BIOS 1722 - Synthesis with Iron Catalysts

Synthesis with Iron Catalysts

By the end of the war the research and development work on iron catalysts had reached such a stage that Ruhrchemie were making plans for full-scale tests before introducing them as standard catalysts for medium pressure synthesis. Lurgi now consider that, as a result of smaller scale tests, they would be prepared to build a plant and give a guarantee of 150 g. liquid and solid products/N.cu.m. Dr. Michael, of I.G. Farben, pointed out, however, that on account of the ease with which iron catalysts bolt, the step up from the semi-technical scale to full-scale might not be quite a straightforward one.

Unlike the case of synthesis with cobalt catalysts, where most of the work has centered round one standard catalyst, a large number of iron catalysts have been developed, which require very different conditions for synthesis and give different types of products. The preparation and performance of the more important of these catalysts will be given in some detail in this section, but before this is done a few more general points, which apply to most iron catalysts, will be considered.

Catalyst Costs

In comparing costs for iron and cobalt catalysts, it must be remembered that the cobalt required to make up for losses is not included in the cost data given by Ruhrchemie. If therefore the steps in the preparation are about equally costly, it is a question of comparing the cost of the cobalt loss with the cost of the extra chemicals needed in the preparation of iron catalysts because the cost of the iron, the raw material, is small. Ruhrchemie consider that it is quite possible for these extra chemicals to outweigh the cobalt loss factor and so to make iron catalysts the more expensive. They gave the following estimate, Table 43, for the manufacture of an Fe100: Ca 10: Cu 5: kieselguhr 50 catalyst, based on small-scale discontinous working. The calculation is for a scale of working to produce 70 reactor charges per month, each of 10 cu.m., and assumes that the used catalyst is worked up as completely as possible, so that additions of iron, copper and calcium are only required to make up for losses. No credit is allowed for recover sodium nitrate.

Pretreatment

Just as cobalt catalysts require reduction in hydrogen before they become active for synthesis, so iron catalysts require pretreatment with water-gas, synthesis gas, carbon monoxide or hydrogen at atmospheric pressure or at reduced pressures before they can be used for synthesis at medium pressures. Ruhrchemie could give no general conclusions on the relative merits of the various methods that have been proposed for carrying out the pretreatment, and they thought that possibly different types of catalyst required different pretreatments. They gave the following rather tentative scheme:-

(1) Pretreatment with hydrogen

(a) Normal reduction. Maximum temperature 300°C., time 30 - 60 mins., gas rate 6,000 litres/itre catalyst/h. using 75% H₂, 25% N₂. This gives good results. For wax-forming catalysts very mild treatment is required.

(b) Longer time and higher temperature. T = 325 - 400°C., time 24 hours, gas rate as before, is a good pretreatment for very concentrated catalysts.

(c) Low gas rate and low temperature. This gives poor results.

(d) High gas rate and low temperature. The first results were good but researches have not been completed.

(2) Pretreatment with hydrogen, followed by carbon monoxide or water-gas. It is the initial hydrogen treatment that determines the activity of the catalyst, the subsequent carbon monoxide or water-gas treatment has little effect.

(3) Pretreatment with carbon monoxide. T = 250 - 325°C., time 24 hours, using pure CO or CO largely diluted with nitrogen. Atmospheric pressure. In nearly all cases this gives good results.

(4) Pretreatment with water-gas. Using gas with 39% CO, 48% H₂ at 250 - 300°C. and gas rates from 200 - 6,000 litres/litre catalyst/h. at atmospheric pressure good results were obtained but the reaches were not completed.

(5) Work was started on pretreatment with iron pentacarbonyl, diluted with nitrogen, at 50 - 250°C., but again, although the first results were good, the work has not been finished.

Dr. Rottig, of Ruhrchemie, considered that the pretreatment is nothing more than a reduction, and that any carbide formation that occurs is merely incidental. After pretreatment the percentage of metallic iron must be less than 5%, calculated on the total catalyst, or less than 10% calculated on the total iron. The metallic iron plus ferrous oxide must not exceed 50 - 60%, on the total iron. Catalysts which are less reduced than this are active, catalysts more reduced are poor. This has been found to be very important, and as a result the extent of reduction is always measured as a routine test, as follows: The metallic iron is determined in the usual way with mercuric chloride. The (Fe + FeO) is determined by boiling for 2 hours under reflux with 2% acetic acid, and the total iron by dissolving in hydrochloric acid in an atmosphere of carbon dioxide.

Lurgi agree that the pretreatment is nothing more than reduction and always use hydrogen. They say that if the catalyst is intended to produce wax it should only contain some 10% metallic iron, but if benzin is the product required, reduction should be carried as for as possible by increasing the time and the gas rate.

Rheinpreussen, however, take the opposite view, and say that pretreatment is not reduction but consists essentially of carbide formation. No advantage is gained by reducing the catalyst before the pretreatment as iron oxide forms carbide much more readily than metallic iron. But in spite of this general conception of pretreatment Rheinpreussen still think that a further study of the reduction of ferric oxide would be very profitable from the point of view of the synthesis.

They agree with the general view that pretreatment at reduced pressure is no better than at atmospheric pressure, and they also say that their own catalysts precipitated on dolomite as a carrier form no free carbon during pretreatment, which is a great advantage. Te K.W.I. agree with this carbide theory of pretreatment and produce thermomagnetic curves to show that during the pretreatment Fe₃O₂ changes to Fe₃C by way of an intermediate compound which is probably a lower carbide. They say that the catalyst only begines to become active when carbide is formed. Their best catalyst, like the Rheinpreussen catalyst, gives no free carbon during pretreatment.

The Utilization Ratio

Since the only oxygen product formed during synthesis with cobalt catalysts is water, the ratio of hydrogen to carbon monoxide used up is approximately 2. Iron catalysts normally produce a mixture of carbon dioxide and water as oxygen products, so that the utilization ratio, X, may be anything between 2 and 0.5, the value it would have if carbon dioxide and no water were formed. The utilization ratio is important because its value controls the composition of the residual gas, and if this gas is to be suitable for use in a second stage of synthesis, the utilization ratio must be approximately the same as the ratio of hydrogen to carbon monoxide in the synthesis gas.

