Synthesis of Iron-Carbide Nanoparticles - Identification of the Active Phase and Mechanism of Fe-Based Fischer–Tropsch Synthesis
At a Glance
Section titled “At a Glance”| Metadata | Details |
|---|---|
| Publication Date | 2020-11-13 |
| Journal | CCS Chemistry |
| Authors | Huabo Zhao, Jin‐Xun Liu, Ce Yang, Siyu Yao, Hai‐Yan Su |
| Institutions | Hefei National Center for Physical Sciences at Nanoscale, Peking University |
| Citations | 89 |
Abstract
Section titled “Abstract”Open AccessCCS ChemistryRESEARCH ARTICLE1 Nov 2021Synthesis of Iron-Carbide Nanoparticles: Identification of the Active Phase and Mechanism of Fe-Based Fischer-Tropsch Synthesis Huabo Zhao†, Jin-Xun Liu†, Ce Yang†, Siyu Yao, Hai-Yan Su, Zirui Gao, Mei Dong, Junhu Wang, Alexandre I. Rykov, Jianguo Wang, Yanglong Hou, Wei-Xue Li and Ding Ma Huabo Zhao† College of Chemistry and Molecular Engineering, Peking University, Beijing 100871 , Jin-Xun Liu† Department of Chemical Physics, School of Chemistry and Materials Science, Hefei National Laboratory for Physical Sciences at the Microscale, iCHEM, University of Science and Technology of China, Hefei 230026 , Ce Yang† Department of Materials Science and Engineering, College of Engineering, Peking University, Beijing 100871 , Siyu Yao College of Chemistry and Molecular Engineering, Peking University, Beijing 100871 , Hai-Yan Su School of Chemical Engineering and Energy Technology, Dongguan University of Technology, Dongguan 523808 , Zirui Gao College of Chemistry and Molecular Engineering, Peking University, Beijing 100871 , Mei Dong State Key Laboratory of Coal Conversion, Institute of Coal Chemistry, Chinese Academy of Sciences, Taiyuan 030001 , Junhu Wang Center for Advanced Mössbauer Spectroscopy, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023 , Alexandre I. Rykov Center for Advanced Mössbauer Spectroscopy, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023 , Jianguo Wang State Key Laboratory of Coal Conversion, Institute of Coal Chemistry, Chinese Academy of Sciences, Taiyuan 030001 , Yanglong Hou *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] E-mail Address: [email protected] Department of Materials Science and Engineering, College of Engineering, Peking University, Beijing 100871 , Wei-Xue Li *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] E-mail Address: [email protected] Department of Chemical Physics, School of Chemistry and Materials Science, Hefei National Laboratory for Physical Sciences at the Microscale, iCHEM, University of Science and Technology of China, Hefei 230026 and Ding Ma *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] E-mail Address: [email protected] College of Chemistry and Molecular Engineering, Peking University, Beijing 100871 https://doi.org/10.31635/ccschem.020.202000555 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesDownload CitationsTrack Citations ShareFacebookTwitterLinked InEmail Despite the extensive study of the Fe-based Fischer-Tropsch synthesis (FTS) over the past 90 years, its active phases and reaction mechanisms are still unclear due to the coexistence of metals, oxides, and carbide phases presented under realistic FTS reaction conditions and the complex reaction network involving CO activation, C-C coupling, and methane formation. To address these issues, we successfully synthesized a range of pure-phase iron and iron-carbide nanoparticles (Fe, Fe5C2, Fe3C, and Fe7C3) for the first time. By using them as the ideal model catalysts on high-pressure transient experiments, we identified unambiguously that all the iron carbides are catalytically active in the FTS reaction while Fe5C2 is the most active yet stable carbide phase, consistent with density functional theory (DFT) calculation results. The reaction mechanism and kinetics of Fe-based FTS were further explored on the basis of those model catalysts by means of transient high-pressure stepwise temperature-programmed surface reaction (STPSR) experiments and DFT calculations. Our work provides new insights into the active phase of iron carbides and corresponding FTS reaction mechanism, which is essential for better iron-based catalyst design for FTS reactions. Download