Modulation of Surface and Interface Properties of Cobalt-Based Fischer-Tropsch Synthesis Catalyst

Yang Linyan; Guo Yupeng; Li Zhengjia; Cen Jie; Yao Nan; Li Xiaonian

doi:10.7536/PC220101

Progress in Chemistry ›› 2022, Vol. 34 ›› Issue (10) : 2254-2266 DOI: 10.7536/PC220101

research-article

Scan to read & Share

Modulation of Surface and Interface Properties of Cobalt-Based Fischer-Tropsch Synthesis Catalyst

Author information +

History +

PDF

Abstract

The conversion of synthesis gas into fuel and high value-added chemicals through the Fischer-Tropsch synthesis (FTs) process is a crucial way to solve the problem of clean utilization of resources such as coal, which has been an important part of the modern coal chemical industry in China. It can reduce the dependence on petroleum imports and ensure national energy strategy security. Cobalt-based catalysts have become one of the most widely studied Fischer-Tropsch synthesis catalysts due to their outstanding catalytic activity, high chain growth factor, low CO₂ selectivity, and long life. How to adjust the surface and interface properties of the catalyst to break the dependence of reduction and dispersion, and improve the reaction activity and product selectivity within a certain carbon chain range is still a significant challenge for the development of high-performance cobalt-based Fischer-Tropsch synthesis catalysts. In this paper, the latest research progress in the regulation of the surface and interface properties of cobalt-based Fischer-Tropsch synthesis catalysts is reviewed from three aspects: structure sensitivity (size and crystal effect), metal-support interaction, and confinement effect, so as to provide a theoretical basis for the design of catalyst microstructure and regulation of reaction performance.

Graphical abstract

Keywords

Fischer-Tropsch synthesis / cobalt-based catalysts / structure sensitivity / metal-support interaction / confinement effect.

Cite this article

Download citation ▾

Yang Linyan, Guo Yupeng, Li Zhengjia, Cen Jie, Yao Nan, Li Xiaonian. Modulation of Surface and Interface Properties of Cobalt-Based Fischer-Tropsch Synthesis Catalyst. Progress in Chemistry, 2022, 34(10): 2254-2266 DOI:10.7536/PC220101

登录浏览全文

4963

注册一个新账户忘记密码

1 Introduction

2 Structural Sensitivity

2.1 Size Effect of cobalt

2.2 Crystal Effect of cobalt

2.3 Co₂C active phase

3 Metal-Support Interaction

3.1 Modulation of Metal Oxide Surface by Carbon

3.2 Surface Functionalization of Carbon Materials

4 Confinement Effect

4.1 Co@MO Catalysts

4.2 Co@CNTs Catalysts

4.3 MOFs-Derived Catalysts

5 Conclusion and Perspectives

1 Introduction

Resources and environment are two basic problems that all countries in the world need to face in their economic development. Green utilization of fossil energy is one of the important ways to solve the shortage of resources and energy and alleviate environmental deterioration. Fischer-Tropsch synthesis (FTs) is a process to convert syngas (CO/H₂) derived from non-petroleum carbon resources (e.g., natural gas, coal, and biomass, etc.) into liquid fuels or chemicals, which is a key step for syngas utilization. It is of great value to promote the efficient and clean utilization of coal and biomass resources, ensure national energy security and stability, and maintain the sustainable development of the national economy^[1].

Group VIII metals have obvious activity for carbon monoxide hydrogenation to hydrocarbons, among which Fe, Co, Ni and Ru have the best activity and the most industrial and commercial value^[2]^[3]. Cobalt-based FTs catalyst has become a hot research topic in the field of FTs because of its high activity, strong chain growth activity, low CO₂ selectivity, long life and other characteristics^[4]. For supported metal catalysts, the smaller the metal particle size is, the larger the exposed active specific surface area is, and the higher the catalytic efficiency is. Small-sized metal clusters or nanoparticles tend to have more unsaturated coordination sites (such as step sites and vertex sites), and for structure-sensitive FTs reactions, the particle size will significantly affect the catalytic activity and selectivity^[5]. However, traditional supported cobalt-based catalysts are mainly prepared by impregnation method with metal oxides (Al₂O₃, TiO₂, SiO₂ and ZrO₂, etc.) as supports, and the catalysts often need to be activated at high temperature for a long period, resulting in large grain size and discrete distribution of active components, especially in the case of industrial application with Co loading up to 20 wt%^[6,7]^[8,9]. Conversely, due to the strong interaction between the metal precursor and the support, an increase in the degree of dispersion often leads to a decrease in the degree of reduction^[10⇓~12]. With the continuous emergence of new materials, carbon materials, such as activated carbon, carbon nanotubes and graphene, have been used as supports for Co-based catalysts. Compared with metal oxide supports, carbon materials have inherent surface inertness.It can effectively weaken the interaction between the active metal oxide and the support and improve the reduction degree of the catalyst, but the Co species are easy to agglomerate and sinter during the reaction, resulting in the decrease of the stability of the catalyst^[13]^[14]^[15]^[16]. How to break the dependence of dispersion and reduction is still an important challenge for the development of efficient FTs catalysts^[17].

In this paper, the latest research progress on the regulation of the surface and interface properties of Co-based FTs catalysts to overcome the common problem of "interdependence of dispersion and reduction" will be reviewed from three aspects:It is expected to provide a theoretical basis for the development of Co-based FTs catalysts with higher activity, stability and product selectivity.

2 Structural sensitivity

2.1 Size effect of cobalt

The FTs response is generally considered to be a structure-sensitive response^[18⇓~20]. Jong et al. Found that when the Co particle size increased from 2.6 nm to 16.0 nm, the conversion frequency (TOF) of the catalyst increased from 1.4×10^-3s^-1 to 23.0×10^-3s^-1, and the

C 5 +

selectivity increased from 51% to 85%. When the particle size of the active phase Co was larger than the critical size of 8 nm, there was no significant change in the FTs catalytic activity and selectivity of the catalyst, which indicated that there was a size effect between the conversion TOF and product selectivity of the Co-based catalyst and the Co particle size^[21]. Subsequently, a large number of studies on the size effect of cobalt-based catalysts have been reported. As shown in Figure 1, when the cobalt grain size is in the range of 2 ~ 10 nm, the TOF value and

C 5 +

selectivity increase gradually with the increase of the grain size of the active phase, while the CH₄ selectivity decreases gradually. However, when the cobalt grain size is beyond this range, the grain size has no significant effect on the performance of FTs^{[22⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓~36]}. When the Co crystallite size is about 8 ~ 10 nm, the catalyst has the highest intrinsic activity.

Other factors such as metal confinement effect, support properties, crystal phase composition and reaction conditions can affect the properties of FTs, resulting in a certain deviation in the correlation between cobalt grain size and properties. However, based on the above analysis of the size effect of different catalytic systems, it can still be explained that the optimal grain size of cobalt-based catalysts is about 8 ~ 10 nm, which provides a reference for the development of high-performance cobalt-based FTs catalysts.

The relationship between the adsorption of each reaction intermediate on the surface of Co/SiO₂ and Co/CNF catalyst and the particle size of the active phase was studied by de Jong et al. Through steady-state isotope transient kinetic analysis (SSITKA)^[36]. When the cobalt particle size is larger than 6 nm, the residence time and surface coverage of CH_x, OH_x and CO intermediates do not change significantly. When the particle size of cobalt is less than 6 nm, a large number of CO molecules with irreversible bonding occupy the active sites of the ladder. The residence time of CO for reversible bonding decreases, while the residence time of CH_x and OH_x intermediates on the catalyst surface increases sharply. The surface coverage of the CH_x,OH_x and CO intermediates drops, while the H surface coverage rises, which leads to a decrease in TOF and an increase in methane selectivity. Further analysis showed that the space-time yield (STY) of :CH₄ and

C 5 +

per active site increased first and then stabilized with the grain size. According to Pestman et al., this is mainly due to the fact that CO needs to be cleaved at the B₅ site to form CH_x monomer, and then migrate to the surface to form CH₄.When the cobalt particle size is less than 6 nm, it is difficult to form B₅ sites or surface reconstruction, which is not conducive to CO dissociation to form CH_x monomers^[37]. Salmeron et al. Also confirmed that the dissociation efficiency of H₂ and Co molecules on larger CO nanoparticles (15 nm) is much higher than that on small nanoparticles (4 nm) using in situ X-ray absorption spectroscopy (XAS)^[38,39]. With the increase of cobalt crystallite size, the number of step sites increases, and the step sites are more favorable for CO cracking to form CH_x monomer, whether it is hydrogen-assisted dissociation or direct dissociation. Therefore, with the increase of cobalt grain size, the increase of CH_x species coverage increases the rate of CH₄ and

C 5 +

formation.