Ruhrchemie gave the following account of the factors controlling the utilization ratio. As far as the composition of the catalyst is concerned, the only factor that affects X is the alkali, and the more alkali that is present, the smaller X becomes. The following example was given to illustrate the extent of the alteration that can be produced. Two catalysts of the composition Fe 100: Cu 5: CaO 10: kieselguhr 120 were precipitated with the same amount of alkali, using 10% excess over the theoretical amount of 8% NaOH solution. One sample was washed with 1,200 ml. boiling water/25 g. Fe and the other with only 600 ml. For synthesis with water-gas the first sample gave a utilization ratio of 1.38 and the second 1.18. The kieselguhr as such does not affect X although it may appear to do so since it may introduce alkali into the catalyst.

Of the synthesis conditions, the most important increase in X is that effected by recirculation. The exact change cannot be predicted, but as a typical example, if X is 0.95 for normal synthesis, recirculation will increase it to 1.25. Brabag say that in general those catalysts which normally form a large proportion of wax in the products respond well to recirculation, and the utilization ratio X can be improved in this way up to 1.2. The second Reichsamtversuch was carried out to investigate the effect of recirculation, and for it a recirculation ratio of 2 was used.

X is increased by increasing the proportion of hydrogen in the synthesis gas, and is decreased by conducting the synthesis at a temperature above the best working value, or by running with a lower CO conversion than normal. This latter is not a perfectly general rule, particularly at low temperatures.

The statement that pretreatment with water-gas increases X has not yet been confirmed but may well be true. It is also possible that X is dependent on the pressure, because in one case it was found to be 1.0. at 10 atm. and 1.10 at 20 atm.

General Synthesis Topics

Some comparison of the performance of a catalyst at atmospheric and medium pressures is important because of the convenience of making a preliminary judgment of a catalyst on the result of a short test at atmospheric pressure. Ruhrchemie said that such tests were useful inasmuch as a catalyst which is active at atmospheric pressure will be more active at medium pressure, and that if one catalyst is better than another at atmospheric pressure it will also be more active at medium pressure.

As far as the effect of pressure on the synthesis is concerned, Lurgi say that the results at 50 atm. are always worse than at 20 atm. At the higher pressure too much of the catalyst surface is covered with products, and what is worse, water is very near its condensation point and damages the iron surface.

Lurgi also point out that the whole question of the second stage is complicated by the fact that the carbon dioxide produced in Stage I has a definitely harmful effect on Stage II. For synthesis with iron catalysts it is not an inert constituent as it appears to be for cobalt, but produces a diminution of activity probably due to oxidation. Lurgi compared water-gas diluted with 30% N₂ and with 30% CO₂ and found that in the second case the yield was 20% less at the same temperature, or to get the same yield 20° higher temperature must be used.

The removal of the carbon dioxide is costly because the benzin has to be removed completely before the Alkazid washing or it is lost and this means that a much bigger active carbon plant would be required than is normally used between the stages.

At the moment therefore, leaving the carbon dioxide in the gas, the conversion U_CO + H₂ cannot be made to exceed 90%.

The Mechanism of the Synthesis with Iron Catalysts

Rheinpreussen considers that the synthesis proceeds by way of a labile carbide and say that the deterioration of the catalyst is due to the change of this carbide into a stable one which cannot act as an intermediate compound for synthesis. This change to the stable carbide is favoured by a high proportion of carbon monoxide in the synthesis gas and also by working at atmospheric pressure. The labile, active, carbide may be one of the higher iron carbides, Fe₂C or higher, and magneto-thermal analysis has shown that Fe₂C does change into Fe₃C and also that catalysts contains more Fe₃C towards the end of their lives than at the beginning.

At atmospheric pressure the formation of carbide is a more rapid reaction than the reduction, and the reverse is true at medium pressures. Hence synthesis gas containing a higher proportion of carbon monoxide must be used at medium pressures.

At atmospheric pressure, free carbon is formed very easily, and Rheinpreussen say that their success in this field is due to a large extent to their preliminary work on the different reactivities of a-and a-Fe₂O₃.

Finally, they say that for iron catalysts the surface area is not nearly so important as it is for other catalysts.

The K.W.I. obtained the following results by thermomagnetic work: Their catalyst with 1% Cu. (see p. 122) after pretreatment with carbon monoxide at reduced pressures contained pure Fe₃C. The catalyst with 20% Cu. (see p. 122) after pretreatment with synthesis gas contains Fe₃C, Fe and Fe₃C₄. Yet both samples are equally active. During synthesis the amount of carbide decreases in the first case but stays constant in the second. The oxygen content of the catalyst increases during synthesis, and this corresponds to a fall in activity.

If synthesis is attempted at medium pressure without pretreatment, no carbide is formed and the catalyst is inactive.

Catalysts for Synthesis at Higher Temperatures

Before interest in Germany turned towards finding an iron catalyst which could be used in the existing medium pressure plants, Ruhrchemie developed several iron catalysts which would work very successfully at 250°C. and higher temperatures. One of these was to be used in the plant designed for the Societa Italiana Carburanti Sintetici at Arezzo. It had the composition Fe 100: Cu 5: CaO 10: kieselguhr 150, and was prepared by rapid precipitation from the nitrates with potassium hydroxide solution at the boiling point. After reduction in dry hydrogen for one hour at 300°C., synthesis was carried out at 20 atm. pressure, with water-gas, recirculation ratio 2.5, and a space velocity of 90. The temperature only required raising from 251 to 257°C. during the course of 120 days, but Ruhrchemie say that on account of the high proportion of kieselguhr it was possible to use it over a wide temperature range. In the experimental work with this catalyst, on which the Arezzo project was based, a two-stage pilot plant was worked to an 83% conversion and gave a yield of liquid and solid products of 115 g./N.cu.m. from this it was calculated that the yield would be at least 140 g. if the plant were operated to a conversion of 93%. It was a wax-forming catalyst, the products containing 35% wax, of which 17% was hard wax of boiling point above 460°C.

Ruhrchemie then found that the expensive step of precipitation with potassium hydroxide could be avoided and the catalyst precipitated with sodium carbonate so long as the sodium salts were washed out thoroughly and a little potassium carbonate or hydroxide incorporated with the washed precipitate. Such a catalyst produced as little methane as a catalyst precipitated with potassium hydroxide. They also found that if they boiled the kieselguhr for a short time with the sodium carbonate solution that was to be used for the precipitation, all the good effects were obtained which Lurgi had previously achieved by adding waterglass to their catalysts (see p. 124). Ruhrchemie claimed that their method gave more effective catalysts and was also cheaper. They always used the same kieselguhr as is used in the manufacture of cobalt catalysts, but they have tested kieselguhr containing both calcium and iron, and they are perfectly good for the preparation of iron catalysts has been found.

