Fischer-Tropsch synthesis

The growing reliance on imported oil gave the synthetic fuels a fresh impetus in the 1980s. The Fischer-Tropsch synthesis (FTS) is the exothermic reaction of carbon monoxide and hydrogen to mainly hydrocarbons together with water and carbon dioxide. It can be represented by the following reaction equation:

Sabatier was the first to react carbon monoxide and hydrogen over a nickel catalyst in 1902. The result was the production of methane and water. Then, in 1923 Franz Fischer and Hans Tropsch of the Kaiser Wilhelm Institute (Germany) developed the Fischer-Tropsch process (i.e. Synthol process), in which a carbon monoxide and hydrogen flow, in the presence of iron, cobalt or nickel catalyst at 180-250℃ and at pressures from atmospheric to 150 bar, produced a mixture of straight hydrocarbons and smaller amounts of oxygenates. The initial objective of this process was gasoline production. Experience at high and medium reactor pressures was disappointing from  an unacceptably high oxygenate product content point of view. Hence, it was adopted an atmospheric reactor pressure.

Cobalt became strongly favored as catalytic element, since iron was less active and deactivated rapidly at atmospheric pressure synthesis operation, while nickel gave high methane selectivity and was affected by its loss due to volatile nickel tetracarbonyls production.

The best catalyst was found to be based on cobalt, supported on Kieselguhr with thoria and magnesium oxide as promoters (100g Co/5g ThO/8g MgO/200g Kieselguhr).

In recent years, we assist to a second renewal of interest in the F-T process for producing liquid hydrocarbons. This new interest centers on making synthetic fuels from natural gas instead of coal.

The process of converting natural gas or coal into marketable liquid hydrocarbons comprises three main elements: 1)synthesis gas production, 2) hydrocarbon synthesis via the F-T conversion process, 3) products work-up (Fig. 1 .4)

In the first step, natural gas is converted to synthesis gas (syngas), a mixture of hydrogen and carbon monoxide, through the commercially known methods (steam-reforming, partial oxidation, or autothermal reforming). The syngas in the second step is converted to hydrocarbons via the F-T synthesis process. In the third step, the primary hydrocarbons in the form of syncrude are worked up to final products consisting mainly of naphtha, diesel fuel and kerosene (middle distillates).

As far as syngas production is concerned, other feedstocks such as coal, heavy residue or shale oil can be used, but the process becomes less economical. Sasol is the largest producer of synfuels and chemicals made by coal gasification (Lurgi's technology is employed). Besides coal plants, since 1993 Sasol has also operated natural gas-based plants at Mossgas, South Africa, with a capacity of 44.000 BPD of fuels. The company is by far the most experienced player in the syngas-based chemical business, as far as reactors design, catalytic formulations, process technology are concerned. Sasol has been opting for iron-based catalyst since 1955. Only in recent years the advantages of cobalt based catalyst for use in slurry phase reactor have been recognized.

For the preparation of Synthol catalyst, Sasol uses iron oxides: the suitable iron oxide is fused together with the required chemical and structural promoters. The fused ingots are milled to a specified particle size range (for optimum fluidization properties). The catalyst is pre-reduced with H₂ at about 400℃ and then loaded in the F-T reactors. Because of the simplicity of the preparation and the low cost of the row materials, the cost of Synthol catalyst is a minor part of the overall process.

The catalyst used in the fixed bed process is more expensive as it is prepared by wet chemical means. Iron oxide is precipitated by adding an alkaline solution to an  iron  nitrate solution. The precipitate is washed, reslurried, silica based support and promoters added and the gel is then extruded. The surface area and pore size distribution of this catalyst is largely determined by the conditions used in the initial iron oxide precipitation step, these effects, such as precipitation time, temperature, order of addition etc.，have been quantified. Because of the high surface area and the presence of reduction promoters (i.e. Cu), the catalyst is pre-reduced under mild condition and then coated with wax to prevent re-oxidation.

The advantages of iron are the following:

    1. it is the cheapest metal to use,

    2. it has a higher selectivity to olefins,

    3. it is an active water-gas shift catalyst.

The latter is important since feed gases with H₂/CO ratio below 2 can be utilized directly. A disadvantage of iron is that it deactivates faster than cobalt due to oxidation and coke deposition.

