Catalyst for Production of Hydrocarbons

The present invention relates to catalysts and more particularly to a catalyst for converting synthesis gas to hydrocarbons.

The reaction to convert carbon monoxide and hydrogen mixtures (defined herein as synthesis gas or syngas) to higher hydrocarbons over metallic catalysts has been known since the turn of the century. This reaction is commonly referred to as the Fischer-Tropsch or F-T synthesis. The F-T synthesis was exploited commercially during World War II in Germany. By 1944 a total of nine F-T plants were operating in Germany, primarily using a catalyst composed of cobalt, magnesium oxide, thorium oxide and kieselguhr, in the relative proportions of 100:5:8:200. Later, most of the thoria was replaced by magnesia, primarily for economic reasons. Currently, commercial Fischer-Tropsch plants are operating in South Africa. These plants use a precipitated iron-based catalyst which contains various promoters to improve the stability and product distribution.

The common F-T catalysts are nickel, cobalt and iron. Nickel was probably the first substance to be recognized as capable of catalyzing the reaction of syngas to hydrocarbons, producing mainly methane (see, for example, "The Fischer-Tropsch Synthesis" by R. B. Anderson, Academic Press (1984), p.2). Iron and cobalt are able to produce longer chain length hydrocarbons and are thus preferred as catalysts for the production of liquid hydrocarbons. However, other metals are also capable of catalyzing the F-T synthesis. Ruthenium is a very active catalyst for the formation of hydrocarbons from syngas. Its activity at low temperatures is higher than that of iron, cobalt or nickel; and it produces a high proportion of heavy hydrocarbons. At high pressures, it produces a high proportion of high molecular weight wax. Osmium has been found to be moderately active, while platinum, palladium and iridium exhibit low activities (see Pichler, "Advances in Catalysis", vol. IV, Academic Press, N.Y., 1952). Other metals which are active, such as rhodium, yield high percentages of oxygenated materials (Ichikawa, Chemtech, 6, 74 (1982)). Other metals that have been investigated include rhenium, molybdenum and chromium, but these exhibit very low activities with most of the product being methane.

Various combinations of metals can also be used for synthesis. Doping cobalt catalysts with nickel causes an increase in methane production during F-T synthesis (see "Catalysis", vol. IV, Reinhold Publishing Co., (1956), p. 29). In U.S. Pat. No. 4,088,671 to T. P. Kobylinski, entitled "Conversion of Synthesis Gas Using a Cobalt-Ruthenium Catalyst", the addition of small amounts of ruthenium to cobalt is shown to result in an active F-T synthesis catalyst with a low selectivity to methane. Thus, these references teach that the combination of two or more metals can result in an active F-T catalyst. In general, the catalysts of these teachings have activities and selectivities which are within the ranges of the individual components.

Combinations of metals with certain oxide supports have also been reported to result in an improved hydrocarbon yield during F-T synthesis, probably due to an increase in the surface area of the active metal. The use of titania to support cobalt or cobalt-thoria is taught in U.S. Pat. No. 4,595,703, entitled "Hydrocarbons from Synthesis Gas". In this case the support serves to increase the activity of the metal(s) toward hydrocarbon formation. In fact, titania belongs to a class of metal oxides known to exhibit strong metal-support interactions and, as such, has been reported to give improved F-T activity for a number of metals (see, for example, S. J. Tauster et al, Science, 211, 1121 (1981)). Combinations of titania and two or more metals have also been shown to yield improved F-T activity. In U.S. Pat. No. 4,568,663, entitled "Cobalt Catalysts. In the Conversion of Methanol to Hydrocarbons and for Fischer-Tropsch Synthesis", combinations of cobalt, rhenium and thoria and cobalt and rhenium supported on titania are claimed useful for the production of hydrocarbons from methanol or synthesis gas. This patent also indicates that similar improvements in activity can be obtained when cobalt-rhenium or cobalt-rhenium-thoria is compounded with other inorganic oxides. However, titania is the only support specifically discussed. The typical improvement in activity gained by promotion of cobalt metal supported on titania with rhenium is less than a factor of 2. We have found that the addition of rhenium to cobalt metal supported on a number of other common supports results in similar improvements in activity.