figure Download PowerPoint Introduction Fischer-Tropsch synthesis (FTS) receives increasing attention because synthesis gas (syngas) can be obtained from coal, biomass, and shale gas, which becomes particularly important for the production of alternative fuel and chemicals.1,2 Despite the extensive exploration of Fe-based catalysts over the past 90 years, its active phases and reaction mechanisms are still controversial.3 The typical iron catalysts, usually produced from thermal reduction and successive activation of iron oxide precursors, contain different phases, including metals, oxides, and carbides, produced during the pretreatment of the catalysts by carbon-containing gases such as CO.4-10 The phase evolution of the iron catalysts during the FTS reaction is even more complicated, and in most cases, a mixture of different phases occurred.11-15 To address this issue, various types of metal or metal carbides were prepared.16-22 Metallic iron23,24 and various phases of iron carbides had all been claimed to be active.4,6,25-38 The complication for FTS comes also from the complexity of the reaction network itself, including CO activation, C-C bond formation, and methane formation.32,39-45 However, the reaction performance dependence on catalyst phases of iron and carbides as well as the catalytic mechanism behind was still elusive, hindering better Fe-based catalyst design for FTS. To address these challenging issues, for the first time, we have successfully synthesized a variety of pure-phase iron and iron-carbide catalysts with similar particle size, including Fe, Fe7C3, Fe3C, Fe5C2, and so forth, which allow us to identify their initial and intrinsic activities as well as the structural evolution of iron-based catalysts during FTS. At the same time, we developed a new experiment, a transient high-pressure stepwise temperature-programmed surface reaction (STPSR) that enabled us to directly explore such complicated and challenging problems in FTS as fundamental knowledge about syngas activation, hydrocarbon and methane formation on pure-phase catalysts, which were previously unavailable. The comprehensive density functional theory (DFT) calculations revealed a deep insight into the intrinsic activity of iron metal and iron carbides on CO activation, C-C bond, and methane formation, rationalizing the kinetic and thermodynamic origin on the structural evolution of different iron-based catalysts during the FTS reaction. This deeper understanding of the active phase of iron-based catalysts and the corresponding FTS reaction mechanism is beneficial for rational design of more effective Fe-based catalysts on FTS by the synthesis of more Fe5C2 catalyst. Experimental Methods Synthesis of Fe7C3 and Fe2C NPs In a four-neck flask, 20 mL of N,N-dimethyloctadecylamine (for Fe7C3) or dodecylamine (for Fe2C) was stirred sufficiently and degassed under 120 °C for 2 h. Then, the system was refilled with NH3 and heated to 180 °C. After that, Fe(CO)5 (0.7 mL, 5.0 mmol) was injected under NH3 atmosphere and kept at this temperature for 30 min. A color change from orange to black was observed during the process, implying the decomposition of Fe(CO)5 and the nucleation of Fe nanocrystals. Subsequently, the mixture was further heated to 350 °C (for Fe7C3) or 260 °C (for Fe2C) at 10 °C/min and kept for 2 h before it was cooled down to room temperature. The product was washed with ethanol and hexane, and collected for further characterization. The as-synthesized nanoparticles (NPs) were kept in an Ar-filled glovebox to avoid exposure to air before further characterization. In the absence of NH3, the NPs would be oxidized to iron oxide ( Supporting Information Figures S1 and S2). Synthesis of Fe5C2 and Fe NPs In a four-neck flask, a mixture of octadecylamine (14.5 g) and Hexadecyltrimethylammonium bromide (CTAB) (0.113 g) was stirred sufficiently. Then, the system was refilled with N2 and heated to 180 °C. Following that, Fe(CO)5 (0.5 mL, 3.6 mmol) was injected under a N2 blanket. The mixture was kept at 180 °C for 10 min. A color change from orange to black was observed during the process, implying the decomposition of Fe(CO)5 and the