Studies performed at Sasol on the commercial spray dried precipitated iron catalyst and Co/Al₂O₃ slurry phase FTS catalyst, resulted in the following conclusions:

1. the cobalt does not show any significant water-gas-shift activity and no water inhibition of the F-T reaction rate,

2. the cobalt catalyst is the preferred option if high per pass conversions are required,

3. desired stabilized intrinsic activity levels can be achieved with cobalt catalyst, implying that extended slurry phase synthesis runs can be realized,

4. cobalt derived hydrocarbon product selectivities show greater sensitivity towards process conditions (i.e. reactor pressure) than that of iron. Iron catalyst, on the other hand, shows marked sensitivity towards chemical promotion. Indeed, the geometric tailoring of pre-shaped support materials can be an optimization tool for effecting  increased wax selectivities with cobalt based catalyst, an approach also suggested by others (Exxon, Shell).

The has conversion of synthesis gas to hydrocarbons (Fischer-Tropsch synthesis) been widely studied and extensively described. A number of synthesis reactions can occur and the whole are quite exothermic,△H=-170 kJ (C atom^-1):

The free energy changes (△G⁰) in the above reactions are such that the hydrocarbon  synthesis is normally favored below about 400℃. Over the temperature range of 200-400℃, the formation of methane is favored. However, since the thermodynamic equilibrium is reached slowly in FT synthesis, it is possible to take advantage of kinetic factors by using suitable catalysts, so that heavier hydrocarbons or alcohols are produced in suitable quantity. The production of hydrocarbons using traditional FT catalysts, such as Fe or Co, is governed by chain growth or polymerization kinetics. The so-called "surface carbide" mechanism is a plausible one for the interaction of CO and H₂ with the catalytic surface and the subsequent synthesis of hydrocarbons. Ample evidence shows that this is the prevalent mode of activation of CO at elevated temperatures on the Group VIII metal catalysts Fe, Co, Ni and Ru. The model can be used as a starting point for understanding the formation of various molecular species during FT synthesis and also for examining hydrocarbon chain growth.

The process is only provided to indicate how chain growth can occur. Several models of chain growth have been discussed in the literature: repeated carbon insertion as indicated above results in chain growth. As a certain stage of chain growth, the hydrocarbon species is desorbed. If an unsatured surface species is desorbed, one obtains an olefinic product. Desorption after hydrogenation results in paraffin species. In some instances, CO insertion into the metal-alkyl bond and subsequent desorption results in an oxygenated species, such as an alcohol. The nature of the product and the product distribution among the carbon numbers will depend upon the catalyst surface, composition (H₂/CO ratio) and the rate of flow feed gas, reaction pressure and the temperature at which FT synthesis is performed. The above parameters will affect the rate of hydrogen and CO dissociation, hydrogenation, degree of polymerization and  desorption of the product species.

At low temperatures, the main primary products are linear 1-alkenes, alkanes, alcohols and aldehydes. The linearity of the product is important for many of their applications. It gives the waxes with high melting point and low viscosity.

The C₉ to C₁₅ olefins are ideal for the manufacture of biodegradable detergents. The C₁₀ to C₁₈ cut is an excellent diesel fuel (with the high cetane number of 75 and zero aromatics). On the other hand, the product linearity is a disadvantage for gasoline production, since a high octane number requires branched alkane and aromatics. Hence, the gasoline requires extensive isomerization  and aromatization.

At higher synthesis temperatures secondary reactions occur, i.e branched hydrocarbons and aromatics are formed. In that way, the diesel cetane number octane  number  increases. Olefins in the at C₃ or C₄ (up to 90%)，to then decrease continuously, the waxes being essentially paraffinic.

FT synthesis catalyst

Most Group VIII metals have significant CO hydrogenation activity, being the product distribution the distinguishing feature. The specific activity of various Group VIII metals was determined: Ru>Fe>Ni>Co>Rh>Pd>Pt, while the average hydrocarbon molecular weight decreased in the order: Ru>Fe>Co>Rh>Ni>Ir>Pt>Pd.

The catalysts that were active for F-T synthesis were compared in order to identify common properties. It was found that:

-they are active for hydrogenation reactions

-they are capable for metal carbonyl formation

-the F-T reaction conditions (T, P) are not far from those where thermodynamics would allow the metals to be converted into metal carbonyl.

From the latter observation, it was suggested that `surface carbonyls' play an essential mechanistic role in the formation of hydrocarbons.

