BACKGROUND OF THE INVENTION Propylene and ethylene are produced commercially by a number of methods, including, for example, steam cracking or pyrolysis of paraffinic materials, petroleum refinery cracking, and alkane dehydrogenation. U. S. Pat. No. 5, 043, 522 explains some of the problems with these methods for making propylene. Disproportionation of olefin mixtures to make ethylene or propylene is also known, and is described, for example, in U. S. Pat. Nos.

4, 180, 524, 4, 499, 328, and 4, 517, 401. These processes generally use amorphous catalysts such as tungsten or molybdenum oxides with other compounds that help to promote the disproportionation.

Steam crackers optimized for naptha feedstocks produce more C4 and higher olefins than steam crackers using lighter feedstocks such as ethane and propane. However, the amount of C4 olefins generated from steam cracking may exceed the need for these olefins in current uses such as paraffin alkylation and methyl tert-butyl ether production. Moreover, naphtha-based steam crackers also produce C5 olefins, which have limited high-value uses. In addition, C5-C9 olefins, which have boiling points up to about 150°C and are readily available, e. g., from catalytic cracking of kerosene or gas oil, also have limited utility. In sum, there is a need for processes that can upgrade C4-C9 olefins, when market conditions require, to propylene and ethylene.

Zeolites are a well-known class of natural and synthetic crystalline aluminosilicates that comprise networks of Si04 and A104 tetrahedra in which the silicon and aluminum atoms are crosslinked by shared oxygen

atoms. Zeolites contain channels or voids of characteristic dimensions. The channel openings, or"pores,"can be rather circular, but more often they are elliptical or irregular in shape. In some zeolites, the channels are one- dimensional and do not interconnect, as in a series of parallel tunnels. In others, the channels intersect and form large cavities at the intersections.

The channels normally contain cations such as sodium, potassium, magnesium, ammonium, or the like, and may contain protons or water molecules. Water can be removed by heating, leaving an active site within the catalyst. The cations can be exchanged, in whole or part, by other different cations, or by protons to make the"H"form of the zeolite.

Zeolites have been used for many types of hydrocarbon transformations, including, for example, toluene disproportionation (see, e. g., U. S. Pat. No. 4, 160, 788), hydrocarbon oil dewaxing (U. S. Pat. No.

5, 000, 840), selective sorption of hydrocarbons (U. S. Pat. No. 4, 423, 280), olefin oligomerization (U. S. Pat. No. 4, 855, 527) or isomerization (U. S. Pat.

No. 5, 177, 281 or 5, 817, 907), gasoline upgrading (U. S. Pat. No. 5, 298, 150), desulfurization of hydrocarbons (U. S. Pat. No. 5, 401, 391), and conversion of C5 linear olefins to ter-alkyl ethers (U. S. Pat. No. 5, 420, 360).

Zeolites have also been used in processes for making propylene or ethylene from mixtures of olefins and paraffinic hydrocarbons. For example, U. S. Pat. Nos. 5, 043, 522 and 5, 026, 936 describe a process for making propylene in which a mixture of 40-95 wt. % paraffinic hydrocarbons (C4 and higher) and 5-60 wt. % olefins (C4 and higher) are heated in the presence of certain zeolites. The examples use ZSM-5, a zeolite that has interconnecting channels in three dimensions with pore size indices greater than 25. U. S. Pat. No. 5, 026, 935 uses ZSM-5 to make ethylene from higher hydrocarbons (including butenes and/or propylene) by a cracking process.

Unfortunately, ZSM-5 is not well-suited for use in a fixed-bed process, which requires a prolonged catalyst lifetime, because it tends to"coke up"from aromatic hydrocarbons forming within its channels. ZSM-5, therefore, would require continuous regeneration if it were used in a fixed-bed process.

