A scale up strategy for bubble column slurry reactors

The hydrodynamics of bubble column slurry reactors are strongly influenced by the scale of operation. We suggest a strategy for scaling up reactors from laboratory scale to commercial size that relies on a fundamental understanding of bubble hydrodynamics, which is incorporated into a computational fluid dynamics (CFD) model.     

Development of Fe Fischer–Tropsch catalysts for slurry bubble column reactors

The effect of two binder systems — a silica-based system and a silica–kaolin–clay–phosphate-based system — on a doubly promoted Fischer–Tropsch (FT) synthesis iron catalyst (100Fe/5Cu/4.2K) was studied. The catalysts were prepared by coprecipitation, followed by binder addition and spray drying at 270°C in a 1 m diameter, 2 m tall spray dryer. The binder silica content was varied from 0 to 20 wt.%. A catalyst with 12 wt.% binder silica was found to have the highest attrition resistance. The FT activity and selectivity of this catalyst are better than a Ruhrchemie catalyst at 270°C and 1.48 MPa. The addition of precipitated silica or kaolin to catalysts containing 10–12 wt.% binder silica decreases attrition resistance and increases methane selectivity. Based on the experience gained, a catalyst has been successfully spray dried in 500 g quantity. This catalyst showed 95% CO conversion over 125 h of testing at 270°C, 1.48 MPa, and 2 NL/g-cat/h and had less than 4% methane selectivity. Its attrition resistance was one of the highest among the catalysts tested.     

Influence of syngas composition on the transient behavior of a Fischer–Tropsch continuous slurry reactor

The influence of syngas composition on the initial behaviour of a Co/Al₂O₃ catalyst in Fischer–Tropsch reaction has been studied in a continuous perfectly mixed slurry reactor for an inlet H₂/CO ratio between 1.6 and 3.35 keeping other conditions constant (T = 220 °C, P = 2 MPa). Significantly different behaviors of initial deactivation for CO conversion have been observed with different H₂/CO ratios. It was observed that the deactivation increases with increase in H₂/CO ratio and in carbon monoxide conversion. The computed liquid concentrations of CO, H₂ and H₂O have shown that water is the most abundant species in the liquid phase of the reactor during our experiments. The concentration of the water produced by the FT reaction seems to be the key parameter responsible of the initial behavior and then of the initial deactivation. For moderate levels of water ( corresponding to P_H2O/P_H2<0.4), a simple kinetic model assuming a reversible oxidation of cobalt active sites by water in competition with their reduction by hydrogen seems to represent satisfactorily the initial behaviour of the catalyst. For higher water concentrations, the irreversible deactivation should be probably taken into consideration.     

Performance of a catalytic membrane reactor for the Fischer–Tropsch synthesis

The study of permeable composite monolith (PCM) membranes for the Fischer–Tropsch synthesis is continued. On the scale of membranes with outer diameter of 42 mm, it is proved that PCM can combine high productivity of hydrocarbons (>55 kg_C5+ ( h)⁻¹ at 0.6 MPa, 484 K), high selectivity towards heavy hydrocarbons (_ASF > 0.85, C₅₊ upto 0.9) as well as high heat-conductivity and high mechanical strength.     

Attrition studies with catalysts and supports for slurry phase Fischer–Tropsch synthesis

Attrition properties of several oxide supports and precipitated iron-based F–T catalyst (100Fe/3Cu/4K/16SiO₂ in parts by weight) were evaluated using ultrasound irradiation test and stirred tank slurry reactor (STSR) test under non-reactive conditions. Attrition by fracture and erosion of the iron-based catalyst was small in both types of tests and its overall attrition strength was better than that of the alumina and silica supports, which were evaluated under the same conditions. Also, attrition studies with four iron-based F–T catalysts were conducted under reaction conditions in the STSR. Catalyst of similar composition, as that used in non-reactive studies, prepared by spray drying technique had the highest attrition strength among all catalysts tested.     

