PROPYLENE

The material 1,2,3-trichloropropane is produced as a byproduct of the epichlorohydrin process. Although some limited commercial use is made of this material, a substantial amount must be incinerated or otherwise disposed of. It is therefore a desirable end to be able to convert 1 ,2,3-trichloropropane on a substantial scale to useful, less chlorinated and unsaturated materials such as allyl chloride and propylene.

Although not directly on point, several published references discuss the catalytic conversion of 1 ,2-dichloropropane (or PDC) to reaction products including propylene.

In German Patent Publication No. 235,630 A1 (DE '630), for example, PDC is converted to propylene in a catalytic gas phase reaction at temperatures ranging from 170 degrees Celsius to 450 degrees Celsius. The catalyst is described as an activated carbon which has been treated with a suspension of iron oxides and/or iron oxide hydrates, and then dried at temperatures in the range of 80 degrees to 2O0 degrees Celsius.

Other methods described in DE '630 include the rhodium-catalyzed conversion at 180-250 degrees Celsius of PDC to propylene , and the dechlori nation at normal temperatures of PDC to a mixture (9: 1 ) of propylene and chloropropylene in the presence of a pure titanium catalyst. The present invention provides for the catalytic conversion of 1 ,2,3-

-trichloropropane to reaction products including allyl chloride or propylene in a commercially substantial proportion (that is, at a yield (defined as the product of the conversion of 1 ,2,3- -trichloropropane and the selectivity to a desired product, on a hydrogen chloride- and hydrogen-free basis) of at least 10 percent, but preferably at least 20 percent and more preferably at least 30 percent), characterized in that 1,2,3-trichloropropane is reacted with hydrogen in the presence of a catalyst including a selected Group IB metal or metals in an elemental or compound form with a selected Group VIII metal or metals, also in an elemental or compound form.

Preferred catalysts will consist essentially of a combination of one or more Group IB metals in elemental or compound form with one or more Group VIII metals in elemental or compound form on a support. More preferably, the catalysts employed in the processes of the present invention will consist of one or more Group IB metals with one or more Group VIII metals on a support.

Preferred Group VIII and Group IB metals include platinum and copper, respectively. More preferably, the Group IB and Group VIII metals will consist substantially entirely of copper and platinum in their elemental or compound forms, and a most preferred catalyst will employ only copper and platinum in their elemental or compound forms as the Group IB and Group VIII metals.

In general terms, copper can be from 0.01 to 20 percent by weight (on an elemental basis) of the catalyst, with platinum comprising from 0.01 to 5.0 percent by weight (also on an elemental basis) of the catalyst. More preferably, copper will be from 0.05 to 15 percent by weight of the catalyst (on an elemental basis) and platinum will be from 0.03 to 3.0 percent by weight of the catalyst. Most preferably, the copper can be from 0.1 to 10 percent by weight of the catalyst (on an elemental basis) and platinum will be from 0.05 to 1.0 percent by weight of the catalyst.

The support can be any of those supports conventionally employed in the art, but is preferably silica or carbon, with carbon being more preferred. A high surface area carbon o support is especially preferred, for example, a carbon having a specific surface area in an unimpregnated condition of 200 m /g or more, especially 400 m /g or more, and most especially 600 m ² /g or more. An example of a commercially-available carbon which has been found to be useful in the present invention is a coal-based carbon produced by Calgon Carbon Corporation under the designation "BPLF3", and may generally be characterized as having a 5 specific surface area of 1 100 m ² /g to 1300 m /g, a pore volume of 0.7 to 0.85 cm3/g, and an average pore radius of 12.3 to 14 angstroms. Based on an X-ray fluorescence analysis of this carbon, a typical bulk composition of the BPLF3 carbon has been determined to be as follows (by weight percent): silicon, 1.5 percent; aluminum, 1.4 percent; sulfur, 0.75 percent; iron, 0.48 percent; calcium, 0.17 percent; potassium, 0.086 percent; titanium, 0.059 percent; magnesium, 0 0.051 percent; chlorine, 0.028 percent; phosphorus, 0.026 percent; vanadium, 0.010 percent; nickel, 0.0036 percent; copper, 0.0035 percent; chromium, 0.0028 percent; and manganese, 0.0018 percent (the remainder being carbon).

