Document 2R8rXLNEQwD6npNa07v0qN5wa

Interoffice Communication To From F. Kennedy W. R. Sorenson ROUGH DRAFT Date Subject June 4, 1974 LARGE VERSUS SMALL REACTORS - THE IMPACT ON THE VINYL CHLORIDE (VCM) TOXICITY QUESTION 5*3 C3 <S5 as 3 av <S3 The suspension polymerization of VCM is the dominant process (80%) of making polyvinyl chloride (PVC). There are certain steps common to all suspension PVC processes. (Henceforth, all references to the polymerization process will mean the suspension process.; 1. Reactor charging 2. Polymerfzation 3. Recovery of unpolymerized VCM 4. lumping of the batch from the reactor 5. Cleaning of the reactor in preparation for the next charge While all suspension operations would involve these steps, they do not necessarily occur in that order in each process, nor would they necessarily be done in the same way. These differences reflect the chemical and engineering methods developed by each manufacturer. We would like to compare large reactors (LR) with small reactors (SR) in two ways, as we have seen it in the Conoco operations: - How the mechanical operations involved in each affect the control of VCM in the work areas. - How the needs of the PVC fabricating industry are served by each. First of all, it is logical to expect that the more mechanical operational steps inherent in a PVC process, the more opportunities for the escape of VCM* Compare then, the resulting mechanical operations in two plants, of say, 180 MM lbs. capacity, one based on four LR's of 18,500-gallon 0O )j A capacity and the other on SR's of 2,200-gallon. It should be understood that the capacity of a PVC plant is based on more than just the number of reactors. Other factors include: the cooling water available, the product mix, and the particular chemical and engineering approach in use. But under the given assumptions, the following emerges: a. Size ratio: IB,500 gallon 2,200 gallon = 8`41 b. Number of SR's needed to equal a four LR plant, assuming productive capacity of a reactor is proportional to size: 8.41 x 4 - 34 e Thus, about thij^ty-f&ur times the number of operations must be carried out in an V SR plant relative to the LR plant; that is, thy-fouir times as many reactors must be charged, a polymerization contained, then stripped of VCM, dumped, cleaned, etc. Clearly, the number of mechanical entities (valves, flanges, pumps, etc.) requiring maintenance is substantially less in an LR plant, a strong positive factor in its favor. In the Conoco LR operations, most of the mechanical steps are carried out by remote control from a panel that permits the operator to keep all phases of the process under observation by means of charts, schematic diagrams, signal lights and closed circuit TV cameras. This system of remote control applies not only to the reactor operations, incidentally, but to the slurry blend tank-dryer operations also. Basically, the only routine reactor operation carried out in the immediate 3 reactor area itself in the LR process is the charging of initiator to each batch. Steps have been taken to reduce the frequency of the latter operation. In the SR plant, by contrast, essentially all the operational steps are carried 5*3 CJ ro <SS S3 --4 1 ro out within the buildings that actually house the reactors. Reactor cleaning has traditionally been accomplished in PVC plants by operator entry. In the Conoco LR process, operator entry is necessary only about once every thirty days; in the SR's, about once every four to five days. All reactor entries are done in every case with proper precautions taken to insure a low prevailing VCM level, but reduced entry frequency is obviously to be preferred until the goal of zero personnel entry can be achieved. It is also evident that many more operators are required for a SR plant of capacity equal to one based on LR's. The ratio is about three and one-half to one. The Conoco LR technology uses an essentially "open air" method of housing the reactors. Except for a roof, the reactors and associated piping and supporting members are enclosed only by a protective metal screen-like structure covering about two-thirds of the total vertical rise involved. The effect is to protect worker and equipment from direct exposure to the elements and to take full advantage of dispersal and dilution factor offered by wind and air currents. The Conoco SR plant and most PVC plants house the reactors in a relatively conventional industrial factory building, with ventilating fans as the source of air turnover and VCM dilution. The same factors that apply to large and small reactor plants are valid in regard to the question of slurry transfer and storage ; in the LR plant slurry blend tanks are very close to the reactors and are essentially outdoors and require no direct personnel involvement in filling them as the reactors are dumped. As the 5*3 k D <S3 number of resins increases; e.g., from one or two to four or five, the number of slurry blend tanks increases. For the SR plant, the contrary of all these factors prevai1s. ro C5J a? O' --.4 rfl yj In the dryer section of the LR plant, there is essential similarity to the SR plant. There need not be bagging in either plant type and hence no warehouse storage, which eliminates an area of concern for VCM control. Storage is in silos of over 160,000 pounds capacity and shipping is generally by rail car. These are features reflecting the marketing concept involved rather than intrinsic aspects of an LR plant. In this connection, however, it is valid to say that the standard spectrum of commercial PVC resins can be made in an LR plant. The commodity resins, i.e., resin suitable for pipe, calendering, wire and cable, clear film, injection molding and other applications can be manufactured. We have had no occasion to manufacture bottle grade resins for marketing reasons. In this respect, the LR plant serves the market needs about as effectively as an SR plant with, as can be expected, considerably greater overall efficiencies. A summarizing element of the difference between LR and SR plants is that the former operates at 97~98~l/2 percent efficiency in VCM-to-PVC conversion in contrast to about 3k percent for the latter. This difference reflects the mechanical operating differences already discussed, in that with fewer operational steps there is less loss of solid PVC and gaseous VCM and the work space is inherently less at risk from VCM in the air. W. R. Sorenson