I-iiint. A Routine J.S. 1 urray I'L-rci-r1 rj 1 r.g & Stores. P. Schweppe tkiiiitenance < Construction L-B, Trenholme Plant Engineering O.A. Holt Gp. Leader Susp, P.V.C. Solv, P,V,C. Monomors  c ij ; * DISCUSSIONS WITH BXL PLASTIC MATERIALS GROUP AYCLIFFE, ENGLAND OCT. 31-NOV. 1, 1966 BXL Personnel: Dr. Tom Love (Resident Mgr. of R and D), Dr. Tom MacEwan (R and D Group Ldr.), Bernard Davis (Suspension Vinyl Dept. Head) UCC Personnel: Dr. F. E. Bailey, Jr., Dean E. Richardson GENERAL ITEMS: 1. Aycliffe is very busy. The demand for their FVC resins greatly exceeds their capacity, and they must buy resin for their fabrication operations. No expansion is foreseen, however, until an advantageous source of monomer is in sight. 2. Meanwhile, the condition of the five Aycliffe autoclaves is a matter of grave concern. The glass lining is spalling badly, and iron contamination from exposed metal areas is a constant worry. The current BXL policy seems to be to run the Aycliffe FVC plant until it falls to pieces, but don't spend any money on it. 3. BXL now has three division: I. Plastics Materials Division (Bakelite) II. Plastic Products Division (Xylonite) III. Consumer Products Division 4. Of principal Interest, the "Shatwell Reports"- reports by Eric Shatwell comparing Aycliffe and Texas City suspension FVC operations following his visit to Texas City during May, 1965: I. "Visit to UCC, May, 1965", by E. E. Shatwell II. "Report of Possible Application of UCC Techniques at Aycliffe", by E. E. Shatwell, It-,10.65. UCC 039868 2 Eric Shatwell is now in polyethylene at Grangemouth. DER has copies of the Shatvell reports, which should answer Dr. Manning's questions concerning applying BXL methods to increase productivity at Texas City. Shatwell believed that a 20 per cent increase was possible at Texas City, principally by using the "water addition" technique. DER will report separately on the Shatwell reports. These reports were believed by BXL personnel to have been sent to UCC. Neither FEB nor DER had seen them and the information in them was a principal reason for the visit to Aycliffe. 5, The general impression was that general purpose, suspension FVC sales price in the U.S. has now slipped below the selling price in Japan, Germany or United Kingdom. In Japan, the price appears to be about 12.5 cents per pound. In Germany, they see some "cheap" Japanese resin. FVC TECHNOLOGY: 1. Tom MacEwan has done preliminary work with a vinyl methyl ether-maleic anhydride copolymer suspending agent, and produced particles which were startling in their sphericity and glassiness. This material is marketed in the U.S. by General Aniline and Film as "Gantrez AN", and we plan to test it at Texas City in early 1967. 2. In order to minimize reaction time, Aycliffe plans to use IPP for all poly merizations lower than 65C, with mixtures of IPP and DLP to be used from 65 to 72C. IPP is readily available. Catalyst mixtures have been extensively studied at Aycliffe as part of a thorough program to maximize FVC production. Received a copy of a report from T. H. MacEwan, "Vinyl Chloride Polymerization with High Activity Initiators - Part III: More Uniform Reaction Rates for Higher Outputs from Vinyl Resin Plant Autoclaves at 58C and 65C", Report 1900, by K. Lodge and T. H. MacEwan, September 5, 1966. System uses a mixture of DLP and IPP (ca. 0.05 phm IPP plus 0.04 phm DLP). UCC 039389 3 3. MacEwan would like to look at CARBOWAX 20M. We promised to send him 10 lbs. b. In discussing resins for blowmolded vinyl bottles, we mentioned that the Diamond and Ethyl resins were currently favored. MacEwan asked for photomicrographs of the two resins. 5. In reply to our question, we were told that porous sintered polyethylene sheets for fluidization were made by Porous Plastics Ltd., Dagenham Dock, Essex, England. Samples have since been obtained. 