As an example of one of their best high temperature catalysts, Ruhrchemie gave the following example. A preparation of the composition Fe 100: Cu 5: CaO 30: kieselguhr 100 was carried out on the semi-technical scale. A solution was first made up to contain 6 kg. iron with the correct amount of copper and calcium in 120 litres condensate water, and heated to boiling. At the same time 22 kg. soda was dissolved in 190 litres condensate water and also boiled. The kieselguhr, 6 kg. was stirred into this solution and boiled for 30 seconds. The solution of the iron was then added with rapid stirring during the course of 2 minutes. A further 3 kg. soda was added to the mixture and the stirring continued for 20 seconds. The precipitate was then filtered off on a filter press and washed for 30 minutes with hot condensate water. The filter cake was put through an extrusion press and the resulting dried catalyst was in a very hard and physically suitable form.

Using ordinary water-gas with 13% inert constituents, the results shown in Table 44 were obtained for synthesis at 20 atm.

For synthesis at higher temperatures Rheinpreussen developed a catalyst of composition Fe 100: Cu 5: K₂CO₃ 0.25, precipitated by sodium hydroxide and containing no kieselguhr. Synthesis was done with gas with H₂/CO - 0.5, but work on these lines has now been stopped.

Ruhrchemie's present opinion about synthesis with such catalysts at 250°C. and above, is that it is less satisfactory on economic grounds, than synthesis with the best low-temperature iron catalysts. In particular, the higher working temperature means that both plant costs and working costs are higher.

Catalysts for Synthesis at Low Temperatures

As already mentioned, the shortage of cobalt in Germany during the war made it very important to develop an iron catalyst which could be used in the existing medium pressure plants, i.e., one which would work at temperatures below 225°C. The position up to the winter of 1943-4 is summed up by the 'Reichsamtversuch', a comparative experiment arranged by the Reichsamt fur Wirtschaftsausbau in which the six best iron catalysts available were tried out simultaneously at the Brabag works at Schwarzheide. Since that time a great deal of progress has been made and in this section an account of the best modern catalysts is given, together with some particulars of several of the Reischsamtversuch catalysts which were not previously available.

Ruhrchemie consider that catalysts must be very dense, and hence contain little kieselguhr, if they are to be active at low temperatures. Their best low temperature catalyst had the composition Fe 100: Cu 5 CaO 8: kieselguhr 30 and was prepared as follows. A solution of iron. copper, and calcium nitrates was made up to contain 50 - 55 g. Fe/litre and with Fe : Cu : CaO in the ratios 100 : 5 : 10, the excess CaO over the amount required in the final catalyst being used to allow for incomplete precipitation. This solution is heated to 98°C. and run into a boiling solution of sodium carbonate of strength 90 - 100 g. Na₂CO₃/litre which is stirred vigorously. The amounts of the solutions are chosen so that after the precipitation the pH is 6.8 - 7.0. The calculated amount of kieselguhr is then stirred in, using preferably a sample of low density which has been heated to 700°C. The mixture is filtered and washed with hot distilled water till the content of NaNO₃ in the finished catalyst is about 0.5 - 0.7% based on iron. From 200 to 220 litres water per kg. Fe are generally sufficient. The wet filter cake is then impregnated with 3.0 to 3.5% KOH, calculated on Fe by repulping it in a solution of KOH and filtering. A concentration of 6.) g. KOH/litre in the mother liquor is usually sufficient to get the correct amount of KOH in the filter cake. After filtration, the cake is formed in a suitable manner and dried at 110°C..

The catalyst is hard and resistant to abrasion. It is reduced at 300°C. for 1 hour with a mixture of 75% hydrogen and 25% nitrogen. The content of metallic iron after reduction should not exceed 8 - 10% and the metallic plus ferrous iron should not exceed 60 - 70% of the total iron present.

The utilization ratio, X, of this catalyst increases from 1.05 for the first month of syntheses to 1.20 after five months, but it is considered that by making small changes in the catalyst these figures could be improved. Small changes in the amount of kieselguhr could increase X a little, and a slight reduction in the amount of alkali might be required to male the catalysts suitable for use on the full-scale.

Dr. Kolbel, of Rheinpreussen, gave the following account of the best iron catalyst he has developed for medium pressure synthesis. Normally it has the composition Fe 100 : Cu 1 - 10 : dolomite 50 - 100 : K₂CO₃ 0.1 - 3, but for work with gases rich in carbon monoxide the dolomite may be replaced by kieselguhr. In general, dolomite is a very satisfactory support for medium pressure iron catalysts as it promotes the formation of wax, prevents the formation of carbon and gives a very hard catalyst. A band of dolomite containing no sulphur or phosphate is chosen in the mine, and the material from this band is roasted at 700°C. and ground to a fine powder. But a synthetic dolomite made simply by mixing magnesium and calcium carbonates was just as good.

The catalyst is prepared as follows:- The iron is dissolved in nitric acid acid to give a 10% solution, the copper nitrate added, the solution boiled, the dolomite added and the precipitation carried out by adding boiling 10% sodium carbonate solution. It is essential that these operations be carried out in the order given above. After filtration and washing, the potassium carbonate, dissolved in water, is stirred into the wet filter cake, which is dried and granulated to 1 - 3 mm. pieces. It is not formed by extrusion through dies. A preparation has been carried out on the large scale in the catalyst factory of Brabag and was perfectly successful, the filtration being easier than for the normal cobalt catalyst.

Pretreatment is done at atmospheric pressure, 240°C., in synthesis gas, H₂/CO = 2 which is carefully freed from carbon dioxide but need not be dried. The gas is passed at 10 times the normal gas rate.

Synthesis is carried out at 205 - 215°C. and 8 - 10 atm. with synthesis gas having H₂/CO = 2. No recirculation is necessary, in fact, it is disadvantageous as it reduces the proportion of wax in the product slightly. Apart from this, the use of synthesis gas is preferable to that of water-gas as there is no formation of carbon, less formation of carbon dioxide, the presence of which makes the active carbon stage of the condensation more difficult, and the wax produced is said require a temperature some 15° higher.

