Effects of Structure and Particle Size of Iron, Cobalt and Ruthenium Catalysts on Fischer–Tropsch Synthesis

This review emphasizes the importance of the catalytic conversion techniques in the production of clean liquid and hydrogen fuels (XTF) and chemicals (XTC) from the carbonaceous materials including coal, natural gas, biomass, organic wastes, biogas and CO2. Dependence of the performance of Fischer–Tropsch Synthesis (FTS), a key reaction of the XTF/XTC process, on catalyst structure (crystal and size) is comparatively examined and reviewed. The contribution illustrates the very complicated crystal structure effect, which indicates that not only the particle type, but also the particle shape, facets and orientation that have been evidenced recently, strongly influence the catalyst performance. In addition, the particle size effects over iron, cobalt and ruthenium catalysts were carefully compared and analyzed. For all Fe, Co and Ru catalysts, the metal turnover frequency (TOF) for CO hydrogenation increased with increasing metal particle size in the small size region i.e., less than the size threshold 7–8 nm, but was found to be independent of particle size for the catalysts with large particle sizes greater than the size threshold. There are some inconsistencies in the small particle size region for Fe and Ru catalysts, i.e., an opposite activity trend and an abnormal peak TOF value were observed on a Fe catalyst and a Ru catalyst (2 nm), respectively. Further study from the literature provides deeper insights into the catalyst behaviors. The intrinsic activity of Fe catalysts (10 nm) at 260–300 ◦C is estimated in the range of 0.046–0.20 s−1, while that of the Co and Ru catalysts (7–70 nm) at 220 ◦C are 0.1 s−1 and 0.4 s−1, respectively.


Introduction
To supply vast energy needs in the world while meeting more stringent environmental regulation to reduce the greenhouse gas emission, the global energy structure dominated by the fossil fuels such as oil, coal and natural gas is required to turn to low carbon system in the next thirty years [1]. For example, based on the US Energy Information Administration (EIA), energy consumption data (see Figure 1), world fossil fuels continuously dominate accounting for 77-84% of energy consumptions by 2040 and 71.5% by 2050 through the reduction of oil and coal usage by 3-7% points. Meanwhile, the renewable energy including nuclear and hydro energy accounts for 17% in 2016 and increase to 22.3% and 28% by 2040 and 2050, respectively. At the present consumption rate, it produces about 30 billion tons of CO 2 per year, mainly from power plants and heavy industries. Therefore, cleaner utilization of the fossil fuels; development of renewable energy primarily based on solar, wind and hydropower; CO 2 utilization technologies; and increase in efficiency of chemical processes will play important role in reducing CO 2 emission. However, despite an expected increase in the fraction of the renewable energy and/or H 2 energy, catalytic conversion of the abundant world reserved and newly produced carbonaceous resources (X) including coal, natural gas, biomass, organic wastes, biogas and CO 2 to liquid and hydrogen fuels One of the heart of the XTF/XTC technologies is the Fischer-Tropsch synthesis (FTS) reaction, which converts a syngas mixture derived from various carbonaceous materials to hydrocarbons and alcohols (Equations (1)-(3)), meanwhile, a side reaction of water gas shift (WGS) is usually accompanied (Equation (4)).
CO + H2  CnH2n + 2 + H2O (ΔH = −165 kJ/mol) (1) CO + H2  CnH2n + H2O (ΔH = −165 kJ/mol) CO + H2  CnH2n + 2O + H2O (ΔH = −165 kJ/mol) CO + H2O = CO2 + H2 (ΔH = −42 kJ/mol) This technology, because of its importance in industry, has received extensive attention since its discovery in 1923, which led to many FT plants in the world based on large number of fundamental and applied studies [2,3]. The most common catalysts for FTS reaction are the transition metal catalysts i.e., iron, cobalt and Ruthenium, among them the Fe and Co catalysts are largely researched and commercialized because of their high activity, high selectivity to heavy hydrocarbons, and low cost [2][3][4][5]. For example, iron and cobalt catalysts can produce hydrocarbons with a chain growth probability over 0.9, and C5+ selectivity surpasses 88% when operated at low temperature ≤230 °C in a slurry phase reactor [5]. In case of Ru based catalysts, many fundamental studies have been also carried One of the heart of the XTF/XTC technologies is the Fischer-Tropsch synthesis (FTS) reaction, which converts a syngas mixture derived from various carbonaceous materials to hydrocarbons and alcohols (Equations (1)-(3)), meanwhile, a side reaction of water gas shift (WGS) is usually accompanied (Equation (4) This technology, because of its importance in industry, has received extensive attention since its discovery in 1923, which led to many FT plants in the world based on large number of fundamental and applied studies [2,3]. The most common catalysts for FTS reaction are the transition metal catalysts i.e., iron, cobalt and Ruthenium, among them the Fe and Co catalysts are largely researched and commercialized because of their high activity, high selectivity to heavy hydrocarbons, and low cost [2][3][4][5]. For example, iron and cobalt catalysts can produce hydrocarbons with a chain growth probability over 0.9, and C 5+ selectivity surpasses 88% when operated at low temperature ≤230 • C in a slurry phase reactor [5]. In case of Ru based catalysts, many fundamental studies have been also carried out irrespective of the extremely high cost of Ru [6][7][8][9][10]. The Ru catalysts have been reported to be most active, high resistant to water and oxygenate-containing environment, and favor the production of high molecular weight hydrocarbons. This, in turn, results in a decreased cost of FTS fuels production, which is an important requisite to successfully convert biomass-derived syngas into hydrocarbons [6,7]. However, in spite of numerous  studies on this topic, there is still no clear picture what the catalysts' potential performance  can reach or what the intrinsic activities per site of Fe, Co and Ru are. Among the vast number of studies on the FTS reaction, there is one important topic regarding the dependence of catalyst performance on catalyst structure (crystal and size). In this review, recent fundamental development in the aspect over Fe, Co and Ru catalysts are discussed, results among different studies are critically reviewed and compared. It is expected that through better understanding the structure effects and the reaction performance of Fe, Co and Ru catalysts, deep insights into catalysis on the catalysts' surface and catalysts' potential activity can be provided, which should be helpful for the design of new generation of catalysts with super activity, stability and selectivity.