6. The Aycliffe FVC Production Unit has gradually moved to higher and higher monomer/water ratios in its efforts to increase productivity. This has been done at no sacrifice in resin quality, and indicates an area which we propose to ex plore at Texas City. 7. Sidney Fisher, Vinyls Technical Manager, joined Drs. Love and MacEwan and Mr. Davis for a discussion by DER of the "Fisher Report" on the Q1TQ-7 process and computer analysis of designed experiments. Fisher was touting BXL's use of a "Simplex" EVOP scheme said to be in use in Birmingham. "Simplex" is one of these methods for checking 13 variables in 16 experiments,, most useful really in making five adjustments in rather well-defined processes. DER also gave a brief review of Dr. J. E. Glass's work on hydrocolloid characterization in suspension vinyl polymerization. Dr. Love was highly impressed by both presentations. He seems to be planning to set up a team to explore computer analysis applications. The discussion of the "Fisher" QYTQ-7 programs answered BXL's (TL and THMacE) numerous questions concerning how we have improved our resin quality. 8. Process-10 (FVP) is of Interest at Aycliffe as is the use of VAZO catalyst. 9. Aycliffe is pursuing the poly(vinyl cetyl ether) copolymer (Kureha type). They plan to make a production batch. They are Interested in the Cumberland (Airco) propylene copolymers, but have not seen any samples yet. They have seen no reports from UCC since early last spring, with the exception of the Bound Brook bimonthlies. They are aware of the problem with Distillers and British Geon. UCn 1+ 10. Dr. Love visited Wacker early in October. DER has a copy of the Aycliffe recipes prepared for Dr. Love to take on his visit to Dr. Bauer. 11. MacEwan's only "new" suspending agent experiments have been with a vinyl methyl ether - maleic anhydride copolymer. 12. Highest polymerization temperature used at Aycliffe is 72C. 13. Alcotex and Japanese polyvinyl alcohol equivalencies appear to be Alcotex 88-25 Gohsenol GM-l^ Alcotex 71-72/6 Gohsenol K8-17 COMPOUNDING: 1. In Howard Williams' talk on compound development, he mentioned that their aim in developing VN-336 for house wiring and armored cable was to decrease the actual stock temperature at the die to less than 200C. Temperatures above 200C were common because of the ever-increasing extruder screw speeds. The decreased melt temperature was obtained by: a) Changing the base resin from VY-18 to DVY-17. This reduced the melt temperature 15C. b) Changing the plasticizer yielded an additional 10C melt temperature decrease. c) Using WINNOWFIL S filler. This 0.75stearic acid-coated CaC03 reduced the melt temperature even further. VN-336 is a Banburied compound, and it was necessary to add back some of the old filler (Microcarb B-^) to obtain sufficient "working" in the Banbury. VN-336 has already taken substantial business from ICI. !JCC 1)39871 5 The formula is; VY-17 Vinyl Resin Mesamoll (obtained from Bayer, a plasticizer based on sulfonated phenol, cresol mixture) Cereclor 4-2 WinrDwfil S Microcarb B-4 Tribase 100 parts 38 parts 17 parts 25 parts 25 parts 3 parts 2. In the case of the rigid injection molding compound, the object was to develop a compound with good moldability in screw preplasticized machines, yet with good gloss and a British Standard softening point of T5C minimum. The compound was designated VX-105, and included a low molecular weight FVC (VY-19), tribase, Paraloid K120N to increase gloss, WMNOWFIL S to boost impact strength. Excellent complex moldings were shown, shot size ranging up to about 4 lbs. The secret was said to be in the balance of lubricants. The Post Office is testing -telephone set housings based on this compound. The formula is: VY-19 100 parts Tribase 5.2 parts Glycerol monostearate 3 parts Paraloid K-120N 2 parts Winnowfil S 5 parts This formulation is being used in rain water drains. It has ousted same ABS. Izod (notch) is 3.2 lbs. British Standard (?) softening point is 75C. Gives moldings up to a 4-lb. shot size. 039872 6 5. The problem in bottle molding compound was tiny entrapped air bubbles when the LeSeuer machine was operated at high speed. This was remedied by using higher bulk density compound in the hopper and lubricating to reduce parison temperature. The high bulk density resin was needed to reduce air entrapment in the hopper, which led to little air bubbles in the blown bottle. VZW-5884 VY-19 100 parts Paraloid K-120N 3 parts Paraloid KM-228 13 parts Dioctyl tin 1.75 parts Later, they