SEARCH

SEARCH BY CITATION

Keywords:

  • Fischer–Tropsch synthesis;
  • heterogeneous catalysis;
  • nanoparticles;
  • nanoporous materials;
  • selectivity control

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. 1. Overview of Catalysts, Reactors, Reaction Mechanisms, and Product Distributions for FT Synthesis
  5. 2. Key Factors Determining the Selectivity and Activity for FT Synthesis
  6. 3. Utilization of Nanoporous Materials for FT Synthesis
  7. 4. Utilization of Zeolites for FT Synthesis
  8. 5. Utilization of Novel Carbon Materials for FT Synthesis and the Size Effect of Metal Particles
  9. Summary and Outlook
  10. Acknowledgements
  11. Biographical Information
  12. Biographical Information
  13. Biographical Information

Fischer–Tropsch synthesis is a heterogeneous catalytic process for the production of clean hydrocarbon fuels or chemicals from synthesis gas (CO+H2), which can be derived from non-petroleum feedstocks such as natural gas, coal, or biomass. Fischer–Tropsch synthesis has received renewed interests in recent years because of the global demand for a decreased dependence on petroleum for production of fuels and chemicals. The product distributions with conventional Fischer–Tropsch catalysts usually follow the Anderson–Schulz–Flory distribution and are typically unselective with regards to the formation of hydrocarbons from methane to waxes. Selectivity control is one of the key challenges of research into Fischer–Tropsch synthesis. This Review article summarizes the effects of key factors on catalytic properties, particularly the product selectivity, and highlights recent developments of novel Fischer–Tropsch catalysts and new strategies with an aim at controlling the product selectivity.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. 1. Overview of Catalysts, Reactors, Reaction Mechanisms, and Product Distributions for FT Synthesis
  5. 2. Key Factors Determining the Selectivity and Activity for FT Synthesis
  6. 3. Utilization of Nanoporous Materials for FT Synthesis
  7. 4. Utilization of Zeolites for FT Synthesis
  8. 5. Utilization of Novel Carbon Materials for FT Synthesis and the Size Effect of Metal Particles
  9. Summary and Outlook
  10. Acknowledgements
  11. Biographical Information
  12. Biographical Information
  13. Biographical Information

Fischer–Tropsch (FT) synthesis is a heterogeneous catalytic process for the transformation of synthesis gas (syngas, CO+H2) into hydrocarbons. This process was first reported more than eighty years ago by two German chemists, Han Fischer and Franz Tropsch.1 The FT process generally includes the following reactions:((1)), ((2))

  • equation image((1))
  • equation image((2))

Both reactions are strongly exothermic (ΔH=−165–204 kJ molequation image). In addition to the formation of alkanes and alkenes, which are usually the target products for FT synthesis, organic oxygenates may be formed to some extent [Equation (3)]. Moreover, the water–gas shift (WGS) reaction, [Equation (4)], also occurs over most FT catalysts.

  • equation image((3))
  • equation image((4))

Syngas can be produced from many non-petroleum resources, such as natural gas, coal-bed gas, landfill gas, coal or biomass, through steam reforming, partial or autothermal oxidation, or gasification processes. The hydrocarbon products of FT synthesis can be sulfur- and nitrogen-free high-quality fuels such as diesel fuels, which have been proven to be more environmentally benign than the petroleum-based fuels,2, 3 and thus may easily meet the increasingly stringent environmental regulations (e.g., low residual sulfur content). Moreover, chemicals such as α-alkenes or C2–C4 lower alkenes may also be directly produced from syngas if a highly selective FT catalyst can be developed. Therefore, FT synthesis is a crucial step for the transformation of non-petroleum resources into super-clean fuels or valuable chemicals from syngas.

The commercialization of FT process began in 1936 in Germany, and many FT plants have since been built for the production of fuels.4, 5 Sasol built the first FT plant in South Africa in 1955, and the second and the third larger-scale plants in 1980 and 1982 because of the extremely cheap domestic coal and the particular state policy in South Africa. In 1993, the Shell middle-distillate synthesis plant came into operation in Malaysia with a capacity of 0.5 million tons per year.5 In 2007, Oryx started a plant with a capacity of roughly 1.4 million tons per year.6 It should be noted that the economic interest of the FT process depends vitally on the oil price. Thus, only a few FT plants could survive in the period of the “oil age”, when a plenty and cheap crude oil supply was possible, whereas the world oil crises prompted the construction of new FT plants. It was estimated that the FT process would be economically preferable when the oil price was above approximately US$20 per barrel.5

Although crude oil supplies may remain for 40 years or more, as expected,7 its price is far above US$20 per barrel and will still rise. Moreover, because light and sweet oil reserves are being depleted rapidly, we are facing the need to exploit lower-quality heavy oils, tar sands, and shale oils, which are highly aromatic and are characterized by high concentrations of heteroatoms, such as sulfur, nitrogen, and metals. These oil resources are not suitable for the production of either clean diesel fuels or linear alkenes. Therefore, the utilization of other carbon resources such as natural gas, coal and biomass to replace the crude oil for energy and chemical productions has become urgent, and has led to a global renaissance of FT synthesis. Many FT plants have been planned or are being constructed by ExxonMobil, Syntroleum, BP, and Chinese companies. Not only gas-to-liquid (GTL) but also coal-to-liquid (CTL) and biomass-to-liquid (BTL) technologies have been developed with FT synthesis as the key step.6, 8, 9 Moreover, the number of publications in 2009 related to FT synthesis (ca. 330 based on Thomson Reuters, ISI Web of Knowledge search using keyword “Fischer–Tropsch”) was almost thrice that in 1998, indicating the resurgent of the interest in FT synthesis also in the academic community.

The development of novel catalysts with high activity and selectivity, especially the latter, is the key to improving FT technologies and is one of the main focuses in the academic community. Many good reviews on FT synthesis are available,46, 822 but few focus on developing catalysts with controllable product selectivity. Herein we analyze the key factors influencing FT product selectivity and highlight recent developments of novel FT catalysts and strategies that aim at regulating product distributions.

1. Overview of Catalysts, Reactors, Reaction Mechanisms, and Product Distributions for FT Synthesis

  1. Top of page
  2. Abstract
  3. Introduction
  4. 1. Overview of Catalysts, Reactors, Reaction Mechanisms, and Product Distributions for FT Synthesis
  5. 2. Key Factors Determining the Selectivity and Activity for FT Synthesis
  6. 3. Utilization of Nanoporous Materials for FT Synthesis
  7. 4. Utilization of Zeolites for FT Synthesis
  8. 5. Utilization of Novel Carbon Materials for FT Synthesis and the Size Effect of Metal Particles
  9. Summary and Outlook
  10. Acknowledgements
  11. Biographical Information
  12. Biographical Information
  13. Biographical Information

The typical active metals used in Fischer–Tropsch catalysts are Fe, Co, and Ru although several other metals, such as Ni and Rh, also exhibit activities for Equations (1) and (2).23 Among these metals, Ru is the most active catalyst for CO hydrogenation, and is capable of working at low temperatures (<150 °C), producing long-chain hydrocarbons.11 Ru can work efficiently without any promoters,11 and thus it may provide more straightforward fundamental insights into the catalyst functioning mechanism and the reaction mechanism. However, high costs and the limited reserves hinder its industrial-scale application. Cobalt and iron have both been employed in industry for FT synthesis. Fe is cheaper than Co, but Co-based catalysts are generally more active and more selective to linear long-chain hydrocarbons. Moreover, Co catalysts are typically more resistant to deactivation by water.10 Thus, Co catalysts have attracted much attention for the synthesis of long-chain linear hydrocarbons, such as wax and diesel fuel.10, 11, 13 On the other hand, Fe-based catalysts can be operated under wider ranges of temperatures and H2/CO ratios without significant rise in CH4 selectivity, whereas Co catalysts only work well under carefully selected temperatures and H2/CO ratios. Moreover, Fe catalysts can not only be used for the production of linear alkane fuels but are also suitable for the production of alkenes or oxygenates, which are important chemical feedstocks. In addition, Fe-based catalysts exhibit much higher activity for the WGS reaction than Co- or Ru-based catalysts. This is helpful for the conversion of syngas with lower H2/CO ratios derived from coal or biomass, but is undesirable for the conversion of H2-rich syngas produced from methane. Therefore, these advantages make the Fe-based catalysts quite attractive for CTL or BTL technology and for the production of alkenes from syngas.14, 15 Generally, heavier modifications are required for Fe catalysts to afford good activities and selectivities. To overcome rapid catalyst deactivation is a big challenge for Fe catalysts.15

FT processes are often divided according to the operation temperatures; low-temperature FT (LTFT) typically operates at 190–260 °C and high-temperature FT (HTFT) at 300–350 °C. The reactors designed for FT synthesis are mostly one of three types; fixed-bed (typically multitubular), slurry-phase, or fluidized bed reactors.12 Typically, fixed-bed and slurry-phase reactors are employed for LTFT processes with either Co or Fe catalysts for the production of linear long-chain alkanes, whereas fluidized bed reactors are used for HTLT processes with Fe catalysts for the production of C1–C15 hydrocarbons and α-alkenes.13 New types of reactors, such as monolith-structured, microstructured, and membrane reactors, have also been studied for FT synthesis.16

The reaction mechanism for FT synthesis is quite complicated, and many reviews have made efforts to describe it.1722 It is now generally accepted that FT synthesis proceeds through a surface-catalyzed polymerization mechanism, which uses CHx monomers formed by hydrogenation of CO. Recent DFT calculations over model Ru or Co surfaces suggested that both direct CO dissociation and hydrogen-assisted CO dissociation via HCO intermediates may occur, depending on the type of surface or site (i.e., terrace or step sites).24, 25 Adsorbed O can be efficiently removed by H to form water, whereas adsorbed C can recombine with H to yield various CHx intermediates (x=0–3). Then, the chain growth through C–C coupling starts, in competition with chain termination through hydrogenation, hydrogen abstraction, or insertion of nondissociatively adsorbed CO to produce alkanes, alkenes, or alcohols, respectively.10 The mechanism for the C–C coupling is still an open question, and many potential reactive chain carriers (true monomers), such as adsorbed CH2,26 CH,27 C,28 and CHδ+, have been proposed.22 Both experimental and theoretical studies may play important roles in elucidating the coupling mechanism.

Although some challenges still remain, many extensive studies have contributed to the development of efficient FT catalysts and elucidation of the FT reaction mechanism. Selectivity control is one of the most important and difficult challenges for FT synthesis. As a result of the polymerization mechanism, the products of FT synthesis generally follow a statistical hydrocarbon distribution, which is known as the Anderson–Schulz–Flory (ASF) distribution.29 In the ideal case, when the chain-growth probability (α), which is determined by the rates of chain growth (Rp) and chain termination (Rt), and is expressed by α=Rp/(Rp+Rt), is independent of carbon chain length, the molar fraction M of a hydrocarbon with a chain length (carbon number) of n can be expressed as:((5))

  • equation image((5))

Therefore, the product distribution is determined by the α value (Figure 1). Such a statistical distribution is nonselective for a desired range of hydrocarbons. For example, the maximum selectivities to C5–C11 (gasoline range) and C12–C20 (diesel range) hydrocarbons are roughly 45 % and 30 %, respectively.

thumbnail image

Figure 1. Product distribution in FT synthesis as a function of the chain growth probability (α).

Download figure to PowerPoint

In this context, the current FT technology generally aims at producing long-chain alkanes (C21+, waxes). In the subsequent steps, the FT waxes are transformed into liquid fuels (mainly diesel fuel) by the hydrocracking process over metal–acid dual-functional catalysts.2, 3 To take full advantage of this two-stage approach, the FT process should be operated at a high α value (>0.9) by minimizing the formation of undesired light products, especially CH4. For this purpose, it is necessary to design and prepare FT catalysts with an excellent selectivity to C5+ hydrocarbons.

As compared with this two-stage process, a one-stage FT process for the direct production of high-quality liquid fuels without hydrocracking units would be more energy efficient. Moreover, the direct production of valued chemicals such as olefins from syngas is also a promising route, but there is still no catalyst that demonstrates enough selectivity for this purpose. Therefore, the development of new strategies and novel catalysts that can tune the selectivity to desired products is an important goal for academic research.

2. Key Factors Determining the Selectivity and Activity for FT Synthesis

  1. Top of page
  2. Abstract
  3. Introduction
  4. 1. Overview of Catalysts, Reactors, Reaction Mechanisms, and Product Distributions for FT Synthesis
  5. 2. Key Factors Determining the Selectivity and Activity for FT Synthesis
  6. 3. Utilization of Nanoporous Materials for FT Synthesis
  7. 4. Utilization of Zeolites for FT Synthesis
  8. 5. Utilization of Novel Carbon Materials for FT Synthesis and the Size Effect of Metal Particles
  9. Summary and Outlook
  10. Acknowledgements
  11. Biographical Information
  12. Biographical Information
  13. Biographical Information

Generally, the nature of the catalyst, the reactor, and the operating conditions are the main factors affecting the product selectivity and the CO conversion activity for FT synthesis. Several reviews and recent publications have dealt in detail with the influences of the reactor types and operating conditions on FT activity and selectivity.16, 20, 3037 In this section, we will focus on the effects of catalyst factors, especially the chemical state of active components and the nature of supports and promoters.

2.1. Effect of chemical state of active components

Insights into the chemical state of the active sites and phases are crucial for the design of highly active and selective FT catalysts. For Co and Ru catalysts, it is generally accepted that the metallic nanoparticles are the active phases.11, 13 Experimental results have indicated that with Co catalysts, FT synthesis only proceeds on metallic (nonoxidized) Co species.13 A recent study using a model Co/SiO2 catalyst formed by depositing Co on a silica film, which combined the reaction and the in situ characterization results, confirmed that metallic Co was required for FT synthesis, and the oxidizing of Co/SiO2, even with a very low pressure of O2 (1.3 mPa), led to negligible FT activity.38

Metallic cobalt can exist in different crystalline forms, including α-Co (hcp) and β-Co (fcc), and the former one is more stable at lower temperatures for bulk cobalt. However, cobalt crystallites with a particle size less than 20 nm are stable as pure fcc phase.39 The sizes of Co particles in many FT catalysts are in this region, and thus, the fcc Co crystallites may play key roles in FT synthesis. There exists a study suggesting that the fcc Co is less active than hcp Co in FT synthesis.40 For Co particles smaller than 3 nm, the number of surface atoms increases over that of interior atoms, and the shape of Co clusters is often that of the ideal Platonian structure composed of similar regular polyhedra such as cuboctahedra.41

Unlike Ru species, which often undergo facile reduction at moderate temperatures (<673 K), the reduction of Co species sometimes may be challenging. Unreduced Co species typically give lower CO conversions and higher CH4 selectivity. Moreover, under working conditions, the transformation of Co0 to cobalt oxides, cobalt carbides, or mixed oxide compounds (e.g., cobalt aluminate or cobalt silicate) may occur as a result of reaction with the supports. Cobalt oxide formation due to oxidation by water formed has been proposed as one of the possible reasons for catalyst deactivation.10, 13 From thermodynamic calculations, although bulk Co0 is not oxidized to CoO or Co3O4 during realistic FT reaction conditions,42 smaller Co crystallites (<5 nm) may be oxidized to CoIIO by H2O formed.43 However, a study using a model Co/SiO2/Si(100) catalyst with well-defined Co crystallites (4–5 nm) demonstrated that no surface oxidation of Co occurred at P(H2O)/P(H2)=1 and 423–673 K.44 In situ characterizations of working FT catalysts represent a significant challenge because of the high temperature, high pressure, and the complex reaction system. Recently, Khodakov and co-workers45 studied the working state of Co species over a 0.1 wt % Pt-promoted 25 wt % Co/γ-Al2O3 catalyst under realistic FT reaction conditions (T=493 K, P=2 MPa, H2/CO=2) by using synchrotron-based in situ time-resolved XRD in combination with STEM-EELS (scanning tunneling electron microscopy electron energy-loss spectroscopy). They observed mainly fcc metallic Co together with some hcp Co under working conditions, and found that the size of fcc Co particles increased from 6 to 10 nm after 3–5 h of reaction. Further prolonging of the time on stream led to the formation of a cobalt carbide, Co2C. No reoxidation of Co species was detected. These observations suggest that sintering and the carbidization are responsible for the deactivation of the alumina-supported cobalt catalyst.

Generally, cobalt carbide species are believed to play an insignificant role in FT synthesis because the rate of carbon diffusion over Co to form carbides is very low. The possibility of formation of cobalt carbides such as Co2C and Co3C is thought to be low under FT synthesis conditions, although these carbides are stable and have been reported occasionally.15, 45 Recently, Co2C was found to form over a La-promoted Co/AC catalyst (AC=activated carbon) operated under the conditions of T=495 K, P=3 MPa, H2/CO=2, and gas-hourly space velocity (GHSV)=1500 h−1.46 This catalyst gave high selectivity (ca. 36 %) to linear higher α-alcohols (C2–C18 alcohols). It is proposed that the Co2C may be responsible for the high selectivity of linear α-alcohols through playing a role in CO insertion into the CnHm–Co species formed at Co0 sites.46 The combination of metallic Co and cobalt carbide species may provide a promising strategy for higher alcohol synthesis.