Catalyst Crystal Structure and Catalytic Performance
The crystal structure of the FTS catalysts plays crucial role in controlling FTS performance. In case of iron catalyst, several iron carbides with different structures, for example θ-Fe 3 C, χ-Fe 5 C 2 , and ε-Fe 2 C/ε'-Fe 2 . 2 C, and Fe 7 C 3 (Figure 2a-d), have been identified as active phases for the FTS reaction [11][12][13][14][15]. Dry [11] has reported phase change in iron for an iron catalyst that was operated in a fluidized bed reactor at 327 • C. The metallic Fe was quickly converted to θ-Fe 3 C and χ-Fe 5 C 2 , while Fe 7 C 3 (Eckstrom and Adcock) was observed at very high pressure i.e., >6.0 MPa. No clear correlation between iron catalyst and carbide phase was reported. In the study of Smit et al. [12], the stability and reactivity of iron carbide phases (θ, χ and ε) in FTS catalysts as a function of relevant reaction conditions were studied experimentally and theoretically. These authors have reported that thermodynamic stability of the carbide phases and the transformation among the carbides are highly dependent on the carbon chemical potential (µc) imposed by the chemical (gas phase) surroundings. Increasing temperature from 150 to 450 • C led to linearly decreasing of the µc, while it increased with increasing pressure from 1 to 2.0 MPa or CO/H 2 ratio from 0.25 to 0.45. The trigonal prismatic (TP) carbides such as θ-Fe 3 C, γ-Fe 5 C 2 , and Fe 7 C 3 are reported to be more stable than the octahedral carbides (O) ε-Fe 2 C and η-Fe 2 C at lower µ C (higher temperature, lower CO pressure). At very low µ C (−7.7 eV) θ-Fe 3 C phase is more stable than ε-Fe 2 . 2 C, and θ-Fe 3 C becomes more stable with respect to γ-Fe 5 C 2 at low µ C (−7.6 eV) conditions; however, Fe 7 C 3 was reported to be stable with respect to γ-Fe 5 C 2 at higher µ C (~7.25 eV). The detailed carbides transformation is outlined in Figure 2b. Davis [13] characterized 100Fe/3.7Si/0.7K catalyst that was collected at different time period during the FTS at 270 • C, 1.3 MPa, H 2 /CO = 0.7 and 10-90% CO conversion. After syngas activation followed by 100 h of FTS reaction, the iron catalyst showed low activity (5-20%CO conversion) and then it was re-pretreated by CO for 24 h. The iron catalyst activity was increased significantly after it returned to normal FTS condition, and a stable CO conversion of 90% was obtained in the next 300 h, which was accompanied by mainly Fe 3 O 4 in the first 100 h and about 40% iron carbides and 60% Fe 3 O 4 in the remaining 300 h (Figure 2c). This study likely indicates that catalyst activity is associated with carbide phase (s). Recently, Change et al. [14] was able to show significant difference in the activities of various iron carbides. The group designed and studied the activity of different carbides on a model Fe/SiO 2 catalyst at 260 • C, 3 MPa and H 2 /CO = 2 in a micro-fixed bed reactor. Different iron carbides including ε-Fe 2 C, Fe 7 C 3 , and χ-Fe 5 C 2 were synthesized by treating the calcined Fe/SiO 2 catalyst with different gas atmospheres (CO, H 2 , or syngas) at 300 • C. The intrinsic FTS activity of Fe 7 C 3 was found to be the highest with a turnover frequency (TOF) of 4.59 × 10 −2 s −1 , while χ-Fe 5 C 2 and ε-Fe 2 C had only about 1/3 of the activity ( Figure 2d); meanwhile, ε-Fe 2 C yielded the lowest methane selectivity. and Fe5C2(010) surface, and inactive CH4 rate on Fe3C(001) and Fe4C(100) surface). These authors found that reaction energy and effective barrier of CH4 formation have a linear relationship with the charge of the surface C atom and the d-band center of the surface (Figure 2e,f). Similar results were also reported on Co and Ru surface by van Santen et al. [17]. CO dissociation energy at dense terraces surfaces of Co or Ru can be much higher than that of step-edge sites (200 vs. <100 kJ/mol), which requires a step-edge site on metals for CO activation. FTS performance of cobalt catalyst strongly depends on cobalt crystal structure as well. Two types of cobalt metal structures (active phases), i.e., face centered cubic (fcc), hexagonal close packed (hcp), have been reported, which were synthesized under different pretreatment conditions, for example, Co-fcc was formed on various Al2O3, SiO2 and TiO2 supported catalysts using hydrogen reduction at high temperature over 350 °C, while the hexagonal close packed cobalt metal (Co-hcp) was preferentially formed when the cobalt catalysts were treated by hydrogen at low temperature below 330 °C, or by syngas reduction or by carburation followed by hydrogenation i.e., H2-CO-H2 pretreatment steps between 230-350 °C [18][19][20][21][22][23][24]. When these cobalt phases are exposed to typical FTS conditions, the Co-hcp and Co-fcc display much different intrinsic activity [19][20][21][22]24]. Ducreux et al. [19] prepared SiO2, Al2O3 and TiO2 supported 11-13%Co catalysts with and without 0.45%Ru in order to explore a possible relationship between microstructure of the cobalt Quantum computation i.e., ab-initio, kinetic Monte Carlo (kMC) and density functional theory (DFT) provided deeper insights into the catalysis phenomena on active metal surface. It was found that the catalyst activity also varied with facet and dimension, corner, edge of active sites. Huo et al. [16] used DFT approach to elucidate why methane formation energy varied with carbide phases (i.e., high CH 4 formation rate on Fe 2 C(011) and Fe 5 C 2 (010) surface, and inactive CH 4 rate on Fe 3 C(001) and Fe 4 C(100) surface). These authors found that reaction energy and effective barrier of CH 4 formation have a linear relationship with the charge of the surface C atom and the d-band center of the surface (Figure 2e,f). Similar results were also reported on Co and Ru surface by van Santen et al. [17]. CO dissociation energy at dense terraces surfaces of Co or Ru can be much higher than that of step-edge sites (200 vs. <100 kJ/mol), which requires a step-edge site on metals for CO activation.