The only other examples in the literature of catalysts involving mixtures of cobalt and rhenium refer to completely different chemical reactions. For example, in Soviet Union Patent 610558, a catalyst composed of cobalt and rhenium supported on alumina is taught to result in improved performance for the steam reforming of hydrocarbons. Steam reforming of hydrocarbons is a process completely different from hydrocarbon production via F-T synthesis and is believed to proceed by a completely different mechanism. Although some steam reforming catalysts can convert synthesis gas to hydrocarbons, such catalysts are not selective for the production of high carbonnumber hydrocarbons (C3 and above) during conversion of synthesis gas. In fact, most commonly used steam reforming catalysts contain nickel as their active metal, and nickel produces mostly methane when used for syngas conversion.

It has been found in accordance with the present invention that synthesis gas comprising hydrogen and carbon monoxide can be converted to liquid hydrocarbons by using a catalyst consisting of cobalt and rhenium supported on alumina and promoted with an alkali. As used herein, alkali refers to one or more of the elements lithium, sodium, potassium, rubidium, and cesium of group IA of the periodic table. The catalyst preferably contains from about 5 to 60% cobalt, has a rhenium content between 0.5 and 50% of the amount of cobalt, and an alkali content between 0.5 and 5 atom % of the amount of cobalt. The alumina preferably is gamma alumina.

It has been found that the addition of small amounts of rhenium to catalysts consisting predominantly of cobalt supported on alumina unexpectedly results in greatly enhanced activity of this catalyst for hydrocarbon production from syngas. This is surprising in light of the fact that rhenium supported on alumina shows very low activity, with most of the product being methane. Furthermore, rhenium addition to cobalt supported on supports other than alumina results in catalysts with much lower activity levels. In addition, the more active cobalt plus rhenium catalyst maintains the high selectivity to higher hydrocarbons and the low selectivity to methane found with an aluminasupported cobalt catalyst. It has also been found that the addition of small amounts of an alkali to these catalysts serves to increase the average carbon number of the products produced during F-T synthesis. Both the high activity and the low methane production of cobalt-rhenium on alumina are unexpected in light of the facts that (1) rhenium shows very low activity for F-T synthesis, (2) the main products from F-T synthesis over a rhenium catalyst are methane and carbon dioxide, and (3) the use of alumina as a support for catalysts containing only cobalt results in no, or at best only a slight, increase in activity compared to the use of cobalt on other supports. Thus, for reasons not fully understood, the combination of cobalt and rhenium plus an alkali supported on alumina results in a catalyst which is significantly more active than either of the two individual metals supported on alumina or the combination of the two metals supported on other inorganic supports, such as silica, magnesia, silica-alumina, titania, chromia or zirconia. Furthermore, the product distribution with a high selectivity to C.sub.2 +hydrocarbons and low selectivity to methane and carbon dioxide would not have been predicted based on the known product distribution from rhenium catalysts. In addition, the addition of an alkali to the catalyst serves to increase the average carbon number of the product. This is advantageous in situations where the market value of the lighter products is lower than that of the heavier products.

FIG. 1 is a graph showing the effect of rhenium content on CO conversion using catalysts containing 12% cobalt;

FIG. 2 is a graph showing the effect on CO conversion of adding rhenium to catalysts containing various amounts of cobalt on an alumina support;

FIG. 3 is a graph showing the effect of potassium to cobalt ratio on the Schulz-Flory .alpha. of the product produced using the catalysts of this invention; and

FIG. 4 is graph showing the effect of potassium to cobalt ratio on CO conversion when using the catalysts of this invention.

The catalyst of the present invention comprises as the active catalytic ingredients cobalt, rhenium and an alkali, that is an element from Group IA of the Periodic Table, supported on alumina with rhenium and the alkali present in relatively smaller amounts than cobalt. This catalyst has been found to be highly active for the conversion of synthesis gas, a mixture of hydrogen and carbon monoxide, into a mixture of predominantly paraffinic hydrocarbons. As indicated above, it has long been known that cobalt is an active catalyst for the F-T synthesis. It is also known that the addition of rhenium to a cobalt catalyst supported on titania gives improved activity, even if rhenium by itself shows very low activity for F-T synthesis and produces methane as the main product. Surprisingly, we have found that the choice of support for the cobalt plus rhenium catalyst is very critical, and that the addition of rhenium to an alumina-supported cobalt catalyst gives a much higher improvement in activity than addition of rhenium to cobalt supported on other inorganic oxides. Also, the alkali increases the average carbon number of the product.