For Fe-catalyzed FT synthesis, the nature of the active sites or phases is still in dispute. Unlike Ru and Co catalysts, iron carbides can be easily formed under working conditions for Fe catalysts, due to the lower activation energy of iron carbide formation (43.9–69.0 kJ mol−1)15, 47 compared to that of FT synthesis over the Fe catalyst (ca. 90 kJ mol−1).48 There are several types of iron carbides with different structures, which can be classified on the basis of the sites occupied by the carbon atoms.15 The presence of carbon atoms in trigonal prismatic interstitial sites leads to carbides such as θ-Fe3C (cementite), χ-Fe5C2 (Hägg carbide) and Fe7C3 (Eckstrom and Adcock carbide), whereas the presence of carbon atoms in octahedral interstitial sites results in ε-Fe2C and ε′-Fe2.2C (hexagonal carbides). FexC (or FexCy) usually denotes iron carbides with poorly defined structures. Among these carbides, hexagonal iron carbides (ε-Fe2C) can be formed by carburizing fine iron or iron oxide powder in CO flow at low temperatures (e.g., 443 K), whereas the formation of χ-Fe5C2 occurs at a higher carburization temperature (ca. 523 K) and θ-Fe3C can be produced by carburizing the reduced iron in syngas at temperatures greater than 573 K. Under working conditions, several iron species, typically α-Fe, γ-Fe, and Fe3O4 (magnetite), may coexist with iron carbides. The functions of these phases in FT synthesis are still ambiguous.15

Fe catalysts usually undergo reconstruction under FT reaction conditions, and long reduction periods are often required to reach the steady states. Schulz and co-workers49 studied the construction of a Fe–Al–Cu/K2O catalyst in H2/CO, and concluded that the FT activity was related to the formation of the iron carbide (Fe5C2), and the metallic iron was less active. Oxidic iron species appear to be responsible for the WGS reaction, one of the main side reactions over Fe catalysts. Recently, in situ or quasi-in situ techniques have been applied to elucidating the possible active phase for Fe-based FT catalysts. By using quasi-in situ TEM-EELS and XRD measurements (without exposure to air), Janbroers et al.50 revealed that the Fe2O3 phase in the fresh K- and Cu-promoted Fe catalyst was transformed into Fe3O4 and an iron carbide with unknown structure in CO flow up to 543 K. They also showed that the carburized catalyst was very sensitive to air exposures including uncontrolled or controlled oxidation. Thus, these authors pointed out that, to characterize the true structure of Fe species under working conditions, the passivation, even under controlled conditions, should also be avoided. De Smit et al.51 reported an in situ characterization of the K- and Cu-promoted Fe catalyst supported on SiO2 using scanning transmission X-ray microscopy (STXM) combined with a nanoreactor. This technique allowed a spatial resolution of approximately 15 nm. They detected Fe2O3 as the sole iron phase in the fresh catalyst and this hematite phase was transformed to Fe3O4 and Fe2SiO4 (a FeII silicate) after H2 reduction. After reaction in syngas (0.1 MPa) at 523 K, Fe3O4 was further converted to Fe0 and Fe2SiO4, and iron carbides (FexCy) were also formed after the reaction. By using combined in situ X-ray absorption fine structure (XAFS) and wide angle X-ray scattering (WAXS) techniques, de Smit et al.52 investigated in detail the structural changes for the K- and/or Cu-modified Fe catalysts with and without SiO2 support during the pretreatments and the FT reactions in a fixed-bed reactor. They found that the unsupported Fe2O3 and Cu-modified Fe2O3 were largely reduced to α-Fe by H2 at 623 K for 2 h, and these samples were readily converted to iron carbides under FT reaction conditions (0.1 MPa and 523 K). Both unsupported catalysts deactivated rapidly during the first 4 h of reaction, and θ-Fe3C was the main phase for both unsupported catalysts, suggesting that θ-Fe3C may contribute to the deactivation. The reduction of the K- and Cu-modified Fe2O3/SiO2 catalyst occurred much slower; after reduction by H2 at 623 K, Fe3O4 and Fe2SiO4 were the main detected phases. This supported iron catalyst was activated very slowly during FT reactions. The pretreatment of the unsupported Fe2O3 and Cu-promoted Fe2O3, and the SiO2-supported K- and Cu-promoted Fe2O3 catalysts in H2/CO led to the formation of fcc γ-Fe and χ-Fe5C2. The catalysts after H2/CO activation showed better stability and activity than those after H2 reduction, suggesting that the γ-Fe or χ-Fe5C2 may be crucial. The Fe-based catalysts are usually sensitive to the pretreatment conditions, which may be due to the formation of different structures of iron and/or iron carbides.

2.2. Effect of the nature of catalyst support

Supports are expected to play the following roles in heterogeneous catalysis: 1) to disperse the active phase giving a high surface area of the catalytically active phase; 2) to stabilize the active phase against loss of surface area during the reaction; 3) to maintain the catalyst mechanical strength and to facilitate the mass or heat transfer in a diffusion-limited or an exothermic reaction. To increase the attrition-resistance of a catalyst by using a proper support is crucial for FT synthesis in slurry phase,53 whereas for FT synthesis in a fixed-bed reactor, the efficient dissipation of heat is crucial. In addition to these physical effects, the chemical interaction between the active phase and the support may also significantly affect the catalytic behaviors. It is accepted that a balanced interaction between the support and the active phase (or the precursor of active phase) is particularly important for FT synthesis. Although too weak an interaction may lead to a poor dispersion of active phase, too strong an interaction will cause difficulty in the reduction of the precursor of the active phase.54 The support may also change the electronic state of the active metal, and thus, affect the CO dissociation ability.13 Furthermore, the pore structure of the support can significantly affect the catalytic performance through changing the reducibility and the size of the active phase or influencing the diffusion of reactants or products. The effect of pore size of the support will be discussed later.

Supports are used for most FT catalysts, especially for Co and Ru catalysts, which are more expensive than Fe catalysts. Oxides, particularly SiO2, Al2O3, and TiO2, are probably the most extensively investigated supports for Co catalysts. Many studies have attempted to clarify the effects of these supports on catalytic properties of supported Co catalysts.13, 14, 5460 In an early paper,55 Bartholomew and Reuel reported a detailed study on CO hydrogenation over Co catalysts supported on different supports (3 wt % Co loading) at 0.1 MPa and 498 K. They found that the specific activity of CO hydrogenation decreased in the order of Co/TiO2>Co/SiO2>Co/Al2O3>Co/C>Co/MgO. They also found that, over a certain support, the specific activity was determined by the Co loading and the dispersion of Co species, and increased with increasing the Co loading (or decreasing the dispersion). The product selectivity was also a function of the support and Co dispersion; higher selectivities to lighter hydrocarbons and CO2 were observed over the catalysts with higher dispersions and lower degrees of reduction.55 Davis and co-workers56 compared the catalytic performances of Co catalysts supported on SiO2, Al2O3, and TiO2. They found that the metal–support interactions affected the reduction of Co species, and the strength of such interactions decreased in the order Al2O3>TiO2>SiO2. The addition of a noble metal promoter, such as Ru or Pt, significantly enhanced the reduction of Co species and thus increased the initial activity of Al2O3- and TiO2-supported Co catalysts.56 By comparing the activity over Co catalysts supported on SiO2, Al2O3, TiO2, and other composite metal oxides, Iglesia10 found that the turnover frequency (TOF) for CO conversion was independent of Co dispersion and support identity in a dispersion range of 0.01–0.12 under typical FT reaction conditions (2MPa and 473 K). By using an in situ-measured rapidly exchanging adsorbed CO to count the number of active Co site, Bertole et al.57 also reached the conclusion that the TOF for CO conversion and the selectivity of CH4 were not affected by the identity of the support (SiO2, Al2O3, or TiO2) although the modification of the support by some basic metal oxides (e.g., Y2O3 or ZnO) might decrease the TOF unexpectedly. Holmen and co-workers58 found that the pore sizes of SiO2, γ-Al2O3, and TiO2 affected the size of cobalt particles; smaller Co particles with agglomeration were found on γ-Al2O3 and SiO2, whereas larger Co particles without agglomeration were detected on TiO2, which possessed a larger pore size. The latter catalyst showed higher C5+ selectivity, especially after modification with Re or in the presence of steam.58 For the slurry-phase reaction, the pore size of different supports was found to be a key factor because the diffusion problem became serious in this case.59 Among the SiO2-, γ-Al2O3-, and TiO2-supported Co catalysts, Co/TiO2 possessed the largest average pore size (ca. 16 nm), which aided an easy diffusion of heavier FT products, and led to the highest C5+ selectivity (ca. 95 % at 2MPa and 493 K) and the largest TOF.59

Besides SiO2, Al2O3, and TiO2, other metal oxides or composite metal oxides have also been employed as supports for Co catalysts. Enache et al.60 compared the catalytic behaviors of Co catalysts supported on γ-Al2O3 and ZrO2, and found that the Co supported on ZrO2 provided a higher TOF than that on γ-Al2O3. Higher selectivity to C5+ hydrocarbons and a lower selectivity to CH4 were obtained over Co/ZrO2. A larger amount of adsorbed hydrogen species, was detected over the ZrO2-supported catalysts, and they might correspond to the hydrogen species on ZrO2 spilled over from Co surfaces, which were assumed to contribute to the higher activity of the Co/ZrO2 catalyst. Bae et al.61 recently demonstrated that the use of amorphous AlPO4 with a mean pore size of 19.4 nm provided a higher CO conversion activity than Co/Al2O3. They also found that the addition of a small amount of phosphorus to Al2O3 could enhance the C5+ selectivity, as well as the TOF.62

In short, the effects of supports on catalytic behavior are complicated for FT synthesis, and comprehensive analyses are often required. It is difficult to simply discuss the effect of the identity of a support. It appears that the interaction between the support and the active metal (or its precursor), the pore structure of support (including the pore size) and the location of metal particles are important factors when the effects of support are concerned. These factors can determine the metal degree of reduction and the metal particle size or morphology. Moreover, the diffusion situation may also be different over catalysts with different pore textures. The utilization of materials with well-defined nanoporous structures as FT catalyst supports has provided new possibilities for tuning the catalytic properties (particularly the selectivity). Recent development in this direction will be highlighted in the later sections.

2.3. Effect of distribution of active component in large catalyst pellets

For the fixed-bed reactor, relatively large catalyst pellets (1–3 mm) are required to decrease the pressure gradients. However, this may lead to the intrapellet transport (diffusion) limitations. The mass transfer may significantly influence the product distributions in FT synthesis.10 Typically, two kinds of diffusion limitations are considered to influence the product selectivity in FT synthesis. Firstly, the limited reactant transport would increase the H2/CO ratio at the catalytic sites because H2 diffuses more quickly than CO. This may enhance the hydrogenation activity and unfavorably increase CH4 selectivity. Secondly, the limited product transport may either restrain the removal of heavier hydrocarbons or enhance the readsorption of α-olefins resulting in higher selectivity to heavier paraffins. The control of distribution of active components in the large catalyst pellets may affect the transport of either reactants or products, and alter the product selectivity.

Iglesia et al.63 found that the increase in the diameter of Co/SiO2 (24.8 wt % Co loading) pellets from 0.36 to 0.86 mm, which contained uniformly distributed Co, increased CH4 and CO2 selectivities. The decrease in CO/H2 ratio due to the increased diffusion limitation in larger pellets resulted in lower C5+ selectivity and lower olefin content. Moreover, the water concentration inside the larger pellets became higher due to the slower removal rate of H2O, leading to a faster WGS reaction and higher CO2 selectivity. The same research group demonstrated that the use of egg-shell catalysts, in which the active Co component is preferentially located near the outer pellet surface, could improve the FT reaction rate and C5+ selectivity.54, 63 They prepared the egg-shell Co/SiO2 (ca. 13 wt % Co loading) catalyst by an impregnation technique using molten cobalt nitrate, and the shell thickness could be controlled by the melt viscosity and the contact time. The maximum C5+ selectivity was achieved over the egg-shell catalyst with a medium shell thickness (100 μm in a 2.2 mm SiO2 sphere). It was believed that this catalyst could avoid an intrapellet H2/CO ratio gradient but still enhance the readsorption of primary α-olefin products. Recently, Ding and co-workers64 prepared egg-shell Co/SiO2 (ca. 2 mm with a shell thickness of ca. 200 μm) using the air entrapped in the hydrophobic SiO2 to inhibit the entrance of the impregnation solution. They also found that the egg-shell catalyst exhibited significantly higher C5+ selectivity (ca. 85 %) than the catalyst with uniformly distributed Co (ca. 78 %).

2.4. Effect of the nature of the promoter

Promoters in heterogeneous catalysts may enhance the activity or modify the selectivity either by a structural effect, changing the structure of the active phase (e.g., the exposed crystalline plane, the size or dispersion of the active phase), or by an electronic effect, modifying the electronic characters of the active phase through electron transfer or electronic interaction. For FT synthesis, Ru-based catalysts can work well for the production of heavier hydrocarbons without modifiers, but Fe- and Co-based catalysts generally require promoters such as alkali metal ions, noble metals, or transition metal oxides to attain optimum catalytic performance.11 The elucidation of the effects and the functioning mechanisms of the promoters in FT synthesis is very important for rational design of selective and active FT catalysts.

2.4.1. Promoter effects for Fe-based catalysts

Promoters have played particularly significant roles in Fe-catalyzed FT synthesis. Most of the Fe-based FT catalysts contain alkali metal ions as promoters. The alkali metal ion is expected to function as an electronic promoter to affect the electronic character of Fe, and can modify the activity and selectivity by enhancing the chemisorption of CO and inhibiting that of H2. The effect of various alkali metal ions on the catalytic performance of a precipitated Fe catalyst (Fe/Si=100:4.6) was investigated in a slurry-phase reaction (T=543 K, P=1.3 MPa).65 The use of K+ and Na+ was found to accelerate the activity for both FT synthesis and the WGS reaction, whereas the addition of Li+, Rb+, or Cs+ decreased CO conversion. The alkali metal ion promoters changed the product selectivity. Modification with alkali metal ions could enhance C5+ selectivity. The addition of Na+ particularly favored the C5–C11 selectivity (the gasoline fraction, ca. 35 wt %), whereas the use of K+ and Cs+ provided the highest selectivities to C19+ (ca. 12 wt %) and C12–C18 (ca. 20 wt %), respectively.65 Alkali metal ions increased ethylene selectivity in C2 hydrocarbons. It is proposed that modification with alkali metal ions decreases the hydrogenation ability and thus accelerates the chain-growth probability and the olefin selectivity, possibly as a result of the increased basicity. The effect of K+ content on catalytic performances of a precipitated Mn-modified Fe catalyst operating in a fixed-bed reactor was investigated in detail.66 The CO conversion increased with K+ content and passed through a maximum at a K+ content of 0.7 wt % (Figure 2). The increase of K+ content to about 1.5 wt % decreased the selectivity to CH4 and increased that to C5+ hydrocarbons, especially to C12+ hydrocarbons. The selectivity to oxygenates decreased on addition of K+. The fraction of olefins in C2–C4 hydrocarbons increased significantly at the same time. XRD and Mössbauer spectroscopy studies suggested that the increase in K+ content increased the fraction of iron carbides (χ-Fe5C2 and ε′- Fe2.2C) after the reaction. In a subsequent study for a K+- and Cu-promoted Fe/SiO2 catalyst in slurry-phase reaction (T=523–533 K, P=1.5MPa, H2/CO=0.67), the presence of K+ was found to increase both the activities of the FT synthesis and the WGS reactions and significantly enhance the C5+ selectivity.67 However, too high a K+ content led to rapid catalyst deactivation. Lohitharn and Goodwin68 showed that the FT reaction activity increased by adding an appropriate amount of K+ to Fe or Fe–Mn catalyst, and the optimum activity was observed at a K+ content of 1.5 mol % (relative to Fe). C5+ selectivity and the ratio of olefin to paraffin were also enhanced by the modification with K+. However, activity towards CO2 formation, possibly including the WGS and Boudouard reactions, increased monotonically with K+ content. By measuring coke deposition, it was clarified that the addition of K+ with higher contents significantly promoted the Boudouard reaction [Equation (6)].

  • equation image((6))
thumbnail image

Figure 2. Effect of K+ content on catalytic performances of precipitated Mn-promoted Fe catalysts.66 Reaction conditions: T=550–573 K, P=2.5 MPa, H2/CO=2, GHSV=1000 h−1.

Download figure to PowerPoint

This reaction was thought to be the main reason for the rapid catalyst deactivation for the catalysts with higher K+ contents.67, 68 By using a steady-state isotopic transient kinetic analysis (SSITKA) technique, it was demonstrated that the addition of K+ does not significantly affect the intrinsic activity of the Fe site but may increase the number of active surface intermediates leading to hydrocarbon products.68

Copper or noble metals may facilitate the reduction of Fe precursors and thus can increase the activity. The presence of Cu or Ru in a K-Zn-Fe catalyst was found to promote the reduction and carburization rates of Fe species during the catalyst pretreatment in H2/CO.69 The activity per gram of catalyst was increased at the same time, but the TOF was not changed. The selectivities to C5+ and CH4 over the Cu- or Ru-promoted catalyst were similar to those over the unmodified catalyst when the comparison was made at similar CO conversions. The addition of Cu accelerated the reduction of Fe species in Fe–Mn–K/SiO2 catalysts and increased the rate of carburization but did not vary the final degree of carburization at the steady state.70 Accordingly, Cu shortened the induction period but did not exert significant influence on the stead-state activity. For the Fe–Mn–K/SiO2 catalysts, the selectivities to CH4 and lighter hydrocarbons (C2–C11) decreased whereas that to heavier hydrocarbons (C11+) increased on Cu modification. The ratio of olefins to paraffins in C2–C4 also increased after the addition of Cu to Fe–Mn–K/SiO2. These observations are quite different from those for the K–Zn–Fe catalyst.69 The basicity of the Cu-promoted catalyst was also found to be enhanced, possibly due to the synergistic effect between Cu and K+, which might be responsible for the changes in product selectivity.