FTS performance of cobalt catalyst strongly depends on cobalt crystal structure as well. Two types of cobalt metal structures (active phases), i.e., face centered cubic (fcc), hexagonal close packed (hcp), have been reported, which were synthesized under different pretreatment conditions, for example, Co-fcc was formed on various Al 2 O 3 , SiO 2 and TiO 2 supported catalysts using hydrogen reduction at high temperature over 350 • C, while the hexagonal close packed cobalt metal (Co-hcp) was preferentially formed when the cobalt catalysts were treated by hydrogen at low temperature below 330 • C, or by syngas reduction or by carburation followed by hydrogenation i.e., H 2 -CO-H 2 pretreatment steps between 230-350 • C [18][19][20][21][22][23][24]. When these cobalt phases are exposed to typical FTS conditions, the Co-hcp and Co-fcc display much different intrinsic activity [19][20][21][22]24]. Ducreux et al. [19] prepared SiO 2 , Al 2 O 3 and TiO 2 supported 11-13%Co catalysts with and without 0.45%Ru in order to explore a possible relationship between microstructure of the cobalt catalysts and catalytic activity. These authors observed complete reduction of CoO to cobalt metal by H 2 on Al 2 O 3 above 550 • C, while it was completed on Co/SiO 2 at lower temperatures. In both cases, mixed Co (hcp) and (fcc) cobalt were formed, but low temperatures tend to form high fraction Co (hcp). When the catalysts were reduced by CO at 230 • C for 15 h followed by H 2 reduction at the same temperature, cobalt carbide (Co 2 C) was decomposed and more Co (hcp) stacking was detected. Examining the activity of different Co/Al 2 O 3 and Co/SiO 2 catalysts indicated that activity of the catalyst, most enriched in Co hcp stacking, was found to be 50% higher than the catalyst most enriched in Co fcc stacking on Al 2 O 3 support (X CO = 30 to 45%, Figure 3a), but the activity difference between the two cobalt phases was enlarged by more than triple fold on a Ru promoted Co/SiO 2 supported catalyst (X CO = 8 to 34%, Figure 3b). Later, Sadeqzadeha et al. [21] and Gnanamani et al. [22] used similar approach i.e., using pure H 2 reduction at 350 • C, or CO (230 • C)-H 2 (230 • C) or H 2 (350 • C)-CO (230 • C)-H 2 (230 • C) reduction procedures to prepare Co-hcp and Co-fcc on 0.1%Pt-25%Co/Al 2 O 3 and 20%Co/SiO 2 catalysts, respectively. Activity of the Co-hcp rich catalyst was over one time higher than the Co-fcc rich 25%Co-0.1%Pt/Al 2 O 3 catalyst (Co STY at 220 • C, 2.0 MPa, H 2 /CO = 2.0, and 40% CO conversion: 0.032 to 0.072 s −1 , Figure 3c), while the TOF was only about 32% higher on 20%Co/SiO 2 (CO rate at 200 • C, 1.99 MPa, H 2 /CO = 2.0, and 3.0 NL/gcat/h: 0.024 to 0.034 mol/gcat/h, Figure 3d). Liu et al. [25] reported lower activation energy for the CO activation on Co (hcp) than Co (fcc), which is consistent well with the higher activity of the Co (hcp) structure. It is likely that the Ru and Pt promoters benefit the formation of Co-hcp phase. Moreover, the two cobalt phases gave different hydrocarbon selectivity, Co-hcp yielded low CH 4 and high C 5+ selectivity [22], indicating relatively high hydrogenation barrier but low C-C coupling energy on the Cohcp phase. Cobalt carbide is another different Co phase, which was reported to be inactive for FTS reaction [21,26,27]. However, Ding et al. [28,29] studied the synthesis of C 1 -C 18 alcohols from syngas over La-Co/AC catalysts at 222 • C, 3 MPa, H 2 /CO = 2.0. High alcohol selectivity i.e., 37-39% was reported under the conditions, which was attributed to La promoting the formation of Co 2 C that benefited the alcohol formation.
Reactions 2021, 2, FOR PEER REVIEW 5 catalysts and catalytic activity. These authors observed complete reduction of CoO to cobalt metal by H2 on Al2O3 above 550 °C, while it was completed on Co/SiO2 at lower temperatures. In both cases, mixed Co (hcp) and (fcc) cobalt were formed, but low temperatures tend to form high fraction Co (hcp). When the catalysts were reduced by CO at 230 °C for 15 h followed by H2 reduction at the same temperature, cobalt carbide (Co2C) was decomposed and more Co (hcp) stacking was detected. Examining the activity of different Co/Al2O3 and Co/SiO2 catalysts indicated that activity of the catalyst, most enriched in Co hcp stacking, was found to be 50% higher than the catalyst most enriched in Co fcc stacking on Al2O3 support (XCO = 30 to 45%, Figure 3a), but the activity difference between the two cobalt phases was enlarged by more than triple fold on a Ru promoted Co/SiO2 supported catalyst (XCO = 8 to 34%, Figure 3b). Later, Sadeqzadeha et al. [21] and Gnanamani et al. [22] used similar approach i.e., using pure H2 reduction at 350 °C, or CO (230 °C)-H2 (230 °C) or H2 (350 °C)-CO (230 °C)-H2 (230 °C) reduction procedures to prepare Co-hcp and Co-fcc on 0.1%Pt-25%Co/Al2O3 and 20%Co/SiO2 catalysts, respectively. Activity of the Co-hcp rich catalyst was over one time higher than the Co-fcc rich 25%Co-0.1%Pt/Al2O3 catalyst (Co STY at 220 °C, 2.0 MPa, H2/CO = 2.0, and 40% CO conversion: 0.032 to 0.072 s −1 , Figure 3c), while the TOF was only about 32% higher on 20%Co/SiO2 (CO rate at 200 °C, 1.99 MPa, H2/CO = 2.0, and 3.0 NL/gcat/h: 0.024 to 0.034 mol/gcat/h, Figure 3d). Liu et al. [25] reported lower activation energy for the CO activation on Co (hcp) than Co(fcc), which is consistent well with the higher activity of the Co (hcp) structure. It is likely that the Ru and Pt promoters benefit the formation of Co-hcp phase. Moreover, the two cobalt phases gave different hydrocarbon selectivity, Co-hcp yielded low CH4 and high C5+ selectivity [22], indicating relatively high hydrogenation barrier but low C-C coupling energy on the Co-hcp phase. Cobalt carbide is another different Co phase, which was reported to be inactive for FTS reaction [21,26,27]. However, Ding et al. [28,29] studied the synthesis of C1-C18 alcohols from syngas over La-Co/AC catalysts at 222 °C, 3 MPa, H2/CO = 2.0. High alcohol selectivity i.e., 37-39% was reported under the conditions, which was attributed to La promoting the formation of Co2C that benefited the alcohol formation.   [19]), (b) FT activity of Co/SiO 2 catalyst (from [19]), (c) FT activity of Co/Al 2 O 3 catalyst (reproduced with permission from [21] Copyright 2011, Elsevier), (d) FT activity of Co/SiO 2 catalyst (reproduced with permission from [22] Copyright 2013, Elsevier). (e) C 1 -C 18 alcohols formation on Co/AC catalyst (reproduced with permission from [28] Copyright 2009, Elsevier), (f) C 1 -C 18 alcohols formation on Co/AC catalyst (reproduced with permission from [29] Copyright 2015, Elsevier).