2.2 Crystal form effect of cobalt

For Co-based FTs catalysts, the active phase is metallic cobalt, which generally has two crystal structures: face-centered cubic (FCC) and hexagonal close-packed (HCP)^[40]. Davis et al. Found that compared with FCC CO, HCP Co catalyst had higher Co conversion activity, lower CH₄ selectivity and higher

C 5 +

selectivity^[41]. Khodakov et al. Used different reduction methods to generate active centers in two different crystal phases: (1) FCC Co phase obtained by pure hydrogen reduction^[42]; (2) FCC Co was obtained by hydrogen reduction, then HCP phase was obtained by Co carbonization, and then HCP Co was obtained by hydrogen reduction, which was mediated by HCP Co₂C to realize the transformation from FCC to HCP phase, also known as reduction-carbonization-reduction process (RCR)^[43]. Subsequently, Li et al. Also reported the difference in activity and selectivity between FCC Co and HCP Co, and the DFT theoretical calculation results further confirmed that HCP Co has higher inherent activity^[44]. In addition to the difference in activity, CO tends to dissociate directly on HCP Co crystals, while it is more prone to hydrogen-assisted dissociation of CO on FCC Co, resulting in a difference in the distribution of reaction products. On the one hand, due to their different crystallographic symmetries, HCP Co has a large number of active sites that FCC Co does not have; On the other hand, due to the different crystal morphology, the density of active sites exposed by the HCP Co crystal face is higher, as shown in Fig. 2.

On the basis of theoretical calculation, Wang et al. First reported the direct experimental evidence of cobalt crystal face effect in FTs reaction by preparing HCP Co and FCC Co catalysts through HCP CoO and FCC CoO intermediates by liquid phase synthesis method^[45]. The results show (as shown in Fig. 3) that the activation energy of Co on HCP Co is 101 kJ/mol, which is significantly lower than 144 kJ/mol on FCC Co, with a higher TOF value (2.1×10^-3s^-1). In order to further reduce the grain size of HCP Co and improve the utilization of active species, Wang et al. Further successfully supported HCP Co on carbon nanofibers by liquid phase synthesis, which reduced the grain size of HCP Co from 85. 3 nm to 23. 1 nm^[40]. Although both theoretical calculation and experimental results show that HCP Co is a better active phase for FTs catalyst, compared with FCC crystal phase, the thermodynamic stability of HCP crystal phase is poor, so it is difficult to prepare nanoparticles with particle size less than 25 nm, and it is almost impossible to prepare HCP CoO nanoparticles with particle size less than 15 nm^[46]. Therefore, how to efficiently and controllably synthesize HCP Co catalysts with smaller grain size based on the research of size effect and crystal face effect is still one of the future research directions of FTs.

2.3 Cobalt carbide active phase

The formation of cobalt carbide (Co₂C) crystal phase can sometimes be observed during the FTs reaction. The formation of Co₂C can lead to the deactivation of the catalyst in the reaction of FTs with long-chain alkanes as the main product^[47]. Khodakov et al. Found that the PtCo/Al₂O₃ catalyst showed a Co₂C crystal phase after long-term evaluation, and the CO conversion of the catalyst decreased from 60% to 20%^[48]. Further experiments showed that the catalyst had no CO conversion activity when the Co⁰ was completely carbonized into Co₂C crystal phase. However, Ding et al. Found that Higher Alcohol Synthesis could promote the generation of oxygenated compounds in the FTs reaction, and Co₂C-Co⁰ constituted a bifunctional active site for Higher Alcohol Synthesis (HAS)^[49,50]. Where Co₂C is the active site for non-dissociative adsorption of CO, Co⁰ is the dissociative adsorption and chain propagation site of CO, and the formation of oxygenates occurs at the interface between the two phases of Co₂C and Co⁰, as shown in Fig. 4. According to the DFT calculation results, the Co₂C(111) plane provides sufficient active sites for the non-dissociative adsorption of CO, and the CO molecules adsorbed on the Co₂C(111)) are more likely to insert into the CH₂ intermediate to form higher alcohols in the presence of the Co₂C-Co⁰ interface. Kruse's research further found that the appropriate addition of alkali metals, especially K, can promote the formation of Co₂C crystal phase, improve the selectivity of oxygenates, and reduce the selectivity of CH₄^[51]. Sun et al. Found that the addition of alkali metal is a necessary condition for cobalt-based catalysts to produce Co₂C crystal phase in syngas atmosphere, and Na can promote the adsorption and cracking of Co on the surface of CO, and promote the formation and stabilization of Co₂C crystal phase^[52]. However, only spherical Co₂C can be formed by introducing Na alone, while nanoprism Co₂C can be obtained in the Na-modified CoMn system after syngas treatment. They first reported that the quadrilateral nanoprism Tropsch to Olefin exhibited excellent catalytic performance in the direct production of light olefins (Fischer-Tropsch to Olefin, FTO) from syngas via FTs. Under mild reaction conditions (T = 250 ° C, P = 1 bar and GHSV=2000 mL·h^-1·

g c a t - 1

), the selectivity of light olefins (C_2-4=) was as high as 60.8% at a CO conversion of 32%, and the product distribution broke the traditional Anderson-Schulz-Flory (ASF) distribution. Compared with the spherical Co₂C, the prismatic Co₂C can expose specific (101) and (020) crystal planes. Further theoretical calculation results show that the cobalt carbide nanoparticles have a strong crystal plane effect, that is, the (101) plane can effectively promote the formation of olefins. At the same time, both (101) and (020) planes can inhibit the formation of methane.

In order to further reveal the growth mechanism of nano-prismatic Co₂C, Li et al. In our research group synthesized Ce-stabilized prismatic Co₂C nanoparticles based on their previous research work, and the additive Ce promoted the formation of Co₂C nano-prisms, which improved the selectivity of light olefins and inhibited the generation of CH₄^[53]. Both the experimental results and the theoretical calculation show that the promoter Ce provides oxygen to the dangling bonds on the surface of Co, reduces the surface energy of Co₂C, promotes the anisotropic growth of Co₂C crystals, and forms nano-prismatic Co₂C with exposed (101) and (020) high-index crystal planes, which has excellent FTO activity. Combined with various analysis and characterization methods such as HRTEM and XRD, various intermediate forms in the growth process of metal Co particles to prismatic Co₂C were identified, and the seven-step formation mechanism of Co₂C nanoprisms was proposed (Fig. 5), which provided theoretical guidance for the controllable preparation of high-performance FTO catalysts.

According to the existing literature, Co₂C is generally considered as an inactive phase in the FTs reaction with long chain alkanes as the main product, but in a specific system, Co₂C can be used as the insertion site of CO to produce oxygenated compounds such as ethanol efficiently, and prismatic Co₂C with exposed (101) and (020) crystal faces can be used as the active site to produce light olefins directly and efficiently through the FTs pathway. To sum up, FTs are structure-sensitive reactions, and their CO conversion activity and product selectivity have strong size effect and crystal plane effect.