Still needed in the art are better ways to make propylene and ethylene from higher olefins. Preferably, the process could use the C4 to C9 olefin streams that are readily available from steam or catalytic cracking. Ideally, the process would retard catalyst coking and deactivation, which hampers productivity in many processes that use heterogeneous catalysts such as zeolites. An ideal process would give valuable ethylene and propylene in favorable ultimate conversions, yields, and selectivities, yet would have a long enough catalyst life to be suitable for use with fixed-bed reactors.

SUMMARY OF THE INVENTION The invention is a process for making propylene and ethylene. The process comprises heating one or more C4 to C9 olefins with a particular zeolite catalyst under conditions effective to produce propylene and ethylene. The catalyst has a pore diameter of 4. 4 to 4. 5 A and one- dimensional, non-interconnecting channels having a pore size index within the range of 23 to 25. Well-known zeolites in this class are MTT (ZSM-23) and TON (THETA 1).

The process of the invention affords high olefin conversions and excellent selectivities to valuable ethylene and propylene. Because catalyst lifetime is long (especially when the catalyst is passivated, as by steam treatment, e. g.), the process is ideal for use with a fixed bed of catalyst.

Major by-products are typically higher olefins that can be recycled to further boost ultimate conversion. Proper selection of the zeolite to those having the particular characteristics noted above minimizes coking and aromatic by- products, and maximizes yields of propylene and ethylene.

DETAILED DESCRIPTION OF THE INVENTION The hydrocarbon feed used as a starting material for the process of the invention is a mixture that contains one or more C4 to C9 olefins.

Suitable olefins are linear or branched isomers that contain four to nine carbons and a single carbon-carbon double bond. Thus, the hydrocarbon

feed can include linear and branched nonenes, octenes, heptenes, hexenes, and the like. Particularly preferred are streams that contain mixtures of C4 and C5 olefins, such as 1-butene, cis-2-butene, trans-2- butene, isobutene, 1-pentene, cis-2-pentene, trans-2-pentene, 3-methyl-1- butene, 2-methyl-2-butene, 2-methyl-1-butene, and mixtures of these.

While the feed can (and often will) contain other types of hydrocarbons, it preferably comprises more than 40 wt. %, more preferably more than 60 wt. %, of C4 to C9 olefins. Hydrocarbon streams suitable for use in the process of the invention include C4 and C5 streams (with or without the isobutene component) from fluid catalytic crackers or hydrocarbon pyrolysis units. A preferred hydrocarbon stream consists essentially of C4 and/or C5 olefins.

The C4 to C9 olefin mixture is converted to propylene and ethylene by contacting it with a particular zeolite catalyst. While many varieties of zeolite catalysts are known, only certain types are useful in the invention.

Useful zeolites are medium-pore zeolites or zeolite-type materials that have a 10-membered ring channel structure and a pore diameter of 4. 4 to 4. 5 A. By"pore diameter,"we mean the diameter of the ring aperture or pore measured at its narrowest dimension, or the smaller of the two major axes for an elliptical pore. Thus, e. g., a catalyst that has channels with elliptical apertures measuring 3. 3 x 5. 0 A does not meet the requirements of the invention because the narrowest dimension (the smaller of the major axes) is not "4.4 to 4.5 Å." Zeolites that have a pore diameter less than 4.4 A are not suitable for use because they are too narrow to permit entry of C4 to C9 olefins into the channels.

Zeolites useful in the invention have one-dimensional, non- interconnecting channels. By"one-dimensional, non-interconnecting" channels, we mean ones that are more or less parallel and non-intersecting.

Zeolite handbooks such as W. M. Meier et al., Atlas of Zeolite Structure Types, 4th Revised Ed. (1996), hereinafter referred to as"the Atlas,"identify such one-dimensional zeolites with a single asterisk (*) in their description of the channels.

As the Atlas explains, each system of equivalent channels is characterized by (1) the channel direction (relative to the axes of the type structure), (2) the number of either T- (usually Si or Al) or 0-atoms, in bold type, that form the rings controlling diffusion through the channels, and (3) the crystallographic free diameters of the channels in Angstroms, based on the atomic coordinates of the type materials and an oxygen radius of 1.35 Å.