Some transition metal oxides are also known to promote the Fe-catalyzed FT reactions. Goodwin and co-workers71 investigated the effect of various transition metal oxides (Cr, Mn, Mo, Ta, V, W, and Zr oxides) on catalytic behaviors of a 100 Fe/5 Cu/17 Si catalyst under reaction conditions of T=553 K, P=1.8 MPa, and H2/CO=1, and found that the addition of these metal oxides except for WOx enhanced the activities of both CO hydrogenation and WGS reactions. The enhancing effect was significant using Cr, Zr, and Mn oxide promoters, and the catalyst modified by MnOx was the most stable against catalyst deactivation. However, the hydrocarbon selectivity was not significantly changed by these promoters. Further studies with the Cr-, Mn-, and Zr-modified Fe catalysts suggested that the presence of these promoters increased the dispersion of Fe species but did not significantly change the reducibility or the BET surface area.72 The intrinsic activity estimated by the SSITKA technique was similar for the Fe catalysts with and without these promoters, implying that the enhanced conversions were due to an increased number of active surface intermediates leading to hydrocarbon products.72 Recently, an X-ray absorption near-edge structure (XANES) study was performed for the Mn-promoted Fe catalysts after reaction (with passivation in a 2 % O2/He mixture) to clarify the function of Mn.73 The result suggested that Mn might substitute for octahedral sites in Fe3O4, corresponding to the formation of (Fe1−xMnx)3O4. The formation of this mixed oxide likely decreased the reducibility of Fe. Typically, Fe-based FT catalysts deactivate due to carbon deposition on larger iron carbide (e.g., θ-Fe3C) particles. The lower deactivation rate of this Mn-modified catalyst may indicate that the (Fe1−xMnx)3O4 phase can be transformed to smaller FexC clusters, which are more active for CO hydrogenation and less prone to deactivation.73

Luo and Davis74 studied the influence of alkaline-earth metal ions on the catalytic behavior of Fe-based FT catalysts, and found that Mg and Ca could suppress the WGS reaction and slightly increase the ratio of olefins in C2–C4 hydrocarbons. Gallegos et al. reported that, as compared to the Fe/SiO2 catalyst, the Fe catalyst supported on MgO-covered SiO2 showed a higher TOF for total hydrocarbon productions and a higher ratio of olefin to paraffin in C2–C4 hydrocarbons.75 Moreover, the presence of Mg at an appropriate content (4 wt %) could suppress the formation of CH4. A systematic study on the modifying effect of Mg on a precipitated Fe–Cu–K/SiO2 catalyst showed that there was an optimum content of Mg for obtaining better catalytic performances.76 At an optimized Mg/Fe weight ratio (0.07), the activity of WGS reaction could be suppressed, whereas the FT activity was enhanced. Mg modification could also effectively shift the product distribution to lighter hydrocarbons, especially to gasoline-range hydrocarbons (C5–C11), and could suppress the hydrogenation of light olefins, resulting in higher selectivity to C2–C4 olefins. The characterizations suggested that the presence of an appropriate amount of Mg could promote the reduction and the carburization of Fe species, and over the catalyst with an Mg/Fe ratio of 0.07, the content of χ-Fe5C2 was the highest. This observation further suggests that the iron carbide (likely χ-Fe5C2) is the active phase for hydrocarbon formation, whereas the oxidic iron species is responsible for the WGS reaction.

Table 1 summarizes the effects of various promoters employed in Fe-based FT catalysts. CO conversions increase with most of these promoters. However, it is unclear whether the intrinsic activity (TOF) of the Fe site can be enhanced. Recent studies appear to indicate that most of the promoters cannot affect the TOF, but are capable of enhancing the reduction and the carburization of Fe species or facilitating the dispersion of active iron carbide species.6576 The product selectivity can be regulated by choosing appropriate modifiers. To achieve the goals of decreasing the selectivities to CH4 and CO2, and of promoting the C5+ or the light olefin formation, we have to combine different modifiers with different functions in one catalyst. Thus, the interactions among different modifiers should also be considered for the rational design of an efficient Fe-based catalyst. Such information as the synergistic effect between different promoters is still in deficiency.

Table 1. Effects of promoters on catalytic behaviors of Fe-based FT catalysts.
PromoterPossible functionsReferences
K+ and alkali metal ions1) Enhancing both FT and WGS activity. 2) Decreasing selectivity to CH4 and increasing that to C5+ hydrocarbons. 3) Increasing the olefin/ paraffin ratio in C2–C4 hydrocarbons.6568
Cu or Ru1) Increasing the activity by facilitating the reduction and carburization of Fe species. 2) Increasing the selectivity to heavier hydrocarbons and the fraction of olefins in lighter hydrocarbons, possibly by enhancing the basicity together with alkali metal ions.69, 70
MnOx1) Increasing the activity by enhancing the dispersion of Fe species. 2) Decreasing deactivation by forming smaller iron carbide species by the formation of mixed oxides with Fe.7173
MgO1) Increasing the FT reaction activity and suppressing the WGS reaction by promoting the reduction and carburization of Fe species. 2) Shifting the hydrocarbon product to lighter range, especially to gasoline range. 3) Increasing the selectivity to olefins.7476
2.4.2. Promoter effects for Co-based catalysts

Typical promoters used for Co-based catalysts are noble metals, transition metal oxides, such as ZrO2 and MnOx, and some rare earth metal oxides.

Many studies have demonstrated that the addition of a noble metal modifier such as Ru or Re can accelerate the reduction of Co precursors into Co0 clusters, which are the active phase for FT synthesis, and may also enhance the dispersion of Co0.1013 The increase in Co site density on a catalyst surface may not only increase the FT activity but also raise the selectivity to C5+ hydrocarbons and decrease that to CH4 due to the increased probability of readsorption of α-olefins, as proposed by Iglesia.10 Moreover, the presence of a noble metal may increase the steady-state activity by retarding the catalyst deactivation, perhaps through the inhibition of reoxidation or of carbon deposition on Co catalysts during the reaction.10

The promoting effect of reducibility of cobalt precursors by a noble metal was particularly significant for the catalysts containing Co species that are difficult to reduce. Khodakov and co-workers77 found that cobalt species in a calcined Co/SiO2 catalyst, prepared by an impregnation using cobalt acetate as the precursor were mainly barely reducible cobalt silicate, and the addition of Ru by co-impregnation with cobalt acetate could increase the fraction of Co3O4 by influencing the oxidative decomposition of cobalt acetate, and significantly enhanced the degree of reduction after H2 reduction at 673 K. However, for the Co/SiO2 prepared by the impregnation using cobalt nitrate, which was mainly transformed to reducible Co3O4 after calcination, the role of Ru modification was to enhance the dispersion of Co0 instead of the degree of reduction. Because of the stronger interactions with support, it is known that the cobalt species on γ-Al2O3 are difficult to reduce. The addition of small amount of Pt, Ru, or Pd to a Co/γ-Al2O3 was found to enhance the reduction of both the surface Co3O4 and the inactive cobalt oxide species, having strong interactions with the support, and thus increased overall activity.78 The activity increased in the order Co/γ-Al2O3<Pd-Co/γ-Al2O3< Ru-Co/γ-Al2O3<Pt-Co/γ-Al2O3, and the C5+ selectivity was also increased in the presence of noble metal promoters due to the increase in the Co0 site density.

Tsubaki et al.79 made a detailed comparison of promoting effects of different noble metals added into a Co/SiO2, and found that the CO hydrogenation rate increased in the order Co/SiO2<Pt–Co/SiO2<Pd–Co/SiO2<Ru–Co/SiO2. The addition of a small amount of Ru increased the degree of reduction of Co and the cobalt–time yield (moles of CO converted per total moles of Co per unit time) remarkably. The TOF based on surface Co evaluated by H2 uptake increased to some extent by Ru modification. The introduction of Ru led to the appearance of bridge-adsorbed CO to a larger extent, indicating the increase in CO dissociation ability. In contrast, Pd and Pt only slightly affected the degree of reduction but enhanced the cobalt dispersion, and decreased the TOF.79 CH4 selectivity was significantly raised over the Pt- and Pd-modified catalysts. The addition of a small amount of Pt (0.1 wt %) into 15 wt % Co/Al2O3 was found not to change the size of Co3O4 particles but to accelerate the reduction of smaller Co3O4 particles, resulting in smaller average size of supported Co0 particles in the reduced catalysts.80 The addition of Pt resulted in a significant increase in the FT reaction rate (cobalt–time yield), but some decreases in C5+ selectivity and increases in CH4 selectivity were detected.80

Rhenium is another widely used promoter for Co catalysts. Early studies suggested that Re increased the Co dispersion on TiO2 by preventing agglomeration of CoOx particles during calcination treatments and/or oxidative regenerations.10 Re/TiO2 itself was approximately two orders of magnitude less active than Co/TiO2 for FT synthesis under the same reaction conditions (T=494 K, P=0.56 MPa, H2/CO=2).57 Whereas some groups reported no significant effect of Re on the activity and selectivity of Co catalysts,57 Storsæter et al.81 observed that modification of Co/Al2O3, Co/SiO2 and Co/TiO2 by Re could significantly increase the hydrocarbon formation rate per gram of catalyst and slightly enhance the selectivity to C5+ hydrocarbons, in turn, decreasing the selectivity to CH4. Martínez et al.82 showed that the introduction of roughly 1 wt % Re into Co/SBA-15 enhanced the reducibility of Co species and increased the activity, but did not alter the intrinsic activity (TOF) of Co sites. The C5+ and especially C10+ selectivity was enhanced by the presence of Re, and CH4 selectivity was decreased at the same time. Another study also argued that the presence of Re could not raise the TOF.83

Besides noble metal promoters, many metal oxides are known to modify the structure and the catalytic behavior of supported Co catalysts.10, 13 ZrO2 was demonstrated to be a good promoter to improve the CO conversion activity and C5+ selectivity of a Co/SiO2 catalyst.84 The promotion of a 20 wt % Co/SiO2 with ZrO2 increased the FT reaction rate, and the sequentially impregnated Zr–Co/SiO2 catalyst (Zr/Co=0.02–0.28) was the most active.85 The TOF based on H2 uptake also increased for the series of Zr–Co/SiO2 catalysts. It was suggested that ZrO2 may create an active interface with Co, which may be responsible to some extent for the enhancement in Co activity by facilitating CO dissociation. A detailed study further confirmed that the increase of Zr content in Zr–Co/SiO2 catalysts reduced the interaction between Co and SiO2 and caused a somewhat weaker Co–ZrO2 interaction, leading to a rise in the degree of reduction of the catalysts after H2 reduction at 673 K.86 The modification by Zr increased the size of Co crystallites but the agglomeration of Co species became less serious, leading to the homogeneous distribution of Co particles in the modified catalysts. The TOF based on the metal dispersion of the freshly reduced catalysts increased on ZrO2 modification, but this appeared to be related to the improved catalyst stability of the larger Co particles.86 The C5+ selectivity increased significantly at a lower Zr content (Zr/Co=0.015), and a further increase in Zr content decreased the C5+ selectivity again. Similar enhancing effects of ZrO2 were also found for a 10 wt % Co/SiO2 catalyst prepared by a sol–gel method.87

The modifying effects of ZrO2 on the catalytic behavior of Co/Al2O3 were also investigated. Goodwin and co-workers88 demonstrated that the addition of ZrO2 to γ-Al2O3 could enhance the reducibility of Co species (20 wt % Co) by preventing the formation of Co surface aluminate, which could not be observed by XRD but was suggested from Raman spectra. Li and co-workers89 observed the formation of CoAl2O4 crystallites from XRD for a 15 wt % Co/γ-Al2O3 catalyst, and found that the addition of ZrO2 could inhibit the formation of crystalline CoAl2O4, enhancing the reduction of cobalt species. These studies all observed the increases in both CO conversion and C5+ selectivity with increasing Zr content.88, 89 The increase in Zr content was found to raise the ratio of olefin to paraffin in C2–C17,89 indicating the decrease in hydrogenation ability for the ZrO2-modified catalysts. The intrinsic site rate estimated by the SSITKA technique was found to be unchanged after ZrO2 modification.88 The higher surface Co site density and the increased coverage of the reaction intermediates were proposed to contribute to the increase in the reaction rate and the C5+ selectivity.88, 89 The addition of ZrO2 was also found to promote CO conversion and C5+ selectivity (especially C5–C20 selectivity) over a Co/AC catalyst.90 After FT reactions, Co crystallites with both hcp and fcc structures were detected for the 15 wt % Co/AC, and the addition of ZrO2 suppressed the formation of hcp Co. Over the AC support, the size of Co crystallites was found to decrease from roughly 20 to 10 nm on ZrO2 modification. Temperature-programmed surface reaction (TPSR) experiments suggested that the CO dissociation ability was enhanced due to the presence of ZrO2.90

Manganese oxide, an efficient promoter for Fe-based catalysts,7173 has also been studied as a promoter for Co-based catalysts. Weckhuysen and co-workers performed a series of studies to elucidate the functions of Mn in the Mn-promoted Co/TiO2 catalyst. The addition of 2 wt % Mn to a 7.5 wt % Co/TiO2 catalyst increased the CO conversions and the cobalt–time yields at different pressures (0.1–1.8 MPa).91 The presence of MnOx also led to a rise in C5+ selectivity and a decrease in CH4 selectivity. The preparation procedure was claimed to be key to obtaining such enhancing effects.92 The preparation of Co/TiO2 by a homogeneous deposition–precipitation (HDP) method followed by incipient wetness impregnation (IWI) to load Mn could produce catalysts with associated Co and Mn. STEM-EELS, combined with XPS and EXAFS characterization, suggested a clear Co–Mn interaction in the calcined catalyst, in which a solid-solution spinel compound (MnxCo3−xO4) may have formed.91, 92 This interaction suppressed the reduction of cobalt species, as indicated by in situ XAS studies.93 After reduction by H2, this solid solution was partially transformed to Co0, and MnO migrated to TiO2, but a certain Co–Mn interaction still remained, since MnO clusters were detected in the vicinity of the Co0 particles. This interaction was proposed to affect the electronic state of cobalt, forming Coδ+ and leading to the lowered hydrogenation ability, and thus decreasing the selectivity to CH4. It was also suggested that the MnO in the vicinity of Co0 may act as new sites for the insertion of CO into the grown alkyl chains, forming CxHyOz intermediates, which were subsequently hydrogenated over the neighboring Co0.91 The ratio of olefin to paraffin in C2–C8 also increased gradually with increasing Mn content in Mn–Co/TiO2 catalysts, further confirming that the addition of MnO decreased the hydrogenation ability of Co species.94 The decrease in electron density on Co particles via direct electronic interaction between Co and MnO was confirmed from diffuse reflectance infrared spectroscopic studies.94 By adopting a strong electrostatic adsorption (SEA) technique, Weckhuysen and co-workers95 succeeded in preparing Mn-promoted Co/TiO2 catalysts with stronger Co–Mn interactions. It was proposed that MnO4 anions were selectively deposited onto the supported Co3O4 [point of zero charge (PZC)=8] but not onto the TiO2 (PZC=3.7) by adjusting the solution of pH. The addition of a small amount of Mn (0.03 wt %) onto 10 wt % Co/TiO2 by this technique could enhance the reduction of cobalt species, resulting in a higher cobalt–time yield. However, higher loadings of Mn hindered the reduction of cobalt species and the uptake of H2 because Mn species may have preferentially covered the Co surfaces. In spite of the decreased activity, a higher content of Mn increased the selectivity to C5+ and decreased that to CH4 (Figure 3). The ratio of olefin to paraffin also increased significantly with Mn contents.

thumbnail image

Figure 3. Effect of manganese content on A) product selectivity and B) the molar ratio of olefin to paraffin for C2, C4 and C6 hydrocarbons over Mn- promoted Co/TiO2 catalysts (Co loading=10 wt %).95 Reaction conditions: T=493 K, P=0.1 MPa, H2/CO=2, Wcat=50 mg, F=12 mL min−1.

Download figure to PowerPoint

Rare earth oxides have also been employed as promoters for Co-based catalysts. The addition of La2O3 to 20 wt % Co/SiO2 (La/Co=0–0.75) was found to increase the cobalt–time yield and decrease the selectivities to CH4 and C2–C4 hydrocarbons under the reaction conditions of T=493 K, P=0.1 MPa, and H2/CO=2.96 With increasing La/Co ratio from 0 to 0.1, the α value increased remarkably, from 0.57 to 0.70, and the ratio of olefin to paraffin in C2–C5 hydrocarbons increased from roughly 1 to 5 at steady states. The SSITKA studies suggested that the presence of La2O3 did not change the intrinsic activity of the Co species, but likely increased the concentration of active sites or active intermediates. Characterizations showed that the presence of La3+ in the aqueous solution for catalyst preparation appeared to moderate the strong Co–support interactions, leading to better reducible cobalt oxides and to a larger number of exposed Co0 atoms.97 The impregnation procedure was found to be important for La2O3-promoted Co/γ-Al2O3 catalysts.98 The impregnation of Co precursor, first onto γ-Al2O3 followed by La impregnation (La/Al=0.013–0.078) had little effect on the catalytic activity and selectivity, and there was little evidence of La–Co interactions in this series of catalysts. For the catalysts prepared by impregnation of La precursors onto γ-Al2O3 followed by Co impregnation, the selectivity to higher hydrocarbons and olefins was increased with increasing La content in the region of La/Al≤0.026. However, a high ratio of La/Al caused significant amount of amorphous La–Co mixed oxides, which were difficult to reduce, and decreased the activity significantly. The enhancement in the selectivity to higher hydrocarbons owing to the modification of Co/γ-Al2O3 by La2O3 was also observed by Vada et al.99 A small amounts of La2O3 (0.7–1.7 wt %) could increase the CO conversion and C5+ selectivity of a 15 wt % Co/AC under reaction conditions of T=503 K, P=2.5MPa and H2/CO=2.100 However, too high a La2O3 loading rather decreased the C5+ selectivity and increased the CH4 selectivity. The addition of La2O3 was found to decrease the reducibility of cobalt species but to enhance their dispersion. The strength of CO chemisorption was increased by La2O3 modification.