3 Metal-support interaction

Supported catalysts are widely used in the field of heterogeneous catalysis because of their simple preparation methods, uniform dispersion of active components and promoters, and high mechanical strength^[54]. The carrier in the supported catalyst not only plays the role of dispersing and stabilizing the metal nanoparticles, but also interacts with the metal particles (MSI).This interaction often leads to interfacial charge transfer, changes in metal structure and molecular adsorption behavior, and then affects the activity, selectivity and stability of the catalyst. Therefore, the interaction between metal and support is an important area of heterogeneous catalysis^[55,56]. For traditional metal oxide supports, SiO₂ and Al₂O₃ are easy to induce the formation of difficult-to-reduce species such as silicate and cobalt aluminate spinel (CoAl₂O₄) after heat treatment (Fig. 6a), resulting in the reduction of the reduction degree of the active component. When TiO₂ is used as the support, the TiO_x species generated during the reduction process are easy to migrate and cover the surface of metal nanoparticles (NPs) with high surface energy (Fig. 6 B), which is called metal-support strong interaction (SMSI), inhibiting the reduction of metal oxides to some extent and blocking the active sites^[55]. Due to the surface inertness of carbon materials, the interaction between the supported metal oxide and the support is usually weak and easy to reduce, but the interaction between the reduced metal NPs and the carbon materials is weak (Fig. 6C), resulting in the easy agglomeration and growth of metal NPs during the reaction process^[57]. In order to balance the contradiction between the metal-support interaction and the functional integrity of the active metal, the previous research work mainly modified the surface of the support to achieve the purpose of regulating the metal-support interaction.

3.1 Carbon modification of oxide surface

Alumina is the most commonly used support material in industrial catalysts, which has many advantages, such as low cost, good thermal stability, high mechanical strength and easy molding. However, when it is applied to Co-based FTs catalyst, heat treatment is easy to induce the formation of difficult-to-reduce species such as CoAl₂O₄, and the introduction of inert components on the surface of Al₂O₃ to block the direct contact between Co species and Al₂O₃ is an effective means to inhibit the formation of species such as CoAl₂O₄. Christopher et al. Successfully introduced a certain amount of SiC inert components on the surface of alumina by adding SiC nanoparticles in Al₂O₃@Al^[58]. Compared with Al₂O₃, the interaction between metal Co and SiC is weaker, and the reduction degree and active site density of SiC-modified CoAC-15 catalyst are significantly improved.It has higher space-time yield of active metal Co (CTY=19.4×10^-6mol_CO·

g C o - 1

·s^-1) and intrinsic reaction activity (TOF=1.49×10^-2s^-1) and

C 5 +

selectivity (92.7%). When the mass ratio of SiC and Al is 0.15, the catalyst exhibits the highest active site density and TOF value. Lee et Al. Prepared Co/CN-Al catalyst by using the method of graphite-phase carbon nitride (g-C₃N₄) coating, and the g-C₃N₄ layer blocked the strong interaction between the metal and the Al₂O₃ support, inhibited the generation of difficult-to-reduce species such as CoAl₂O₄ spinel, and improved the reduction degree of the catalyst and the dispersion of metal Co on the surface of CN-Al^[59]. Compared with the conventional Co/Al₂O₃ catalyst, Co/CN-Al has higher activity for FTs. In addition, the g-C₃N₄ layer was pyrolyzed during the high temperature heat treatment to produce CO reducing gas, which reduced cobalt oxide to metallic cobalt, and the catalyst could be activated efficiently without further H₂ pretreatment. Recent studies have shown that coating a layer of carbon material on the surface of TiO₂ can effectively inhibit the migration of TiO_x species and weaken the strong metal-support interaction^[60,61]. As shown in Fig. 7, Li et al. Modified the surface of TiO₂ by hydrothermal carbonization using glucose as the carbon source to obtain the C-TiO₂ support^[60]. The surface of the hydrothermal carbon layer (HTC) has abundant oxygen-containing functional groups, such as — COOH, — OH, which provide sufficient anchoring and nucleation sites for cobalt species to achieve high dispersion of Co species on the surface of the C-TiO₂ support. Compared with Co/TiO₂(11.6nm), the Co/C-TiO₂ catalyst has a smaller Co crystallite size (8.6 nm) but exhibits a higher degree of reduction, breaking the dependence between the degree of reduction and crystallite size. Meanwhile, the Co/C-TiO₂ catalyst exhibited higher catalytic activity (CTY=2.5×10^-5mol_CO·

g C o - 1

·s^-1) and

C 5 +

selectivity (83.5%). Hong et al. Used urea as carbon and nitrogen sources to generate carbon nitride (CN) layer on the surface of TiO₂ in situ by pyrolysis at high temperature to regulate the surface properties of TiO₂ and metal-support interaction^[61]. On the one hand, carbon nitride can stabilize cobalt nanoparticles, improve the dispersion of Co, and prevent active sites from being coated by TiO₂, which significantly improves the stability of the catalyst while improving its activity. To sum up, modifying the surface of metal oxide support with carbon materials can effectively adjust the surface and interface properties of the catalyst, break through the limitation of dispersion and reduction to a certain extent, and balance the contradiction between activity and stability.

The carbon layer may cover the active metal sites and reduce the activity of the catalyst. During the reaction at high temperature and high pressure, the hydrogenolysis of carbon materials may also cause structural damage and loss. Ordomsky et al. Impregnated a SiO₂ with a certain amount of glucose and obtained a carbon modified support after pyrolysis at high temperature. The functional groups on the surface of the carbon material not only played a role in dispersing Co species, but also promoted the crystallization of cobalt oxide as nucleation centers^[62]. When the catalyst was calcined in air to remove the carbon layer, the grain size of cobalt oxide only increased from 5 ~ 6 nm to 7 ~ 8 nm, and the reduction degree of cobalt oxide was still 50%, and the metal-support interaction was weaker than that of the unmodified catalyst. Li et al. Synthesized a series of Ru/TiO₂-T catalysts through the dopamine sacrificial coating strategy, that is, after impregnating a certain amount of dopamine on the surface of Ru/TiO₂, the dopamine was first pyrolyzed at high temperature in an inert atmosphere to generate a carbon layer, and then further calcined in air atmosphere to remove the carbon layer, while Ru was oxidized to RuO₂, and the catalyst was activated by H₂ and then applied to the FTs reaction^[63]. This strategy effectively inhibited the migration of TiO_x species and the coating of Ru nanoparticles, reduced the formation of Ru-TiO₂ interface, significantly increased the number of Ru active sites exposed on the surface, and the catalyst had higher catalytic performance and stability for Fischer-Tropsch synthesis. Further study shows that the thermal stability of CO adsorbed on the surface exposed Ru sites is better, and the TOF value is 5.3 times higher than that of the Ru sites at the interface. In addition, the surface-exposed Ru site enhances the bond strength of the CH_x intermediate and the metal site, promotes chain propagation, improves

C 5 +

selectivity, and suppresses CH₄ production.