Worked in two stages, with a space velocity of 500 - 1000/h. for Stage I, the yield is 150 -155 g./N.cu.m., excluding methane, which corresponds to a production of 5.5 t. products/full-scale reactor/day. The life of the catalyst is 6 months or more. The utilization ratio, X, is 1.5 and is better in Stage I than in Stage II. It is known to depend on the alkali in the catalyst when worked with gas rich in carbon monoxide, but it is not known whether there is any such effect when working with synthesis gas. The methane formed amounts to 5 - 10% of the products, which is less than the amount formed by medium pressure synthesis with a cobalt catalyst. The products are in general similar to those obtained from cobalt catalysts at atmospheric pressure, in particular, in having only a small proportion of wax. The wax can be increased by using more alkali in the catalyst, a higher proportion of carbon monoxide in the synthesis gas, or by working at higher pressures. In this latter case, however, more organic oxygen compounds are produced.

The catalyst used by Rheinpreussen in the Reichsamtversuch at Schwarzheide in 1943 was an earlier and less effective catalyst of this type, of composition Fe 100 : Cu 7.5 : dolomite 50 - 65 : K₂CO₃ 2, pretreated in synthesis gas at reduced pressure. The yield was not good as the experiment had to be done in one stage.

Work at the K.W.I. had been on rather different lines, and their best catalyst for synthesis at low temperatures is one containing no kieselguhr or other support. It has the composition Fe 100 : Cu 20 : K₂CO₃ 1 and is made from a mixture of 75% FeCl₂ and 25% FeCl₃. An amount of these salts corresponding to 50 g. iron is dissolved in water, made up to 3 litres, the solution heated to 70°C. and precipitated rapidly at that temperature by the corresponding amount of sodium carbonate dissolved in 1 litre water. A slight excess is used so that the solution is just alkaline at the end of the precipitation. The mixture is then heated to 100°C., filtered, washed free from alkali, the evaporated and dried at 105°C. No special precautions against oxidation are taken during the drying and, in fact, drying in air is better than in an inert atmosphere. More ferric salt must be used in the original mixture if it is desired to dry out of contact with air. The dried catalyst is granulated on a sieve.

It is pretreated with synthesis gas, H₂/CO = 2, at 225°C. and atmospheric pressure, at a space velocity of 500/h. The gas is neither dried nor freed from carbon dioxide. This pretreatment lasts for 2 days. At the end of this period it may become inactive, on account of wax formation, and if so, it must be extracted with high boiling Kogasin (230 - 300°C.) at 225°C., and the 2 days pretreatment repeated. After a further extraction it is ready for synthesis.

Up to now no attempts have been made to start synthesis at medium pressure directly but two weeks work at atmospheric pressure has always been done first (for details see p. 140). After this two weeks, the pressure is increased to 2 atm. and synthesis continued for another 2 weeks at 212°C., using water-gas. The pressure is then put up to 10 atm. and synthesis can be done at a constant temperature of 198 - 200°C. for three months. No extraction with solvent is necessary although much more wax is being formed than at atmospheric pressure. The average yield is 143 - 148 g./N.cu.m., including methane, which latter is, however, vanishingly small. The utilization ratio X - 0.7 - 0>8. The effect of recirculation has not yet been tried.

A good medium pressure catalyst is alos obtained by reducing the 20 parts of copper in the above preparation to 1 part. When this is done the pretreatment is best if carried out with carbon monoxide alone, at 1/10 atm. pressure. The synthesis temperatures are rather higher than if 20 parts of copper are used.

Referring back to the Reichsamtversuch, the K.W.I. said that the experimental reactor used at Schwerzheide had a wider tube than their own reactors, and as a result, their first catalyst gave too much liquid and gaseous products. This was rectified by preparing a second catalyst containing more alkali and this one ran successfully for 3 months. A third catalyst, with still more alkali, was tried for 3 weeks.

Lurgi point out that the K.W.I. catalyst produced a great deal of C₂ hydrocarbons and that these are included in the yields quoted. In practice these hydrocarbons could not be recovered and hence the actual yields would be lower. The final gases contain 40 - 50% CO₂ and 1 - 2% CO₂ Hydrocarbons, and if these are to be recovered the CO₂ must be removed by Alkazid washing, which in turn causes losses of hydrocarbons. Thus the recovery of C₂ is always too dear. Although the results obtained in the Reichsamtversuch with all the six catalysts are well-known, the compositions of the Lurgi, the Brabag and the I.G. Farben catalysts were not available. These have now been obtained.

The Brabag catalyst had the composition Fe 100: Cu 20 : Zn 20 : K₂CO₃ 1, the zinc being in the form of zinc oxide. It was a normal precipitated catalyst, made from the nitrates and sodium carbonate. It was pretreated with water-gas at 235 - 240°C. at atmospheric pressure. When the production of carbon dioxide falls off the pretreatment is considered to be finished, and it usually requires some 48 hours. If disintegrates during synthesis.

The first of these catalysts used in the Reichsamtversuch was too soft, disintegrated, and came out of the reactor with the products. The second attempt was spoilt because the temperature was increased too rapidly and deposition of carbon occurred. The third attempt was successful.

The Lurgi catalyst had the composition Fe 100 : Cu 10 : Al₂O₃ 9 : SiO₂ 30 : K₂CO₃ 2, the K₂CO₃ being added to give 2% K₂O. Subsequent Lurgi catalysts had 25 parts copper but behaved essentially in the same way as the above except that the utilization ratio was improved. These catalysts are prepared by precipitation with sodium carbonate from a hot solution of the mixed nitrates; freshly precipitated silica, from waterglass and dilute mineral acid, is them stirred in and the mixture filtered, washed and impregnated with potassium carbonate.

Lurgi say that pretreatment may equally well be done with hydrogen, water-gas or diluted carbon monoxide but they prefer hydrogen as this avoids all trouble due to synthesis setting in, bolting and carbon formation. For full-scale working hydrogen would be the obvious choice for two reasons. Firstly, if water-gas were used the treatment would have to be done in the reactor itself because of the danger of synthesis setting in, and the reactors are not capable of being used at temperatures high enough for the pretreatment. And secondly, on account of the very high gas rate required, the water-gas would have to be recirculated and before it could be passed through recirculated and before it could be passed through recirculation pumps all the synthesis products would have to be scrubbed out of it.

The pretreatment actually used is carried out with hydrogen at a space velocity of 1,000/h. for 1 - 3 hours at 280°C., the exact time depending on the alkali and copper content. The reduction is slower the more alkali present. The temperature required varies from 250 to 300°C., according to the amount of copper present, and in itsabsence temperature above 300°C. are required. If pretreatment is done with gases containing carbon monoxide, under certain conditions temperatures as low at 180 - 200°C. may be used. Temperatures above 300°C. always give bad results.