The introduction of CeO2 (4.5–38.2 wt %) into 25 wt % Co/SiO2 was found to change the product selectivity, although the CO conversion and TOF were not significantly changed.101 The distribution of hydrocarbons in the C5+ fraction was modified after the addition of CeO2; the selectivity to C5–C13 became higher (ca. 33 % at its highest) whereas that to C22+ (wax) decreased. However, the selectivity to CH4 and C2–C4 hydrocarbons increased significantly in the presence of CeO2 modifier. It was proposed that the presence of CeO2 accelerated the hydrogenation ability. Moreover, the hydrogenation of the formate species formed on partially reduced CeO2−x may also have contributed to the increased CH4 selectivity. Shen and co-workers102, 103 reported quite different modifying effects of CeO2 in CeOx–Co/SiO2 (10.5–15 wt % Co; Ce/Co=0.2–0.6) catalysts prepared by a co-impregnation method. The addition of CeO2 to Co/SiO2 was found to increase CO conversion and the selectivity to heavier hydrocarbons. The presence of CeO2 decreased the percentage of gasoline (C5–C10) fraction and increased those to middle distillates (C11–C16) and C16+ (Figure 4).102 The presence of CeO2 slightly inhibited the reducibility of cobalt species but enhanced Co dispersion. The TPSR studies of H2 with the chemisorbed CO suggest that the modification by CeO2 accelerates the dissociation of CO on catalyst surface and increases the concentration of active intermediates for chain growth.103

thumbnail image

Figure 4. Change of distribution of C5+ hydrocarbons after the addition of CeOx into a Co/SiO2 catalyst. Reaction conditions: T=483 K, P=1.2 MPa, H2/CO=2.

Download figure to PowerPoint

Other metal oxide promoters, such as MgO, Al2O3, and TiO2, are known to enhance the catalytic performances of Co/SiO2 catalysts.104106 The addition of MgO to Co/SiO2 was found to increase the CO conversion to some extent and decrease the selectivity of CO2 significantly.104 The addition of Al2O3 or TiO2 to Co/SiO2 could enhance the dispersion of Co species although the reducibility was somewhat decreased.105, 106 CO hydrogenation activity in the slurry-phase reactions was enhanced by the addition of Al2O3 or TiO2 to Co/SiO2. In the case of Al2O3 modification, CH4 selectivity also decreased slightly. Interestingly, a reversible promoting effect also existed; the addition of a small amount of SiO2 (5–20 wt %) to Co/Al2O3 also increased CO conversion and decreased CH4 selectivity.107 The reducibility of Co species was enhanced in this case. These observations further confirm that the balance between reducibility and dispersion of Co species is crucial for FT synthesis.

The effects of promoters typically used for Co-based FT catalysts are summarized in Table 2. Besides the identity of these promoters, many studies have shown that the introduction method or procedure of the promoters (preparation technique) is a key for obtaining the optimum promoting effect.92, 95, 98, 101, 102 The technique capable of creating direct interactions between the promoter and the cobalt species has been proven to be crucial for MnOx-modified Co/TiO2 catalysts.9195 Appropriate location of the promoters over the supported Co catalysts should be considered in the future design of effective catalysts.

Table 2. Effects of typical promoters on catalytic behaviors of Co-based FT catalysts.
PromoterPossible functionsReferences
Noble metals such as Ru and Re1) Increasing the CO conversion activity by either enhancing the reduction of Co precursors or facilitating the dispersion of Co species by preventing the aggregation of CoOx or Co0 during calcination or reduction. 2) For Ru or Re, increasing the selectivity to C5+ and decreasing that to CH4, owing to the increased Co site density; for Pt and Pd (with higher hydrogenation ability) decreasing C5+ selectivity. 3) Increasing the bridge-chemisorbed CO and thus enhancing the CO dissociation ability. 4) Retarding deactivation by inhibiting the reoxidation of Co0 or carbon deposition.1013, 57, 7783
ZrO21) Increasing the CO conversion activity by enhancing the Co reducibility through replacing stronger metal–support interactions with the weaker Co-ZrO2 interaction. 2) Enhancing the C5+ selectivity because of the increased Co site density or the lowered hydrogenation ability. 3) Increasing the olefin/paraffin ratio due to the decreased hydrogenation ability. 4) Increasing Co0 size over SiO2- or Al2O3-supported catalysts but decreasing Co0 size over AC-supported catalysts; Enhancing CO dissociation ability over the AC-supported catalyst. (5) Increasing the concentration of active intermediates but not changing the intrinsic Co activity in many cases.8490
MnOx1) Increasing the selectivity to C5+ and decreasing that to CH4, possibly by decreasing the hydrogenation ability through electronic modification of Co species to form some Coδ+; the existence of Co–Mn interactions (e.g., forming MnxCo3−xO4 solid solution in catalyst precursor) is essential. 2) Increasing the ratio of olefin to paraffin because of the decreased hydrogenation ability. 3) Possibly increasing the CO conversion activity if the Mn content is low.9195
Rare earth oxides such as La2O3 and CeO21) Increasing the C5+ selectivity and the olefin/paraffin ratio of in the case of La2O3 by modifying the Co–support interactions; 2) The effect of CeO2 on product selectivity is complicated and depends on catalyst preparation technique. An appropriate preparation method may result in higher selectivity to middle distillates. 3) Raising CO conversion activity with a appropriate content and preparation technique.96103

3. Utilization of Nanoporous Materials for FT Synthesis

  1. Top of page
  2. Abstract
  3. Introduction
  4. 1. Overview of Catalysts, Reactors, Reaction Mechanisms, and Product Distributions for FT Synthesis
  5. 2. Key Factors Determining the Selectivity and Activity for FT Synthesis
  6. 3. Utilization of Nanoporous Materials for FT Synthesis
  7. 4. Utilization of Zeolites for FT Synthesis
  8. 5. Utilization of Novel Carbon Materials for FT Synthesis and the Size Effect of Metal Particles
  9. Summary and Outlook
  10. Acknowledgements
  11. Biographical Information
  12. Biographical Information
  13. Biographical Information

3.1. FT synthesis over catalysts supported on mesoporous materials

Ordered mesoporous materials, typified by MCM-41 and SBA-15, have shown many unique characteristics as hosts for the design of structure-defined catalysts with active components as encapsulated metal nanoparticles, single-site metal ions or included metal complexes.108115 The mesoporous materials typically possess high surface areas (>500 m2 g−1) and uniform porous channels with controllable pore diameters (2–30 nm) and pore lengths,114 which are suitable for the preparation of suitably dispersed metal particles. Because active metal (Fe, Co or Ru) nanoparticles with suitable dispersion can lead to better catalytic performances in FT synthesis, the mesoporous materials, particularly mesoporous silicas, have attracted much attention as promising catalyst supports for FT synthesis.116 Moreover, the mesopores in which the active metal particles are located may function as a nanoreactor to control the chain length, either by shape selectivity or by enhancing the readsorption of α-olefin intermediates. In other words, the nanospaces of mesoporous materials can be expected to regulate the product selectivity.

Wang et al.117 prepared SBA-15 with different pore sizes (3.6–12 nm) and investigated their potential as supports for Co catalysts for FT synthesis. Khodakov et al.118 performed a detailed study on the catalytic behaviors of Co catalysts supported on MCM-41 and SBA-15, and found that the Co catalysts supported on mesoporous silicas with mean pore sizes greater than 3 nm showed higher CO conversion rates and C5+ selectivities (60–70 %). Otherwise, lower CO conversions and higher CH4 selectivities (>20 %) were obtained using mesoporous silicas with smaller pores as the supports. The lower reducibility of Co species in the smaller silica pores was believed to be responsible for the lower performances.

Thus, to prepare a mesoporous material-supported catalyst with enhanced reducibility of active metal precursors is important. Several research groups have clarified that the precursor of cobalt is an important factor affecting its reducibility because the nature of cobalt precursor would strongly affect its interaction with the support.77, 119123 Ohtsuka et al.121 reported that the impregnation of SBA-15 possessing an average pore diameter of 8.6 nm with cobalt acetate (20 wt % Co), followed by air calcination at 773 K and H2 reduction at 673 K, resulted in an inactive catalyst for FT synthesis at 503 K and 2MPa because of the very low reducibility of Co species. However, the use of cobalt nitrate or the mixture of nitrate and acetate provided much higher CO conversions. The 20 wt % Co/SBA-15 prepared using nitrate or the mixture of nitrate and acetate precursor provided higher selectivity (ca. 30 %) and higher space-time yield (STY=260–270 gCequation image h−1) to C10–C20 hydrocarbons (diesel fuel).121, 122 The Co/SBA-15 was further demonstrated to exhibit higher C10–C20 selectivity (ca. 40 %) than the Co catalysts supported on other mesoporous materials or microporous materials.124 As compared to Co/Cab-O-Sil (a nonporous silica), mesoporous material-supported catalysts gave lower C21+ selectivity and higher C10–C20 selectivity (Figure 5), implying the limitation of chain growth in the nanospace of SBA-15.124

thumbnail image

Figure 5. Catalytic behavior of Co catalysts supported on mesoporous materials and on Cab-O-Sil (Co loading=5 wt %). Reaction conditions: T=523 K, P=2 MPa, H2/CO=2, Wcat=0.80 g, F=20 mL min−1.

Download figure to PowerPoint

The addition of noble metal promoters is an effective strategy to enhance the reducibility of Co species located in mesoporous channels. Several groups have shown that the presence of Ru or Re promoter can enhance the reducibility of Co species over MCM-41 or SBA-15 and can significantly increase CO conversions and/or C5+ selectivity.125128 The 0.5 wt % Ru–14 wt % Co/MCM-41 provided a higher CO conversion rate than the corresponding SiO2-supported catalyst, although the reducibility of the former catalyst was still lower than the latter (Table 3), showing that mesoporous materials are promising for FT synthesis.125 More significant promoting effects of Ru on Co reducibility, CO conversion and C5+ selectivity were found for the cobalt catalyst located in smaller mesopores (3–4 nm) than that in larger mesopores.128 It was demonstrated that, under the same reaction conditions, the Co/SBA-15 provided higher CO conversions than the conventional Co/SiO2.123 The presence of Re significantly enhanced the degree of reduction of Co species in the 20 wt % Co/SBA-15 catalysts, resulting in a further rise in CO conversion and C5+ selectivity (Table 3).123 When compared to a commercial silica with an average pore diameter of 33 nm, the Co catalyst loaded on a SBA-15 with a pore diameter of roughly 9 nm was demonstrated to afford a higher cobalt–time yield.129

Table 3. Comparison of catalytic behaviors of Co catalysts supported on mesoporous silica with those supported on conventional SiO2.
CatalystCo reducibility [%]Co0 dispersion [%]CO conv. [%]Selectivity [%]Selectivity to a certain range of productsReference
    CH4C5+  
  1. [a] Reaction conditions: T=250 °C, P=0.2 MPa, H2/CO=2; [b] not available; [c] reaction conditions: T=220 °C, P=0.2 MPa, H2/CO=2; [d] reaction conditions: T=220 °C, P=2 MPa, H2/CO=2; [e] reaction conditions: T=190 °C, P=0.1 MPa, H2/CO=2; [f] cobalt-site yield (10−4 s−1); [g] reaction conditions: T=483 K, P=2MPa, H2/CO=2; [h] reaction conditions: T=503 K, P=2MPa, H2/CO=2.

5 wt % Co/SBA-15[a]69n.a.[b]9214ca. 76C10–C20: 40 %124
5 wt % Co/SiO2[a]n.a.[b]n.a.[b]4112ca. 83C10–C20: 31 %124
0.5 % Ru-14 % Co/MCM-41[c]384.87.819.434.4 125
0.5 % Ru-14 % Co/SiO2[c]587.71120.431.3 125
20 % Co/SiO2[d]899.117.717.963.0 123
20 % Co/SBA-15[d]6211.223.119.564.7 123
1 % Re-20 % Co/SBA-15[d]8313.543.014.174.2 123
25 % Co/SBA-15[e]96n.a.[b]2.56[f]15.268.0 129
25 % Co/SiO2[e]100n.a.[b]1.87[f]16.562.4 129
15 % Co/HMS[g]45n.a.[b]88.27.584.2C5–C18: 69.2 %137
15 % Co/SiO2[h]91n.a.[b]66.26.486.1C5–C18: 59.2 %137

The introduction of other modifiers may also enhance the reduction and/or dispersion of Co species located in mesoporous channels. Okabe et al.130 reported that the incorporation of some heteroatoms (e.g., Al, Ti, Zr and V) into the framework of mesoporous silica by direct hydrothermal synthesis before loading Co and a noble metal promoter (Ir) could significantly enhance the C5+ selectivity, probably due to the increased Co dispersion. For the slurry-phase reaction under the reaction conditions of T=503 K, P=1 MPa and H2/CO ratio=2, the introduction of these heteroatoms increased the C5+ selectivity from approximately 60 % to 82–91 %. The presence of Zr in the framework of mesoporous silica increased the fraction of fcc Co in the used catalyst, and this may also contribute to the increase in C5+ selectivity.131 A Zr-grafted SBA-15 was demonstrated to be a superior support for Co catalysts.132 The Co species located on ZrO2 grafted on SBA-15 were more reducible than those located on the silica wall of SBA-15.

Another interesting strategy for improving the reducibility or dispersion of cobalt species is the hydrophobic modification of mesoporous silica.133 Either the water vapor formed in the confined mesopores during the H2 reduction or the strong interaction between cobalt precursors and the silanol (Si[BOND]OH) groups may retard the reduction of cobalt species.125, 133 Eyring and co-workers133 demonstrated that the silylation of a SBA-15 (mean pore diameter≈8 nm) with hexamethyldisilazane (HMDS), followed by impregnation of (cyclooctadiene)(cyclooctenyl)cobalt [Co(C8H12)(C8H13)], air calcination at 823 K and H2 reduction at 723 K resulted in a Co/SBA-15 catalyst with higher degree of Co reduction (45–50 %) as compared to that without silylation (15 %). The 6 wt % Co/SBA-15 catalyst prepared by this procedure exhibited significantly higher C5+ selectivities and CO conversions (Figure 6).133The selectivity to CH4 and C2–C4 hydrocarbons decreased significantly by using the silylated SBA-15. Interestingly, most of the C5+ hydrocarbons were in the C5–C10 range, and the C5–C10 selectivity reached approximately 60 %.

thumbnail image

Figure 6. Effect of silylation of SBA-15 by hexamethyldisilazane (HMDS) on the catalytic performance of Co/SBA-15 (Co loading=6 wt %). Reaction conditions: T=538 K, P=0.69 MPa, H2/CO=2, GHSV=2100 h−1.

Download figure to PowerPoint

Other kinds of mesoporous silicas have also been applied to FT synthesis. MCM-48, which possesses a cubic 3D porous structure, is expected to favor mass transfer. A 10 wt % Co/MCM-48 provided C5+ selectivity around 74 % at a CO conversion around 27 % under the reaction conditions of T=503 K, P=1 MPa, and H2/CO=2.134 A hexagonal mesoporous silica (HMS) with a pore diameter (ca. 3 nm) similar to MCM-41 but less-ordered wormhole-like pore structures and smaller domain sizes with shorter channels and larger textural mesoporosity, is also expected to be beneficial to the diffusions of products and reactants.135 When compared to the 15 wt % Co/MCM-41, 15 wt % Co/HMS gave higher CO conversion and higher C5+ selectivity.136 The weight ratio of wax to oil was also higher over the Co/HMS catalyst, reaching 8.3 at 503 K (P=2MPa and H2/CO=2). The modification of HMS with ZrO2 before loading Co further enhanced the CO conversion and C5+ selectivity. The superior performance of the Co/HMS was further confirmed by a comparison with Co/SiO2, and it was shown that the 15 wt % Co/HMS provided a higher CO conversion even at a lower temperature than the 15 wt % Co/SiO2 although the former catalyst possessed a lower Co reducibility (Table 3).137 The higher Co dispersion in Co/HMS was believed to account for its higher activity. Although the C5+ selectivity over Co/HMS was not higher than that over the Co/SiO2, the former catalyst afforded higher C5–C18 selectivity and lower C19+ selectivity (Figure 7).137 A latter study showed that, as compared to the Co/SiO2 catalyst, the Co/HMS catalysts showed lower CO conversions and lower C5+ selectivity possibly because of the lower Co reducibility.138 However, the fraction of C11–C18 in C5+ hydrocarbons became higher over the Co/HMS. These studies imply that the mesoporous structure of HMS may hinder the chain growth, tailoring the product distribution toward the diesel fraction.

thumbnail image

Figure 7. Comparison of product selectivities obtained over the Co/HMS and Co/SiO2 (Co loading=15 wt %). Reaction conditions: T=483 K (for Co/HMS) or 503 K (for Co/SiO2), P=2.0 MPa, H2/CO=2, GHSV=500 h−1.