have used a mixture of emulsion resin with suspension; 90 parts VY-19 10 parts Corvic H3559 (ICI emulsion resin) FVC blown bottles are not a profitable item in U.K., but they are coming into use with tin stabilizers. In U.K., tin is not approved but Is currently allowed, 4. Tom Young - technical service activities in vinyls for BXL at Aycliffe - discussed powder compounds for extrusion - rain water pipe, etc. A powder compound was successfully developed: DVZ-900: VY-17 Dibasic lead stearate Calcium carbonate Stearic acid Paraloid 120N It was mixed In a Henschel at 120 and air-cooled. When they shifted the cooling to a double-stage Papenmeier, the powder density began to fall from 36/38 to 34 lbs,/ cu.ft. VY-17 has a good particle shape and particle distribution. To get higher density, VY-22 was developed. VY-22 has a poor particle distribution but a higher bulk den sity. DVX-901 uses VY-22.  7 ICI sells diced compound with a "British standard" softening point of 76-77. BXL's powder blend has a B.S. softening point of 84. By vising a 6C>/kO mixture of VT-20 and VY-22, BXL can get an 87 softening point which is apparently used in DVZ-901. The chief effort in vinyls at Aycliffe is to achieve higher softening points. They were curious why this was not a higher priority point with us. They are interested in chlorinated PVC routes to higher service temperature vinyls. 5. Williams and Young presented a problem that they recently solved in making resin for thick-walled, large-bore, flexible vinyl tubing (use obscure, but ap parently related to a mining operation). Tubing had a rough and rippled interior surface. Processor was temporarily solving the problem by double-processing. BXL solved the problem, a feat over which they were highly pleased, by blending a lower and a higher M.W. resin to give higher back pressure in the extruder. UNION CARBIDE INTBRNAL. RRMP NDIN I IT = 1 1966 CHEMICALS DIVISION P. O. BOX 471, TEXAS CITY. TEXAS campmmr isNm c*rT ( Dr. F. E. Bailey, Jr. - 511 Dr. W. R. Manning - 511 Mr. R. N. Wheeler - 5lk Mr. J. H. Field/Mr. N. A. Gimber/ Mr. J. F. Erdmann S**M November 50, 1966 I am attaching two reports by Eric Shatvell of BXL Plastic Materials Group at Aycliffe, England. Copies of the reports were apparently forwarded to 270 Park Avenue but never progressed further. The information contained in the reports makeB possible certain comparisons of Aycliffe productivity with that experienced at Texas City, but as Mr. Shatwell points out, pre cautions must be taken in making any such comparisons. The reader must bear in mind that Aycliffe is pushed for every possible ounce of resin production, that the glass linings in the Aycliffe autoclaves are badly worn, necessitat ing cleaning after every run, and that production at Texas City is currently limited primarily by the dryers rather than the autoclaves. Nevertheless, there are certain areas where Aycliffe practices could be of value. These include the use of mixed catalyst systems in the middle temperature ranges (55-70C), higher moncmer-to-water ratios, and, as a last resort, water injection. I noted during my recent visit to Aycliffe that they have modified their recipes slightly to permit them to operate at VCl/water ratios much higher than outb, without adverse effect on resin quality. Ad ditionally, it appears that a close study of Texas City conversions (monomer to resin) is in order, particularly for the low molecular weight homopolymers. Your suggestions in any of the above areas will be welcomed. Yours very truly, DER/sj Attachment Dean E. Richardson -------- ---- ' REfOaT Ok 71ID POSSIBLE APPLICATION OF U.C.C. TECHNIQUES AT AfCLIFFE 1. INTRODUCTION AND SUbNARY In the report on my visit to Union Carbide Corporation1 s P.V.C. Plant at Texas City, I described their philosophy for obtaining maximum output rate. I carried out some calculations in the Appendix, based on extrapolations from data available at that time, which indicated that with a high degree of auto mation , as in existence at Texas City, their philosophy was justified. I also indicated that it might well be that if v;e had the same degree of automation at Aycliffe, it would pay us to adopt their