60/364,483, filed March 15,2002, entitled METHOD FOR CONVERTING METHANOL INTO ETHYLENE.

This invention pertains to methods for converting methanol into mixtures of olefins, mainly ethylene and propylene.

Research of conversion of methanol into olefins has gone on worldwide since the discovery of the mixed silico-alumina-phosphorous type catalysts by Union Carbide around 1980. Initial work was carried out at Union Carbide by Jeffrey M. O. Lewis and Giacomo Corvini. This work was at first carried out in a micro-reactor with a fixed bed of unbound SAPO catalyst, the particles of which were most likely of a size of about 1 micron. Tests with low conversion of methanol showed a high ratio for production of the C2 olefin over the C3 olefin, but this ratio decreased especially at higher conversion of methanol. Finally at very high conversion of methanol the C3 yield was even higher than that of ethylene. When they later switched to a larger catalyst aggregate, the time on stream (TOS) became smaller. This led to a switch to fluid bed operation. The work was then soon taken over by UOP and described in many patents and also in the literature.

In the publications of UOP the selectivity of the different olefins produced in the conversion of methanol over this new catalyst has been reported as about 49% ethylene, about 30% propylene and about 10% butylenes. Most of the research on the use of SAPO-34, as reported in the literature, has led to very similar or only marginally higher ethylene yields. While in some isolated cases much higher ethylene selectivities have been published, scale-up of these experiments has not been reported at all and seem to have led to difficulties.

Silicoaluminophosphate (SAPO) molecular sieves serve are particularly desirable catalytic materials in converting methanol to olefin compositions. They are particularly good catalysts for making olefins such as ethylene and propylene from methanol. The advantage of using SAPO based catalysts, particularly SAPO-34 based catalysts, is that such catalysts have relatively high ethylene and propylene selectivities.

However, SAPO catalysts undergo relatively rapid deactivation due to coke formation.

Silicoaluminophosphate molecular sieves are generally classified as being microporous materials having 8,10, or 12 membèred ring structures. These ring structures can have an average pore size ranging from about 3.5 A to about 15 A. Preferred are the small pore SAPO molecular sieves having an average pore size of less than about 5 A, preferably an average pore size ranging from about 3.5 A to about 5 A, more preferably from about 3.5 A to about 4.2 A. These pore sizes are typical of molecular sieves having 8 membered rings. It is preferred to use the SAPO-34 catalyst in a form in which the original very small catalyst crystals can easily be contacted; either the catalyst be used as made, or only in larger aggregates, wherein the pores in the catalyst aggregates particles leading to the original SAPO crystals should be appreciable larger than the actual catalyst pore size; preferably they should not present extra diffusion barrier, but should be in the form of "highways"in the larger aggregate particle.

This invention preferably uses the earlier rejected fixed bed process. This is now made possible by the attainment of much longer times-on-stream (TOS) than earlier deemed possible. In this operation it also has been found possible to attain much higher selectivities for the more desirable ethylene.

This application is about the conversion of methanol into olefins, using as catalyst SAPO-34 or similar catalysts, characterized by having a narrow pore in the actual catalyst. A number of rules were found to be necessary or at least helpful in attaining the desired goals. These rules were presented in a Provisional Application Serial No.

60/364,483, filed. March 15,2002, entitled METHOD FOR CONVERTING METHANOL INTO ETHYLENE, which Provisional Application is incorporated into this Application in its entirety.

The rules given for use of methanol only were and are: 1. The inlet partial pressure of methanol at the inlet of a catalyst contact is to be lower than 0.05 atmospheres absolute (ata), preferably lower than 0.03 ata, and most preferably lower than 0. 015 ata; as the catalyst becomes more and more aged, the limits can be moved upward gradually in tandem with the decline of catalyst activity to a maximum of twice the given values; i. e. , 0.1 ata, 0.06 ata and 0.03 ata, for an aged catalyst; 2. The methanol partial pressure at the exit of a catalyst is higher than 0.002 ata, preferably higher than 0.005 ata and most preferably higher than 0.008 ata.; 3. Preferably a multi-stage fixed bed reactor (s) is used, most preferably in adiabatic operation; 4. Each of the stages is plugflow, if desired with a recycle over one or more stages; 5. After the first stage and before each of the following stages methanol is added to keep the methanol partial pressures into and out of the catalyst beds within the set limits; the temperature and even the physical state of the added methanol can be chosen so as to compensate for the temperature rise in each stage; 6. The catalyst used is SAPO-34 or any other silico-alumina catalyst with similar behavior, defined as having some minor or severe diffusion barrier at the entrance to the inside of the catalyst elementary structure, combined with a tendency to produce olefins from methanol, probably even known to produce these olefins at pressures of 1 ata or higher, but up to now with substantial after-reactions to higher olefins, possibly even to aromatics and coke; the catalyst is in a form which allows open access to the small crystals initially formed in the preparation of the catalyst; as one of the possibilities unbound catalyst can be used with an average size of around one micrometer; 7. The number of stages is chosen, so that together with the other conditions mentioned, a final ethylene yield is obtained equivalent to a maximum partial pressure of 0.06 ata, preferably to a maximum partial pressure of 0.04 ata, most preferably to a maximum partial pressure of 0.025 ata; in case of use of recycle the number of stages can be limited to a single stage, but it is preferable to use several stages; for example a single stage with appropriate recycle can be used, followed by several stages without recycle; if no recycle is used, the number of stages can be substantial, up to 20 stages or more in order to achieve the desired olefin partial pressure; 8. When recycle is used, it can be effected with recycle compressors, but it is preferred to effect the recycle with jets, fed by the feed mixture ; 9. It is preferred to use a diluent of the methanol; steam is the preferred diluent ; a steam to methanol inlet ratio to the first stage can be as low as 2 to 1, but a ratio of 4 to 1 is preferred; it is possible to use higher ratios up to 200 to 1 ; when high ratios are used in the inlet to the first stage the following stages can be fed with pure methanol or with methanol-steam ratios, which contain much less steam ; 10. Especially when aiming at lower maxima for the final ethylene partial pressure a total system pressure of less than 1 ata, preferably less than 0.6 ata, most preferably less than 0.4 ata can be used; 11. The temperature of reaction is between 350 and 450 C, with a temperature rise due to the MTO reaction of no greater than 50° preferred ; regeneration is with oxygen-containing gases; the regeneration is with increasing temperatures, starting at 350 C and going up to a maximum of 550 C; 12. As additional improvement to further improve the production of ethylene the catalyst can be pre-treated to cause partial aging; this can for instance be achieved by contacting the catalyst with a methanol-or propylene- containing gas mixture with a methanol or propylene partial pressure of between 0.1 and 0.3 ata at temperatures between 425 and 475 C; after this pretreatment a light regeneration may be used, which removes not more than 20 to 50% of the contained coke. The treatment may be repeated to achieve a better performance as to the ethylene to propylene ratio.