Yao et al. Found that nitrogen-doped carbon species, mainly containing pyrrole nitrogen, could be generated on the surface of Co/SiO₂ catalyst by a simple co-impregnation method using dicyandiamide as carbon and nitrogen sources^[64]. The formation of CN_x can improve the reducibility of the catalyst and transfer electrons from the N atom to the neighboring Co⁰ particle, which promotes the dissociation process of CO, so that Co/SiO₂-CN_x has a higher space-time yield and TOF value than Co/SiO₂. The addition of a certain amount of organic compounds, such as dicyandiamide and polyethylene glycol (PEG), during the impregnation of Co can significantly change the surface and interface properties of Co/Si

O 2 [64]

, CoRu/Si

O 2 [65]

, CoRuO_x/Al₂

O 3 [66]

, CoRu/ZSM-5 and other catalysts^[67]. Yao et al. Found that the addition of PEG in the co-impregnation process could significantly reduce the crystallite size of Co₃O₄ (8.4 nm), but its reduction degree was similar to that of large-particle Co₃O₄(21.5nm) prepared without PEG, which broke the reduction-dispersion dependence^[66]. In addition, when macroporous γ-Al₂O₃(21.2nm) was used as the support, the addition of polyethylene glycol (PEG) could reduce the particle size of Co₃O₄ to 4. 6 nm, which was significantly smaller than the pore size of the catalyst (14. 5 nm), breaking the dependence between pore size and particle size faced by the supported catalyst prepared by impregnation method.This structure enables the reduced catalyst to have both a high Co⁰ dispersion (31.6%) and a large pore size, which can therefore provide more Co⁰ active sites and enhance CO diffusion, significantly improving the CTY and

C 5 +

selectivity of the catalyst. Yao et al. Found that the addition of PEG in the catalyst preparation process could not only improve the reducibility of Co₃O₄ particles, but also promote the formation of RuCo bimetallic small particles, and the synergistic effect of promoter Ru and Co active sites weakened the adsorption of H species on the catalyst. Compared with Co/ZSM-5-P, CoRu/ZSM-5-P catalyst had higher selectivity for gasoline fraction hydrocarbons, and could break through the Anderson-Schulz-Flory distribution control of FTs products^[67]. The preparation method has also been successfully applied to Ni-based methanation catalysts^[68⇓~70].

3.2 Carbon material surface functionalization

Carbon materials have high specific surface area, excellent mass and heat transfer ability and good pore structure, but the interaction between graphite basal plane and deposited metal is weak Van der Waals force^[71]. Therefore, when FTs are used as supports, there is often a problem that the metal-support interaction is weak, which leads to the agglomeration and sintering of metal nanoparticles in the reaction process, and ultimately reduces the stability of the catalyst. Therefore, in the FTs system with carbon materials as supports, surface functionalization is generally used to improve the stability and reaction performance of the catalyst. There are many synthetic methods for functionalization of carbon materials, including heteroatom doping, surface oxidation and activation, grafting and sulfonation^[72,73].

One of the important means of surface functionalization of carbon materials is to integrate heteroatoms into the framework of carbon materials. On the one hand, the doping atom provides the anchor position of the active metal, and on the other hand, its unique electronic properties can control the electronic state of the active metal. Therefore, heteroatom doping not only directly affects the grain growth process, but also acts as a promoter of the catalyst. Nitrogen doping is the most representative heteroatom doping in carbon materials, and N atoms exist in various forms such as pyridine nitrogen, pyrrole nitrogen, and graphite nitrogen, as shown in fig. 8^[74]. The nitrogen atom acts as an electron donor to improve the FTs performance of the catalyst by strengthening the metal-support interaction and increasing the active metal dispersion.

Sun et al. Prepared a series of N-doped ordered mesoporous carbon materials (OMCs) with nitrogen content ranging from 6.5 wt% to 10.5 wt% by post-synthesis method using cyanamide as nitrogen source, and studied the effect of nitrogen doping on the catalytic performance of Co/NMC catalyst FTs^[27]. Various characterization results show that the doped nitrogen atoms act as anchor sites of cobalt species, which can enhance the metal-support interaction and significantly improve the dispersion of Co active components. With the decrease of grain size, the interaction between metal oxide and support is strengthened, which leads to the decrease of reduction degree of the catalyst, and the CTY of the nitrogen-doped catalyst shows a volcano-type curve distribution with the increase of N-doping amount. Similarly, Coville et al. Introduced nitrogen-containing functional groups on the surface of hollow carbon spheres through melamine pyrolysis, which enhanced the metal-support interaction and highly dispersed Co species (6.4 nm). Although the reduction degree of the catalyst was reduced to a certain extent, N doping inhibited the agglomeration and sintering of Co particles, and the activity and stability of the catalyst were significantly improved^[75].

In addition to the post-doping of existing carbon materials, a large number of studies have also tried to directly prepare nitrogen-containing carbon materials. Li et al. Prepared N-containing carbon nanospheres (NCS-T) by pyrolysis of polydopamine nanospheres at different temperatures. With the increase of pyrolysis temperature from 500 ℃ to 800 ℃, the content of pyrrole nitrogen in NCS-T support gradually decreased, while the content of pyridine nitrogen and graphite nitrogen increased, and the TOF value of its supported cobalt-based FTs catalyst decreased from 10.0×10^-3s^-1 to 7.6×10^-3s^-1^[74]. Compared with pyridine nitrogen and graphite nitrogen, pyrrole nitrogen has higher anchoring ability and can provide more active sites. In addition, DFT theoretical calculations show that both pyridine nitrogen and pyrrole nitrogen can make the local electronic States below the Fermi level, especially pyrrole nitrogen^[76]. Therefore, the pyrrole nitrogen induced electronic state has the highest energy level and charge density among all the N species. At the same time, the increase of electron cloud density of active metal CO further weakens the C — O bond, promotes the cleavage and activation of Co, and then improves the activity of FTs^[77].

It is another important way to enhance the metal-carbon interaction by endowing the surface of carbon materials with oxygen-containing functional groups and introducing defect sites through oxidizing solution or gas treatment. Acid treatment is mainly to introduce oxygen-containing groups (hydroxyl, carbonyl and a large number of carboxyl) and defects on the surface of carbon materials by wet chemical oxidation, which enhances the cation exchange between the metal and the functionalized carbon materials, so the introduction of oxygen-containing functional groups and defects can play a role in anchoring metal particles. Dalai et al. Functionalized carbon nanotubes (CNTs) by nitric acid pretreatment and used them as supports to prepare a series of supported cobalt-based FTs catalysts^[78]. It is found that higher nitric acid concentration and higher treatment temperature contribute to the formation of more oxygen-containing functional groups and defects on the CNT surface, which play a decisive role in the dispersion of cobalt, so the activity of the catalyst is also significantly improved with the increase of functional groups and defect sites on the CNT surface. Chernyak et al. Prepared different oxidized carbon nanotube supported cobalt catalysts by adjusting the oxidation time of nitric acid, and found that the oxidation time had a great influence on the texture properties of the support and there was an optimal oxidation condition^[22]. In addition, the crystallite size of metallic Co decreases with the increase of the oxidation degree of the support, and the catalyst with the highest degree of functionalization presents the best reaction performance for FTs. Chu et al. Used nitric acid to pretreat CNTs to prepare supported cobalt-based catalysts. The surface of the carrier treated by nitric acid formed more oxygen-containing functional groups, which made the catalyst show better FTs performance^[79]. This method has also been widely used in the regulation of metal-support interactions in FTs catalytic systems such as Fe and Ru^[80,81].

To sum up, N-doping and wet chemical surface oxidation are more widely used means for functionalization of carbon materials, in which the electron-donating effect of doped nitrogen species enriches metal electrons, thereby enhancing the interaction between metal and C carriers.CO adsorption and dissociation were promoted, and then the activity of FTs was improved. The effect of pyrrole nitrogen on CO hydrogenation activity was the most significant, followed by pyridine nitrogen, while the effect of graphite nitrogen was not obvious. The main function of acid pretreatment is to introduce oxygen-containing functional groups and defect sites on the surface of carbon materials, enhance the cation exchange strength between the metal and the support, and achieve the purpose of anchoring the active metal, thereby improving the dispersion of active species and enhancing the catalytic performance of FTs.

4 Confinement effect

It is an effective strategy to improve the stability of supported metal catalysts by using the encapsulation and confinement effect of the support to inhibit the migration and agglomeration of metal particles. The effect of confinement on the performance of Co-based FTs catalyst is mainly divided into geometric effect and electronic effect. The geometric space effect is mainly reflected in two aspects: one is that the pore structure significantly affects the diffusion form, path and rate of reactants and intermediates, as well as the accessibility of reactant molecules to the active sites in the pore; Secondly, the confined structure provides a physical barrier for nanoparticles to inhibit the migration of metal atoms and prevent the agglomeration and growth of particles. In addition to the geometric effect, the confinement effect of carbon materials is also reflected in a certain electronic effect. On the one hand, it affects the interaction between carbon materials and H₂ and CO molecules, and changes the local gas concentration. At the same time, the electronic properties of the active metal are directly affected by the electron transfer between the metal and the support.