philosophy. At the time when my report was being issued, further data came into my hands, when Research & Development Report No. 1824 by Mr. J. R. Wallace was circulated, giving details of experiments carried out in the 50 gallon autoclave to determine the reaction rate curves with I.P.P. at a great many different concentration levels. This led me to examine the original reaction rate curves and these have been plotted on graph No. 1 (see Appendix). Using this graph and the reaction time-concentration curves on Pig. 5 of 1. d D. Report No. 1824 (which is repro duced as graph No. 2), calculations were carried out to check whether, in fact, an application of Union Carbide's philosophy would be of profit to Aycliffe. These calculations, which are given in detail in Section 2 of this report, led to th6 following conclusions. (i) (ii) With present outcycle times, the direction in whicn ve have been currently progressing at Aycliffe will give the best output rate. (Section 2.5) If the average outcycle time could be reduced by use of automation or other techniques to 5 hours per batch, the best output rate could be obtained either by (a) using present methods or (b) running to 81# conversion instead of 89-90% - the resin would then not be VY.18 however. Tho calculation of (a) + (b) is given in 2.6 tuid 2.8, the alight difference between them be In,/ within the range of error of tho data and uasumv>tiona made. Since those calculations did not shaft ary benefit at Aycliffo by using U.C.C.'s techniques and philosophy, the calculations in the Appendix of the report on ny visit to the United States were checked and done again, using the new data which had become available. These are reproduced under Section 2 and show that the errors in extrapolations made in calculating the original results were such that they nullified one of the conclusions drawn from the calculations, i.e. that Union Carbide's philosophy was the best for their plant. In fact, both methods would lead to almost identical output rates, and the method to be chosen would depend upon the properties required in the final resin and also the diminution in viryl chloride efficiency, which would be expected when handling larger quantities of recovered monomer. Tliis can be seen by comparing the calculations under 3.2 with those under 3*1- It does not, however, invalidate the main con clusion of the calculations, which was that Union Carbide would obtain at least 20$ increase in output rate by using our water addition techniques. A check was then made on what output rates could be expected from 5,500 gallon autoclaves, as proposed in the design summary for the extension of the V,,R. These are shown in Section 4- They confirm the scale-up factor used in the Outline of Project Report No. A.45 of Hay 1964 of 1.5 between 5,500 gal. and 3,000 gal. autoclaves, and indicate that output rates of 800,000 to 900,000 lb ./week of VY.18 should be obtainable from the 5 proposed new autoclaves. The reason for the error in extrapolation in the original report was because it was not realised that the initial reaction rate at different'I.P.P. concentrations was almost the ."ome. It had been assumed, from what evidence was then available, that the curves ran parallel to each other at increasing I.P.P. concentrations. 2 U.C^ Techniques v, Aycliffe Techniques for Aydiffe 3.000 Imp, gal. Autoclave 2.1 Assumptions made (based on previous measurements). -3 - The overall heat transfer coefficient U = 52 BTlXhr)(ft^)(F). The reaction temperature is 540C. The cooling water temperature is 14C. 2 Effective jacketed area of autoclave = 250 ft. Temperature rise of cooling water at maximum flow = 8C, Reaction rates at various I.F.P. concentrations follow curves in Fig. I. 2.2 Calculation of Maximum Tolerable Reaction Rate r q = U.A.vtm. Ati = 54-14 = 40C ) 'ito = 54-22 = 32C ) ati > 2 -ito * * 36C. 2 q = 52 x 250 x 36 x 1.8 = 840.000 BTU/hr. Since uHr = 700 BTU/lb. Max. reaotioh rate tolerable = 840.000 700 = 1.200 lb./hr. 2.3 Calculation of output rate per autoclave hour using V U.C.C. level of 75? full, making VY.18 at 90$ conversion, and with H.P. water addition to make up volume contraction and also to fill autoclave up to 9C$ full. Charge wt. a 7,600 lb. t Volume contraction = 350 x 7.600 9,140 = 2^0 Imp, gal. (350 gal. is contraction at 9,140 lb. charge wt.) .