The cobalt is added to the alumina support in some amount up to about 60 wt % of the catalyst, including cobalt. Preferably, amounts between 5 and 45 wt % are used; and more preferably between about 10 to about 40 wt %. The content of rhenium is between about 0.5 and 50 wt % of the cobalt content; preferably between 1 and 30 wt %; and more preferably from about 2 to around 20 wt %. The content of alkali is between about 0.5 and 5 atom % of the cobalt content.

In addition to cobalt, rhenium and alkali, it may be beneficial to include a small amount of an additional metal oxide promoter in an amount between about 0.1 and 5 wt %, and more preferably between about 0.2 and 2 wt %, based on the weight of the complete catalyst. The promoter is suitably chosen from elements in groups IIIB, IVB and VB of the periodic chart, the lanthanides and actinides. The promoter oxide can be chosen from, for example, Sc.sub.2 O.sub.3, Y.sub.2 O.sub.3, La.sub.2 O.sub.3, Ce.sub.2 O.sub.3, Pr.sub.2 O.sub.3, ZrO.sub.2, Ac.sub.2 O.sub.3, PaO.sub.2, Nd.sub.2 O.sub.3, CeO.sub.2, V.sub.2 O or Nb.sub.2 O.sub.5. The most preferable oxide is La.sub.2 O.sub.3, or a mixture of lanthanides, rich in lanthanum. Oxides like MnO or MgO can also be included. While not essential, the use of these metal oxides is common in the art, since they are believed to promote the production of products with higher boiling points, while maintaining or improving catalytic activity. However, the catalyst is highly active and selective without the addition of one or more of these metal oxide promoters.

The catalytically active metals, the alkali, and the promoter metal oxide, if present, are distended on alumina. Although other supports may be used, it has been found, for example, that silica, titania, chromia, magnesia, silica-alumina and zirconia produce catalysts with much lower activities.

To be most effective when used as a support, alumina should be characterized by low acidity, high surface area, and high purity. These properties are necessary in order to enable the catalyst to have high activity and a low deactivation rate, and to produce high molecular weight hydrocarbon products. The surface area of the alumina support is at least, and preferably greater than, about 100 m.sup.2 /g; and more preferably at least about or greater than 150 m.sup.2 /g. The pore volume is at least, and preferably greater than, about than 0.3 cm.sup.3 /g. The catalyst support must be of high purity. That is, the content of elements, e.g. sulfur and phosphorous, that have a deleterious effect on catalytic activity must be kept low. The sulfur content of the catalyst support should be kept below 100 ppm and preferably below 50 ppm. Although gamma alumina has generally been used and is preferred, a number of alumina structures, if prepared properly, can meet these conditions and are suitable supports. For example, etaalumina, xi-alumina, theta-alumina, delta-alumina, kappa-alumina, boehmite and pseudo-boehmite can all be used as supports.

The method of depositing the active metals, the alkali and the promoter oxide on the alumina support is not critical, and can be chosen from various methods well known to those skilled in the art. One suitable method that has been employed is known as incipient wetness impregnation. In this method the metal salts are dissolved in an amount of a suitable solvent just sufficient to fill the pores of the catalyst. In another method, the metal oxides or hydroxides are coprecipitated from an aqueous solution by adding a precipitating agent. In still another method, the metal salts are mixed with the wet support in a suitable blender to obtain a substantially homogeneous mixture. In the present invention, if incipient wetness impregnation is used, the catalytically active metals and the alkali can be deposited on the support using an aqueous or an organic solution. Suitable organic solvents include, for example, acetone, methanol, ethanol, dimethyl formamide, diethyl ether, cyclohexane, xylene and tetrahydrofuran. Aqueous impregnation is preferred when Co(No.sub.3).sub.2 is used as the salt, while an organic solvent is the preferred solvent when the catalyst is prepared from cobalt carbonyl.

Suitable cobalt compounds include, for example, cobalt nitrate, cobalt acetate, cobalt chloride and cobalt carbonyl, with the nitrate being the most preferable when impregnating from an aqueous solution. Suitable rhenium compounds include, for example, rhenium oxide, rhenium chloride and perrhenic acid. Perrhenic acid is the preferred compound when preparing a catalyst using an aqueous solution. Suitable alkali salts for incorporating the alkali into the catalyst include, for example, the nitrates, chlorides, carbonates, and hydroxides. The metal oxide promoter can suitably be incorporated into the catalyst in the form, for example, of the nitrate or chloride.

