Unravelling the Anomalous Coking Resistance over Boron Nitride-Supported Ni Catalysts for Dry Reforming of Methane

Open AccessCCS ChemistryRESEARCH ARTICLES30 Nov 2022Unravelling the Anomalous Coking Resistance over Boron Nitride-Supported Ni Catalysts for Dry Reforming of Methane Jiang Deng†, Min Gao†, Jun-ya Hasegawa, Xiaoyu Zhang, Aiyong Wang, Aling Chen and Dengsong Zhang Jiang Deng† International Joint Laboratory of Catalytic Chemistry, College of Sciences, Shanghai University, Shanghai 200444 †J. Deng and M. Gao contributed equally to this work.Google Scholar More articles by this author , Min Gao† Institute for Chemical Reaction Design and Discovery (WPI-ICReDD), Hokkaido University, Sapporo 001-0021 Institute for Catalysis, Hokkaido University, Sapporo 001-0021 †J. Deng and M. Gao contributed equally to this work.Google Scholar More articles by this author , Jun-ya Hasegawa Institute for Catalysis, Hokkaido University, Sapporo 001-0021 Google Scholar More articles by this author , Xiaoyu Zhang International Joint Laboratory of Catalytic Chemistry, College of Sciences, Shanghai University, Shanghai 200444 Google Scholar More articles by this author , Aiyong Wang International Joint Laboratory of Catalytic Chemistry, College of Sciences, Shanghai University, Shanghai 200444 Google Scholar More articles by this author , Aling Chen International Joint Laboratory of Catalytic Chemistry, College of Sciences, Shanghai University, Shanghai 200444 Google Scholar More articles by this author and Dengsong Zhang *Corresponding author: E-mail Address: [email protected] International Joint Laboratory of Catalytic Chemistry, College of Sciences, Shanghai University, Shanghai 200444 Google Scholar More articles by this author https://doi.org/10.31635/ccschem.022.202202342 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesDownload CitationsTrack Citations ShareFacebookTwitterLinked InEmail Metal oxides have been used as the supports for heterogeneous catalysis for many years, but they still suffer from coking in some high-temperature applications. The main reasons for coking are the uncontrollable dissociation of C–H and the overbalance between carbon deposition and removal. Herein, we find a boron nitride (BN)-immobilized Ni catalyst shows unprecedented coking resistance in dry reforming of methane via the incomplete decomposition of methane. Unlike the Ni-based catalysts supported by traditional metal oxides, BN-supported Ni accelerates the first C–H dissociation while inhibiting the breaking of the final C–H bond; hence, the suppression of the complete decomposition of methane thoroughly addresses the coking issue. This work reveals the fundamental reason for the coking resistance over BN-supported Ni catalysts is selective activation of the C–H bond, which can provide an inspiring idea for other applications. Download figure Download PowerPoint Introduction Supported catalysts have been used for heterogeneous catalysis for a long time.1,2 The original intent of using supported catalysts is the high dispersion and exposure of active sites enabled by the support while maintaining the heterogeneity.3 Realistically, the performance of supported catalysts is significantly influenced by the properties of the supports because they can be involved in the catalysis process through the adsorption/activation of reactants and/or regulation of the electron structure of active metals. Metal oxides, such as γ-Al2O3, MgO, SiO2, ZrO2, CeO2, La2O3, and TiO2, which are equipped with plentiful surface acidity/basicity and/or excellent reducibility, have been widely used as the supports of catalysts.4–6 Moreover, the strong metal–support interaction (SMSI) and electronic metal–support interactions have been particularly explored based on some metal-oxide supported catalysts due to their effective regulation of the activity, selectivity, and stability of catalysts.7–9 Nevertheless, it remains a great challenge to address the coking issue among catalytic processes requiring high temperature such as the Fischer–Tropsch synthesis,10,11 dehydrogenation of alkane,12 and dry reforming of methane (DRM).13,14 Principally, carbon deposition happens when the C–H bond undergoes uncontrollable decomposition during alkane