*. Total quantity of water to add = 290 + = 290 + 450 a 740 Imp. gal. 3,000 ucc 039378 -4 - If this is added over the last hour of reaction and the H.P. process water temperature = 10C. Heat absorbed = 7*4 * (54-10) x 1.8 BTU/hr. * Reaction rate it will control a 700 . lb -/hr. = 7.400 x 44_x lj8 x 100 700 x T$00 ll^hr. Maximum tolerable reaction'rate = 7,oOO = 15.8 + 11 f 11 26.8%/hr From Fig. 1 Reaction time = With full automation outcyde time = 7 hrs. 3 hrs. .*. Total cycle time 10 hrs. Output rate * 7.600 x .9 10 684 lb ./auto.hr. And on 5 autoclaves = 575.000 lb ./week.. 2.4 Present output rate of VY.13 using current Avcliffe methods i.e. 9C$> full 90^ conversion charge at 9,140 lb. Reaction time Outcycle time Cycle time = 8 hrs. = 4*5 hrs. = 12.5 hrs. Output rate = 9.140 x .9 12.5 = 657 lb./auto, hr. and on 5 autoclaves ~ 550,000 lb ./week ucc 039879 -5- 2.5 Output rate at present conditions, with full automation giving 3 hrs, outcycla time. Reaction time Outeycle time Cycle time 8 hrs. 3 hrs. 11 hrs. Output Rate = = 74-8 lb ./auto. hr. and on 5 autoclaves = 627.000 lb./wk. < 2.6 Output rate using present methods, but increasing charge weight to condition where termination is on temperature rise with a slight pressure drop. This was done during W/E. 11.9.64-. Charge wt. 9,800 lb. Reaction time Outeycle time Cycle time Output Rate Conversion = 895*6 7.5 hrs. 4.5 hrs* 12.0 hrs. 9.800 x 0.89 12 728 lb ./auto, hr.. and on 5 autoclaves 610.000 lb./week 2.7 As 2.6 but with automation to reduce outeycle times to 3 hours Output rate 9.800 x 0.89 10.5 830 lb./auto.hr, and on 5 autoclaves 696.OOO lb./week 2.8 U.C.C. run their QYTQ7, equivalent to VY.18, to 81$ conversion. If we ran VY.18 to 81$ conversion, it would be no longer VY.18, but probably much more like VY.ll. However, output rates have been calculated as follows, assuming it was an acceptable proposition. ucc 039880 6- - In fact, it is doubtful whether the drying plant could cope with it. 2.9*1 U.C.C. level of of 75$ full* making up contraction in volume and falling to 90$ by H.P, water addition. This is in fact t2.3 to 81$ conversion. Reaction time 6.5 hrs. Outcycle time 3*0 hrs. Cycle Time 9*5 hrs. Output Rate = 7.600 x 0.81 9.5 = 648 lb./auto.hr. and on 5 autoclaves * 544.000 lb ./week 2*9*2 Present best Aycliffe method with full automation (like 2*7) to 81$ conversion. Reaotion time 6.5 hrs. Outcycle time 3*0 hrs. Cycle time 9.5 brs. Output Rate 9.800 x 0.81 9.5 " 835 lb./auto.hr. and on 5 autoclaves = 710.000 lb./week 2.9*3 Present best Aycliffe method (2.6) to 81$ conversion. Reaction time 6.5 hrs. Outcycle time 4.5 hrs. Cycle time 11.0 hrs. Output rate = ?,8?.*nT"~' ^u = 724 lb./auto, hr. and on 5 autoclaves 606.000 lb./week ijrr 039881 Recalculation of Results in Appendix I of Report on visit to America. The same assumptions are made hut new reaction data on I.P.P. used (from graph No.l). Maximum Reaction Rate - no change = 1.670 lb./hr. Using U.C.C. charge wt. 11,500 Max. rate of reaction - same = Reaction time = 11.5 hrs. U.C.C. Outcycle time = 3.5 hrs. Cycle time = 15.0 lira. Output rate a jl.500 x .9 15 = 6j&. lb ./auto, hr. Check on U.C.C.1 s rate if they used I.P.P. - figures are correct, except that I.P.P. concentration to give this would be 0.0175 phm. If U.C.C. ran at 90% full with same H/V ratio (a) Pull pressure drop. Reaction time Outqycle time Cycle time 13-0 hrs. 3.5 hrs. 16.5 hrs. Output Rate 13.600 x 0.9 16.5 756 lb ./auto,hr. (b) Going to 81% conversion. Reaction time Outcycle time Cycle time 10.0 hrs. 3.5 hrs. 13.5 hrs. Output rate 13.800 x .81 13.5 828 lb./auto.hr. -8 - 3.5 Run at 9095 full with increased 16/W ratio (Aycliffe) (a) Pull pressure drop. Reaction time Outpycle time Cycle time 13.0 hrs. 3-5 hrs. 16.5 hrs. Output rate = ikiiPO.x ^ I6.5 782 lb./auto, hr. (b) 81# Conversion Reaction time Outpycle time Cycle time 11.0 5.5 14.5 Output rate 3 14.300 x .81 14.5 794 lb./auto, hr. 3.6 Run at 75% full with Aycliffe V*V/ ratio. - Original was correct = 772 lb./auto.hr. 3.7 With Water Addition 3.7.1 At 75% full 3.71*1 U.C.C.'s present method using I.P.P. Only difference in cycle times. Reaction time Outcycle time Cycle time 6.75 hrs. 3.5 hrs. 10.25 hrs. Output rate = 11.500 x .81 10.25 = 910 lb./auto.hr. -9- 3.7.I.2. If made up to 909# full. Reaction time Outcycle time Cycle time Output rate . 