After aqueous impregnation, the catalyst is dried at 110.degree. to 120.degree. C. for 3 to 6 hours. When impregnating from organic solvents, the catalyst is preferably first dried in a rotary evaporator apparatus at 50.degree. to 60.degree. C. under low pressure, then dried at 110.degree. to 120.degree. C. for several hours longer.

The dried catalyst is calcined under flowing air by slowly increasing the temperature to an upper limit of between 200.degree. and 500.degree. C., preferably between 250.degree. and 350.degree. C. The rate of temperature increase is preferably between 0.5.degree. and 2.degree. C. per minute, and the catalyst is held at the highest temperature for a period of 2 to 5 hours. The impregnation procedure is repeated as many times as necessary to obtain a catalyst with the desired metals content. Cobalt, rhenium, alkali and the metal oxide promoter, if present, can be impregnated together, or in separate steps. If separate steps are used, the order of impregnating the active components can be varied.

Before use, the calcined catalyst is preferably reduced with hydrogen. This can suitably be done in flowing hydrogen at atmospheric pressure at a flow rate between 30 and 100 cm.sup.3 /min when reducing about 2 g of catalyst. The flow rate should suitably be increased for larger quantities of catalyst. The temperature is increased at a rate between 0.5.degree. and 2.degree. C. per minute from ambient to a maximum level of 250.degree. to 450.degree. C., preferably between 300.degree. and 400.degree. C., and maintained at the maximum temperature for about 6 to 24 hours, more preferably 10 to 24 hours.

After the reduction step, the catalyst may be oxidized and rereduced before use. To carry out the oxidation step, the catalyst is treated with dilute oxygen (1-3% oxygen in nitrogen) at room temperature for a period of 1/2 to 2 hours before the temperature is increased at the same rate and to the same temperature as used during calcination. After holding the high temperature for 1 to 2 hours, air is slowly introduced, and the treatment is continued under air at the high temperature for another 2 to 4 hours. The second reduction is carried out under the same conditions as the first reduction.

The reactor used for the synthesis of hydrocarbons from synthesis gas can be chosen from various types well known to those skilled in the art, for example, fixed bed, fluidized bed, ebullating bed or slurry. The catalyst particle size for the fixed or ebullating bed is preferably between 0.1 and 10 mm and more preferably between 0.5 and 5 mm. For the other types of operations a particle size between 0.01 and 0.2 mm is preferred.

The synthesis gas is a mixture of carbon monoxide and hydrogen and can be obtained from any source known to those skilled in the art, such as, for example, steam reforming of natural gas or partial oxidation of coal. The molar ratio of H.sub.2: CO is preferably between 1:1 to 3:1; and more preferably between 1.5:1 to 2.5:1. Carbon dioxide is not a desired feed component for use with the catalyst of this invention, but it does not adversely affect the activity of the catalyst. All sulfur compounds must, on the other hand, be held to very low levels in the feed, preferably below 1 ppm.

The reaction temperature is suitably between 150.degree. and 300.degree. C.; and more preferably between 175.degree. and 250.degree. C. The total pressure can be from atmospheric to around 100 atmospheres, preferably between 1 and 30 atmospheres. The gaseous hourly space velocity, based on the total amount of synthesis gas feed, is preferably between 100 and 20,000 cm.sup.3 of gas per gram of catalyst per hour; and more preferably from 1000 to 10,000 cm.sup.3 /g/h, where gaseous hourly space velocity is defined as the volume of synthesis gas (measured at standard temperature and pressure) fed per unit weight of catalyst per hour.

The reaction products are a complicated mixture, but the main reaction can be illustrated by the following equation:

where (--CH.sub.2 --).sub.n represents a straightchain hydrocarbon of carbon number n. Carbon number refers to the number of carbon atoms making up the main skeleton of the molecule. In F-T synthesis, the products are generally either paraffins, olefins, or alcohols. Products range in carbon number from one to 50 or higher.