Download figure to PowerPoint

There are only a few reports on the utilization of mesoporous materials as supports of Fe-based catalysts for FT synthesis. A 5 wt % Fe/MCM-41 catalyst, which contained 24 % Fe0 (average size≈1.3 nm) and 76 % Fe2+ after H2 reduction at 698 K for 26 h, mainly produced CH4 (selectivity>80 %) in CO hydrogenation possibly because of the lower reducibility and the very small size of Fe0.139 The comparison of catalytic behaviors of Fe catalysts supported on various kinds of mesoporous and microporous materials revealed that Fe/MCM-41 and Fe/MCM-48 with a Fe loading of 10 wt % were almost inactive for FT synthesis (CO conversion<5 %) under the reaction conditions of T=543 K, P=2 MPa, and H2/CO=2.140 Fe/SBA-15 was active under the same reaction conditions, providing a CO conversion of 17 %. However, the catalytic performance of Fe/SBA-15 was very similar to that of the Fe/SiO2. An interesting enhancing effect of the incorporation of small amount of Al into SBA-15 was reported by Eyring and co-workers.141 The presence of small amount of Al (Al/Si=0.01–0.033) did not affect the pore structure of SBA-15, but the reducibility of supported iron oxides was significantly enhanced. The iron species in sample with an Al/Si ratio of 0.01 was the most reducible. The Fe/SBA-15 without Al after H2 reduction pretreatment at 773 K for 10 h required a long reduction period (ca. 25 h) to reach a significant activity level (CO conversion, ca. 20 %). The incorporation of Al into SBA-15 reduced the induction period. CO conversion increased to a higher steady-state value (ca. 35 %) after about 7 h of reaction for the Fe/Al-SBA-15 with an Al/Si ratio of 0.01. Significant changes to the product selectivity also took place after the inclusion of Al. The incorporation of Al with an Al/Si ratio of 0.01 increased the C11+ selectivity from ca. 50 % to over 70 % (Figure 8).141 Further increases in Al content were detrimental to the production of heavier hydrocarbons as well as the CO conversion. EXAFS and Mössbauer spectroscopic studies suggested that the catalyst after FT reactions contained Fe0, nonmagnetic iron oxides, and iron carbides, and the incorporation of Al in the framework of SBA-15 accelerated the formation of iron carbides. As mentioned earlier, the beneficial effects of heteroatoms (including Al) in the framework of mesoporous silicas were also observed for Co-based FT catalysts.130, 131

thumbnail image

Figure 8. Effect of the presence of Al in SBA-15 on the product selectivity of Fe/SBA-15 catalysts (Fe loading=20 wt %). Reaction conditions: P=0.69MPa, H2/CO=2.

Download figure to PowerPoint

3.2. Effect of pore size of support on FT synthesis

The catalytic performances of FT catalysts, especially the product selectivities, are determined by a complex interplay among diffusion, reactions at active sites, and secondary reactions. It can be expected that the size of support pores may affect a) the reducibility and the dispersion of cobalt species, b) the diffusion of products and reactants, and c) the probability of secondary reactions through α-olefin readsorption in a confined nanospace. These factors are all crucial in determining FT catalytic behavior. Thus, discussion on the sole effect of the pore size (i.e., the dimension of nanospaces inside which active metals are located) is quite difficult. The use of mesoporous materials with narrow pore size distributions as catalyst supports has provided some new insights into the pore size effects. In this section, some recent research in this area is summarized.

Some studies have contributed to clarifying the pore size effect using conventional amorphous SiO2 or γ-Al2O3. A comparison of catalytic performances of 20 wt % Co catalysts supported on SiO2 with average pore diameters of 2, 4, 6, 10, or 15 nm demonstrated that both the C5+ selectivity and CO conversion were the highest over the Co/SiO2 with an average pore size of 10 nm.142 The degree of reduction and the size of Co also changed with the pore size of SiO2. Aside from the sample with the smallest pore size (2 nm), the degree of reduction and the Co particle size increased with the pore size. A similar phenomenon was observed for the Co/SiO2 catalysts with the pore size of SiO2 changing from 2.4 to 15.8 nm; the catalysts with mean pore sizes of 6–10 nm displayed higher FT activity and higher C5+ selectivity.143 A smaller pore size led to lower reducibility of Co species, whereas a larger pore size caused the formation of larger Co particles. Thus, the balance between the reducibility and the particle size of Co resulted in the optimum activity and selectivity for catalysts with a medium pore size.

Borg et al.144 performed a detailed study using 13 γ-Al2O3 supports with average pore diameters varying from 5.9 to 26.7 nm as supports for 20 wt % Co catalysts promoted with 0.5 wt % Re. The degree of reduction evaluated from temperature-programmed reduction by H2 (H2-TPR) and the Co particle size estimated from H2 chemisorption both increased with the pore size of γ-Al2O3 and varied in the ranges of 80–96 % and 10.4–14.9 nm, respectively (Figure 9 A). C5+ selectivities over these Co/γ-Al2O3 catalysts were compared at an equal conversion of CO (50 %). The C5+ selectivity increased monotonically with the pore size of γ-Al2O3 and, at the same time, the CH4 and C2–C4 selectivities decreased accordingly (Figure 9 B). This tendency is different from that for the Co/SiO2 described above, whereupon the maximum C5+ selectivity was for the catalyst with a medium pore size.142, 143 It was proposed that not only the higher extent of α-olefin readsorption in wider pores but also the larger Co particles contributed to the higher C5+ selectivity over the catalyst with a larger pore size.144 The effect of metal particle size on product selectivity will be discussed in section 5.2.

thumbnail image

Figure 9. Effect of the pore size of γ-Al2O3 on A) the degree of reduction (○) and particle size of Co (▴) and B) the selectivity towards C5+ (•) and CH4 (□) for the Co/γ-Al2O3 catalysts (Co loading=20 wt %). Reaction conditions: T=483 K, P=2 MPa, H2/CO=2.

Download figure to PowerPoint

It can be expected that mesoporous materials with ordered porous channels and changeable pore diameters are more suitable for the elucidation of pore size effects. MCM-41- and SBA-15-type mesoporous silicas with pore diameters ranging from 2 to 9.1 nm were synthesized and applied as Co catalyst supports (5 wt % Co loading).145 The size of Co3O4 was well controlled by the pore size of mesoporous silicas, and an increase in the pore size led to increases in both Co3O4 particle size and the reducibility. Under the reaction conditions of T=463 K, P=0.1 MPa, and H2/CO=2, the catalysts with pore sizes smaller than 3 nm exhibited significantly lower CO conversions and lower C5+ selectivities, mainly due to the lower reducibility.118 With changing the pore diameters of SBA-15 from 3.7 to 16 nm, the degree of reduction of the 30 wt % Co/SBA-15 catalysts were found to slightly rise from 50.0 to 57.5 %.146 At the same time, the size of Co particles increased. Under the reaction conditions of T=383 K, P=2 MPa and H2/CO=2, with increasing pore size of SBA-15, CO conversions passed through a maximum at a pore size of about 9 nm. The C5+ selectivity increased from about 75 % to 87.5 % as the pore size of SBA-15 rose to 9 nm, and then did not change significantly.146 Since the change in the degree of reduction (50–57.5 %) was not very significant, it was proposed that the size of Co particles mainly determined the C5+ selectivity. Similar to the result of Borg et al.,144 the larger cobalt particles favored the formation of C5+ hydrocarbons. An increase in the intensity of the IR band ascribed to bridge-adsorbed CO was observed with increasing pore size of SBA-15 (or Co particle size).146 Thus, the CO dissociation ability become stronger over larger Co particles. A comparison of 20 wt % Co catalysts supported on SBA-15, Al-MCM-41 and INT-MM1 with pore diameters of 4.9, 3.2, and 2.6 nm, respectively also showed that the Co/SBA-15 with a wider pore and larger Co particles exhibit the highest C5+ selectivity (79.6 %) and the lowest CH4 selectivity (15.6 %).147

Other mesoporous metal oxides with different pore sizes have also been used as supports for Co catalysts to elucidate the pore size effect. The C5+ selectivity was higher over the Co catalyst supported on mesoporous Al2O3 with wider pores.148 Mesoporous ZrO2 (meso-ZrO2) samples with pore sizes ranging from 2.9 to 12.6 nm were synthesized by using zirconium(IV) n-propoxide as the Zr source and P127, a triblock copolymer (EO20PO70EO20; EO=ethylene oxide, PO=propylene oxide), as the template, and were applied to FT synthesis.149 Similarly, the size of Co3O4 before reduction was controlled by the pore size of meso-ZrO2. The Co reducibility increased from about 60 % to 94 % with increasing the pore size from 2.9 to 12.6 nm. The CO conversion and C5+ selectivity increased significantly with the pore size for the 10 wt % Co/meso-ZrO2 catalysts (Figure 10). Particularly, the catalyst with the largest pore size (12.6 nm) could provide a high selectivity (ca. 32 %) to the diesel fuel (C12–C18). The bridged-type CO species were observed over the catalysts with larger pore size of meso-ZrO2, indicating that the dissociation of CO was favored over these samples. The increased Co reducibility, the larger Co particles, and the interface between Co and ZrO2 may all contribute to the high C5+ and C12–C18 selectivity of the catalyst with a pore size of roughly 12 nm.

thumbnail image

Figure 10. Effect of the pore size of mesoporous ZrO2 on catalytic performances over the Co/ZrO2 catalysts (Co loading=10 wt %). Reaction conditions: T=503 K, P=2 MPa, H2/CO=2, GHSV=1000 h−1.

Download figure to PowerPoint

Only very few studies have contributed to elucidating the effect of pore size of support on the catalytic behavior of Fe- and Ru-based catalysts. Okabe et al150 investigated 10 wt % Ru/SiO2 catalysts with average pore sizes changing from 4 to 8.4 nm for FT synthesis in slurry-phase reactions. Whereas CO conversion remained almost unchanged, the selectivity to CH4 decreased and that to C5+ increased with increasing the pore size. This trend was explained mainly by the more facile diffusion within the catalysts with larger pores. Li and co-workers151 succeeded in preparing Ru nanoparticles with similar sizes (2.8–3.6 nm) confined in the pores of SBA-15 with different pore diameters (3.6–13.1 nm) by using an immobilization method. This allowed them to analyze the pore effect more directly. On changing the pore size from 3.6 to 7.3 nm, the ruthenium–time yield increased, whereas the selectivities to both C5+ and CH4 did not change significantly. However, a further increase in the pore size to 13.1 nm decreased the selectivity to C5+ and increased those to CH4 and C2–C4 (Figure 11). Moreover, they confirmed that the Ru particles with a similar mean size (4.1 nm) but located outside the mesopores exhibited significantly lower C5+ selectivity although the CO conversion activity became higher. The products over the catalyst with confined Ru particles did not obey the ASF distribution. The confinement may enhance the repeated readsorption of the α-olefins, leading to the higher C5+ selectivity in the catalysts with mesopores of sizes 7.3 nm and below.

thumbnail image

Figure 11. Effect of the pore size of SBA-15 on catalytic performances over the Ru/SBA-15 catalysts (Ru loading≈4 wt %). Reaction conditions: T=508 K, P=1 MPa, H2/CO=2.

Download figure to PowerPoint

In short, many of the studies reported to date appear to suggest that the larger pore size results in better C5+ selectivities as well as higher CO conversion activities, although a few suggest that there exits an optimum pore size. However, these observations are generally the result of complex interplay among many factors induced by changing the pore size of the support. These factors include not only the confinement effect, which leads to changes in the readsorption probability of α-olefins in the confined space and the diffusion situation, but also the effect of changes in the site density, such as the changes in reducibility and the particle size of Co. The latter effect may be more significant in some systems. Only limited studies have provided information about the sole effect of pore size. More insights are expected if we are able to prepare catalysts with a fixed reducibility and particle size of active metals (particularly Co), which are located in mesoporous materials with changeable pore sizes.

3.3. Utilization of bimodal porous supports for FT synthesis

Although catalysts with wider pores are beneficial to the diffusion behavior, which is particularly crucial for the slurry-phase reactions, the dispersion of supported metals may be lower because of the smaller surface areas of the larger-pore supports. A bimodal porous support containing larger pores for rapid transportation of products and reactants and smaller pores for dispersion of metals would be promising for FT synthesis.

Tsubaki and co-workers synthesized bimodal porous supports containing a larger-pore SiO2 (Q-50; ca. 45 nm) and a smaller-pore SiO2 (3–6 nm), Al2O3,and ZrO2 by impregnating Q-50 with SiO2 or ZrO2 sol or Al(NO3)3 as a polyethylene glycol solution, and found that these bimodal supports possess distinct features for FT synthesis in the slurry phase.152156 As compared to the Co/Q-50 catalyst, the Co catalysts supported on bimodal supports, SiO2–Q-50, ZrO2–Q-50, or Al2O3–Q-50, possess smaller Co size and higher Co dispersion. These bimodal porous material-supported Co catalysts displayed significantly higher CO conversions, although the degrees of reduction of Co were lower than that for the Co/Q-50 catalyst (Table 4). The TOFs for CO conversion, calculated based on surface Co atoms, were also significantly higher over the ZrO2-Q-50- and Al2O3-Q-50-supported catalysts. CO conversions and TOFs over the bimodal catalysts were also much higher than those over the Co catalyst supported on SiO2 with only smaller pores (Q-3; ca. 3 nm). The selectivities to CH4 and CO2 over the bimodal catalysts were significantly lower than those over the Co/Q-3 catalyst but slightly higher than those over the Co/Q-50 catalyst. The C5+ selectivity was the highest over the Co/ZrO2–Q-50 catalyst. The higher TOF over Co/ZrO2–Q-50 also suggests the possibility for chemical promotional effects by using ZrO2 besides the spatial effect of bimodal catalysts.

Table 4. Features of Co catalysts supported on bimodal porous supports (Co loading=10 wt %).
CatalystDegree of reduction [%]Co size [nm]Co dispersion [%]CO conversion [%]TOF [10−2 s−1]Selectivity [%]α
      CH4CO2 
  1. [a] From Ref. 154; reaction conditions: T=513 K, P=1.0 MPa, weight/flow-rate (W/F)=10 g h mol−1, H2/CO=2; [b] from Ref. 156; reaction conditions: T=513 K, P=1.0 MPa, W/F=10 g h mol−1. H2/CO=2.

Co/Q-50[a]ca. 9935 (XRD), 37 (TEM)2.5173.517.13.20.86
Co/Q-50[b]ca. 9935 (XRD), 37 (TEM)2.513.56.17.12.10.86
Co/Q-3[a]60–624.5 (XRD), 1.4 (TEM)38220.9422.020.50.84
Co/SiO2-Q-50[a]86–8821 (XRD), 23 (TEM)3.6334.3910.12.70.86
Co/ZrO2-Q-50[a]8824 (XRD), 22 (TEM)4.78613.4111.03.20.87
Co/Al2O3-Q-50[b]9015 (XRD), 23 (H2 chemisorption)4.33912.110.21.30.87

The impregnation of Q-50 with a mixed aqueous solution of nitrates of iron, copper, and potassium followed by calcination at 673 K simply produced a bimodal porous catalyst with the smaller pores (6.2 nm) built by the deposited iron, copper, and potassium species. This catalyst showed higher CO conversions than those without the bimodal structure.157 Hollow mesoporous silica spheres (HMSS), which exhibited a bimodal pore distribution with one maximum at approximately 40 nm and the other at around 2.7 nm, were also promising for FT synthesis.158 The 30 wt % Co/HMSS showed a high C5+ selectivity (ca. 94 %) at a CO conversion of about 84 % under the conditions of T=493 K, P=2.0MPa, H2/CO=2, and GHSV=1000 h−1, whereas the C5+ selectivity was about 85 % at a CO conversion of about 70 % over the 30 wt % Co/MCM-41 under the same reaction conditions. The C5–C18 selectivity over Co/HMSS (ca. 53 %) was also higher than that over Co/MCM-41 (ca. 41 %).

4. Utilization of Zeolites for FT Synthesis

  1. Top of page
  2. Abstract
  3. Introduction
  4. 1. Overview of Catalysts, Reactors, Reaction Mechanisms, and Product Distributions for FT Synthesis
  5. 2. Key Factors Determining the Selectivity and Activity for FT Synthesis
  6. 3. Utilization of Nanoporous Materials for FT Synthesis
  7. 4. Utilization of Zeolites for FT Synthesis
  8. 5. Utilization of Novel Carbon Materials for FT Synthesis and the Size Effect of Metal Particles
  9. Summary and Outlook
  10. Acknowledgements
  11. Biographical Information
  12. Biographical Information
  13. Biographical Information

4.1. FT synthesis over zeolite-supported catalysts

It is well known that zeolites possess shape-selective features that do not allow the formation of products, intermediates, or transition states larger than the size of the cavities or channels of the zeolites. These features may limit chain growth, leading to the formation of lighter hydrocarbons. On the other hand, confinement in a nanovoid may enhance the readsorption probability and the secondary reactions of α-olefins, leading to long-chain hydrocarbons. Furthermore, the acidity of zeolites may catalyze the secondary cracking, isomerization, and aromatization reactions of the primary FT products, contributing to adjustment of the product distributions.

The hydrogenation of CO over zeolite-supported mono- or bimetallic catalysts has been highlighted in several reviews.116, 159 Faujasite zeolites (X and Y) possess supercages (diameter≈1.3 nm) and have attracted particular attention for the preparation of small metal clusters in a confined environment. Guczi et al.159 found that, under the conditions of T=502 K, P=2.1 MPa, and H2/CO=2, Ru/NaY was very active for CO conversion (86 %), but the main products were CH4 and CO2. On the other hand, Co/NaY exhibited a very low CO conversion. The addition of Ru to Co/NaY, forming a bimetallic catalyst, significantly enhanced CO conversion from 1.1 % to 9 %, due to an enhancement in the degree of reduction of Co species. Moreover, the selectivity to CH4 decreased from 44 % to 26 % and that to C5+ increased from 22 % to 38 %.