6.25 hrs. 3.5 hrs. 9.75 hrs. 11.500 x .81 9.75 965 lb./auto.hr. 3.7.1.3 Running to 90% conversion - make up of volume contraction. c Reaction time 8.0 hrs. Outqycle time 3.5 hrs. Cycle time 11.5 hrs. Output rate 11.500 x 0.9 11.5 961 lb./auto.hr. 3.7.1*4 'with make up to 90% full Reaction time = 7.0 hrs. Outcycle time a 3-5 hrs. Cycle time '= 10.5 hrs. Output rate - H.500 x .9 10.5 a 998 lb./auto.hr. c 3.7.2 At 90% full. 3.7.2.1 At 81% Conversion Reaction time Outcycle time Cycle time Output rate 7.25 3.5 10.75 13.800 x .81 10.75 1,010 lb./auto.hr ucc -10- 3.7*2.2 At SO1/* conversion Reaction time Outcycle time Cycle tiiue Output rate 8.5 3.5 12.0 12.0 1.035 lb./auto.hr. 4 Expected output rate from proposed 5500 Imp. gal. Autos, under Aycliffe conditions. c 4.1 Tolerable Reaction Rate from jacket cooling. Assumptions: U = 52 BTU/hr. ft? P. Reaction temperature = 540. Others as per 2.1 Effective jacketed area = 340 ft.^ q * U A Cktm. = 52 x 340 x 36 x 1.8 = 1.135.000 BTU/hr. R = 1.620 lb./hr. G 4.2 Using normal methods for VY.18 Charge wt. = x 9,140 = 16,800 lb. Volume contraction = 550 x 16.800 = 644 Imp. gal. 9,140 Juckot controllable reaction rate = 1.620 100 167800 X = 9.65$/hr. by water addition in last hour = fAj? 1,8 700 x 16,800 =. 4.35?^hr. ijr-t039885 -11- Total acceptable reaction rate Reaction time Outcycle time Cycle time 11.0 hrs. 4.5 hrs. 15.5 hrs. 14.0 hr3. Output rate 16.800 x .9 15.5 970 lb./auto.hr. and on 5 autoclaves = 815.000 lb./wk. 4.3 Running with increased charge wt, to 89# conversion Charge wt. = 9.800 x 16,800 9,140 18,000 lb. Volume contraction rate control = 18.000 4.35 16,800 X 3 4.65%/hr. Jacket controllable rate = 1.620 100 18,000 x 3 9.0^/hr. Max. acceptable reaction rate = 15.65%/hr Reaction time Outcycle time Cycle time Output rate 10.5 hrs. 4.5 hrs. 15.0 hrs. 18.000 x .89 15.0 for 5 autoclaves 1.110 lb./auto.hr. 954.000 lb,/week Ratio of Output rato of 5,500 gal. to 3,000 gal. (of. 2.6) 1.110 728 1.53 Confirming the assumptions made in the design report. k C -12- 4.4 With reduced outcycle time to U.C.C. value Cycle time = I4.O Output rate = 14.0 for 5 autoclaves 1,150 lb./auto.hr. = 1.000.000 lb./week C ucc 039887 039888 i r' 1 0 1 <U U] u. $ 1 0 l vu CWrt IHtlt N1101 ^Iic^1(l4t1f QWIJ iHf IT N |M1 | * I Cf* H< I I C/fty INTERNAL CORRESPONDENCE CHEMICALS AMD PLASTICS To (Name) Division locaUort Copy to Mr- L. C. Lovenstein (2) New York Office - 29 Mr. M. E. Eisenhour Dr. A. B. Steele Mr. L. S. VanDelinder Mr. R. W. Wheeler P. O. BOX 8361, SOUTH CHARLESTON \^$f^ngGINIA 25303 A/3^ t Date February 111, 1970 received Originating Dept. BXL's Expansion Proposal for PVC Resin R. N. Wheeler Dear Larry: Thank you for the opportunity to review the engineering details of BXL's expansion proposal for PVC resin. As you suggested, I submitted the proposal to Mr. L. S. VanDelinder of our Material Engineering Group, and he has commented at length on the proposed materials of construction. I am attaching two copies of his comments - one for your files and one to be transmitted to BXL. If BXL has questions about Mr. VanDelinder's comments, I suggest they correspond directly with him. In addition, I offer the following comments or questions: 1. Why do they find it desirable to dilute the resin slurry before feeding it to the centrifuge: This seems to be a step in the wrong direction, since the water is to be removed in the centrifuge. If they could achieve a higher solids content of the resin leaving the centrifuge, obviously the capacity of the drying system would be increased. 2. BXL is to be complimented on the apparent simplicity of their monomer recovery system. 