The catalyst Lurgi used in the Reichsamtversuch started synthesis successfully at the first attempt and worked to a utilization ratio X = 0.7 without recirculation or 1.0 with recirculation. Their latest catalyst gives X = 1.3 and they are confident that this could be increased to 1.6 if necessary. Ruhrchemie and Rheinpreussen got as good a utilization ratio but only at the expense of producing more methane. The silica gel is the factor which gives a good utilization ratio.

The alkali silicate, that is, the K₂O plus an equivalent amount of silica, is responsible for the good physical form and the long life. It stabilizes the paste during the preparation of the catalyst and prevents thixotropic liquefaction, and as a result, the catalyst threads after drying are hard and break like glass. A normal iron catalyst becomes more and more reduced during continued synthesis but this alkali silicate catalyst does not.

The life is very good; thus, during the 90 days run at Schwarzheide the temperature had only to be raised 2°, from 216 to 218°C. to maintained activity.

The I.G. Farben low-temperature iron catalysts is a melted catalyst of the composition Fe : Al₂O₃ : CoF₂ : K₂O the calcium fluoride being added to increase the formation of light hydrocarbons. It is a very remarkable fact that such a catalyst was just as active as the other catalysts used in the Reichsamtversuch, all of which were made by precipitation and pretreatment under carefully controlled conditions. To produce wax, the calcium fluoride is left out, and preparations of the composition Fe : K₂O : MgO or Fe : K₂O : Al₂O₃ are used, the activating oxides being melted with the iron in oxygen. At 230 - 250°C. they produce products containing more than 60% wax. The cost of preparation is about the same as that of precipitated catalysts.

Catalysts based on Luxmasse

Ruhrchemie say that it is difficult to make such catalysts hard enough for use on the full-scale. They are essentially catalysts which tend to form light products and not wax, and they require a high temperature. They are considered on all these grounds o be inferior to the low-temperature catalysts already referred to.

As an example, the catalyst Fe 100 : Cu 3.5 : bleaching earth 100 : kieselguhr 10 may be described, as it was prepared and tested by Ruhrchemie in 1940. A volume of 1,100 ml. 10.3% KOH, specific gravity 1.085 was heated to boiling, 3.5 g. kieselguhr 120 added and the boiling continued for 1.5 minutes. Then, with continuous intensive stirring a suspension of 200 g. Luxmasse (35 g. Fe), 35 g. bleaching earth (Granosil) and 10.3 ml. copper nitrate solution (1.25 g. Cu) in 500 ml. hot distilled water was added and heated with continued stirring till boiling began. The mixture was then filtered, washed with three portions, each of 200 ml., of distilled water at 90°C., and the filter cake dried at 110 - 115°C. After granulation in the usual way the bulk density was 500 g./litre.

Synthesis was started with water-gas at atmospheric pressure, at a space velocity of 100/h. and the temperature was increased until a contraction of 20% was obtained. The pressure was then put up to 15 atm. and synthesis carried out with a recirculation ratio of 1.5 to 2.5, at 255°C. The utilization ratio, X, was 1.27 and the conversion, U, 60%. The yield, calculated for a 90% conversion, was 95 g. liquid products and 21 g. gasol per N.cu.m., and the liquid product had the following boiling range:- below 200°C. 39%, 200 - 320°C. 31%, wax above 320°C. 30%. The product boiling below 200°C. had an octane number of 72.

Iron Catalysts for Gas Rich in Carbon Monoxide

All iron catalyst development done by Ruhrchemie was in the direction of catalysts to work with water-gas, and they had no immediate interest in catalysts suitable for gases containing a higher proportion of carbon monoxide. Using such gas, more olefines are produced, but at the cost of a higher synthesis temperature, which means bigger plant costs and working costs. Carbon formation and dark coloured wax also occur. The effect of recirculation on these matters was not known.

The essential requisite for synthesis with gas rich in carbon monoxide is a supply of cheap oxygen, because suitable synthesis gas economically.

An example of a very successful catalyst for a gas with H₂/CO 2 : 3 is the K.W.I. catalyst of composition Fe 100 : Cu 1 : K₂Co₃ 0.25, prepared by dissolving iron in nitric acid under fixed conditions such that a little ferrous nitrate is formed, and then precipitating in the normal way. It works at 235 - 245°C. and 10 - 15 atm.

Lurgi consider that if such a gas could be prepared as cheaply as ordinary water-gas, the best ratio of H₂/CO is 1 as it gives more wax, more olefines and less methane than ordinary water-gas. But they consider that gases richer in carbon monoxide are not desirable.

The I.G. High-Velocity Gas Recycle Process

The following additional information about this method, in which the heat of reaction is removed by recycling cooled residual gas, at a very high rate, was obtained from Dr. Michael who was responsible for its development.

The physical state of the catalyst is very important. In the first place it must be very or otherwise it breaks up in the rapid gas stream and the resistance of the bed becomes uneven, with the result that part of the catalyst is only exposed to a slow gas stream and so process overheated. Hence sintered carbonyl iron catalysts were normally used but precipitated catalysts will do equally well if they are sintered. Irregular pieces of size above 10 mm. were used and with these, good gas distribution can always be attained by using a large excess pressure for the recirculating gas. It was realized, however, that if regular shaped pieces could be used the excess pressure might be reduced to 1/5th of the present value, with a corresponding saving of power.

The design of the reactors must be arranged so that the end where the gas enters is streamlined. This was not done in a 4 cu.m. reactor made by I.G. and turbulence occurred in such a way that in certain places the gas flow was not large and hot spots developed. If another large reactor had to be built for this purpose, one with about six alternate layers of catalyst and cooling tubes would be used and with such a design a recirculation ratio of 25 would be sufficient instead of 100. Figure 6 is a sketch of such a plant and was obtained from the I.G. Farben at Ludgwigshafen-Oppau. Another interesting type of reactor that has been used for 5 litres of catalyst had a high-speed fan at the top and a central tube through the catalyst bed, and the fan was arranged to circulate gas up through the central tube and down through the catalyst bed. Both fan and motor were enclosed in the pressure resisting casing of the reactor. In this way an internal recirculation could be carried out in addition to the main external cooling recirculation, and the temperature distribution in the reactor was thereby improved. Dr. Michael thought it might also be a good idea to arrange the catalyst is an annular space between two perforated walls and pass the gas through, radially, from the inside to the outside. With such an arrangement catalyst could be removed continuously from the bottom, reactivated, and put in again at the top.

