Return to FIAT 1267 Report Table of Contents

The CO - H₂ Synthesis at I. G. Farben A.G. - 1949

Dr. H. Zorm
Under the appeciation of
Dr. W. F. Faragher

Office of Military Government for Germany (US)

Foreword to

The CO - H₂ Synthesis

    The reports assembled in this publication were compiled by German technicians of the French occupation zone and represent the part played by I. G. Farben A. G. in the field of CO - H2 synthesis (Volume I).  They will supplement the reports on this subject made under the direction of Professor Dr. Friedrich Martin, former president of the Board Ruhrchemie A. G., Oberhausen-Holten-Ruhr, who brought together under his authority technologists of the Ruhrchemie, Lurgi, Brabag, and Rheinpreussen companies.

    In addition to the CO - H2 synthesis studied at Ludwigshafen, there appear in volume II the report of Dr. Winkler on the possibilities for practical application of his gas producer, and in volume III reports on polymerized gasolines.

    The initiative for this project belongs to Dr. W. F. Faragher, European Chief of the Synthetic Fuels and Lubricants Section of the American Field Intelligence Agency, Technical who with the cooperation of "Institut du Petrole de Paris" effected a Franco-American cooperation in collecting documents indispensable to the progress of science.

    The French team wishes to render him thanks for the energy and understanding he showed in the undertaking and prosecution of this vast project.

Jacques Foucher
Military Government of the French Zone
Enseigne de Vaisseau, Administration Branch
for I. G. Farben A. G. Factories at
Ludwigshafen-Oppau/Rhine

Volume I:

CO - H₂ Synthesis

Table of Contents

      A. The researches on CO - H₂ synthesis in the Gasification Section of the BASF Laboratory from 1925 to 1945 (June 6, 1947 by dr. Duftschmidt).

I.  The development of fused-iron catalysts (according to research by Dr. Lincckh and Dr. Klemm).

II.  The oil-circulation process (on the basis of research by Dr. Duftschmidt).

III.  The high-pressure gas-circulation process for the synthesis of ethyl alcohol from waste gases (on the basis of research by Dr. Linckh and Dr. Klemm).

    Appendix: Fused-iron catalysts for the CO- H₂ synthesis. (Report of Dr. Klemm, dated July 1, 1944).

    B.  Hydrocarbon synthesis from carbon monoxide and hydrogen.  Researches of the Ammonia Laboratory from 1935 to 1944 (by Dr. Arno Scheuermann, June 1947).

    C.  Hydrocarbon synthesis from CO - H₂ with iron catalysts.  Gas circulation process and foam process (by Dr. Michael, July 10, 1947).

Volume II

Winkler Generators

    The historic development of the gasification of fine coal according to the process of Dr. Winkler.

October 1947

Volume III

Polymerized Gasoline

Table of Contents

Report of Dr. Franz Duftschmid on the Researches on the CO - H2 Synthesis with Generator Experiments of the BASF from 1925 to 1945

1. The development of fused-iron catalysts (on the basis of researches by Dr. Eduard Linckh and Dr. Richard Klemm).
2. The oil-circulation process (on the basis of researches by Dr. Franz Duftschmid).
3. The high-pressure gas-circulation process for the synthesis of ethyl alcohol from waste gases (on the basis of researches by Dr. Eduard Linckh and Dr. Richard Klemm).

Appendix:  Fused-iron catalysts for CO - H₂ synthesis (report of Dr. Klemm, July 1, 1944).

                                                                                   Oppau, May 30, 1947.

Low Pressure Division, Oppau
Gasification Experiments Section
Dr. Fritz Winkler

Introduction to

Part I     CO - H₂ Experiments

    Research was conducted at the Badische #nilin- und Soda- Fabrik by Dr. Mittasch and Dr. Pier on the conversion of CO - H₂ to methanol (1921-1922) and by Dr. Karl Hochschwender on the conversion to isobutyl alcohol (1924-1925).  Experiments on the production of hydrocarbons only from CO - H₂, and in particular, of C₂H₄, C₃H₆, and C₄H₈ from CO - H₂ were started in January 1926 at the instance of Dr. Fritz Winkler by his collaborator Dr. Eduard Linckh.  A slight conversion of CO - H₂ to C₂H₄ etc, was found to occur with copper catalysts at ordinary pressure and at 100° C.  On April 1, 1926 Franz Fischer published (for the first time) an article in the Zeitschrift für Brennstoffchemie on the production of oils etc. from CO - H₂ at ordinary pressure with iron and cobalt catalysts.  Furthermore, Linckh discovered that silver, gold, and zinc catalysts act like copper catalysts (French patent 635950, English patent 293185), but that the yields were very small.  In the fall of 1926 he turned to the preparation of oils etc. from CO - H₂ under pressure with fused iron-oxide catalysts.  Linckh investigated more than 1000 catalysts, all in the gas phase under a pressure of 100-200 atmospheres.  Special arrangements of the catalyst in the reactor were tested for the purpose of conducting away the large quantities of heat evolved during the conversion of CO - H₂ to oils.  It was extraordinarily difficult, however, to avoid soot formation on the catalyst in the gas phase under pressure.  The solution of this problem was arrived at by Duftschmid: conversion of CO - H₂ with Linckh's catalyst in its own oil.

    Dr. Klemm conducted experiments on the preparation of ethyl alcohol from CO - H₂ under 200 atmospheres pressure in the gas phase.  He was able to avoid the deposition of carbon on the Linckh catalyst by copper plating a part of the catalyst, by arrangement of the catalyst in a thin bed at the tube wall (the reactor consisted of a bundle of tubes); and by using large quantities of circulating gas.

Dr. Fritz Winkler

The Development of Fused-Iron Catalysts for the CO - H₂ Synthesis

(On the basis of researches by dr. E. Linckh and Dr. R. Klemm)

    In 1925 Dr. F. Winkler brought forward the suggestion of making gaesous olefins from coal or water gas.  To this end Dr. Eduard Linokh had begun to conduct experiments in 1926 for the purpose of producing gaseous hydrocarbons with catalysts, in particular ethylene from water gas.

    At this time there already existed long-established BASF patents dealing with the preparation of liquid hydrocarbons along with oxygen-containing compounds with the aid of catalysts under high pressure.  Orlov was supposed to have obtained gaseous olefins with nickel catalysts.  The initiated experiments were conducted with this report in mind, but the information could not be verified.  However, Dr. Linckh soon succeeded in producing noticeable quantities of gaseous olefins under ordinary pressure and at 100° C. with copper catalysts for the first time.  An increased yield, however, could not be achieved.  The experiments were therefore extended to include iron catalysts.  Substantial quantities of gaseous olefins were thereby obtained under high pressure.  Final gases containing 13 per cent by volume of olefins were supposed to have been obtained.  Later, however, interest reverted to the recovery of liquid hydrocarbons.  As a result of these researches Dr. Linckh succeeded in developing serviceable fused-iron catalysts for the synthesis.  The technique of preparing these catalysts depends almost entirely on the production of the conventional ammonia synthesis catalysts.  Carbonyl iron powder was always employed as the starting material for the preparation of a catalysts.  It is intimately mixed with added substances having a promoting action and melted in water-cooled crucibles with a blast of oxygen.  The resulting molten lumps contain the iron in the form of oxides, principally as Fe₃O₄.  The molten cakes are broken up into suitable particle sizes.

    They are treated with hydrogen to prepare them for the synthesis.  This reaction takes place at temperatures ranging from 450 to 650 degrees.  The reduction temperature, however, can be still further lessoned by the use of high pressure.

    In this manner Dr. Linckh investigated more than 1000 assorted catalysts.  Unfortunately, the detailed results of these researches were lost as a result of the war and of the death of Dr. Linckh toward the end of the war.

    The catalysts covered by German patent 708512, November 14, 1935, French patent 812290, English patent 465668 have remained of lasting value for us. The catalyst identified by Expt. No. 997 became the standard catalyst for all our later technical development researches.

This catalyst is produced by admixing

25 grams of silicon powder

25 grams of titanium oxide

50 grams of potassium permanganate

50 grams of water

to 1 kilogram of iron powder.  In place of patassium permanganate the equivalent amount of manganous oxide, pyrolusite (MnO₂), or manganese powder and potassium hydroxide can be used without impairing the effectiveness of the catalyst.  If these catalysts are treated at 650 degrees, the reaction will proceed rapidly and, on completion of reduction and cooling, will be rather insensitive to air.  However, these catalysts are only sufficiently reactive for the synthesis at pressures of 100 atm. gage and over.  On the other hand, if the reaction is conducted at temperatures ranging from 450 to 500 degrees, the catalysts will develop a reactivity favorable for synthesis at medium pressures ranging from 15 to 25 atm. gage.  However, since the catalysts are markedly pyrophoric, special precautionary measures must be applied in this instance to prevent the catalyst from burning in air.  Transference in the synthesis reactors therefore takes place with exclusion of air in the presence of carbon dioxide, or in case of synthesis, in a liquid medium by soaking or dipping in oil.  Reduction at low temperatures possesses the additional disadvantage of a duration of 5 - 6 days.  The hydrogen is recirculated above the catalysts and is dried before admittance into the reaction vessels.  This drying is advantageously accomplished in a silica gel drying installation.

    Dr. Linckh endeavored to develop the iron catalysts to a point where the synthesis could be conducted at ordinary pressure with a good yield.  He believed that this was achieved by the addition of antimony or arsenic to the iron catalysts.  However, as a result of the war, practical tests on these catalysts were no longer carried out.

    Linckh made the interesting proposal that etched metal disks and the like be employed instead of catalysts (German patent 721359, French patent 805696, English patent 458035).

    The fused iron catalysts were developed further by Dr. Richard Klemm, to be used for a specific purpose, viz., to attain the greatest possible yields of ethyl alcohol.

    It is no longer necessary to abstract these researches at this point because a relevant original report can be appended (see report of Dr. Klemm, July 1, 1944).

The Oil Circulation Process

1.  General Comments on the Oil Circulation Process

    In the synthesis of hydrocarbons from carbon monoxide and hydrogen, large quantities of heat are evolved which must be conducted away from reaction chamber in a suitable manner.

    This problem had heretofore been technically solved in the Fischer-Tropsch process by dividing the catalyst into relatively thin beds and by charging it into narrow tubes and pockets.  The reaction heat is thereby indirectly conducted away by water under pressure.

    The use of iron instead of cobalt catalysts poses greater problems of heat removal because iron catalysts only react at temperatures higher than those of cobalt catalysts, the danger of spontaneous carbon monoxide decomposition consequently becoming appreciably greater.  In addition, the known iron catalysts react principally at elevated pressures thus further promoting formation of methane and carbon monoxide decomposition.

    These considerations led to the development of oil circulation process.  This process (discovered by F. Duftschmidt, E. Linckh and F. Winkler) and developed by the Fritz Winkler experimental team since 1935 at the I. G. Farbenindustrie at Oppau is differentiated from the Fischer-Tropsch process principally by the fact that the quantities of heat evolved in this markedly exothermic reaction are directly taken up and removed by a liquid medium located in the reaction chamber.

    It should be mentioned at this point that a second process was later developed by the I. G. Farbenindustrie at Ludwigshafen by Dr. Michael in which work was also done in the liquid phase with circulation of oil.  This is the so-called foam process, which is most clearly differentiated from the oil circulation process dealt with here, by the fact that finely divided catalyst suspended in a circulating oil is employed in the foam process.  This process was developed in Ludwigshafen since about 1940.  These two processes embody differing viewpoints in still another respect.  The Oppau oil circulation process employs oil evaporation for the removal of the reaction heat and is therefore characterized as a synthesis in a boiling liquid.  The Ludwigshafen foam process, however, substantially features a non-evaporating liquid medium, thereby representing a further development of the principle already proposed in 1928 by the I. G. Farbenindustrie A. G. (M. Pier) of conducting the synthesis in a high-boiling, non-evaporating liquid medium.

    The I. G. Farbenindustrie interests, and in particular the BASF works, immediately obtained the basic rights and also acquired technical leadership in this field by the development of the two processes.

    The present report deals only with the oil circulation process since the foam process has received separate treatment.

    The oil circulation process operates with a fixed granular catalyst.  As a result of the immediate removal of the heat it is possible to do without division of the catalyst into thin beds.  The synthesis reactor consequently acquired a much simpler shape without the necessity of any built-in installations.  In the indirect heat removal process the quantities of heat conducted through the wall of the catalyst chamber per unit of the time are quite limited.  However, this limitation is largely done away with in the direct heat removal method of the oil circulation process.  As a result a greater capacity can be achieved in the oil circulation process with regard to time and reactor space.  Also, power consumption in operating the oil circulation is slight and posseses hardly any importance in determining production costs.

    In 1936 and 1937 the process was developed to a semi-commercial stage (200 liter reactor), and run under high pressure (100 atm. gage).  In 1938 it was developed into a medium pressure process (15-20 atm. gage) and worked out as a two-stage process similar to the Fischer synthesis.  The first stage of the medium pressure process was tested in a 1.5 m³ reactor for 5 months with an output of 9 ton/month.  The second stage was successfully tried for a period of one month in a makeshift 0.2 m³ reactor (unfortunately a complete two-stage pilot plant was unavailable).

    The process was thus proved to be practical and safe.  The robustness of the catalysts is characteristically indicated by an experiment with the afore-mentioned 1.5 m³ reactor in which the catalysts were used without harm after the reactor had been in disuse for two years, the spent catalyst lying in the oil residue of the reactor during this time.

    Fast experimental results, however, justified plans of a reactor height of about 18 meters for large-scale units.  It was also planned to conduct the synthesis in one stage.

2.  The Principle of Heat Removal by Means of Oil

    The quantity of heat evolved during the conversion is computed according to equations

   2 CO + H₂ -------- CH₂ + CO₂ + 46.9 calories

CO + 3 H₂ --------- CH₄ + H₂O (steam) + 48.9 calories

to roughly 4000 calories per kilogram of liquid hydrocarbon produced if it is assumed that the conversion proceeds to 80 per cent according to the first equation with 20 per cent methane formation, i.e., if about 165-170 grams of liquid hydrocarbons are formed, including liquifiable gas, for every 1 m³ of converted CO - H₂, which is the case in the procedure dealt with here.

    The principle of working in the liquid phase consists in collecting this evolved heat by the oil itself and removing it.  This can be accomplished in various ways; by simply conducting away the heat in the heated oil in the form of sensible heat, or by removing it from the reaction chamber by evaporation of the oil in the form of latent heat of evaporation.

    In the first case the liquid medium consists of oil which is not vaporizable under the reaction conditions; in the second case it consists of oil containing vaporizable consituents under the conditions of the reaction, or containing an oil fraction which is vaporizable under these conditions.

    If the heat is removed only in the form of sensible heat, a rise in temperature within the reaction chamber will be the precondition for removal of the reaction heat.

    The quantity of oil to be put into circulation per kilogram of made products is calculated simply from the equation:

    If the quantity of final gas G - 2.7 m³ and the quantity of the made water W = 0.23 kgs. for 1 kg. of product at 90 per cent conversion are included in the above equations, the amounts of oil to be circulated are calculated as follows:

t	Q = kg 01/kg product
= 5°	1600
= 10°	800
= 25°	320
= 50°	156

    The quantities of circulating oil are appreciably higher in this case as compared to the second case since a part of the circulating oil is evaporated.

    If the quantity of circulating oil is divided into an evaporating portion Q1 and a non-evaporating portion Q2, we get the following equation:

The quantities of circulating oil can be calculated on the basis of this equation, as was done, for example, in the table for various  D t and various portions of vaporizable oil.

    The oil circulation process is carried out by the following procedures:

(Diagrams I and II on page 133o)

    In the procedure according to Diagram I the circulating oil is cooled within temperature range t₂ - t₁ before being returned to the reactor.  This procedure approaches that of the above-mentioned procedure in that the oil medium contains relatively few vaporizable constituents.  The extent to which this precondition is carried out depends on the temperature range  Dt and the quantity of the final gas.  It is evident that the boiling curve of the circulating oil cannot be freely chosen for even if relatively high-boiling circulating oils are employed, the oil will acquire a specific composition by condensation in apparatus K, depending on the working procedure.

    The procedure according to Diagram II, on the other hand, permit cooling in the cooling device K to the extent that hydrocarbons are condensed in a wider boiling range, depending on the selection of temperature t_o.  A lower limit for t_o is substantially conditioned by the fat that high-boiling paraffins are formed during the synthesis which naturally enrich the circulating oil.  It is expedient, in general, not to exceed a temperature of 80 - 100 degrees in order to avoid obstructions in the tube of the cooling device of the steam producer caused by solidifying of these paraffins.  It is also advantageous to cool below the strictly necessary temperature for low-boiling hydrocarbons would otherwise condense which evaporate on being reheated to t₁ before entering the reaction chamber.  The heat exchanger, which is a section of the cooling device and is used to reheat the circulating oil, would then be unnecessarily large.  The Oppau experiments which were conducted in order to develop the oil circulation process were always carried out according to the procedure shown by Diagram II.  However, it was planned to revert to Procedure I for new plant units.  In this connection, provision was to be made for the injection of lighter-boiling fractions, which were removed from the final gas in the form of condensates, into a supplementary circulation in the reaction chamber, as is shown by Diagram III:

(Diagram III on page 134)

3.  Catalysts

    After the granular fixed-bed catalyst was proved to be practicable, the oil circulation process was developed on this basis since this procedure shows considerable simplicity and distinction as compared with the catalyst suspended in oil.

    Certain minimum requirements had to be met by the catalyst with regard to firmness and durability when the oil circulation procedure was adopted.

    These requirements were fulfilled by the iron catalysts developed by Dr. E. Linekh, which were produced by melting iron with promoters in a current of oxygen into ammonia systhesis catalysts in the form of oxides (Fe_sO₄).

    The fused cakes thus obtained are broken up into 6-12 mm. pieces and the catalyst is reduced with hydrogen at 450-500 degrees for the medium pressure synthesis, and up to 650 degrees for the high pressure synthesis.  Deficient reduction of the pieces is to be avoided because a bursting of the nucleus, caused by a splitting off of carbon might occur.  The catalysts are substantially composed of iron, with slight additions of silicon, alkali, manganese, and titanium oxide.  The exact composition of catalyst No. 997, which is most often used as a standard in experimental work, is given in Part I, page 2 of this report.

    The catalyst reduced in a reduction furnace installation apart from the synthesis reactor is charged into the reactor after cooling in the presence of carbon dioxide, or, particularly if the catalyst is reduced at temperatures below 600 degrees and is very pyrophoric, it is discharged from the reduction vessel into oil in the absence of air and is charged into the reactor.

    The following figures are assumed in the plans for a catalyst  reduction installation:

Reduction temperature	470 - 500° C,
Hydrogen current	200 - 300 liter per hour kilogram catalyst
Bulk density of the oxidic catalyst	2.5 kg./liter
Reduction time	6 - 8 days

    The hydrogen is recirculated in the reduction installation.  The hot hydrogen issuing from the reduction vessels is first conduct into a heat exchanger in a countercurrent to the cold hydrogen which recycled to the reduction vessels.  After passing out of the heat exchanger the hydrogen is cooled as far as possible and then, after separation the condensed liquid, passed through a silica gel dryer for complete drying before being returned to the reduction vessels.

    Moreover, the given conditions are dependent on the type catalyst and its ease of reduction.  The reduction temperature can be raised to 650 degrees for conducting the synthesis under high pressure.  A substantial shortening of the reduction time is thereby attained.

4.  Procedure and Equipment

   After granular, fixed-bed catalysts in the reactor proved to be feasible, the oil circulation process was developed on this basis it offers a substantial simplification in contradistinction to the suspended in oil.

    An upright high pressure tube of 200 or 500 mm. internal and a height of 6 meters served as the synthesis reactor.  Apart from central tube for the thermo elements, the reactor had no built-in exchanger.

    The empty reactor chamber was charged with catalyst.  The synthesis gas runs through the reactor (S) from bottom to top along with the circulating oil in a direct current.  The hot oil and gas leaving the reactor surrenders its heat to the cold circulating oil in a heat exchanger (R).  After passing out of the heat exchanger the oil is again cooled in a condenser or a steam generator (K1) to about 120 degrees.  At this temperature the final gas is separated from the oil in a separator (A1).  The made oil excess and if necessary the reaction water is drawn off from the separator (R).  The oil is again conducted to the reactor (S) via the heat exchanger (R) and a supplementary pre-heater (W) by a circulation pump, the oil in the separator in the meantime being maintained at a specific level.  The supplementary pre-heater is merely used for starting.  The final gas is cooled to approximate external temperature in another cooler (K2), gasoline as well as alcohol-containing water being separated in another separator.  If the operation is carried out in two stages, then the final gas is conducted to the second stage which is coordinated with the first stage.  In the second stage the final gas is either depressured and conducted to the activated charcoal plant (AK) or washed with oil without being depressured so that the light gasoline and O₃ - C₄ hydrocarbons still retained in it may be recovered.