4.1 Co @ MO catalyst (MO: Metallic oxide)

In general, Co nanoparticles tend to gradually aggregate and agglomerate with the FTs reaction, resulting in a decrease in catalytic activity. The confined structure of the catalyst can stabilize and anchor the metal particles, which is an important means to overcome the contradiction between activity and stability. Li et al. First prepared a series of Co₃O₄ nanoparticles with different particle sizes by hydrothermal method using tetradecyltrimethylammonium bromide (TTAB) as a capping agent.Then a layer of tetraethoxysilane (anion) was coated on the surface of the TTAB (cation) -terminated Co₃O₄ nanoparticles by using the electrostatic interaction between the anion and cation particles, and the SiO₂ coated cobalt-based catalyst with a "watermelon seed" structure was obtained after the TTAB was removed by calcination at high temperature (as shown in Figure 9)^[29]. The catalyst prepared by the method has uniform size distribution of Co nanoparticles, and the spatial confinement structure can effectively inhibit the agglomeration and sintering of Co particles in the FTs reaction process. In addition, the contact time between the reaction intermediate and the active site in the confined space is prolonged, which further promotes the formation of long-chain hydrocarbons. With the increase of cobalt crystallite size from 7.2 nm to 11.4 nm, the TOF value of the catalyst increased from 3.9×10^-2s^-1 to 6.4×10^-2s^-1, and the FTs product changed from diesel fraction (66.2%) to gasoline fraction (62.4%). It is worth emphasizing that compared with the large-sized Co particles, the small-sized Co particles can strongly adsorb a large number of C^* species, so that there is a high enough coverage of CH_x in the confined space to promote the chain growth of the reaction through the carbide mechanism. Therefore, the smaller the grain size, the larger the chain growth factor, and the higher the heavy hydrocarbon selectivity.

The traditional preparation methods of Co-based FTs catalysts (such as wet impregnation, deposition precipitation and ion adsorption) are insufficient in controlling the micro-morphology of metal Co. Li et al. Found that the core-shell structure can well stabilize the single crystal cobalt core and expose the highly active HCP Co (1011) crystal face during the reaction^[82]. With p-Co(Co₃O₄Nanoplates) as the inner core, the SiO₂ and C shells were introduced by hydrolysis of tetraethyl orthosilicate (TEOS) and pyrolysis of polyvinylpyrrolidone (PVP) to form a core-shell type Co@C-SiO₂ catalyst, as shown in Fig. 10. The reduction of Co₃O₄ to metallic Co is accompanied by a drastic volume contraction, and when the inward contraction of particles is pulled by the strong interaction in the opposite direction, cobalt single crystals split into polycrystalline particles. The presence of the carbon layer endows the catalyst with a moderate metal-support interaction, which makes the Co core shrink as a whole during the phase transition, thus maintaining the single crystal structure. At the same time, the presence of the carbon layer physically blocks the direct contact between Co and SiO₂, avoiding the formation of difficult-to-reduce cobalt silicate and other species. Further study showed that the silica shell had a decisive effect on the stability of the catalyst, and only when the silica and the intermediate carbon layer existed at the same time, the Co single crystal core with exposed highly active (10-11) crystal face could be maintained during the reaction. The Co@C-SiO₂ catalyst not only provides a highly active crystal face, but also suppresses the splitting and agglomeration of single crystal Co during the reaction, thus exhibiting excellent catalytic activity (TOF=4.0×10^-2s^-1) and stability (TOS = 168 H), as well as higher selectivity to

C 5 +

(88.9%) and lower selectivity to methane (5.3%). Li et al. Prepared core-shell Co@SiO₂-PVP nanocatalysts by in-situ coating method. Although the silica coating reduced the reduction degree of the catalyst, it enhanced the dispersion and stability of metal Co, and the catalytic activity did not decrease^[83]. Compared with the Co/SiO₂-ACT catalyst, the Co@SiO₂-PVP catalyst with core-shell structure has lower selectivity for

C 19 +

and higher selectivity for methane, and the SiO₂ shell not only restricts the diffusion of reactants, but also hinders the secondary hydrogenation of olefins and shifts the reaction products to light hydrocarbons. To sum up, the confinement effect of metal oxides is mainly reflected in the inhibition of metal particle agglomeration and the modulation of molecular diffusion performance, that is, the geometric space effect, and the regulation of metal-support interaction is very limited.The coupling of metal oxide and carbon layer can endow the catalyst with moderate metal-support interaction, which can further improve the FTs catalytic performance of the confined catalyst.

4.2 Co @ CNTs catalyst

Carbon nanotubes (CNTs) are hollow tubes composed of graphene layers, and the curling of graphene layers changes the electron cloud density of large π bonds, resulting in the enrichment of electrons on the outer surface of CNTs and the lack of electrons on the inner surface^[84]. On the one hand, the electron-deficient inner surface can enhance the interaction of carbon materials with H₂ and CO molecules, and the interaction with CO molecules is stronger. This effect can change the local concentration of gas-phase molecules on the surface of the catalyst, resulting in a lower H₂/CO inside the CNTs than in the outer gas phase^[85]. On the other hand, the electron transfer effect between the metal and the support can directly affect the electronic properties of the metal and change the adsorption and activation ability of the metal to the reactant molecules^[77]. Therefore, the confinement effect of CNTs is not only the geometric effect, but also the electronic effect.

Li et al. Found that the spatial confinement structure of CNTs inhibited the grain growth, and the grain size of Co₃O₄ decreased with the decrease of the inner diameter of CNTs^[28]. Dalai et al. Found that after 480 H of FTs reaction, the size of cobalt particles inside the CNTs tube remained at about 8 nm, while the Co NPs outside the tube were severely sintered, with the average particle size increasing from 11.5 nm to 25 nm^[86]. In addition, the inner cavity of CNTs can also change the redox properties of metal oxides. Li et al. Used three kinds of CNTs with average pore diameters of 5, 11 and 17 nm as carriers to prepare supported cobalt-based catalysts. The H₂-TPR results showed that the Co₃O₄ particles inside the tube were easier to be reduced than those outside the tube, and the FTs activity was higher. The reduction ability decreased with the increase of CNT diameter, which was mainly due to the decrease of electron transfer ability on the surface of non-planar carbon nanotubes caused by the increase of CNT diameter^[87]. How to encapsulate metal particles into CNTs with high selectivity is one of the main challenges in the development of the above catalysts. Akbarzadeh et al. Treated CNTs with nitric acid and high temperature to produce more defects, which can be used as the anchoring of Co particles to control the Co content and grain size in CNTs channels^[88]. When the heat treatment temperature was increased from 600 ℃ to 900 ℃, the Co content in the CNTs channels increased from 31% to 70%.With the increase of Co content in the channels, the catalyst showed higher reaction activity,

C 5 +

product selectivity and stability. Although CNTs materials have good application prospects in the field of Fischer-Tropsch synthesis, the high preparation cost limits their large-scale industrial application.

4.3 MOF-derived catalyst

The effect of metal loading on the activity of the catalyst is very important. The maximum metal loading of the catalyst prepared by impregnation method is generally less than 30 wt%. It is of great significance to develop alternative preparation methods to obtain FTs catalysts with high metal loading and optimal particle size distribution. MOF materials are a class of novel crystalline porous materials with tunable structure and function, which are formed by coordination self-assembly of metal clusters/ions and ligands. They have special three-dimensional pore structures and have been widely used in catalysis, biomimetics, fluorescence, chemical sensing, gas adsorption and separation, and other fields^[89]. Therefore, MOF-derived carbon-based materials with controllable morphology and pore structure are easy to be doped with a variety of heteroatoms through the selection of organic ligands, which has become one of the current research hotspots^[90].