Preferably, however, either a coconut-based carbon such as produced by Calgon Carbon Corporation under the designation PCB (having a published or advertised specific 5 surface area of from 1150 m ² /g to 1250 m ² /g and a pore volume of 0.72 cubic centimeters (cc) per gram ) or a wood-based carbon such as produced by Calgon Carbon Corp. as WSIV Special carbon (with a published or advertised specific surface area of 1400 m ² /g, and a pore volume of 1.25 cc/g) is employed for the conversion of 1 ,2,3-trichloropropane to propylene because of the lower rate of catalyst deactivation observed with the use of these catalyst supports in this 0 process as compared to the aforementioned BPLF3 carbon.

It will further be preferred, with either the coal-based BPLF3 carbon support or the above-mentioned coconut- or wood-based carbons, to employ hydrogen chloride in the feed to further reduce coking and catalyst deactivation rates (as exemplified below). In this regard, the rate of conversion loss is preferably no more than about 0.03 percent per hour, and 5 especially is no more than about 0.01 percent per hour.

Preferably the selected catalyst will additionally have been pretreated by exposure to a chloride source, for example, hydrogen chloride, to improve initial selectivity to propylene over propane and to make the product stream immediately useful in an allyl chloride

process (wherein impurity propane can react with chlorine to produce 1-chloropropane, which because of the similarity of its boiling point to allyl chloride's boiling point is difficult to remove therefrom), for example, or to minimize venting in a propylene oxide process of any propane in the product stream fed to the propylene oxide process. In this regard, catalysts of the present invention which have been impregnated, dried and reduced under hydrogen have been observed to produce propane with decreasing selectivity and propylene with increasing selectivity overtime. The initial chloride source pretreatment is expected to substantially reduce, however, the amounts of venting or downstream processing involved in making use of the product stream in a propylene oxide process or allyl chloride process, respectively. o Catalysts that have been pretreated by exposure to a chloride source and not reduced, or which have been merely dried and started up (and which have been in essence treated with hydrogen chloride in situ on start up), have been observed to produce the least amount of propane initially, albeit with a substantial penalty in conversion. For this reason, it is considered that it will generally be preferable to both reduce and pretreat (by exposure to a 5 chloride source) a given catalyst, as opposed to reducing the catalyst solely, or not reducing the catalyst at all in favor of chloride-source pretreatment or treatment with HCI in situ on start up.

A variety of reaction conditions can be employed for conducting the inventive process, depending for example on whether the process is carried out in the gas phase or in the liquid phase (in either a batchwise or continuous mode). In general, however, in a gas phase 0 process reaction pressures can range from atmospheric up to 1500 psig (10.3 Pa (gauge)), with temperatures of from 100 deg. C. to 350 deg. C, residence times of from 0.25 seconds to 180 seconds, and hydrogen to 1 ,2,3-trichloropropane feed ratios ranging on a molar basis from 0.1 : 1 to 100: 1. More preferably, reaction pressures will range from 5 psig (0.03 Pa (gauge)) to 500 psig (3.4 Pa (gauge)), with temperatures of from 180 deg. C. to 300 deg. C, residence times 5 of from 0.5 seconds to 120 seconds, and hydrogen to 1,2,3-trichloropropane feed ratios of from 0.5: 1 to 20: 1. Most preferably, reaction pressures will range from 40 psig (0.28 Pa (gauge)) to 300 psig (2.1 Pa (gauge)), with temperatures of from 200 deg. C. to 260 deg. C, residence times of from 1 second to 90 seconds, and hydrogen to 1 ,2,3-trichloropropane molar feed ratios of from 0.75: 1 to 6: 1. In a liquid phase process, it is anticipated that reaction pressures can range 0 from atmospheric pressure up to 3000 psig (20.6 Pa (gauge)), with temperatures ranging from 25 degrees Celsius to 350 degrees Celsius, residence times of from one to 30 minutes and hydrogen to 1,2,3-trichloropropane molar feed ratios of from 0.1 : 1 to 100: 1.

It is presently believed that the mechanism involved in the process of the present invention involves the conversion of 1,2,3-trichloropropane to allyl chloride, which is in turn 5 converted to propylene. Given this information and further given the examples provided herein, those skilled in the art will be well able to select those conditions and process parameters which will provide the desired products in the desired proportions.

Illustrative Examples Examples 1 and 2

In Examples 1 and 2, 1,2,3-trichloropropane was reacted with hydrogen over a supported bimetallic platinum/copper catalyst of the present invention containing 0.25 percent by weight of platinum on an elemental basis and 0.50 percent by weight of platinum on an elemental basis.