(Diagrams on pages 135 and 136)

    The fresh gas is also pre-heated in a suitable manner and admitted to the reactor from below.

5.  The Oil Circulation

    The oil produced in the synthesis itself is employed as circulating oil.  In this procedure, as distinct from similar procedures, the use of high-boiling fractions, which remain almost completely liquid under the reaction conditions, is deliberately given up.  It is desirable for higher performance that vaporizable constituents be present in the circulating oil under the reaction conditions if possible.  The boiling of the liquid in the synthesis reactor creates the conditions, on the one hand, for good contact of the synthesis gas with the catalyst, thus achieving a higher capacity, and, on the other hand, for favorable heat removal and temperature regulation.

    The boiling range of the circulating oil may be regulated as desired by the condensation temperature of the separator (A) placed in the oil circulation.  Two examples are cited as typical for the boiling range of the circulating oil:

        (Table on page 137)

    The reaction can safely be kept at the desired degree of conversion and at the desired temperature by regulation of the quantity of recirculated oil.

    The quantity of circulating oil is increased with the charge rate to the reactor.  In the case of the plant units hitherto operated the quantity of circulating oil was adjusted for 70 - 80 liters per hour per kilogram of made product in the same period.  In the case of larger units, where heat losses are of less importance, the quantity of circulating oil could probably be adjusted to a higher level.  Future plans call for a quantity of circulating oil of 120 - 150 kg./hour per kg. of product.

6.  Evaluation of the Experimental Results of the Plans for a Large Scale Reactor

(Graph on page138)

    It is evident from the two curves representing the temperature course and the distribution of the conversion in the experimental reactor that the procedure is especially more favorable for larger reactor units.  It is clear that the temperature t₁ in the lower part of the 6 meter long experimental reactor was not very much in excess of the temperature at the start of the reaction, and that conversion in the lower part of the reactor therefore sets in very slowly, increasing by degrees with the gradual rise in temperature as a result of the slight conversion.  Thus, the lower sector of the reactor was poorly utilized.  Just as soon as temperature t₂ (sic) is reached, conversion also attains a corresponding increase.  After temperature t_y near the upper end of the catalyst bed is passed, the conversion undergoes a similar rise with rapidly increasing temperature.  This means that the proportion of vaporizable fractions present in the range of temperature is too small to conduct away satisfactorily the best by evaporation.  Since the boiling point distribution of the circulating oil is not subject to any arbitrary influencing, the process has therefore been expanded to this end ^(x) so that vaporizable "cool oil" is fed to t_y in order to check this undesirable temperature rise.

    An experiment with a cold feed of this kind conducted in a 200 mm. reactor achieved the desired success.  Since another oil pump was unavailable, this operation could only be carried out by a makeshift branching off from the main oil circulation with valve regulation.  It was possible by careful attention to arrest the undesired temperature rise in the upper end of the reactor to the desired extent.  However, only an approximate 55 per cent conversion of the gas was attained because of the low (6 meter) height of the experimental reactor.

    In planning a large-scale reactor unit the improvement of the process outlined in Diagram III, page 5 is offered, the temperature cou? being shown in the following chart:

(Chart on page 139)

    The temperature range  D t, which was previously maintained at 50 degrees, and often at 100 degrees, in order to attain a 50 per cent conversion in the 6 meter reactor, can be reduced to 25 degrees to insure the same degree of conversion with a reactor height of 9 meters.  With reactor height of 18 meters it will be possible to carry out the entire 50 per cent conversion in one stage.  It cannot be decided here whether the one-stage or the two-stage process is economically more advantageous for large scale production.  The procedure which was conducted in the ? plant up to 75 per cent conversion only produced a specific yield of 140-150 grams in the second stage as compared with a yield of 165-170 grams in the first stage.  This difference in yield should have decreased as a result of the equalization of the reactor temperature and the low-peak temperature at the reactor exit which could be accomplished because of the better control of the process possible in a large-scale reaction.  An increase in yield can therefore be counted on in large-scale reaction. However, in the following sections only those yield figures are given which were actually obtained in past experiments.

7.  Synthesis Gas

    In general a CO:H₂ ratio of 55:45 or 1:0.8 is necessary for the synthesis.  Gases rich in hydrogen can also be used; however, they cause a shift in the composition of the product with proportionate increase of the hydrocarbons boiling in the gasoline boiling range and the C₃ and C₄ hydrocarbons.

    The following table assembles the results of comparative experiments in which the CO:H₂ ratio was modified, conditions in other respects being the same.  These comparative experiments were carried out in pressures of 15 to 20 atm. gage.

(Table on page 140)

    The table shows the decrease in C₃ and C₄ formation with rising CO content of the synthesis gas and the simultaneous increase in the formation of high-boiling products.

    The gas rich in CO which is necessary for the synthesis can be directly produced by oxygen gasification in Winkler generators.  The desired CO:H₂ ratio can also be attained in the customary cyclic water gas production from coke, if a part of the steam is replaced by CO₂.  A gasification medium consisting of 27 per cent by volume of carbon dioxide and 73 per cent by volume of steam is needed to attain a CO:H₂ ratio of 55:45 or of 1:0.82 according to experiments with Pintsch-Drehrost (rotary grate) generators.  The water gas will then be composed of 12.7 per cent CO₂, 47.7 per cent CO, 39.0 per cent H₂ and 0.2 per cent CH₄.  The synthesis itself produces the CO₂ necessary for gasification if the resdual gas put through a CO₂ scrubber.  This quantity of CO₂ is sufficient to produce the necessary quantity of synthesis gas rich in CO.

    A final gas good for heating purposes is simultaneously obtained with the CO₂ scrubbing of the residual gas of the synthesis.

    A gas rich in CO can also be produced from natural gas by using the carbon dioxide resulting from the synthesis process.  It should be mentioned at this point that most of our experiments were conducted with synthesis gas poor in inert gases (2% N₂).  However, we conducted a number of experiments with gases containing a higher proportion of inert gases (13%).

    In the following example the composition of the final gases the two-stage procedure are given with a synthesis gas free of inert gas.

(Table on page 141)

    The composition of gases when gases containing inert gases are used can easily be calculated from these figures if it is assumed that inert constituents take no part in the conversion.

    The synthesis gas must be purified from inorganic and organic sulfur.  According to our experience the catalysts are not as sensitive as in the Fischer process.  Fine purification to a sulfur content of 4-5 mg/m² should suffice.

8.  Pressure

    The medium pressure synthesis is conducted under 15-25 atm. gage.  The results obtained at this high pressure have been cited in the individual sections of this report in comparison with the results of the medium pressure synthesis.  Apart from the high capacity attainable, however, these results have little commercial significance since the compression costs cannot be compensated for by increased yields or by improvement of the properties of the product.

    Pressures up to 200 atm. gage are briefly referred to at the conclusion of the report.  These were used, however, in the formation of oxygen-containing products.

9.  Performance

    The medium pressure synthesis is capable of a continuous hourly output of 20-25 grams total product per liter catalyst.

    The hourly output of the high pressure process was stepped up to 160 grams total product per liter.

10.  Temperature

    The temperature in the reactor rises from the bottom to the top.  It is conditioned by the oil circulation.  A temperature rise of 50 degrees C. was established in general in the experiments conducted here so that the following temperatures were recorded with maximum reactor charge rate:

(Table on page 142)

        However, as was mentioned before, an extensive temperature equalization is possible by using a more advantageous procedure, a temperature difference of 15 to 20 degrees C. being possible with a reactor unit temperature of 250-255 degrees C. in larger reactor units.

11. Yield and Conversion

    145 - 150 grams of total product per Nm³ CO - H₂ were obtained on medium pressure reactors with a 90 per cent conversion in two stages.

    This total product consisted of

        13 - 15% liquifiable gaseous hydrocarbons

        32 - 35% gasoline fraction, boiling range to 175° C.

        18 - 20% middle oil fraction, boiling range to 175 - 320°

        10 - 12% soft paraffin fraction, boiling range 320 - 400°

        16 - 18% hard paraffin

        5 - 6% water-soluble alcohols

    The following page shows a chart giving conversions for two-stage conversion.  The results are converted to an oxygen-free synthesis gas (ideal gas).

(Chart on page 143)

12.  Properties of the Product

    The products have a marked olefinic character.  The olefin content drops with increasing molecular size, as is shown by the table:

(Table at top of page 144)

    The oils also contain oxygen-containing constituents (alcohol esters).  The oxygen content of the low-boiling gasoline fractions up 1? 100 degrees runs to about 1.5 - 2.0 per cent.  However, it can exceed 1? amount if the water-soluble alcohols produced are separated out in a concentrated form with the gasoline fraction, for the gasoline partially dissolves these water-soluble alcohols (C₂H₅-OH).  However, these water-soluble alcohols can be removed from the gasoline by treatment with water.  The oxygen content declines with rising boiling point to 0.5 per cent in the fractions ranging from 200 - 400 degrees C.

    The primary gasoline has an octane number of 68 (research) which is increased to 85 by addition of 0.1 per cent lead.  After hydrogenation the octane number suffers a considerable drop.

    The middle oil fraction yields Diesel oil directly.  The cetane number is 75 - 78.  A middle oil test gave the following values:

        The hard paraffin has a melting point in excess of 90° C. and can be used for all practical purposes (emulsifying wax and the like).  The hydrocarbon products are branched to a certain extent.

(Tables on pages 145 and 146)

13.  Results of Experiments with Synthesis Gases Having an Appreciable Excess of Hydrogen

    The influence of an appreciable excess of hydrogen on the carbon monoxide has already been mentioned in the previous section on "Synthesis gas".  The results of the experiments will be dealt with here in greater detail.  Circulating depressured gases of the methanol or isobutyl alcohol synthesis were used as the synthesis gas.

(Table follows on page 147)

    The carbon monoxide is extensively converted when these hydrogen rich gases are used; however, the greater part of the hydrogen remains the final gas.

    The working up of such hydrogen-rich gases frequently obtain as residual gases of other syntheses under pressure might very well considerable commercial interest although the yields with respect to ? of fresh gas would be relatively low.  However, since only the CO are portion of the H₂ is withdrawn from the synthesis gas, the CH₄ conte? the final gas being increased, commercial application of the synthesis be assured in many instances if the residual gas is worked up in a ? decomposition plant and again made available for other syntheses (N₂ ? synthesis).  The yields in the above table are given as specific yields m³ of converted CO - H₂, in order to facilitate comparative evaluation.

    It can be seen that water-soluble alcohols, consisting of  ?% ethyl alcohol, are principally formed under high pressures addition, the formation of C₃ and C₄ hydrocarbons is very great.

    The formation of ethylene is not included in the yields c? above although ethylene formation in individual experiments has reached a point where it can no longer be neglected.

    In the following the example of Experiment 35a is subject to more intense consideration:

Procedure according to Experiment 35a

    In the large-scale experiment conducted under a working p? of 15 atm. gage (one-stage conversion), 1000 cbm. of isobutyl recycle gas, containing 200 m³ of CO and 640 m³ of H₂, resulted in conversion of 130 m³ of CO and 165 m³ of H₂, with 770 m³ of residual gas containing 70 m³ of CO and 475 m³ of H₂.

        The composition was

        12.5 % CO₂

        1.9 % CnH₂n

        9.1 % CO

        61.4 % H₂

        9.0 % CnH₂n+2

        6.1 % N₂

        From this 56.1 kgs. of product were obtained, composed of

        7.8 % ethylene

        12.9 % propylene

        6.8 % butylene

        19.5 % propane and butane

        15.0 % light gasoline

        30.0 % gasoline and oil

        5.6 % ethanol

        2.4 % methanol

    The synthesis was carried out in one stage under medium pressure of 15 atm. gage in such a manner that for 5000 m³/hour of isobutyl recycle gas a reactor unit with 12 m³ of catalyst space was used.

    Catalyst life was estimated at one year.  Not the slightest evidence of catalyst fatigue was observed in an experimental period lasting one and one-half months, using hydrogen-rich synthesis gas under a pressure of 20 atm. gage.

    If 5000 m³/hour = 43.2 million m³/year of isobutyl recycle gas is passed through a synthesis installation of this kind, then 3840 m³/h = 33.15 million m³/year of residual gas will again be available after the synthesis and removal of the products in an activated charcoal plant.  This can again be worked up for the NH₃ synthesis according to the decomposition process of Dr. Sachse.

    From the conversion of the 1475 m³/hour CO-H₂ withdrawn from the isobutyl recycle gas, 2400 tons/year of the above-mentioned composition are obtained.

Procedure according to Experiment 25c

    Experiment 25c can be treated similarly to the example of Experiment 25a.  A working pressure of 150 atm. gage was used, the other working conditions remaining the same.  If 1000 m³ gas containing 922 m³CO-H₂ are converted according to the oil circulation process, 777 m³ of residual gas are obtained.

        The analysis follows:

        12.9 % CO₂

        0.9 % CnH₂n

        6.2 % CO

        72.3 % H₂

        4.7 % CnH₂n+2

        3.0 % N₂

A product of 43.7 kg. was obtained, consisting of 17.70 kg. water-soluble alcohols (of which 12 kg. are ethyl alcohol 11.05 kg. C₃ and C₄ hydrocarbons (about 40 per cent olefinic) 14.95 kg. oxygen-containing oils Moreover 0.6 per cent by volume, i.e., 5.8 kg. of ethylene are contained in the residual gas which is not counted in the yield.

      383 tons of product per year would be obtained if 1000 m³ of synthesis recycle gas were worked up hourly.  777 m³ of residual gas of the above-mentioned composition would be delivered hourly to the synthesis decomposition plant for working up.

14.  Experiments in the Production of Oxygen-Containing Products

    The experiments described in the above section lead to the formation of oxygen-containing products.

    When the researches of Ruhrchemie on the Oxo reaction became known at the start of the war, the question was universally raised whether such products could not also be obtained in the CO - H₂ synthesis in one operation.

    In 1940 we received the assignment from Dr. Müller-Contadi of investigating the preparation of higher-molecular weight alcohols by the oil circulation process.

    At the same time investigations were initiated in the Leuna Works which lead to the discovery of the Synol synthesis.  The actual circumstance that led to the Synol synthesis was the fact that a method of treatment of fused iron catalysts was discovered which permitted lowering of the synthesis temperature to 195-215 degrees.  In this connection it was shown that the low synthesis temperature was the decisive factor in obtaining higher-molecular weight alcohols.

    The experiments which we were conducting at Oppau at that time assumed that elevated pressure and an excess of hydrogen in the synthesis gas was the most important factor in guiding the synthesis toward the formation of oxygen-containing products.

    The experiments dealt with in the preceding section confirm the conception to the extent that increasing yields of water-soluble alcohol are obtained with hydrogen-rich synthesis gas and rising pressure.  Both factors, however, forced the synthesis into the formation of low-molecular products.

    We have already reported in the previous section that we obtained products relatively poor in oxygen when synthesis gas rich in carbon monoxide was used.  These were the conditions under which we initially developed the hydrocarbon synthesis according to the oil circulation process at 1? atm. gage in 1935.  We had already turned to experimentation with the o? circulation process at 200 atm. gage in a pilot plant at Leuna.  The re? of this pressure rise to 200 atm. gage was only a further shifting of the products toward the short chain lengths and an increase in methane formation.  Besides, a very rapid catalyst fatigue was observed at high pressure.  In 19 (illegible) we made the observation that we actually arrived at ?ed oxygen-containing compounds with gas rich in carbon monoxide under high pressure if we reduced the fused iron catalysts at 450 - 500 degrees C. instead of the previous 650 degrees for the high pressure procedure, as was already done in the medium pressure synthesis.

    The following considerations were adopted in our experiments in the production of higher-molecular alcohols:

1.  The reduction of the fused iron catalysts was to occur at the lowest possible temperature.

2.  Conversion in the synthesis should be divided into more than two systems.

3.  The catalyst space velocity should be kept as low as possible.

4.  Overheating in the oil circulation must be avoided to protect any higher-molecular weight alcohols from decomposition.

    All of these four considerations attempt to lower the synthesis temperature as much as possible in order to counteract the tendency toward formation of lower-molecular weight compounds as the result of high pressure.

    The application of these conditions had the desired effect, to some extent, as is shown by the appended experimental date.  It is beyond doubt, however that our results have been surpassed by the Leuna Synol process and our experiments outstripped.  The progress of the Launa process is based on the fact that, with catalysts which can be used at 200 degrees C. and middle pressure, the molecule-reducing use of high pressure can be avoided, since the formation of oxygen-containing compounds can also be attained at these low synthesis temperatures without excessive increase in pressure.

    The application of this Leuna know-how to the oil circulation process was obvious.  An experiment was also conducted in which our fused iron catalysts were reduced with a great excess of hydrogen according to the experience of the Synol synthesis.  However, it was found out that an excess of hydrogen is not effective in catalyst reduction if the particle size exceeds 2-3 mm.  In any case an experiment conducted with a particle size of 8 mm. did not have the desired effect of lowering the synthesis temperature to about 200 degrees.  It was not possible to conduct further experiments because of the progress of the war.

    In this connection we present the results of one of the many experiments in the preparation of higher-molecular weight alcohols according to the oil circulation process.  In this experiment the same catalyst was used that was employed in most of the earlier experiments in the production of hydrocarbons.

    A report on experiments with other fused iron catalysts which according to the results of the investigators, is especially favorably the formation of oxygen-containing products, is unnecessary since the results of the experiment are basically the same.

Procedure according to Experiment 25

    A specific yield of 195 grams of product per 1 Nm³ of conve? product was obtained, using synthesis gas (CO=H₂ = 1.16), an hourly output of 40 grams of product/1 catalyst and a degree of conversion of 3? per cent of CO - H₂ in the first stage.  Temperature 236 - 260 degrees pressure 180 atm. gage.  The product consisted of

24.0 % C₃ and C₄ hydrocarbons

21.6 % oil boiling to 150 degrees

12.7 % oil boiling at 150-250 degrees

9.0 % oil boiling above 250 degrees

32.7 % water-soluble oxygen compounds (C₂H₅OH).

        A detailed investigation conducted by Dr. Leithe led to the following approximate composition of the total product:

8.5 % methanol

21.0 % ethanol

10.0 % propanol

6.5 % alcohols C₁ - C₁₂

2.5 % alcohols C₁₂ - C₂₀

11.5 % water-soluble fatty acids

5.0 % fatty acids C₅ - C₁₁

1.5 % fatty acids C₁₂ - C₂₀

26.5 % hydrocarbons to 200 degrees

3.5 % hydrocarbons 200 - 300 degrees)   about half olefinic

3.5 % hydrocarbons over 300 degrees)

List of Patents for the Oil Circulation Process

    Diagrams I and II (referenced from page 15)

The High Pressure Gas Circulation Process for the Synthesis of Ethyl Alcohol from CO - H₂ - containing Waste Gases

    The experiments of Dr. Eduard Linkh in the prepararation of hydrocarbons from CO - H₂ under elevated pressure with fused iron catalyst furnished the groundwork for this process.  In this connection appreciated quantities of reaction water were formed containing water-soluble alcohol especially ethyl alcohol.  The experiments were continued by Dr. Klemm in order to increase the proportion of alcohols.  At the same time Dr. Klemm initiated experiments for the study of fused iron catalysts and to de? new catalyst compositions which would especially favor the formation water-soluble alcohols.

The Gas Circulation Process

    The conversion of carbon monoxide-hydrogen mixtures with fu? iron catalysts under pressure necessitates special preparations in order to avoid an excessive reaction in the direction of methane or soot formation.  If such measures are not taken, the reaction will be concentrated into catalyst layer situated near the gas entry to the accompaniment of ma? heating.  As a result of the high reaction heat the reaction continue to proceed rapidly in the narrowest space and spontaneously leads to carbon monoxide decomposition and to loss of control of the reaction temperature.

    After several experiments the following measures were taken to control the reaction:

    1. The catalyst is arranged annularly in a thin bed inside of a tube.  The outside of the tube is jacketed with water under pressure in order to surrender a portion of the reaction heat to the water through the tube.

    2. The synthesis gas is recirculated so that the carbon monoxide concentration in the circulated gas does not exceed 10 per cent by volume.

