Efficient and Stable Ammonia Synthesis by Self-Organized Flat Ru

Oct 7, 2016 - Efficient and stable catalysts for ammonia synthesis under mild conditions are required to meet the strong demand for NH3 as an importan...
0 downloads 8 Views 6MB Size
Download PDF
Research Article pubs.acs.org/acscatalysis

Efficient and Stable Ammonia Synthesis by Self-Organized Flat Ru Nanoparticles on Calcium Amide Yasunori Inoue,†,¶ Masaaki Kitano,‡,¶ Kazuhisa Kishida,‡ Hitoshi Abe,§,∥,⊥ Yasuhiro Niwa,§ Masato Sasase,‡ Yusuke Fujita,† Hiroki Ishikawa,† Toshiharu Yokoyama,‡ Michikazu Hara,*,†,⊥,# and Hideo Hosono*,‡,⊥,# †

Institute of Innovative Research, Tokyo Institute of Technology, 4259 Nagatsuta, Midori-ku, Yokohama, Kanagawa 226−8503, Japan Materials Research Center for Element Strategy, Tokyo Institute of Technology, 4259 Nagatsuta, Midori-ku, Yokohama, Kanagawa 226−8503, Japan § High Energy Accelerator Research Organization, KEK, 1-1, Oho, Tsukuba, Ibaraki 305−0801, Japan ∥ Department of Materials Structure Science, School of High Energy Accelerator Science, SOKENDAI, The Graduate University for Advanced Studies, 1-1 Oho, Tsukuba, Ibaraki 305-0801, Japan ⊥ ACCEL, Japan Science and Technology Agency, 4-1-8 Honcho, Kawaguchi, Saitama 332-0012, Japan # Laboratory for Materials and Structures, Tokyo Institute of Technology, 4259 Nagatsuta, Midori-ku, Yokohama, Kanagawa 226−8503, Japan ‡

S Supporting Information *

ABSTRACT: Efficient and stable catalysts for ammonia synthesis under mild conditions are required to meet the strong demand for NH3 as an important precursor chemical and hydrogen carrier. Here we report that during ammonia synthesis, flat-shaped Ru nanoparticles with a narrow distribution (2.1 ± 1.0 nm) and self-organized on Ca(NH2)2 exhibit high catalytic performance far exceeding alkalipromoted Ru-based catalysts in yield and turnover frequency (TOF). This catalyst enables continuous NH3 production, even at 473 K under ambient pressure. During ammonia synthesis, Ru nanoparticles are distinctly anchored on the surface of Ca(NH2)2 by strong Ru−N interaction, which leads to the epitaxial growth of Ru on the support surface. The high catalytic performance is due to the formation of high-density flat-shaped Ru nanoparticles and high electron donor ability at the Ru/ Ca(NH2)2 interface. The catalytic stability is significantly improved by Ba-doping of Ca(NH2)2, and no degradation was observed after ca. 700 h of operation. KEYWORDS: ruthenium, self-organization, ammonia synthesis, Ca(NH2)2, XAFS

■

INTRODUCTION The reduction of dinitrogen (N2) to ammonia (NH3) under mild conditions is one of the most challenging topics in catalysis, while the conventional industrial process (Haber− Bosh process) requires high temperatures (673−873 K) and pressures (20−40 MPa).1 The formation of ammonia is thermodynamically favorable under low-temperature and high-pressure conditions. However, the reaction rate is moderate at low temperatures because a large amount of thermal energy is necessary to dissociate N2 on a conventional catalyst surface.2 Therefore, it is essential to develop a new catalyst that can cleave N2 even at low temperatures to improve the catalytic performance for ammonia synthesis. Although ammonia production under ambient conditions has been performed using organometallic complexes,3−5 the system is still far from appropriate for practical application due to the poor stability and low reaction rate of such catalysts. In addition, this reaction system generally requires an excess © XXXX American Chemical Society

amount of strong reducing agents and extra proton sources to afford NH3. It was recently reported that silica-supported tantalum hydride and a trinuclear titanium polyhydride complex induce N2 cleavage at 523 K and ca. 300 K, respectively, under atmospheric pressure without additional reducing agents or proton sources.6,7 However, continuous ammonia production has not been achieved with these systems. We have focused on an electride material, in which electrons function as anions, as the support material of a Ru catalyst for ammonia synthesis. Both the 12CaO·7Al2O3 and Ca2N:e− electrides were demonstrated to boost N2 dissociation on the Ru catalyst due to their strong electron-donating ability, which results in much higher catalytic performance in terms of the turnover frequency (TOF) than the conventional Ru Received: July 12, 2016 Revised: September 21, 2016

7577

DOI: 10.1021/acscatal.6b01940 ACS Catal. 2016, 6, 7577−7584

Research Article

ACS Catalysis

Figure 1. (a) Time course of NH3 synthesis over Ca(NH2)2 and 8 wt % Ru/Ca(NH2)2 at 613 K and 0.1 MPa. (b) Temperature dependence of NH3 synthesis rate over 10 wt % Ru/Ca(NH2)2 and 10 wt % Ru−Cs/MgO at 0.1 MPa. (c) Time course of NH3 synthesis from 15N2 and H2 over Ru/ Ca(NH2)2 at 473 K and 45 kPa. The inset shows the time course of the 15N2 concentration. (d) NH3 synthesis rate at 613 K over 10 wt % Ru/ Ca(NH2)2 and 10 wt % Ru−Cs/MgO as a function of reaction pressure.