Alumina-supported cobalt-molybdenum catalyst for slurry phase Fischer–Tropsch synthesis

An evaluative investigation of the Fischer–Tropsch performance of two catalysts (20%Co/Al₂O₃ and 10%Co:10%Mo/Al₂O₃) has been carried out in a slurry reactor at 2 MPa and 220–260 °C. The addition of Mo to the Co-catalyst significantly increased the acid-site strength suggesting strong electron withdrawing character in the Co-Mo catalyst. Analysis of steady-state rate data however, indicates that the FT reaction proceeds via a similar mechanism on both catalysts (carbide mechanism with hydrogenation of surface precursors as the rate-determining step). Although chain growth, , on both catalysts were comparable (  0.6), stronger CH₂ adsorption on the Co-Mo catalyst and lower surface concentration of hydrogen adatoms as a result of increased acid-site strength was responsible for the lower individual hydrocarbons production rate compared to the Co catalyst. The activation energy, E, for Co (96.6 kJ mol⁻¹), is also smaller than the estimate for the Co-Mo catalyst (112 kJ mol⁻¹). Transient hydrocarbon rate profiles on each catalyst are indicative of first-order processes, however the associated surface time constants are higher for alkanes than alkenes on individual catalysts. Even so, for each homologous class, surface time constants for paraffins are greater for Co-Mo than Co, indicative that the adsorption of CH₂ species on the Co-Mo surface is stronger than on the monometallic Co catalyst.     

A highly active and stable Fe-Mn catalyst for slurry Fischer–Tropsch synthesis

A highly stable and active Fe-Mn catalyst for slurry Fischer–Tropsch synthesis (FTS) was prepared and scaled up for the application in the industrial pilot plant at Institute of Coal Chemistry (ICC), Chinese Academy of Sciences (CAS). One Lab-scale catalyst and one scaled-up catalyst are introduced in the present paper. The particle size of the Lab-scale catalyst is about 5–15 μm, while it is increased to 30–90 μm for the scaled-up catalyst. Simultaneously, the morphology of the catalyst was greatly improved after the catalyst being scaled up. Both the Lab-scale and scale-up catalysts show high FTS activity. CO conversion of the Lab-scale catalyst and the scaled-up one are over 70.0% (H₂/CO = 0.67, 275 °C, 1.5 MPa and 3000 h⁻¹) and 55.0% (H₂/CO = 0.67, 260 °C, 1.5 MPa and 2000 h⁻¹), respectively. The catalysts also possess excellent stability, no obvious deactivation was observed during stable run of 4200 h and 1200 h on stream for the two catalysts. However, the Lab-scale catalyst produced more methane (about 8–10 wt%) and C_2–4 (22–30 wt%) and less C5⁺ hydrocarbon (55–70 wt%). Meanwhile, the hydrocarbon distribution of the catalyst was greatly improved for after the catalyst being scaled up, and the distribution of hydrocarbon products become much preferable. The selectivity to methane was well controlled at about 5 wt%, and the sum of and was increased to 91–93 wt%. On the whole, the scaled-up catalyst satisfies the requirements of the application for FTS in the industrial pilot plant of slurry bubble column reactor (SBCR) at ICC, CAS.     

Construction of the Fischer–Tropsch regime with cobalt catalysts

Changes of activity and selectivity during the initial phases of Fischer–Tropsch (FT) synthesis have been measured with three promoted cobalt catalysts. It is shown that the FT regime is formed in situ in a slow process lasting several days. A “construction” of the “true FT catalyst” is therefore assumed. Taking into account complementary investigations, this construction is assigned to the segregation of the catalyst surface caused by strong CO chemisorption. This process would be accompanied by an increase of the number of active sites and their disproportionation into sites of higher and lower coordinations, which would exhibit different catalytic properties. The observed initial activity and selectivity changes are well to be explained with this concept.     

Overview of reactors for liquid phase Fischer–Tropsch synthesis

The following overview is divided roughly into three sections. The first section covers the period from the late 1920s when the first liquid phase synthesis was first conducted until about 1960 when the interest in Fischer–Tropsch synthesis (FTS) declined because of the renewed view of an abundance of petroleum at a low price. The second period includes the activity that resulted from the oil shortage due to the Arab embargo in 1972 and covers from about 1960 to 1985 when the period of gloomy projections for rapidly increasing prices for crude had faded away. The third section covers the period from when the interest in FTS was no longer driven by the projected supply and/or price of petroleum but by the desire to monetize stranded natural gas and/or terminate flaring the gas associated with petroleum production and other environmental concerns (1985 to date). These sections are followed by a brief overview of the current status of the scientific and engineering understanding of slurry bubble column reactors.     