    3.  The catalyst at the gas entry side is copper plated in order to diminish the danger of soot formation.

    4.  The recirculated synthesis gas is conducted throughout the entire length of the catalyst bed but mainly passes the space free of catalyst in the middle of the tube so that it only sweeps along the annular catalyst, partially penetrating it.

    The ration of the quantity of circulating gas to the fresh gas was 3 : 1.

    20 - 25 per cent of the fresh gas was constantly released from circulation as final gas.

    In general, the circulating depressured gas of the isobutyl alcohol synthesis was used as fresh gas; however, other gas mixtures were experimented with.  The experiments were conducted at a pressure of 180 atm. gage and a temperature of 290-320 degrees.

Equipment

    The reactor consisted of a high pressure tube of 120 mm. internal diameter with a built-in seamless boiling tube of 82 mm. internal diameter.  Water, activated by a thermosyphon, was circulated in the outer annular space between both tubes.  The inner boiling tube contains a central wire net hose which keeps the middle of the tube free from catalyst for the catalyst is charged into the annular space between the boiling tube and the hose.  The diameter of the bed was 10 - 12 mm.

    The synthesis gas enters the top of the reactor, passes from top to bottom of the central boiling tube containing the catalyst and conducted to a hot separator via a small separator, a heat exchanger and a hot cooler.  It then passes through a "cold cooler", the cold separates and is washed by oil.  The circulating gas is returned to the reactor by means of a gas circulating pump after depressuring a part of the circulated gas and adding make-up gas.  Before entry into the reactor the circulating gas passes through a pre-heater and finally the heat exchanger where it is heated to the reactor entry temperature by the hot circulating gas.  The depressured gases of the compressed oil wash, as well as the circulated depressured oil, were led through an activated charcoal plant in order to recover highly volatile hydrocarbons.

Catalysts

    Fused iron catalysts that were obtained by melting the following mixture in a current of oxygen were used in the pilot plant:

89.5 % iron powder

2.24 % titanium dioxide

2.24 % silicon powder

2.24 % manganous oxide

2.03 % copper powder

1.80 % KOH dissolved in some water

    If the mixture is carefully added to the melt the copper will burn well and will enter the melt.  On the other hand, when copper o? was employed it was reduced to metal and separated out in the form of molten metallic globules.

Reduction of the Catalysts

    The oxidic fused catalyst cakes were crushed to a particle size of (sic) mm. and then reduced with hydrogen at 650 degrees C. for 3 days.  A hydrogen quantity of 500 - 700 liters is used hourly for every liter of catalyst.

    The reduction temperature of 650 degrees is the most favorable.  Higher temperature markedly impairs catalyst activity.  A lower temperature during reduction means greater activity and shows that the synthesis takes place at low temperatures.  In this connection, however, the formation of water-soluble alcohols diminishes, and the formation of oil increases.

Copper Plating of the Catalyst

    It was observed over and over again when working in the gas phase that more or less marked formation of soot readily occurred at the side of the catalyst charge turned to the gas entrance.  To correct this trouble, which causes catalyst decomposition, copper-plated catalyst was charged into the upper sixth of the reactor.  Copper-plating of the catalyst is accomplished by treatment with a weak (5%) nitric acid copper nitrate solution.

    In this manner soot deposition was avoided without impairment of activity.

Synthesis Gas

    a. Type of Gas   

        Three different types of fresh gas were used in the various experiments.

    1.    A mixture of 20 per cent pure CO and 80 per cent zero (Nullgas) gas (2 per cent N₂).

    2.  Butyl residue gas enriched to about 45 per cent with pure CO; rediluted to 20 - 22 per cent CO by addition of zero gas (2 per cent N₂).

    3.  Pure butyl residue gas (6 - 6.5 per cent CO₂, 0.5 - 1 per cent HnH₂n, 20 - 22 per cent CO, 57 - 64 per cent H2, 5 - 6 per cent CnH₂n+2, 5 - 6 per cent N₂.

    In these investigations an approximately constant proportion 20 per cent CO was maintained so that differences in result could not attributed to divergencies in this constituent.  The hydrogen proportion showed appreciably greater differences.  In the mixture of CO with zero gas it was 80 per cent.  In the case of the butyl (Tanol) residue, gas treated according to 2., it was 70 per cent, and in the case of pure butyl (Tanol) residue gas, it was 60 per cent.  This gives a ratio of CO : H ? = 1 : 4, 1 : 3.5, 1 : 3.

    The third important difference lies in the "inert gas content" particularly with saturated and unsaturated hydrocarbons.  In this instance the butyl residus gas clearly has the greatest proportion.

    According to the experimental results there is no noticeable difference if the mixed gas consisting of CO - H₂ or the rediluted, er? Tanol residue gas is used.  However, the use of pure Tanol residue gas immediately resulted in appreciably better yields.  This was shown with ?cular clarity in an experiment in which rediluted gas was used in the period, iso-gas in the second, and rediluted gas again in the third.  The first period yielded 64.35 grams/m³ fresh gas product, the second 82.4 grams/m³, the third 60.75 grams/m³.  This increase in yield which occurred in several experiments is explained by the proportion of low-molecular carbons and similar compounds which were then synthesized or built up larger molecules.  It is not clear to what extent  the CO² and methane participate in the conversion.  It can be said in general that the use of 1? Tanol residue gas is most advantageous.  In addition, there is the omition  of compression since the gas can be used with the obtained pressure.

    b. Space Velocity

    The throughput of fresh gas is of further importance in increasing the yield.  This is shown by several experiments.  Calculations are for a space velocity of m³ per liter of catalyst per hour.  Three  experiments were conducted with a special catalyst arrangement ("lower bed installation" with 50 cups") with Tanol flue gas under similar conditions.  The use of 0.65 m³/liter catalyst resulted in 93.3 grams/m³ fresh gas; 0.75 m³/liter gave 62.28 grams.  Another experiment, conducted with mixed gas consisting of CO-zero gas (Nullgas), resulted in 58.25 g/m³, using 2.15 m³/liter of catalyst in the first period; a space velocity of 3.4 m³/liter in the second period gave only 41.54 grams.  Experimental conditions were almost the same.

    c.  Pressure

    Increase in pressure acts in two directions.  Under similar conditions the yield increases with rising pressure and the tendency toward formation of hydrogen-containing products simultaneously increases.  An experiment with a bundle of tubes with circulation of water, using a space velocity of 3.4 m³/hour, gave 41.94 grams/m³ at 100 atm. gage, and using a space velocity of 3.7 m³/hour gave 44.08 grams/m³ at 180 atm. gage.  In the first case the yield consisted of 30.59 per cent of alcohol; in the second case it consisted of 84.5 per cent of alcohol.

    d.  Circulating Gas

    The quantity of circulating gas is kept as high as necessary, to insure that the entrance gas does not exceed 10 per cent of CO on an average by dilution with fresh gas.  Larger quantities of circulating gas are too large to pass through and produce yields that are too little.  As a result of increasing enrichment of the CO in the incoming gas the conversion is altered in favor of the oil content, and finally the reactor tends to get out of control.

    In addition to removing a part of the heat of reaction the circulating gas should also act as a safety gas.  It is assumed that the first stage of the conversion produces oxygen-containing compounds, in particular alcohols.  These compounds, which call for speedy removal, are safeguarded by dilution with non-reactive or slightly reactive compound.  A decrease in the quantity of circulating gas led to an increase in hydro-carbons and similar compounds.

    With respect to the formation of alcohols, it was shown to be advantageous to free the circulating gas of volatile compounds by the ? wash under pressure before returning to the reactor since these compounds again participate in the formation of hydrocarbons and other similar substances during the reaction.  The total yield was higher if it was not ?

    e.  Final Gas

    The final gas carries with itself a greater or lesser amount highly volatile hydrocarbons, particularly liquefiable hydrocarbons, depending on pre-treatment.  It is therefore purified with charcoal, consisting afterwards essentially of hydrogen, methane and nitrogen, along with a percentage of CO₂ and some non-condensable hydrocarbons.  The remainder of the unconverted CO is also contained in it.  It is available for further working up for fuel purposes.  The remainder is about 30 - 40 per cent the volume of the fresh gas.  The quantity of substances which can be removed by active charcoal ranges between rather wide limits, between about 20 grams/m³ to about 80 grams/m³, depending on conditions.

    5.  Products

    The following products are obtained by the synthesis in the gas phase:

    a.  Oils from the cold separator.

    b.  Light oils as distillates from the wash oil.

    c. Activated charcoal gasoline.

    d.  Liquefiable gas.

    e.  Alcohols from the cold and hot separator.

    A total yield of 53.31 grams/m³ of the following composition was found as means value in the 7 large-scale experiments:

Oils from the cold separator	-    12.61	g/m3	-    23.28 %
Light oils from the wash oil	-    3.14	"	-    5.62 %
Activated charcoal gasolines	-    6.56	"	-    12.64 %
Liquefiable gases	-    4.2	"	-    8.49 %
Alcohols	-    25.58	"	-    49.47 %
	58.31	g/m3	- 100%

        In this connection fluctuating small quantities of semi-solid, paraffin-containing products, which may contain up to 40 per cent of higher alcohols, are obtained from the hot separator.  Because of their small quantity they are calculated as part of the oils from the cold separator.

    These fresh, light-colored oils range from colorless to light yellow.  They turn brown if kept in the presence of air.  They have a pungent odor similar to esters of alcohols.  Mean density 0.771.  Flakes of small amounts of higher hydrocarbons often float in it.  Boiling begins at about 45-50 degrees.  90 - 95 per cent is distilled at 300 degrees, the residue consisting of dark brown, semi-solid paraffinoid masses.  The acid number is 10-12; saponification number 30-35; OH number 120-140.  The oils still contain appreciable quantities of water-soluble alcohols which can be scrubbed out in quantities of 18-25 per cent by stirring with water.  ? per cent of the oil is evaporated at 90 degrees during distillation.  ? distillate is divided into two parts.  The lower layer (8-10 per cent ) with a density of 0.896 contains about 75 per cent alcohol in water. ? upper layer with a density of 0.720 contains about 23-25 per cent of alcohol (water-soluble).  The residue of 48-50 per cent of the crude is dark brown and contains about 5-6 per cent water-soluble alcohol. ? was possible by careful distrillation to recover the bulk of the alcohol, especially on adding slight quantities of water.  The light distillation corresponds substantially to activated charcoal gasoline.  The remainder consist principally of fractions corresponding to illuminating oil and ? oil.  The residue consists principally of paraffins with a V.S. of 16?.  Together with oils of similar origin the oils can be worked up to gas illuminating oils and liquefiable gas.

    B.  Light Oils from Wash Oil

    A light colorless distillate (d = 0.680 - 0.702) is obtained from the wash oil of the oil scrubber.  The distillate stands between oil from the cold separator and the activated charcoal gasoline, but resembles the latter a great deal.  It is therefore mixed into that tar?.  Boiling starts at 30-35 degrees.  This fraction which contains about ? per cent of the yield can be worked up with activated charcoal gasoline.

    c.  Activated charcoal gasoline

    A condensate with a density of 0.660-0.680 and composed of 85-95 per cent of low-boiling, paraffinic and olefinic hydrocarbons if obtained when the activated charcoal towers are purged with superheated steam.  This amounts to 10-15 per cent of the yield.  Boiling starts 30-35 degrees and 40 per cent is evaporated at 45 degrees.  This distillate forms two layers: the lower layer with a density of 0.962 is composed of 55-60 per cent alcohols and amounts to about 1 per cent of the gasoline; the upper layer (38 per cent) with a density of 0.670 is composed of 16 per cent of alcohols and the distillation residue (d=0.722) another 13-14 per cent of alcohols.  The crude gasoline contains roughly 17 per cent alcohol.  After preliminary hydrogenation this product can be worked up with other gasolines.

    d.  Liquefiable Gases

    The liquefiable gases, the C₃ - C₅ hydrocarbons, are driven off along with the activated carbon gasoline.  They form 8-10 per cent of the yield and, according to current analyses, contain about 90-95 per cent olefinic hydrocarbons.  Since no compression is available, they have heretofore not been recovered but their presence merely determined in a part of the residue gases.  The bulk of the liquefiable gas (about 90 per cent) is contained in the let down gases of the pressure scrubber along with a large part of the activated charcoal gasoline.  In a large plant provision could be made for continuous fractional distillation of the wash oil with fractionated depressuring.  The final gasolines and the liquefied gas could be continuously obtained under pressure in a liquified form.  It is planned to return the liquefiable gases to the process by injection.

    e. Alcohols

    The product alcohol is obtained in amounts of 45 - 55 per cent of the yield, in individual cases up to 60 per cent.  The bulk is contained in the water from the cold separator which consists of 45-50 per cent alcohol.  A solution of 3-10 % alcohol is obtained from the hot separator.  The alcohols are pre-concentrated to a concentrate of 90 per cent.  The crude alcohol contains about 10 per cent water.

Acetaldehyde                2 - 3 %

Acetone                        3 - 5 %

Methanol                        10 - 15 %

Ethanol                            60 - 70 %

Higher alcohols                10 - 15 %

    Since the separation of aldehyde and acetone from the alcohol, especially ethyl alcohol, by distillation is incomplete, these imp? should be converted to the corresponding alcohols by a previous ca? hydrogenation.  Experiments show that the aldehyde and ketone reaction thereby completely disappears.

Literature:

French patent 833302	)
English patent 478318	)Discovered by Dr. Ed. Linckh, Dr. Fr. Winkler
American patent 2148099	)
German patent 76490, December 23, 1943	)Discovered by Dr. R. Klemm, Dr. E. Linckh, Dr. Fr. Winkler

Oppau, June 6, 1947                                                            Duftschmidt

Ammonia Laboratory                                                                                        June 1947

Dr. Arno Scheuermann

The Hydrocarbon Synthesis from Carbon Monoxide and Hydrogen

Researches of the Ammonia Laboratory from 1935 - 1944

Researches of the Oppau Ammonia Laboratory in the Fischer-Tropsch Synthesis

I.  Investigations with Cobalt Catalysts

    a.    Experiments at Atomospheric Pressure

    The Ammonia laboratory started its investigations in the f? of pressureless hydrocarbon synthesis according to Fischer-Tropsch i? by setting itself the task of reexamining the relevant material in the literature.  Dr. Lorenz was able to prove by numerous experiments that when the cobalt-thorium-kieselgur and nickel-manganese-aluminum-kieselgur catalysts proposed by Fischer were used, the values checked in respect to performance and life of the catalysts, type of product, and susceptibility of the synthesis to temperature and catalyst poisons.  He further decided that there was little choice in the individual products so that a single adaptation to the market conditions, i.e., a purposeful shifting in the ?position of the product, was not possible.  It was possible, for example to roughly triple the paraffin proportion only by a substantial increase of the contact time of the gas with the catalyst to about 6-7 times the technically conventional period at a synthesis temperature kept down to 160-170 degrees.  This was only possible at the expense of the catalyst life.  But even this short-lived success was dependent on the type of kieselgur used, for other types, even under these extreme conditions did not give these high paraffin yields without some restrictions on the Kieselgur.  When the kieselgur was replaced with silica gel, the formation of higher-boiling products was greatly impaired.  The reproducibility of the catalysts remained unsatisfactory despite expenditure time and effort.  Statements on the influence of experiment variation cannot be made with any certainty until this indispensable condition is met.

    The author, together with Dr. Meisenheimer, in October 1936 undertook to continue the investigations at this juncture.  In this connection it seemed advantageous to examine the part played by the individual components of this catalyst by temporarily restricting ourselves to catalysts containing cobalt as the basic metal.  In order to proceed gradually a cobalt catalyst was produced at first without additions in order to check on the influence of various precipitation conditions, reduction temperature, and alkali content.  The result of the initial investigations, whose success is attributable to our close collaboration with the X-ray laboratory, can be summed up as follows:

    1.    Pure cobalt metal produced under specific precautionary measures, with a catalyst space velocity of 1 liter gas/1 gram Co hourly, yields approximately 110 grams/Ncbn synthesis gas as compared to 25 grams/Nm³ in the gasoline synthesis.  This is the value given in the literature.  The lower the reduction temperature the better the catalytic activity of these catalysts without promoters or carriers.  Catalytic activity does not occur at high reaction temperatures due to premature sintering.

    2.    However, if the finished catalyst contains noticeable quantities of Co₃O₄ along with basic carbonate before reduction, it will be substantially less susceptible to the reaction temperature as a result of the difficulty of reducing this oxide, which acts as a stabilizer.   

Red. Temp	grams yield/m3
	without Co3O4	with
350	0	30
300	1	-
270	15	108
225	39	135

    3.    The following was proved to be reliable as a catalyst production recipe:

    Cobalt nitrate containing water of crystallization is dissolved in water and, with a little more than the calculated quantity of potassium carbonate, subjected to a cold drop by drop precipitation within 40-50 hours.  The precipitation is washed out by a suction filter until there was no longer any evidence of nitrate in the filtrate.  The moist catalyst is dried for 24 hours at 110 degrees and, in order to attain a sufficient content of Co₃O₄ by the oxidation of CoO, again moistened and again dried at 110 degrees for 10 hours.  The catalyst is reduced at 225 degrees.  Time about 10 hours.

    4.    The recrystallation which rapidly set in caused a short for this unstabilized catalyst.

    5.    Carbide formation in the spent catalysts could not be obtained with X-rays.

    At the time that this information was discovered a new prob? was added, namely, the increase of the paraffin yield.  This task was closely connected with the work being done at Oppau in the field of paraffin oxidation to soap-stock fatty acids, or wax.  As a result of this we were forced to pursue our activities in several directions which frequently intersected.

    The simplest manner of stabilizing cobalt obtained in an a? form was deposition on carriers. For this purpose we initially used kieselgur, the extreme case simultaneously occurring in which we for once used a Fischer catalyst with a thorium content of 0.  This resulted in a catalyst which naturally did not quite approach the standard Fischer catalyst with its 18 per cent thorium in respect to activity, but when displayed an excellent life.  It was used about 4 months without being regenerated.  This was attributed to the fact that it only formed paraffin at the start; later, after about 14 days, practically no more paraffin was formed.  This direction, however, did not conform to our wishes, but a least is showed us the part played by one of the three components, the thorium oxide.    

    It accomplishes two functions in the catalyst combination:

1.    Stabilization of the catalyst in an active form, which is only possible otherwise by a special drop by drop precipitation.  2.  Guiding the synthesis into the production of long-chain products.  This information led us to investigate catalysts with a high content of thorium oxide for the purpose of an increased paraffin yield.  As was to be expected, an increased percentage of paraffin was obtained.  However, since the total yield in conjunction with the greater thorium content, at first drops slowly, then more rapidly, the actual increased yield of paraffin may be ignored.  However, the combination of a thorium-rich and a normal catalyst in a two-stage operation proved to be good.  The use of this catalyst resulted in a paraffin yield of 35 grams/Ncbm of a total yield of 130 grams/Ncbm.  The experiment lasted 2 months, hydrogen regeneration at 220 degrees occurring after each 14 days.  The experiment was discontinued after two months without any visible falling off of activity.  Further efforts were made to increase the paraffin contents by the pressureless procedure.  Alkalization of the catalyst or modification of the gas composition (increasing CO content) were tried.  It was shown that each individual measure or a combination of both tended towards increased paraffin formation.  However, this increase only reached moderate proportions and does not compare with the increase attained by raising the thorium content.  Since a decrease in the total yield is tied up with these measures, the increase in the percentage of paraffin is no yardstick for the actual formation of paraffin.

    However, all these efforts to achieve an increased yield of paraffin are outstripped by another measure, namely, the application of pressure.  It is therefore not necessary to go into the above experiments in greater detail.  However, before turning to these pressure experiments we should like to report briefly on our efforts to find other promoters and carriers for the promotion and stabilization of the cobalt.

    At first magnesium oxide seemed to us to be a suitable promoter addition since it forms a spinel with cobalt oxide (Co₂o₃), and forms an unbroken series of mixed crystals with CoO.  Because of the attainable homogeneity this appears to us to be a favorable precondition for catalytic activity after the reduction.  It was determined as a result that addition ? of 2-15 per cent MgO without addition of kieselgur were approximately equivalent among each other, but gave a total yield of only about 80 grams.  This type of catalyst is very dense and, since it possesses a higher metal content, with respect to the same quantity of metal, it occupies a substantially smaller volume than the kieselgur catalyst; for example, 4 grams of cobalt are contained in 10 c.c., it would otherwise require about 40 c.c.  This results in a four-fold volume space velocity since the catalyst charge rate is referred to the gram quantity of cobalt (1 liter/1 gram Co/hour).  However, if it is desired to employ the usual volume charge with the catalyst then four times as much catalyst metal is needed and a yield of 100 grams/1? is  obtained after two months without regeneration.  This corresponds to the life of the usual cobalt-thorium-kieselgur catalysts.