reforming or the dehydrogenation processes.15–17 The carbon deposition is usually overcome by passivating the active sites, which leads to inferior activities.18,19 Other options include accelerating the gasification of formed carbon by introducing oxidized gas or improving the adsorption and activation of CO2, and so on.20–22 In these cases, it is very tricky to balance the carbon deposition and removal and the compromised metallic active sites resulting from oxidized gas induced deactivation.23 More recently, De Jong et al.24 found the iron dispersed on a relatively inert support such as α-Al2O3, SiC, and carbon nanofibers show high activity, robust stability, and less coking than on γ-Al2O3. The weakly interactive support was proposed to contribute to the formation of iron carbides, which facilitated the reaction and alleviated the coking problem.24 More recently, Song et al.25 proposed that NiMo supported by single-crystal MgO, which has fewer defects, shows better activity and coking resistance than on traditional polycrystalline MgO. Additionally, Xie et al.26,27 found the porous single-crystalline MgO and CeO2 with higher surface-area supported Ni catalysts exhibited super coking resistance and robust stability. Although the mechanism for the coking resistance is still unclear, the unexpected performance over the less defective supports suggests a promising way to solve the carbon deposition by inert supports in heterogeneous catalysis. Hexagon boron nitride (BN), known as “white graphite,” shows excellent heat conductivity and robust thermostability.28,29 The ideal BN has a relatively inert surface, and the absence of O in BN compared with metal oxides relieves the effect of active atmosphere. Recently, BN exhibited a distinctive performance for oxydehydrogenation of alkane.30,31 Moreover, a Ni-based catalyst supported by BN was reported to display good coking-resistance properties for the DRM reaction.32 The Ni-based catalysts supported by the defective BN proposed by our group displayed super coking resistance and sintering resistance for DRM, but the underlying coking-resistance mechanism is still unclear.33,34 Dong et al.35 even proposed the reaction-induced SMSI effect over the BN-supported Ni catalysts, which improves the DRM performance. Although many investigations have shown BN-supported metal catalysts show better coking resistance compared with active-oxide supported catalysts, a lack of comprehension of the underlying specific mechanism remains. Moreover, the weak metal–support interaction would lead to particle sintering (>10 nm) during catalyst preparation or reaction process in which the support effects would be weakened with large Ni nanoparticles (NPs), making it difficult to understand the coking-resistance mechanism. Herein, functional BN was selected as the support and Ni NPs with size around 3.5 nm can be confined by self-orientation onto the boundary of BN (Ni/BNf). The DRM reaction was used to evaluate the performance of the Ni-based catalysts. The Ni/BNf catalysts show 1.2 times and 13 times higher turnover frequency of CH4 ( TOF CH 4 ) than that of α-Al2O3- and SiO2-supported Ni catalysts, respectively. The higher intrinsic activity of the Ni/BNf catalyst is attributed to the lower activation energy of the C–H decomposition and better activation of the CO2. Compared with the traditional oxide-supported Ni-based catalysts, the first C–H dissociation is accelerated while the craking of the final C–H bond of CH4 is inhibited in Ni/BNf, thus suppressing the complete decomposition of the CH4. Moreover, the preferential adsorption and activation of CO2 in the presence of CH4 also reduce the probability of interaction between CH4 and metallic Ni. As a result, the coking issue can be exhaustively overcome showing no carbon deposition even using pure CH4 as reactants. Our results suggest the BN can serve as a novel support by tuning the reaction pathway, which might be available in many other challenging reactions. Experimental Methods Materials and chemicals Nickel nitrate (99.5%), SiO2 (A.R.), and α-Al2O3 (A.R.) were bought from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Pristine boron nitride (BNp) and functionalized BN (LEAU3002) (BNf) were purchased from