Gascon et al. First reported the preparation of FTs catalyst by MOF derivation strategy in 2015, and found that the Fe @ C catalyst prepared by Basolite F300 (Fe-BTC) pyrolysis showed excellent catalytic activity and stability for Fischer-Tropsch synthesis^[91]. The introduction of heteroatom N by nitrogen-containing organic ligands can further improve the catalytic performance of FTs^[92]. Subsequently, Co-MOF series with different topologies were applied to the preparation and structure-activity relationship study of FTs catalysts, as shown in Fig. 11^{[93⇓⇓~96]}. Isaeva et Al. Loaded Co nanoparticles into the microporous channels of MIL-53 (Al) to prepare a novel Co @ MIL-53 (Al) FTs catalyst. The framework structure of MOF significantly inhibited the migration and agglomeration of Co nanoparticles. Compared with the traditional Co/Al₂O₃, Co @ MIL-53Al exhibited higher

C 5 +

selectivity and lower CH₄ selectivity^[97]. In order to increase the internal and external surfaces of MOF materials to enhance their adsorption properties and shorten the diffusion path of reactants and products, Isaeva et Al. Used microwave irradiation to synthesize nanoscale metal-organic framework MIL-53 (Al). Compared with micron-scale Co @ MIL-53 (Al), nanoscale Co @ MIL-53Al catalyst has higher activity and product selectivity^[98].

Limited by the thermal stability of MOF materials, more research work has chosen MOF as a self-sacrificial template to prepare coated cobalt-based FTs catalysts by high temperature pyrolysis. Recently, ZIF-67-derived Co @ C catalysts with different Co crystallite sizes (8.4 ± 1.9 nm to 74.8 ± 11.6 nm) were used for size effect studies^[25]. The results show that the growth process of Co grains in the ZIF-67-derived catalyst is more controllable, and the grain size distribution is more uniform. Meanwhile, the physical barrier of the carbon layer inhibits the agglomeration and reoxidation of Co grains during the FTs reaction, showing excellent stability. Gascon et al. Utilized the above characteristics of MOF derivatives to prepare the Co@SiO₂ catalyst by coupling the ZIF-67 material with SiO₂ through hydrolysis, pyrolysis and high temperature calcination using tetraethyl orthosilicate (TEOS) as the silicon source, which can still maintain a small grain size (12.3 nm) and a high degree of reduction (DOR: 78%) with a Co loading up to 51%. The catalyst preparation process is shown in Figure 12^[96]. The confinement effect plays a crucial role in the control of Co crystallite size during the pyrolysis of ZIF-67. Compared with the catalyst prepared by the impregnation method, the Co@SiO₂ catalyst has a higher density of active center sites, showing higher FTs activity,

C 5 +

selectivity and stability. In addition, Li et al. Impregnated different amounts of TEOS on the surface of Co-MOF-71, and obtained Co-Si @ C catalysts with high density and dispersion of Co active sites by adjusting the addition of SiO₂.Meanwhile, the incorporation of Si also promoted the partial conversion of Co nanoparticles to hcp Co through the Co₂C intermediate, and the above catalysts showed good FTs activity and outstanding diesel selectivity, with a space-time yield of

C 5 +

reaching 1.45 G ·

g c a t - 1

·h^-1, which is significantly higher than that of the traditional supported catalyst^[94].

MOF-derived catalysts have unique advantages in heteroatom doping, and Co @ NC and Co @ C were prepared by Zou et al. Using nitrogen-rich ZIF-67 and nitrogen-free Co-MOF-74 as self-sacrificial templates^[93]. Co @ C catalyst has wider mesopore distribution, which is beneficial to gas diffusion and mass transfer, and has higher Co conversion (30%) and

C 5 +

selectivity (65%), while Co @ NC catalyst has relatively small pore size, which on the one hand hinders the migration of long-chain hydrocarbons,On the other hand, N acts as an electron-donating group to enhance the electron cloud density around cobalt and promote the formation of short-chain hydrocarbons, so the CO @ NC catalyst shows lower Co conversion (10%) and

C 5 +

selectivity (31%), but higher C₂-C₄ selectivity (37%).

During the high-temperature pyrolysis of Co-MOF, Co can catalyze the graphitization of derived carbon materials, which may lead to the graphitized carbon layer around Co nanoparticles is not porous, which seriously hinders the diffusion of reactants and the contact with active sites.In order to overcome the above shortcomings, Li et al. Developed a new MOF derivation strategy, using chemical vapor deposition of acetylene to control the reduction of Co nanoclusters and the spontaneous formation of porous carbon shells in Co-MOF.In that prepare Co @ C catalyst, the crystal size of the metal Co can reach 4 to 6 nm, the crystal phase is between hexagonal close packing and face-centered cubic, and the carbon lay has a porous structure, which is beneficial to the diffusion of reactants and products,The catalyst containing both FCC and HCP Co exhibited the highest FTs activity (CTY=254.1~312.1μmol_CO·

g C o - 1

·s^-1) and hydrocarbon yield (STY=4.0~5.2 kg_HC·k

g c a t - 1

·h^-1),The selectivity of

C 5 +

is about 70% ~ 80%^[95]. The unique structural properties, geometric steric effects and electronic effects of MOF and its derivative materials provide a new way for the development and application of Co-based FTs catalysts with high active site density and high stability.

5 Conclusion and prospect

In order to solve the common problem of "dispersion-reduction dependence" of cobalt-based FTs catalysts, this paper reviews the latest research progress in the regulation of interfacial properties of cobalt-based FTs catalysts in recent years from the aspects of structural sensitivity, metal-support interaction and confinement effect. With the development of nanotechnology in catalyst preparation and characterization, significant progress has been made in the understanding of the regulation of the surface and interface properties of FTs catalysts.However, there are still a lot of related problems to be further explored and solved: (1) Although both theoretical calculations and experimental results show that HCP Co has better catalytic performance than FCC Co for FTs, its thermodynamic stability is poor, and the efficient and controllable synthesis of HCP Co catalysts with small grain size is still facing great challenges; (2) Carbon materials have attracted much attention due to their controllable surface chemical properties and excellent thermal conductivity. However, compared with metal oxide supports, carbon materials generally have low mechanical strength, and may decompose through hydrogenolysis reaction in the reaction process at high temperature and high pressure, resulting in the loss of heteroatoms or amorphous transformation.In addition, the defect sites and functional groups on carbon materials are very complex, so it is still difficult to reveal their effects on catalytic performance at the atomic level. (3) How to realize the composite of carbon materials and metal oxides, on the one hand, weaken the interaction between metal precursors and supports, promote the reduction of catalysts, improve the dispersion and intrinsic activity of catalysts, but at the same time maintain high mechanical and chemical stability, is still a field and direction worthy of exploration and in-depth study in the future. (4) that in situ reduction of active specie is realized through the pyrolysis of organic complex material such as MOFs, the dependence of the reduction degree on the dispersion degree can be broken through, the limitation of the reduction degree on the utilization rate of the active species can be eliminated,The proof of concept has been completed in the development of Fischer-Tropsch synthesis catalysts with high active site density, which has important theoretical guiding significance, but the high price and low mechanical strength of MOFs materials are important challenges limiting their industrial large-scale application.

References

Publishing order | Descend order by publishing year | Descend order by cited within

[1]	Xu J, Yang Y, Li Y W. Fuel, 2015, 152: 122.

[2]	Khodakov A Y, Chu W, Fongarland P. Chem. Rev., 2007, 107(5): 1692.

[3]	de Smit E, Weckhuysen B M. Chem. Soc. Rev., 2008, 37(12): 2758.

[4]	Chen Y P, Wei J T, Duyar M S, Ordomsky V V, Khodakov A Y, Liu J. Chem. Soc. Rev., 2021, 50(4): 2337.

