Document 2qvz7Dx4Ed7J7Bd7jXvzJ3vzp

(conoco) Interoffice Communication Kennedy To Charles M. Starks From R. J. Convers, R. Pate June 9, 1981 Subject Potential Plans Plant ^jLfjP'--------------- ~r - - -------- y % Jtvwix . due* SX.C. I. INTRODUCTION This document is a sequel to the 5/5/81 IOC by R. J. Convers des cribing oxy catalyst problems at the VCM plant. The suggestions here are an outgrowth of discussions about plant operating data, lab oxy work by R. J. Convers and discussions by Exploratory Sec tion personnel concerning the best approach for precluding the , 4-5 month turnaround times which seem quite likely when O2 based I oxy and the VCM expansion become realities for Conoco (March 1983 - tenrt--ative ly)_._______-______ Sl*4**-f II. GENERAL APPROACH In the absence of, or the likelihood of having, oxy pilot facili ties before a plant demonstrated need, we propose a series of plant test runs in which minor catalyst changes are made in a specified fashion based on our knowledge of oxychlorination. The success of the plan hinges on plant cooperation, a general concen sus that the approach is technically sound (free from catastrophic operational problems and associated economic detractors) and the ability to obtain suitably mnrlifipH catalysts. In essence success depends on our ability to sell VCM plant personnel on the idea that potential advantages would far outweigh the costs and that run times of 5-7 months (currently being experienced) would be the most disasterous thing which would happen during the sequence of tests. From preliminary considerations we specu late that additional operating costs to the plant might be in the $100-200M range, a fairly small price to pay for extending turnaround times from 5-7 months at reduced rates (with now avail able catalysts) to 9-12 months at 112% of design (with a modified Harshaw system). The benefit, calculated on the basis of annualiz ed PED turnaround cost data as a function of turnaround time could be as much as $3.5MM for the expanded oxy operations. In the plan all experimental justification comes from historic plant operating data assuming only minor insight from our lab oxy work. III. PROPOSED PLAN The heart of our proposal is based on plant test runs described below. Due to time restrictions only two of the three runs could be made prior to startup of .the VCM Expansion (O2 based oxychlorina tion) scheduled for March 1983. This assumes ~9 month run times for each test run. CCR 000018276 Charles M. Starks June 9, 1981 Page two Scs Run 1: A test run using theo old Harshaw oxy catalyst (having porosity mainly in the 120-160 A range) system in which the Tecosil diluent (broken glass) is replaced by a 3/16" tableted inert dilu- ent" Run 2: A test run using a new Harshaw catalyst, actually a margin ally modified version of the old Harshaw oxy catalyst in which the porosity is mainly in the 170-200 A range. Again, use of this catalyst would be coupled with use of 3/16" tableted, inert diluent in place of Tecosil. Run 3i A test run using another modified version of the catalyst used ^in Run 2 such that catalyst porosity is mainly in the 220300 A range. As before, catalyst use would be coupled with use of 3/16" tableted, inert diluent in place of Tecosil. Logic: a) Historic plant run data depicts 9-10 month oxy run times at 112% of design capacity using old Harshaw catalyst (120-160 X pore range) when run with Tecosil diluent in R-301. This is a significant improvement over current run times 5-7 mo. with SCS catalyst (at reduced capacity) or historic run times with BASF catalyst. The old Harshaw-Tecosil system exhibits high AP problems early in the run presumably because of poor packing properties ascribed to Tecosil in the presence of 3/16" tablets (i.e., Harshaw catalyst) - interstitial flow properties are inherently reduced using irregular shaped/sized Tecosil diluents. This system nonetheless gives 9-10 month run times because of slow coking processes with the Harshaw system. ,h) Historic plant data using BASF 3/16" tablets (no Tecosil) ended prematurely after ~7 mo. (at 1127. of design) due to rapid coking which causes excessive AP problems. However, the initial stages of this run did not have high AP problems like those of the Harshaw-Tecosil system in a) above packed with uniformly shaped 3/16" tablets). From a) and b), use of inert diluents shaped into 3/16" tablets in combination with Harshaw 3/16" tableted catalyst should exhibit a major operating advantage. Lower AP performance early in the run (10--12 psig) should extrapolate additively to longer plant run times with the Harshaw catalyst than pre viously experienced with the historic Harshaw-Tecosil systems. The historic Harshaw-Tecosil system characteristically operated for 9-10 months at 112% of design in air based oxy (a signifi cant improvement over the 5-7 month runs experienced with BASF catalyst at similar rates as for the Stauffer SCS system at reduced rates). Using the old Harshaw catalyst with 3/16" tableted diluent we conservatively estimate plant run times of 11--12 months at 112% of design. jU-Car^Z CCR 00001&277 Charles M. Starks June 9, 1981 Page three d) BASF and Stauffer SCS catalysts (now the only catalysts available