Baowang
Lu
*,
Mitsuhiro
Inoue
and
Takayuki
Abe
Hydrogen Isotope Research Center, Organization for Promotion of Research, University of Toyama, Gofuku 3190, Toyama 930-8555, Japan. E-mail: baowangl@ctg.u-toyama.ac.jp
First published on 10th January 2018
Traditional low-temperature H2 production from methane needs high operating temperatures (above 500 °C), generates CO2 emissions, and shows poor stability. This study focused on more low-temperature stable H2 production from methane using a carbon-supported Ru catalyst (BS-Ru/C). This catalyst could stably generate H2 without COx at 200 °C over 15 days. The temperature for clean H2 production in conjunction with carbon fixation was found to be 200–250 °C, far lower than around 550 °C for the conventional low-temperature methane catalytic decomposition (MCD). The BS-Ru/C catalyst enabled the occurrence of steam reforming of methane (SRM) at 260 °C, which far exceeded other previous low-temperature catalysts reported at around 500 °C. Although CO2 was produced during the SRM process from 260 to 500 °C, no CO was produced. This lower-temperature stable H2 production technology using the BS-Ru/C catalyst with outstanding ability and stability will provide a helping hand to build a H2 society in near future.
MCD:
CH4 → C + 2H2 ΔH = +74.8 kJ mol−1 |
SRM:
CH4 + H2O ↔ CO + 3H2 ΔH = +206 kJ mol−1 |
Generally, for preparing dispersed metal nanoparticles on supports, conventional methods often result in liquid waste emission and require a heating process. Therefore, a “dry” technique without any liquid waste emission and heating, a “polygonal barrel-sputtering method,” has been developed as an environment-friendly catalyst preparation method.19–22 Additionally, the metal nanoparticles prepared by the polygonal barrel-sputtering method exhibit higher catalytic performance than those obtained by conventional methods.22 Based on the above considerations, in order to achieve success in lower-temperature stable H2 production from methane via SRM and MCD processes, the polygonal barrel-sputtering method was employed to prepare sputtered Ru nanoparticles on carbon as a catalyst (BS-Ru/C), and lower-temperature stable H2 production without COx emissions (clean H2) or with less COx emissions at 200–500 °C using it was investigated.
A reference Ru/C catalyst (referred to as Ref.-Ru/C, with a Ru deposition amount of 0.8 wt%) was also prepared using a conventional wet process for comparison purposes. In this process, 0.060 g of ruthenium(III) chloride n-hydrate (RuCl3·nH2O, Ru: 40 wt%, Kanto Chemical, Japan) was dissolved in 7 ml of water, after which 3 g of carbon powder was added to the solution and the resulting slurry was dried at 80 °C for 24 h. Finally, the sample was reduced by heating to 350 °C at a rate of 100 °C h−1 and keeping it at this temperature for 4 h in an 80% H2/N2 gas stream (flow rate: 50 ml min−1).
The chemical state of the Ru particles was evaluated using X-ray photoelectron spectroscopy (XPS, Thermo Fisher Scientific, ESCALAB 250X, T) at 25 W and 25 °C with monochromatized Al Kα excitation. Samples for XPS were prepared by mixing Au particles purchased from Niraco with the Ru/C, and the resulting spectra were normalized relative to the 84 eV Au 4f7/2 peak. The compositions of the Ru species were estimated from the areas of each peak, using a sensitivity factor associated with the XPS instrumentation.
Catalysts before and after use were assessed by temperature-programmed reduction (TPR) with a BEL-CAT Catalyst Analyzer (BEL Japan Inc.). Each sample (approximately 50 mg) was pretreated for 1 h at 120 °C in an Ar flow (30 ml min−1) followed by cooling to 40 °C in an Ar stream (30 ml min−1). After Ar was replaced by a mixture of 3 vol% H2 in Ar (total flow rate: 30 ml min−1), the temperature was increased from 50 to 950 °C at 10 °C min−1. All TPR data were normalized to the respective sample weight and expressed in arbitrary units.
The deposition state of Ru particles was observed using conventional transmission electron microscopy (TEM, JEOL, JEM2100 and JEM2000EXII) at 200 kV. Catalyst use was also examined by Cs-corrected scanning transmission electron microscopy (Cs-corrected STEM) at 80 kV using a high-resolution TEM instrument (JEOL, JEM-ARM200F).