3. What is removed in the carbon filters? Are they really necessary? I am returning herewith the copy of BXL's memo which you sent to me. Very truly yours, 4ft' $ylf**U**l*4 W. R. Manning * WRM:rw j ) --p- 039891 f - j INTERNAL CORRESPONDENCE CHEMICALS AND PLASTICS ft* ?0 ~Tlm *1. w. P. O. BOX 8361, SOUTH CHARLESTON, WEST VIRGINIA 25303 To (Name) D.vision Location Copy to Mr. W. R. Manning Building 2000-3329 Technical Center Mr. J. H. Howell Dare Originating Dept. Subject February 2, 1970 Engineering Materials of Construction for English Suspension Vinyl Resin Plant Dear Ray: The proposed plans for expansion of the suspension PVC resin plant in England have been reviewed from the standpoint of materials selection. Before making comments on specific items of equipment, some general statements regarding corrosion by vinyl chloride are in order. Wet vinyl chloride is corrosive. At the lower temperatures (ambient and below), the rate of attack is not excessive. Contamina tion of the product contained in steel will result, rusting of the surface will occur if any oxygen is present, and an uneven corrosion will develop. On the stainless steels, severe pitting can result, and if perforation by pitting of the metal does not occur first, stress-corrosion cracking will ultimately develop. These comments are generally applicable to conditions of low HC1 acidity with available water whether contained in a monomer or derived from a moi resin. Vinyl chloride monomer which has been caustic washed and taken to a -20C dew point will not produce such corrosion. The product can be handled in the stainless steels at the higher tempera tures without corrosion of any type. Most problems produced in handling such a dry monomer have been occasioned by the breakthrough of wet product. The stress-corrosion cracking of the stainless steels (18-8 or austenitic types of the 300 series) can be a particularly expensive maintenance item if it is proposed to use these in moist HC1 environments. The time to initiate cracking these alloys in such an environment is controlled by the temperature, the magnitude of stress in the metal (residual or imposed), the availability of an oxidant, the availability of crevices or other areas where the chlorides can concentrate, and certain other factors of lesser importance. Consequently, no corrosion rate or time-to-failure can be cited for an application beforehand. As a general guide, the use of the stainless steels in such moist chloride environments will not be attractive above 90C. At lower temperatures, the type of equip ment and ease of repair are factors in deciding whether the alloy should be used. Certain designs of equipment such as pump castings, centrifuge parts, and piping lend themselves to stress relief before installation. With the lower stress levels in the metal, a satis factory life can be achieved. This is not an infallible solution I ICC 033892 2- but should be considered whenever stress-corrosion cracking is considered to be a possibility. We have been using a significant amount of stainless steel in chloride resin service. Varying results have been obtained, so no firm statement can be made except that, if trouble-free operation is desired, do not use the stainless steels. If higher than average maintenance costs are not objectionable, a savings in capital cost can be achieved by using the stainless steels. The following comments on the specific items of equipment proposed. 5.1.1 Tanker Unloading and Storage Unloading pump - OK in stainless Transfer pump - OK in stainless Monomer storage - Bare steel can be tried for this dry vinyl chloride if desired, but we would recommend a baked phenolic coating on the interior. The coating can be applied at a later date if found to be necessary, but you are sure of clean monomer if you do it now. Assure that the phenolic coating is well baked. 