The preparation of Co0 nanoparticles confined in the supercages of faujasites is still not very convenient because the reduction of Co2+ exchanged in zeolites is difficult at moderate temperatures (<800 K) due to strong interactions between the metal cations and the anionic zeolite framework. Tang et al.160, 161 succeeded in preparing faujasite-confined Co nanoparticles by the following procedures: a) ion-exchange of Co2+ into the zeolite; b) precipitation of Co2+ with NaOH aqueous solution within the supercages of the zeolite; c) calcination; d) H2 reduction. The introduction of a precipitation step resulted in the conversion of Co2+ to CoOx clusters after calcination. Most of the CoOx clusters, mainly in the CoO state, could be included in the supercages by choosing proper preparation conditions. Some mesopores were also formed after the treatment of zeolites by NaOH, and some CoOx particles might also exist in the formed mesopores. These CoOx clusters showed higher reducibility than the Co2+ located in the ion-exchanging positions by H2. The formed small Co0 nanoparticles were active in FT synthesis, providing higher CO conversions than the larger Co particles outside the supercages. The catalysts with smaller confined Co particles, however, exhibited lower C5+ selectivity and relatively higher CH4 selectivity.161 Metallic Co nanoparticles could also be prepared with narrow size distribution by the reduction of the Co2+-exchanged zeolite with NaBH4 aqueous solution.162 When compared to Co/SiO2 or Co/faujasite, with larger Co particles outside the supercage, the faujasite-confined Co catalysts prepared by this method exhibited lower selectivity to C21+ but higher selectivity to C5–C20. The supercage likely exerts a cut-off effect for heavier hydrocarbons.

In addition to faujasites, MCM-22, a layered zeolite with MWW structure, was also employed as a support for Co catalysts for FT synthesis.163 One distinct feature of zeolite MCM-22 is its high external surface area and low fraction of micropores. The selectivity of a 10 wt % Co/MCM-22 catalyst in FT synthesis depended on the Si/Al ratio and the degree of H+ exchange. The exchange of Na+ with H+ significantly increased CH4 selectivity. The selectivity of CH4 decreased and that of C5+ increased gradually with increasing Si/Al ratio. Under the conditions of T=280 °C, P=1.25 MPa, and H2/CO=2, the Co/Na-MCM-22 (Si/Al=200) showed a lower CH4 selectivity and a higher C5+ selectivity than Co/SiO2, even though the size of Co particles in the former catalyst was smaller.163

Zeolites ITQ-2 and ITQ-6, which were derived by delamination from layered MCM-22 and ferrierite, respectively, have also been applied to FT synthesis.164 The delaminated zeolites are formed by thin zeolite layers (ca. 0.9–2.5 nm) with very high accessible external surface areas (typically >600 m2 h−1), containing irregular mesopores due to the interparticular condensation. These delaminated zeolites possess only little or even no microporosity. The Co species supported over the pure silica ITQ-2 and ITQ-6 (20 wt % Co loading) exhibited relatively higher reducibility (88–89 %), whereas those over MCM-41 were difficult to reduce (38 % reducibility). The Co particles over ITQ-6 (mean size of Co3O4, 9.3 nm) were smaller than those over ITQ-2 (mean size of Co3O4=11.7 nm) and over SiO2 (mean size of Co3O4=10.6 nm). Co/MCM-41 showed a bimodal Co size distribution with maxima at 3 nm and 19 nm (mean size of Co3O4=9.2 nm), suggesting that some Co species are located in mesoporous channels and some are deposited on the external surface. The Co/ITQ-6 provided a higher CO conversion during FT synthesis due to its higher reducibility and higher Co dispersion. The product selectivities over different catalysts were quite different. Under the conditions of T=493 K, P=2 MPa, H2/CO=2 and similar CO conversions (20–24 %), the C5+ selectivities over the ITQ-2 and ITQ-6 supported Co catalysts were approximately 72 %, significantly higher than those over Co/SiO2 (ca. 65 %) and Co/MCM-41 (ca. 45 %).164 The higher C5+ selectivity of the ITQ-supported catalysts was ascribed to the higher concentrations of coordinatively unsaturated Co0 sites over these catalysts. The peculiar nanoporous structure of ITQ zeolites may also contribute to the higher C5+ selectivity.

Among many zeolite-supported Fe catalysts, the Fe/NaX and Fe/NaY showed higher CO conversions and unique product distributions with significantly lower selectivities to CH4 and C2–C4 and higher C5+ selectivity (Table 5).140 The selectivity to C10–C20 was particularly higher over the faujasite-supported Fe catalysts. Use of Li+-exchanged zeolite Y (LiY) as the support further increased the selectivities to C5+ and C10–C20 and decreased that to CH4. C5+ and C10–C20 selectivities of 69 % and 26 % could be attained over 10 wt % Fe/LiY without any other promoters. The formation of CO2, which likely came from the WGS reaction, was also significantly suppressed by using LiY as the support. The role of Li is unclear, and further elucidation of the effect of the cavity of faujasite zeolites is needed. A zeolite K-LTL-supported Fe catalyst was found to show unique behaviors in FT synthesis.165 A CO conversion of around 40 % was obtained over this catalyst with 77 % selectivity to hydrocarbons and 23 % selectivity to CO2 under the reaction conditions of T=543 K, P=2 MPa, and H2/CO=2. The fractions of CH4, C2–C4 and C5–C16 in hydrocarbons were approximately 5 %, 40 %, and 50 %, respectively. This catalyst could give a high ratio of olefins; the ratio of olefin/paraffin in the C2–C4 range was about 2.2 and that in the C5–C16 range was about 1.4. Characterizations suggested that some Fe0 nanoclusters were possibly located in the voids of zeolite LTL, and these Fe0 nanoclusters were retained during the reaction, whereas the Fe0 located outside the zeolite was carburized in FT synthesis. The small Fe nanoclusters might be responsible for the high olefin selectivity, whereas the iron carbides likely accounted for the formation of heavy hydrocarbons.

Table 5. Catalytic behaviors of zeolite-supported Fe catalysts.[a]
Catalyst[b]CO conv. [%]Selectivity [%]
  CO2CH4C2–4C5–9C10–20C21+
  1. [a] Reaction conditions: T=543 K, P=2 MPa, H2/CO=2; [b] Fe loading=10 wt %.

Fe/SiO222121540294.10.1
Fe/Na-ZSM-53.01914481900
Fe/Na-MOR29201235237.03.2
Fe/Na-β5.82416461401
Fe/NaX48305.61824185.1
Fe/NaY49217.82323178.9
Fe/LiY408.66.716272616
Fe/K-Y75368.323199.73.9

Besides the nanospace, the acidity of acid-form zeolites has also been used to regulate the product selectivity in FT synthesis. Bessell166 demonstrated that the Co catalysts supported on ZSM-5, zeolite Y, mordenite, SiO2, Al2O3, and bentonite showed similar selectivities to CH4 (17–26 %) and CO2 (1–2 %), but the distribution of higher hydrocarbons depended strongly on the support acidity. The straight-chained FT products were obtained over catalysts with the nonzeolitic and low-acidity supports. On the other hand, the most strongly acidic ZSM-5 supported Co catalyst showed the highest selectivity to gasoline-range hydrocarbons (81.4 % in liquid hydrocarbon products) and the largest fraction of branched products. No shape selectivity of zeolites was observed. A further comparison of the Co catalysts supported on pentasil zeolites with different pore structures, including ZSM-5, ZSM-11, ZSM-12, and ZSM-34, revealed that the use of ZSM-12, which possesses the largest pore channels (0.57×0.61 nm) but weaker acidity, provided the highest fraction of gasoline in liquid hydrocarbons and the lowest fraction of n-paraffins.167 In contrast, ZSM-34, with a more constrained channel structure, produced heavier products containing more n-paraffins. The chain growth occurred on Co particles, and the resultant primary hydrocarbons then underwent secondary reactions at the accessible zeolite acid sites to produce lighter products with more iso-paraffins. Since most of the Co particles may not be located inside the small zeolite channels, the accessible acid sites mainly include the external acid sites and the internal acid sites close to the pore mouths of the zeolite. The degree of secondary restructuring was the greatest over ZSM-12-supported catalyst, indicating that the accessibility of primary FT products to the internal acid sites is an important factor.

4.2. Design of zeolite-containing hybrid and core–shell catalysts for FT synthesis

As described above, the combination of FT active metals or catalysts with acidic zeolites can produce high-octane branched hydrocarbons in the gasoline range. Such bifunctional processes have been studied intensively in different ways, for example, by using zeolite-supported Co catalysts described above, using a hybrid catalyst containing the mixture (typically a physical mixture) of a FT component and a zeolite component in a single reactor, or by using a dual reactor arrangement with the two functional catalysts in separate reactors.168 Shell developed a two-step process for producing branched hydrocarbons from syngas. A conventional FT catalyst was employed in the first step, and the obtained wax products were hydrocracked separately in the second step. This process is not only complicated but catalyst deactivation also occurs easily due to the deposition of wax. The coexistence of an acidic zeolite with the conventional FT catalyst may inhibit the deactivation of the first-step catalyst by simultaneous hydrocracking of waxes.

Higher temperatures (e.g., ≥573 K) were typically required for efficient secondary reactions, such as hydrocracking, isomerization, and aromatization, over a zeolite catalyst. Since Co catalysts could not work at such high temperatures, Fe-based catalysts were previously thought to be more suitable for the design of a hybrid catalyst. However, a physical mixture of a fused Fe catalyst and H-ZSM-5 (Si/Al=30 or 280) showed quicker deactivation and lower C5–C11 (gasoline fraction) selectivity in a Betty microreactor at 573 K as compared to the dual layer configuration with a wire mesh between the two catalysts to avoid direct contact.169 The alkali migration from the Fe catalyst to the zeolite occurred, leading to a shift in the selectivity of the FT catalyst towards light paraffins, which could not be converted over H-ZSM-5. In the case of dual layer configuration, the use of a higher acidic H-ZSM-5 (Si/Al=30) resulted in the formation of aromatics together with light paraffins. The deactivation of H-ZSM-5 occurred rapidly. The deactivation rate of the H-ZSM-5 with a high Si/Al ratio (280) was slower, and the increase of operation temperature to 623 K could improve the performance and the catalyst stability.170

A more detailed investigation on hybrid catalysts prepared by physically mixing a K–Fe–Co (K/Fe/Co=45:3:1) catalyst and H-ZSM-5 with different Si/Al ratios and crystal sizes (1:1 weight ratio) was reported.171 Under the reaction conditions of T=583 K, P=2 MPa, H2/CO=1/1 and time on stream of 15–17 h, the base K–Fe–Co catalyst gave a CO conversion of 79.1 % and selectivities to hydrocarbons, alcohols, and CO2 of 58.3 %, 3.4 % and 38.3 %, respectively. The addition of H-ZSM-5 increased the CO conversion to greater than 90 % and increased the selectivity to hydrocarbons by suppressing the formation of alcohols. The fractions of C5–C12 and C13+ in the hydrocarbons were, respectively, about 30 % and 35 % over the base catalyst. The addition of zeolite suppressed C13+ formation, and increased the C5–C12 fraction significantly to 50–80 %, depending on the properties of the zeolite used. The increase in selectivity to gasoline by addition of H-ZSM-5 resulted not only from the cracking of the heavier hydrocarbons but also from the formation of aromatics, mainly benzene and C7–C10 alkylbenzenes, by consecutive oligomerization, cyclization, and dehydrogenation of the primary short-chain α-olefins. The concentration of aromatics in the total hydrocarbons reached 35–40 % at the initial stage over the hybrid catalysts containing H-ZSM-5 with a lower Si/Al ratio (15–50), but it decreased with time on stream due to the poisoning of the acid sites by coke deposition. The use of H-ZSM-5 with a smaller crystal size (ca. 100 nm) or the addition of small amount of Pd could suppress the deactivation. Similar phenomena were observed over a hybrid catalyst composed of H-ZSM-5 (SiO2/Al2O3=83.7) and a Fe-based catalyst (Fe/Cu/Mg/Ca/K=200:10:300:30:10); the C11+ heavier hydrocarbons disappeared, and the selectivity of C4–C6 branched hydrocarbons increased significantly.172

Hybrid catalysts, composed of Co catalysts and zeolites can also provide gasoline-fraction hydrocarbons selectively under proper conditions. The addition of H-ZSM-5 to Co/SiO2 by physical mixing was found to suppress the formation of C11+ and increased the selectivity of C4–C10iso-paraffin under the reaction conditions of T=503–523 K, P=1.0MPa, H2/CO=3.173 A further addition of Pd/SiO2 as a third component in the hybrid catalyst could decrease the selectivity of CH4 from 18 % to about 11 % at 523 K and increase the catalyst stability. The use of a two-stage reactor containing a hybrid catalyst of Co/SiO2 and H-ZSM-5 in the first reactor working at a lower temperature (513–523 K) and another hybrid catalyst of H-β and Pd/SiO2 in the second reactor working at a higher temperature (573–593 K) could terminate the products at a carbon number of 6.174 By choosing suitable temperatures for both reactors, a selectivity of 64.4 % for C4–C6iso-paraffins was achieved. With the hybrid of H-ZSM-5 in the first-stage reactor, heavier hydrocarbons (waxes) deposited were removed from Co/SiO2, maintaining its stability, while with the Pd/SiO2 hybrid in the second-stage reactor, the acid sites of H-β were maintained and strengthened by a hydrogen spillover effect. A detailed study further demonstrated that, under typical FT synthesis conditions for Co catalysts (T=523 K, P=2.0MPa and H2/CO=2), the zeolite could effect cracking of the C13+ hydrocarbons formed on the Co catalyst mainly to gasoline-range (C5–C12) branched products.175 No aromatic products were detected over the Co-based hybrid catalyst. This is quite different from the Fe-based hybrid catalysts, whereby a large fraction of aromatics were obtained.169171 The positive effect of zeolite on the stability of the Co/SiO2 catalyst was not ascertained. The yield of branched hydrocarbons decreased with time on stream. Among the zeolites examined, the deactivation rate was found to increase with the pore dimensions, (i.e., H-ZSM-5<H-MOR<H-β<USY). Coke molecules were mainly 2- and 3-ring aromatics in larger pore zeolites (USY and H-β), which might be formed from light olefins produced in FT synthesis via consecutive oligomerization, cyclization, and dehydrogenation reactions. Further comparisons among different kinds of zeolite cocatalysts, including H-ZSM-5, MCM-22, ITQ-2, and IM-5, with Co/SiO2 revealed that the initial yield of iso-C5–C8 hydrocarbons correlated well with the external acidity rather than the total amount of the Brønsted acid sites, indicating the existence of limitations in the diffusion of the long-chain n-paraffin within the zeolite channels under reaction conditions.176

The contact between the FT catalyst and the zeolite and their spatial arrangement are expected to be crucial for the bifunctional process. To enhance this process, Tsubaki and co-workers177182 developed core–shell- or capsule-type catalysts for the direct conversion of syngas to gasoline-range iso-paraffins. They coated zeolite membranes on pellets of a conventional FT catalyst, such as Co/SiO2 or Co/Al2O3, by a hydrothermal synthesis method. The time for crystallization of the zeolite over the FT catalyst could be used for adjusting the thickness of the enwrapped zeolite membrane. It is expected that the reactants, that is, syngas, diffuse through the zeolite membrane (shell) to reach the FT catalyst (core), on which long-chain normal paraffins are formed, and then the normal paraffins diffuse through the zeolite membrane, where they can be cracked and isomerized (Figure 12). Because of the unique spatial configuration of this core–shell catalyst, all straight-chain hydrocarbons formed on the FT core catalyst can enter the zeolite channels and the efficiency of the secondary reactions is expected to be higher than the hybrid catalyst.

thumbnail image

Figure 12. Schematic representation of the modified FT synthesis over core–shell catalysts.177, 178

Download figure to PowerPoint

The experimental results showed that the coating of H-ZSM-5 (Si/Al=80) on 10 wt % Co/SiO2 to form a core–shell catalyst suppressed C11+ formation under the conditions of T=533 K, P=1.0MPa, H2/CO=2, whereas there were still heavier paraffins, up to C20, formed on the mechanically mixed hybrid catalyst under the same conditions.177, 178 CH4 selectivity was higher over the core–shell catalyst composed of H-ZSM-5 and Co/SiO2 and increased with the thickness of H-ZSM-5 membrane shell (Figure 13), possibly because H2 diffuses more quickly than CO inside the small zeolite pores or channels, leading to a higher H2/CO ratio over the core catalyst. The selectivity of iso-paraffins increased with zeolite thickness, and the ratio of the iso- to n-paraffin for C4+ increased to 1.88, whereas it was only 0.49 over the hybrid catalyst with a similar composition. It is of interest that the ratio of olefin to paraffin was also increased with the thickness of the H-ZSM-5 membrane. The core–shell catalyst was found to be very stable in a 100 h test reaction.178 Further studies showed that the use of smaller Co/SiO2 pellets favored the growth of H-ZSM-5 membrane, and increased the iso-paraffin selectivity as well as the CO conversion.179 By combining a hydrogenation catalyst, such as Pd/SiO2, with the core–shell catalyst in a single reactor as a dual-bed catalyst, the olefins formed from the core–shell catalyst were hydrogenated, mostly converted to iso- paraffins.180

thumbnail image

Figure 13. Influence of the thickness of H-ZSM-5 shell on catalytic behavior of the H-ZSM-5/Co/SiO2 core–shell catalysts.178 Reaction conditions: T=533 K, P=1 MPa, H2/CO=2.