Saint-Gobain (Shanghai, China). BNf was annealed at 600 °C for 3 h in air at a ramp rate of 10 °C/min and the obtained sample was denoted as A-BNf. The other supports were used as received. Preparation of catalysts BNf, A-BNf, BNp, α-Al2O3, and SiO2 supports (0.5 g) were dispersed into 30 mL deionized (DI) water, and 1 mL of Ni(NO3)2 (0.0257 M) solution was added, and the mixture was stirred for 12 h at room temperature. The suspension liquid was vaporized through rotary evaporation at 45 °C and dried at 80 °C overnight. The powder was calcined at 550 °C for 6 h in air at a ramp rate of 2 °C/min. The metal loading was varied by adjusting the amount of Ni2+ solution and was determined by inductively coupled plasma-optical emission spectroscopy (ICP-OES). Before reaction, all the catalysts were reduced by 10% H2:N2 at 30 mL/min at 750 °C for an hour at a ramp rate of 10 °C/min. The samples were named as Ni/BNf, Ni/A-BNf, Ni/BNp, Ni/SiO2, and Ni/α-Al2O3 corresponding to specific supports. The loading amount of SiO2 in Ni/BNf and Ni/A-BNf was around 0.4 wt % as determined by the ICP-OES. Catalytic performance evaluation The catalytic activity and stability tests of the catalysts were conducted in a quartz fixed-bed tubular reactor with a total gas-flow rate kept constant at 50 mL/min (CO2∶CH4 = 25∶25, WHSV = 25,000 mL g−1 h−1). Catalysts (0.12 g, 20–40 mesh) were diluted with 0.6 g quartz sand (20–40 mesh), and the diameter of the reactor was 8 mm. The stability test was carried out at 750 °C for 10 and 100 h, respectively. An online gas chromatograph equipped with a thermal conductivity detector analyzed the products. The conversion of CH4 and CO2 was calculated based on the following equation: C CH 4 = F CH 4 , in − F CH 4 , out F CH 4 , in × 100 % C CO 2 = F CO 2 , in − F CO 2 , out F CO 2 , in × 100 % where F CH 4 , in , F CO 2 , in , F CH 4 , out , and F CO 2 , out are the inlet and outlet volume flow rates of CH4 and CO2, respectively. The TOF CH 4 was tested at weight-hourly space velocity (WHSV) of 108,000 mL g−1 h−1. Catalysts (50 mg) were diluted by 0.6 g 20–40 mesh quartz and put in a quartz fixed-bed tubular reactor with a total gas flow-rate kept constant at 90 mL/min (CO2∶CH4 = 45∶45, mL/min). Before reaction, the calcined sample was reduced at 750 °C for 1 h at a ramp rate of 10 °C/min. TOFs (s−1) were calculated from the following: TOF CH 4 = n [ CH 4 ] Weight of catalyst × w Ni ÷ M Ni × Ni dispersion × 60 where wNi is the weight percent of Ni over catalyst, MNi is 58.69 g/mol, and Nidispersion is determined by H2 chemisorption experiments. Characterization of Catalysts In situ diffuse reflectance infrared Fourier transform spectra In situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTs) experiments were carried out on a Nicolet is50 FT-IR spectrometer (Massachusetts, USA). For the in situ DRIFTs transient reaction of CH4 and CO2, the catalysts preadsorbed CO2 (2 mL/min) for 1 h, and then a CH4 (2 mL/min) stream was introduced with flowing N2 (40 mL/min) during the whole test. The reaction temperature was 550 °C, and spectra were collected. Second, the catalysts preadsorbed CH4 (2 mL/min) at 500 °C, and then CO2 (2 mL/min) stream was introduced to examine the active species. In situ DRIFTS CO2 and CH4 temperature-programmed reactions were carried out in a gas mixture with CH4∶CO2∶N2 = 2∶2∶40 mL/min from 300 °C to 700 °C after pretreatment under N2 for 30 min. The CH4 and CO2 temperature-programmed decompositions were performed under CH4∶N2 (CO2∶N2) = 10∶50 mL/min. Operando Raman spectrometry Samples (10–20 mg) were put into the ceramic tube of the Linkam reactor. The Raman experiments was carried out on the LabRAM HR Evolution by using a 325 nm laser (the ND filter was 10%, hole value was 100). The spectra were acquired in the range of 1000 to 2000 cm−1. The spectral resolution employed was 4 cm−1. The operando Raman was conducted in a gas mixture with CH4:CO2 = 10∶10 mL/min. The off-gas was detected by on-line MS (Pfeiffer Omnistar, Asslar, Germany). Computational details The theoretical calculations were done by density functional theory (DFT) with the functional of Wu and Cohen implemented in SIESTA code.36–38 Double-ζ plus polarization function (DZP) basis sets are