catalyst.8−12 Nevertheless, there remains several obstacles to be resolved for practical application of a Ru-loaded electride, such as the small surface area of the electride (1−3 m2 g−1), which increases the size (20−30 nm) of deposited Ru particles. B5 step sites, which consist of five Ru atoms formed on the steps of Ru particles, are considered to be the primary active point for ammonia synthesis. The density of B5 step sites increases with a decrease in the Ru particle size because of the increase in particle edges.13,14 The optimum Ru particle size for a high density of B 5 step sites has been theoretically and experimentally demonstrated to be in the range 1.8−3.5 nm.15−17 A current challenge is how to realize a highly active catalyst for ammonia synthesis under mild conditions on the basis of this principle. Herein, we demonstrate that the Ru-loaded Ca(NH2)2 catalyzes the reduction of N2 to NH3 at temperatures as low as 473 K under ambient pressure. Flat-shaped Ru nanoparticles with a diameter of ca. 2.1 nm are self-organized on Ca(NH2)2 during ammonia synthesis. The driving force is the strong interaction between Ru and N, and the epitaxial growth of Ru nanoparticles on Ca(NH2)2, where the resultant catalyst exhibits a much higher activity for ammonia synthesis than conventional Ru catalysts by an order of magnitude.

343 K.20 When Ca(NH2)2 is heated at 613 K under N2+H2 flow, the ammonia formation rate rapidly decreases with the reaction time (Figure 1a) and reaches 0 after reaction for 20 h. The total amount of ammonia generated from Ca(NH2)2 (0.1 g, 1.38 mmol) used for 20 h reached 1.35 mmol. Thus, this ammonia production is due to the decomposition of Ca(NH2)2 (Ca(NH2)2 → CaNH + NH3). On the other hand, Ru-loaded Ca(NH2)2, referred to as Ru/Ca(NH2)2, functions as an efficient and stable catalyst for ammonia synthesis without noticeable deactivation. The total amount of ammonia produced for 45 h was estimated to be ca. 37.8 mmol, which far exceeds that estimated by the decomposition of Ca(NH2)2 and confirms the efficiency of Ru/Ca(NH2)2 as a catalyst for ammonia synthesis. Figure S1b shows the relationship between the catalytic performance of Ru/Ca(NH2)2 and the Ru loading. The ammonia synthesis rate over Ru/Ca(NH2)2 increases significantly with the Ru loading and reaches a maximum (9.6 mmol g−1 h−1) at 10 wt % Ru. Figure 1b demonstrates that the dependence of the catalytic activities of Ru/Ca(NH2)2 on the reaction temperature. The results for Cs−Ru/MgO, the most active conventional Ru catalysts for ammonia synthesis, are also shown for comparison.21,22 The catalytic activity of Ru/Ca(NH2)2 is superior to that of Cs−Ru/MgO over the entire temperature range examined under ambient pressure. Ammonia production over Cs−Ru/MgO is not detected below 523 K, while it was confirmed over Ru/Ca(NH2)2 at 473 K. Figure 1c shows the reaction time course for ammonia synthesis from heavy

■

RESULTS AND DISCUSSION Calcium amide (Ca(NH2)2) is an ionic compound with the anatase structure that consists of Ca2+ and NH2− ions18,19 (Figure S1a), and readily decomposes to produce NH3 above 7578

DOI: 10.1021/acscatal.6b01940 ACS Catal. 2016, 6, 7577−7584

Research Article

ACS Catalysis Table 1. Catalytic Performance of Various Ru Catalystsa catalyst

Ru content (wt %)

SBET (m2 g−1)

Ru/Ca(NH2)2

10

50−60

Ru/C12A7:e− Cs−Ru/MgO (Ru/Cs = 1)

1.2 10

1−2 20−30

D (%) b

61 2.7c 3.2c 25b 24c

db (nm) 2.1 ± 1.1 41 5.2 ± 2.2

TOF (s−1) b

0.0074 0.17c 0.025c 0.0015b 0.0015c

NH3 syn. rate (mmol g−1 h−1) 15.8 0.34 1.28

a Reaction conditions: catalyst weight 0.1 g, flow rate 60 mL min−1 (H2/N2 = 3), pressure 0.8 MPa, temperature 573 K. bThe mean sizes of Ru particles (d), the Ru particle dispersion (D), and TOF were estimated by averaging the particle size distribution measured using STEM (Figure 4). c These values were evaluated using active site numbers determined by the CO pulse chemisorption method (Ru/CO = 1.0).