    The attempt to deposit the cobalt-magnesium catalyst combination on kieselgur resulted in a catalyst with a yield about 10 per cent poorer than the standard catalyst with a similar paraffin yield.

    Further investigations on the replacement of magnesium oxide by manganese, ceric or lanthanic oxide resulted at best in equivalent but not better catalysts than those proposed by Fischer.

    The cobalt-nickel-manganese-uranium-kieselgur catalyst proposed by the Japanese, with the same yield as the standard catalyst, is quite noteworthy because of the marked direction of its reaction towards paraffin (28 grams paraffin/Ncbm in over four weeks).  On the basis of our own experiments we attribute this condensing property to the influence of the uranium.

    Although we do not ignore the good catalytic properties of the standard catalyst we received the impression during our investigations that the kieselgur might easily lead to trouble.  The fault lies in the undefined character of the material since we have not reached a point where we can exactly analyze its make-up.  Its physical structure is doubtless of importance along with its chemical composition so that, in the end, only catalytic experimentation can decide upon its utility.  We therefore spent a great deal of time in looking for another material to replace kiesslgur, such as aluminum oxide, kaolin, magnesium oxide, silica gel, talc, porcelain, etc.  Our experiments showed that kaolin and aluminum oxide, after a pre-treatment with temperatures of 400-600 degrees or 800-1000 degrees, yield a catalyst practically equivalent to the kieselgur catalyst (life four months), all other carriers reacting unfavorably.  Nevertheless, the fact that we retained the kieselgur in general is explained by the observation that a less favorable type (4 S) was substantially improved by igniting in a current of air at temperatures of 500-700 degrees.  It even outstripped type 20, the type generally employed by Fischer which was supplied by the German Kieslgur-Werks, Hanover after treatment by ignition.  The treatment at Hanover, however, does not occur at a specific temperature, many particles therefore being overburned.  A further observation was made to the effect that if slight additions of magnesium oxide are added to the kieselgur to the extent of 1-2 per cent with respect to the cobalt metal, fluctuations in the behavior of the various kieselgurs could be practically eliminated.  Nevertheless, experiments undertaken from time to time with the aim of replacing kieselgur with silica gel finally achieved success by the observation that a silica gel preheated for about five hours in a current of air at 800 degrees to avoid silicate formation can completely replace kieselgur if a fine-pored gel is employed.

    We can summarize the results of our numerous experiments by stating that we cannot propose a more serviceable catalyst than the cobalt-thorium-kieslegur catalyst.  We consider of equal value for synthesis performance the slowly precipitated thorium-free Co catalyst with about 1 per cent Ag on fine-pored silica gel which offers the advantages of a saving of thorium, of reproducibility in the synthesis reactor, and of better catalyst life.  The weaknesses of the thorium-free containing standard catalyst lie in the fact that it contains kieselgur with its unpredictable qualities and that its reproducibility is still somewhat uncertain.  The flactuations in the kieselgur, however, can be equalized by the addition of slight quantities of magnesium oxide and by heat pre-treatment at a specific temperature that is not too high.  Reduction in receptacles other than synthesis reactors is also technically objectionable.  Transporting the reduced catalyst under carbon dioxide causes the most active spots to be lost as a result of surface oxidation.  Therefore we consider the addition of about 1 per cent of silver to the cobalt catalyst to be one of the advantages of our procedure.  The reduction temperature is thereby lowered so far that we can reduce directly in the synthesis reactor without the necessity of taking a decrease in the yield as is the case with copper.  In conclusion, the following experiences gleaned in catalyst preparation may be mentioned:

    Potassium bicarbonate was proved to be better than potassium carbonate as the precipitating agent for the Fischer catalyst that was quickly precipitated and subsequently boiled since the catalysts become more active and more readily reproducible.  On the other hand, potassium carbonate must be used for the thorium-free catalysts since these catalysts are slowly precipitated without subsequent heating.  Under these circumstances a complete precipitation cannot be achieved with bicarbonate.

    Furthermore, washing with a minimum quantity of wash water was proved to be necessary for otherwise, despite freedom from alkali, less active catalysts would be obtained.

    We know from experience that the size of the excess of precipitating agent is of negligeable importance.  A direct relationship of catalyst efficiency with the PH of the precipitation was not noticeable.  The type of drying, whether in the air at 110 degrees, in a current of carbon dioxide, or in a vacuum, is equally unimportant.

    b.  Experiments with Medium Pressure

    As was already mentioned, the synthesis conducted under a slightly increased pressure brought about a substantial improvement in the paraffin yield.  Nevertheless, more than a year's work was necessary before the technical development of the laboratory apparatus had reached a point where we could obtain with certainty a total yield of 130-140 grams/Ncbm with a paraffin proportion (boiling point above 320 degrees) of upwards of 60 per cent.

    The pressure used by us in the laboratory pressure experiments ran from 12 atmospheres to a maximum of 20 atmospheres; no difference in paraffin production was found within this pressure range.  The charge velocity was the same as in the pressureless experiments: 1 liter gas/? 1 gram metallic cobalt/hour.  Experience shows that our procedure for attaining the highest possible paraffin yields with cobalt-kieselgur catalysts free of additions, which almost stopped producing paraffin after several days without pressure, turned out to be more reliable than the thorium-rich catalysts which are good paraffin producers without pressure.  Catalyst life was also satisfactory; after ten weeks of operation the catalyst still retained more than 90 per cent of its maximum activity.  The explanation for the inferiority of the thorium-containing catalysts under pressure to the catalysts free of thorium is based on the fact that former forms very high-molecular weight paraffin which leaves the catalyst to a lesser degree and therefore leads to the block of the active spots.

    The most important innovation in the experiments was the instillation for the first time of pressure flow manometers which permit exact measurement under pressure of the even flow of the smallest quantities of gas (e.g., 12 liters of depressured gas/hour = 1 liter gas/hour at 12 atmospheres).  The requisite pressure capillaries with fused manometer le? were made from special Jena glass.  The connection to the high-pressure piping was accomplished by means of stuffing boxes fitted with special soft rubber packing.  The pressure flow manometers constructed in this manner withstood a test pressure of 20 atmospheres, the glass and metal joints being found sufficiently tight.  The pressure capillaries were tested by putting the entire apparatus under pressure after installation before the start of the experiment.  Then through an analysis valve behind the synthesis reactor a constant volume of gas, e.g., 12 meters/hour, was let down through a flow manometer.  After a little while the proper manometer level could be read off on the pressure flow manometer.  The constancy of the result was tested by repeated readings over a period of time.  Varied volumes of inlet gas under pressure can be accurately gaged by adjusting the quantity of exit gas.  The corresponding quantity of let down gas is then stated to be the inlet gas volume.  Comparison of the inlet gas volume calculated by m    of the oxygen balance with the inlet gas volume measured by flow capillaries demonstrated that measurement of the inlet gas volume under pressure was far superior in most instances to calculation of the inlet gas volume.  An additional advantage lies in the fact that a more uniform operation of the reactors can be attained by keeping the inlet gas volume constant with the pressure capillaries.

    Later operation showed that a by-pass in the inlet gas capillaries was advantageous.  By cutting off the narrow capillary path it enables the product to be forced out of the pressure separator without risk, and makes possible an easy interchange of capillaries without interrupting the inlet gas feed.  This was very convenient when the pressure capillaries were gaged during operation.

    A brief survey of the performance of cobalt-kieselgur-catalyst employed by us, which was prepared by slow drop by drop precipitation of a cobalt nitrate solution in the cold with potassium carbonate, will be given in the following in order to facilitate a comparison with the iron catalysts developed by us at a later date for the carbon monoxide-hydrogen synthesis.  In general the precipitation lasted for more than 24 hours.  Catalyst life lasts for several months so that regeneration with hydrogen seldom comes into question.  

    In the following we give several values obtained by various methods of preparing our catalyst:

Expt. no.

Cat. no.

Temp. °C.

CO consum.

in exit % CO2

gas %CH4

%k2O/grams cat. for 100 Co

Yield solid + liquid product Nm3

% Par. >320

% Straight chain

(Table on page 148)

186°	71	1.0	7.7	=  0.324 kg/liter	Kt/day
appear as mean values				135        61       98
Division of the primary product into boiling ranges

(Table on page 148)

    Mean values

    If a lower yield is desired, especially if only one synthesis reactor is available (technically the process is run in two stages) then it is possible to manage with lower temperatures:

    Experiments with a lower CO conversion (charge velocity as usual 1 liter/1 gram Co)

Expt. no.

Cat. no.

Temp.

CO consum.

in the  % CO2

exit gas %CH4

%K2O in catalyst 100 Co

grams yield solid + liquid product Ncbm

(Table on page 149)

Mean values:

    It can be seen that the methane content is still proportionately high even with this slight CO conversion.  This is attributed to the good hydrogenation power of the cobalt.  One would think therefore that the catalyst could be adjusted to another cobalt concentration by kieselgur, thereby possibly checking the hydrogenating influence of the cobalt.  However, this was not successful, as is shown by the following table:

Cat. no. Expt. Content no. %K2O %Co

Temp.

CO consump.

%CO2 %CH4 in the exit gas

grams yield of solid plus liquid prod. Ncbm

% par. > 320°

% straight-chain

(Table on page 149)

    There are no regular great differences, in particular the methane content shows no progression of values.  There is a progression in the case of paraffin production which increases with rising Co content, whilst, as is evidenced by the boiling point figures, the proportion of hard paraffin rises.

(Table on page 149)

    The following series of experiments answers the question: what effect does alkali in the catalyst have on methane formation?

Expt. no.

Cat. no.

%K2O/cat..

Temp.

CO consumption

%CO2 %CH4 in the exit gas

g. yield solid plus liquid pr. Ncbm

%Par. 320°

% straight-chain

(Table on page 150)

      It is evident that with a rising proportion of alkali the synthetic temperature must be markedly increased to produce the same CO conversion.  This, however, assists in the formation of methane, as can be seen from the corresponding values.  At the same time it diminishes paraffin formation.  This remains even with a decreased CO conversion:

(Table on page 150)

    In conclusion, let us offer one more comparison; the behavior of a cobalt catalyst with a CO:H₂ - 1:1 synthesis gas.  All other experiments were conducted with a CO:H₂ = 1:2 synthesis gas:

(Table on page 151)

    The table shows: at 1:1 there is a smaller yield of primary product, a higher proportion of olefin and alcohol in the primary product, less straightchanedness of the paraffin, and also less methane formation.  In the case of synthesis gas 1:2, methane content diminishes in the course of time if the catalyst is kept at the same temperature.  This being the case, it is important in evaluating the methane value to know at what time the sample was taken:

(Table on page 151)

    II.    Investigations with Iron Catalysts

    a.    Experiments at Atmospheric Pressure

    As a result of the external circumstances caused by the start of the war in 1939 the experiments which were being conducted to synthesize hydrocarbons from carbon monoxide and hydrogen with iron catalysts, which heretofore had been of subordinate interest and had been undertaken in a modest way in connection with cobalt catalysts, became of major interest.  Since subsequent investigations were conducted exclusively under medium pressure, the experiments which were conducted previously at one atmosphere will be briefly presented here.

    Since we were dealing with what amounted to a new field of activity, the experiments must be evaluated in the light of an initial exploration in the field of catalysts, especially in view of the fact that many questions still remained unanswered.  However, this field did not appear urgent enough to devote more work to it.  It became  quite clear to us even in the early part of our work that it would not be possible, without intensive and systematic work, to develop an iron catalyst capable of completely replacing the cobalt catalyst in the atmospheric pressure process.

    The catalysts investigated for the most part consisted largely of mixtures of the following constituents: iron-copper-alkali with magnesium oxide or with aluminum oxide.

    Since the copper, according to our viewpoint at that time (future data on the role of the copper will be given later) served solely to facilitate the reduction of the iron, i.e., to lower the reduction temperature, it was present in all catalysts in the constant ratio of Fe:Cu = 4:1.

    Both catalyst combinations were investigated for the most favorable combination of alkali and metallic oxide, as well as their behavior with respect to variation of the CO:H₂ ratio in the synthetic gas and response to the reduction temperature.  It was found that, even under the most favorable conditions of composition and experiment con? at best yields of 60 grams of solid and liquid products/Ncbm synthesized gas were obtained over several weeks.  Contrasted with this were yield of about 130 grams/Mcbm with the cobalt standard catalyst (with thor?) Of the remaining investigated catalyst combinations only two more ser? will be stressed. The first aimed at progressive replacement of the cobalt by iron in the cobalt-kieselgur catalyst customarily employed by us.  In the second series, several iron-nickel (1:1) - aluminum oxide catalysts with assorted methods of preparation were the object of our investigations.

    The first series demonstrated that cobalt-kieselgur catalyst containing up to 40 per cent iron, with respect to the cobalt, evidenced good effectiveness, the yields, however, decreasing as the proportion or iron rose.  At the same time the optimum synthesis temperature rose. The content of cobalt was controlling in the reaction.  This is inferred by the fact that only water was formed as a by-product during the reaction whereas iron catalysts give carbon dioxide for the most part.

    The catalysts of the second series, iron-nickel catalysts, promoted with 5-15 per cent aluminum oxide, gave yields of about 80 gms of solid and liquid products, falling far short of the normal cobalt catalysts.  The type of precipitation for these catalysts is of little importance with respect to the total yield; however, it is possible to obtain catalysts by slow reverse precipitation (letting the nitrate solution of the metals drop into the potassium carbonate solution) in the cold which are characterized by an extraodinarily small bulk weight as compared with catalysts produced by other means (0.2 and 1.1).  Great economies of metal can be effected with the same catalyst space without influencing the yield.  In addition to the economies in metal there is still another advantage, namely, that they are easy to use because of their slight sensitivity to temperature.  With such catalysts the reaction proceeds mainly in the direction of water formation.  In essence this points to action on the part of the nickel which is neutralized by the addition of the iron.

    The catalysts with high and low bulk weights are clearly differentiated from each other with respect to their external characteristics.  The heavy catalysts are compact, vitreous, dark-brown; the light ones, on the other hand, are light-brown particles which display a tendency toward decomposition as a result of their soft, loose structure.

    So much for the experiments with iron catalysts in synthesis at atmospheric pressure.

    b.    Experiments with Iron Catalysts at Medium Pressure

    The aim of these investigations was the synthesis of a paraffin (boiling point, straight chained, iodine number) suitable as a starting material for paraffin oxidation.  The problem, therefore, consisted in finding a suitable catalyst capable of replacing to a great extent the cobalt catalyst developed for this purpose by us.  This catalyst was to be produced from native raw materials or from raw materials abundantly available to us.

    It is not necessary to discuss details of equipment at this point since the reactors were the same as those employed for the pressure experiments with cobalt catalysts.  It should again be stressed that we are des?ing here with small scale experiments, using catalyst charges of 30-100 c

    The working pressure - insofar as the contrary is not expressly stated - was always 12 atmospheres since we attained the best results at this pressure with the cobalt catalysts.  The catalyst charge rate was not based on the weight of the metal of the catalyst, as was the case with the cobalt catalysts, but on the volume, and in general, was 240 parts by volume of synthesis gas (CO : H₂ = 1 : 2) : 1 part by volume of catalyst/hour.  Since we were more or less concerned with catalyst testing in these investigations carried out solely on a laboratory scale, we ran only one operating stage, seeing to it that the gas was extensively converted (app. 70 - 80 % consumption).  We were quite clear about the necessity for a process of several stages or of recycle operation in conversion to a commercial scale.  It was also clear to us that a CO : H₂ = 1 : 1 synthesized gas was more advantageous for iron catalysts if care was taken to see that if such a gas mixture was used, the conversion proceeded in another ratio up to CO : H₂ = 1.6 : 1.

Catalyst Tests

    The point of departure for our catalyst investigations was the experience obtained with cobalt catalysts, namely, that catalysts yielding the least possible paraffin in the pressureless synthesis behave especially well in the medium pressure synthesis in regard to paraffin yield, synthesizing a great deal, but not an excessively, hard paraffin which leaves the reactor without difficulty and permits undisturbed operation for mon? without regenerating the catalyst.

    The initial pressure experiments (December 1939) with a promoted precipitated catalyst in connection with the pressureless experiments conducted with it show that it is possible with iron catalysts, too, in principal to obtain a satisfactory, straight-chained paraffin suitable for oxidation to fatty acids which would be equal in quality to the good products obtained with cobalt catalysts.  The same total yields and paraffin yields, however, were not attained.

    The catalyst used was obtained by precipitating an iron-copper-aluminum oxide-potassium catalyst drop by drop for 48 hours, and was used without carriers.  Its reaction temperature ranged from 210 to 220 degrees, viz., within the temperature range up to 230 degrees, which is the maximum allowable for commercial reactors with steam cooling constructed for the synthesis with cobalt catalysts.

    Using numerous individual observations, the catalyst combination used in the initial experiments was worked up into a catalyst which fulfilled our requirements in regard to paraffin yield to a large extent.  Its composition is roughly equivalent to the following parts by weight: Fe - Cu - A1₂O₃ - K = 100 - 25 - 100 - 10.

    Such catalysts were used by us for months at a time always at the same reaction temperature of 208 degrees, the longest run being 135 days, without showing any signs of fatigue at the end of this period.  However, after several weeks the paraffin content in the primary product dripped from about 60 to 45 per cent in favor of the gasoline fraction.

    The involved method of preparation of this catalyst may be considered a disadvantage.  Not merely the two-day precipitation, but also the difficulty of washing the catalyst free of alkali, or of establishing the desired alkali content, requires a great deal of time and care.  Therefore, after obtaining good results with magnesium oxide instead of aluminum oxide, it occurred to us to test this promoter, too, in our precipitation catalysts at medium pressure.

    The influence of various quantitative ratios upon one another was investigated in a large number of experimental series.  This was at first with catalysts without carriers.  Certain regular patterns became evident; however, certain catalysts, after a temporarily good start, revealed an early drop in activity after a lengthy run.  This difficulty, however, could be eliminated with the use of carrier catalysts, kieselgur proving itself to be very reliable as in the case of the cobalt catalysts.  In the case of these catalysts, too, the individual constituents were compared to each other.  At the conclusion of these investigations in paraffin synthesis two catalysts were available which, by way of comparison, showed the following numerical values:

 (Table on page 152)

    The table shows that there are certain differences, even though they are not very great, in the results attained with these two catalysts.  The catalyst promoted by MgO behaves somewhat more favorably with respect to the total yield of solid and liquid products in a once-through gas operation.  However, this advantage is counterbalanced by the fact that it tends toward the formation of higher-boiling products as a result of the higher synthesis temperature.  Also, the straightchainedness of the paraffin obtained with it is not quite as good as that obtained with the paraffin produced by aluminum catalysts.  It was finally possible to eliminate these disadvantages of the magnesium catalyst by producing catalysts, by a high gas throughput during reduction with hydrogen or synthesis gas (5000:1) at temperatures of 180-250 degrees, which required a working temperature reduced by 15-20 degrees as compared with the previous temperature.

    We have the following working hypothesis for this catalyst behavior during running in:  The conversion of carbon monoxide and hydrogen is not a simple process.  Both polymerization and hydrogenation are probably involved, both types of reaction certainly not being accelerated by the same promoters.  Furthermore, it is assumed that the methylene group, which is the precondition for polymerization, is formed by the hydrogination of iron carbide, whereas the hydrogenation of this methylene group to methane, or the hydrogenation of a longer unsaturated hydrocarbon to the corresponding saturated hydrocarbon, proceeds by way of the metallic iron points.  The extent of probability whether a formed methylene radical will be hydrogenated to methane, or whether it will have the opportunity to combine with other radicals to form larger complexes, will depend on the metal:metallic carbide ratio.  The lower the temperature, the greater the tendancy towards chain formation.  The catalyst must also be run in in such a manner that the iron will be in a condition - reduced or unreduced - capable of forming carbide at low temperatures when the carbon monoxide and hydrogen come together, i.e., the synthesis must be capable of completion at low temperatures.  The probability is thereby created for an increased yield of paraffin and a smaller loss through undesired gas formation.