commercially) both coke up and deactivate much faster than the old Harshaw system. Present thought ascribes this condition to differences in pore size distribution of the support. BASF and Stauffer SCS catalysts have porosity distri buted mainly in the 60 X range. This contrasts the pore distribution in the old Harshaw catalyst which centers in the 120- QJ 160 A range. Support for this position also comes from lab oxy data. Lab oxy runs using a steam calcined Catapal alumina support (pore structure centered in the 220-250 A range) gave short term performance data in which reduced levels of both EtCl and heavy ends (1,1,2-TCE) were observed relative to SCS or BASF systems. The latter may be particularly signifi cant because: 1) heavy ends materials may inherently be formed in smaller quantities on larger pored catalysts. 2) once formed heavy ends materials likely escape more readily from larger pored systems than from smaller pored systems hence forming coke at dimished rates. 3) combined effects from 1) and 2) could be responsible for reduced coking and longer run times in a now unspecified ` c fashion. The test runs suggested should dependence of coking and run times on eras in question. from the suggested plant test runs would estimate $100-200M of incremental plant e test runs, assuming 9 month vs 6 month >f design, more than pay for themselves, conservatively estimate annualized savings sting operations. In addition very timely ita would be obtained for downstream use s with even larger dividends. If the plant will not accept the plan above we could develop accel erated oxy catalyst life tests in our shop. The qualities of this approach are: * Additional time/expense in an open ended experimental approach. Questionable results would be obtained. Even under the best of conditions predictions of catalyst run times in the plant would be doubtful. Results could be obtained rapidly and cheaply. There would be no risk to plant operations. OCR 000018278 idrKS-- - -- .!. vj June 9, 1981 Page four Another approach would be arrange a cooperative research program with Shell to take mutual advantage of our combined resources on a mutal problem in a setting where we could take advantage of their pilot facilities. If it were possible to arrange such a venture the data obtained would be: * Expensive * Highly credible (likely) * Slow to obtain * Risk free to plant operations Such a program with Shell may be difficult or impossible to arrange but nonetheless contact by some process should be made with Shell to determine the extent to which they (as developers of the general C>2 based oxy technology) are concerned with, or are handling, the problems we are experiencing and foresee for the future. Contacts with Shell in this regard should probably not be made until the VCM plant's tactical position is determined. V. TIMING The VCM plant has been "limping along" at reduced rates with an SCS catalyst system. They will likely have a turnaround due to catalyst deactivation in mid-July. The plant plans to charge a mixed DASF-SCS system which likely will also run for only 3- 6 mos. Unless we can rapidly implement our plans, Lhe December 1981-January 1982 time frame is the earliest any of the test runs could be made at the plant. With the run times expected (9-11 mo. using Harshaw systems) one and possibly two of the suggested runs could be made prior to the March 1982 start-up for the VCM expansion. Obviously, if the plant is agreeable no time should be wasted in contacting Harshaw about preparing catalysts of inte rest and pursuing any necessary support functions in Exploratory Research should development of a large pore support be needed. Implementation of beneficial results from the test runs proposed would likely not occur until after the March 1983 startup which will surely be made with Stauffer SCS catalysts to obtain contract ual guarantees. Assuming an initial 4 month turnaround July 1983 would be the earliest time to implement results from the test runs proposed. Exploratory Work Needed in Support of the Proposal Before we approach the VCM plant with these ideas we must demon strate that the 3/16" tableted diluent of choice does not adversely affect selectivity in the oxy system with the Harshaw catalyst in question. Such data would come from comparative oxy lab runs. In similar fashion demonstration of initial AP effects for the Harshaw--Tecosil vs Harshaw--tableted diluent systems would be desire- CCR 000018279 k I * 1^ Charles M. Starks June 9, 1981 Page five VII. able. These data would come from model experiments where oxychlorination is not occurring and should help establish the credibility of our proposal to the plant. Assuming the VCM plant is receptive to the suggested format, Harshaw must be approached concerning the preparation of the old Harshaw catalyst as well as the larger pored versions suggested. If Harshaw can prepare catalysts with the features desired we must at least screen them in the lab. If they cannot prepare them, we may have to work independently, then with them to develop the large pored support needed and test the catalysts in our lab re actor. Apart from the catalyst support development work mentioned above we should attempt to steam