In addition, with many catalysts, for example those based on iron, the water gas shift reaction is a well known side reaction:

With cobalt catalysts the rate of this last reaction is usually very low. However, it is found that, even though rhenium catalysts exhibit a relatively high selectivity to carbon dioxide, the cobalt plus rhenium catalyst of this invention surprisingly does not have a higher selectivity to carbon dioxide than the cobalt only catalyst.

The hydrocarbon products from Fischer-Tropsch synthesis are distributed from methane to high boiling compounds according to the so called Schulz-Flory distribution, well known to those skilled in the art. The Schulz-Flory distribution is expressed mathematically by the Schulz-Flory equation:

where i represents carbon number, .alpha. is the Schulz-Flory distribution factor which represents the ratio of the rate of chain propagation to the rate of chain propagation plus the rate of chain termination, and W.sub.i represents the weight fraction of product of carbon number i. This equation shows that an increased .alpha. results in a higher average carbon number of the products. Higher .alpha. values are desirable when heavier products, such as diesel fuel, are relatively more valuable than lighter products, such as naphtha.

The products produced by the catalyst of this invention generally follow the Schulz-Flory distribution, except that the yield of methane is usually higher than expected from this distribution. This indicates that methane is apparently produced by an additional mechanism.

Catalysts promoted with alkali in the manner prescribed in this invention produce a product having a higher average carbon number than the products from non-alkali promoted catalysts. That is, the Schulz-Flory .alpha. for the product from an alkali promoted catalyst is higher than for a non-alkali promoted catalyst.

It is well known, and also shown in one of the following examples, that rhenium alone is a low activity catalyst for Fischer-Tropsch synthesis, producing a product which is predominantly methane, and that the alkalis have no activity for F-T synthesis. On the other hand, cobalt is a well known catalyst for producing higher carbon number hydrocarbons. In U.S. Pat. No. 4,568,663, it has been shown that adding small amounts of rhenium to cobalt supported on titania improves the catalytic activity. In the present invention, it has been found that the hydrocarbon yield obtained by adding rhenium is surprisingly much larger for an alumina supported cobalt catalyst than that obtained from cobalt and rhenium on several other inorganic supports. The improved activity and increased selectivity to heavier hydrocarbons is followed by no deleterious effect on the selectivity to methane.

The catalyst of this invention is further described in the following examples.

The following examples describe the preparation of various catalysts and the results obtained from testing these catalysts for conversion of synthesis gas into hydrocarbons.

Before being tested, each catalyst was given a pretreatment consisting of reduction by passing hydrogen over the catalyst at a rate of 3000 cm.sup.3 /g/h while heating the catalyst at a rate of 1.degree. C./min to 350.degree. C. and maintaining this temperature for 10 hours. In the tests, synthesis gas consisting of 33 vol % carbon monoxide and 67 vol % hydrogen was passed over 0.5 g of the catalyst at atmospheric pressure at temperatures of 185.degree., 195.degree. and 205.degree. C. according to the following schedule:

9 hr. 50 min. at 195.degree. C.

4 hr. 20 min. at 205.degree. C.

4 hr. 30 min. at 185.degree. C.

9 hr. 50 min. at 195.degree. C.

The flow rate of synthesis gas was 1680 cm.sup.3 /g of catalyst/h. Products from the reactor were sent to a gas chromatograph for analysis. Catalysts were compared based on the results over the period from 10 to 30 hours on stream.

This example describes the preparation of a control cobalt catalyst which was used for comparative purposes. This catalyst was prepared as follows:

A solution was prepared by dissolving 17.03 g of cobalt nitrate, Co(NO.sub.3).sub.2.6H.sub.2 O, and 0.76 g of mixed rare earth nitrate, RE(NO.sub.3).sub.3, where RE stands for rare earth with a composition of 66% La.sub.2 O.sub.3, 24% Nd.sub.2 O.sub.3, 8.2% Pr.sub.6 O.sub.11, 0.7% CeO.sub.2, and 1.1% other oxides (Molycorp 5247), in 30 ml of distilled water. The total solution was added with stirring to 25 g of Ketjen CK300 gammaalumina which had been calcined 10 hours at 500.degree. C. The prepared catalyst was then dried for 5 hours in an oven at a temperature of 115.degree. C. The dried catalyst was then calcined in air by raising its temperature at a heating rate of 1.degree. C./minute to 300.degree. C. and holding at this temperature for 2 hours. The finished catalyst contained 12 wt % cobalt and 1 wt % rare earth oxide with the remainder being alumina. This catalyst is referred to as preparation "a" in Table I. The above procedure was repeated to produce preparation "b" catalyst in Table I.