used to treat the 2s22p1, 2s22p3, 2s22p4, 1s1, and 4s23d8 valence electrons of B, N, O, H, and Ni atoms, respectively.39,40 The remaining core electrons are represented by the Troullier–Martins norm-conserving pseudopotentials41 in the Kleinman–Bylander factorized form.42 The h-BN surface was represented by a single layer slab with 7 × 7 element of h-BN. Periodic boundary conditions were used for all systems. The energy cutoff of 200 Ry was chosen to guarantee convergence of the total energies and forces. All calculations were spin polarized. A common energy shift of 10 meV was applied. Bader (atoms in molecules [AIM])43,44 method was used to analyze the charge transfer. The adsorption energy of Ni10 on BN is defined as Eb = E(Ni10) + E(BN) − E(Ni10/BN) where E(X) denotes the electronic energy. The adsorption energy is defined as Eb = E(M/Ni10@BN) − E(Ni10@BN) − E(M), where M denotes CHx (x = 0–4) or CO2. The reaction energy is defined as Erec = Etot + yENi10-H – ECHx – (y + 1)ENi10 (y = 0–4). More negative values of adsorption energy and reaction energy indicate a higher stability of the total system. All the calculations were performed at Г point of the Brillouin zone. The energy cutoff of 500 Ry was used to calculate charge density. The visualization of electron density difference was done by Visualization for Electronic and Structural Analysis (VESTA).45 GRRM17 was used to calculate transition state.46,47 Results and Discussion Activity performance of catalysts for DRM reaction The BNf is the hexagonal BN functionalized with dimethylsiloxane–ethylene oxide block copolymer (DMSEO) which were purchased from Saint-Gobain and used as received. The DMSEO underwent decomposition at 200 °C as confirmed by the thermogravimetric analysis (TGA) ( Supporting Information Figure S1). Compared with the BNp (Saint-Gobain), the BNf shows better hydrophilism ( Supporting Information Figure S2), which is beneficial for Ni loading via the impregnation processes. When the BNf was annealed at 600 °C, the functional groups of BNf decomposed according to the TGA results, and the obtained support, which was labeled as A-BNf, still possessed the hydrophilic property ( Supporting Information Figure S2). Based on the atomic force microscopy results ( Supporting Information Figure S3), the thickness of the BNf decreased from 2.96 ± 0.34 to 2.26 ± 0.32 nm for A-BNf, which further confirms the decomposition of functional groups in BNf. The conversion of the CH4 increased linearly with the metal loading ranging from 0.1 to 0.5 wt % but with only a slight increase when the mass loading increased by 5 wt % because of the limitation of gas diffusion at the current mass space velocity ( Supporting Information Figure S4). Hence, the Ni/BNf catalyst with 0.3 wt % Ni was used as the benchmark compared with the other commercial support-loaded Ni catalysts (Ni/A-BNf, Ni/BNp, Ni/α-Al2O3, and Ni/SiO2) to exclude the limitation of gas diffusion. As shown in Figure 1a, the conversion of the CH4 over the Ni/BNf catalyst increased along with the reaction temperature, and the gap between the thermodynamic equilibrium also decreased. The initial conversion rates of CH4 are in following order: Ni/BNf ≈ Ni/α-Al2O3 > Ni/A-BNf > Ni/BNp > Ni/SiO2 (Figure 1b). The ratio of H2:CO was around 0.90 for the Ni/BNf catalysts, which suggests the reverse water gas reaction happened ( Supporting Information Figure S5). The H2:CO ratio of the Ni/SiO2 catalyst was much lower than 1, which suggests a serious side reaction. Actually, the metal dispersion over the BN-supported catalysts is similar with that of Ni/α-Al2O3 and Ni/SiO2 catalysts ( Supporting Information Table S1). The conversion rates of the CH4 over the Ni/BNf catalyst remained constant during the 10 h DRM test, whereas a gradual decline occurred with the Ni/α-Al2O3, Ni/A-BNf, Ni/BNp, and Ni/SiO2 catalysts. According to the TGA results of the spent catalysts ( Supporting Information Figure S5d), the carbon accumulation is only observed for Ni/BNp and Ni/α-Al2O3. The decline of activity might be due to the carbon accumulation and metal sintering. To better understand the intrinsic activity, the TOF CH 4 was measured. The TOF CH 4 of the Ni/BNf catalyst was 8.7 s−1 at 650 °C, which is slightly