Because methane catalytic decomposition could not stably occur below 180 °C using the BS-Ru/C catalyst due to deactivation, its catalytic stable decomposition was carried out from 200 to 500 °C. In the inset in Fig. 1(A), at the S/C ratio of 0, no H2 stable generation occurred from 200 to 340 °C, while at the S/C ratio of 28.6, H2 was stably produced at a surprisingly low temperature of 200 °C, and its yields increased with increasing temperature. This temperature of stable H2 generation with the BS-Ru/C catalyst is far lower than that (around 550 °C) previously reported for the MCD process.7–9 As shown in Fig. 1(A), increasing the temperature from 200 to 250 °C, the CH4 conversion and H2 yield increased from 8.6% and 2.0% to 15.9% and 4.8%, respectively. Unexpectedly, no COx was obtained at this temperature region. It can be considered that water did not supply active oxygen even though it was used, showing that a complete low-temperature MCD process occurred. According to a previous report,23 our H2 product at this temperature region can directly applied for polymer electrolyte membrane fuel cells (PEMFCs) and alkaline fuel cells (AFCs) without COx removal. To the best of our knowledge, this is the first-ever demonstration of clean H2 production in conjunction with carbon fixation, in which neither CO2 nor CO was produced in the low temperature range of 200–250 °C. From 260 °C, CH4 conversion and the H2 yield further increased with increasing temperature, while CO2 was also generated with a gradually increasing yield, indicating that water began to supply active oxygen and the SRM process occurred. Therefore, CO2 production derived from CO through water gas shift reaction (WGSR) was dependent on temperature. The temperature at which this SRM process occurred is far lower than previous reports (around 500 °C).10–16 At 500 °C, CH4 conversion, the H2 and CO2 yields reached 100%, 75%, and 60%, respectively. Although CO2 production could not be avoided from 260 °C, it is worthy to note that the BS-Ru/C catalyst did not generate CO, which reduces the performances of PEFCs, irrespective of temperature. This result suggests that although CO was produced, it was completely converted to CO2 through WGSR.
WGSR:
CO + H2O ↔ CO2 + H2 ΔH = −41 kJ mol−1 |
As shown in Fig. 1(B), the H2/CH4 ratio increased linearly with increasing temperature, while increased slowly in the temperature range of 420–440 °C, and then decreased slowly with increasing temperature from 440–500 °C. A H2/CH4 ratio of <2 strongly indicates that methane was partially decomposed, and hydrocarbon simultaneously resulted as the main solid product other than the H2 gas product, while a H2/CH4 ratio of ≥2 means that methane was completely decomposed. When the H2/CH4 ratio was ≈2, carbon was the main solid product. From 260 to 420 °C, the H2/CH4 ratio gradually increased from 1.5 to 3, suggesting that a mixed process (MCD + SRM) occurred, but was essentially constant at 3 at 440–500 °C, indicating a complete SRM process. Therefore, the H2/CH4 ratio and the reaction process were dependent on temperature. The activation energies were determined to be 45.4 kJ mol−1 at 200–250 °C (MCD process), 30.4 kJ mol−1 at 260–420 °C (MCD + SRM process) and 11.6 kJ mol−1 at 440–500 °C (SRM process), which are approximately one half to one tenth of those previously reported for MCD (59–92.4 kJ mol−1)24–27 and SRM processes (91–144 kJ mol−1).11–13 These very low activation energies clearly confirm the very high catalytic activity of the BS-Ru/C catalyst.
Subsequently, the catalytic stability was tested at 200, 340, and 400 °C. In Fig. 2, H2 generation was stable over 15 days at 200 °C, although its yield was relatively low at 2.5%. At 340 °C, the H2 yield increased to approximately 45% and kept constant over 15 days, while at 400 °C, the highest yield of about 60% was obtained after 9 days. It is well known that the best catalyst reported in the MCD process was maintained for only 70 h,28 therefore, our BS-Ru/C catalyst exhibited excellent catalytic stability, as well as high catalytic activity, which lowers the temperature required for H2 generation. It is also worth noting that the H2/CH4 ratio was 0.5 at 200 °C, 2 at 340 °C, and 2.5 at 400 °C (the inset of Fig. 2), indicating the temperature dependence of the H2/CH4 ratio, further suggesting that the H2 generation pathway may change with temperature.