Kinetics of Fischer–Tropsch synthesis on titania-supported cobalt

Rate data have been obtained for CO hydrogenation on a well-characterized 11.7% Co/TiO₂ catalyst in a differential fixed bed reactor at 20 atm, 180–240°C, and 5% conversion over a range of reactant partial pressures. The resulting kinetic parameters can be used to model precisely and accurately the kinetics of this reaction within this range of conditions. Turnover frequencies and rate constants determined from this study are in very good to excellent agreement with those obtained in previous studies of other cobalt catalysts, when the data are normalized to the same conditions of temperature and partial pressures of the reactants. Based on this comparison CO conversion and the partial pressure of product water apparently have little effect on specific rate per catalytic site. The data of this study are fitted fairly well by a simple power law expression of the form −r_CO=kP_H2^0.74P_CO^−0.24, where k=5.1×10⁻³ s⁻¹ at 200°C, P=10 atm, and H₂/CO=2/1; however, they are best fitted by a simple Langmuir–Hinshelwood (LH) rate form −r_CO=aP_H2^0.74P_CO/(1+bP_CO)² similar to that proposed by Yates and Satterfield.     

C-tracer study of the Fischer–Tropsch synthesis: another interpretation

The ¹³C-tracer results from the introduction of ¹³C₂H₄ into syngas prior to conversion with a rhodium catalyst have been used to support a surface vinyl mechanism for Fischer–Tropsch synthesis. The results were first interpreted by a mechanism that involved a decrease in ¹³C species on the surface as the carbon number increased. This model is shown to be incorrect. Considering only the ¹³C-labeled products, the data are consistent with earlier tracer studies showing that the added ¹³C₂H₄ initiates chains.     

Supercritical phase process for direct synthesis of middle iso-paraffins from modified Fischer–Tropsch reaction

A new concept on the direct synthesis of middle iso-paraffins through the modified supercritical Fischer–Tropsch (FT) reaction was proposed and experimentally demonstrated with the combination of Co/SiO₂ and palladium-supported β zeolite (Pd/β) catalysts both in one-stage and two-stage fixed-bed reaction systems. Depending on the reaction conditions, the selectivity of C₄⁺ iso-paraffins mostly in gasoline range was 60–80% due to hydrocracking and hydro-isomerization function of Pd/β. Irrespective of reaction conditions, the hydrocarbons produced from FT reaction were preferentially hydro-converted over Pd/β catalyst while the supercritical solvent of n-hexane could only be slightly hydro-converted under severe conditions. The production of iso-paraffin in two-stage process was tested for 100 h without observable deactivation by feeding additional hydrogen in lower reactor, which is ascribed to the inhibition of coke deposition over Pd/β as revealed from TG analysis of the used catalysts.     

Transient studies of the elementary steps of Fischer–Tropsch synthesis

The pulse transient method has been used to study the kinetics of several key steps of Fischer–Tropsch (FT) synthesis over cobalt supported catalysts. These elementary steps involve chemisorption of hydrogen and propene, and chemisorption and hydrogenation of carbon monoxide. It is found that at the conditions of Fischer–Tropsch synthesis, hydrogen chemisorption is reversible and quasi-equilibrated, while carbon monoxide adsorption is generally irreversible. Chemisorption of propene on cobalt metal sites results in its rapid autohydrogenation to propane and simultaneous formation of C_xH_y surface species.

The transient response curves produced during hydrogenation of carbon monoxide pulses in a flow of hydrogen have been analyzed using the modified Kobayashi model, which involves irreversible chemisorption and dissociation of carbon monoxide, quasi-equilibrated adsorption of hydrogen and reversible adsorption of water. The kinetic analysis suggests that oxygen-containing species are probably the most abundant surface intermediates. Desorption of water from the catalysts seems to be much slower than hydrogenation of surface carbon species.     