In case of Ru FTS catalyst, the same face centered cubic and hexagonal close packed Ru metal structures as cobalt catalyst have been reported. Interestingly, the Ru catalysts show distinctive reverse activity trend with the Ru phases. Both experiment and theoretical DFT calculation by Li et al. [30] consistently indicated that the activity of the Ru-fcc was significantly higher than the Ru-hcp at the same particle size 6.8 nm, CO rate at 160 • C was found to be 37.8 vs. 11.8 molCO/mol-Ru/h for the Ru-fcc and Ru-hcp, respectively. This could be due to the fact that fcc and hcp Ru bulks exhibited very different morphologies, the fcc Ru was octahedron like populated mainly by close-packed (111) facets, and the hcp Ru was a dihedral-like shape populated mainly by open facets, which gave greater amounts of Ru-fcc open facets and more active sites for CO dissociation, even though the intrinsic activity of hcp phase is higher than fcc due to the low energy barrier for CO dissociation on the hcp phase on both Co and Ru metals (Co-hcp 1.07 ev vs. Co-fcc 1.47 ev [25], Ru-hcp 1.0 ev vs. Ru-fcc 1.2ev) [30].
It should be noted that the structure of the catalysts that plays crucial roles on the FTS reaction are not only controlled by pretreatment conditions [14,[18][19][20][21][22][23][24], but also are adjusted greatly by promoters and supports/binders [31][32][33][34][35]. For example, K [31,32] and Mn [33,34] were reported to promote and suppress iron carburization, respectively, leading to different amounts of iron carbides. Incorporation of SiO 2 to iron catalysts restrained carburization of iron catalysts but promoted heavier hydrocarbons formation and catalyst stability [35]. Most recently, Lu et al. [15] reported that SiO 2 or Al 2 O 3 supports showed electronic impact on iron atomic status, resulting in an improved amounts of C-rich iron carbide active phase (ε-Fe 2 C) that was claimed to have more positive roles on the activity of supported Fe/Al 2 O 3 and Fe/SiO 2 catalysts under 280 • C, 1 MPa, H 2 /CO = 1, and 10 NL/g-cat/h. The similar notable support effects on Ru catalyst have also been evidenced [6,36,37]. The discussion of additional support effects on the properties of FTS catalysts is beyond the scope of this review.
In summary, the crystal structure (phase and morphology) of iron, cobalt and Ru catalysts, that can be controlled by adjusting pretreatment conditions, supports/binders and promoters, remarkably impact on FTS performance. The influence are very complicated not only because of many active phases of each metal present during FTS, but also because of different morphologies formed including shape, facet and orientation, edge and corner on each phase, which could change amounts of active sites, and CO adsorption mechanism (multiple reaction mechanisms could occur on each catalyst surface), consequently leading to different catalyst activity, selectivity and stability [16,25,30]. Further improving FTS catalyst performance towards better stability and higher heavier hydrocarbon yield is challenge and this requires tremendous efforts including experimental optimization of the catalyst structure coped with the advanced technique of molecular-level catalyst design i.e., DFT.

Metal Particle Size Effects on Catalyst Performance
Metal particle size effect, which is related to the structure sensitiveness of the catalyst, has been extensively studied in the past decades [8,. Since Boudart et al. [38] proposed the concept and assumed FTS is a structure insensitive reaction, it took long time to amend the conclusion. Iglesia's group used Co or Ru catalysts with the metal particle size over 8 nm and had drawn the conclusion of independent of catalysts performance on the metal particle size/dispersion (structure insensitive) [8,40,41]. This conclusion is not in line with the earlier particle size studies on iron catalysts by Vannice [42] and Bartholomew groups [43], which noticed lower activity on smaller Fe particles in less than 9 nm particles. A conclusive cobalt particle size effect on FTS has not been made until de Jong group [44] studied the effect in a tiny to big particle size range (2.6-27 nm). In contrast, the Fe and Ru particle size effects are less consistent even though it has attracted significant interests from academics. The review and discussion on the recent exploration of the particle size effects on Co, Fe and Ru catalysts are given below. Important insights into the particle size effects are also provided.

Examples of Cobalt Catalyst
As mentioned above, Iglesia [41] found that the CO turnover frequency (TOF) remained constant at 0.021 s −1 (independent of Co dispersion and support type) in the Co particle size range of 8-96 nm under 200 • C, 2 MPa, H 2 /CO = 2.05, and 55-65%CO conversion (Figure 4a), apparently, the effect of the smaller Co size less than 8 nm was ignored. A comprehensive Co particle size effect study was represented by the recent study of Bezemer et al. [44], who used 1.0-22%Co/CNF (Carbon Nanofiber), prepared by impregnation (IWI) and homogeneous deposition precipitation (HDP), that yielded 2.6-27 nm Co particles. These authors reported that Co TOF was only 0.0014 s −1 at 2.6 nm, which increased with increasing Co particle size up to 8.0 nm, reaching at a constant value of 0.023 s −1 at 210 • C, 3.5 MPa, and H 2 /CO = 2.0 ( Figure 4b). Therefore, both the Iglesias and Bezemer groups found that Co TOF was independent of Co particle size for the catalysts with sizes greater than 8 nm at typical FT conditions. C 5+ selectivity and olefin selectivity were reported to increase with increasing Co particle size up to 16 nm, accompanied with a decrease in CH 4 selectivity. In several subsequent studies on Co particle size effect, Breejen et al. [45] used the steady-state isotopic transient kinetic analysis (SSITKA) technique to elucidate nature of the small Co size effect (<8 nm). It was postulated that small particles block edge/corner sites and show a low intrinsic activity at the small terraces; furthermore, small Co particles were assumed to possess higher H 2 coverages, leading to higher CH 4 selectivity. Holmen et al. [46] obtained the maximum C 5+ selectivity (84.2%) under 210 • C 2.0 MPa, H 2 /CO = 2.1, and 43-46%CO conversion at an average Co particle size of 8 nm on Al 2 O 3 support, above 10 nm, C 5+ selectivity was nearly unchanged. Prieto et al. [47] also reported the two regions of Co particle size i.e., < >10 nm, that showed different activity trends over 10%Co/ITQ-2(zeolite) catalysts. The maximum and constant TOF (0.0086 s −1 ) at 220 • C, 2 MPa, H 2 /CO = 2, and 10%CO conversion was observed at the Co size range of 10-141 nm. Small cobalt particles (5.6-10 nm) showed lower activity, that was ascribed to the formation of partially oxidized Co species (Co d+ ) at the Co-support interface. Gavrilovic et al. [62] and Park et al. [64] also conducted Co particle size effect in 6-13 nm and 3-16 nm ranges using 0.5%Re20%Co/Al 2 O 3 and 5%Co/Al 2 O 3 , respectively. Similar activity and C 5+ selectivity trends were obtained in the cobalt particle size ranges smaller and larger than 9.3-10 nm.