Starting with synthesis gas with H₂/CO = 0.8 to 1.2, Dr. Michael said that the high-velocity gas recycle process will give an output of 800 kg./cu.m. catalyst/h. at a conversion of U_{CO + H2} = 75 - 80% and at a temperature of 300 - 330°C. The reaction goes in such a way as to produce four moles carbon dioxide to one of water, as the oxygen products. Of the gas converted, 20% goes to form methane and the composition of the products higher than methane is given in Table 45.

The benzin has an octane number of 68 and the much higher figures frequently quoted refer to the product obtained by refining the benzine with alumina or bleaching earth. The higher oil is an inferior lubricating oil. Very little wax is formed, but what little there is, is branched chain and resembles ceresin.

The life of the catalyst should be some 6 months on the large scale, but on the smaller scale, with internal circulation as described above, lives of 10 months have been obtained.

At lower temperatures, 280°C. for example there is less methane formed but the quality of the benzin is worse. More wax is also formed, and the lower the temperature the higher the percentage of oxygen in the wax. At these lower temperatures the catalyst requires treatment with hydrogen at 400°C. every three days or so, as it gets covered with oxygen-containing resins.

It was stated that when a distribution curve is plotted for the products of this process, showing yield of each individual chain length, the yeild falls off steadily as the chain length increases, approaching zero asymptotically for large chain lengths, and has no maximum and minimum at C_5-6 and C₂ respectively, as in the case for the products of the normal synthesis. This is very surprising.

The final position with regard to the high-velocity gas recycle process was that it was generally acknowledged to be too expensive for the production of benzin. To get heavier products more active catalysts were required but precipitated catalysts were found to be much too soft to stand up to the high recycle conditions.

In several respects, the conditions are similar to those that hold good when the fluidized bed technique is used, and Dr. Michael had actually carried out experiments with a fluidized bed in the reactor with internal circulation. Results were very good at first, but the particles of catalysts, soon stuck together and the method ceased to work. During the initial period the rate of production of products was ten to twenty times the normal rate.

As is well-known, the catalysts used in the work with high-velocity gas recirculation were made by sintering carbonyl iron. It was thought to be useful therefore to find out how I.G. Farben produced their carbonyl iron, and the following information was obtained. Type A iron is produced by decomposing pure iron carbonyl vapour at 250°C. The particles formed contain 1.2% carbon and are of sizes ranging from 3 to 10 microns. Type E iron in prepared by adding 10% ammonia to the carbonyl and then decomposing it as before. The amount of carbon is less than in type A but the powder contains a little nitrogen. All the carbon in these preparations can be removed by hydrogen at 400.

Synthesis in the Liquid Phase

The following is an account of the work of Dr. Kolbel of Rheinpreussen in this field. A catalyst without a carrier was used, of composition Fe 100 : Cu 0.2 - 0.5 : K₂CO₃ 1, precipitated in the way described for the Rheinpreussen medium pressure catalyst. The 1% alkali prevents the formation of methane, and if less alkali were used the main product would be low-boiling hydrocarbons instead of wax. Pretreatment is done for 1 day with a gas with H/=0.5 at a temperature 10 - 20° above the initial synthesis temperature.

This catalyst is suspended in Diesel oil so as to give a suspension containing 10 20 % Fe, and synthesis is done in one stage at 230 - 245°C., 3 - 10 atm., with synthesis gas having II₂/CO = 0.5, at a space velocity of about 75/h. Recirculation is not necessary. The utilization ratio, X, is 0.5 as the reaction forms carbon dioxide and no water. It is interesting that part of the Diesel oil reacts to form larger molecules, and in particular, that not only the olefines in it react in this way, but also the paraffins. As a result, the amount of wax obtained per cu.m. synthesis gas is extremely high, being 130 - 140 g. The yield of products produced from the synthesis gas, that is, not including the weight of material derived from the Diesel oil, is 180 g./cu.m., and this does not include methane because none is formed. Working to this yield, the catalyst life is 1 - 2 months, but for smaller yields the life is longer. The products contain 50 - 60% wax, b.p. > 290°C., 50 - 60% olefines, and 10 - 25% alcohols etc. The more alkali in the catalyst the more alcohols are formed.

The synthesis products are obtained by withdrawing oil and catalyst, filtering and returning the catalyst, all this being done in an atmosphere of synthesis gas. After 1,000 hours, the liquid in the reactor is renewed, being replaced by fresh Diesel oil.

The experimental plant was a tube 150 mm. diameter and 3 - 4 m. high, with cooling coils either inside or outside. This plant used 3 cu.m. gas/h. A pilot plant was constructed with internal cooling coils to deal with 750 cu.m./h. but this had not yet been used. For a full scale plant, very large reactors could be used, each capable of producing 7,000 t. products/year, on account of the simplicity of the design.

For this reason, coupled with the high yield of wax and negligible formation of methane, Rheinpreussen consider that this is the most promising of all the proposed modifications of the Fischer-Tropsch synthesis.

Ruhrchemie have also done a large amount of experimental work on liquid-phase synthesis and they gave the following account of their results. Before 1941 they satisfied themselves that if the gas-rate were chosen to get a reasonable conversion it was insufficient to keep the catalyst suspended and that for this reason alone recirculation was necessary, so that a high gas velocity through the converter could be combined with the normal low throughput. With recirculation, results were very promising and work was continued on this basis.

Two reactors were available (see Figure 1 of BIOS Final Report 447) each consisting of a wide vertical tub, 6m. high fitted with jackets through which the heating oil circulated. The first contained 55 litres of the oil-catalyst mixture. The second contained 85 litres and differed in having a wider head so that less oil was carried off in the gas stream, particularly at higher gas rates. Whatever the gas rate, however, much oil is always carried off and the level of liquid in the converter sinks. Hence a condenser is arranged to condense out all the oil in the gas and return it continuously to the converter. The water also condenses here and is separated from the oil before the latter is returned. Only benzin and gasol remain in the gas and are absorbed in the active carbon scrubbers. The oil level in the reactor, as shown by a sight-glass, is then kept constant by drawing off the liquid synthesis products periodically through a filter stick immersed in the oil. As the oil level is altered by very small alterations in the pressure, if the latter is not kept absolutely constant the amount of oil drawn off, and hence the yield calculated from it will be very variable. Hence not only must the pressure be kept as constant as possible but yields must never be calculated for periods shorter than several days, or a week. In making alterations in the pressure, for starting or stopping an experiment, sudden changes must be avoided or the oil-catalyst mixture will froth over into the condenser and separator.