Some evidence refuting the alkenyl mechanism for chain growth in iron-based Fischer–Tropsch synthesis

Recently, Maitlis et al. [J. Catal. 167 (1997) 172] proposed an alternative reaction pathway for chain growth in the Fischer–Tropsch synthesis. In this mechanism, chain growth is assumed to occur by methylene insertion into a metal–vinyl bond, forming an allyl species that will subsequently isomerise to a vinyl species. Organo-metallic allyl complexes, Fe{[η⁵-C₅H₅](CO)₂CH₂CH=CH₂} and Fe{[η⁵-C₅(CH₃)₅](CO)₂CH₂CH=CH₂} were synthesised. Under thermal treatment, the decomposition of these complexes was observed, instead of the isomerisation. In a hydrogen atmosphere, the reduction of the iron–carbon bonds and the hydrogenation yielding iron–alkyl species was observed. This clearly shows that the proposed vinyl–allyl isomerisation is unlikely to occur in mono-nuclear iron complexes. Hence, it might be expected that the reaction mechanism proposed by Maitlis et al. [J. Catal. 167 (1997) 172] is unlikely to be the main route for chain growth in the Fischer–Tropsch synthesis.     

Attrition resistance of cobalt F–T catalysts for slurry bubble column reactor use

There exists much current interest in the use of supported Co catalysts and slurry bubble column reactors (SBCR) for the conversion of natural gas to higher hydrocarbons via the Fischer–Tropsch (F–T) synthesis. Catalyst attrition resistance is extremely important in the operation of slurry-phase reactor systems because of potential problems with plugging of system filters and/or contamination of the liquid products. This paper addresses the effects of different supports, promoters, and preparation methods on the attrition resistance of Co F–T catalysts for SBCR use.

The calcined supports had attrition resistances (inversely related to % fines <11 μm generated during attrition testing) as follows:

−Al₂O₃>TiO₂(rutile)SiO₂

Loading of Co onto the supports improved the attrition resistances of both alumina and silica significantly. It has essentially no effect on titania. The resulting catalysts had attrition resistances in the order

Co/Al₂O₃>Co/SiO₂>Co/TiO₂(rutile)>Co/TiO₂(anatase)

The addition of small amounts of metal (Ru, Cu) and oxide (La, Zr, K, Cr) promoters had mainly small effects on the attrition resistance of the supported Co catalysts. However, it would appear that the addition of Zr to Co/alumina had a negative impact on its attrition resistance. The different preparation methods used in this study (aqueous impregnation, non-aqueous impregnation, and kneading) did not appear to have a significant effect on catalyst attrition resistance.     

Fischer–Tropsch synthesis on monolithic catalysts of different materials

Monolithic structures made of cordierite, γ-Al₂O₃ and steel have been prepared as catalysts and tested for Fischer–Tropsch activity. The monoliths made of cordierite and steel were washcoated with a 20 wt.% Co–1 wt.% Re/γ-Al₂O₃ Fischer–Tropsch catalyst whereas the γ-Al₂O₃ monoliths were made by direct impregnation with an aqueous solution of the Co and Re salts resulting in a loading of 12 wt.% Co and 0.5 wt.% Re. The activity and selectivity of the different monoliths were compared with the corresponding powder catalysts.

Higher washcoat loadings resulted in decreased C₅₊ selectivity and olefin/paraffin ratios due to increased transport limitations. The impregnated γ-Al₂O₃ monoliths also showed similar C₅₊ selectivities as powder catalysts of small particle size (38–53 μm). Lower activities were observed with the steel monoliths probably due to experimental problems.     