nucleation of Fe nanocrystals. Subsequently, the mixture was further heated to 350 °C (for Fe5C2) or 300 °C (for Fe) at 10 °C/min and kept for 10 min before it was cooled down to room temperature. The product was washed with ethanol and hexane, and was kept in an Ar-filled glovebox to avoid exposure to air before further characterization. Preparation of supported catalyst The NPs obtained from high-temperature liquid-phase synthesis had been washed with n-hexane and ethanol several times and dispersed in n-hexane under N2 protection. Afterward, the dispersion of iron-carbide NPs was added into a certain amount of silica (N2 adsorption-desorption isotherm is shown in Supporting Information Figure S3) under stirring. After evaporating the solvent at room temperature, the supported catalyst was prepared. The amount of iron determined by inductively coupled plasma (ICP) was around 8%. Catalysis reaction The catalytic performance of the catalysts was evaluated in a fixed bed reactor. About 80 mg catalyst was loaded in a quartz-lined stainless steel reactor. The feed gas was a mixture of 32% CO, 64% H2, and 4% Ar. In a typical reaction, the pressure and gas flow rate were set for 30 bar and 20 mL/min (gas hourly space velocity, GHSV = 15,000 mL/g·h), respectively. Then, the reaction tube was heated from room temperature to 270 °C at 5 °C/min, and the reaction was conducted at 270 °C. It is worth noting that no reduction or carburizing pretreatment was carried out before the reaction. The gas-phase products were analyzed by an Agilent 6890 gas chromatography (GC) equipped with a flame ionization detector (FID) and thermal conductivity detector (TCD), with 4% Ar as inert standard. The heavier hydrocarbons were cooled down and collected in a trap, and analyzed offline by an Agilent GC 7820, with a HP-5 capillary column and FID. The products selectivity was calculated on a carbon basis. STPSR experiment Prior to the STPSR experiments, the Fe catalyst was treated in H2 (20 mL/min) and the Fe5C2 catalyst was treated in 10% C2H4/H2 mixture at 300 °C for 2 h to remove the surface contaminants. After cooling down to room temperature, the gas flow was switched to synthesis gas of 20 mL/min, and the pressure was raised to 30 bar. Then, the reactor was heated to 150 °C at 20 °C/min, and kept steady for 2 h. Afterward, the temperature was elevated 20 °C higher and held for 2 h at that temperature. The process was repeated until the reaction temperature reached 270 °C. A Pfeiffer Omnistar mass spectrometer (MS) was used to analyze the reactants and products online. The M/e value detected as follows: 2 for hydrogen; 15 for methane; 18 for water; 26, 27, and 30 for C2 products (acetylene, ethylene, and ethane); 28 for CO; 42 for C3; 44 for CO2; 56 for C4+; and 70 for C5+. Catalyst characterization The transition electron microscopy (TEM) experiments were conducted at a FEI Tecnai F30 transmission electron microscope (HRTEM) operating at 300 keV. The X-ray photoelectron spectroscopy (XPS) experiments were carried out on an AXIS Ultra imaging photoelectron spectrometer with Al kα as X-ray source. The binding energy of graphite carbon was calibrated to 284.8 eV. Powder X-ray diffraction (XRD) data were collected at a Rigaku D/Max-2400 equipped with Cu Al Kα radiation. The Raman characterizations were performed on a Renishaw 1000 Raman imaging microscope system with an excitation wavelength of 632.8 nm. The 57Fe Mössbauer effect spectra of as-synthesized iron-carbide NPs were collected by a Topological 500A spectrometer and a proportional counter at room temperature. The γ radiative source was a 57Co (Rh) moving with constant acceleration mode. The X-ray absorption fine structure (XAFS) spectra were collected at beam line 14 W of Shanghai Synchrotron Facility (SSRF) in transmission mode with a Si (111) monochromator. The samples for characterization were prepared and transferred under protection of Ar. As metallic iron and iron carbide were very sensitive to oxidation, the oxidation of the sample was sometimes unavoidable. For TEM study, the catalyst used was passivated in 0.5% O2 in Ar at room temperature for 1 h before the measurement. Calculation