The results of the tests with this catalyst are shown in Table I. In this and the following tables, selectivity is defined as the percent of the carbon monoxide converted that goes to the indicated product.

This example shows that a cobalt catalyst exhibits good selectivity to ethane and longer chain length hydrocarbons and low selectivity to methane and carbon dioxide.

This example describes a rhenium catalyst prepared for comparative purposes. The procedure employed was the same as for Example 1 except that the solution contained 0.33 g of perrhenic acid, HReO.sub.4 as 82.5% aqueous solution, and 0.54 g of rare earth nitrate to make 24 ml of solution which then was added to 20 g of calcined alumina. The finished catalyst contained 1 wt % rhenium and 1 wt % rare earth oxide with the remainder being alumina.

The results of the tests with the catalyst of Example 2 are shown in Table II.

Repetition of the procedure from Example 2, except that 0.83 g of perrhenic acid were used, gave a catalyst containing 4 wt % rhenium. The results of the tests with the catalyst of Example 3 are shown in Table III.

The results from Examples 2 and 3 show that catalysts containing rhenium but no cobalt have very low activity for producing desirable liquid hydrocarbons from synthesis gas. Furthermore, about half the product is carbon dioxide, and most of the hydrocarbon product is methane.

The preparation procedure of Example 1 was employed except that varying amounts of perrhenic acid were added to the solution. This produced a series of catalysts containing 12 wt % cobalt and 0.1, 0.2, 0.3, 0.5, 1.0, 2.0, 4.0, and 8.0 wt % rhenium in addition to 1.0 wt % rare earth oxide.

The results of the tests with the catalysts of Examples 4 through 11 at 195.degree. C. are shown in Table IV and further illustrated in FIG. 1. FIG. 1 shows the effect on carbon monoxide conversion of adding rhenium to catalysts containing 12% cobalt.

As can be seen from comparison of the results in Table I with Table IV and FIG. 1, the addition of small amounts of rhenium to a cobalt supported on alumina catalyst significantly increases the conversion of the carbon monoxide in the feed. Levels of rhenium as low as 0.1 wt % result in approximately doubling the CO conversion. The exact level of Re for optimum activity is very important, as the rate of carbon monoxide conversion increases rapidly at low rhenium addition levels, reaches a maximum and then decreases gradually at levels greater than 1 wt % rhenium. However, even at the highest rhenium level investigated (8%), a clear improvement in conversion is evident when compared to the catalyst not containing rhenium.

It is important that the increase in activity occur without a corresponding increase in either the methane or the carbon dioxide selectivities. Table IV shows that the increase in carbon monoxide conversion is not accompanied by any substantial change in either the selectivities to methane or carbon dioxide. Thus, after rhenium addition the principal reaction products are still desirable hydrocarbons.

The preparation procedure of Example 1 was employed except that varying amounts of cobalt nitrate and perrhenic acid were added to the solution. This produced a series of catalysts containing from 3.0 to 40 wt % cobalt and from 0 to 5.0 wt % rhenium in addition to 1.0 wt % rare earth oxide.

The results of the tests with the catalysts of Examples 12 through 25 at 195.degree. C. are shown in Table V.

The results in Table V show that for cobalt catalysts without rhenium, there is a significant increase in activity in going from 3% cobalt to 6% cobalt. However, only modest increases in activity occur from this point up to cobalt loadings of as high as 40%. At a cobalt loading of 3%, the addition of rhenium does not improve the catalytic activity, but the improvment upon rhenium addition is significant for higher cobalt loadings. In fact, the improvement in activity due to the addition of rhenium increases as the cobalt content increases as shown in FIG. 2.

To illustrate the use of promoters other than rare earth oxides, the following catalysts were prepared. The preparation procedure used to prepare the catalyst of Example 8 was used except that zirconium nitrate, Zr(NO.sub.3).sub.4, or vanadyl oxalate, VO(C.sub.2 O.sub.4 H).sub.3, was substituted for the rare earth nitrate. The results of tests at 195.degree. C. with the catalysts of Examples 26 and 27 are shown in Table VI. In addition to the promoter, these catalysts contained 12% cobalt and 1% rhenium and were supported on alumina.