dinitrogen (15N2) and H2 over Ru/Ca(NH2)2 at 473 K. The m/ z = 18 and 17 signals increased with the reaction time, accompanying a decrease in 15N2 (m/z = 30). In contrast, the intensity of the m/z = 16 (NH2, a fragment of NH3) signal was not noticeably changed, even after 20 h of reaction, which implies that NH3 is not formed from the decomposition of Ca(NH2)2. Accordingly, the m/z = 18 and 17 signals are assignable to 15NH3 and 15NH2 (a fragment of 15NH3), respectively, which provides clear evidence of ammonia formation from gas phase nitrogen and hydrogen at 473 K. Figure S2 shows ammonia production over Ru/Ca(NH2)2 at 473 K for a prolonged time. Ammonia production increased linearly with the reaction time and reached 4.94 mmol after 480 h of reaction, which is more than ca. 3.6 times as much as that produced due to the decomposition of Ca(NH2)2. These results indicate that both N2 dissociation and N−H bond formation proceed at 473 K over the Ru/Ca(NH2)2 catalyst. The difference in the ammonia synthesis rates over Ru/ Ca(NH2)2 and Cs−Ru/MgO increases significantly with the reaction pressure (Figure 1d); distinct hydrogen poisoning is observed for the latter but not for the former.23 The reported Cs−Ru/MgO catalyst exhibits high catalytic activity, and the effluent NH3 mole fraction (ENH3) is estimated to be ca. 0.8 vol % at 613 K under atmospheric pressure (weight-hourly space velocity (WHSV) = 17400 mL gcat−1 h−1).22 ENH3 with Ru/ Ca(NH2)2 (0.61 vol % at 613 K and 0.1 MPa) is comparable to that of Cs−Ru/MgO,22 even under a high WHSV of 36000 mL gcat−1 h−1. Furthermore, the catalytic activity of Ru/Ca(NH2)2 increased linearly with the reaction pressure, reaching 31.7 mmol g−1 h−1 (ENH3 = 2.17 vol %) at 0.9 MPa. In contrast, the catalytic activity of Cs−Ru/MgO was almost independent of the reaction pressure.22 Therefore, it can be concluded that Ru/ Ca(NH2)2 is far superior to the Cs−Ru/MgO catalyst. Table 1 summarizes the catalytic performance of the tested catalysts at 573 K and 0.8 MPa. TOFs were obtained using the numbers of surface Ru atoms estimated from both CO chemisorption24 and average Ru particle sizes determined by transmission electron microscopy (TEM) observations. Both the ammonia synthesis rate and TOF with Ru/Ca(NH2)2 were much higher than those with Cs−Ru/MgO. It should be noted that the Ru dispersion in Ru/Ca(NH2)2 estimated by CO chemisorption is much smaller than that based on TEM observation. This large discrepancy suggests that not all surface Ru atoms of the Ru particles on Ca(NH2)2 are available for ammonia synthesis. Actually, STEM-EDX analysis revealed that Ca−N (no information on hydrogen) species derived from Ca(NH2)2 support is formed on the Ru surfaces after the ammonia synthesis reaction (Figure 2). This is due to the strong metal−support interaction (the detail is discussed in the following section) that is commonly observed for noble metal catalysts supported on TiO2 after high-temperature reduction in H2.25 Therefore, low Ru

Figure 2. STEM-EDX analysis for 10 wt % Ru/Ca(NH2)2 after NH3 synthesis reaction. The red line indicates the scanning path for the EDX line analyses.

dispersion for Ru/Ca(NH2)2 determined by CO chemisorption originates from the coverage of the Ru surface with the Ca−N species, resulting in a large difference in Ru dispersion by TEM observation from that by CO chemisorption. As a result, Ru/ Ca(NH2)2 is superior to Ru/C12A7:e−, which has exhibited the highest TOF among the Ru-based catalysts reported to date, not only in ammonia formation rate, but also TOF (based on CO chemisorption).8 While it is difficult to estimate the accurate quantity of active sites on Ru/Ca(NH2)2, we can conclude that Ru/Ca(NH2)2 has the highest ammonia synthesis rate and TOF among all the Ru-catalysts at low reaction temperature. The high TOF value of Ru/Ca(NH2)2 may be attributed to the electron donation from Ca(NH2)2 support to Ru. The electron donating property of Ca(NH2)2 was examined through N2 isotopic exchange reaction ( 15 N 2 + 14 N 2 = 2 15 N 14 N). Generally, the N 2 dissociation reaction on the Ru surface is strongly enhanced by electron donation from the support material, lowering the energy barrier for this reaction.10,26,27 As shown in Table 2, Ru/ Ca(NH2)2 exhibits the highest reaction rate among the tested catalysts, and the activation energy of Ru/Ca(NH2)2 is almost half that of Cs−Ru/MgO, which indicates that Ca(NH2)2 is much superior to Cs-promoter in electron donating ability. On the other hand, Ru/Ca(NH2)2 has a relatively larger activation energy than Ru/C12A7:e−. This result may be understood by considering that the electron-donating power of Ca(NH2)2 is weaker than that of C12A7:e− with a very low work function 7579

DOI: 10.1021/acscatal.6b01940 ACS Catal. 2016, 6, 7577−7584

Research Article

ACS Catalysis

discussed later. A similar metal−support interaction was reported for highly dispersed metal catalyst supported on metal oxide, in which the metal−oxygen bonds were observed in EXAFS spectra when small metal nanoparticles were strongly bonded with the oxide support.30,31 The structural parameters extracted for the Ru/Ca(NH2)2 samples through EXAFS curve-fitting analyses are listed in Table S1. The 5 and 10 wt % Ru samples were fitted well by the Ru metal model. A Ru−N shell was not incorporated for these two samples. Both Ru−Ru and Ru−N correlations were considered to fit the 1 and 2 wt % Ru EXAFS data. The procedure for curve-fitting analysis of the 2 wt % Ru/Ca(NH2)2 is shown in Figure S4. The absorption spectrum (Figure S4a) was processed to extract the k2-weighted EXAFS oscillation, k2χ(k) (Figure S4b). k2χ(k) over the range of 2.3−14.0 Å−1 was transformed to obtain the FT shown in Figure S4c. The peak at 1.3 Å should be carefully discussed to be assigned to Ru−N or Ru−O bond. A possibility was considered that the peak is related to amorphous Ru oxide species, but this idea was discounted on the basis of the following results: We tried fitting by the amorphous Ru-oxide model, and an outcome was a mixture of 80% metal and 20% oxide. This ratio of oxide, however, was inconsistent with the XANES spectra in which the edge position shows the sample is in metallic state as discussed above (Figure 3b). Thus, there is no possibility of an existence of significant Ru oxide species, and it is not possible to assign the peak to the Ru−O bond. On the other hand, the Ru−N species model fits the peak well (R factor = 0.04057). More carefully, the FT was fitted by 3 shells, Ru−Ru, Ru−N, and Ru−O (Figure S4c). If there is a Ru−O bond rather than a Ru−N bond, we should have an outcome with more Ru−O CN than Ru−N CN. However, the result contained no Ru−O CN, and the Ru−N CN was retained as above. The CNs of Ru−Ru and Ru−N shells were determined to be 2.92 and 1.43, respectively (Table S1), while CN of the Ru−O path was not yielded. The Ru−O and Ru−N species were treated equally in addition to the Ru−Ru path in the fitting procedure as well, but the obtained result consisted of metallic Ru−Ru and Ru−N bonds only. This result indicates that a possibility of Ru−O bond formation in Ru/Ca(NH2)2 can be ruled out. We conclude that the major contribution to this peak is not due to Ru-oxide species but the Ru−N bond. The XAFS results for the 1 and 2 wt % Ru/Ca(NH2)2 catalysts are in good agreement