Iron and cobalt are the only metals used until now for industrial application. Nickel tends to form volatile nickel carbonyl at elevated pressure and is most selective for methane relative to Co, Fe and Ru catalysts. Ruthenium is the most active F-T catalyst, working at the lowest reaction temperature of about 150℃, and it is selective towards high molecular weight products. However, it is most active in the pure form, i. e. supports and/or promoters appear to have no beneficial effect. Even under conditions favorable for high wax yields, Ru tends to have high methane selectivity. Ruthenium has a high potential as catalyst for converting synthesis gas to a variety of hydrocarbons, but its high price and limited world resources exclude possible industrial application.

Other Group VIII metals, namely Rh, Re, Os, Pd, Pt and Ir, yield mostly oxygenated compounds partly because CO does not chemisorb dissociatively on these metals. Mo is not a Group VIII metal, but it exhibits moderate F-T activity and its nitride and carbide show excellent alkene synthesis rate. It is attractive because of its sulphur-resistant characteristics.

Iron catalysts are commonly used because of their low costs. They have a high selectivity to olefins and a high water-gas shift activity. The latter aspect is important since it means that CO₂ can also be hydrogenated to hydrocarbons and that feed gases with  H₂/CO ratios below 2 can be utilized directly. A disadvantage of Fe is that it deactivates faster due to oxidation and coke deposition. In fact, metallic iron is extremely active for the F-T reaction but it is not thermodynamically stable under F-T conditions and is rapidly converted to more stable, but less active, iron carbides. With time, the carbides are also slowly oxidized by water vapor to inert magnetite. Another factor that contributes to deactivation is poisoning by S-containing compounds such as HZS. Poisoning by chemisorbed S is permanent. Chlorine, although to a lesser extent, also deactivates iron catalysts.

Cobalt catalysts do not oxidize or carburize under normal F-T conditions and because of this should deactivate less rapidly than iron catalysts. Co catalysts can give high  yields of liquid hydrocarbons and waxes but under similar conditions the methane yields are usually higher and the olefin yields are lower than for iron catalysts, i.e. they are more hydrogenating. The yields of oxygenated compounds are lower for cobalt catalysts and this can be an advantage if they are not desired.

The relatively high cost of Co makes of fundamental importance its distribution on suitable supports (i.e. Al₂O₃, TiO₂, SiO₂). The activity of supported Co catalysts increases linearly with loading in the range 5 to 30 mass%Co and so a compromise between cost and activity has to be made. Under F-T conditions, Co is a poor water gas shift catalyst and so the feed gas needs to be adjusted to about 2, as required by the stoichiometry of the F-T reaction. Poisoning by S-containing compounds causes the permanent deactivation of cobalt-based catalysts. Cobalt is the most widely Group VIII used in natural gas-to-liquids research project.

F-T catalysts are often prepared by precipitation, impregnation, ion exchange, and synthesis from organometallic compounds and vapor phase deposition in which the metal precursor is loaded onto the support surface. This is then followed by drying, calcination and catalyst activation (via reduction of the metal precursor to generate the metallic phase). The interplay of catalyst composition and preparation conditions determines the activity and selectivity behavior for a given set of process parameters. The selection of the `best' catalyst is the most crucial step in F-T technology.

Thermodynamics and reaction mechanism

The F-T synthesis is a highly exothermic reaction:

Thermodynamically, the formation of methane and other hydrocarbons is energetically favorable, that means negative Gibbs energy. For example,△G⁰ values at 227℃ for the formation of ethane and propane are -29.2 and -37.3 kcal/mol respectively, while at 200℃△H⁰ values for the formation of methane, ethane and propane are about -50，-42, and -29 kcal/mol, so confirming the exothermic nature of the process.

In  Fig. 1.5 standard Gibbs energy variation for the production of hydrocarbons and alcohols is reported as a function of temperature. From the diagram reported in figure 4, it can be deduced that methane formation is highly favored over that of the alcohols, olefins and hydrocarbons of heavier molecular weight. Moreover, hydrocarbons are favored over olefins and alcohols. Finally, as far as olefins are concerned, ethylene is less favored at temperature below 430℃.

Fischer-Tropsch synthesis has long been recognized as a non-trivial polymerization reaction with the following steps as the key sequence:

1. reactants adsorption

2. chain initiation

3. chain growth

4. chain termination

5. products desorption

6. re-adsorption and further reaction

Since its discovery, many efforts have been made to identify the surface species that lead to chain initiation and chain growth. Fischer and Tropsch proposed the earliest ideas in 1926. They hypothesized CO reaction with the metal of the catalyst to form bulk carbide (CO dissociative chemisorption), which subsequently undergoes hydrogenation to form methylene groups (see Scheme1).These species were assumed to polymerize to form hydrocarbon chains that then desorbs from the surface as saturated and unsaturated hydrocarbons (see Scheme 4).