The number of asterisks indicates whether the channel system is one-, two-, or three-dimensional. Interconnecting channels are separated by a double arrow (<-->). A vertical bar (1) means that there is no direct access from one channel system to another.

MTT (ZSM-23), a zeolite useful in the invention, has a one- dimensional channel structure. The Atlas describes its channels as follows : [001] 10 4. 5 x 5. 2*. The boldface 10 indicates a 10-membered ring structure, the 4. 5 and 5. 2 refer to pore diameter (in Angstroms ; two numbers because of the non-circular apertures), and the asterisk identifies the channels as one-dimensional and non-interconnecting. In contrast, MFI (ZSM-5), a zeolite that is not useful in the invention, has three-dimensional, interconnecting channels. The Atlas describes its channels as follows : { [010] 10 5. 3 x 5. 6 <--> [100] 10 5. 1 x 5. 5} ***. The triple asterisk denotes a three-dimensional structure, and the double arrow indicates that the channels interconnect.

Zeolites useful in the invention have a pore size index within the range of 23 to 25. By"pore size index,"we mean the product of the dimensions (in Angstrom units) of the two major axes of the pores. This is the definition used, e. g., by Haag et al. in U. S. Pat. No. 5, 177, 281. The pore dimensions are simply multiplied together to get the pore size index. For example, TON (THETA 1), a zeolite catalyst useful in the invention, has elliptical pores measuring 4.4 x 5.5 Å. Multiplying these numbers gives a pore size index for TON of 24. 2. Cancrinite, on the other hand, a catalyst not useful in the invention, has pores measuring 5. 9 A, and a pore size index of 34. 8 (i. e., greater than the upper limit of 25 for catalysts useful in the invention).

Information about many zeolites is now available"online"courtesy of the Structure Commission of the International Zeolite Association. The website address is : www. iza-sc. ethz. ch/IZA-SC.

What the zeolites useful in the process of the invention have in common are channels large enough to admit the C4 to C9 olefins and large enough to allow propylene and ethylene to diffuse out. On the other hand, the channels are generally small enough to minimize formation of hydrocarbon coke precursors within the channels.

Most zeolites are not useful in the process of the invention. For example, MTW (ZSM-12) has a one-dimensional, non-interconnecting channel structure, but it has a pore size index of 32. 5 (i. e., outside the range of 23 to 25). MEL (ZSM-11) is not suitable for use because it has a three- dimensional, interconnecting channel structure. Similarly, MFI (ZSM-5), another catalyst with a three-dimensional channel system, would not be suitable. Ferrierite (FER) (pore diameter < 4.4 Å), chabazite (CHA), Linde Type A (LTA) (both with three-dimensional channels), and many other common zeolites are also unsuitable.

The zeolites used in the process of the invention are usually powders.

To facilitate their use in fixed-bed reactors, the zeolites are optionally combined with one or more binders. Suitable binders are well known in the art and include, for example, natural clays (e. g., montmorillonite, kaolin, bentonite), silicas, aluminas, and the like. Aluminas and silicas are preferred. When a binder is used, it is typically present in an amount within the range of about 0. 5 to about 40 wt. % based on the combined amounts of binder and zeolite catalyst. The binders can be used in any convenient form, including powders, slurries, gels, or the like. If desired, the catalyst powder and/or binder can be combined with water and mulled using commercial mullers such as the Lancaster Mix Muller to produce a catalyst- containing paste. In the alternative, the zeolite catalyst can be used alone in powder or pelletized form.

While zeolite catalysts are normally synthesized in the alkali metal form, it is generally preferred to use the hydrogen form of the zeolite in the

process of the invention. Zeolites are conveniently converted to the hydrogen form by ion exchange with ammonium halide or nitrate solution, followed by calcination. It has been found that this procedure prolongs catalyst lifetime. Moreover, the amount of aromatics generated from the process tends to decrease with time.