    The mean values obtained with catalysts promoted by M_gO with different types of reductions (synthesis gas CO:H₂ = 1:2, catalyst space velocity 240 liters/liter catalyst/hour) are given in the following in order to make possible a quantitative comparison:

 

Gas Reduction Temp. Charge

Time Synth. temp.

CO consump.

Grams solid and liq. prod.

% Par. > 320° C.

Grams CH Ncbn

without reduction.

(Table on page 153)

    The table shows that the synthesis temperature for catalysts without previous reduction, or reduction at high temperatures, is practically the same, whereas a so-called "flooding" with hydrogen or synthesis gas low temperature makes possible a reduction in temperature.  Despite the higher working temperature, however, the unreduced catalyst, insofar as its catalytic activity is concerned, behaves in a manner much more similar to the catalyst reduced at low temperature in regard to the consumption ? of HO:H₂, and the formation of methane, olefin and alcohol, than to the catalyst used at the same temperature, but which was formed at a higher reduction temperature.  This supports the conclusion that the iron newly obtained from the oxide is particularly capable of forming the necessary carbide with the carbon monoxide of the synthesis gas, more readily, in fact, than a reduced but already sintered iron. Even though the purpose of this running in is to create an especially reactive iron or iron oxide we are not saying that the iron oxide of the starting material must be thoroughly reduced.  On the contrary, certain X-ray analyses showed that highly-active catalysts were reduced only in traces before use.  The question is then immediately raised whether in that case the added copper is at all necessary to facilitate the reduction.  In this connection the experiment we conducted on the basis of the X-ray investigations of Dr. Herbst will be found interesting.  They show that the added copper performs a second function in the catalytic behavior that is at least just as important.  As was discovered for the first time in the Ammonia Laboratory (Dr. Halle), when iron catalysts reduced at low temperature are employed a hexagonal iron carbide appears which, just as in the case of Hägg's iron carbide, probably has the formula Fe₂C on the basis of its structure and chemical analysis.  However, this has hitherto not been described in the literature.  This carbide is of particular interest because it has the same lattice structure as the analogous iron nitride Fe₂N and because it has very similar lattice dimensions.  Thus, carbon, under the conditions of the hydrocarbon synthesis, can be placed into the iron lattice under certain circumstances just as nitrogen is placed into the iron lattice in iron nitride formation.  This "new, hexagonal carbide" seems to us to be a necessary, if not in itself sufficient pre-requisite, for especially good catalytic activity.  Therefore, the information obtained by X-ray investigations (Dr. Herbst) to the effect that copper-containing iron catalysts are less sensitive to temperature and that this "new hexagonal" iron carbide in the presence of an addition of copper, which is always found in our catalysts, can stand a temperature around 100 degrees higher without into Hägg's carbide, is significant.  The initial and barely noticeable conversion only begins to occur at 350 degrees with copper-containing catalysts and the conversion is completed at 450 degrees, whereas in the case of catalysts free of copper conversion begins to set in at 290 degrees and is complete at 330 degrees.

    Let us now, after this discussion of problems of catalyst formation, return to our two catalyst combinations.  In judging the respective advantages and disadvantages of the two catalysts we gave our preference to the ? promoted with MgO because of the simplicity and ease of its production.  The catalyst is rapidly precipitated.  This is expediently accomplished by running together the nitrate solution of the metals (Fe, Cu, Mg) and alkali carbonate solution in a mixing nozzle (constant P_H).  The finished precipitation mixture is collected in a receiver with the kieselgur suspended in H₂O or with the fine-pored silica gel, heated for a short time (about 10 minutes) and then decanted.  Since the precipitation quickly settles and the catalyst is readily washed, its preparation requires o? a fraction of the time necessary for the preparation of the catalyst promoted with aluminum oxide.  Catalyst efficiency is independent of the manner and duration of drying.

    The catalyst, in the form employed by us at the present time, has the following composition:  Fe - Cu - MgO - K - kieslegur (or silica gel) : 100 - 25 - 50 - 6/8 - 50.  In unreduced form broken into pieces of 1-3 mm. it has a bulk weight of 0.4 - 0.45.  The combination can und? considerable variation within broad limits in regard to the proportions of magnesium oxide, potassium as well as kieselgur and silica gel without revealing any basic change in catalyst efficiency.  This is shown by the following table which contains a small selection of such precipitated catalysts (catalyst span velocity 240 liters gas/liter catalyst/hour, CH₂ = 1:2).

Catalyst number		for 100 Fe and 25 Cu
	MgO	SiO2	K

(Table on page 154)

Cat. nr	Synth. temp.	CO consum.	g.yield/Ncbm	% par. >320°C	% straight.-1 chain	g CH4 /Ncbm	Fraction 320-450
							% olefins	% alcohol

(Table on page 155)

1)    Determination of the degree of branching is accomplished by a purely empirical method worked out by schaarschmidt with antimony pentachloride.  This method was perfected by Dr. Leithe in the Ammonia Laboratory for use in transference to higher-boiling fractions and improved by Dr. Kotzschmar.

    The catalyst was repeatedly tested in four-stage experiments in order to test its effectiveness on a somewhat larger scale.  We will discuss these investigations in greater detail since the results seem quite remarkable.

    Equipment:  The small-scale laboratory experiments were conducted in pressure tubes of 15-16 mm. internal diameter situated in an electric furnace, but in the course of several years the tendency grew to embed these high pressure tubes in an oil bath in order to achieve a more uniform distribution of heat over the whole catalyst bed.  The catalyst tube was then heated by the circulating oil which was pre-heated in an electric furnace.  Temperature recording, which at first was undertaken by thermo-elements situated in the catalyst, was accomplished when using oil circulation by catalyst thermometers in the oil at the entrance and exit to the catalyst tube, after establishing the fact that a temperature loss of about 0.2 degrees occurred throughout the entire catalyst length in a 5 m. tube with equal oil and catalyst temperature.

    The reactor system for the four-stage process consisted of 4 to 5 m. long iron catalyst tubes with an internal diameter of 16 mm.  The contact zone was 4.50 m; the quantity of catalyst in the unreduced state amounted to 0.8-1.0 liters.  Removal of the considerable amount of reaction heat was accomplished by oil circulation which was set up separately for each reactor.  The high pressure gas lines were so arranged that in case of need each reactor could be removed from the series.  If desired, the synthesis gas could pass to the next reactor without separation of the lighter boiling products.  As usual a gas mixture obtained by mixing the constituents in a CC:H₂ = 0.9-1:1 ratio served as synthesized gas.  It contained no carbon dioxide and only about 2 % N₂.  The gas is largely free of sulfur, but to be certain, it was subjected to an additional purification in a copper preliminary reactor at 180 degrees and in 3 coupled towers with activated charcoal - caustic potash - activated cha?. Gas recording under pressure was accomplished by dry gas meters set in pressure-resistant chambers.  In addition, a pressure flow manometer was placed before the first reactor in order to note the constancy of flow.  Separation of the solid and liquid products was accomplished after every reactor:

Hot separator	120°	Paraffin - middle oil
Cold separator	20°	Middle oil - gasoline - water
Cooling by ice	0°	Gasoline - water
Low cooling	-70°	Gasoline - liquefied gas

    Furthermore, it was possible to take gas analyses after every reactor.  MgO catalysts produced in various ways served as catalysts.  They were reduced in the furnace with H₂ at 220 degrees for 24 hours at a catalyst space velocity of 1000:1.

    The following table summarizes the data obtained in three different experiments:

(Table on page 156)

    The duration of the experiments differed greatly.  Whilst two experiments ran for 124 and 141 days, the final experiment was interrupted after 42 days on account of trouble with the equipment.

    The balance shows that the synthesis gas which was fed to the catalyst in a CO:H₂ = 0.9-1:1 ratio was converted in a 1.1-1.2:1 ratio.

    The very remarkable yield of primary product, which is about a third higher than would otherwise be obtained in a reactor ⁽¹⁾ under similar conditions of catalyst space velocity (120:1), can only be explained for the present by the varying high linear velocity of the gas with equal volume charge.  This causes a better heat removal and dispersion throughout the entire catalyst area.  The linear velocity of a catalyst tube of 16 mm internal diameter is per sec./12 at/227° C.:

in 1 tube of 5 m length :	2.5 cm
in 4 tubes each 5 m     :	10.0 cm.

    The last value lies at the border between the laminar and the turbulent flow area.

    If the intermediate separation of the products which do not separate freely is omitted, then almost no important differences occur.  The slight shifting in the boiling points of the higher boiling products to the extent to which it is real - only meets our desire to obtain long-chain products.

    The following table contains several values for purposes of comparison.

(Table on page 157)

The influence of the Experiment Conditions:  Several of the influences felt in the preparation of the precipitated catalyst are mentioned here without in any way laying claim to exhaustiveness:

    Method of preparation:    Selection of precipitating agent

                                            Concentration of precipitating solutions

                                            Temperature of the precipitation solutions

                                            P_H of the precipitation

                                            Speed of the precipitation

        Promoter:                    Time of addition

                                            Form of addition

                                            Formation of mixed crystal or compound

                                            Reciprocal influencing of several promoters

            Carrier:                    Purification

                                            Temperature pre-treatment

                                            Pore size

            Alkalization:             Method of introduction of desired alkali compound

            Washing water:        Quantity and temperature

            Drying:                    Temperature

                                           Vacuum or gas current

                                            Sintering

                                            Bulk weight (pressed or broken up)

                                            Particle size

                Reduction:            Temperature and time

                                            Gas space velocity

                                            Pre-treatment with foreign gases

    These examples give an idea of the extraordinary abundance of variations possible in the preparation of the catalyst, even when its composition is known.  Attention was directed to the special influences described above in the preparation of the MgO catalyst insofar as they were distinct and observed.

    In addition to these problems arising in the preparation of catalysts, there is also the multiplicity of the experimental conditions which must be considered in evaluating the behavior of the catalyst which will be discussed here briefly.  The difficulty in obtaining exact data is caused by the fact that several factors are always simultaneously at work so that it is often not at all practical to attribute this or that manifestation to a specific experimental condition.

Pressure:  As is shown by the following examples selected from widely separated pressure stages, a pressure rise - with the same catalyst, equal catalyst space velocity and the same synthesis gas - acts in such a manner that, in order to achieve equal CO conversion, the temperature must be raised.  At equal temperature and lesser pressure there is a better CO conversion.  This can also be expressed in another way, viz., with equal temperature the lesser partial pressure of the ideal gas ⁽¹⁾ (CO:H₂ in the synthesis gas in relation to the consumption in the synthesis) attains a better CO conversion.

(Table a on page 158)

    In connection with the pressure rise there occurs a rise in gasification and, in general, a shortening of the hydrocarbon chains with decrease of olefins and an increase of alcohols.  It remains uncertain to what extent the chain shortening is a function of the pressure for the temperature rise requisite for attaining the same CO conversion acts in the same direction.

Temperature:  This is made clear by the following brief summary:

(Table b on page 159)

    It can be seen from the table, furthermore, that the difference in the olefin and alcohol contents are very small in the same boiling ranges.  Straightchainedness seems to drop with a rise in temperature.  However, the values are not given since this is not at all certain.

Composition of the Synthesis Gas:  A feed gas with a CO:H₂ ratio of 1:? is more suitable for an iron catalyst, which directs the reaction toward the carbon dioxide side

                                            3 CO + 3 H₂ = 2 CH₂ + CO₂ + H₂O

than a gas rich in hydrogen, e.g., 1:2, which is used with the cobalt catalyst and is converted by it

                                                            CO + 2 H₂ = CH₂ + H₂O

    It remains uncertain whether the iron catalyst, too - several indications point to it - acts toward the water side in the first stage the water gas process taking place only in the second stage.  In any case, a hydrogen-rich gas contains too much hydrogen, which is available for hydrogenation, for the total reaction with the iron catalyst.  Thus, a 1 : 2 gas contains only 66 per cent ideal gas for the iron catalyst in? as the conversion gas goes to 1 : 1.  It is therefore in accord with other previous working hypothesis if, under such experimental conditions the ? is preponderant in a metallic carbide:metal ratio, and if we find increased methane and liquefiable gas formation, more saturated products, and sho? chained products with the same temperature and the same catalyst.  This is confirmed by experiment.  The richer the gas in hydrogen, the more marked the effect.  Comparative experiments are still lacking.

(table c on page 160)

The Influence of the Synthesis Gas Ratio on the Experimental Results

    Since a higher CO conversion is obtained (see the same findings under "pressure") at the same temperature with hydrogen-rich gas, whose ideal gas partial pressure is lower, an argument can also be put forward to the effect that the temperature for the catalyst is already so high under these synthesis gas conditions that the shortening of the hydrocarbon chains, as well as the increased methane formation, is conditioned by it.  We are inclined toward the first viewpoint.

    The influence of higher inert gas contents has not been investigated.  Since we are dealing here also with a reduction in ideal gas partial pressure, we should expect behavior as in hydrogen-rich gases, i.e., a shortening of the hydrocarbon chains and a better CO conversion at the same temperature than with a gas free of inert gas.

    The problem of time contact, i.e., of catalyst space velocity per unit of time, is the last point of the external experimental conditions to be mentioned.  This effect is made clear by the following table:

(Table on page 161)

A smaller CO conversion is obtained with the same temperature and a shorter contact time, a maximum of long-chain products being formed with space velocities of 720:1 and 960:1, whilst substantial differences between olefin and alcohol content are not pronounced.  When the CO conversion is the same, for which the temperature must be increased with a higher volume space velocity, the olefin and alcohol content remain almost the same in equal boiling fractions, while the chain length sometimes increases and sometimes decreases.  In this connection there are various contradiction experimental results.

(Table e on page 162)

    These discrepancies, as well as the above-mentioned maximum formation, may be explained by the more equalized heat relations in the case of a higher linear velocity of the gas (Cf. comments on the four-stage process).

    In concluding these researches on paraffin recovery with iron precipitated catalysts, it should be mentioned for the sake of thoroughness that we had advanced so far toward the end of the war that we were able to run a 600 liter experimental reactor filled wiht MgO catalyst which was produced batchwise, formed on a grooved cylinder (designed by Dipl. Ing. Misenta) and dried beforehand.  Frequent troubles with the synthesis apparatus prevented the catalyst from coming to full production.  Three and a half tons of paraffin recovered during a period of temporary uninterrupted operation were lost by fire during paraffin oxidation in an air raid.  As a result of the effects of the air raid, which also precluded resumption of plant operation, a new catalyst installation was not tried again, although the apparatus had been perfected to such a point toward the end of the or? and only experiment that it ran uninterruptedly and gave good balances.  The primary product obtained with unsatisfactory CO conversion contains 63 per cent of snow-white paraffin with a boiling point in excess of 320 degrees.

Investigations on the Recovery of Olefins

    The recovery of a primary product which would be rich in olefins in all boiling ranges is presented by the prospect of advantageously using certain fractions where saturated products can not be employed to any great extent, i.e., the middle oil fraction (B.p. 195-320 degrees) for the Oxo? reaction or sulfonation if products of this boiling range are considered for working up to detergents.  In this connection the possibility would always remain of purposefully converting the olefins of the higher-boiling fractions into saturated hydrocarbons by a simple hydrogenation.  It is therefore perfectly comprehensible if, along with the researches on paraffin recovery, the problem of olefin and alcohol recovery is again and again put into the forefront.

    Olefins ⁽¹⁾ and alcohols are now produced in noteworthy quantities with the above-mentioned iron precipitated catalysts, working with CO-rich gases.  Even more favorable results were obtained with fused catalysts even though here too conversions to saturated products occurred depending on composition.

    The use of such fused catalysts, produced by the oxidizing-melting of iron in the presence of promoters in an oxygen current after the manner of the ammonia catalyst, occurred to us.  They represent an excellent material because of the speed with which they can be produced and the ease and the exactness with which they can be reproduced.

    It would be going too far afield to cite all the experiments directed toward the recognition of favorable combinations of catalysts.  The investigations did not merely confine themselves to the modification of the ammonia catalyst by replacing the aluminum oxide by other oxides, such as MgO, CaO, BeO, etc. which were unreducible under the reduction and experiment conditions, but also studied the influence of additions of metallic oxides which are reducible like the iron, the actual catalyst metal: copper, silver, nickel and cobalt.

    The result of these experiments was that thenceforth we turned for the most part to fused catalysts containing magnesium oxide and alkali as promoters.  Relinquichment of aluminum oxide as promoter was caused by the observation that with it smaller total outputs with less straight-chained hydrocarbon chains were obtained.

    In investigating the iron-magnesium oxide-potassium-fused catalysts we had to consider two limiting cases.

1.    Catalysts in which the MgO content was 0, that is, which only contained iron and alkali, and

2.    Catalysts in which the alkali content was 0, that is, which contained no further addition apart from MgO.

1.    The catalyst free of MgO yielded products with the highest olefin value.  This may very well be explained by the high reaction temperature of 320 - 3? degrees requisite because of the failure of other promoters.  It is noteworthy that no soot formation occurred despite this stated high temperature.

2.    The catalyst free of alkali yielded a small amount of olefins, about 20 per cent ⁽¹⁾, and only few products boiling over 320 degrees the amount of which can be improved by increased additions of MgO.  With these catalysts we soon approached the products of the paraffin-producing precipitated catalysts.  This became especially marked when we turned to introducing acid additions into the catalysts.  Products containing only up to 10 per cent of olefins are obtained by melting in SiO₂.  The proportion of the total yield boiling above 320 degrees, however, was slight, and the 70 per cent straightchanedness of the obtained product left too much to be desired for this type of catalyst to be used in our investigations directed toward the production of paraffin.

    The two limiting cases show the markedly reaction-directing influence of the alkali addition toward olefin production.  Even a slight proportion of alkali in the iron-magnesium oxide mixture causes a speedy rise of the olefin content in the primary product, as well as an increase in the chain length.  This rise in olefin content, however, does not increase linearly with the alkali content but, starting with a catalyst free of alkali which produces 20 per cent olefins, rapidly attains an end value of about 75 per cent olefins.  This is a value that can hardly be out-stripped even with higher alkali contents.

    It is noteworthy that the modification of product quality in the above described manner is not dependent on the absolute alkali or magnesium content (i.e., dependent on the actual catalyst metal iron); the MgO:K ratio is decisive.

    This result was confirmed by a series of further experiments with higher MgO contents and also correspondingly higher alkali contents.  As long as the ratio was kept the same the same products were always obtained.

    In summary we give the result of several series of experiments:

Experiment Series with Iron-Magnesium Oxide-Alkali

    Catalyst space velocity 480:1; synthesis gas: 33 % CO

Cat. Comp. Fe MgO K %

Synth. CO temp. convers.

gram yield /Nm3

% Par. 320°

Straight chain

Par. %olef.

frac. % alc.

?

    In all the above catalysts used Hägg's carbide Fe₂C was shown X-rays to be present.

    If the alkali in the iron-magnesium oxide-fused catalyst is replaced by an acid constituent (SiO₂)

(Tables on page 163)

Cat. Comp. %Fe %MgO %SiO

Synth.  temp.

CO convers.

g. yield /Nm3

% Par. >320°

% Straight- chain

Paraffin % olef.

frac. % alc.

the olefin content will drop somewhat to rise again with greater SiO₂ additions, producing the following graph:

            Dependence of the Olefin Yield on the Acid and Alkali 

Content with the Iron-Magnesia-Fused Catalyst

(Graph on page 164)

    After the synthesis X-rays show magnetite (Fe₃O₄) along with Hägg's carbide to be mainly recognizable in these alkali-free fused catalyst.

    Roughly the same behavior can be established in regard to olefin content if, instead of completely leaving out the alkali in the MgO fused catalyst, its quantity is kept constant while the quantity of silicic acid is increased.

Cat. Comp. %MgO %K %SiO2

Synth.  temp.

CO convers.

g. yield /Nm3

% Par. >320°

% Straight- chain

Paraffin % olef.

frac. % alc.

(Table at bottom of page 164)

    Similar behavior can be expected with such fused catalysts where the quantity of silicic acid was kept constant and the quantity of alkali was varied.  As can be seen from the following table, however, the differences in the olefin content of the soft paraffin fraction are not very cnsiderable even though a slight rise in the expected direction occurred with an increasing amount of alkali.  A synthesis gas containing 50 per cent of CO as compared with 33 per cent in the previous experiments was used in this new series of experiments; the catalyst space velocity was 240:1 instead of 480:1.

(Table on page 165)

    The X-ray findings bear notice.  The catalyzing substance alters with the increasing alkali content.  Whereas magnetite predominated at the start, more and more Hägg carbide appears along with the magnetite until finally, at the point where only Hägg carbide was expected, the new hexagonal carbide made its appearance.