Thus, as shown in Fig. 3, it is concluded that H2 generation over BS-Ru/C includes three pathways: (I) partial methane direct decomposition (partial MCD process): CH4 → CHx + ((4 − x)/2)H2 in the low temperature range of 200–250 °C, different from the conventional MCD process (complete methane direct decomposition);7–9 (II) simultaneous methane direct decomposition and steam reforming in the moderate temperature region of 260–420 °C, and complete methane direct decomposition occurring only at the H2/CH4 ratio of 2; and (III) low temperature methane steam reforming in the temperature region of 440–500 °C. The H2 yields were recalculated based on the maximum H2/CH4 ratios in each temperature region (200–250 °C: 2, 260–500 °C: 3). As a result, the H2 yield in Fig. 3(A) was determined to be approximately 100% at 500 °C. To our knowledge, there have been no reports of such a high H2 yield at 500 °C. The stability test data in Fig. 4(B) demonstrates that the clean H2 yield at 200 °C corresponds to H2 generation of 10% while the values at 340 and 400 °C were 40% and 75%, respectively.
Fig. 3 Recalculated data for the reaction over the BS-Ru/C catalyst based on Fig. 1 and 2 ((A): H2 yield versus reaction temperature, (B): H2 yield versus reaction time). Temperature regions (I), (II), and (III) are explained in the text. The data were determined from the following equation: H2 yield = (H2 amount (mol)/CH4 feed amount (mol))/X × 100 (X = 2 (200–250 °C), 3 (>260 °C)). |
To clarify the reason for the enhanced activity and stability of the BS-Ru/C catalyst, we examined the surface physical properties of both catalysts. Fig. 5 presents the Ru 3p2/3 spectra obtained by X-ray photoelectron spectroscopy (XPS). A broad peak observed at approximately 463 eV for both catalysts could be deconvoluted to give peaks attributed to Ru(0) (461.6 eV), RuO2 (463.2 eV), and RuO3 (465.3 eV).29 However, the proportion of Ru(0), serving as the active sites11 in BS-Ru/C (31.5%), is lower than that in Ref.-Ru/C (43.3%). In Fig. 6, the temperature programmed reduction (TPR) profile of Ref.-Ru/C contained a large peak assigned to the reduction of dispersed RuOx at 162.2 °C with a small crystalline RuO2 reduction peak at approximately 260 °C.30 In contrast, the BS-Ru/C catalyst gave two large reduction peaks of well-dispersed RuOx and crystalline RuO2.29–31 Based on the XPS and TPR results, the biggest difference between both catalysts is their surface composition (atomic arrangements). Therefore, the special surface characteristic of the BS-Ru/C catalyst can be considered as the main factor that influences its improved stability and activity.
To fully understand the reaction pathway, the physical properties of BS-Ru/C catalysts before and after the stability test were assessed by TPR (Fig. 8). For the fresh catalyst, compared with the carbon support, the first two peaks were attributed to Ru particles, while the later broad peak in the range of 320–720 °C was attributed to the carbon support. After the stability test at 200 °C, a new peak at approximately 350 °C appeared. According to ref. 33 and 34, it was assigned to the Cβ carbon in the hydrocarbon. After the stability test at 340 °C, a peak appeared in the vicinity of 550 °C ascribed to Cγ, indicating graphitic carbon formation.33,34 However, the catalyst used at 400 °C showed two peaks at 350 and 550 °C, indicating that both hydrocarbon and graphitic carbon were generated. The hydrocarbon must have originated from the recombination of the CHx species dissociated from the methane on the catalyst surface.35,36 Although hydrocarbon was formed at both 200 and 400 °C, the hydrocarbon should have different hydrogen proportions due to different H2/CH4 ratios (Fig. 2), i.e., abundant and poor hydrogen-containing hydrocarbons were produced at 200 and 340 °C, respectively.