In the preparation of this platinum-copper catalyst, an aqueous H ₂ tCl6 stock solution was prepared by dissolving H ₂ PtCl66H ₂ 0 (J. T. Baker, Inc.; Baker Analyzed Grade, 37.6 percent Pt) in deionized and distilled water. An amount of CuCI ₂ (Aldrich Chemical Company, o Inc., 99.999 percent purity) was placed in a 250 mL Erlenmeyer flask, and a proportionate amount of the H ₂ f ^> tCl6 stock solution was added with swirling to dissolve the CuCI ₂ . The solution was then diluted with deionized, distilled water and swirled. Calgon BPLF3 activated carbon (6 x 16 mesh, Calgon Carbon Corp., Pittsburgh, Pa.) was added to the flask, and the flask was agitated rapidly so that the carbon carrier was evenly coated with the aqueous Pt Cu 5 solution. The catalyst preparation was dried in an evaporating dish in air at ambient temperatures for 18 hours, and then further air-dried in an oven at 120 degrees Celsius for 2 hours before being charged to the reactor, dried under a nitrogen purge and reduced in a manner described hereafter.

The 1,2,3-trichloropropane was pumped via a high pressure syringe pump 0 through 1/16 inch (1.6 mm) (O.D.) Monel'" nickel alloy tubing (unless specifically noted below all of the components, tubing and fittings of the test reactor apparatus were also made of Monel ^™ nickel alloy (Huntington Alloys, Inco Alloys International, Inc.)) into a packed sample cylinder serving as a feed evaporator.

The 1/16th inch tubing extended almost to the center of the packed cylinder, 5 which was heated to a vaporizing temperature of 180 degrees Celsius using electrical heat tracing. Vaporization of the 1,2,3-trichloropropane feedstock was accomplished in the feed line, so that the 1,2,3-trichloropropane was superheated when combined with the hydrogen feed stream. Thermocouples were used to monitor the skin temperature of the feed evaporator and the temperature of the gas exiting the feed evaporator. 0 The hydrogen feed stream was metered to a preheater using a Model 8249 linear mass flow controller from Matheson Gas Products, Inc. Secaucus, N.J., with the preheater consisting of a packed sample cylinder wrapped with electrical heat tracing. Thermocouples were used to monitor both the skin temperature of the preheater and the temperature of the gas exiting the preheater. The preheater temperature was set and maintained at 140 degrees 5 Celsius.

Vaporized 1,2,3-trichloropropane exiting the evaporator was mixed with the hydrogen gas from the preheater in a 2 foot (0.61 meter) long section of 1/4 inch (0.64 cm) tubing maintained at a temperature of 140 degrees Celsius. The mixed gases then were passed

into and reacted within a tubular reactor (1/2 inch (1.27 cm) O.D., 12 inches (30.5 cm) in length) located within an aluminum block equipped with a cartridge heater for achieving a desired reaction temperature with computer control.

The catalyst charge (0.6 grams in each of Examples 1 and 2) was generally placed in the tubular reactor over a glass wool support contained in the center of the reactor tubing. The catalyst was then covered with a plug of glass wool.

The catalyst was dried for from 8 to 24 hours at 150 degrees Celsius under a nitrogen purge, and thereafter reduced by passing hydrogen through the reactor at a flow rate of 34 mL/minute for 24 hours. The reactor temperature was then lowered to the temperature setpoint of the particular catalyst run. The reactor temperature and hydrogen gas flow were allowed to equilibrate for about 1 hour before the liquid 1,2,3-trichloropropane flow was started into the apparatus.

After reacting the 1 ,2,3-trichloropropane and hydrogen in the vapor phase in the tubular reactor thus prepared, the products from the reaction were passed to a gas sampling valve, which provided gaseous aliquots for online gas chromatographic analysis in a Hewlett- Packard Model 5890 Series II gas chromatograph (Hewlett-Packard Company). The gas chromatograph was equipped with a flame ionization detector, and used 30 meter by 0.53 millimeter (I.D.) 100 percent methyl silicone/fused silica and 30 meter by 0.53 millimeter (I.D.) porous polymer-lined fused silica columns to separate the various reaction products. Response factors were conventionally determined by injections of gravimetrically-prepared standards of the individual reaction products. These response factors were applied in conjunction with individual peak areas and the total mols of all reaction products to determine the mol percents of individual components in the reactor effluent, and the selectivity to individual reaction products (mols of product divided by mols of 1 ,2,3-trichloropropane converted, multiplied by 100). Percentage conversion of the 1,2,3-trichloropropane was determined by subtracting the mol percent of 1,2,3-trichloropropane remaining in the test reactor effluent from 100, excluding from the calculation any unreacted hydrogen and the hydrochloric acid produced by the reaction.