[5]	van Santen R A, Markvoort A J, Filot I A W, Ghouri M M, Hensen E J M. Phys. Chem. Chem. Phys., 2013, 15(40): 17038.

[6]	Savost’yanov A P, Yakovenko R E, Sulima S I, Bakun V G, Narochnyi G B, Chernyshev V M, Mitchenko S A. Catal. Today, 2017, 279: 107.

[7]	Xu R, Hou C P, Xia G F, Sun X, Li M F, Nie H, Li D D. Catal. Today, 2020, 342: 111.

[8]	Eschemann T O, Oenema J, de Jong K P. Catal. Today, 2016, 261: 60.

[9]	Nabaho D, Hans Niemantsverdriet J W, Claeys M, van Steen E. Catal. Today, 2016, 261: 17.

[10]	Guo S P, Niu C C, Ma Z Y, Wang J G, Hou B, Jia L T, Li D B. ChemCatChem, 2021, 13(5): 1375.

[11]	Guo Q, Huang J, Qian W X, Zhang H T, Ma H F, Ying W Y. Catal. Lett., 2018, 148(9): 2789.

[12]	Zhao N, Chen Y, Li X, Nisa M U, Jiang X N, Dai L Y, Li Z H. Appl. Surf. Sci., 2021, 570: 151127.

[13]	Du H, Jiang M, Zhao Z A, Li Y H, Liu T, Zhu H J, Zhang Z C, Ding Y J. Catal. Lett., 2021, 151(12): 3632.

[14]	Nakhaei Pour A, Housaindokht M R, Kamali Shahri S M. Ind. Eng. Chem. Res., 2018, 57(41): 13639.

[15]	Wang B, Han Y Y, Chen S F, Zhang Y H, Li J L, Hong J P. Catal. Today, 2020, 355: 10.

[16]	Ghogia A C, Nzihou A, Serp P, Soulantica K, Pham Minh D. Appl. Catal. A Gen., 2021, 609: 117906.

[17]	Shi Y M, Yao N, Lu C S, Fan Y B, Liu H Z, Li X N. Prog. Chem., 2009, 21(10): 2044.

[18]	石玉梅, 姚楠, 卢春山, 范洋波, 刘化章, 李小年. 化学进展, 2009, 21(10): 2044. ).

[19]	Boudart M, McDonald M A. J. Phys. Chem., 1984, 88(11): 2185.

[20]	Khadzhiev S N. Pet. Chem., 2016, 56(6): 465.

[21]	Chen W, Lin T J, Dai Y Y, An Y L, Yu F, Zhong L S, Li S G, Sun Y H. Catal. Today, 2018, 311: 8.

[22]	Bezemer G L, Bitter J H, Kuipers H P C E, Oosterbeek H, Holewijn J E, Xu X D, Kapteijn F, van Dillen A J, de Jong K P. J. Am. Chem. Soc., 2006, 128(12): 3956.

[23]	Chernyak S A, Suslova E V, Ivanov A S, Egorov A V, Maslakov K I, Savilov S V, Lunin V V. Appl. Catal. A Gen., 2016, 523: 221.

[24]	Xiong H F, Motchelaho M A M, Moyo M, Jewell L L, Coville N J. J. Catal., 2011, 278(1): 26.

[25]	Eschemann T O, Lamme W S, Manchester R L, Parmentier T E, Cognigni A, Rønning M, de Jong K P. J. Catal., 2015, 328: 130.

[26]	Luo Q X, Guo L P, Yao S Y, Bao J, Liu Z T, Liu Z W. J. Catal., 2019, 369: 143.

[27]	Yang Y F, Jia L T, Hou B, Li D B, Wang J G, Sun Y H. J. Phys. Chem. C, 2014, 118(1): 268.

[28]	Yang Y F, Jia L T, Hou B, Li D B, Wang J G, Sun Y H. ChemCatChem, 2014, 6(1): 319.

[29]	Fu T J, Lv J, Li Z H. Ind. Eng. Chem. Res., 2014, 53(4): 1342.

[30]	Cheng Q P, Tian Y, Lyu S S, Zhao N, Ma K, Ding T, Jiang Z, Wang L H, Zhang J, Zheng L R, Gao F, Dong L, Tsubaki N, Li X G. Nat. Commun., 2018, 9: 3250.

[31]	Melaet G, Lindeman A E, Somorjai G A. Top. Catal., 2014, 57(6/9): 500.

[32]	Yang Y F, Jia L T, Hou B, Li D B, Wang J G, Sun Y H. Catal. Sci. Technol., 2014, 4(3): 717.

[33]	Fischer N, van Steen E, Claeys M. J. Catal., 2013, 299: 67.

[34]	Zhong M, Wang J G, Chen C B, Ma Z C, Jia L T, Hou B, Li D B. Catal. Sci. Technol., 2019, 9(21): 6037.

[35]	Prieto G, Martínez A, ConcepciÓn P, Moreno-Tost R. J. Catal., 2009, 266(1): 129.

[36]	van Deelen T W, Su H, Sommerdijk N A J M, de Jong K P. Chem. Commun., 2018, 54(20): 2530.

[37]	den Breejen J P, Sietsma J R A, Friedrich H, Bitter J H, de Jong K P. J. Catal., 2010, 270(1): 146.

[38]	Pestman R, Chen W, Hensen E. ACS Catal., 2019, 9(5): 4189.

[39]	Herranz T, Deng X Y, Cabot A, Guo J G, Salmeron M. J. Phys. Chem. B, 2009, 113(31): 10721.

[40]	Tuxen A, Carenco S, Chintapalli M, Chuang C H, Escudero C, Pach E, Jiang P, Borondics F, Beberwyck B, Alivisatos A P, Thornton G, Pong W F, Guo J H, Perez R, Besenbacher F, Salmeron M. J. Am. Chem. Soc., 2013, 135(6): 2273.

[41]	Lyu S, Peng B, Kuang T, RappÉ K G, Zhang Y H, Li J L, Wang L. ACS Appl. Nano Mater., 2019, 2(4): 2266.

[42]	Gnanamani M K, Jacobs G, Shafer W D, Davis B H. Catal. Today, 2013, 215: 13.

[43]	Karaca H, Safonova O V, Chambrey S, Fongarland P, Roussel P, Griboval-Constant A, Lacroix M, Khodakov A Y. J. Catal., 2011, 277(1): 14.

[44]	Tsakoumis N E, Patanou E, Lögdberg S, Johnsen R E, Myrstad R, van Beek W, Rytter E, Blekkan E A. ACS Catal., 2019, 9(1): 511.

[45]	Liu J X, Su H Y, Sun D P, Zhang B Y, Li W X. J. Am. Chem. Soc., 2013, 135(44): 16284.

[46]	Lyu S, Wang L, Zhang J H, Liu C, Sun J M, Peng B, Wang Y, RappÉ K G, Zhang Y H, Li J L, Nie L. ACS Catal., 2018, 8(9): 7787.

[47]	Deng X, Yang D Z, Tan G G, Li X H, Zhang J W, Liu Q F, Zhang H L, Mellors N J, Xue D S, Peng Y. Nanoscale, 2014, 6(22): 13710.

[48]	Li X, Liu X W, Jiang D, Wen X D. Journal of Fuel Chemistry and Technology, 2018, 46(4): 459.

[49]	李啸, 刘兴武, 姜东, 温晓东. 燃料化学学报, 2018, 46(4): 459.).

[50]	Karaca H, Hong J P, Fongarland P, Roussel P, Griboval-Constant A, Lacroix M, Hortmann K, Safonova O V, Khodakov A Y. Chem. Commun., 2010, 46(5): 788.

[51]	Zhao Z A, Lu W, Yang R O, Zhu H J, Dong W D, Sun F F, Jiang Z, Lyu Y, Liu T, Du H, Ding Y J. ACS Catal., 2018, 8(1): 228.

[52]	Pei Y P, Liu J X, Zhao Y H, Ding Y J, Liu T, Dong W D, Zhu H J, Su H Y, Yan L, Li J L, Li W X. ACS Catal., 2015, 5(6): 3620.