The original alkali metal can also be replaced by other suitable metal cations, such as other alkali metals, calcium, magnesium, or the like. These metals can regulate the effective pore size index of the zeolite. For example, converting a zeolite to the hydrogen form generally increases the effective pore size index relative to the alkali metal form of the catalyst, while substituting the alkali metal with a metal tends to decrease the effective pore size index. Whether or not the zeolite catalyst is modified by converting it to the H-form, or by metal substitution, the effective pore size index needs to be within the limits defined herein.

The alkali metal can be replaced by a transition metal if desired to improve selectivity or modify the product mixture. For example, we surprisingly found that replacing some or all of the alkali metal with nickel or manganese reduces the proportion of aromatic compounds in the product mixture when olefins are converted to propylene and ethylene.

The catalysts can be prepared for use by any number of methods, which are now well known in the art. For example, the patent literature provides synthetic methods for MTT (U. S. Pat. Nos. 4, 076, 842 and 4, 490, 342) and TON (U. S. Pat. Nos. 4, 556, 477 and 5, 342, 596).

The zeolites can be used essentially"as is."Usually, however, the zeolites are calcined prior to use to remove traces of water, preferably by heating them at a temperature within the range of about 200°C to about 750°C, more preferably from about 300°C to about 650°C, and most preferably from about 450°C to about 600°C. When the zeolites are combined with a binder (e. g., alumina), calcination usually follows the pelletization process. Pellets are usually made by extrusion. If desired, one or more peptizing acids (e. g., nitric acid, citric acid, or acetic acid) are included with an extrusion aid (e. g., hydroxypropyl methylcellulose) to make

a plastic, extrudable material. Trace amounts of an oxidizing metal such as Pd or Pt can be used, if desired, to promote coke removal during catalyst regeneration (see, e. g., U. S. Pat. No. 5, 648, 585).

The zeolites are preferably passivated prior to use ; steam treatment is one way to passivate the zeolites. We found that steam treatment significantly prolongs catalyst lifetime. Prolonged catalyst lifetime is important for making the catalyst valuable for a fixed-bed process.

Steaming is normally done after preparation, pelletization, and drying of the catalyst. In one convenient procedure, steaming is performed just prior to using the catalyst for hydrocarbon conversion. Water is slowly fed to the reactor by any convenient means while the reactor is at a temperature greater than 100°C, usually from about 110°C to about 200°C, and the temperature is slowly elevated. Steaming is usually performed at elevated temperatures of about 450°C to about 700°C for a time ranging from several hours to several days.

In practicing the process of the invention, the mixture of C4 to C9 olefins is contacted with the zeolite catalyst under conditions effective to produce propylene and ethylene. Preferably, the process is performed in the vapor phase by bringing a heated olefin mixture into contact with the zeolite catalyst. Either the catalyst or the olefin mixture (or both) can be heated. Preferably, the reaction is performed at a temperature within the range of about 400°C to about 750°C, more preferably from about 450°C to about 700°C, and most preferably from about 500°C to about 600°C. While the reactor pressure is not usually critical, it is preferred to perform the process at a total reactor pressure within the range of about 0. 5 to about 10 atmospheres, more preferably from about 1 to 4 atmospheres. Any suitable feed rate can be used. Generally, it is preferred to use a hydrocarbon weight hourly space velocity (WHSV) within the range of about 0. 5 to about 1000 h-1, more preferably from about 1 to 50 h-1. If desired, hydrogen or a diluent can be added to the feedstock.

The process of the invention can be practiced in a batch, continuous, semi-batch, or semi-continuous manner. A continuous process is preferred.