In the first experiments the gas was passed into the reactor through a ceramic plate which divided up the gas stream into bubbles. After interruptions of the synthesis which allowed the catalyst to settle, this plate became blocked and pressure differences of up to 8 atm. across it were necessary to re-establish to correct gas rate. It was replaced therefore by a simple copper capillary tube. This never blocked and although the gas was not distributed nearly as well the conversion was just as good.

If the reactor has to be shut down for any purpose the whole contents are drawn off from a valve at the bottom, while gas is still passing to keep the catalyst in suspension. On cooling the mixture sets solid, and in doing so, partial separation occurs. But on remelting and putting the mixture back in the reactor the activity of the catalyst is undiminished. Emptying and refilling may be done many times in this way without affecting the activity.

At the end of an experiment the contents of the reactor are emptied as just described, filtered hot, and the wax obtained used as the liquid medium for the next experiment. This is better than using fresh Diesel oil each time because it takes from 2 to 4 weeks, according to the nature of the catalyst, for this oil to be replaced by wax.

All the following data refer to synthesis with water-gas, and the temperatures required are the same as for ordinary medium pressure synthesis, being from 230 - 250°C. In general no increase in temperature is necessary during an experiment after the end of the running-in period. A pressure of 15 atm. is always used as there is no reason to suppose that the best pressure is different from that for ordinary synthesis.

The conversion of carbon monoxide, U_CO, is 70 - 75% in one stage, with recirculation ratios of 1 to 3, but at times it may be as high as 80 - 82%. If synthesis were being conducted in several stages, however, as it would be on a large scale, it is better to restrict the conversion in Stage I to 65 - 75% so as to minimize gas formation and get the best possible catalyst life. The residual gas from the first stage is perfectly suitable for use in a second stage, where a conversion of 70 - 75% can be obtained. In a two-stage test a total conversion of U_CO = 93 - 94% was obtained.

The formation of methane is in all cases smaller than for the corresponding ordinary synthesis. During the early part of the catalyst life it is particularly low, being 1.5 - 2.0% or even less than 1% of the total products. For a catalyst which produces a high proportion of wax, the average amount of methane formed is 3 - 4% when working at a total conversion of 70 - 90%. In the later stages of the catalyst life the formation of methane increases.

The utilization ratio, H2/CO, is bigger than for the ordinary synthesis under otherwise similar conditions. For Stage I it is usually above 1.3, being 1.4 - 1.5 in two examples quoted. Strangely enough it is hardly affected by the recirculation ratio, and changing this from 1.7 to 7 has little effect. Even without recirculation it is still above 1.3. The utilization ratio in Stage I will be smaller, and is in fact about 0.90. For the two stages together, the utilization ratio is 1.24 - 1.28 which corresponds well with the composition of the water-gas. Further work has the aim of making the utilization ratio equal to this for each stage separately because in any stage, deviation of the utilization ratio from the composition of the gas supplied leads to a diminution in yield.

The effect of the gas rate and of the concentration of the catalyst in the oil is shown in Table 46. The volume of catalyst is the volume of dry granules, before grinding, which is mixed with with the requisite amount of oil to fill the reactor. When 10 litres of catalyst are used, 50 litres of oil are required, and when more catalyst is used rather less oil is employed. All results are for a recirculation ratio of 3.

The yield, calculated from 90 - 93% conversion obtained for two stages, is 170 g./cu.m. but direct measurements of the yield have not yet been made, on account of difficulties in measuring the amounts of gases concerned.

Just as in ordinary synthesis, the nature of the products is dependent, in the first place, on the catalyst used, and in the second, on the synthesis conditions. A catalyst which gives a high proportion of wax by ordinary synthesis will also do so in the liquid phase synthesis. It is uncertain whether hard wax is cracked slowly, to soft wax and Diesel oil due to its long contact with the catalyst. There is one difference from ordinary synthesis, and that is the products are more saturated.

The longest test so far carried out last for 7 1/2 weeks without any marked deterioration of the catalyst, although the production of methane increased from 2 - 2.5% to 6 - 8% of the total products. It can be said therefore that the life of a catalyst for liquid phase synthesis is no less than for ordinary synthesis.

The effect of adding various substances to the oil-catalyst suspension has been investigated. In order to decrease the utilization ratio to the right value for water-gas, alkali was added, in the form of soda finely powdered and suspended in Diesel oil, sufficient to form 10% by weight of the iron present. This decreased the utilization ratio from 1.30 to 1.08 - 0.96 and did not change the conversion or the gas formation. But after 6 days the wax produced became brown, and was finally brownish black, jelly-like and only half solid.

In another experiment 5 litres C_10-18 fatty acids were added. This put the utilization ratio up to 1.81 from the original value of 1.15, but it decreased again rapidly and after a few days was 1.05. The fatty acids had by this time been destroyed, as shown by zero acid or ester numbers for samples of the oil in the reactor. While the acids were still effective, the amount of methane formed increased to 6 - 10% and the conversion diminished somewhat.

The data in Table 47 give the results of a normal two-stage test.

I.G. Farben have also investigated liquid phase synthesis, and Dr. Michael gave the following account of their work. The catalyst is prepared from carbonyl iron by burning in oxygen. Some 1% K₂CO₃ or borax is added and the mixture reduced at 250°C. until it contains 2/3 metallic, iron 1/3 ferrous oxide. It is then mixed with the oil and milled with it to a size of 30 microns. Synthesis is carried out at 20 atm. pressure in a simple tube of capacity 1.5 cu.m., fitted with a porous plate. The oil is recirculated at such a rate as to change the contents of the reactor twenty times an hour, and with this rate there is a 10°C. rise of temperature through the reactor. The synthesis gas is passed in through the porous plate and a gas recirculation ratio of 1 is used. The resistance of the porous plate does not increase during synthesis. The progress of the synthesis can be judged from the analyses in Table 48.

A conversion of U_{CO + H2} = 65% was used, for a single stage, and this gave an output or products of 350 - 450 g./litre catalyst/day/ From each N.cu.m. gas converted, the following products are obtained (Table 49). The benzin was of poorer quality than that from the high gas recycle synthesis, but more wax than the amount stated could be produced if required.

Tests of this process have lasted three months, but after about two months, running trouble was experienced from deposits of ferrous carbonate which formed on the upper parts of the reactor walls and then fell off and interfered with the smooth working of the synthesis. It was thought that this trouble might be avoided by using shorter and wider reactors.