Molecularly adsorbed CO (associative adsorption) was hydrogenated to form a hydroxycarbene or enol, M=CH(OH), which then underwent further hydrogenation to produce a methylene group.  The growth of hydrocarbons occurs by polymerization of these latter species. The formation of  enol intermediates was suggested that chain growth takes place by the condensation of enol groups with the concurrent elimination of water (see Scheme 2).

Another mechanism for chain growth was proposed: molecular CO was postulated to insert into the metal-carbon bond of an adsorbed alkyl species. Hydrogenation of the resulting acyl group was assumed to produce water and a new alkyl group containing an additional methylene unit. The chain is initiated by methyl groups formed by stepwise hydrogenation of molecularly adsorbed CO (see Scheme 3).

In Scheme 4 the three different proposals are resumed and a global view of chain growth and chain termination to various types of products is given.

The mechanism proposed by Fischer (`carbide' theory) did not explain the production of relatively large amounts of oxygenated products, i.e. alcohols. Mechanisms described by the Schemes 2 and 3, explained the formation of oxygenated products, nevertheless recent experimental evidence indicated the mechanism of Scheme 1 is possible.

In drawing Scheme 4, it was suggested that non-oxygenated, rather than oxygenated, intermediates play a dominant role in the synthesis of hydrocarbons from CO and H₂. Moreover, results obtained from the fields of surface science, organometallic chemistry, and catalysis  strongly support that hydrocarbon synthesis is initiated by the dissociation of CO.

F-T synthesis plants

Sasol  is  the  largest  producer  of  synfuels  and  chemicals  made  by  coal gasification. Besides coal plants, since 1993 Sasol has also operated natural gas-based plants at Mossgas, South Africa. The company is by far the most experienced player in the syngas-based chemical business, as far as reactors design, catalytic formulations, process technology are concerned.

Since 1955, Sasol has operated with the Fischer-Tropsch synthesis at either low temperatures (LTFT) to produce a syncrude with a large fraction of heavy, waxy hydrocarbons, or at high temperatures (HTFT) to produce a light syncrude and olefins. Until the nineties, two types of F-T reactors have been used, the tubular fixed bed (TFB) reactor for LTFT and the circulating fluidized bed (CFB) reactor for HTFT.

The TFB, also known as Arge reactor, has been in use at its Sasolburg plants in South Africa for the production of wax since 1955. This type of reactor consists of a shell containing 2050 tubes, 12 m long, 5 cm in internal diameter and operates at a shell side temperature range of 220-2500C and a reactor pressure of 25 bar for the earlier reactors and 45 bar for a reactor commissioned in  1987. It produces heavy waxy hydrocarbons that are subsequently hydrocracked and/or isomerized for diesel production (Fig. 1.6-A).

The TFB reactor makes use of extruded on silica precipitated iron-based catalyst, promoted with a small amount of potassium and packed into the tubes. Syngas is passed downward through the catalyst bed and  is converted to hydrocarbons. The exothermic heat of the F-T reaction is removed through the tube wall to produce steam on the shell side of the reactor. That gives rise to axial and radial temperature profiles. For maximum reaction rates, a maximum average temperature is required.  This is limited  however by the maximum allowable temperature peak, which cannot be exceeded in order to prevent catalyst deactivation and carbon formation on the catalyst. Carbon formation causes break up of the catalyst with a consequent loss in conversion efficiency and necessity to catalyst replacement.

Product selectivities are temperature dependent and flexibility with respect to temperature control would be advantageous. The choice of temperature level is however limited by the need to not exceed the maximum peak temperature. Pressures drops across TFB reactor are high, from 3 to 7 bar; that gives rise to considerable compression cost.

An advantage of the TFB reactor is that there are no problems in separating the wax product from the catalyst, on the other hand, the reactor is complex and has a high capital cost. At the prevailing reactor conditions, scale up is mechanically difficult,  the  tube  sheets  become  very  heavy.  In  addition  to  this,  special removable grids are required for the removal and replacement of the catalyst, making the reactor design more complex and costly. The catalyst replacement itself is cumbersome and maintenance and labor intensive, and it causes down time and disturbances in plant operations.