Download figure to PowerPoint

The formation of a core–shell catalyst by coating H-β on a 10 wt % Co/Al2O3 also resulted in a cut-off of heavier hydrocarbons (C12+) and a significant increase in the selectivity of middle (C4–C11) iso-paraffins (Figure 14).181, 182 The ratio of iso-paraffins to n-paraffins for C4+ reached 2.34 over the H-β/Co/Al2O3 core–shell catalyst under the conditions of T=533 K, P=1.0 MPa and H2/CO=2. Moreover, this core–shell catalyst exhibited a lower CH4 selectivity (13.6 %) than the Co/Al2O3 itself (21.9 %) and the physical mixture of H-β and Co/Al2O3 (16.6 %) with the same compositions. This observation is different from that for the core–shell catalyst composed of H-ZSM-5 and Co/SiO2.177, 178 The factors influencing the formation of the H-β/Co/Al2O3 core–shell catalyst were investigated in detail.182 Prior to the hydrothermal synthesis in the H-β precursor solution for H-β membrane growth, the pretreatment of the Co/Al2O3 pellets under refluxing tetraethylammonium hydroxide (TEAOH) solution followed by the impregnation of the pretreated Co/Al2O3 with ethanol was found to be crucial. The possible functions of the hot TEAOH treatment were proposed as follows: a) to clean the surface of Co/Al2O3 pellets using its strong basicity; b) to increase the hydroxyl groups over the surface of the core Co/Al2O3 catalyst, and the Al[BOND]OH and Si[BOND]OH groups in the zeolite precursor solution may react with these hydroxyl groups to form stable Al[BOND]O[BOND]Al and Al[BOND]O[BOND]Si bonds for anchoring the H-β shell; c) to corrode the surface of Co/Al2O3 catalyst pellets to make a rough surface easier to coat.

thumbnail image

Figure 14. Product distribution over Co/Al2O3 (A), hybrid catalyst (B) and core–shell catalyst (C).182 Reaction conditions: T=533 K, P=1 MPa, H2/CO=2.

Download figure to PowerPoint

5. Utilization of Novel Carbon Materials for FT Synthesis and the Size Effect of Metal Particles

  1. Top of page
  2. Abstract
  3. Introduction
  4. 1. Overview of Catalysts, Reactors, Reaction Mechanisms, and Product Distributions for FT Synthesis
  5. 2. Key Factors Determining the Selectivity and Activity for FT Synthesis
  6. 3. Utilization of Nanoporous Materials for FT Synthesis
  7. 4. Utilization of Zeolites for FT Synthesis
  8. 5. Utilization of Novel Carbon Materials for FT Synthesis and the Size Effect of Metal Particles
  9. Summary and Outlook
  10. Acknowledgements
  11. Biographical Information
  12. Biographical Information
  13. Biographical Information

5.1. Novel carbon materials as efficient supports for FT catalysts

Carbon materials such as carbon nanofibers (CNFs) and nanotubes (CNTs) and mesoporous carbons have recently attracted much attention as catalyst supports because of their unique properties, such as high purity, high mechanical strength, good electrical conductivity, high thermal stability, and high surface area.183190 These carbon materials provide new possibilities to tune the metal–support interactions and to prepare size-controllable metal nanoparticles, either in their interior nanospaces or on their external surface. CNFs or CNTs possess high external surface areas without micropores, which may be beneficial to FT reactions with problems of mass transfer. The ordered porous channels in carbon nanotubes or mesoporous carbons may also exert a confinement effect, resulting in unique catalytic properties.188, 190 Different functional groups (e.g., carboxylic acid groups) can also be introduced onto carbon surfaces, which are useful for immobilization of catalytically active sites or cause additional catalytic functions (e.g., acid catalysis). Furthermore, hydrogen species adsorbed on carbon nanotubes formed via the spillover effect may participate in hydrogenation reactions with unique properties.191

Carbon materials have received much research interests in FT synthesis. Some early studies demonstrated that carbon-supported Fe catalysts were promising for the synthesis of light olefins.192195 However, these studies were generally performed at lower pressures (e.g., 0.1 MPa) and suffered from lower CO conversions (<10 %). Several recent studies have investigated the effect of preparation techniques for introducing Fe catalysts, as well as promoters such as Cu and K, onto CNTs, and both the impregnation and the deposition–precipitation method have produced active Fe/CNT catalysts.196, 197 The modification of Fe/CNT with Cu enhanced both the FT and the WGS reaction rates, but no significant changes in product selectivity were observed by the addition of Cu. On the other hand, the modification by K+ decreased the selectivities of CH4 and C2–C4 and increased those of C5–C11 from roughly 42 to 52 % and of C12+ from roughly 2 to 8 %.197 The fraction of C2H4 in total C2 hydrocarbons also increased significantly from about 10 to 70 %. The presence of K+ (K/Fe weight ratio=0.5–0.7:9) decreased the hydrogenation ability of the Fe/CNT catalyst. The addition to K+ to a CNT-supported Fe–Ru bimetallic catalyst also significantly enhanced C5+ formation and the ratio of olefins to paraffins.198 Although a previous study showed the deactivation of Fe/CNT catalyst,196 recent results demonstrated that the CNT-supported Fe or Fe–Ru bimetallic catalyst was quite stable.197199 The CNT surfaces may contain docking stations to keep the small metal nanoparticles from sintering.199

Guczi et al.200 found that, compared to Co/CNT, the Fe/CNT prepared by the impregnation was a more active catalyst for FT synthesis. The fraction of C2–C6 olefins in the total C2–C6 was quite high (>85 %). The Fe/CNT catalyst also exhibited lower CH4 selectivity and higher C2–C4 and C5+ selectivities as well as better olefin selectivity. It is of interest to note that the large and aggregated Fe particles over the CNTs after H2 reduction become smaller (10–20 nm) and homogeneously dispersed after the FT reaction, suggesting the reconstruction of the Fe species, possibly by carburization.200

Bao and co-workers performed an interesting study to elucidate the confinement effect of CNTs on the structural and catalytic properties of Fe catalysts.201 They developed techniques to selectively prepare Fe nanoparticles with similar sizes (4–8 nm) but with different locations, inside and outside the CNTs (denoted as Fe-in-CNT and Fe-out-CNT, respectively). The use of CNTs with closed caps for impregnating iron nitrate aqueous solution produced the Fe-out-CNT, whereas the CNTs with the caps opened by strong acid pretreatment were used for the preparation of the Fe-in-CNT. Autoreduction or H2 or CO reduction of iron species was significantly enhanced when they were located inside the CNTs.201203 Through in situ XRD studies under near-reaction conditions (543 K, 0.1–0.95 MPa), the transformation of metallic Fe and Fe3O4 to iron carbides (FexCy, most likely Fe5C2 and Fe2C) and FeO was observed. The relative ratio of the integral XRD peaks of FexCy/FeO was about 4.7 for the Fe-in-CNT catalyst, whereas that for the Fe-out-CNT catalyst was about 2.4, suggesting that the formation of FexCy was facilitated inside the CNTs. This resulted in significant differences in FT synthesis (Table 6). Under the conditions of T=543 K, P=5.1 MPa, H2/CO=2, CO conversion over the Fe-in-CNT was 40 %, being 1.4 and 2.4 times greater than those over the Fe-out-CNT and a Fe/AC, respectively. The C5+ selectivity was also enhanced by confining the Fe inside the CNTs. A C5+ space–time yield of 440 g kgequation image h−1 was achieved over the Fe-in-CNT, significantly higher than those over the Fe-out-CNT and the Fe/AC (210 and 61 g kgequation image h−1). The Fe-in-CNT was also very stable, and the Fe particle size was not significantly changed even after 200 h of reaction. In contrast, the Fe particle size for Fe-out-CNT grew significantly during the reaction. The Fe particles over Fe/AC were approximately 8–12 nm in size. It was proposed that the better reducibility and the higher FexCy concentration played key roles for the higher activity and C5+ selectivity over the CNT-confined Fe catalyst. The reactions confined in the CNTs may also have enhanced the readsorption of α-olefins and the chain-growth probability.

Table 6. Catalytic properties of Fe catalysts located inside and outside the CNTs.
CatalystCO conv. [%]Selectivity [%]Size of Fe [nm]
  CO2CH4C2−4C5+ 
  1. [a] From Ref. 201; Fe loading=10 wt %; reaction conditions: T=270 °C, P=5.1 MPa, H2/CO=2; [b] from Ref. 204; Fe loading=ca. 12 wt %; reaction conditions: T=270 °C, P=2 MPa, H2/CO=2, TOS=125 h; [c] CO2 selectivity was calculated separately from the hydrocarbon selectivity.

Fe-in-CNT[a]4018124129Before and after: 4–8
Fe-out-CNT[a]2912155419Before: 6–10 After: 12–16
Fe/AC[a]17515719Before and after: 8–12
in-Fe/CNT[b]ca. 8538.9[c]25.638.236.2Before: 6–11 After: 6–12
out-Fe/CNT[b]ca. 7939.5[c]40.535.723.8Before: 5–9 After: 6–24

Dalai and co-workers204 also prepared CNT-supported Fe catalysts with similar Fe particle sizes but different locations. They succeeded in preparing catalysts with iron oxides located mostly (ca. 80 %) inside the CNTs (denoted as in-Fe/CNT) or mainly (ca. 70 %) outside the CNTs (denoted as out-Fe/CNT) by using the CNTs pretreated with the same procedure. The following conclusions, similar to those made by Bao and co-workers,201 were obtained by using these catalysts: a) the iron oxides confined in the CNTs were reduced more easily than those outside the CNTs; b) in-Fe/CNT exhibited lower CH4 selectivity and higher C5+ selectivity than the out-Fe/CNT (Table 6); c) in-Fe/CNT was more stable than out-Fe/CNT; d) the Fe particles over out-Fe/CNT underwent serious aggregation whereas those over in-Fe/CNT remained almost unchanged.

CNF- and CNT-supported Co catalysts have also been studied for FT synthesis. Stable activity was observed over a herringbone-type CNF-supported Co catalyst for reactions at 2.8–4.2 MPa, and a C5+ selectivity of 86 % was obtained over a 12 wt % Co/CNF catalyst.205 A simple two-step impregnation method produced Mn-promoted Co/CNF catalysts containing Mn directly associated with Co, as indicated by STEM, EELS and XPS characterization.206, 207 For the Mn-promoted Co/TiO2 catalyst, the migration of MnOx to the TiO2 support occurred readily during the reduction.91 However, over CNFs, MnO remained close to the Co particles after reduction, probably because of weak interactions between MnO and the CNFs.207 The small amount of Mn significantly enhanced the C5+ selectivity and the TOF of the Co/CNF catalyst. Thus, the CNFs can provide a way for preparing FT catalysts with direct contact (or strong interaction) between the active metal and the modifier without interference of support effects.

Co/CNF catalysts containing uniformly distributed Co nanoparticles were successfully prepared by a homogeneous deposition–precipitation (HDP) from a basic cobalt solution containing [Co(NH3)6]2+ followed by ammonia evaporation.208 The size of Co nanoparticles over such prepared catalysts were roughly 8 nm, whereas that for the catalyst prepared by HDP from an acid cobalt solution using urea hydrolysis was about 25 nm. The former catalyst showed higher CO conversion activity and higher C5+ selectivities in FT synthesis.

Some studies have contributed to comparing catalytic performances of the Co catalysts loaded on the CNFs or CNTs with those loaded on the conventional metal oxide supports. The catalytic performances of Co catalysts supported on CNFs with two different structures, i.e., herringbone and platelet were compared with those on Al2O3.209 The Co particles over the Co/CNF platelets were smaller than those over the Co/CNT herringbone, and the former catalyst exhibited a higher CO conversion. The C5+ selectivities were similar over both catalysts. The C5+ selectivities over Co/CNF catalysts (81–82 %) were slightly higher than that over Co/γ-Al2O3 (ca. 79 %) at an equal conversion (40 %). Tavasoli et al.210, 211 clarified that, as compared to γ-Al2O3, the CNT as a support could reduce the metal–support interactions and enhance the Co reducibility. At the same time, higher Co dispersion or smaller Co size could be achieved by using CNTs. Therefore, the use of CNTs can facilitate both the reducibility and the dispersion of Co species. The CO conversion was higher over the Co/CNT catalyst than that over the Co/γ-Al2O3 in both the fixed-bed and the continuous stirred tank reactors. The C5+ selectivity was slightly lower and the C2–C4 selectivity was slightly higher over the Co/CNT catalyst as compared to those over the Co/γ-Al2O3 catalyst in the fixed-bed reactor under reaction conditions of T=493 K, P=0.1 MPa and H2/CO=2.210 In the continuous stirred tank reactor, a slight shift of products to lighter hydrocarbons was observed over the Co/CNT.211 The smaller size of Co particles over the CNTs may be responsible for the shift to lighter hydrocarbons. An increase in Co loading on the CNTs could increase the size of Co particles and enhance the C5+ selectivity. A recent study demonstrated that, although Co/CNT was not more selective toward C5+ formation than Co/Al2O3, the Co catalyst loaded on CNTs grown on MgO (Co/CNT–MgO) exhibited higher C5+ selectivity and a markedly increased olefin/paraffin ratio of C2–C3 (11.2).212

The effect of acid pretreatment for CNTs on the catalytic behavior of the supported Co catalysts was investigated.213 HNO3 pretreatment opened the caps of the closed tubes, producing catalysts with a larger fraction of Co particles homogeneously dispersed within the tubes. The reducibility of Co species increased and the size of Co particles decreased. As expected, the CO conversion increased significantly by using the acid-treated CNT as the support. However, the C5+ selectivity became lower and the CH4 selectivity became higher. These tendencies are unlike those of the CNT-supported Fe catalyst, with which a higher C5+ selectivity was achieved over the confined Fe particles.201, 204 Another study showed that the Co was mainly anchored on the outer surface of the acid-pretreated CNTs and had irregular shape.214 The average size of cobalt oxide particles before reduction was about 11–15 nm, slightly depending on the outer diameter of the CNTs. The catalysts also displayed high cobalt reducibility, which was only slightly influenced by the nitric acid pretreatment and CNT outer diameters. Acid pretreatment resulted in a 25 % increase in hydrocarbon yield, mainly owing to the removal of impurities and oxidation of the CNT surfaces.214

Very few studies concerned Ru/CNT catalysts for FT synthesis. A Ru–Mn/CNT exhibited a comparable FT activity to Ru–Mn/γ-Al2O3 in a slurry-phase reaction.215 Kang et al.216 compared the catalytic performances of Ru catalysts loaded on various supports for the conversion of syngas with a H2/CO ratio of 1.0. Compared to Ru/SiO2, the Ru/CNT catalyst exhibited a slightly higher CO conversion (Figure 15). More significantly, Ru/CNT demonstrated a markedly higher C10–C20 selectivity (60 %). The acid pretreatment for CNTs was found to be key to obtaining the high C10–C20 selectivity; it caused the generation of acidic functional groups, which might lead to the cracking of C21+ hydrocarbons. Larger concentrations of adsorbed hydrogen species were detected over the CNT- and zeolite β-supported Ru catalysts, which also gave lower C21+ selectivities. Thus, both the acidic functional groups and the unique hydrogen species over the CNT-supported catalyst played roles in the selective cracking of heavier hydrocarbons to C10–C20.

thumbnail image

Figure 15. Comparison of catalytic performances of some supported Ru catalysts (Ru loading=3 wt %). Reaction conditions: T=533 K, P=2.0 MPa, H2/CO=1, Wcat.=0.5 g, F=20 mL min−1.

Download figure to PowerPoint

5.2. Effect of Size of Metal Nanoparticles on FT Synthesis

Elucidation of the size effects of metal nanoparticles in FT synthesis is important for the design of highly active and selectivity-controllable FT catalysts. Boudart and McDonald summarized the results on structure sensitivity of FT synthesis over some Group 8 metal catalysts.217 CO hydrogenation reactions do not show a clear-cut structure-sensitive or -insensitive behavior. For example, for supported Ru catalysts, some authors reported increases in FT activity (TOF) on increasing the particle size or decreasing Ru dispersion, but others found no particle size effect.218, 219 Early work on supported Fe catalysts showed that the TOF increased with Fe crystallite size.194, 195, 217 Variation in the structure of iron carbides, which are believed to be an active phase in FT synthesis, with the particle size may complicate the particle size effect for these catalysts. Many other factors such as the metal–support interactions and the preferential deactivation of certain sites during reaction conditions can make the interpretation of size effect in FT synthesis quite complicated.

For Co-catalyzed FT synthesis, Iglesia demonstrated that the TOF was not changed with the Co particle size in the range of 9–200 nm.10 Early results on the size effect for catalysts containing smaller Co particles over metal oxide supports are controversial because the strong metal–support interaction may mask the intrinsic size effect by influencing the reducibility of cobalt species.220 Therefore, to find a catalyst system with a weak metal–support interaction and variable Co particle sizes is crucial for the elucidation of cobalt size effect.

As described above, the use of CNFs or CNTs as the supports of Co could result in both higher reducibility (weaker metal–support interactions) and higher Co dispersion.205214 A comprehensive study on the Co size effect using Co/CNF catalysts with average Co particle sizes ranging from 2.6 to 27 nm was reported by de Jong and co-workers.220 They confirmed that cobalt species over these catalysts were in metallic state before and after FT reactions, and no cobalt oxide or carbide phases were detected after reactions. FT reactions were performed under both low (0.1 MPa) and high (3.5 MPa) pressures. In both cases, the TOF first increased with increasing Co particle size from 2.6 nm up to about 6 nm (0.1 MPa) or 8 nm (3.5 MPa; Figure 16), and then remained almost unchanged with further increases in Co size. The product selectivity also depended on the size of Co particles; the C5+ selectivity first increased with the Co particle size from 2.6 to around 6 nm or 8 nm, and then remained unchanged or increased slightly. The selectivity of CH4 underwent the reverse change at the same time. For catalysts with Co particle sizes greater than 8 nm, a shift of product distribution to heavier hydrocarbons was observed on further increasing the Co size to 16 nm, accompanied by an increase in α value. Larger Co particles may favor the production of heavier hydrocarbons.

thumbnail image

Figure 16. Changes of TOF for CO conversion and C5+ selectivity as a function of Co particle size over the Co/CNF catalysts.220 Reaction conditions: T=483 or 523 K, P=3.5 MPa, H2/CO=2.