Reviewing the liquid phase synthesis, Lurgi point out the following objections:- It is very difficult to separate catalyst dust from the oil, and if this dust remains in it and is recirculated, acids are formed which attack the catalyst. There are also corrosion difficulties. On a large scale the very big mass of hot oil constitutes a fire hazard which is particularly dangerous in war time.

When the costs of the oil filtration and recirculation plant are considered, it is doubtful if the process is cheaper than the conventional, medium pressure, synthesis.

The Oil-Recirculation Process

Some of the advantages of the liquid phase synthesis can be retained and some of the disadvantages avoided if a catalyst is arranged in a fixed bad and oil circulated over it to remove the heat of reaction and the products. Such a method was developed by Dr. Diftschmidt, of the I.G. Farben, and the present opinion of the company is that this is the best method of conducting synthesis with iron catalysts, although it does not give quite such high yields as synthesis in the usual Ruhrchemie type of reactor. At the end of the war they were planning a 40,000 t./year plant for this process.

A melted iron catalyst was to be used in 8 - 12 mm. pieces, and reduced with hydrogen. The rate of oil recirculation used on the small scale scale was 550 - 750 litres/hour for a reactor of capacity 60 litres. The pressure was 20 atm. and carbon dioxide the main oxygen product.

It was pointed out that this process, in which the whole reactor is filled with oil is very different from that developed by Ruhrchemie in which a fixed bed of catalyst is sprayed with oil to keep the catalyst free from wax and to remove the heat of reaction by the latent heat of evaporation of the oil.

Direct Synthesis of Alcohols

Ruhrchemie gave the following details of a semi-technical scale test of the catalyst Fe 100 : Cu 5 : CaO 10 : kieselguhr 5: KOH 3, whose laboratory scale preparation and testing have already been described (BIOS Report 447, Appendix VI).

The catalyst was precipitated with sodium carbonate, half of the kieselguhr being added before and half after the precipitation. It was filtered on the filter press and washed for 30 minutes and then repasted with 70 litres hot per kg. iron and filtered. The repasting and filtering were repeated twice. After the last of these filtrations, washing was continued for 100 minutes. The filter-cake was impregnated with the 3% KOH in the extrusion press and then dried at 140°C. or less and the threads broken to 1 - 3 mm. pieces.

Pretreatment was done at 225°C. with 75% hydrogen ( the rest nitrogen) for 24 hours at a space velocity of 6,000/h., and this caused reduction so that 71.8% of the iron was soluble in 2% acetic acid.

Synthesis was carried out in a semi-technical scale plant with 16 full sized double tubes, using water-gas at a space velocity of 100/h., pressure 10 atm., temperature 200 - 214°C. and no recirculation. For the 750 hours of the experiment the conversion U = 60 - 65%, the utilization ratio X = 0.65 - 0.72, and of the gas converted, 14% went to form methane. The yield was 95 g./N.cu.m. and of this, about 15 g. was gasol. The composition of the products is shown in Table 50.

Synthesis at Atmospheric Pressure

Dr. Kolbel, of Rheinpreussen, gave the following particulars of his latest results in this field. The best catalyst had the composition Fe 100 : Cu 1 - 10 : kieselguhr 50 - 100 : K₂CO₃ 1/2 - 15, although attempts were being made to replace the kieselguhr by dolomite as in the other Rheinpreussen iron catalysts. It is prepared by dissolving the iron in nitric acid to give a 10% solution, adding the copper nitrate, boiling, adding the kieselguhr, precipitating with boiling 10% sodium carbonate solution, filtering and washing. The potassium carbonate is dissolved in water and is stirred into the filter cake which is then dried and granulated to 1 - 3 mm. pieces. It is not submitted to any process involving forcing through dies.

The success or failure of this catalyst depends on the pretreatment, which is done for 1 or 2 days at 230 - 235°C. with synthesis gas, H₂/CO = 2, which has been carefully freed from both carbon dioxide and water vapour, and is passed at 20 times the normal synthesis gas rate either directly or with recirculation (recirculation ratio 3), at atmospheric pressure.

Synthesis can either be done with water-gas using recirculation, or with a gas richer in hydrogen, approaching the composition of ordinary synthesis gas, without recirculation, and in both cases the results are essentially the same. The longest run so far was with water-gas, and lasted for 2 months at 218°C. without any sign of the activity falling off. No regeneration treatment was done during this period and it is estimated that the catalyst life will be at least 4 months, taking into account the fact the iron catalysts do not respond to regeneration treatments nearly as well as cobalt ones. The yield of products, including methane, is given as 130 - 135 g./N.cu.m. (CO + H₂) for one stage at a space velocity of 250/h., and for two stages at a total space velocity of 100/h. it is 160 g./N.cu.m. (CO + H₂), including methane, or 150 g. without methane. Half of the oxygen product is water, half carbon dioxide, and the utilization ratio X is about 1. It is rather higher in Stage I than in Stage II, and is not affected very much by the amount of alkali in the catalyst. The conversion of carbon monoxide, U_CO, is 85 - 90% for Stage I and 97% for the two stages together.

The products contain as high a proportion of wax as those from cobalt medium pressure synthesis, and the amount of olefines in the total products is very much higher, being 40 - 50%. The more alkali there is in the catalyst, the bigger the proportion of wax in the products; the total yield of products is however independent of the amount of alkali in the catalyst.

For synthesis at atmospheric pressure, the K.W.I. used the Fe 100 : Cu 20 : K₂CO₃ 1 catalyst whose preparation and pretreatment have already been described (see p. 122). Synthesis was done with water-gas at 205°C. with a space velocity of 50/h. or at 215°C. with a space velocity of 100 - 150/h. For either set of conditions the conversion was U_CO = 90 - 95% and the yield, including methane, 140 - 145 g./N.cu.m. inert-free gas. The methane amounted to 2 - 4 g. for the lower space velocity and to 5 - 6 g. for the higher one. This yield was maintained for 3 months by carrying out a solvent extraction every 2 or 3 days, and by allowing the temperature to rise 5° by the end of the 3 months. The formation of methane increases a little during this period.

The products contain some 30 - 40% wax, of b.p. > 300°C., and they vary in colour from yellow to bright brown due to resins formed from unsaturated products. More oxygen compounds are formed than in synthesis with cobalt catalysts, the amount being the same as in the medium-pressure iron Reichsattversuch experiments. The utilization ratio X = 0.6 so that no second stage is possible.