Table 2. Nitrogen Isotopic Exchange Reaction over Various Ru Catalysts catalyst

Ru (wt %)

activity (mmol g−1 h−1)a

Ea (kJ mol−1)b

Ru/Ca(NH2)2 Ru/C12A7:e− Cs−Ru/MgO

10 1.2 6

3.6 0.79 0.52

75 58 139

a

Reaction temperature, 613 K; 15N2/14N2 = 1/4; total pressure, 20 kPa. bTemperature range, 533−613 K.

(2.4 eV). The details of electronic promoting effect will be discussed in the later section. The local structure of Ru in the particles was investigated using X-ray absorption fine structure (XAFS) measurements. Figure 3a shows Fourier transforms (FTs) of the k2-weighted extended X-ray absorption fine structure (EXAFS) spectra (note these spectra are not corrected for phase shift) for Ru/ Ca(NH2)2 with various amounts of Ru-loading after ammonia synthesis. A major peak is located at 2.3 Å, which corresponds to the Ru−Ru bond. The intensity of this peak increases with the Ru-loading amount. As shown in Table S1, the coordination number (CN) around Ru loaded on Ca(NH2)2 is much smaller than that of bulk Ru (CN = 12), which indicates that the nanometer-sized Ru particles are well-dispersed on Ca(NH2)2. The mean diameter of Ru particles for 5 wt % Ru/Ca(NH2)2 was estimated to be ca. 3.6 nm from the coordination number (7.06) by application of a previously reported procedure.28 Another peak was also observed at 1.3 Å for the 1 and 2 wt % Ru samples. Although this peak position is close to that for the Ru−O bond in RuO2 (Figure S3), the FTs and X-ray absorption near-edge structure (XANES) spectra for 1 and 2 wt % Ru/Ca(NH2)2 are distinctly different from those for RuO2 (Figure 3b). If the oxidized species are present on the Ru surface, the edge shift of XANES to higher energy will be observed.29 The XANES spectra reveal that Ru on Ca(NH2)2 remains almost in the neutral (metal) state. The first peak at 1.3 Å is attributed to a correlation between Ru and nitrogen atom in the Ca(NH2)2 support. If we assume that this peak is due to Ru−N, then the peak intensity should increase for smaller Ru particles. The areal ratio of the first to second peak increases with a decrease in the Ru particle size or the loading amount of Ru on Ca(NH2)2. Therefore, the first peak at 1.3 Å can be assigned to a Ru−N bond. The details of the analysis are

Figure 3. (a) FTs of the k2-weighted EXAFS oscillations for various amounts of Ru-loaded Ca(NH2)2 after NH3 synthesis. (b) Ru K-edge XANES spectra for 1 and 2 wt % Ru/Ca(NH2)2 after the NH3 synthesis reaction. Spectra for Ru and RuO2 are also included. 7580

DOI: 10.1021/acscatal.6b01940 ACS Catal. 2016, 6, 7577−7584

Research Article

ACS Catalysis

Figure 4. (a) TEM and (b) HAADF-STEM images of Cs-promoted 10 wt % Ru-loaded MgO catalyst after the NH3 synthesis reaction. (c) TEM image of 10 wt % Ru/Ca(NH2)2 catalyst before reaction. (d) HAADF-STEM image of 10 wt % Ru/Ca(NH2)2 catalyst after NH3 synthesis. (e) Enlarged image of Ru particle in panel (d). The insets of (a) and (d) show the particle size distributions of Cs−Ru/MgO and Ru/Ca(NH2)2, respectively.

reaction, that is, Ca(NH2)2 immediately after Ru deposition by pyrolysis of Ru3(CO)12. Individual Ru particles cannot be distinguished in the image, whereas a high density of small and flat Ru patches appears on the surface of Ca(NH2)2 after the reaction (613 K, 50 h); no conventional round shape grains of Ru were observed in Ru/ Ca(NH2)2, even after the reaction. The average Ru particle size was estimated to be 2.1 nm, and the particle size distribution is markedly narrower than that of Cs−Ru(10 wt %)/MgO. The lower Ru-loaded samples had extremely small Ru nanoparticles that are distributed on Ca(NH2)2 support (Figure S8). The high Ru dispersion on Ru/Ca(NH2)2 is attributed to the strong interaction between Ru atoms and the N atoms of Ca(NH2)2 on the basis of the XAFS results. Such a strong interaction prevents the aggregation of Ru particles during ammonia synthesis. It is important to note that the lattice fringes of Ru (2.23 Å) on Ca(NH2)2 are parallel to those of the underlying support material (2.94 Å). This result implies that there is an epitaxial relationship between the Ru nanoparticles and the support surface. A possible explanation is a higher order or domain matching epitaxial relationship;32,33 that is, the three-plane spacing (2.94 × 3 = 8.82 Å) of the support match that of the four planes (2.23 × 4 = 8.92 Å) of Ru. Similar observations were reported for oxide-supported metal catalysts with strong metal−support interactions.34,35 Such an epitaxy is considered to be the origin of the self-organized flat Ru nanoparticles with narrow size distribution formed during the reaction. Lattice matching at the subnanometer scale is considered favorable for the formation of such nanoparticles. The presence of critical strain from the accumulation of small mismatches results in the