Fig. 8 TPR profiles obtained from the carbon support and the BS-Ru/C catalyst before and after stability tests at 200, 340, and 400 °C. |
The BS-Ru/C catalyst used at 340 °C was observed by Cs-corrected STEM. In Fig. 9, the two carbons (the deposited carbon and support carbon) can be clearly distinguished. The left image contains a highly dark layer with 100 lattice fringes around the dark Ru particles, which are assigned to the carbon deposited on the support carbon with a gray layer, such as carbon fibers and tubes.37 Fortunately, the deposited carbon growth behaviour was also observed (Fig. 9 (right)). A Ru particle was loaded on the edge of the carbon support by the polygonal barrel-sputtering method. After carbon growth, this Ru particle had left the carbon support, and was obviously seen on the end of the carbon fiber/tube, clearly indicating that these formed carbons did not cover the Ru surfaces, also showing that the catalyst surface kept clear from the beginning to the end of the reaction. The growth behaviour of this deposited carbon was very similar to the previous observation using Ru catalyst37 and Ni catalyst.38 It can be considered that a carbon atom was formed on the Ru catalyst surface through methane dissociation, and then moved to its edge through diffusion caused by concentration gradient, similar to the previous report using Fe, Ni and Ni-containing catalysts.39,40 When the carbon atom at the Ru edge achieved supersaturation, a carbon nucleus formed and resulted in the formation of carbon layers after growth. This observation clearly confirms that after the suppressing methane dissociation rate using a large amount of water, there would be enough time to achieve good carbon diffusion and growth, leading to a lack of cover on the Ru catalyst surfaces. Although the Ru particles in BS-Ru/C were very small (1.8 nm), carbon diffusion and growth occurred, unlike the previous report that carbon fiber/tube can be formed only when Ru critical size is not less than 30 nm.37 Compared with the fresh catalyst (Fig. S2†), no change in the Ru particles size was found, different from our previous observation.22 The dissociated carbon expanded laterally, resulting in Ru particles to be surrounded and encapsulated by the deposited carbon, further suppressing Ru particle growth even above 200 °C. The amount of carbon formed during the stability test period using the BS-Ru/C catalyst at 340 °C (1850 g C per g Ru) was significantly higher than that using the conventional catalyst (<801 g C per g metal);41 however, BS-Ru/C is far from deactivated, so long-term H2 stable production could be expected. The material balance (carbon balance and hydrogen balance) of the reaction at 340 °C was calculated as following: H: 1.02, C: 1.00 (C: 0.88 in the deposited carbon; C: 0.12 in CO2).
In principle, according to the mechanisms of the conventional MCD17 and RSM13 processes, both processes firstly begin from CH4 adsorption and its catalytic decomposition to hydrogen and carbonaceous species. The carbonaceous species further dissociates to carbon and H atoms during the MCD reaction, and while it further dissociates to carbon and H atoms, it also reacts with the surface oxygen derived from steam to form CO during the SRM reaction. Simultaneously, CO is converted to CO2 through WGSR. Of course, hydrocarbon formation via the recombination of carbonaceous species35,36,42,43 also cannot be avoided. In both conventional MCD and SRM processes, the rates of methane dissociation, carbon atom diffusion and growth should be equal at the steady state. Only at this condition, carbon does not accumulate on the active surface, which adsorbs and dissociates methane, and each process can proceed smoothly due to the clean active surface employed. If this balance breaks, carbon forms a covering layer on the active surface and the catalyst deactivates. In this research, we achieved the above balance in the low-temperature range of 200–500 °C using a large amount of water to suppress methane dissociation, which resulted in stable hydrogen production. In addition, CO was completely converted to CO2via WGSR, resulting in no CO production.
The H2 yield was not influenced by the Ru amount, so the CH4 step-wise dissociation was reversible, that is, CH4 step-wise dissociation and hydrogenation of the dissociated intermediate CHx occurred simultaneously and reached equilibrium, unlike the previous report.44 The activation energy of methane dissociation reported at low temperatures is 5.9–8.5 kcal mol−1,42,45,46 which is far lower than that of the H2 yield obtained in this research, indicating that methane dissociation was not a rate determining step (RDS), different from the previous literature,47 while carbon diffusion and growth should be RDS. The maximum YCO2 (CO2 yield)/CCH4 (CH4 conversion) ratio (≈0.22) remained almost unchanged at a certain temperature range. In addition, high CH4 concentration ≈ 10% caused the production of another co-product CO at high temperature; however this did not cause any change to the hydrogen amount and the maximum CO2/H2 ratio (≈0.22). Thus, CO2 formation through WGS should be considered as a RDS. Based on the proposed molecular mechanisms summarized in literature,48 the reaction mechanism was considered, and shown in Fig. 10. Cβ and Cγ in Fig. 10 could be reduced by H2.
Fig. 10 The proposed molecular reaction mechanisms for methane decomposition and steam reforming at low temperature. |
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c7se00572e |
This journal is © The Royal Society of Chemistry 2018 |