Spin-polarized DFT calculations have been performed by using the Vienna ab initio simulation package (VASP).46,47 Throughout the calculations, projector augmented wave (PAW) potential48 and the generalized gradient approximation (GGA) with the Perdew−Burke−Ernzerhof (PBE) exchange−correlation functional49 were adopted. The planewave cutoff energy was set by 400 eV. The force and energy convergence standards were 0.02 eV/Å and 1 × 10−4 eV, respectively. Monkhorst−Pack50k-points sampling of 3 × 7 × 7 and 10 × 10 × 10 were adopted for Fe5C2 and Fe bulk calculations with monoclinic (space group C2/c) and body-centered cubic (BCC) crystal structures, respectively. The optimized lattice constant of Fe5C2 is a = 11.55 Å, b = 4.50 Å, and c = 4.99 Å with β = 97.6°, which are consistent with experimental findings that a = 11.59 Å, b = 4.58 Å, and c = 5.06 Å with β = 97.75°.51 The calculated lattice constant (2.83 Å) for Fe with BCC crystal structure also agreed very well with the experiment.52 Fe-terminated Fe5C2 (100) surface was simulated by a slab of seven-layered Fe atoms and three-layered C atoms. In the calculations, the topmost four Fe and one C layers were fully relaxed, whereas the remaining atoms were fixed in their bulk positions. For the BCC-Fe (310) surface, a 10-layered Fe atoms slab model was used, and only the top five Fe layers were allowed to relax. A p(2 × 2) unit cell was utilized for the two slab models considered. We have used Monkhorst−Pack mesh k-points of 3 × 3 × 1 for Fe5C2 (100) and 5 × 5 × 1 for the BCC-Fe (310) surface. The vacuum region along the z direction was specified by 15 Å, and the dipole correction was considered in our calculations. Force reversed method53 was used to locate the transition states (TSs), and a force tolerance of 0.03 eV/Å was applied without zero-point correction. Some TSs along the minimum-energy reaction pathways were also reaffirmed by using the climbing image-nudged elastic band (CI-NEB) method.54 For a given elementary reaction, we considered the separately adsorption of intermediates at their most favorable adsorption sites as the initial and final states for the reaction barrier calculations. Results and Discussion In fabrication of both Fe7C3 and Fe2C NPs, NH3 was chosen as the atmosphere as well as the inducing agent while Fe(CO)5 was used as the precursor. In particular, Fe7C3 NPs were obtained in N,N-dimethyloctadecylamine solvent under 350 °C for 2 h, and Fe2C NPs were produced in dodecylamine solvent under 260 °C for 2 h (see Supporting Information). α-Fe and Fe5C2 NPs were synthesized via a bromide-induced process as described elsewhere.19 Figures 1a-1h present the powder XRD patterns and corresponding TEM images of as-synthesized Fe7C3, Fe2C, Fe5C2, and α-Fe NPs. According to the XRD results, the peaks in each sample consistent with the standard patterns, which suggests the generation of a single phase in each of the four samples. Furthermore, TEM images of the four samples indicated that they were all spherical NPs with diameters around 18 nm. Figure 1 | XRD patterns and TEM images of (a and e) Fe7C3, (b and f) Fe2C, (c and g) Fe5C2, and (d and h) α-Fe NPs. It is clear from the TEM images and XRD profiles that all the obtained Fe and iron-carbide NPs have the pure-phase structure and similar size (around 18 nm). Download figure Download PowerPoint The XAFS and Mössbauer spectra data also supported this conclusion. Fe K-edge XANES suggested that the iron-carbide particles exhibit very low oxidation states as compared with the metallic Fe foil. The relatively low-frequency oscillation of the postedge features indicated that the Fe central atom had neighbors with small bond lengths. Further extended XAFS (EXAFS) fitting results confirmed that all of the Fe-carbide particles synthesized had Fe-C coordination shells near 2.0 Å. The average first Fe-Fe shell bond length expanded from 2.46 to around 2.60 Å due to the incorporation of carbon into the BCC lattice of α-Fe (Figures 2a and 2b). Furthermore, no features of Fe oxides and Fe were observed, suggesting all the particles were pure carbide. (The Mössbauer spectra and corresponding fitting results are shown in