with the model considering the Ru−Ru and Ru−N bonds, as shown in Table S1. Such low CNs may be interpreted to indicate that the Ru catalyst exists as a very small cluster and that the Ru atoms are strongly bonded to the N atoms of the support. Powder X-ray diffraction (XRD) patterns for Ca(NH2)2 and Ru/Ca(NH2)2, before and after reaction (613 K, 50 h), are shown in Figure S5. Although Ca(NH2)2 is completely converted into CaNH after the heat treatment, the Ca(NH2)2 phase remains in the Ru/Ca(NH2)2 catalyst after the reaction. Therefore, Ru/Ca(NH2)2 after the reaction consists of a mixture of Ca(NH2)2 and CaNH phases, as evidenced by the CaNH phase in the XRD pattern. However, Ru-loaded CaNH prepared from Ca(NH2)2 by heating at 613 K exhibits fairly lower catalytic activity than Ru/Ca(NH2)2 (Figure S6), which suggests that the high catalytic performance of Ru/Ca(NH2)2 is not due to the presence of CaNH but to other Ca−N−H species formed at the interface between Ru and Ca(NH2)2. From these results, the strong Ru−N interaction at the interface between Ru and such Ca−N−H species is expected to form fine Ru nanoparticles on the support surface. High-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) analysis was conducted to investigate the microstructure of Ru nanoparticles on Ca(NH2)2 after the reaction. The size of the Ru particles on the Cs−Ru(10 wt %)/MgO catalyst is widely distributed in the range of 2 to 10 nm, which is larger than the Ru particles on the Cs−Ru(2 wt %)/MgO, because higher metal loading typically leads to larger metal particles (Figures 4 and S7). Figure 4c shows a HAADF-STEM image of Ru/Ca(NH2)2 before the 7581

DOI: 10.1021/acscatal.6b01940 ACS Catal. 2016, 6, 7577−7584

Research Article

ACS Catalysis

of hydride species (H−), which are easily exchangeable with an electron. Reversible exchange between hydride ions and electrons occur at the Ru−support interfaces, imparting high tolerance to hydrogen poisoning (Table S2).12 This idea implies that such a hydride species to enhance electron donation to Ru is also formed on Ru/Ca(NH2)2 during reaction. A plausible route for generation of such H− ion during the reaction is via an F center with a low work function at the interface, e− (F center) + H0 → H−, resulting from the formation of NH2− vacancy and subsequent electron trapping at the vacancy site. Further effort to examine this idea is ongoing. Ba-doping (3 wt %) of Ca(NH2)2 was effective to improve the activity and long stability of Ru/Ca(NH2)2. Figure 5a shows

formation of a very narrow distribution of flat shaped particles with fixed separation rather than the large flat plate formation obtained by conventional epitaxial growth. Furthermore, the observed lattice spacing of Ru particles (2.23 Å) on Ca(NH2)2 is close to that (d111 = 2.21 Å) of the face-centered-cubic (fcc) (111) plane of Ru, which differs from the conventional structure of Ru particles. Figure 4b shows well-defined lattice fringes with a d-spacing of 2.14 Å in the large spherical Ru particles of Cs−Ru/MgO, which are attributed to hexagonal close-packed (hcp) Ru (d002 = 2.142 Å). Although fcc Ru is a metastable phase compared with hcp Ru,36 fcc Ru nanoparticles would be stabilized by strong Ru−Ca(NH2)2 adhesion. The adhesion energy between metals and supports also has a considerable effect on the shape of a supported metal nanoparticles, i.e., two-dimensional (2D) growth (flat-shaped metal particles) prevails over 3D growth (spherical metal particles) if the support material has a very strong adhesion energy for metal nanoparticles.37,38 As shown in Figures 4d and S8, flat-shaped Ru nanoparticles are formed on Ca(NH2)2. Vlachos et al. has recently demonstrated from both density functional theory (DFT) calculations and experiments that flat Ru particles on Al2O3 have a higher number of active sites (B5type step sites) and exhibit higher catalytic activity for ammonia decomposition than a hemispherical Ru particle catalyst.39 This suggests that there is a high density of B5-type step sites on the flat-shaped Ru nanoparticles on Ca(NH2)2. On the basis of these results, we have concluded that the flat-shaped Ru nanoparticles are self-organized by the strong-metal support interaction between Ru and the Ca(NH2)2 support, and by the higher order epitaxial relationship, which accounts for the high catalytic performance and long-term stability during ammonia synthesis. The effect of electron promotion is also considered to be an important factor in the catalytic activity of Ru/Ca(NH2)2. The catalytic activity of Ru/Ca(NH2)2 is ca. 10 times higher than that of Cs−Ru/MgO (Table 1), although Ru/Ca(NH2)2 has only 2.4 times more surface Ru sites than Cs−Ru/MgO. Therefore, the effect of electron promotion on the catalytic activity cannot be ruled out. Kinetic analyses were conducted to elucidate the electron promotion effect of Ca(NH2)2 in ammonia synthesis, and the results are summarized in Table S2. The kinetic data for Ru/Ca(NH2)2 are similar to those of Ru/C12A7:e− and Ru/Ca2N:e−, distinct from those of Cs−Ru/ MgO, a representative conventional Ru catalyst. The apparent activation energy (59 kJ mol−1) and reaction order for N2 (0.53) on Ru/Ca(NH2)2 are almost half those of Cs−Ru/MgO catalyst (113 kJ mol−1, 1.0), and the values are very similar to those of Ru/C12A7:e− and Ru/Ca2N:e−. In such Ca-based electride catalysts, strong electron donation from the electrides to Ru nanoparticles facilitates N2 cleavage, and N2 dissociation is no longer the rate-determining step of ammonia synthesis over these Ru-electride catalysts, which results in 0.5 of the reaction order for N2 and 50−60 kJ mol−1 of the apparent activation energy.8,11,12 The results of nitrogen isotopic exchange reaction also support that Ru/Ca(NH2)2 catalyst facilitates the N2 dissociation reaction. (Table 2) These results reveal that Ca(NH2)2 also has high electron-donating ability to Ru, which is equivalent to or less than those of C12A7:e− and Ca2N:e−. However, no species with strong electron-donating power are present for stoichiometric Ca(NH2)2. Therefore, it is natural to assume that such a species is created at the Ru/ Ca(NH2)2 interface during the reaction. A plausible explanation for strong electron donation on Ru/Ca(NH2)2 is the formation