5.1.2 Autoclaves, blowdown and dilution Autoclaves - Glassed steel is definitely to be preferred. I believe both American and German sources would fabricate autoclaves to meet the required specifications. If glassed steel is not to be used, a stainless clad steel can be tried. Do not use solid stainless construction with cooling water used on the exterior of the shell. Cracking from the water side can occur after only one or two steamings "of the interior. "Provide adequate treatment of the cooling water and have steel exposed on that side of the shell. Unfortunately, using a stainless-clad steel provides high metal stresses on the interior and stress-corrosion cracking can be expected at some indeterminate time. Stress relief of the plate after fabrication is not of value ' in this case, since the stainless will go into tension after such a treatment of the plate. Keep the bulk temperatures down, the interior wall as cool as possible, the walls as clean as possible, and do not steam the interior- if other methods of cleaning can possibly be devised. If the latter is impossible, steam as briefly and infrequently as possible. Blowdown and Dilution Tanks - Could baked phenolic coated steel tanks be used here? If not, a glassed steel is recommended. Keep your use of stainless steel to a minimum in this area. 5.1.3 Drying and Handling Flash and rotary dryers - Cracking of the austenitic stain less steels in this area is a certainty. However, the alternative 039893 -3 - materials are not attractive economically. Texas City has had cracking and repairs, but the equipment continues to provide satisfactory service. Stress relieve all heavier stainless steel parts, such as the centrifuge bowl and internals. Stainless steel must not be used for the hot air heating coil or rapid failure will occur. Use cupro-nickel or comparable finned tubing for the heat exchanger. For smaller parts and flat removable surfaces consider the use of baked phenolic coated steel again. Also, the use of glass-reinforced polyester or epoxy plastic construction should be evaluated thoroughly for the dryer ductwork and lighter gauge construction. Unless static electricity build-up can be foreseen in such construction, we see no reason why this lesser expensive, clean material of construction should not find a much more extensive use for such applications. Product Sifter - Do not place aluminum in contact with high-chloride vinyl resins, pitting of the aluminum develops rapidly and the resulting attack can produce most undesirable metallic contamination of the resin. Stainless steel or coated steel should be used. At this point, where essentially no monomer exists, the use of a catalyzed epoxy coating on the steel can be started. A heavy coating of epoxy (10-20 mils) over steel can be used to handle the resin from this point on except where moving parts are required. Intermediate Storage Bins - Coated steel as specified. Centrifuge - Stainless steel (stress relieved parts) as specified. For this and other stainless steel designated, use type 316L wrought alloy or CF3M castings. Do not use any 400 series stainless alloys. Even the 316L alloy may rust in areas eventually and require cleaning to restore the desired surface. Bagging machine - Stainless steel is OK but consider where coated steel and glass-reinforced plastic might be used to economical advantage. 5.1.4 Monomer Recovery Rather than comment on the individual items listed here, I refer you to my original comments regarding the use of stainless steels in wet monomer. Such a wet monomer is handled in stainless steel equipment at Texas City without incidence of corrosion. To do this it must be emphasized that the temperatures must be low and the acidity maintained as low as possible. Aqueous layers with the vinyl chloride should be drained constantly from the vessels and not allowed to increase in acidity with time. The reference to a carbon filter to be constructed of stainless steel raises questions also. If the monomer going through this trap is wet, a significant pitting of the stainless steel can ucc 039394 4 result from galvanic action with the carbon. A phenolic-coated steel vessel would be preferred if the acidity of the monomer is not sufficient to attack the steel greatly at pinholes that might exist in the coating. 5.1.5 Suspending Agent and Catalyst Storage items. Stainless steel is appropriate for construction of these 5.1.6 Se rvices The only comment relates to the deionizing unit. What type of "deionizing" product is used in the vessel? Is back washing with an acid necessary? If so, a coated steel will obviously be required. We trust these comments will be helpful. information is requested, please contact me. If additional Very truly yours, LSV/bh U'X 039895