    In summary, we can say on the basis of the olefin values cited in the preceding tables that the only catalyst of importance for the recovery of unsaturated hydrocarbons ⁽¹⁾ is the fused catalyst, without addition of silicic acid, promoted with MgO and alkali and composed of Fe - MgO - K = 100 : 5 : 2.5.  The catalyst is reduced at about 500 degrees in about 48 hours with an hourly catalyst space velocity of 2000:1.  This is not to say that a thorough reduction is necessary.  Our investigations showed that catalysts with a degree of reduction of 30 - 100 per cent give practically equal yields of products of very similar composition ⁽¹⁾.  A reduction of synthesis temperature with fused catalysts - as in the case of precipitated catalysts - is attained by catalyst formation by means of a higher linear flow velocity of the reducing gas.  Nevertheless temperatures of 225-245 degrees are still necessary when synthesis gas is used with a space velocity of 200:1.

    In this last respect the behavior of A1₂O₃-containing fused catalysts is more favorable.  Their synthesis temperature is again lower by 10-20 degrees.  The olefin content of the primary product, however, was substantially lower ⁽¹⁾ than the olefin content attained with the MgO fused catalyst.  In this respect it departed from the series of the olefin-produced catalysts.

    Just as the addition of A1₂O₃ lowers the synthesis temperature, in like manner other additions to the fused catalysts act in other directions.  In this connection the action of fluoride additions in the form of CaF₂ ⁽²⁾ A1 F₃: FeF₃ ⁽²⁾ were of interest.  The former largely direct the synthesis toward low-boiling products at a low synthesis temperature, the other two, however, are rather characterized by the fact that they direct the synthesis toward the formation of low-boiling products at temperatures of 170-200 degrees however - and this is of scientific interest - in the case of these low temperatures, i.e., with slight CO conversion, only water appears as a side product and not carbon dioxide, as is usual with iron catalysts.  The latter only begins to appear increasingly with a rise in the CO conversion, i.e., at higher temperatures.

Alcohols:

    We can summarize very briefly the experiments conducted by us for the production of a primary product containing a substantial content of alcohols ⁽¹⁾ since the results achieved up to now are quite unsatisfactory.  Even when a higher percentage of alcohols occasionally appeared in one or another fraction, the fraction as such was so small that the gram yield of alcohols was of practically no importance as compared with other products. 

    The first experiments disclosed that rather large amounts of alcohols were formed especially at the start, i.e., at low temperatures; furthermore, that alcohol formation was promoted by a rise in pressure.  Therefore precipitated catalysts as well as fused and roasted catalysts were tested at high and medium pressure.  With none of these types did  we achieve our hoped for aim of substantially improving the quantity of alcohol in grams/Ncbm.  We strove to attain a catalyst which we could use at a low temperature since this was connected with a decrease in methane formation and an increase in total output as well as the formation of higher alcohols.  The latter is obtained on the level of previous investigations in particular with fused and precipitated catalysts, the former being superior to the precipitated catalysts.  The roasted catalysts are especially suitable for the synthesis of lower alcohols up to C₄; at lower temperatures, (slight consumption of CO) however, higher alcohols, which are of importance as starting materials for detergents, are formed up to 25 per cent of the total product.

Dr. A. Scheuermann

    In addition to the author, the following persons participated for varying periods of time in the researches of Ammonia Laboratory on hydrocarbon synthesis from carbon monoxide and hydrogen at atmospheric and medium pressure:

Dr. Bartholome	Dr. Marecek
Dr. Döll	Dr. Meisenheimer
Dr. Kärtkemeyer	Dr. Schmole
Dr. Kotzschmar	Dr. Vorbach

Table (1) (referenced on page 58)

233	1096a	190	71	1,0	6,5	0,8	139	61	100
269	1136	189	71	0	8,0	0,2	135	67	97
275	1138	182	67	0,6	5,6	0,4	130	60	97
297	1133	188	73	1,2	8,8	0,1	137	61	99
673	1444	180	75	1,0	9,5	0,1	133	58	95
Ale Mittelworte ergeben sich:					=0,324 kg/Ltr. Kt/Tag
	186°	71	1,0	7,7	135	61	98

Table (2) (referenced on page 58)

Die Verteilung des Primärproduktes auf die Siedebereiche

Vers. Nr.	-195	-320	-450	>450
233	16, 6	22, 1	23, 9	37, 4
269	15, 3	17, 4	26, 0	40, 3
275	19, 5	20, 2	26, 2	33, 9
297	18, 7	18, 5	24, 7	36, 6
673	17, 9	28, 4	24, 4	28, 8
Kittelwerte:	18%	21%	25%	36%

Table (referenced on top of page 59)

Vers. Nr.	Kont. Nr.	Temp.	Co-Verbr.	im %CO2	Abgas %CH4	5K2O im Kont. 100 Co	g Ausbente feste + fl. Prod./ncbm.
233	1096a	181	51	0	3,2	0,8	95
275	1138	174	42	0,2	2,8	0,4	86
297	1133	177	50	0,4	3,2	0, 1	98
673	1444	160	55	0	3, 1	0, 1	77
Mittelwerte:		173	50		3, 1	0, 1	89

Table (referenced middle of page 59)

Vers. Nr.	Kont. Nr.	Gehalt %K2O	%Co	Temp.	Co-Verbr.	im %CO2	Abgas %CH4	? Ausb. feste + flüss.Pr. /Ncbm	% Par. >320°	%Garad- kattigk.
325	1193	0, 1	40, 2	188	69	1, 2	6, 4	121	72	94
298	1132	0, 01	37, 6	177	79	0, 8	12, 1	139	67	100
297	1133	0, 1	30, 0	188	73	1, 2	8, 8	137	61	99
303	1134	0, 1	25, 1	182	74	0, 5	8, 9	236	62	100
299	1135	0, 2	20, 8	180	76	3, 4	9, 0	136	58	98

Table (referenced on bottom of page 59)

	-195	-320	-450	>450
325	12, 8	13, 9	19, 7	52, 7
298	17, 1	14, 2	21, 0	46, 2
297	18, 7	18, 5	24,7	36, 6
303	18, 4	18, 8	26, 6	35, 6
299	21, 9	19, 7	24, 1	33, 7

Ludwigshafen, July 10, 1947

Hydrocarbon Synthesis from Carbon Monoxide and 

Hydrogen With Iron Catalysts

(Gas Circulation Process and Foam Process)

Dr. W. Michael

 

Introduction

    The following material deals with researches conducted by the author from 1935 to 1944.  These notes were put down largely from memory because very little data is still available.  Valuable documents such as laboratory books and diaries of the pilot plant were completely lost.  They were shipped for the most part to Hassmersheim near Heilbronn to safeguard them from bomb damage and there they completely disappeared in the war disorders.  A small but very important part was lost in February 1945 when the house of the author was completely destroyed by bombs and fire.  Also missing, unfortunately, are many of the abridged reports and often the most important, which were preserved at other places.  Much of the material was recreated by questioning former collaborators who have been partly scattered throughout the country.

    For, these reasons we re compelled to omit the once abundant statistical material.  Nevertheless, the following description of the process should be sufficient.  What could no longer be verified was omitted.

    The first of the following synthesis processes that are described, the gas circulation, was reported in detail to the Standard Oil Co. of New Jersey, to Mr. Keith of the Kellogg Co., New York, and to the Shell Co. together with the pilot plant data.

I. The Gas Circulation Process

    The synthesis of hydrocarbons from carbon monoxide-hydrogen mixtures with cobalt catalysts after Fischer is characterized by gasolines with poor motor behavior and Diesel oils with good motor behavior.  Since the predominance of gasoline motors resulted in a demand for good gasolines, the production of synthesis products was based with this goal in mind.  From 1935 onwards the experiments conducted by the author also pursued this end.  Orientation experiments demonstrated that fundamentally better gasolines could be obtained by the use of iron catalysts, but their use presents us with disproportionately great difficulties.  The solution of these difficulties was still to be discovered.  These difficulties stem from the quite considerable heat of reaction of the hydrocarbon synthesis which necessitates an extensive division of the reaction chamber for heat removal.  This is effected in the case of the cobalt catalyst, which works below 200 degrees, by a narrow bundle of tubes combined with a parallel set of plates.  The iron catalyst, however, offers us greater difficulties.  Temperatures of 300 degrees and higher and medium pressures must be used if it is desired to produce a good grade of gasoline with the iron catalyst.  However, this is not possible with the Fischer apparatus because difficulties will arise after a little while which will make continued operation of the reaction impossible.

    These difficulties consist of two deleterious reactions which in principal may also occur with the cobalt catalyst if sufficient care is not employed.  However, these difficulties occur more readily and more intensely in the case of the iron catalyst.  These are: 

Methane formation and decomposition of carbon monoxide to soot.

    The more rigorous working conditions necessary for gasoline production by means of iron catalysts promote the appearance of these detrimental reactions.

    These difficulties, to which the iron catalyst is liable, furnish the reason why this metal which is so important as a catalyst in many reactions has heretofore not been employed in this synthesis.

The Reactions

    The basic reactions of hydrocarbon formation are as follows:

1.    CO + 2 H = CH₂ + H₂O + 41.0 calories

2.    2 CO + H₂ = CH₂ + CO₂ + 50.4 calories

    Reaction 1 occurs almost exclusively when cobalt catalyst is employed; reaction 2 predominates when iron catalyst is used.

    The heat of reaction of 2 exceeds that of 1 by about 9 calories, that is, by about 20 per cent.  The heat of reaction when iron is used, therefore, is appreciably greater than in the case of cobalt.  Equation 2 may be considered to occur by equation 1 and the following water gas reaction:

3.    H₂O + CO = H₂ + CO₂ + 9.4 calories

    At temperatures of 300 degrees and below the last equilibrium tends almost entirely in the direction of carbon dioxide.

    The detrimental reactions are as follows:

4.    CO + 3 H₂ = CH₄ + H₂O + 48,7 calories

5.    or 2 CO + 2 H₂ = CH₄ + CO₂ + 58.1 calories

6.    and 2 CO = CO₂ + C + 40.1 calories.

    With rising temperature the hydrogenation velocity increases more markedly than the aggregation velocity with which the CH₂ radicals are strung together in chains.  Consequently, CH₄ formation begins more and more to predominate with rising temperatures.

    Reaction 6 which results in the formation of soot starts at temperatures upwards of 350 degrees as a result of the decomposition of primarily-formed iron carbide.

    Methane formation, like soot formation, is markedly promoted by pressure.  They appear whenever heat accumulation and, as a result, overheating occurs due to deficient heat removal.  In this connection it must be noted that a certain amount of methane formation occurs at all temperatures where carbon monoxide reduction takes place and that its percentage proportion in the total production of the hydrocarbons drops in accordance with the drop in synthesis temperature.  Hydrogenation always occurs to a limited extent whenever iron catalysts are used.  Also a small amount of saturated hydrocarbons, possessing more than one C atom, are obtained with progressive increase of the degree of saturation toward the larger molecules.

    When iron catalysts are used the products contain a little oxygen.  The occurrence of this oxygen is attributable to the Oxo reaction which also occurs with the same catalysts at 120 degrees under pressure with olefins, CO and H₂, aldehyde groups -CHO being added to the double bond.  Since the double bond, as is shown by the cases investigated, is terminal, either straight-chain aldehydes or aldehydes with a CH₃ group in alpha position will be formed, depending on whether the aldehyde group is terminal or at the next to the last C atom.  Since the reaction conditions under which the synthesis takes place are not optimal for the Oxo reaction, only a partial formation of Oro products occurs, the aldehydes being hydrogenated to alcohols for the most part.  In addition, the Oxo reaction is paralleled by the acid formation by addition of CO and H₂O to olefins discovered by Reppe.  Occurrence of Canizzaro acid formation from the aldehydes formed by the Oxo reaction by disproportionation into alcohols and acids is also conceivable.  The ketones that occur are probably formed from the acids, the esters from the acids and alcohols or, according to Reppe, by addition of CO and alcohols.  It must be noted in this connection that, under the prevailing conditions, the higher alcohols readily split off water and turn into olefins.

Catalyst Conditions

    The Fischer Synthesis takes place at temperatures below 200 degrees and is largely conducted at ordinary pressure.  At these temperatures the higher boiling products are separated out on the catalyst in a liquid state.  They form a protective film on the catalyst which acts as a brake against the appearance of harmful reactions on the catalyst.

    When pressure is employed the proportion of higher-boiling products in the total product grows.  The film of liquid on the catalyst is thereby reinforced and a counterinfluence is created to the appearance of detrimental reactions promoted by pressure.

    Higher temperatures and pressure must be used if iron catalysts are employed for gasoline.  We begin to approach more closely conditions at which soot formation can occur.  The iron is obtained in a pyrophoric, i.e., highly dispersed state, in the reduction.  The synthesis reaction proceeds with a tendency toward the formation of CO₂, the heat of reaction rising as much as 20 per cent as compared with the CO catalyst.  The content of higher-boiling products decreases the higher the synthesis temperature selected.  Consequently, the conditions for the presence of a satisfactory film of liquid on the iron catalyst are not favorable.  Moreover, the reaction density, i.e., the amount of product and the production of heat in the time unit pro volume unit, is appreciably greater than in the Fischer Synthesis.  Hazards are thereby considerably increased.

    If the circumstance does occur in which these violent reactions take place, they will happen in the following manner:  Somewhere in the catalyst there is a narrow, as it were, point-sized, highly active area where a markedly heightened reaction velocity prevails.  In that spot the production of heat is extraordinary, but the heat removal, for some reason or other, is not satisfactory.  The temperature rises, the reaction thereby proceeding even more violently and, consequently, the temperature becomes even higher with a progressive increase in the formation of methane and soot until these are produced exclusively.  The soot acts as a heat insulator.  The overheating reaches over to the neighboring catalyst sections and leads to the identical result.  In spots temperatures may occur which are several hundred degrees higher than the requisite temperature. In short, the reactor must be shut down and the catalyst removed.

Experiences in Small-Scale Experiments

    These were the conditions at the time the first fundamental experiments were conducted on a small scale in tube reactors with a diameter of 15 mm. which were placed in an electrically heated aluminum core with automatic temperature control.

    Since the reactor output was too little without pressure, a pressure of 10 atmospheres was applied.  In time this was raised to 20 atmospheres and this pressure was maintained.

    A CO : H₂ = 4 : 5 mixture, as it occurs in water gas, was used as gas.  This gas composition roughly corresponds to the ratio at which CO and H₂ are consumed in the catalysis with iron.  Conversion of CO with H₂O therefore permits a saving of H₂.  The nitrogen content was largely 1 - 2 per cent.  There was no carbon dioxide of any importance.  Sulfur was removed from the gas to about 1 mg. per cbm.

    To retain any iron carbonyl which might be present since it decomposes to form a highly active catalyst in the reactor which forms soot, the gas was run into a vessel containing activated charcoal.  After that it was passed through a dust filter.

    Highly active precipitated catalysts containing reduction-resistant oxides as well as some alkali apart from iron were used as catalysts.

    It was demonstrated that at temperatures of 300 degrees and higher - such temperatures were necessary to obtain good gasolines - catalyst decomposition and soot formation occurred after a time, e.g., one day.

    An attempt was made to counteract this difficulty by adding all sorts of additions to the iron.  However, success was only achieved with those additions which caused a decrease in output, i.e., the "success" consisted only of the fact that the reaction was subdued and thus the tendency toward uncontrolled temperature increase was diminished.

    As a result of the arbitrary arrangement of the catalyst particles the gas does not follow a straight course in the Fischer tube, but - as was shown by experiments with aerosols in glass tubes - travels from the center to the wall and then back, heat transport occurring therefore not so much by radiation and conduction as by convection.  There is naturally no regularity.  It can therefore come about that the gas in the tube axis may remain for a long period of time in the middle of the tube before again traveling to the wall to give up heat.  In such instances the temperature difference between center and wall is higher than usual and may become dangerously high for the iron.

    The circulation conditions naturally are improved if higher flow velocities are used, gas turbulence being promoted in this manner.  At the same time, however, there is increasing danger that any hot points present may be fanned and thus be heated to a dangerously high point, just as a glowing spark may be fanned to bright heat by blowing.

The Iron - Sinter - Catalyst

    At this point an attempt was made to eliminate the difficulties by physical means, namely, in the following way.  The catalyst was to be assured of good heat-conducting properties in all its parts until no more isolated highly-active catalyst particles with too little heat capacity existed.  In this way any point-sized overheating which might occur would be rendered impossible by the removal of the heat.

    This effect was achieved by the reduction with hydrogen at high temperatures of a catalyst prepared by precipitation.  A temperature of 800 - 850 degrees was found suitable at which reduction was conducted for about 4 hours (File numbers 9131, 9261, German patient 729290; file numbers 9919, 9960, 10189, 10190, German patient 763688).  The catalyst was sintered thereby and showed rounded corners and edges and the entire texture became denser.  The fine structure was coarsened.

    This catalyst justified expectations.  It achieved the same output as the unsintered catalyst and reaction conditions became so stable that for months on end there were no difficulties in the way of operation.

    Catalyst preparation was simplified by using iron powder as starting material which was obtained by the decomposition of iron carbonyl in heat.  It consists of iron globules of about 2 - 5 / u diameter and under the microscope revealed a polished surface like that of drops of mercury.

    This powdered iron was pasted with water containing about one per cent dissolved borax, formed into flat cakes, and then cut into cubes with a lateral length of 1 cm.  It was then sintered in a cylindrical electric oven with many trays for several hours at 850 degrees in a hydrogen atmosphere.

    Catalysts which were prepared in the described manner met all requirements in regard to output and stability.  Addition of third substances, such as alumina, etc. did not result in any improvement but, if used to any great extent, caused a regression, and was therefore omitted thenceforth.

    Also, the iron fused catalyst, as it is obtained for the ammonia synthesis by burning iron to Fe₂O₃, introduction of alumnia and alkali into the melt and reduction with hydrogen, is suitable.  It is recommended, however, that it be sintered before use in order to attain greater stability.

The Gas Circulation Process

    In the case of the sintered catalyst an investigation was conducted to see if it might not be possible to use wider tubes in order to cheapen the design of the tube reactors.  None of these experiments, despite partial successes, led to any satisfactory result so that a new method of heat removal was undertaken.

    It consists of the following: The interval of time between heating and removal of heat, which in the tube reactor is left to change, is now controlled.  The gas is conducted from the catalyst to the condenser in short regulated intervals, overheating thus being avoided with certainty.

    This is accomplished by a rapid circulation between catalyst space and condenser.  The sojourn in the reaction chamber is regulated to so short a space of time that possible heating of the gas cannot exceed a narrow temperature range, for example, 10 degrees.  The gas then passes to the condenser and is there cooled down to a point where it is the same as at the entrance to the catalyst space (file number 10824).  (Drawing on page 166)

    The sketch shows the system.  If heating of 10 degrees is permitted the gas must be recirculated for about 100 times until it has finished reacting or, what amounts to the same thing, 1 per cent of the circulating gas is constantly diverted and 1 per cent of make up gas supplied in its stead.  Consequently, nowhere in the system is virginal gas found, and this fact alone renders it impossible for the reactor to get out of control.  Nevertheless, the capacity of the reactor is not inferior to that of a tube reactor with an equal catalyst volume.

    Contact time of the gas in the reaction chamber is only about 1/2 of 1 second.  Appreciable gas velocities are necessary therefore, and it is advantageous not to make the catalyst bed to high.  In general, measures should be taken to keep the flow resistance of the gas in the catalyst arrangement as low as possible.

    The catalyst space can be enlarged as much as desired at right angles to the gas flow.  Its shape is extremely simple, consisting of an empty cylinder with one or more nets for the reception of the catalyst.

    For reasons of economy the by-pass tube through which the gas is recycled should be constructed more narrowly than the reaction chamber.  1/10 of the reactor cross section was selected.  In this way the gas velocity is stepped up tenfold.  In its passage into the further catalyst space with its much slower flow the shape of the space must be so chosen that no hazardous eddies are created which might impair the required laminar flow through the catalyst bed.  However, laminar flow is necessary, for only that guarantees that the gas will flow through the catalyst bed with uniform speed throughout the entire cross section and will not stay too long in one spot causing it to become too hot.  Above all, as experience has demonstrated, change in direction of the route of the gas and widening of the cross section must not coincide or violent eddies will arise, rendering impossible a uniform flow through the whole cross section.  This will lead to local gas stagnation and heat accumulation in case the catalyst bed is situated directly under an eddying area of this kind.  Such a case will be discussed later in describing the experimental reactors.  To keep the gas path as simple as possible an axial blower in preference to a centrifugal should be selected in order to maintain a laminar flow in the whole circulation system if possible.  In carrying out this reactor system on a large scale it is advantageous, for the sake of better control and attendance, to place the motor on the outside of the pressure chamber and to pack the shaft leading into the gas chamber by means of labyrinths.  The motors in the pilot plants were placed in the pressure chamber to ensure simplicity.