higher than that of the Ni/α-Al2O3 catalyst (7.0 s−1) and is 13 times higher than that of the Ni/SiO2 catalysts (Figure 1c). Moreover, the TOF CH 4 of the Ni/BNf catalyst is three times and six times higher than Ni/A-BNf and Ni/BNp catalysts, respectively (Figure 1c). Additionally, the Ni/BNf catalyst is better than the other oxide-supported Ni catalysts reported in the literature ( Supporting Information Table S2). The DRM reaction mechanism over the Ni/α-Al2O3 catalyst is sensitive to the temperature,48 and the activity decreased more rapidly along the reaction temperature than that of Ni/BNf catalysts ( Supporting Information Figure S6). Arrhenius plots ( Supporting Information Figure S7) and the apparent activation energy (Ea) for catalysts are shown in Figure 1d. The Ea calculated based on the CH4 conversion is 30.6 ± 3.2 kJ/mol for the Ni/BNf catalyst, which is much lower than that of Ni/α-Al2O3 (60.6 ± 0.4 kJ/mol), Ni/A-BNf (68.2 ± 3.3 kJ/mol), and Ni/BNp (82.6 ± 1.7 kJ/mol) catalysts indicating the fast kinetics of the DRM reaction in the Ni/BNf catalyst. As proposed by Cui et al.,48 surface oxygen species can react with CHx species resulting in low activation energy while the reaction of CHx with CO2 the reaction The much lower activation energy of the Ni/BNf catalyst ( Supporting Information Table suggests the reaction mechanism is from other metal oxide-supported Ni catalysts. Additionally, when the Ni/BNf catalyst was with for 100 h, the conversion of the CH4 over the Ni/BNf catalyst remained constant (Figure The robust stability of the Ni/BNf catalyst for the DRM reaction as a catalyst to the reaction mechanism of the Figure 1 of CH4 conversion over Ni/BNf catalyst as a function of reaction temperature with WHSV of 25,000 mL of conversion rate of CH4 tested at 750 °C with WHSV of 25,000 mL g−1 h−1. of TOF CH 4 over Ni/BNf, Ni/BNp, Ni/A-BNf, Ni/α-Al2O3, and Ni/SiO2 catalysts tested at WHSV of 108,000 mL of activation energy calculated (Ea) based on the conversion of CH4 over Ni/BNf, Ni/BNp, Ni/A-BNf, Ni/α-Al2O3 catalysts. of CH4 conversion of Ni/BNf catalyst as a function of on stream in which the catalyst was after the first 50 h and then for 50 h test. Download figure Download PowerPoint Characterization of supported Ni-based catalysts is that the Ni NPs have been into the of the BN with an size of 3.5 ± nm as displayed in the electron microscopy (Figure and Supporting Information Figure According to the the Ni NPs on the Ni/BNf catalyst the with the of and the BNf the (Figure to The of the support a structure of hexagonal BN (Figure The of Ni NPs was further via coupled with the energy spectroscopy (Figure A of and was found for the support and the Ni NPs were at the boundary of the Moreover, and O were with which suggests the SiO2 from decomposition of the DMSEO on the BNf, the of Ni NPs ( Supporting Information Figure the of Ni NPs with an size of ± over the Ni/A-BNf catalyst ( Supporting Information Figure suggests the of the functional groups for the formation of Ni NPs on the boundary of the BNf. The similar particle size of Ni NPs in the Ni/BNp ( Supporting Information Figure with that of the Ni/BNf catalyst suggests the in Ni/BNf and Ni/BNp catalysts. DRM reaction, the size of the Ni NPs increased by ± nm ( Supporting Information Figure which shows better sintering resistance than that of the Ni/A-BNf and Ni/BNp catalysts ( Supporting Information Figure The slight sintering of Ni in Ni/BNf the conversion of which is also observed in other ( Supporting Information Table S4). Dong et al.35 proposed that the reaction-induced strong interactions happened for the BN-supported Ni catalysts with the of the which is also observed in the spent Ni/BNf catalyst after DRM reaction ( Supporting Information Figure DRM reaction, were still some Ni NPs on the boundary of the BN ( Supporting Information Figure The Ni NPs in the Ni/SiO2 and Ni/α-Al2O3 catalysts are also where the Ni sintering happened to the Ni/SiO2 and Ni/α-Al2O3 catalysts ( Supporting Information Figure Figure 2 and of Ni/BNf fast Fourier transform of the of the support in Ni/BNf of Ni/BNf and the corresponding of Ni/BNf catalyst with of N, B, and O spectra of Ni over Ni/BNf and spent Ni/BNf catalysts after 10 h DRM