Download figure to PowerPoint

To understand the particle size effect observed for the Co/CNF catalysts, a study on the surface coverages and residence times of carbon-, oxygen- and hydrogen-containing intermediates and reactants was performed using steady-state isotopic transient kinetic analysis (SSITKA) experiments (0.185 MPa, H2/CO=10).221 The results suggest that the lower TOF for the smaller Co particles (<6 nm) may be attributed to the longer CHx residence time and its lower surface coverage. A significant amount of irreversibly bonded CO was found to exist on the smaller Co particles, likely causing blocking of the Co surfaces and also contributing to their lower activity. The higher CH4 selectivity of the smaller Co particles was proposed to arise from their higher hydrogen coverage.

A subsequent analysis suggested that the Co particles with sizes of 4.7±0.2 nm could provide the maximum cobalt–time yield (CTY, 10−5 molCOequation image s−1).222 Since the broadening of particle size distribution in an actual case may decrease the calculated CTY, the preparation of supported Co particles with a narrow size distribution around 5 nm would provide an optimized CTY value. By using an NO/He (1 % NO) gas for the calcination of a Co/SiO2 prepared via incipient wetness impregnation of cobalt nitrate, it was possible to prepare SiO2-supported Co particles with narrow size distributions.223 The prepared Co/SiO2 catalyst with Co particle sizes at 4.6±0.8 nm exhibited a significantly higher CTY (4.8×10−5  molCOequation image s−1 under T=493 K, P=0.1 MPa and H2/CO=2) than the Co/SiO2 prepared by conventional air calcination (2.41×10−5  molCOequation image s−1).222 However, the Co/SiO2 catalyst prepared by the NO/He calcination showed a lower C5+ selectivity.

Similar to CNFs, ITQ-2 also possesses a large external surface area. Co/ITQ-2 has been proven to be an efficient FT catalyst.164 A new technique for preparing Co/ITQ-2 catalysts with both high reducibilities and narrow Co particle size distributions was developed by combining reverse micelles containing cobalt in the core with a surface-silylated ITQ-2 delaminated zeolite.224 Using the Co/ITQ-2 catalysts with mean Co sizes ranging from about 5 to 11 nm, Martínez and co-workers performed a detailed study on the Co size effect.225 Under the reaction conditions of T=493 K, P=2.0MPa and H2/CO=2, the TOF was found to increase from 1.2×10−3 to 8.6×10−3 s−1 with increasing mean size of Co particles from 5.6 to 10.4 nm. The TOF for a Co/SiO2 catalyst, which had a much larger mean size of Co (141 nm), was 8.2×10−3 s−1 under the same reaction conditions. Thus, an increase in Co size from 10.4 to 141 nm does not significantly change the TOF. This trend of variation of TOF with the mean size of Co particles is similar to that for the Co/CNFs.220 The C5+ selectivity for the series of Co/ITQ-2 catalysts changed only slightly from 68.8 % to 61.6 % with increasing Co particle size from 5.6 to 10.4 nm.225

The influence of Co particle size on the catalytic performances was also investigated using Co/γ-Al2O3 or Co/α-Al2O3 with different mean Co sizes.226 With changing the Co particle size in the range of 3.1–18 nm, the TOF varied in the range of 3.1×10−2–6.3×10−2 s−1 under the reaction conditions of T=483 K, P=2 MPa and H2/CO=2.1. It appears that there is no obvious correlation between the TOF and the Co particle size. In contrast, when the C5+ selectivity for the Co/γ-Al2O3 series of catalysts at the same CO conversion was plotted against the Co particle size, a volcano-like curve was obtained, and the optimum size was 7–8 nm. This trend is different from those for the Co/CNFs, where the C5+ selectivity increased monotonically with Co particle size (2.6–27 nm), and for Co/ITQ-2, with which the C5+ selectivity decreased slightly on increasing the Co particle size (5.6–10.4 nm).

The size effect in Ru-catalyzed FT synthesis is also of fundamental interest. An early study indicated the increase of TOF for CO hydrogenation on decreasing the dispersion of Ru.218 Although the reduction of Ru species is generally more facile than that of Co species, the strong metal–support interaction (SMSI) can also complicate the understanding of the intrinsic Ru size effect. For example, the SMSI effect for the Ru/TiO2 after reduction caused the covering of Ru particles by amorphous TiOx, which exerted significant influences on both TOF and product selectivity.227, 228 With increasing the coverage of TiOx over Ru particles, the TOF passed through a maximum, while the α value and the olefin-to-paraffin ratio increased monotonically.

Kang et al.216 studied the effect of Ru size on FT catalytic behavior of the Ru/CNT catalysts, which exhibited high C10–C20 selectivity. Under reaction conditions of T=533 K, P=2.0 MPa and H2/CO=1, CO conversions over the Ru/CNT catalysts with mean Ru sizes of 2.3–10 nm changed only slightly in the range of 25–35 %, whereas the product selectivity varied significantly. On changing the mean size of Ru particles from 2.3 to 10.2 nm, both the C5+ and the C10-C20 selectivities passed through maxima, and the optimum Ru size was 7-8 nm (Figure 17). The TOF increased with the size of Ru particles up to about 6 nm, and then decreased slightly with a further increase in Ru particle size.

thumbnail image

Figure 17. Effect of mean size of Ru particles on catalytic behavior of Ru/CNT catalysts.216 Reaction conditions: T=533 K, P=2.0 MPa, H2/CO=1.

Download figure to PowerPoint

In short, recent studies have demonstrated that CNFs or CNTs are good catalyst supports for studying the particle size effect. For both Co and Ru catalysts, in the smaller particle size region (<6–10 nm), the TOF increases with the particle size, whereas in the larger particle size region (>6–10 nm), TOF may be kept unchanged or slightly decreased with further increasing particle sizes. These observations coincide generally with the so-called class I surface sensitivity,41 whereby the activation of molecular CO requires a reaction center with a unique configuration of several metal atoms and the step-edge sites. Future studies are required to understand why such larger particle sizes (ca. 6–10 nm) are required to obtain optimized TOFs. Concerning the effect of metal particle size on the product selectivity, more systematic studies are needed. A current consensus appears to be that larger metal particles result in higher C5+ selectivity,144, 220, 222 but there exist exceptions.216, 225, 226 Whether it is possible to control the product selectivity by tailoring the size of metal particles is still ambiguous. Nevertheless, insights into the metal size effect obtained to date have provided important clues for the design of efficient FT catalysts.

Summary and Outlook

  1. Top of page
  2. Abstract
  3. Introduction
  4. 1. Overview of Catalysts, Reactors, Reaction Mechanisms, and Product Distributions for FT Synthesis
  5. 2. Key Factors Determining the Selectivity and Activity for FT Synthesis
  6. 3. Utilization of Nanoporous Materials for FT Synthesis
  7. 4. Utilization of Zeolites for FT Synthesis
  8. 5. Utilization of Novel Carbon Materials for FT Synthesis and the Size Effect of Metal Particles
  9. Summary and Outlook
  10. Acknowledgements
  11. Biographical Information
  12. Biographical Information
  13. Biographical Information

Fischer–Tropsch (FT) synthesis has received renewed interest because of its important position in the transformation of nonpetroleum feedstock to environmentally benign fuels and valuable chemicals. However, very wide product distributions are generally obtained over conventional FT catalysts. Selectivity control remains one of the most important and difficult challenges in the research area of FT synthesis. Development of efficient catalysts with controlled selectivity or tuned product distribution is a highly desirable goal.

The understanding of key factors determining the catalytic behaviors (especially the selectivity) is crucial for rational design of a selective and active FT catalyst. The present review has analyzed the catalyst factors in detail. The following factors are proposed to play crucial roles in controlling the product selectivity.

Firstly, the nature of active components determines the product distribution. Ru catalysts possess the highest activity and outstanding chain growth probability, and are of interest for fundamental research. Co catalysts are normally favorable for the production of linear long-chain alkanes, which can be used as super clean diesel fuel or lubricants after further hydrotreatments. Ru0 and Co0 nanoparticles are the active phase in FT synthesis. For Co-based catalysts, the reducibility and the dispersion are two keys for obtaining optimum activity and C5+ selectivity. Note that a recent report has demonstrated that the combination of Co0 and Co2C can produce C2–C18 α-alcohols with good selectivity (ca. 36 %).46 In contrast, not only long-chain linear alkanes but also olefins and oxygenates can be produced over Fe catalysts. Many studies indicate that iron carbides are the active phase, but the nature of the active and selective carbides is still unclear. A few recent reports appear to suggest that χ-Fe5C2 may be responsible for hydrocarbon formation.52, 76, 200 When large catalyst pellets (1–3 mm) are used, the distribution of active component in the pellet also affects the product selectivity because of the diffusion limitation, and the use of egg-shell Co/SiO2 catalysts has been demonstrated to enhance the C5+ selectivity.63, 64

Secondly, promoters are essential for modifying the selectivity to target products especially in the case of Co and Fe catalysts. Noble metals such as Ru and Re can enhance the reducibility and/or dispersion of Co species and can thus increase the C5+ selectivity of Co catalysts. Transition metal oxides particularly ZrO2 and MnOx may decrease the selectivity of CH4 and increase that of C5+ hydrocarbons by decreasing the hydrogenation ability and/or increasing the CO dissociation probably via regulating the electronic state of Co species. Rare earth oxides, such as La2O3 and CeO2, also show promoting effects for Co catalysts. For Fe catalysts, alkali metal ions are typically required for decreasing CH4 selectivity and increasing the chain growth probability. The ratio of olefins to paraffins can also be raised by modification with alkali metal ions. Cu or Ru added to Fe catalysts can facilitate the reducibility of the catalysts, reducing the induction period and/or increasing the activity. MnOx and MgO efficiently promote Fe catalysts by suppressing catalyst deactivation and the WGS reaction, respectively. The ratio of olefins to paraffins can also be raised by MgO modification.

Thirdly, the catalyst support affects the product selectivity by enhancing the dispersion of metal particles and facilitating heat and mass transfer. The choosing of support with a proper interaction with the active metal (or metal precursor) is crucial because the balance between the reducibility and the dispersion determines the FT catalytic performances. The nanopores of supports may also function as nanoreactors in which the readsorption of primary α-olefins can be enhanced.

We have shown in this review that ordered mesoporous materials are promising FT catalyst supports. The Co or Fe catalysts supported on mesoporous materials such as SBA-15 can provide higher C5+ selectivity and CO conversion than those on conventional SiO2 owing to the enhanced Co dispersion and Co site density. The pore size and structure (or texture), which may determine the reducibility and dispersion of metal particles, the diffusions of reactants and products, and the probability of the secondary reactions, are key parameters influencing the product selectivity. Many reports have shown that the larger pore size leads to higher reducibility and larger size of metal particles, resulting in higher C5+ selectivity and higher CO conversion rates.144149 Some studies have demonstrated that the use of mesoporous materials with proper pore sizes can produce middle distillate fuels with higher selectivities.122, 124, 133, 136138, 141 The key problem for using mesoporous materials with smaller pore sizes is the lower reducibility of Co or Fe species introduced. Useful strategies such as the addition of promoters (e.g., noble metals and ZrO2), incorporation of heteroatoms (e.g., Zr and Al) into the framework of mesoporous silica, or hydrophobic pretreatment (e.g., silylation) of surfaces of mesoporous materials have been proposed to solve this problem. However, little information is available on the sole effect of pore size without variation of reducibility and size of metal particles. The development of techniques for the preparation of catalysts with a high reducibility and a certain size of metal nanoparticles but changeable pore sizes remains a challenge.

Carbon nanofibers (CNFs) or nanotubes (CNTs) with high external surface areas are another type promising supports for FT synthesis. The CNT- or CNF-supported catalysts have been demonstrated to be stable in FT reactions. One of the most attractive features of these carbon materials is that both a high degree of reduction and high dispersion (small particle size) of supported Co or Fe can be achieved. Thus, the CNT- or CNF-supported catalysts usually exhibit higher CO conversion rates.201, 209211 The key parameters influencing the product selectivity of CNF- or CNT-supported catalysts involve the size of metal nanoparticles, the location of active metals, and the pretreatment of CNFs or CNTs. Comprehensive studies using Co/CNFs and Ru/CNTs with average Co and Ru sizes of 2.6–27 nm and 2.3–10 nm, respectively, have broadened our knowledge on the size effect in FT synthesis.216, 220222 It has been demonstrated that relatively larger particle sizes (6–8 nm) are required for obtaining higher TOFs and C5+ selectivities for both Co- and Ru-based catalysts. Future studies are still needed to gain further insights into the size effect on product selectivities in different systems. Two studies have demonstrated the importance of the location of Fe nanoparticles in CNTs.201, 204 As compared to the Fe particles located outside the tubes of CNTs, the Fe particles confined in CNTs are more reducible and can be transformed into iron carbides more facilely in FT reactions. Thus, the confinement of Fe catalysts inside the tube of CNTs results in higher chain growth probability (C5+ selectivity) and higher activity. This CNT-confined Fe catalyst has also shown promising performances for production of light olefins,229 which is a difficult challenge in FT synthesis. However, whether the CNT- or CNF-supported or confined Co catalysts exert positive effects on product selectivity is still ambiguous. The pretreatment of CNTs with acids may alter the product selectivity of CNT-supported catalysts.213, 214, 216 The generated acidic sites on CNT surfaces may work for the selective cracking of heavier (C21+) hydrocarbons, providing an outstanding C10–C20 selectivity (ca. 60 %) over the Ru/CNT catalyst.216

Zeolites have also attracted much attention for FT synthesis. In addition to the shape-selective feature of zeolites, which may exert a cut-off effect for heavier hydrocarbons, the acidity of zeolites has resulted in the development of a series of bifunctional or modified FT catalysts, which can produce iso-paraffins or aromatics in the gasoline range. The combination of an H-form zeolite (typically H-ZSM-5) with a Fe catalyst to form a hybrid catalyst working at temperatures higher than 573 K can provide C6–C12 aromatics with high selectivity. The use of hybrid catalyst composed of a Co catalyst (e.g., Co/SiO2) and an H-form zeolite working at about 523 K can also suppress the formation of C11+ and increase the selectivity to C4–C10iso-paraffins. The development of core–shell-type catalysts containing a FT-catalyst core, such as Co/SiO2 or Co/Al2O3, and a zeolite shell further enhances the bifunctional process, producing more concentrated C4–C10iso-paraffins. Further modification of the bifunctional catalysts may lead to the commercialization of this novel FT process with tailored product selectivity in the near future.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. 1. Overview of Catalysts, Reactors, Reaction Mechanisms, and Product Distributions for FT Synthesis
  5. 2. Key Factors Determining the Selectivity and Activity for FT Synthesis
  6. 3. Utilization of Nanoporous Materials for FT Synthesis
  7. 4. Utilization of Zeolites for FT Synthesis
  8. 5. Utilization of Novel Carbon Materials for FT Synthesis and the Size Effect of Metal Particles
  9. Summary and Outlook
  10. Acknowledgements
  11. Biographical Information
  12. Biographical Information
  13. Biographical Information

The authors gratefully acknowledge financial support from the National Natural Science Foundation of China (Nos. 20625310 and 20923004), the National Basic Research Program of China (Nos. 2005CB221408 and 2010CB732303), the Research Fund for the Doctoral Program of Higher Education (No. 20090121110007), and the Key Scientific Project of Fujian Province (2009HZ0002-1).

Biographical Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. 1. Overview of Catalysts, Reactors, Reaction Mechanisms, and Product Distributions for FT Synthesis
  5. 2. Key Factors Determining the Selectivity and Activity for FT Synthesis
  6. 3. Utilization of Nanoporous Materials for FT Synthesis
  7. 4. Utilization of Zeolites for FT Synthesis
  8. 5. Utilization of Novel Carbon Materials for FT Synthesis and the Size Effect of Metal Particles
  9. Summary and Outlook
  10. Acknowledgements
  11. Biographical Information
  12. Biographical Information
  13. Biographical Information

Ye Wang received his B.S. and M.Sc. degrees from Nanjing University in China in 1986 and 1989, and obtained his Ph.D. degree in 1996 from Tokyo Institute of Technology in Japan. He worked as research associate at Tokyo Institute of Technology from 1996 to 1997 and at Tohoku University from 1997 to 2000. He moved to Hiroshima University in 2000 and was promoted to associate professor in 2001. He became a full professor at Xiamen University in China in the August of 2001. His main research interests are heterogeneous catalysis for selective oxidation and energy- related processes such as Fischer–Tropsch synthesis and cellulose conversions. He has published more than 110 papers in international journals.

Thumbnail image of

Biographical Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. 1. Overview of Catalysts, Reactors, Reaction Mechanisms, and Product Distributions for FT Synthesis
  5. 2. Key Factors Determining the Selectivity and Activity for FT Synthesis
  6. 3. Utilization of Nanoporous Materials for FT Synthesis
  7. 4. Utilization of Zeolites for FT Synthesis
  8. 5. Utilization of Novel Carbon Materials for FT Synthesis and the Size Effect of Metal Particles
  9. Summary and Outlook
  10. Acknowledgements
  11. Biographical Information
  12. Biographical Information
  13. Biographical Information

Qinghong Zhang received her B.S. and M.Sc. degrees from Nanjing University in China in 1989 and 1992, and obtained her Ph.D. degree from Hiroshima University in Japan in 2002. Since the October of 2002, she has been appointed as associate professor at Xiamen University. Her research interests include the synthesis and characterizations of novel nanostructured materials with advanced catalytic properties. She has published more than 60 papers in international journals.

Thumbnail image of