Figure 5. (a) Reaction rate for NH3 synthesis over various 10 wt % Ru-loaded amides. (b) Time course of NH3 synthesis on Ru/Ba− Ca(NH2)2, Ru/Ca(NH2)2, and Cs−Ru/MgO. Reaction conditions: catalyst weight, 0.1 g; synthesis gas, H2/N2 = 3, flow rate, 60 mL min−1; pressure, 0.1 MPa; temperature, 613 K.

that Ba-doping distinctly improves the catalytic activity of Ru/ Ca(NH2)2, despite the low activity of Ru/Ba(NH2)2. Ru/Ba− Ca(NH2)2 also exhibited excellent stability during reaction for 700 h (almost 1 month). (Figure 5b) In contrast, the catalytic activities of Ru-loaded pristine Ca(NH2)2 and Cs−Ru/MgO decreased after prolonged reaction (>100 h). The initial activity of Ru/Ba−Ca(NH2)2 was lower than that of Ru/Ca(NH2)2, but increased with the reaction time, which suggests that the strong metal−support interaction in Ru−Ba−Ca(NH2)2 is achieved during the initial reaction period. Further investigation is ongoing to elucidate reasons for the excellent stability of the catalyst. 7582

DOI: 10.1021/acscatal.6b01940 ACS Catal. 2016, 6, 7577−7584

Research Article

ACS Catalysis

■

CONCLUSIONS Ru/Ca(NH2)2 functions as an efficient catalyst for ammonia synthesis with excellent stability. The ammonia synthesis rate and TOF of Ru/Ca(NH2)2 are superior to those of alkalipromoted Ru catalysts and Ru-loaded C12A7 electrides in the whole range of reaction temperatures (473−613 K). Flatshaped Ru nanoparticles with a narrow distribution centered at 2 nm are grown on the Ca(NH2)2 surface. It is considered that such unique shape and the distribution of Ru nanoparticles are explained in terms of domain matching epitaxy driven by a strong chemical interaction between Ru and N atoms. The high catalytic performance is attributed to the formation of a high density of flat-shaped Ru nanoparticles and the electron promotion effect of Ca(NH 2 ) 2 . The present findings demonstrate that the size and shape of Ru nanoparticles can be controlled by strong metal−support chemical interactions and the epitaxial relationship between the metal and nitridebased support material.

■

613 K and the pressure was in the range of 0.1−1.0 MPa. The NH3 produced was trapped in 5 mM H2SO4 aqueous solution, and the amount of NH4+ generated in the solution was estimated using an ion chromatograph (LC-2000 plus, Jasco) equipped with a conductivity detector. Each catalyst was repeatedly tested at least three times to evaluate the reproducibility. Ammonia synthesis from 15N2 and H2 was conducted at 473 K over 15 wt % Ru/Ca(NH2)2 (0.25 g) pretreated at 613 K for 24 h using a U-shaped glass reactor connected to a closed gas circulation system. The mixture of 15N2 and H2 gases (total pressure: 45 kPa, 15N2:H2 = 1:3) was introduced into the glass system. The change in the composition of the circulating gas was monitored with a quadrupole mass spectrometer (M101QA-TDM, Canon Anelva Corp.), and He was used as a carrier gas. The m/z = 2, 16, 17, 18, 28, 29, and 30 mass signals were monitored as a function of time to follow the reaction. Nitrogen isotopic exchange reaction was also conducted at 613 K over 10 wt % Ru/Ca(NH2)2 (0.05 g) pretreated at 613 K for 24 h using same closed gas circulation system. The mixture of 15N2 and 14N2 gases (total pressure: 20 kPa, 15N2/14N2 = 1/ 4) was adsorbed on the catalyst until an adsorption equilibrium. The masses 28, 29, and 30 m/z were monitored as a function of time to follow the exchange reaction.