Reactions 2021, 2, FOR PEER REVIEW 8 Several reasons are proposed to be responsible for the discrepancy: (1) support/transport effect, (2) deactivation, and (3) experimental error. The cobalt initial TOF value of 0.1 s −1 reported by Ma et al. [48] was obtained in a CSTR using fine particles i.e., 45-90 μm, thus the value reflects the intrinsic Co activity because the internal and external mass transfer were eliminated and the deactivation was ignored; while the studies of Iglesia [41}, Bezemer et al. [44] and Prieto et al. [47] employed large catalyst particles i.e., 170 mm, 500 mm or 250-400 mm, respectively, which included the mass transfer effect in the reaction, consequently, it decreased the catalysts' activities. However, the activity difference resulted by the difference in mass transfer of the 170-500 mm particles is estimated to be less than 17-30% at initial and steady state periods in term of a recent study on the sieve size effect using 20-390 mm Co/Al2O3 catalyst particles [49]. Thus, the mass transfer effect is unlikely the main cause while the deactivation should be primarily responsible for the different TOF values obtained in these studies.
Further analysis of the data of Ma et al. [48] indicates that after 120-200 h of testing, the Co TOF value based on the initial Co dispersion was decreased by 34-52% and the higher loss percentage was observed on the smaller Co particles. Assuming that the intrinsic activity of Co sites at a given set of conditions was unchanged, the loss in 34-52% activity might indicate a loss in the same amount of Co active sites, or even loss in more amounts of cobalt active sites i.e., 50-90% in other studies [41,44,47]. These data suggest that the majority of cobalt active sites did not participate the reaction in the lab testing or in the commercial practice [50] due to site blocking, or deactivation. Therefore, design of new generation of cobalt FTS catalyst by preserving as much as Co active sites during FTS reaction and maximizing cobalt particles' activity is an urgent and important task, which could dramatically improve product yields and reduce capital cost in FTS.  [41] Copyright 1997, Elsevier), (b) the influence of cobalt particle size on the TOF at 3.5 MPa, 210 °C and H2/CO = 2.0 (reproduced with permission from [44] Copyright 2006, Elsevier), (c) the influence of cobalt particle size on Co TOF at 2.2 MPa, 220 °C and H2/CO = 2.1 (data from [48]).

Examples of Iron Catalyst
The iron particle size effects on FTS reaction were studied by many researchers as well, but the conclusion is still debatable. Earlier studies on the topic were conducted on 2.5-6.0%Fe/Carbon (C) catalysts by Vaccine group [42], and on 1-10%Fe on C or/Al2O3 catalysts by Bartholomew group [43]. Both groups reported consistently that smaller Fe particles displayed lower activity after H2 pretreatment at 350 °C, for example, the TOF for CO hydrogenation at 220 °C, 0.1 MPa and H2/CO = 2.0 increased from 0.86 to 4.4 × 10 −3 s −1 in a study by Vannice et al. [42] and 0.69 to 18 × 10 −3 s −1 in a study by Bartholomew et al. [43] with increasing Fe particle size from 0.6 to 9 nm. These authors did not provide  [41] Copyright 1997, Elsevier), (b) the influence of cobalt particle size on the TOF at 3.5 MPa, 210 • C and H 2 /CO = 2.0 (reproduced with permission from [44] Copyright 2006, Elsevier), (c) the influence of cobalt particle size on Co TOF at 2.2 MPa, 220 • C and H 2 /CO = 2.1 (data from [48]).
An attention that must be paid is that the Co TOF varies significantly in different studies although researchers reach a consensus on the Co particle size effect. Iglesia [41] obtained an apparent Co TOF of 0.021 s −1 on various supported catalysts at 200 • C and 2.0 MPa in a fixed bed reactor, which is equivalent to 0.05 s −1 at 220 • C; the TOF value of Bezemer et al. [44] on CNF support at 210 • C and 3.5 MPa was 0.023 s −1 , equivalent to 0.038 s −1 at 220 • C. Prieto et al. [47] reported the lowest Co TOF values of 0.0086 s −1 on Co/ITQ-2-zeolite under similar process conditions. In the study of Park et al. [64], a high cobalt initial TOF value (24-40 h) of 0.08 s −1 was obtained on a 5%Co/Al 2 O 3 catalysts with mean cobalt particle size of 9.3 nm under 220 • C, 1.0 MPa, H 2 /CO = 2.0 and 30% CO conversion. Recently, Ma et al. [48] studied the effects of 0.5% noble metals (Pt, Re, Ru) on 25%Co/Al 2 O 3 catalyst for FTS reaction under 220 • C, 2.0 MPa, H 2 /CO = 2.1 and 50% CO conversion in 1-L CSTR. By taking into account the chemisorption results on the cobalt catalysts, they obtained a high initial Co TOF of 0.1 s −1 in the cobalt particle size range 10.0-18.6 nm on the Pt, Re, Ru promoted and unpromoted 25% Co/Al 2 O 3 catalysts (Figure 4c), which is in line with Park's study. At steady state, the Co TOF based on the initial Co dispersion was reduced to 0.05-0.067 s −1 , which is close to the apparent TOF values in the study of Iglesia [41], but much greater than that in the studies of Bezemer and Prieto.
Several reasons are proposed to be responsible for the discrepancy: (1) support/transport effect, (2) deactivation, and (3) experimental error. The cobalt initial TOF value of 0.1 s −1 reported by Ma et al. [48] was obtained in a CSTR using fine particles i.e., 45-90 µm, thus the value reflects the intrinsic Co activity because the internal and external mass transfer were eliminated and the deactivation was ignored; while the studies of Iglesia [41], Bezemer et al. [44] and Prieto et al. [47] employed large catalyst particles i.e., 170 mm, 500 mm or 250-400 mm, respectively, which included the mass transfer effect in the reaction, consequently, it decreased the catalysts' activities. However, the activity difference resulted by the difference in mass transfer of the 170-500 mm particles is estimated to be less than 17-30% at initial and steady state periods in term of a recent study on the sieve size effect using 20-390 mm Co/Al 2 O 3 catalyst particles [49]. Thus, the mass transfer effect is unlikely the main cause while the deactivation should be primarily responsible for the different TOF values obtained in these studies.