calcine some preformed commercial oxy catalysts and assess the value of such an approach. This is simple, we can do it, and the technology to make the modified catalysts may be patentable. PREDICTIONS FROM PLANT AND LAB DATA Based on plant and lab data the following predictions can be reason ably made. Again, most are extensions of demonstrated past perfor mance . The catalyst concoction to be implemented by the plant in mid-July will have a 5-7 month run time. Since neither the BASF nor the SCS catalyst systems have performed well indivi dually, there is no reason a combination should be better. Results from proposed Run 1 should easily mimic old Harshaw results (i.e., 9-10 month run times @ 1127. of design) in the plant. Coupled with removal of the Tecosil diluent run time could be extended to 11-12 months. Even 9 month run times would be a quite favorable benefit relative to plant opera tions with SCS and BASF catalysts. A prediction of 11-12 month run times implies the absence of significant AP problems early in the run. Results from lab data using a steam calcined alumina support indicated the formation of significantly less ethylchloride and 1,1,2--trichloroethane than routinely seen over BASF or SCS catalysts. This trend may well be reflected in the plant setting and at least indicates no disasterous consequences from switching to a larger pored catalyst system. Predicting results from the larger pored catalyst runs (Runs 2 & 3) is somewhat dubious but it seems quite likely that run times in excess of 12 months could definitely be possible. Such benefits would likely occur due to diminished rates of catalyst fouling we believe would be associated with the larger pored catalysts. CCR 000018280 Charles M. Starks June 9, 1981 Page six VIII. PROPOSED TACTICAL SEQUENCE 1) Discuss recommendations and decide on best plan for hand ling the problems foreseen - In progress. 2) If plan is favorable in eyes of Chem. Res. personnel, demon strate utility of 3/16" formed diluents in lab oxy studies and gather AP data as a function of gas velocity in model experiments. 3) If results in 2) are as predicted, discuss plan for test runs with VCM plant personnel - decide go/no go and/or modify plan so they accept if needed. 4) If plant agrees to plan, immediately approach Harshaw to determine feasibility for generating large pored catalysts in their shop. If yes, decide go/no go with plant and R&D input -- get cost data. 5) If 4) is possible at Harshaw and we go, screen catalysts prepared by Harshaw to get some idea of performance expecta tions prior to use in the plant. 6) If 4) is not possible at Harshaw, provide experimental assistance from Exploratory Res. to develop large pored supports - transmit information to Harshaw and work with them to facilitate development of desired catalyst(s) screen catalyst(s) so prepared as in 4) - coordinate pre paration of commercial sized batch of catalyst at Harshaw (costs? quantities? Plant involved - R&D involved - decide go/no go with catalyst preparation) - must be done by "6/82 (assuming Run 1 is done with a 12 month run time) to assure catalyst availability for turnaround in 1/83. 7) Apart from considerations 1--6, steam calcine a small sample of preformed, preloaded SCS catalyst (and perhaps others)test catalyst(s) in lab oxy unit - decide utility of approach to large pored catalyst preparation - perhaps implement in 6 above if technically successful. IX. POSITIVE AND NEGATIVE PLANT CONSIDERATIONS Negatives: 1) Additional operating expenses to plant Added cost for specially prepared catalysts ~ $100--200M 2) A very remote possibility that plant operations could be impaired marginally. 00001-8281 CCfc Charles M. Starks June 9, 1981 Page seven Positives: 1) Good dividend potentially obtained for a minimum investment * Plant savings of '$3.5MM/yr seen a nominal reality for business plans based on expanded VCM operations. * Test runs would more than pay for themselves ('$1.2MM/yr @ 112% design + 9 month run times). 2) Catalyst life data obtained - highly credible. 3) Worst case turnaround times would likely be 5-7 months under current operating conditions. No worse than present operations. 4) Low market demand presently - a good time to experiment for times when optimized operations will be essential. X. ECONOMIC SENSITIVITY TOPLANT TURNAROUND TIMES OR CATALYST LIFE The magnitude of the economics is depicted from absolute turnaround costs assessed annually as a function of catalyst life. The values ascribed to turnaround costs were obtained from G. L. Fryar's IOC (01-11--80; Technology Evaluation Report--Shell/Stauffer OxygenBased Oxychlorination). Total catalyst replacement is assumed in all cases. CCR 000018282 Charles M. Starks June 9, 1981 Page eight Catalyst Life (months) 4 5 6 7 8 9 10 11 12 Annualized Costs ($MM) Existing Conoco operations (670MM lb/yr VCM) 5.42 4.35 3.63 3.11 2.72 2.42 2.18 1.98 1.81 Annualized Costs ($MM) Expanded Conoco Operations (1MMM lb/yr VCM)_________ 7.33 5.86 4.88 4.19 3.66 3.26 2.93 2.66 2,44 Obviously there is a great deal in incentive to extend catalyst lifetimes to the maximum degree possible. Robert M. Owens Group Leader rvwt'U} Ronald J. Convers Senior Research Chemist CCR 000018283