Supporting Information Figure S4 and Table S1.) The sextet peaks indicated the formation of Fe2C, Fe7C3, and Fe5C2 NPs, while the weak doublet peaks in Fe5C2 suggested that the Fe5C2 NPs might have better crystallinity compared with Fe2C and Fe7C3. Therefore, the single-phase nature of these α-Fe and iron-carbide NPs, along with their similar morphology, provides us with an ideal platform for the investigation of their intrinsic catalytic behavior and structural evolution in the FTS process. At the same time, because the cementite (Fe3C) was reported to have poor activity in FTS,30,32 it was not discussed in this work. Figure 2 | (a) XANES and (b) EXAFS spectra of Fe5C2, Fe7C3, and Fe2C. The XANES and EXAFS spectra of Fe foil and standard pattern of Fe3O4 are shown by black and green curves, respectively. The curves composed by black circles in (b) indicate the experimental data of each iron carbide, and the colored curves indicate the simulation curves. The XAFS samples were prepared and transferred under protection of Ar. Download figure Download PowerPoint The α-Fe and iron-carbide NPs were dispersed on silica support and directly used in the FTS reaction (3 MPa syngas, 270 °C). For a 40 h reaction, the product distribution on Fe and Fe5C2 was similar except Fe had higher C5+ selectivity (49%). Fe2C had the highest selectivity toward CO2 (22.7%), while Fe2C and Fe7C3 showed considerably high selectivity of 20.1% and 19.5% toward methane, respectively ( Supporting Information Figure S5). Both CO2 and CH4 are undesired products. CO conversion on prepared catalysts with time on stream is shown in Figure 3 with FTS activity shown in Table 1. Figure 3 | CO conversion as a function of time over Fe, Fe2C, Fe7C3, and Fe5C2 catalysts (reaction conditions: 270 °C, 30 bar, 20 mL/min syngas). Download figure Download PowerPoint Table 1 | The Surface-Specific Activity (TOF) and Metal-Mass-Based Activity (Activity) Excluded CO2 Formation over Various Iron and Iron-Carbides Catalysts Catalysts Temperature (°C) Pressure (bar) TOS (h) TOFFTS (s−1) Activity (1 × 10−4 MolCO/gFe·s) References Fe 270 30 1 0.13 1.1 This work Fe 270 30 30 0.20 1.7 This work Fe5C2 270 30 1 0.29 2.3 This work Fe5C2 270 30 30 0.22 1.8 This work Fe7C3 270 30 1 0.16 1.6 This work Fe7C3 270 30 30 0.18 1.7 This work Fe2C 270 30 1 0.16 1.8 This work Fe2C 270 30 30 0.09 1.0 This work FexOy@C 270 20 — — 0.31 [55] Fe/SiO2 270 20 — — 0.20 [56] Fe-in-CNT 270 50 — — 2.5 [57] [email protected] 340 20 — 0.11 Fe 30 — — time on At the Fe5C2 the highest CO conversion (around by Fe2C and Fe7C3 while α-Fe had the CO conversion The initial CO the intrinsic catalytic of Fe2C, Fe7C3, and the reaction the activities of the four catalysts showed different For Fe5C2 and Fe2C, the CO conversion in the first h For Fe7C3, CO conversion in the first 5 h, and a small it the that of the α-Fe catalyst in the first 7 h, and constant the conversion for all catalysts the steady Fe5C2 the most active with a conversion of and the activities of Fe7C3 and Fe and were whereas Fe2C the active The (TOF) and activity of those Fe-based catalysts are in Table 1. Fe5C2 had the highest and activities and 1.8 × 10−4 After reaction, the of used Fe5C2 and Fe2C catalysts were with phases into a mixture of iron oxide and carbide, whereas for Fe7C3, the formation of Fe5C2 was observed ( Supporting Information Figures and the phase of used α-Fe catalyst was into Fe5C2, in to Fe3O4 ( Supporting Information Figure been a on and metal carbide is the active phase for FTS. To we Fe5C2 as a for the carbide phase and compared its catalytic behavior in the of FTS with that of α-Fe catalyst. We a high-pressure STPSR to allow the of reaction kinetics at high This STPSR experiment enabled us to the by reaction The formation of various products during the of syngas over the Fe5C2 catalyst is in Figure 150 °C, no products were which that syngas be over Fe5C2 at 150 °C. the reaction temperature 150 °C for 20 the formation of C2 hydrocarbons and and CO2 were the temperature was raised to °C, with the of the amount of and C2 was considerably to our methane was not detected until The for methane only the reaction temperature reached °C. The formation of methane at a higher temperature that of C2 hydrocarbons that the of toward methane is favorable C-C the temperature reached °C, by the formation of C5+ hydrocarbons 20 min Figure | STPSR on (a) Fe5C2 and (b) Fe Download figure Download PowerPoint To our the STPSR of α-Fe catalyst showed a behavior as compared with that of Fe5C2 At 150 °C, formation was observed, whereas no C2 hydrocarbons be detected in 2 h. C2 hydrocarbon only at around 10 min the temperature reached °C. The formation temperature of methane, and CO2 was at °C. and at and °C, which was very similar to those of C2 hydrocarbon on α-Fe at a higher temperature that on Fe5C2 the formation of methane on both catalysts a temperature of °C or methane formation on both catalysts compared with the formation of and C2 The formation of in STPSR process is as the on the catalysts was a pretreatment at 300 °C before Therefore, the can only be from the reaction and on the catalyst. are two for the product from CO or the from the gas phase that was not by pretreatment before In the source can be by a experiment ( Supporting Information Figure This that the source of on the α-Fe catalyst can only be the from Therefore, it is that CO at 150 °C over Fe which is also the for the Fe5C2 catalyst. carbon would be the At 150 °C, the formation of the hydrocarbons was observed on Fe5C2, not on α-Fe the first 2 h. This that carbon would on and might into the sites of the and the bulk α-Fe carburizing α-Fe toward carbide. Fe NPs can be into iron whereas the from the single and of Fe to iron carbide is which can be to the for carbon atoms into the sites by the of Fe atoms. this is the corresponding formation energy unit of Fe5C2 under FTS at this is in Supporting Information). The was confirmed by in XRD experiments under FTS reaction conditions MPa syngas, Figure It was that under a syngas which to a diffraction with a value of was to the iron-carbide phase value of most Fe5C2) increasing the reaction temperature higher °C. the reaction was not the same as in the STPSR experiments, this that in the FTS process, α-Fe is to with surface carbon from CO, iron carbide. we the supported α-Fe catalyst with or the mixture of and to supported iron-carbide catalyst as by XRD (Figures and we observed that the initial activity of the catalyst very to the initial activity of the pure-phase Fe5C2 catalyst. The activity evolution of the Fe catalyst that of Fe5C2 that it in the first and relatively stable around 15 h of reaction. The observed on the α-Fe catalyst be to the process of the was and Fe5C2 was the catalysts would showed a higher activity because of the higher intrinsic activity of Figure 5 | In XRD patterns of Fe NPs treated with 2 MPa syngas at various was in beam line of The the from in Download figure Download PowerPoint For the of activity of Fe5C2 catalyst with reaction time it was to the oxidation of iron carbide by the products such as and CO2 to the iron This was confirmed by the activity of pure-phase Fe5C2 treated with CO2 at different (see Supporting Information Figure both the initial crystal phase of the catalyst and the reaction atmosphere the structural evolution and the structure of iron-based catalysts under reaction Figure | (a) CO conversion of Fe5C2, and α-Fe NPs XRD pattern of α-Fe NPs catalysts (b) before and It is clear that carburizing the supported α-Fe catalyst with or the mixture of and at 350 °C for 1 h would α-Fe to Download figure Download PowerPoint To the experimental results presented it is essential to study the intrinsic FTS activity of the iron and the with the iron carbide. DFT calculations were performed to study the FTS including CO activation, C-C and methane formation. The Fe (310) and Fe-terminated Fe5C2 (100) ( Supporting Information Figure were used to model the corresponding iron and catalysts, respectively. The Fe (310) was chosen because the Fe (310) surface a of surface of the iron by and it higher activity for CO as compared with the and (111) Fe5C2 (100) was it is one of the under operating FTS conditions 10 bar, = on Fe5C2 a Fe5C2 surface, be no active sites for CO activation or the activity of CO activation be