If necessary, the catalyst can be regenerated using conventional techniques such as treatment with air diluted with an inert gas such as nitrogen. The process can be used with any desired kind of reactor system, including, for example, a fixed-bed, moving-bed, or fluidized-bed reactor system. The catalysts, when pelletized or combined with a binder and extruded, are particularly useful in a fixed-bed reactor system.

Long catalyst life is the key to a process that will be useful with a fixed-bed reactor. As Table 1 shows, the process of the invention is characterized by long catalyst life. Even after 300 hours of continuous operation, the MTT zeolite gives high olefin conversions and excellent selectivities to propylene and ethylene. In fact, the propylene selectivity actually improves throughout the experiment. At the 300-hour mark, propylene and ethylene make up a combined 92. 4 wt. % of the C1-C3 products. Moreover, because the C5 olefins and C6-C8 nonaromatics that comprise up to 80% or more of the heavy (> C4) components can be recycled, yields of ethylene and propylene can be boosted even more.

The following examples merely illustrate the invention ; those skilled in the art will recognize many variations that are within the spirit of the invention and scope of the claims.

General Experimental Conditions Apparatus The laboratory pilot unit is a semi-automated unit that can control temperature and pressure. The reactor is located within a convection oven to control temperature in the reactor. Hydrocarbon flow is controlled by a flow meter and flow control valve. The pressure of the system is controlled by the vent header pressure at 0-3 psig. Process variable data is collected and stored using software on a Local Area Network. Reactor effluent compositions are measured periodically by an on-line gas chromatograph (GC).

The reactor is constructed from 1/2"stainless steel tubing and fittings.

The reactor consists of two zones : a preheat section and a reaction section.

The preheat section is 1-2"long and is packed with 8-14 mesh tabular corundum. A layer of glass or quartz wool separates the preheat section from the reaction section. A"tee"fitting is situated at the start of the reaction zone. This fitting allows the radial insertion of a thermocouple to measure fluid temperature as it enters the reaction zone. The reaction section is 3-5"long and contains about 2-4 grams of catalyst. The catalyst is a pressed powder that is broken into pieces and sieved. The pieces that are retained between 12 and 20 mesh screens are used in the reactor. A plug of glass or quartz wool is placed on top of the catalyst bed and the remainder of the reactor is filled with tabular corundum to secure the catalyst in its position and prevent any fluidization during normal operations.

Catalyst Conditioning After the reactor is mounted within the convection oven, the catalyst can be subjected to"conditioning"steps, namely drying and/or steaming. To dry the catalyst, a stream of nitrogen is passed over the bed, while the oven temperature is slowly ramped from ambient temperature to 104°C and held there for at least one hour. Steaming is done after drying. Water is fed to the reactor via a high pressure syringe pump at a rate less than 1 mL/h while the reactor is at a temperature greater than 100°C. The oven temperature is increased to obtain the desired steaming temperature (as measured by the thermocouple situated at the beginning of the catalyst bed). Steaming lasts from 4 to 60 hours.

Hydrocarbon Conversion To start the reaction part of the experiment, the reactor temperature is adjusted to run conditions (in the range of 450-750°C) from the steaming conditions while water is flowing. Once reaction temperature is reached, the water flow is terminated and hydrocarbon feed (a mixture of butenes and butanes) is started. Nitrogen head pressure on a liquid reservoir of the feed material is used to drive the feed through the reactor. Once the feed has reached the tubing within the oven, it is vaporized. This vaporized feed

passes over the catalyst bed and is reacted. A portion of the vaporous product mixture is directed to a GC sampling valve through a heated transfer line. Periodically, this stream is analyzed for composition by the GC.

EXAMPLE 1 Preparation of Propylene and Ethylene from Butenes using MTT Catalyst Preparation A sample of MTT zeolite (20. 47 g of Na form, 67/1 Si02/AI203) is mixed with 200 mL of a solution of ammonium nitrate (2. 23 N, 18. 1 eq. of ammonium per eq. of sodium) and 200 mL of deionized water, and the mixture is stirred for 2 h at reflux. The catalyst is filtered to remove the liquid. This exchange process is repeated four more times. After the fifth exchange, the zeolite is washed with 400 mL deionized water and filtered.