EXPERIMENTAL DETAILS

Catalyst Preparation. Ca(NH2)2 was synthesized by the following method. Ca (99.99%, Aldrich Co.) metal shot was placed in a stainless steel autoclave, and NH3 gas was then introduced into the autoclave at 227 K to form liquid NH3. After stirring the solution for 1 h, the autoclave was heated at 373 K for 1 h. The autoclave was then opened in an Ar-filled glovebox to obtain Ca(NH2)2 powder. Ba(NH2)2 and Badoped Ca(NH2)2 were prepared using the same method, with Ba (99.99%, Aldrich Co.) and a mixture of Ca and Ba metal shot as source materials. 1−20 wt % Ru-loaded samples were prepared by chemical vapor deposition using Ru3(CO)12 (99%, Aldrich Co.) in accordance with a previous report.8 Characterization. N2 adsorption−desorption isotherms were measured at 77 K using a specific surface area analyzer (BELSORP-mini II, MicrotracBEL) to determine the Brunauer−Emmett−Teller (BET) specific surface area of samples. The Ru dispersion and mean particle size were determined by CO pulse chemisorption at 323 K with a He flow of 30 mL min−1 and 0.09 mL pulses of 9.88% CO in He using a catalyst analyzer (BELCAT-A, MicrotracBEL); a stoichiometry of Ru/ CO = 1 was assumed. The samples were reduced at 613 K under H2 flow (50 mL min−1) for 15 min before CO-pulse chemisorption. Scanning transmission electron microscopy (STEM) and high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) images were captured using a microscope (JEM-ARM 200F, Jeol) equipped with an energy dispersive X-ray spectroscopy (EDX) detector operating at 200 kV. X-ray absorption fine structure (XAFS) experiments were performed using the NW-10A beamline at the Photon Factory. A Si(311) double-crystal monochromator was used to obtain the monochromatized X-ray beam, and spectra were obtained in transmission mode. Ru/Ca(NH2)2 and BN (dried at 573 K) were mixed in an Ar-filled glovebox, and the obtained mixture was pressed with a hand-press apparatus to obtain a pellet sample. The pellet sample was then sealed in a plastic bag. XAFS spectra were analyzed using the Athena and Artemis software packages.40 The FEFF6 code was used to calculate the theoretical spectra.41 Ammonia Synthesis Reaction Test. NH3 synthesis reactions were performed in a silica glass or stainless-steel reactor with a synthesis gas (H2:N2 = 3:1) at a flow rate of 60 mL min−1. The reaction temperature was varied from 373 to

■

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acscatal.6b01940. Experimental methods, characterization, kinetic analysis, and supplemental data, as noted in the text (PDF)

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. Author Contributions ¶

Y.I and M. K. contributed equally.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by a fund from Accelerated Innovation Research Initiative Turning Top Science and Ideas into High-Impact Values (ACCEL) of the Japan Science and Technology Agency (JST). A portion of this work was supported by a Kakenhi Grant-in-Aid (No. 15H04183) from the Japan Society for the Promotion of Science (JSPS). The authors appreciate the technical assistance of E. Sano, Y. Takasaki, M. Okunaka, and S. Fujimoto. The XAFS experiments were performed under the approval of the Photon Factory, High Energy Accelerator Research Organization (KEK) (Proposal No. 2013S2-002). The authors express thanks to Drs. K. Nakajima, K. Kamata and S. Matsuishi for helpful discussions and suggestions.

■

REFERENCES

(1) Liu, H. Ammonia Synthesis Catalysts: Innovation and Practice, 1st ed.; World Scientific Publishing Co. Pte. Ltd.: Singapore, 2013. (2) Prieto, G.; Schuth, F. Angew. Chem., Int. Ed. 2015, 54, 3222− 3239. 7583