Fischer–Tropsch synthesis. Conversion of alcohols over iron oxide and iron carbide catalysts

Both iron oxide (Fe₂O₃) and iron carbide catalysts are active for the dehydration of tertiary alcohols; the oxide catalyst is not reduced nor is the bulk carbide oxidized by the steam generated during the dehydration reaction. Secondary alcohols are selectively converted to ketones plus hydrogen by both the iron oxide and carbide catalyst. Fe₂O₃ is reduced to Fe₃O₄ during the conversion of secondary alcohols. Both iron carbide and oxide catalysts dehydrogenate a primary alcohol (C_n) to an aldehyde which undergoes a secondary ketonization reaction to produce a symmetrical ketone with 2n−1 carbons. These results plus those of our earlier ¹⁴C-tracer studies suggest that dehydration of alcohols to produce olefins makes a minor, if any, contribution during Fischer–Tropsch synthesis with an iron catalyst at low and intermediate pressure conditions.     

On the effect of water during Fischer–Tropsch synthesis with a ruthenium catalyst

The addition of water during Fischer–Tropsch synthesis over a supported ruthenium catalyst led to a significant increase in product formation rates and significant changes in product selectivity, in particular lower methane selectivity and improved chain growth. Upon increasing water partial pressures, the total product distribution shifted from ASF distributions, with typical deviations due to olefin reinsertion, to a much narrower distributions. Such distributions can mechanistically not be explained by sole C₁-wise chain growth. An additional product formation route considering combination of adjacent alkyl chains to form paraffins (“reverse hydrogenolysis”) has been proposed. The findings are discussed with regard to the crucial mechanistic role of water as a moderator in the kinetic regime of the Fischer–Tropsch synthesis.     

Interpretation and kinetic modeling of product distributions of cobalt catalyzed Fischer–Tropsch synthesis

Carbon number distributions of Fischer–Tropsch products on iron and cobalt catalysts show deviations from the ideal Anderson–Schulz–Flory (ASF) distribution. For products obtained on cobalt catalysts these deviations are traced back by many authors to re-adsorption and incorporation of 1-alkenes followed by subsequent chain growth. In the present work, it could be shown by means of model calculations and based on experiments with co-feeding of ethene and 1-alkenes that such subsequent chain growth cannot be regarded as the main reason of observed deviations of the carbon number distribution from the ideal ASF distribution. The co-feeding experiments suggest that these deviations are the consequence of two different mechanisms of chain growth causing a superposition of two ASF distributions.

Consequently, the carbon number distributions are represented by this superposition. In order to describe distributions as a function of reaction conditions the model parameter, the growth probabilities ₁ and ₂ as well as μ₁, the fraction of distribution (₁) are presented as function of the partial pressures of hydrogen and carbon monoxide. Finally, the typical model parameters of products formed on cobalt and iron are compared.     

Co from partial reduction of La(Co,Fe)O3 perovskites for Fischer–Tropsch synthesis

Small Co clusters (d<10 nm) supported over mixed La–Co–Fe perovskites were successfully synthesized. These catalysts were active for Fischer–Tropsch (FT). Depending on the Co to Fe ratios the mixed perovskite exhibited two different forms: the rhombohedral phase of LaCoO₃ is maintained for the mixed perovskite when x>0.5, the orthorhombic phase of LaFeO₃ is found for x<0.5. Interestingly only one of these structures is active for the FT reaction: the orthorhombic structure. This is most likely due to the capacity of this material to maintain its structure even with a high number of cation vacancies. These cations (mostly Co) were on purpose extracted and reduced. Magnetic measurements clearly showed their metallic nature. Rhombohedral Co–Fe mixed perovskites (x≥0.5) cannot be used as precursors for Fischer–Tropsch catalysts: their partial reduction only consists in a complete reduction of Co³⁺ into Co²⁺.

The partial reduction of orthorhombic perovskites (x<0.5) leads to active Fischer–Tropsch (FT) catalysts by formation of a metal phase well dispersed on a cation-deficient perovskite. The FT activity is related to the stability of the precursor perovskite. When initially calcined at 600 °C, a maximum of 8.6 wt.% of Co⁰ can be extracted from LaCo_0.40Fe_0.60O₃ (compared to only 2 wt.% after calcination at 750 °C). The catalyst is then composed of Co⁰ particles of 10 nm on a stable deficient perovskite LaCo_0.05³⁺Fe_0.60³⁺O_2.40. Catalytic tests showed that up to 70% in the molar selectivity for hydrocarbons was obtained at 250 °C, 40% of which was composed of the C₂–C₄ fraction.