Further analysis of the data of Ma et al. [48] indicates that after 120-200 h of testing, the Co TOF value based on the initial Co dispersion was decreased by 34-52% and the higher loss percentage was observed on the smaller Co particles. Assuming that the intrinsic activity of Co sites at a given set of conditions was unchanged, the loss in 34-52% activity might indicate a loss in the same amount of Co active sites, or even loss in more amounts of cobalt active sites i.e., 50-90% in other studies [41,44,47]. These data suggest that the majority of cobalt active sites did not participate the reaction in the lab testing or in the commercial practice [50] due to site blocking, or deactivation. Therefore, design of new generation of cobalt FTS catalyst by preserving as much as Co active sites during FTS reaction and maximizing cobalt particles' activity is an urgent and important task, which could dramatically improve product yields and reduce capital cost in FTS.

Examples of Iron Catalyst
The iron particle size effects on FTS reaction were studied by many researchers as well, but the conclusion is still debatable. Earlier studies on the topic were conducted on 2.5-6.0%Fe/Carbon (C) catalysts by Vaccine group [42], and on 1-10%Fe on C or/Al 2 O 3 catalysts by Bartholomew group [43]. Both groups reported consistently that smaller Fe particles displayed lower activity after H 2 pretreatment at 350 • C, for example, the TOF for CO hydrogenation at 220 • C, 0.1 MPa and H 2 /CO = 2.0 increased from 0.86 to 4.4 × 10 −3 s −1 in a study by Vannice et al. [42] and 0.69 to 18 × 10 −3 s −1 in a study by Bartholomew et al. [43] with increasing Fe particle size from 0.6 to 9 nm. These authors did not provide CH 4 selectivity data, but an estimation based on the available CH 4 and CO rates suggested lower CH 4 selectivity on smaller iron carbides; furthermore, smaller Fe carbides were reported to have lower activation energy (84-95 kJ/mol or 64-105 kJ/mol). These results are presumably due to the difference in electronic density on the small and large particles, different Fe-C interaction, and/or a decrease in the corner and edge sites on large crystallites that were responsible for the increases in both turnover frequency and activation energy with increasing crystallite size. Small Fe particles i.e., 1-2 nm led to high deactivation rate [42,43].
Recently, a progress on the study on Fe particle size effect was made by several research groups [51][52][53]. van Steen et al. [51] and Park et al. [52] explored the effects of Fe particle size in the ranges 3.0-19.2 nm over 2.8-4.7%Fe/C, 4.6-5.6%Fe/g-Al 2 O 3 and 2-12 nm over 5%Fe/Al 2 O 3 catalysts, on FTS performance, respectively. After H 2 reduction at 350 • C, lower initial TOF on smaller Fe particles (3-10 nm by Steen, 2.0-6.2 nm by Park) was noticed, then it remained nearly constant up to the particle size of 19.2 nm or 12 nm. For example, the stable TOF values for the Fe/Al 2 O 3 and Fe/C catalysts with larger Fe particles in van Steen study under 270 • C, 3.0 MPa, H 2 /CO = 2.0, 7.2 or 18 NL/g-cat/h (40 and 8% CO conversion) were 3.1 and 6.0 mmol/m 2 -Fe/s, respectively (Figure 5a), and 0.187 s −1 in Park study under 300 • C, 1 MPa, 3.6 L/gcat/h, and H 2 /CO = 2 (Figure 5b). Thus, part of the van Steen and Park results in the small Fe particle size region are consistent with the previous studies of Vannice et al. [42] and Bartholomew et al. [43], and a most recent study by Gu et al. [72], which reported an increase in Fe TOF with increasing iron particle size in the 2.5-12 nm range under 350 • C, 0.1-10 MPa, and H 2 /CO = 1.0 and 60% CO conversion. However, there is a difference in the Fe particle size threshold (10 vs. 6.2 nm), at which the TOF starts to be stable, in the studies. The difference could arise from the support effects and/or errors in the characterization and activity measurements. Further comparison of various studies on Fe catalysts suggests that the TOF values of the Fe particles over 7 nm reported by Park et al. [52], Chang et al. [14] and Ma et al. [5] under 260-300 °C, 1-3 MPa, H2/CO = 1-2 and 20-50% CO conversion are essentially consistent, and they fall into the range 0.046-0.20 s −1 . These high TOF values should be close to the Fe intrinsic activity due to the consistency of the results and the ignored mass transfer effect by using fine particles in these studies. Thus, the Fe TOF values reported can serve as reference for the design of highly active Fe catalysts. Recently, Keyvanloo et al. [54] prepared highly active 40%Fe/AlSi and 40%Fe/AIG (including 5%La) supported catalysts. The activity of the supported Fe catalysts with 8 nm FexC under 260 °C, 2.0 MPa, H2/CO = 1 was estimated to be 0.045 s −1 , which is consistent with the reported TOF value. However, the CO TOF for the Fe catalyst with 7 nm Fe particle size reported by Torres et al. [53] was about 10 times lower, 0.017 s −1 at 300 °C (corrected from 0.06 s −1 at 340 °C and assuming Ea = 90 kJ/mol), while the activity on the Fe/C and Fe/Al2O3 catalysts reported by Barkhuizen et al. [51] was significantly beyond the range (80-200 times higher), 12-51 s −1 at 260-300 °C (estimated from the reported 3-6 mmol/m 2 Fe/s at 270 °C and assuming Fe density 20.4 atoms/nm 2 ). The primary causes for the large difference in the TOF in the studies [51,53] are not known.

Examples of Ruthenium Catalyst
The effects of Ru particle size on FTS were explored by many researchers and the results are more similar to that of the Co particle size effect as discussed above, but some differences from study to study still exist. Systematic early studies on the effect of Ru particle size on FTS performance were reported by Kellner et al. [55] and Iglesia [8], which Interestingly, de Jong group [53,73] studied the conversion of syngas to light olefins (FTO) over 1, 2, 5, 10 and 20% Fe supported on carbon nano fiber (CNT) with and without Na and S promoters in a micro fixed bed reactor. A different Fe particle size effect from that in the van Steen et al. [51] and Park et al. studies [52] was reported in the Fe size range of 2-17 nm [53] or 3-9 nm [73]. After the iron catalysts were reduced in H 2 at 350 • C for 2 h, it led to a decreased trend of CO to hydrocarbons TOF from 0.1 to 0.06 s −1 with increasing Fe x C size from 2 to 7 nm at 340 • C, 2 MPa, H 2 /CO = 1, and 5.45 NL/g-cat/h [53] (Figure 5c). At a low pressure 0.1 MPa and 350 • C, the same trend was also obtained, however, the CO conversion remained nearly unchanged (9-11%) for the unpromoted 1, 2, 5, 10 and 20%Fe/CNT catalysts at a constant flow rate (6 mL/min vs. 0.02 g cat). With the absence of the Fe reduction degree data, it is estimated from the available rates and Fe dispersion data that the extent of Fe reduction is relative constant despite significant change in Fe loading from 1 to 20%. These results are unusual that cannot be explained by the common knowledge of metal loading effect. It is likely that many iron active sites on high Fe loading catalysts did not participate in reaction in the low pressure. The authors also reported higher CH 4 selectivity on smaller Fe particles, which is consistent with the study of Park et al. [52], but against the results of Vannice et al. [42] and Bartholomew et al. [43]. The reasons for the opposite Fe particle size effects obtained by Torres et al. [53] and Gu et al. [73] is not known. Additional efforts are needed to clarify the issue of iron particle size effect.