Example In a 3/4"ID reactor two layers of catalyst were present. Each layer consisted of small powder SAPO-34 catalyst of approximately 1.1 micron size, diluted with a slight excess of also 1.1 micron-sized alpha alumina. On the bottom and the top of each layer glass wool was present to contain the fine powder. A total of 1.22 g SAPO-34 was used together with about 1.62 g of a alumina powder. The rest of the reactor was filled with alumina particles of 2 mm size. After the reactor the effluent was cooled and separated water was bled out. The remaining gas was partially recycled with a gas pump.

The reactor was deactivated by treatment with nitrogen, containing 0. 3 wt% CS2 for 30 minutes. Then the catalyst was deactivated with a feed of 20 mole% methanol and 80 mole% steam for 1 hour. The catalyst was partly regenerated with steam for 30 minutes at 400°C.

The feed to the reactor was 62 g of a methanol-water mixture per hour, containing 1 wt% methanol, entering the catalyst bed at 400°C and at slightly above atmospheric pressure in order to avoid entrance of air at the exit, which air then could enter the recycle pump. The recycle pump was set at a recycle of 22.6 ml/minute.

The following results were obtained at about a 50% conversion of the methanol: At TOS C2 H4 C3 H6 C4 H8 C3 H8 C02 H2 DME Total moles Selectivity (hrs) C2+C3 0.25 77.17 17.17 1.56 0 1.15 2.54 0.41 100 0.962 1 76. 9 20.4 0.38 0.91 0.18 1 0.23 100 0.976 2 75.98 20.06 0.19 0.78 0.19 2.61 0.19 100 0.983 As the analysis had not determined the amount of saturates, the number of total moles was closed to 100% by assuming propane to be the lacking component. This species would be the most detrimental of the likely candidates (these are methane, ethane and propane; at the low amount of butylenes, the presence of saturated butanes is less likely).

As not much change was noted during the short time on stream and in line with other longer operating similar experiments, the expectation is well founded of a possible TOS of 1 day.

After practically all the published research on standard SAPO-34 catalyst has resulted in acceptance of the two"facts", namely the impossibility of improving the selectivity picture for the lowest two olefins and secondly the impossibility of operating the reaction cleanly in a fixed bed operation with a reasonable time on stream, it is very surprising that with a rather simple limitation on the partial pressures of the main reactant such different results have been attained. While it is not the intent to limit the value of this invention to any one explanation, it is still important to try and provide some possible understanding for these unusual phenomena.

The reason may well be in the peculiarity of a fast reacting catalyst, which is partially closed off from the outside world. In the beginning the different theoreticians who have worked on this new reaction did not attach enough significance to the effect of a diffusion barrier between a fast reacting catalyst and the outside world. In analogy with the reaction of methanol over ZSM-5 it seems logical to expect a very fast reaction of methanol over the catalyst. As there is a flow of co-produced steam out of the catalyst, the outside steam is not directly available for cooling the reaction. The heat removal therefore has to be taking place by the slower heat exchange with the outside gases. The complex picture of outward diffusion of the different species of products out of what most likely is often an equilibrium mixture has defied analysis so far. The present findings are a clue of the importance of these remarks.

As stated before, Lewis and Corvini found a poor selectivity for ethylene at high conversion of methanol. The conversion in the Example was stopped at about a partial pressure for methanol of 0.025 ata, which still allows an overall high selectivity for ethylene. The level at which the after-reaction of ethylene becomes critical is most likely dependent on the ethylene partial pressure reached, as the side reaction is expected to be proportional to the C2 partial pressure. Therefore when using substantially higher final ethylene partial pressure, for instance by staging, a higher critical value for the final ethylene partial pressure should be used.