For comparison with alumina, several catalysts were prepared on other supports. The preparation procedure used to prepare the catalyst of Example 8 was repeated, but without the addition of rare earth oxide. The titanium-supported catalysts were prepared on titania calcined at both 500.degree. C. and 600.degree. C. After calcination at 600.degree. C., the titania is mainly in the crystalline rutile form; while after calcination at 500.degree. C. the anatase:rutile ratio is about 1:1. The catalysts prepared on the titania support calcined at these two temperatures showed exactly the same catalytic activity.

The supports used were: Davison Grade 59 silica; Degussa P25 titania; Alpha Chemicals No. 88272 chromia; magnesia prepared by calcining Fischer basic magnesium carbonate; American Cyanamid AAA Silica-Alumina; and Alpha Chemicals 11852 zirconia (containing 2% alumina). Information on the composition of the catalysts prepared on the different supports is given in Table VII.

A series of tests was conducted to evaluate the activities of the catalysts of the above examples in converting synthesis gas into hydrocarbons. The results of the tests with the catalysts of Examples 28 through 41 at 195.degree. C. are shown in Table VIII. The results from catalysts prepared on alumina are included for comparison.

The catalysts in Table VII were prepared to test the teaching that various inorganic supports are acceptable for preparing cobalt plus rhenium F-T catalysts. An examination of the data in Table VIII leads to the surprising conclusion that the type of support is extremely important and that vast differences in activity exist between catalysts prepared on one support and catalysts of the same catalytic metals content on another support. More surprisingly, only cobalt plus rhenium on alumina showed a commercially attractive activity level and selectivity.

Catalysts on magnesia and chromia exhibited extremely low activities, both with and without rhenium. Catalysts on zirconia and silica-alumina showed somewhat higher activities, but selectivity to C.sub.2 + hydrocarbons was poor. These catalysts showed only modest improvements in activity upon the addition of rhenium.

Catalysts without rhenium supported on silica and titania showed activity levels close to comparable cobalt on alumina catalyst. However, upon addition of rhenium, the alumina catalyst showed a surprising increase in activity from about 15% carbon monoxide conversion to 33% carbon monoxide conversion; whereas, the silica supported catalyst showed only a very small increase in activity from 11% carbon monoxide conversion to 12% carbon monoxide conversion, while the titania supported catalyst showed a larger, but still modest, gain in activity from 11% carbon monoxide conversion to 17% carbon monoxide conversion.

From these examples, plus those presented previously, it can be concluded that the catalytic activity of a cobalt catalyst supported on alumina is greatly improved by adding minor amounts of rhenium, as long as the cobalt level is greater than about 5 wt %. Although improved activity from rhenium addition is also observed for some other supports, the activity level achieved by adding rhenium to a catalyst supported on alumina is much higher than for other supports. This result is surprising and would not have been predicted based on teachings in the prior art.

To demonstrate the advantages of including an alkali in the catalysts of this invention, the catalysts of Examples 42 through 55 were prepared and tested.

The catalysts of Examples 42 to 50 were prepared on Harshaw Al 4100P alumina which had been screened to 100-270 mesh. The catalysts of Examples 51 to 56 were prepared on Ketjen CK300 gamma alumina, screened to 20-40 mesh. Both supports were calcined overnight at 500.degree. C. before use. The technique used for preparing the catalysts was incipient wetness, as described in Example 1. The amounts of materials used in the preparation of each catalyst are shown in Table IX. The prepared catalysts were then air dried for 5 to 24 hours in an oven at a temperature of 120.degree. C. The dried catalysts were then calcined in air by raising the temperature at a heating rate of 1.degree. C./minute to 300.degree. C. and holding at this temperature for 2 to 16 hours. The compositions of the finished catalysts are shown in Table IX.

The results from testing these catalysts are shown in Table X. As is clear from these results, the addition of an alkali to the catalyst serves to increase the average molecular weight of the product, as evidenced by an increase in the Schulz-Flory .alpha.. Higher levels of the alkali result in higher .alpha.'s as illustrated in FIG. 3. However, activity decreases as the alkali content increases, as illustrated in FIG. 4. Therefore, for any particular situation there is an optimum alkali level that balances desired average product moleular weight and catalyst acitvity. Also, the effectiveness of the alkali varies from one alkali to another with, for example, potassium being more effective than lithium.