Further comparison of various studies on Fe catalysts suggests that the TOF values of the Fe particles over 7 nm reported by Park et al. [52], Chang et al. [14] and Ma et al. [5] under 260-300 • C, 1-3 MPa, H 2 /CO = 1-2 and 20-50% CO conversion are essentially consistent, and they fall into the range 0.046-0.20 s −1 . These high TOF values should be close to the Fe intrinsic activity due to the consistency of the results and the ignored mass transfer effect by using fine particles in these studies. Thus, the Fe TOF values reported can serve as reference for the design of highly active Fe catalysts. Recently, Keyvanloo et al. [54] prepared highly active 40%Fe/AlSi and 40%Fe/AIG (including 5%La) supported catalysts. The activity of the supported Fe catalysts with 8 nm Fe x C under 260 • C, 2.0 MPa, H 2 /CO = 1 was estimated to be 0.045 s −1 , which is consistent with the reported TOF value. However, the CO TOF for the Fe catalyst with 7 nm Fe particle size reported by Torres et al. [53] was about 10 times lower, 0.017 s −1 at 300 • C (corrected from 0.06 s −1 at 340 • C and assuming E a = 90 kJ/mol), while the activity on the Fe/C and Fe/Al 2 O 3 catalysts reported by Barkhuizen et al. [51] was significantly beyond the range (80-200 times higher), 12-51 s −1 at 260-300 • C (estimated from the reported 3-6 mmol/m 2 Fe/s at 270 • C and assuming Fe density 20.4 atoms/nm 2 ). The primary causes for the large difference in the TOF in the studies [51,53] are not known.

Examples of Ruthenium Catalyst
The effects of Ru particle size on FTS were explored by many researchers and the results are more similar to that of the Co particle size effect as discussed above, but some differences from study to study still exist. Systematic early studies on the effect of Ru particle size on FTS performance were reported by Kellner et al. [55] and Iglesia [8], which suggested structure insensitive of the FTS reaction, i.e., both studies indicated Ru time yield was independent of support type (Al 2 O 3 , SiO 2 , and TiO 2 ) and increased linearly with increasing Ru dispersion. The Ru time yield changed from 0.0012 to 0.0065 s −1 at 203 • C, 0.56 MPa, H 2 /CO = 2.1 and 45-60% CO conversion when Ru dispersion increased from 8.5 to 46% (Figure 6a) [8]. Based on the relationship between the Ru particles size and dispersion (d Ru = 132/D), 8.5 to 46% Ru dispersion corresponds to the Ru particle size of 3.0 to 15.5 nm, in which the slope yields a constant Ru TOF of 1.41 × 10 −2 s −1 . In recent years, there has been an increasing interest in exploring the Ru particle size effect, probably inspired by the early Ru works and/or by the recent studies of Co particle size effect. A similar trend (constant TOF) at large Ru particles of 10-23 nm (0.108 s −1 at 250 • C, 0.2 MPa, and H 2 /CO = 10) using Ru/Al 2 O 3 catalyst or at 7-12 nm (0.105 s −1 at 240 • C, 1.5 MPa, and H 2 /CO = 2) using 3.6-17.3%.Ru/C catalyst was reported by Carballo et al. [7] and Eslava et al. [56], respectively; at Ru sizes smaller than 7 or 10 nm, Ru TOF was found to increase with increasing Ru size (Figure 6b). Apparently, the results of Carballo and Eslava about the Ru particle size effect were quite similar to that of the Co particle size effect (Co size threshold: 8-10 nm) [41,[44][45][46][47]64] and some of the Fe particle size effect studies (Fe size threshold: 7-10 nm) [51,52]. The smaller nano Ru particles showed lower activity, which was explained by strong adsorption of surface species and blocking of active sites. Later, Carballo et al. [57] reported a different Ru particle size effect on the initial activity in the Ru particle size range 4 to 71 nm over 3%Ru/Al 2 O 3 (Figure 6c). The initial TOF values increased with Ru particle size, but the incremental pace became slower at the Ru size greater than 10 nm. This unique trend of the initial activity was presumably due to loss in the step-edges sites on the large Ru particles i.e., >10 nm, for CO dissociation, while the terraces sites might remain unchanged. The small Ru particles likely favored the production of CH 4 and lowered C 5+ hydrocarbons. of 6.9 mol/mol Ru/h at a low temperature 150 °C and 3 MPa, H2/CO = 2.0 in a batch reactor, which was 6-35 times higher than that of other Ru/SiO2 and Ru/Al2O3 catalysts. The results suggest a strong support effect on the Ru/SiO2 and Ru/Al2O3 catalysts. More interestingly, an unprecedented peak TOF value of 12.9 h −1 at the Ru size of 2.0 nm was reported, after that Ru size point, the TOF increased gradually with increasing Ru size from 2.5 to 4.0 nm (Figure 6e). This result is rather unique, that no atomic-level explanation is currently available to rationalize the variation.  Recently, Wang group [58,75] studied the Ru size effects in the range of 2.4-10 nm with 3%Ru on various supported catalysts (SiO 2 , Al 2 O 3 , TiO 2 , MaO, ZrO 2 , HY, NaY, H-beat, CNT, AC, graphite) at a high temperature 260 • C, and 2.0 MPa H 2 /CO = 1.0, 2.4 NL/g-cat/h. Such Ru size effect namely a constant TOF beyond 7 nm Ru size as discussed above was not observed, instead a peak TOF (0.193 s −1 ) ( Figure 6d) and a peak selectivity to C 10 -C 20 (~62%) were reported at the Ru size of 6.3 nm [58]. The TOF was decreased in the large Ru size range between 7 and 10 nm, which is different from the studies of Carballo [7], Iglesia [8] and Kellner [55]. The catalyst deactivation and/or support effect could be responsible for the varied results. Xiao et al. [59] recently studied the effects of Ru size on FTS activity in aqueous phase. Authors obtained a narrow Ru particle size 1.8-4 nm by changing Ru loading from 0.5 to 5% on PVP. The authors observed a maximum activity of 6.9 mol/mol Ru/h at a low temperature 150 • C and 3 MPa, H 2 /CO = 2.0 in a batch reactor, which was 6-35 times higher than that of other Ru/SiO 2 and Ru/Al 2 O 3 catalysts. The results suggest a strong support effect on the Ru/SiO 2 and Ru/Al 2 O 3 catalysts. More interestingly, an unprecedented peak TOF value of 12.9 h −1 at the Ru size of 2.0 nm was reported, after that Ru size point, the TOF increased gradually with increasing Ru size from 2.5 to 4.0 nm (Figure 6e). This result is rather unique, that no atomic-level explanation is currently available to rationalize the variation.