The conditions and results of Examples 1 and 2 are summarized in Table 1 below, and demonstrate a yield of 76.4 percent of propylene for the 180 degree run ((98 x 78)/100) and a yield of 31 percent of allyl chloride in the second, highertemperature run ((50 x 62)/100):

TABLE 1

Residence Conversion

Catalyst Feed T(°C) H2/Feed Time(s) [%] Products & Selectivities (% ⁾ b

0.25Pt/0.50Cu//C 1,2 _f 3TCPaa 180 16.9 8.6 98 78% C ₃ H ₆ , 10% C ₃ H ₈

220 6.6 5.7 50 62% allyl chloride, 20% C ₃ H ₆

a = 1,2,3TCPa = 1,2,3-trichloropropane σx b = Balance to 100% is various hydrocarbons

Examples 3-7

The experimental method and apparatus of Examples 1 and 2 above were employed (except as indicated below) in these Examples for determining the effects in slowing the rate of catalyst deactivation in converting 1,2,3-trichloropropane to propylene of adding hydrogen chloride (HCI) to a 1 ,2,3-trichloropropane feed , and of omitting HCI from the feed but using a different wood- or coconut-based carbon support.

Bimetallic platinum/copper catalysts were prepared for these examples in the ratios indicated in Table 2 below, by dissolving H PtCl66H ₂ 0 U- T. Baker, Inc.; Baker Analyzed Grade, 37.6 percent Pt) in deionized and distilled water. An amount of CuCI ₂ (Aldrich Chemical Company, Inc., 99.999 percent purity) was placed in a 250 mL Erlenmeyer flask, and the H ₂ PtCl6 stock solution was added with swirling to dissolve the CuCI ₂ . The solution was then diluted with deionized, distilled water and swirled. An activated carbon (Calgon BPLF3 carbon, 6 x 16 mesh, Calgon Carbon Corp., Pittsburgh, Pa. for the runs with HCI in the feed, Calgon PCB coconut-based activated carbon or Calgon WSIV Special wood-based activated carbon) was added to the flask, and the flask was agitated rapidly so that the carbon carrier was evenly coated with the aqueous Pt/Cu solution.

The various catalysts were charged to the reactor and dried under flowing nitrogen at 90 cubic centimeters per minute, at temperatures increasing from 25 degrees Celsius to 120 degrees Celsius at a rate of 3 degrees Celsius per minute with the 120 degree temperature being held for an hour. These catalysts were then reduced with flowing hydrogen at 90 cubic centimeters per minute, with the temperature being raised from 120 degrees Celsius to 220 degrees Celsius at 3 degrees per minute and the 220 degree temperature being held for 2 hours. The reaction temperature was set at the same 220 degrees Celsius, with a reaction pressure of 20 pounds per square inch, gauge (0.14 Pa (gauge)), a residence time of 3.5 seconds, a liquid hourly space velocity of 0.91 hr-i (based on the volume of liquid 1,2,3-trichloropropane fed per hour to the test reactor divided by the packed bed volume of the particular catalyst in the reactor) and a hydrogen to 1 ,2,3-trichloropropane (TCP) molar feed ratio of 6.0 to 1. Hydrogen chloride when added was metered in as a gas in a 3.0 to 1 ratio with respect to the 1 ,2,3-trichloropropane.

Selectivities to allyl chloride and to propylene, as well as percent conversion of the 1,2,3-trichloropropane, were assessed initially for all of the catalysts, and at a later period in time when the percent conversion reached 20 percent for the standard BPLF3 carbon-based catalyst as well as forthe runs of the BPLF3 carbon-based catalyst involving HCI in the feed. The results are as indicated in Table 2:

TABLE 2

Initial Percent (Pet.) 20 Percent Pet. Conversion

00

While various embodiments of the processes and catalysts of the present invention have been described and/or exemplified herein, those skilled in the art will readily appreciate that numerous changes can be made thereto which are nevertheless properly considered to be within the scope or spirit of the present invention as more particularly defined by the claims below.