    Apart from the above-described "hot" circulation a "cold circulation" was also maintained on a smaller scale.  In this a small portion of the recycle gas was cooled down to water temperature in order to separate the water and the higher-boiling products from the recycle gas.  It was carried out in such a manner that, outside of the one per cent of the circulating gas which was supposed to leave the reactor constantly, three more per cent was conducted into the product separator and there, after cooling to the temperature of cold water, returned to the hot circulation in a heat exchange with the outgoing gas.

    The conversion in the reactor can be intensified if the carbon dioxide is removed from the returning gas after cooling.  If the fresh gas contains little inert gas, such as N₂, CH₄, the gas in one stage can be converted to considerably more than 90 per cent with eventual intensification of the recycling.  It would be more advantageous to add another synthesis stage only if there were an increase in the content of inert gas.

Expenditure of Power for the Circulation

    To obtain a picture of the amount of power requisite for the recirculation it is expedient to compare it with the compression power.

    If a mol of a gas is compressed to 20 atmospheres in two stages, theoretical compression energy will amount to 2.8 watt hours.  If the ? which is under a pressure of 20 atmospheres is subjected to a more ?se compression of 0.1 atmospheres by the blower after completion of ?cle in order to compensate for the loss of pressure, energy of the ? of 0.7 watt hours will be consumed.  This will be increased to watt hours if a second stage is added.

    The ratio:  circulation power
                    compression power 

?ts to about 30 per cent and dwindles to about half if the gas resistance diminished, that is, to 15 per cent.  This is the case with the muti-? reactor which will be described later.  Additional savings of power ?e effected by special catalyst arrangement as well as by increase of temperature span in the reaction chamber.

The Steam Boiler

    In the Fischer apparatus there is a temperature difference of a few degrees between the extensively subdivided catalyst space and surrounding steam boiler water.  In the gas circulation process the recurrence can be considerably increased, for example by 50 degrees.  This way it is possible to keep the heat exchange surface with the ?ng water relatively small and to cheapen the steam boiler considerably without the stability of the temperature suffering.  In this way the steam temperature is lower but its temperature, as a result of the appreciably higher reaction temperature in gasoline operation, is nevertheless much higher than in the Fischer Process.  This is of use when it is employed for power.

Operation of a Gas Circulation Reactor on a Laboratory Scale

    Circulating reactors of five liters capacity were constructed with an "internal circulation" for laboratory purposes.

    The gas circulation, as is shown by the attached diagram, (p 167) : accomplished by a centrifugal blower above the catalyst chamber.  The blower forces the gas through an annular catalyst chamber and sucks it back through an axial hollow chamber.  Regulation of the circulation speed is accomplished by a throttle valve situated in the hollow space.  The motor is installed in a special chamber connected with the reaction chamber.  The driving shaft passes through the narrow connecting space with so little play that the fresh gas flowing from the motor chamber into the reaction chamber suffices to prevent the circulating gases from penetrating into the motor chamber.  Water cooling was installed at the constriction to protect the motor chamber from the heat of the reaction chamber.  The reaction chamber is surrounded by an annular electric heater.  Heat removal at this reactor size is accomplished by radiation.

    The reactor was equipped with automatic temperature regulation so that is was possible to maintain the reaction temperature at the same point for months at a time.  The motors were of the alternating current type with short circuit rotors.  In one case one of these motors ran for 10 months without interruption.

    The reactor gases were withdrawn at the base of the reactor.  This was followed by a steam-heated separator for the reception of the high-boiling products and the middle oil which would solidify at ordinary temperature.  After that for the recovery of the gasoline there came a water-cooled separator after which pressure was let down.  The let down gas was then dried in a cloride of lime tower and then passed through a glass separator to remove the light gasoline and the liquid gases.  The glass separator was cooled with a carbon dioxide-alcohol mixture.  Beyond this was situated the gas meter which measured the amount of the outgoing gas.  The quantity of the feed gas could be calculated from this.  Later the quantity of feed gas was directly recorded by dry gas meters constructed for pressure.  The gases were frequently subjected to an additional cooling with liquid air in order to recover completely the hydrocarbons - even to the methane which was incompletely retained.  The abundantly present carbon dioxide was removed beforehand in order to avoid any difficulties with it.

    As for the purification of the fresh gas, it can be stated that it was received practically completely purified of sulfur (down to 1 mg/cm³).  It was repeatedly sent through activated charcoal, caustic potash and finally through cellulose wadding.  The composition, which corresponded largely to that of water gas, namely, CO : H₂ = 4 : 5, was produced by mixing the pure gases or by admixture of CO to the fresh gas of the isobutyl oil plant which was prepared in a CO : H₂ = 1 : 2 ratio.  The nitrogen content amounted to about 1 - 2 per cent.

Operation of the Reactors on a Commercial Scale

    In large scale operation we departed from the "internal circulation" because in an exterior circulation the individual parts of the apparatus could be controlled more closely and are more readily accessible.  The system, it is true, will then require more space and the steel requirements in general will be greater.  On the other hand, the construction of the individual apparatus parts becomes very simple.  A cylindrical vessel which is resistant to pressure with one or more built-in grates on which the catalyst can be charged suffices as a reactor.  We can manage with a relatively small tube condenser for cooling purposes.  This condenser must be designed for greater team pressures as a result of the higher working temperature.  An axial lower is recommended for maintenance of the gas circulation at a pressure difference of about 1 m. water column for technical flow reasons.

    The first pilot construction was a reactor for 400 liters of catalyst.  The arrangement of reactor, condenser and blower was the same as in the schematic drawing of the gas circulation process.  Just as in the laboratory, the blower motor was placed in the pressure chamber for the sake of simplicity.  The fresh gas was conducted into the motor housing in order to render it impossible for the circulating gas to flow into it.  The catalyst had a bulk height of 80 cm.  The gas flowed through the catalyst bed from top to bottom.  Where the gas from the circulation returns to the reactor violent ?dies might be caused.  For this reason various devices have been installed in order to cause the gas to flow through the catalyst bed in a uniformly minor manner.  (See the attached drawing.p. 168)

    The fresh gas was added to the circulating gas in an amount of roughly 1 per cent.  The gas entered the catalyst chamber at 320 degrees and left at 330 degrees.  In the subsequent condenser it was cooled down to 323 degrees, and by admixture with the cold fresh gas it lost another 3 degrees, returning to the reaction chamber at 320 degrees.  1 per cent of the circulating gas is constantly removed after the condenser and conducted to the separators ⁽¹⁾. The first separator was maintained at 100 degrees with steam, the second at the temperature of cold water, and the third was cooled with ammonia to -40 to -45 degrees.  However, the steam was removed from the exit gas with calcium chloride beforehand in order to avoid clogging of the low temperature condenser by formation of ice.  At this point the gas was naturally not depressured before low temperature cooling as was the case in the laboratory apparatus.  For the exit gas was to be conducted to a second synthesis stage or be recycled to the first reactor for the purpose of complete utilization after removal of the carbon dioxide with a high pressure water wash.

    In starting, the reactor was brought to the temperature in the following manner:  steam at a pressure of 20 atmospheres was conducted through the condenser - steam of higher pressure was not available - and the circulation was set in operation.  The missing temperature span, up to the point where the reaction set in, was bridged by an electric heating element placed in the gas circulation.  Once the reaction was in operation and the coooling regulated, the temperature remained steady and the entire system had so much inertia that it took a few hours before a slight regulation was needed.  Gas conversion ran to 78 - 82 per cent and the CO₂ content of the exit gas 30 - 45 per cent.  The product obtained per liter of catalyst space daily amounted to 0.8 kg with an hourly synthesis gas feed of 270 liters per liter of catalyst space.

    Occasionally a second synthesis stage was attached to see what degree of conversion could be attained.  To this end a pressure water wash as attached in order to remove CO₂ and a 5 liter circulating reactor was ?ressed into service for want of a suitable apparatus.  A suitable gas stream was branched off from the total exit gas.  The total conversion attained together with this second reactor amounted to 92 - 93 per cent.  The output per liter catalyst space per day was only 0.6 kg product as a result of the higher content of inert gas.  This is 3/4 of the output of the first stage.  The exit gas consisted largely of CO₂, CH₄ and N₂.  The CO disappeared almost entirely.  157 grams of utilizable crude product was obtained per thermal cubic meter of ideal gas (CO and H₂ without inert gases) passed to the reactor.

    Noticeable amounts of gaseous olefins are present in the circulating ? after intense cooling and these are lost in the water wash since the solubility of these olefins in water is not inconsiderable; indeed, the larger the molecule the greater the solubility.  For this reason experiments with alkasid liquor were conducted, the alkazid liquor possessing a pronounced effective solubility for carbon dioxide.  It can also be regenerated by heating.  ? quantity of water employed in this instance is so slight that no noticeable ?ses of gaseous hydrocarbons can occur.  The experiment was conducted for about 8 weeks without noticing a drop in the action of the liquor.

    Toward the end of a two-month operating period the catalyst generally showed a distinct decline in productivity.  Its surface had become coated with material which could be extracted with hot benzol or xylol, whereby it recovered its original activity.  In operation at 300° and lower temperatures this condition occurred in a shorter time.  On this basis the previously described ? recycle was incorporated.  Thereby we succeeded in keeping the catalyst ?ace clean for a much longer time.

4 Cbm Reactor

    As a further step in approaching the conditions of large-scale production, a reactor for 4 cbm catalyst was constructed.  Its design, which is shown by the attached diagram, was fixed by the viewpoint of the best possible space utilization with the least gas resistance.  For this reason the reactor and the catalyst bed were so arranged that the catalyst bed made up the bulk of the available inner space.  The height of the catalyst bed was 80 cm., and the entire catalyst bed contained 4 m³ of catalyst.  It was assumed that the pressure difference existing between the upper and the lower side of the catalyst bed would cause a uniform penetration of the catalyst bed in all its parts.  Here too cooling was effected by water which was supplied in a high pressure coil situated in the gas circulation.  Steam was produced at a pressure of 60 atmospheres.  Heat control was effected by regulation of the steam pressure. (Diagram p. 169)

    In this operation provision was also made for a so-called cold circulation which, as was already explained above, consists in a small portion of the gases from the hot circulation being cooled down to the temperature of cold water in order to separate out water and high-boiling products.  During the contact time in the reactor system the gas was passed about three times through the cold circulation.  In this manner the accumulation of high-boiling products on the catalyst was greatly protracted.

    The reactor was started by conducting 20 atm. steam into the condenser and into an electric heater.  The gas circulation was set going and in a few hours the apparatus reached a temperature at which the reaction began to set in, whereupon the normal working temperature was attained.  The reactor produced about 3 tons daily.  After 7 - 8 weeks the CH₄ content of the exit gas rose with simultaneous falling off of the yield.  A thermoelement showed abnormally high temperature.

    The reactor was therupon shut down and opened.  At the lower part of the catalyst bed, more toward the gas entry side, there was discovered a pocket of soot which was encroaching more and more on the surrounding territory.  At the center of the heat development the catalyst particles were found thickly encased in soot, the particles being partially decomposed with formation of scale.  It was evident that the gas velocity in flowing through the catalyst bed was not uniform and was so slight at the spot in question that the foreseen temperature rise of 10 degrees was far exceeded.  As a result carbon monoxide decomposition and marked methane formation set in in time.  The soot which was in process of formation rendered the gas stagnation worse and the deleterious reactions began to spread to the surrounding territory.  The following factors are of importance in explaining this manifestation:

    The passing of the circulation pipe line into the reactor caused a tenfold increase of the flow cross section.  Moreover, the gas was forced to change its direction at the same spot.  The sudden retardation of velocity evidently caused the formation of violent eddies which, promoted by the space arrangement, led to a marked disturbance of the flow in the catalyst bed.  An above-normal gas velocity probably prevailed in one sector, whereas the gas almost stood still in another part and became too hot as a result.

    To eliminate this difficulty baffle plates were installed to curb the violent gas movements.  Moreover, the catalyst bed was divided into 3 horizontal layers separated by thin empty spaces for pressure equalization.  Conditions were bettered as a result but difficulties were not entirely eliminated.  After about two months of operation the beginnings of local soot formation were again noticed at approximately the former spot.

    It was therefore concluded that the form of the reactor in use was unsuitable in principal and that it must be replaced by another which would take into consideration the knowledge acquired.  The following design was contemplated:  An upright cylinder to contain alternating layers of catalyst and cooling elements, for example, 7 one upon another, or 7 systems superimposed one upon another in series, as is shown in the attached drawing.  In this way it is possible to use high and wide cylinders and to maintain a laminar flow.  An irregular gas flow can only occur at the top where the recirculated gases enter.  This hazard can be eliminated by installing baffle plates.  (Drawing p. 170)

    Another way of carrying away the reaction heat would be to cool with admixed cooling gas, thereby surrendering the use of the heat of reaction.  After a temperature rise of 10 degrees about 3 per cent cooling gas would have to be supplied and thoroughly mixed with the aid of baffles.  Since a temperature rise of 10° is equivalent to a gas conversion of about 1 per cent, only one per cent of the 3 per cent cooling gas need be make-up gas.  The 2 per cent, which is withdrawn from the gas in process of conversion, is cooled and again mixed into the hot gas.

    Cooling by cooling gas can be combined with the gas circulation, it being thus possible to omit the installation of cooling elements between the catalyst beds in multi-stage reactors.  The cooling elements can be replaced by mixing baffles.  In addition another solution was contemplated.  The catalyst could be used for a very short time, with drawn and regenerated, or replaced by a cheap catalyst.  For this purpose reactor designs should be used which would permit charging and removal of the catalyst without interrupting operation.  Attempts were made to regenerate spent catalyst containing carbide and oxide by oxidation with air and subsequent reduction (File number 11217, German patent 756609).  Good results were attained insofar as the oxidation, which was carried out at 500 - 600 degrees, penetrated to the inmost core of the entire catalyst particle.  Otherwise the particle decomposed at once when it was used again with attendant scale formation.  An alternative procedure consists in burning the spent catalyst with pure oxygen with attendant melting in order to produce a fused catalyst, such as is employed in the ammonia synthesis.  In every one of these cases the question must be investigated whether, in view of the cheapness of iron and of the moderate requirements in regard to purity necessary if the contact time in the reactor is kept short, it pays in general to regenerate spent catalyst, especially if a cheap processing of compact iron represented by the fused catalyst is contemplated.

    It was not possible to undertake experiments with new reactor designs as a result of wartime restrictions.  Moreover, the requirements of war, as will be explained below, directed the synthesis researches into new paths.

    Special steels are not necessary as construction material for the reactors.  Ordinary steel can be used as long as it possesses the tensile strength necessary for the temperatures in question.  Experiments which extended over a period of years showed that neither carbonization or decarbonization of the steel occurs under these conditions.

Composition of the Products of the Gasoline Synthesis

    The products obtained in the synthesis showed certain variations in their composition.  The following table gives an average of the reaction conditions:

Pressure:	20 atm.
Mean temperature:	325°
CO : H2                                           =	4 : 5
Output per day per liter catalyst space:	0.8 kg.
Catalyst:	Sintered catalyst fr carbonyl iron with 1% of sodium borate

Composition of the Useful Products (in percentages by weight)

    In addition there is a gasification (methane and ethane) to the extent of 18 - 20 per cent (Total product counted as 100).

Description of the Products

    The considerable yield of markedly unsaturated gases and the not inconsiderable amount of iso-butylene is noteworthy.  Polymerized gasoline can be produced from the olefins, as well as isoctane from the iso-butylene.  The butane can be used for alkylation.

    45 per cent of the gasoline constituents boil up to 100 degrees.  In the crude state it possesses an octane number of 70 - 75 and still contains 2 - 4 per cent of oxygen in the form of alcohols, acids, esters, aldehydes and ketones.  It needs refining before it can be used in a motor.  This is accomplished by passing the gasoline in the vapor phase over active alumina and, in addition, passed over alumina (1) mixed with Cr₂O₃ and 2nO at 380 - 400 degrees ⁽²⁾ in order to improve the odor.  The oxygen in the form of H₂O, CO and CO₂ can be removed leaving an insignificant remainder of less than 1/2 per cent.  The loss of weight of the gasoline amounts to 5 - 6 per cent.

    By this treatment the octane number (Research) rose to 84 - 86 and the octane number (Motor) to 75 - 78.  Subsequent refining with granosil at 200 degrees ⁽³⁾ and admixture with a stabilizer ⁽⁴⁾ produced a satisfactory storage-stability which was confirmed by a two-year period of storage during which time it stood up to all tests.

    The middle oil as Diesel oil had a cetane No. of 48 - 50.  Viscosity, Engler degrees at 20 degrees was 1.16.  The oxygen content approximated 2 per cent.  The small quantity of high-boiling products had the character of a mixture of paraffin and a light machine oil.

II.  The Foam Process

    After the outbreak of the war a shortage of soaps and detergents was soon felt which was caused by the falling off of imports of fats and vegetable oils.

    To make up this shortage alcohol sulfonates were produced in Leuna which possessed excellent detergent action and insensitivity toward lime.  The raw material was middle oil of the Fischer Synthesis which was completely saturated by hydrogenation.  It was subjected to solfochlorination in the presence of ultraviolet light.  In this manner alcohol sulfonates were obtained.

    Since the middle oil obtained with the iron catalyst is largely unsaturated with the double bond terminal, as was shown by investigation, an attempt was made at direct sulfonation with sulfuric acid.  The results were successful.  Sulfuric acid esters were obtained which were the equal of alcohol sulfonates in regard to detergent action and, on the whole, cheaper in price.    

    The gasoline process was not at all suited to the production of middle oil for this purpose.  The yield is relatively low and experiments demonstrate that the middle oil produced at lower temperatures is more straight-chained and therefore qualitatively 

 

    The middle oil as diesel oil had a cetane No. of 48 - 50.  Viscosity, Engler degrees at 20 degrees was 1.16.  The oxygen content approximated 2 per cent.  The small quantity of high-boiling products had the character of a mixture of paraffin and a light machine oil.

II. The Foam Process

    After the outbreak of the war a shortage of soaps and detergents was soon felt which was caused by the falling off of imports of fats and vegetable oils.

    To make up this shortage alcohol sulfonates were produced in Leuna which possessed excellent detergent action and insensitivity toward lime.  The raw material was middle oil of the Fischer Synthesis which was completely saturated by hydrogenation.  It was subjected to sulfochlorination in the presence of ultraviolet light.  In this manner alcohol sulfonates were obtained.

    Since the middle oil obtained with the iron catalyst is largely unsaturated with the double bond terminal, as was shown by investigation, an attempt was made at direct sulfonation with sulfuric acid.  The results were successful.  Sulfuric acid esters were obtained which were the equal of alcohol sulfonates in regard to detergent action and, on the whole, cheaper in price.

    The gasoline process was not at all suited to the production of middle oil for this purpose.  The yield is relatively low and experiments demonstrate that the middle oil produced at lower temperatures is more straight-chained and therefore qualitatively superior for this purpose.  Since appreciably more middle oil is simultaneously formed at a lower temperature, we were naturally faced with the problem of reducing the synthesis temperature as much as possible.

    An economical gas conversion with the sintered catalyst used in the gas circulation can no longer be achieved below 280 degrees.

    It was therefore decided to carry out the reaction in the liquid phase which would permit the use of the iron catalyst without hazard in a highly activated form in a finely-divided, to a certain extent, colloidal dispersion which would guarantee absolute temperature control.  In addition, the catalyst is continuously extracted of its own accord.  Particular attention, however, must be directed to an effective contact between liquid and gas since the solubility of the gas in oil under the conditions of the synthesis is very slight.  In consequence the gas must be mechanically dispersed in the liquid in the form of very fine gas bubbles.

Experiments with Nozzles

    Gas dispersion was first attempted by means of fine nozzles.  Variously arranged nozzles were placed into the reactor at different spots.  However, satisfactory fineness of the gas bubbles was not attained.

The Stirring Reactor

    At this point an attempt was made to apply the principle of the stirring reactor which had been proved to be reliable in other processes.  An upright cylindrical reactor was used with a shaft projecting settling of the catalyst which would otherwise occur.    