DOI: 10.1021/acscatal.6b01940 ACS Catal. 2016, 6, 7577−7584

Research Article

ACS Catalysis (3) Pool, J. A.; Lobkovsky, E.; Chirik, P. J. Nature 2004, 427, 527− 530. (4) Arashiba, K.; Miyake, Y.; Nishibayashi, Y. Nat. Chem. 2011, 3, 120−125. (5) Anderson, J. S.; Rittle, J.; Peters, J. C. Nature 2013, 501, 84−87. (6) Avenier, P.; Taoufik, M.; Lesage, A.; Solans-Monfort, X.; Baudouin, A.; de Mallmann, A.; Veyre, L.; Basset, J.-M.; Eisenstein, O.; Emsley, L.; Quadrelli, E. A. Science 2007, 317, 1056−1060. (7) Shima, T.; Hu, S.; Luo, G.; Kang, X.; Luo, Y.; Hou, Z. Science 2013, 340, 1549−1552. (8) Kitano, M.; Inoue, Y.; Yamazaki, Y.; Hayashi, F.; Kanbara, S.; Matsuishi, S.; Yokoyama, T.; Kim, S.-W.; Hara, M.; Hosono, H. Nat. Chem. 2012, 4, 934−940. (9) Inoue, Y.; Kitano, M.; Kim, S.-W.; Yokoyama, T.; Hara, M.; Hosono, H. ACS Catal. 2014, 4, 674−680. (10) Kitano, M.; Kanbara, S.; Inoue, Y.; Kuganathan, N.; Sushko, P. V.; Yokoyama, T.; Hara, M.; Hosono, H. Nat. Commun. 2015, 6, ArticleNo. 6731. (11) Kanbara, S.; Kitano, M.; Inoue, Y.; Yokoyama, T.; Hara, M.; Hosono, H. J. Am. Chem. Soc. 2015, 137, 14517−14524. (12) Kitano, M.; Inoue, Y.; Ishikawa, H.; Yamagata, K.; Nakao, T.; Tada, T.; Matsuishi, S.; Yokoyama, T.; Hara, M.; Hosono, H. Chem. Sci. 2016, 7, 4036−4043. (13) Dahl, S.; Törnqvist, E.; Chorkendorff, I. J. Catal. 2000, 192, 381−390. (14) Dahl, S.; Sehested, J.; Jacobsen, C. J. H.; Törnqvist, E.; Chorkendorff, I. J. Catal. 2000, 192, 391−399. (15) Jacobsen, C. J. H.; Dahl, S.; Hansen, P. L.; Törnqvist, E.; Jensen, L.; Topsøe, H.; Prip, D. V.; Møenshaug, P. B.; Chorkendorff, I. J. Mol. Catal. A: Chem. 2000, 163, 19−26. (16) Honkala, K.; Hellman, A.; Remediakis, I. N.; Logadottir, A.; Carlsson, A.; Dahl, S.; Christensen, C. H.; Nørskov, J. K. Science 2005, 307, 555−558. (17) Raróg-Pilecka, W.; Miśkiewicz, E.; Szmigiel, D.; Kowalczyk, Z. J. Catal. 2005, 231, 11−19. (18) Juza, R. Angew. Chem., Int. Ed. Engl. 1964, 3, 471−481. (19) Bouclier, P.; Portier, J.; Turrell, G. J. Mol. Struct. 1969, 4, 1−13. (20) Hino, S.; Ichikawa, T.; Leng, H.; Fujii, H. J. Alloys Compd. 2005, 398, 62−66. (21) Aika, K.-i.; Takano, T.; Murata, S. J. Catal. 1992, 136, 126−140. (22) Rosowski, F.; Hornung, A.; Hinrichsen, O.; Herein, D.; Muhler, M.; Ertl, G. Appl. Catal., A 1997, 151, 443−460. (23) Siporin, S. E.; Davis, R. J. J. Catal. 2004, 225, 359−368. (24) Larichev, Y. V.; Moroz, B. L.; Zaikovskii, V. I.; Yunusov, S. M.; Kalyuzhnaya, E. S.; Shur, V. B.; Bukhtiyarov, V. I. J. Phys. Chem. C 2007, 111, 9427−9436. (25) Tauster, S. J.; Fung, S. C.; Garten, R. L. J. Am. Chem. Soc. 1978, 100, 170−175. (26) Hikita, T.; Kadowaki, Y.; Aika, K. J. Phys. Chem. 1991, 95, 9396−9402. (27) Hinrichsen, O.; Rosowski, F.; Hornung, A.; Muhler, M.; Ertl, G. J. Catal. 1997, 165, 33−44. (28) Pirkkalainen, K.; Serimaa, R. J. Appl. Crystallogr. 2009, 42, 442− 447. (29) Bera, P.; Priolkar, K. R.; Gayen, A.; Sarode, P. R.; Hegde, M. S.; Emura, S.; Kumashiro, R.; Jayaram, V.; Subbanna, G. N. Chem. Mater. 2003, 15, 2049−2060. (30) Nagai, Y.; Hirabayashi, T.; Dohmae, K.; Takagi, N.; Minami, T.; Shinjoh, H.; Matsumoto, S. i. J. Catal. 2006, 242, 103−109. (31) Qiao, B.; Wang, A.; Yang, X.; Allard, L. F.; Jiang, Z.; Cui, Y.; Liu, J.; Li, J.; Zhang, T. Nat. Chem. 2011, 3, 634−641. (32) Narayan, J.; Larson, B. C. J. Appl. Phys. 2003, 93, 278−285. (33) Kwak, B. S.; Erbil, A.; Budai, J. D.; Chisholm, M. F.; Boatner, L. A.; Wilkens, B. J. Phys. Rev. B: Condens. Matter Mater. Phys. 1994, 49, 14865−14879. (34) Bernal, S.; Botana, F. J.; Calvino, J. J.; Cauqui, M. A.; Cifredo, G. A.; Jobacho, A.; Pintado, J. M.; Rodriguez-Izquierdo, J. M. J. Phys. Chem. 1993, 97, 4118−4123.

(35) Gatica, J. M.; Baker, R. T.; Fornasiero, P.; Bernal, S.; Blanco, G.; Kašpar, J. J. Phys. Chem. B 2000, 104, 4667−4647. (36) Kusada, K.; Kobayashi, H.; Yamamoto, T.; Matsumura, S.; Sumi, N.; Sato, K.; Nagaoka, K.; Kubota, Y.; Kitagawa, H. J. Am. Chem. Soc. 2013, 135, 5493−5496. (37) Shao, X.; Prada, S.; Giordano, L.; Pacchioni, G.; Nilius, N.; Freund, H.-J. Angew. Chem., Int. Ed. 2011, 50, 11525−11527. (38) Liu, J. ChemCatChem 2011, 3, 934−948. (39) Karim, A. M.; Prasad, V.; Mpourmpakis, G.; Lonergan, W. W.; Frenkel, A. I.; Chen, J. G.; Vlachos, D. G. J. Am. Chem. Soc. 2009, 131, 12230−12239. (40) Ravel, B.; Newville, M. J. Synchrotron Radiat. 2005, 12, 537− 541. (41) Zabinsky, S. I.; Rehr, J. J.; Ankudinov, A.; Albers, R. C.; Eller, M. J. Phys. Rev. B: Condens. Matter Mater. Phys. 1995, 52, 2995−3009.

7584

DOI: 10.1021/acscatal.6b01940 ACS Catal. 2016, 6, 7577−7584