This washing process is repeated four times.

The washed zeolite is transferred to a porcelain crucible, which is placed in an oven at 100°C for 14 h. The oven temperature is increased at 100°C per hour to 550°C, and is held at that temperature for an additional 8 hours. The zeolite is removed from the oven and allowed to cool in a dessicator.

A portion of the zeolite is pressed into wafers. The wafers are broken into pieces with a mortar and pestle. The pieces are sieved and the portion that is retained between 12 and 20 mesh (U. S. Standard) is used to load the reactor.

Catalyst Conditioning The reactor is loaded (from bottom to top) with 1/2"pad of quartz wool, 1'/2"of tabular corundum,'/2"pad of quartz wool, 2. 74 g of MTT zeolite pieces,'/2"pad of quartz wool, approximately 12"of tabular corundum and 1/2"pad of quartz wool. A thermocouple is inserted axially into the quartz wool pad below the catalyst bed.

The reactor is placed in the oven and connected to the feed and effluent sections of the pilot unit. Nitrogen is swept through the reactor as the catalyst is dried by raising the temperature and holding as follows : Temperature, °C Duration, h 82 1 91 1 104 1 After drying, the zeolite is steamed. With nitrogen flowing and the reactor at 104°C, water is injected into the system at a rate of 0. 25 mL/h.

The temperature of the reactor is raised to 600°C and is held there for 24 h.

The reactor is then cooled to approximately 510°C.

Conversion of Butenes to Propylene and Ethylene Nitrogen and water feeds are discontinued and the hydrocarbon feed is started at approximately 7 g/h. The composition of the feed is given below : Component Wt % isobutane 3. 7 n-butene 38. 2 n-butane 24. 6 trans-2-butene 19. 6 cis-2-butene 13. 5 The reactor effluent is analyzed periodically (every 2-3 h) by on-line GC. After 24 h, the reactor temperature is raised to approximately 525°C.

Typical results, demonstrating this catalyst's ability to crack olefinic feeds to propylene and ethylene in a fixed bed reactor, are shown in Table 1.

The preceding examples are meant only as illustrations ; the following claims define the scope of the invention.

Table 1.

Conversion of Butenes to Propylene and Ethylene with MTT Zeolite Time on stream (h) 52 98 152 202 300 WHSV (h-1) 2. 6 2. 5 2. 6 2. 5 2. 5 Temperature (°C) 526 526 525 526 526 Butene conversion (wt. %) 76. 4 74. 8 71. 0 70. 5 68. 9 Lights (< C4) yield (%) 38. 338. 735. 835. 735. 1 Lights composition (wt. %) methane 1. 7 1. 4 1. 1 1. 3 1. 2 ethylene 20.0 19.3 15.7 15.6 14.2 ethane 1.5 1.1 0.7 0.7 0.7 propylene 63. 8 68. 3 75. 7 76. 1 78. 2 propane 13. 0 10. 0 6. 7 6. 2 5. 6 Butanes (wt. %) 6. 8 5.7 5.4 5.3 5.2 Heavies (> C4) yield (%) 31.4 30.6 30.0 29.7 28. 8 Heavies composition (wt. %) C5 olefins 40. 5 42. 6 49. 7 50. 5 53. 9 C6-C8 nonaromatics 29. 6 30. 9 31. 2 29. 7 28. 3 benzene 2. 1 2. 2 2. 4 2. 3 2. 5 toluene 9. 0 8. 0 5. 6 6. 5 5. 7 C8 aromatics 10. 9 10. 2 7. 0 7. 3 6. 7 C9 to boiling pt. = 400oF 6.6 5.5 3.9 3.4 2.9 Boiling point > 400oF 1.2 0.5 0.2 0.3 0.1