    The best foam (dispersion) plates were those sintered together from quartz sand with glass as the cementing agent.  They were mostly employed in a round conical shape, 4 - 5 cm thick.  Asbestos foil served as packing material.  At first the varying heat expansion of iron and foam plate caused some difficulties.  It was possible to place several such plates on the bases of reactors with large cross sections.  Moreover, a strong iron grating was ground onto each plate in order to enable it to withstand the existing pressure differences without breaking.  A pore size of 0.1 and 0.15 mm. is suitable.  The pore diameter of 0.2 mm, shows a noticeable drop in output.

    Attempts were also made to employ filter candles in place of filter plates.  However, it was established that an upright pore wall, and, to an even greater extent, a downwardly inclined pore wall produces larger bubbles.  But since a filter candle contains such surfaces whether it is in a vertical or horizontal position, its use for this purpose was abandoned.

The Reactors on a Laboratory Scale

    The reactors for the foam process on a laboratory scale possessed the essential features of the process:  Gas dispersion by means of dispersion plates, slurry circulation and degassing of the slurry.  The slurry circulation takes place by means of a central tube introduced into the axis of the tube.  The fine gas bubbles produced by the dispersion plate ascend in the annular part of the reactor cylinder.  Because of this the slurry in this part of the reactor became lighter than in the central tube.  A circulation pattern following the principle of the "Mammoth" pump occurred, the slurry ascending in the annular part, being degassed at the top, and again returned to the bottom through the central tube.  The temperature was determined by means of thermocouples which were placed at varying levels in the reactor and from these it was possible to establish whether the slurry circulation was functioning or not.

    Apart from these 3 m high reactors with a capacity of 3 liters, 10 and 25 liter reactors constructed on the same principle were employed.

Reactors on a Semi-Works Scale

300 Liter Reactors

    On completion of a 300 liter reactor constructed according to the same design as the laboratory reactors it was found that automatic slurry circulation could no longer be attained with certainty and that if the circulation was broken for one reason or another, it could no longer be set in motion because of the sinking of the catalyst in the meantime.

    The following occurred then:  As a result of the gradual settling of the catalyst the catalyst concentration in the reactor, especially in the lower part, was higher and the slurry was therefore heavier than in the circulation in which the slurry is supposed to flow downwards.  As a result the buoyancy of the gas bubbles no longer sufficed to set the circulation in motion.

    A circulation pump was therefore placed in the slurry circulation which was shifted to the outside of the reactor for readier accessibility.

1.5 cbm Reactor

    The 1.5 cbm reactor was constructed according to the dispersion principles already described.

    The reactor was a cylinder 8 m in height with a capacity of 1.5 cbm; the gas separator had a capacity of 600 liters.  The slurry circulation line had a cross section 1/10 of the reactor.

    Two circulation pumps coupled in parallel, one serving as a reserve, were used to pump the slurry around.  Each pump could be interchangeably disconnected.  At first some trouble was had by the stuffing boxes because hot reactor liquid and catalyst penetrated the stuffing boxes and acted as an abrasive on the shaft.  Moreover, there was also the danger that the slurry which passed to the outside would ignite as a result of the pyrophoric iron contained in it if it ran into absorptive lining, e.g., the insulation of the reactor.  The catalyst would dry up as a result of the absorption of the liquid and ignite.  This difficulty was solved by perforating the packing and slowly forcing into it middle oil which stemmed from its own production and was not a foreign substance.  A portion of the forced-in oil went inside and thus actively prevented penetration by the reactor slurry into the packing box.  From time to time it was necessary to withdraw the stuffing box and replace the packing at intervals.

14 cbm Reactor

    Finally, a 14 cbm reactor was constructed for large-scale operation.  It can be visualized by the attached drawing.  The reactor was not put into operation because a sufficient supply of synthesis gas requisite for the operation of the reactor could not be released on account of the war.

(Drawing p. 173)

The Catalyst

    The catalyst in the liquid phase can be used in the most highly ?ted and the most finely divided form without fear of danger, as is the case in the gas phase.  This intense activity can be attained in several ?.  The catalyst, for example, can be produced by the conventional method percipitation.  Finely divided colcothar, as obtained by combustion of carbonyl vapor, can be used ⁽¹⁾.  Iron carbonyl vapor can also be ?osed directly in the reactor slurry ⁽²⁾.  The latter is the simplest and the cheapest process and gives the best catalyst.

    The catalyst was produced from colcothar in the following way:  ?har was pasted in water containing 1 - 2 per cent sodium- or potassium ?, kneaded, formed into cakes, dried, cut into smaller pieces and ?d at the lowest possible temperature with considerable hydrogen ?p the water vapor pressure low.  The pyrophoric catalyst, after cooling withdrawn from the reactor protected by carbon dioxide, introduced middle oil of its own production and ground to a highly divided ?e fineness of 2/u.  Greater fineness could not be attained by grinding.  Settling velocity in middle oil was tested in standing cylinders.  In ?ay a gage for the fineness of the grinding is obtained.

    The requisite amount of catalyst suspension was charged into ?actor ⁽³⁾ into which a previous charge of liquid had been introduced.  When the reactor was first put into operation it was charged with purchased paraffin.  As production progressed it was more and more displaced by the new formation of slurry.  When iron carbonyl was used it was decomposed in the slurry itself.  The requisite alkali was run into the slurry in the form of alcoholic potassium hydroxide or sodium hydroxide solution.

    Additions of other substances to the catalyst were tested in various instances.  However, their addition did not result in any noticeable change in the results.

Slurry Circulation with Fixed-Bed Granular Catalyst

    Attempts were also made to use the catalyst in a fixed granular form, to fix it in the reactor, and to effect the liquid circulation by a central tube.  If this is done a suspended catalyst is not used, the liquid is much lighter and has less friction, and the liquid circulation therefore proceeds automatically without any trouble according to the principle of the mammoth pump since the liquid charged with gas bubbles which flows upward through the catalyst bed is specifically lighter than the degassed liquid that flows downward through the central tube.  An important simplification of the apparatus was achieved by omission of the pump.

    The experiments on a laboratory scale produced good results, not at all inferior to those attained with a suspended catalyst.  The liquid circulation was conducted with absolute safety.

    A 300 liter reactor was constructed in order to test the process on a larger scale according to the same principle.  The liquid was recycled axially.  The catalyst was placed on a number of perforated plates placed one on top of the other.  The reactor was never put in operation on account of the war.

Foam Process Experiences

    The temperature inertia of the apparatus in the foam process is ordinarily great and upsets are not likely even if attendance is ?ed.

    The gas mixture was established with an excess of CO in order in a higher percentage of olefins in the middle oil.  A CO : H₂ = as a composition was usually used.  At 250 degrees 0.35 - 0.45 kg product contained per liter catalyst space per day.  At 240 degrees about 0.35 kg are obtained.

    However, a difficulty occurred after two months operation.  The ? began to dwindle and irregularities appeared in the temperature ?ution.  The reactor was shut off and opened.  The following picture showed: A ring-shaped coating of catalyst several cm. in thickness had formed at the reactor wall.  The thickness of the coating lessened below disappeared in the vicinity of the dispersion plate.

    Deposits had also been formed in the gas separator.  Catalyst ? to parts of the separator free of liquid.  This catalyst deposit ?ed an opportunity of further reaction to the already reacted gas.  ?ed to overheating as a result of the lack of heat removal.  A portion ?i the cooler which contained a cooling coil.  No deposit appeared in slurry circulation where the slurry, as a result of the narrower cross ?n, had a higher velocity.

    The deposit was examined and was found to consist chiefly of ?i FeCO₃, that is, iron compounds, which are stable in the reaction gas prevailing in the upper section of the reactor and which contains an abundance of CO₂ and H₂O vapor(1).

    The deposit was readily crushed and did not give the impression that it was held together by crystalline forces.  We were tempted to assume the formation of an adhesive, high-molecular, organic substance, probably containing oxygen, which was sparingly soluble in the reactor slurry and which was stable in the atmosphere of the upper part of the reactor.  The attempt to find such a substance, however, was not successful.

    An attempt was made at this point to overcome this difficulty by installing stirrers which were constantly to slowly scape the reactor wall.  In this way the formation of deposit was to be prevented.  However, after about three more months it turned out that the reactor jacket itself remained clear but that thick deposits were formed on the stirrer rods and that these deposits in part dropped onto the dispersion plate and were detrimental to the gas dispersion.  There was naturally much less opportunity for the formation of deposit than before but another device would have been needed to keep the stirrer clear.  More and more methods were supposed to be tried to obviate this difficulty but they could no longer be put into effect as a result of the accumulating war exigencies.

    The reactor was to be made shorter in order not to force excess gas conversion in one pass and to avoid a high percentage of H₂O and CO₂.

    It was assumed that then the conditions for deposition of the catalyst would no longer be favorable.

    Furthermore, an H₂-rich mixture with a possible excess of hydrogen, or even a CO : H₂ = 1 : 2 ratio was to be tried since we knew by experience that hydrogen-rich mixtures are detrimental to the formation of resinous substances.

    Or a small portion of the slurry was to be constantly run over an absorbent in order to remove high-molecular, resinous substances or, in case the suspended catalyst caused difficulties, a hydrogenation with hydrogen was to be affected with such care that the resin-formers would be hydrogenated and not the olefins. 

    Moreover, the catalyst was to be withdrawn from the reactor continuously or in regular intervals of time, purified and again returned to the reactor, or replaced by a new batch.  This could be readily done in the liquid phase without interrupting operation.

    An attempt was made to reemploy a sample of the reactor deposit as a catalyst.  It was again ground with middle oil.  The sample was run in an experimental reactor and it turned out that after two days it approximated the output of a moderately good new catalyst.  The table follows:

    The higher degree of reduction in the laboratory reactor must be attributed to the fact that a more rapid liquid circulation was maintained here than in the large-scale experiment for external reasons, with a consequent decreased gas conversion.  Catalyst deposition was never observed in the small-scale reactor.

The Products of the Foam Process

    The foam process at 240 - 250 degrees is differentiated for the most part from the gas circulation (process) in regard to products by the following features:

    The higher fractions are present in about equal proportions with the gasoline fractions.  The products have a more straight-chained construction.  The oxygen content is somewhat higher.  Gas production decreases more markedly.  The gasoline does not attain the resistance to knocking of the gasolines produced at a higher synthesis temperature; conversely, the Diesel oil is superior for the same reasons.  Gas formation constantly dwindles.

    This tendency persists to the lowest attainable synthesis temperatures.

    The product yield can be varied within certain limits so that in one instance the higher-boiling products can be promoted, and in the other instance the lower-boiling ones.  The following is an average composition of the product ⁽¹⁾.

	% by weight	Olefin content (2)	Oxygen content
Liquid gas (C3, C4)	10	75 - 80 %
Gasoline to 200 degrees	30 - 40	70 %	6 - 8 % 3)
Middle oil 200-325 degrees	25 - 35	50 - 60 %	3 - 5 $ 4)
Paraffin above of 325 degrees	20 - 30	-	-
Alcohols in the product water (mainly ethanol)	5	-	-

    Gasification in the form of methane and ethane runs to 5 - 8 per cent.

    The proportion of paraffin can be increased at the expense of the lower-boiling constituents by the use of markedly alkalized catalyst.

Use of the Products

    The gasoline obtained at a synthesis temperature of 240 - 250 degrees does not have a high octane number even in a refined state.  The octane number (research) is 60 in an unrefined state, and 70 - 75 in a refined state.  For this reason, if gasoline is to be the main product, a higher synthesis temperature should be selected.

    On the other hand, since it contains 70 - 75 per cent terminal olefins, it is suitable for the production of aldehydes and alcohols by means of addition under pressure of CO and H₂ by the Oxo reaction and subsequent hydrogenation at 180 degrees, or for the production of fatty acids by addition of CO and H₂O according to Reppe.

    To this end it is advantageous to use narrow fractions.  The addition products thus fall out of the boiling range and are readily isolated by distillation.  C₆ and C₇ alcohols are thus obtained from the 50 - 100° fraction with a 65 per cent yield; C₈ - C₁₁ alcohols are obtained from the 100 - 150 degrees fraction with a similar yield.  Moreover,  20 per cent higher alcohols of about C₂₀ are obtained.  The residual gasoline together with the unused gasoline fraction gives a gasoline with an octane number (research) of 60.  Since the CO addition takes place partly at the terminal atom and partly at the next C atom, alcohols are partially obtained with a methyl group in the alpha position.

    The alcohols obtained in this way can be advantageously used in the production of lacquer.

Middle Oil

    The middle oil obtained at a low synthesis temperature is more suitable as a Diesel oil than that obtained at a higher temperature.

    Whilst the latter possesses a cetane number of only 40 - 50, the former possesses a cetane number of 75 - 80.  The Engler viscosity at 20 degrees is 1.16.

    The suitability of the middle oil obtained at a low temperature for production of sulfuric acid esters and alcohol sulfonates was the real cause for the development of the synthesis in the liquid phase.  The chains of 13 - 18 carbon atoms give valuable fine detergents.

    The olefins have the advantage that they can be converted directly into sulfuric acid esters with sulfuric acid.  40 per cent of the middle olefins can be directly transformed.  The best product is obtained from the 230 - ? degrees fraction.  The separation of the sulfonates form the unconverted olefins is accomplished by ethanol, etc.  If necessary the residual oil, after hydrogenation, can be converted into alcohol sulfonates by sulfochlorination, or converted in the same way to Igepal NA with benzene.

    The olefin contents and the straightchainedness ⁽¹⁾ of the middle oil fraction were determined as follows:

Fraction	Olefin contents	Straightchainedness
200 - 250°	70 %	69 %
250 - 300°	56 %	70 %
300 - 350°	44 %	74 %

    Another use of middle oil is for the production of lubricating oil.  35 - 40 per cent of a lubricating oil with a Viscosity Index of 85 is obtained by condensation with AlCl₃.  Heating oil amounts to 10 - 15 per cent and a Diesel oil with a cetane number of 60 - 70 amounts to 50 per cent.

Paraffin

    The obtained paraffin has an oxygen content of 2 - 4 per cent.  It can be readily freed of oxygen and simultaneously saturated by hydrogenation.  Its appearance will then be pure white; it is hard, no longer sticky and useful for the most varied purposes.  For example, it is suitable for polishing wax and shoe polish.  Where lower quality is satisfactory it can be used without hydrogenation.

    Two-thirds of the paraffin boils above 450 degrees and forms high-melting paraffin.  The first third can be used for paraffin oxidation directly after hydrogenation whereas the high-melting paraffin must first be subjected to careful thermal cracking to obtain a chain length suitable for paraffin oxidation.

    Furthermore, cracking turns 70 per cent of the total paraffin into middle oil containing 70 per cent olefins.  This is also a means of obtaining olefin-rich middle oils.  The following was obtained by careful cracking of the paraffin at 400 - 420 degrees:

Gasoline to 200°	15 %	Olefin content
Middle oil 200 - 250°	15 %	69
Middle oil 250 - 350°	55 %	51
Residue above 350°	8 %
Coke	5 %
Gas	1 - 2 %

It is used in the same way as the original middle oil.

Sintered Catalysts

for Low Synthesis Temperatures

    The sintered catalysts that have been further described in the preceding pages are sintered in a reduced, i.e., metallic condition.  Their activity is thus neutralized to a certain extent.

    Another category of sintered catalysts is obtained if the sintering is conducted before the reduction and the reduction carried out at the lowest possible temperature with a great deal of hydrogen.  Special effects will be obtained if substances such as tungsten oxide, barium oxide, etc. are introduced ⁽¹⁾.  Catalysts of this kind possess a remarkable thermal stability and a high activity which permits of a lowering of the synthesis temperature to 230 degrees and beyond, and thus makes possible the production of almost exclusively straight-chained paraffins.  They can be used in the conventional Fischer reactor.

    A special characteristic of these catalysts is their preference for the formation of paraffin.  The following was produced in one instance:

Liquefied gas (C3, C4)	4 %
Gasoline up to 200°	17 %
Middle oil to 350°	19 %
Paraffin	60 %

    The catalyst was prepared in the following manner:

    Iron powder, obtained by thermal decomposition of iron carbonyl, wax mixed with several per cent HO₃, pasted with a little water containing 1 per cent potassium borate (basis iron) and kneaded.  This was pressed into pellets and ignited at 850 degrees with air for 4 hours.  They were then reduced at 450 degrees with a great excess of hydrogen.

    It is often more advantageous to use precipitated iron as the starting material or to undertake the precipitation with the other substance.  Among others, the following substances are suitable as the other substance: barium, magnesium, chromium and molybdenum.

Dependence of the Composition and Quality of the Products 

on the Reaction Conditions

    In general the following regularities depending on the external conditions can be set up:

    Higher pressure steps up the oxygen contents of the products.  If the pressure is raised from 20 to 100 atmospheres, for example, it will increase almost threefold.

    Higher synthesis temperature gives better gasolines, inferior Diesel oils, branching of the products is greater, and the gaseous products and the gasoline become more important at the expense of the higher boiling products.

    Lower synthesis temperature gives less valuable gasolines, but better Diesel oils.  The straightchainedness is higher.  Approximately the same quantities of gasoline, middle oil and paraffin are obtained.  If special catalysts are introduced paraffin will become the main product.  The formation of gaseous products progressively dwindles.  The oxygen content increases.  The products are more suitable for further chemical working up.

    Gas mixtures richer in carbon monoxide yield a somewhat higher degree of unsaturation in the products.  The use of hydrogen-rich gases has a favorable effect on the catalyst life.  A higher alkali content promotes the formation of higher-boiling products.

                                                                                                                                    Michael

Listing of Several Reports and Memoranda

14456	Discussion of the State of the CO - H2 Synthesis on April 7, 1941 at Lu. 558.
1367	Discussion of the CO - H2 Synthesis held in Berlin on July 1, 1941.
14438	Report on the Session of July 16, 1942.  State of the Foam Process.
06808	Michael                                                                          May 23, 1942
		Hydrocarbon Synthesis from CO and H2 according to the Foam Synthesis with Iron Catalyst.  Utilization and Work-Up of the Products.
06797	Report on the Session of July 16, 1942.  Hydrocarbon Synthesis.
		State of the Foam Process.
06305	Michael                                                                                                         July 3, 1942
		Performance and Yields of the Foam Process on the Basis of Experiments Conducted with a 1.5 cbm Reactor.
14491	Michael                                                                                                         May 7, 1940
		State of the Synthesis Oil Experiments in the Production of Detergent Starting Oils.
14474	Carbon Monoxide Hydrogenation Olefins as Starting Materials for Detergent Raw Materials and Textile Dying Aids.  June 4, 1940
14504	Michael                                                                                                          May 7, 1940
		Middle Oil Process
?6770	Michael                                                                                                          March 12, 1941
		Products of the Foam Process
?6765	Fürhäuser - Haussmnn                                                                                    March 28, 1941
		Review of the Synthesis Oils Obtained from High Pressure Experiment Section (Michael) and of the Textile Dying Aids (Sulfuric Acid Esters or Sulfonic Acids) Produced Therefrom.
?6762	Investigation of Product Samples.                                                                    July 2, 1941
?6827	Michael                                                                                                          September 13, 1941
		Alcohol Synthesis from CO - H2 Mixtures by means of the Gas Circulation Process.
?6815	Bueren                                                                                                            February 3, 1942
		The Working-Up of the Michael Product from Alcohol
06815	Michael                                                                                                           June 19, 1942
		Work-Up of the Synthesis Products to Alcohols, Fatty Acids and Sulfonates.
06747	Haussmann                                                                                                      March 21, 1942
		Foam Process = Olefins
03745	Bueren - Christmann                                                                                        June 1, 1942
		Refining of Michael Paraffin and its Work-Up to Cylinder Oil or Pour Point Depresser.
06743	Refining of Michael-Middle Oil and Gasoline by Extraction with 70 per cent Sulfuric Acid.         June 5, 1942
06783	Michael - Ehrmann                                                                                           February 18, 1943
		The Troubles Occurring in the Foam Process for Hydrocarbon Synthesis and their Elimination.
14517	Memorandum on the Discussion in Ludwigshafen on December 21, 1939 on the I. G. Treibstoffwerk, Upper Silesia (Gas Circulation Process)
14509	Production of Safety Motor Fuel at the I. G. Treibstoffwerk, Upper Silesia (on the basis of the Gas Circulation Process).
14486	Compilation of the Estimates of Construction Costs for the Production of 100,000 tons/year working with Middle Oil on the Basis of the Synthesis Process (Michael).                                          May 7, 1940