The above studies suggest more complicated Ru particle size effects despite many of the studies tend to show similar Ru size effect to the Co catalyst (two size regions' effects). Additional efforts are needed to elucidate the unusual trends on the Ru catalyst. On the other hand, it is difficult to compare the Ru activity per active site (TOF) in different studies because temperatures used were varied. Corrections in the Ru TOF obtained at different temperatures to the values at a given temperature i.e., 220 • C, assuming the Ru activation energy for CO hydrogenation is 65 kJ/mol, make it possible to understand the catalyst's working efficiency. It is found that the estimated Ru TOF values in the Ru size 7-40 nm under 220 • C, 0.56-4 MPa, H 2 /CO = 2.0 in most Ru studies [7,8,[56][57][58][59], fall into 0.025-0.075 s −1 . However, the results in several other studies over Ru catalysts are different. Based on the initial Ru activity of a 3%Ru/γ-Al 2 O 3 with Ru size of 70 nm in the study of Carballo et al. [57], and a study with Pt promoted Ru shell catalyst (Pt/Ru = 1/9) in aqueous FTS reaction by Li et al. [30], the estimated Ru TOF at 6.8 nm at 220 • C and 2-4 MPa was about one order of magnitude higher, reaching 0.4-0.44 s −1 . This Ru TOF value is slightly higher than that in a recent study with an active 2.2% Ru/TiO 2 catalyst having similar Ru particle size [60], which gave Ru TOF of 0.3 s −1 at 220 • C and 2 MPa. Comparing these high Ru TOF values, it is concluded that the intrinsic activity of the Ru catalyst at 220 • C should fall into 0.4-0.44 s −1 . The low activity of the Ru catalysts suggests that most Ru research catalysts were far from the high efficiency, and majority of Ru active sites in catalysts had not participated in the FT reaction. In terms of the above data analysis, the design of new Ru catalyst should aim at taking advantage of every Ru sites for the FTS reaction, so that potential productivity of the Ru catalyst can be reached. The study further suggests that the intrinsic activity of Ru catalyst is about 3 times higher than that of Co catalyst (TOF: 0.4 vs. 0.1 s −1 at 220 • C).

Conclusions
Fossil fuel energy will continue to be the dominant energy resource in the world over the next several decades; thus, catalytic conversion of carbonaceous containing materials including coal, natural gas, biomass, organic wastes, biogas and CO 2 to liquid and hydrogen fuels (XTF) and chemicals (XTC) can serve as a sustainable, efficient and clean approach to partly supply nations' energy need and reduce CO 2 emission. FTS is one of the important XTF/XTC processes that play a key role in cleaner utilization of the carbon resources for fuels and chemicals production.
The recent studies on the effects of FTS catalysts' structure (crystal and size) on FTS performance is comparatively reviewed and summarized. The crystal structure of the catalysts plays decisive roles on FTS performance, but the effects are so complicated. Not only the crystal type but also morphology including shape, facet, edge and corner, affect catalyst activity and selectivity. Due to differing carbon chemical potential, one active iron carbide site that is good for high activity may not favor the formation of the desired products. For example, Fe 7 C 3 is the most active, while ε-Fe 2 C gives lower methane selectivity. In the case of Co and Ru catalysts, close packed metal (hcp) and face centered cubic (fcc) have been evidenced, but two phases in the two types of catalysts show opposite activity trends at the same particle size, i.e., Co: hcp > fcc, Ru: fcc > hcp, which reflects complex effects of morphology and facet on the Co and Ru catalysts. Multiple reaction pathways could occur on catalyst step-edge or terraces sites due to different CO/H dissociation energy on the sites. The phase transformation for all three types of FTS catalysts during the reaction (Fe x C Fe y C Fe 3 O 4 , hcp fcc) make the FTS catalysis even more complicated. Therefore, further improving stability, activity and selectivity of FTS catalyst is challenging and should consider the environment that the catalyst is exposed to. It is possible to successfully tune catalyst structure toward the target products with the aid of the advanced technique of molecular level catalyst design plus massive experimental trials.
The recent studies related to particle size effects were thoroughly compared and discussed in this review. Based on the extensive studies on the topic, it was concluded that there are two metal size regions in which the FTS activity (TOF) of Co, Fe and Ru catalysts changed with particle size in a different manner. In the small particle size region, for example, below 8 nm for cobalt particles or 7 nm for iron and ruthenium particles, the activity of the catalysts started to increase dramatically, whereas in the big particle size greater than the threshold, the activity was independent of metal particle size. The threshold of the particle size reported in some studies was extended to 10 nm. The reasons why small metal particles offered low activity has been explained by site blocking at particles' edge or corner, and/or oxidation of small nanoparticles. However, inconsistent results in the small particle size region for the iron and ruthenium catalysts present in some studies i.e., opposite activity trend on small iron particles and peak TOF value at 2 nm ruthenium particle, were observed. Thus, the effects of particle size on the FTS activity of iron and ruthenium catalysts have not reached a consensus. Additional studies are needed to clarify the discrepancy.
The activity of the Co, Fe and Ru catalysts were studied by analyzing activity data of many researchers. Based on the literature studies, the intrinsic activity of Co and Ru particles of 7-70 nm at 220 • C were estimated to be 0.1 s −1 and 0.4-0.44 s −1 , respectively, while that of the Fe catalysts of 10 nm at 260-300 • C was 0.046-0.20 s −1 . However, most of the research catalysts or some industrial catalysts only contributed 10-50% of potential activity due to the support, electronic or deactivation effects. Therefore, further improvement of activity of the FTS catalysts to a higher level is an important and urgent task in future study.