test. Download figure Download PowerPoint The energy of the and the of after the heat indicating a interaction happens to the SiO2 and the where the SiO2 only as the ( Supporting Information Figure When the Ni NPs were the energy of and on the Ni/BNf catalyst in similar to the Ni/A-BNf catalyst. the reason for this difference to the boundary of the BN in the hence, some of the Ni NPs are in the SiO2 but the Actually, is no between the Ni2+ and the BNf ( Supporting Information Figure and the of the BN would the solution to on the boundary of BN that is by the The DMSEO underwent decomposition in which the formed SiO2 the Ni NPs during the process ( Supporting Information Figure the spectroscopy of Ni were the of the because of the low Ni the property of Ni NPs in the Ni/BNf and spent Ni/BNf catalysts after DRM reaction were via spectroscopy as shown in Figure The of Ni is at corresponding to the metallic DRM reaction, the Ni which suggests the Ni NPs the metallic be that the BNf the DRM reaction because no conversion of CH4 occurred over BNf in the absence of Ni ( Supporting Information Figure In the Ni NPs in the Ni/BNf catalyst were on the boundary of the which would the metal sintering during the DRM reaction, and the oxidized of Ni NPs these effects on the carbon Operando Raman and surface spectrometry on coking-resistance To and the carbon operando Raman was conducted over the Ni/BNf catalyst. The Ni/α-Al2O3 catalyst was used as the sample because of initial DRM activity to that of Ni/BNf catalyst. was no conversion of CH4 by catalysts due to the formation of the carbon decomposition happened due to the properties of ( Supporting Information Figure The and in the Raman spectra from and of was used to the formation of the For the Ni/BNf catalyst, the of the was used to the formation of carbon because of the BN at which with the CH4 and CO2 were during the first min. According to the results of the mass spectrometry (Figure and Supporting Information Figure H2 and CO were detected with the of CH4 and CO2, indicating the DRM reaction. The conversion of CH4 and CO2 over the Ni/BNf catalyst was calculated as and which was higher than that of the Ni/α-Al2O3 catalyst of CH4 and of Nevertheless, the was observed over the Ni/BNf catalyst during the DRM reaction process (Figure a after 10 over the Ni/α-Al2O3 catalyst (Figure and then This that the CH4 and CO2 underwent the DRM reaction mechanism where the formed carbon with CO2. According to the thermodynamic the decomposition of CH4 the main reason for the carbon deposition at 750 Hence, the decomposition of CH4 over the Ni/BNf and the Ni/α-Al2O3 catalysts was further explored by with pure CH4 stream under 750 The lower mass spectrometry of the CH4 and the higher MS of the H2 compared with the over Ni/BNf and Ni/α-Al2O3 catalysts (Figure and Supporting Information Figure indicate the decomposition of the CH4 where the lower MS indicate better decomposition activity of the CH4 over the Ni/BNf catalyst. The conversion of the CH4 over the Ni/BNf catalyst was which is higher than that of the Ni/α-Al2O3 catalyst the decomposition of the was in the Raman spectra of the Ni/BNf catalyst, which suggests a structure and no carbon deposition (Figure the over the Ni/α-Al2O3 catalyst and the increased the also along with the the CH4 underwent complete decomposition over the Ni/α-Al2O3 catalyst but the Ni/BNf catalyst. Figure 3 Operando Raman spectra over Ni/BNf and Ni/α-Al2O3 catalysts. Operando Raman spectra with on-line MS results during operando Raman in which the and is the and the on-line results, respectively. Ni/BNf catalyst in flow of mL/min) at 750 °C for and then the CO2 for min. Operando Raman spectra over Ni/α-Al2O3 catalyst in flow of CH4:CO2 mL/min) at 750 °C for and then the CO2 for min. and corresponding Raman based on the of the of spent Ni/BNf catalyst after decomposition of CH4 at 750 °C for 5 and corresponding Raman based on the of the of spent Ni/α-Al2O3 catalyst after decomposition of CH4 at 750 °C for 5 at the gas flowing and and the corresponding results of over Ni/BNf and Ni/α-Al2O3 catalysts. Download figure Download PowerPoint To better understand the decomposition mechanism of the the Ni/BNf and Ni/α-Al2O3 catalysts were with the