[53]	Xiang Y Z, Kruse N. Nat. Commun., 2016, 7: 13058.

[54]	Zhang C, Li S G, Zhong L S, Sun Y H. J. Phys. Chem. C, 2021, 125(11): 6061.

[55]	Li Z J, Yu D M, Yang L Y, Cen J, Xiao K, Yao N, Li X N. ACS Catal., 2021, 11(5): 2746.

[56]	Pardo-Tarifa F, Cabrera S, Sanchez-Dominguez M, Boutonnet M. Int. J. Hydrog. Energy, 2017, 42(15): 9754.

[57]	Bertella F, ConcepciÓn P, Martínez A. Catal. Today, 2017, 289: 181.

[58]	Wolf M, Fischer N, Claeys M. Chem Catal., 2021, 1(5): 1014.

[59]	Chen W Y, Ji J, Duan X Z, Qian G, Li P, Zhou X G, Chen D, Yuan W K. Chem. Commun., 2014, 50(17): 2142.

[60]	Matsubu J C, Zhang S Y, DeRita L, Marinkovic N S, Chen J G, Graham G W, Pan X Q, Christopher P. Nat. Chem., 2017, 9(2): 120.

[61]	Park H, Kim K Y, Youn D H, Choi Y H, Kim W Y, Lee J S. ChemCatChem, 2017, 9(21): 4098.

[62]	Liu C C, He Y, Wei L, Zhang Y H, Zhao Y X, Hong J P, Chen S F, Wang L, Li J L. ACS Catal., 2018, 8(2): 1591.

[63]	Hong J P, Wang B, Xiao G Q, Wang N, Zhang Y H, Khodakov A Y, Li J L. ACS Catal., 2020, 10(10): 5554.

[64]	Cheng K, Subramanian V, Carvalho A, Ordomsky V V, Wang Y, Khodakov A Y. J. Catal., 2016, 337: 260.

[65]	Lyu S S, Cheng Q P, Liu Y H, Tian Y, Ding T, Jiang Z, Zhang J, Gao F, Dong L, Bao J, Ma Q X, Yang Q H, Li X G. Appl. Catal. B Environ., 2020, 278: 119261.

[66]	Li S, Yao N, Zhao F D, Li X N. Catal. Sci. Technol., 2016, 6(7): 2188.

[67]	Han Q, Yao N, Shi Y M, Ma H F, Li X N. Catal. Commun., 2012, 22: 52.

[68]	Ma C C, Yao N, Han Q, Li X N. Chem. Eng. J., 2012, 191: 534.

[69]	Yao M, Yao N, Shao Y, Han Q, Ma C C, Yuan C K, Li C G, Li X N. Chem. Eng. J., 2014, 239: 408.

[70]	Wang J, Yao N, Liu B, Cen J, Li X N. Chem. Eng. J., 2017, 322: 339.

[71]	Yuan C K, Yao N, Wang X D, Wang J G, Lv D Y, Li X N. Chem. Eng. J., 2015, 260: 1.

[72]	Liu B, Yao N, Li S, Wang J, Lv D Y, Li X N. Chem. Eng. J., 2016, 304: 476.

[73]	Dai H J. Surf. Sci., 2002, 500(1/3): 218.

[74]	Stein A, Wang Z Y, Fierke M A. Adv. Mater., 2009, 21(3): 265.

[75]	Lin Q W, Zhang J, Lv W, Ma J B, He Y B, Kang F Y, Yang Q H. Small, 2020, 16(15): 1902603.

[76]	Cheng Q P, Zhao N, Lyu S S, Tian Y, Gao F, Dong L, Jiang Z, Zhang J, Tsubaki N, Li X G. Appl. Catal. B Environ., 2019, 248: 73.

[77]	Dlamini M W, Phaahlamohlaka T N, Kumi D O, Forbes R, Jewell L L, Coville N J. Catal. Today, 2020, 342: 99.

[78]	Li X Y, Pan X L, Yu L, Ren P J, Wu X, Sun L T, Jiao F, Bao X H. Nat. Commun., 2014, 5: 3688.

[79]	Lu J Z, Yang L J, Xu B L, Wu Q, Zhang D, Yuan S J, Zhai Y, Wang X Z, Fan Y N, Hu Z. ACS Catal., 2014, 4(2): 613.

[80]	Vosoughi V, Badoga S, Dalai A K, Abatzoglou N. Ind. Eng. Chem. Res., 2016, 55(21): 6049.

[81]	Zhang H, Lancelot C, Chu W, Hong J P, Khodakov A Y, Chernavskii P A, Zheng J, Tong D G. J. Mater. Chem., 2009, 19(48): 9241.

[82]	Cheng K, Ordomsky V V, Legras B, Virginie M, Paul S, Wang Y, Khodakov A Y. Appl. Catal. A Gen., 2015, 502: 204.

[83]	Kang J C, Zhang S L, Zhang Q H, Wang Y. Angew. Chem. Int. Ed., 2009, 48(14): 2565.

[84]	Qin C, Hou B, Wang J G, Wang G, Ma Z Y, Jia L T, Li D B. ACS Appl. Mater. Interfaces, 2019, 11(37): 33886.

[85]	Zeng B, Hou B, Jia L T, Li D B, Sun Y H. J. Mol. Catal. A Chem., 2013, 379: 263.

[86]	Yang F, Deng D H, Pan X L, Fu Q, Bao X H. Natl. Sci. Rev., 2015, 2(2): 183.

[87]	Guan J, Pan X L, Liu X, Bao X H. J. Phys. Chem. C, 2009, 113(52): 21687.

[88]	Tavasoli A, TrÉpanier M, Dalai A K, Abatzoglou N. J. Chem. Eng. Data, 2010, 55(8): 2757.

[89]	Xie W, Zhang Y H, Liew K, Li J L. Sci. China Chem., 2012, 55(9): 1811.

[90]	Akbarzadeh O, Mohd Zabidi N, Abdul Wahab Y, Hamizi N, Chowdhury Z, Merican Aljunid Merican Z, Ab Rahman M, Akhter S, Rasouli E, Johan M. Symmetry, 2018, 10(11): 572.

[91]	Bao J, Yang G H, Yoneyama Y, Tsubaki N. ACS Catal., 2019, 9(4): 3026.

[92]	Yang W P, Li X X, Li Y, Zhu R M, Pang H. Adv. Mater., 2018: 1804740.

[93]	Wezendonk T A, Santos V P, Nasalevich M A, Warringa Q S E, Dugulan A I, Chojecki A, Koeken A C J, Ruitenbeek M, Meima G, Islam H U, Sankar G, Makkee M, Kapteijn F, Gascon J. ACS Catal., 2016, 6(5): 3236.

[94]	An B, Cheng K, Wang C, Wang Y, Lin W B. ACS Catal., 2016, 6(6): 3610.

[95]	Qiu B, Yang C, Guo W H, Xu Y, Liang Z B, Ma D, Zou R Q. J. Mater. Chem. A, 2017, 5(17): 8081.

[96]	Pei Y P, Li Z, Li Y W. Aiche J., 2017, 63(7): 2935.

[97]	Zhang C H, Guo X X, Yuan Q C, Zhang R L, Chang Q, Li K, Xiao B, Liu S Y, Ma C P, Liu X, Xu Y Q, Wen X D, Yang Y, Li Y W. ACS Catal., 2018, 8(8): 7120.

[98]	Sun X H, Suarez A I O, Meijerink M, van Deelen T, Ould-Chikh S, Zecevic J, de Jong K P, Kapteijn F, Gascon J. Nat. Commun., 2017, 8: 1680.

[99]	Isaeva V I, Eliseev O L, Kazantsev R V, Chernyshev V V, Davydov P E, Saifutdinov B R, Lapidus A L, Kustov L M. Dalton Trans., 2016, 45(30): 12006.