Ammonia Synthesis on Wool-Like Au, Pt, Pd, Ag, or Cu Electrode

Aug 29, 2017 - Developing an ammonia synthesis process from N2 and H2 is of interest in the catalysis and hydrogen research community. Wool-like metal...
3 downloads 0 Views 1MB Size
Download PDF
Research Article pubs.acs.org/acscatalysis

Ammonia Synthesis on Wool-Like Au, Pt, Pd, Ag, or Cu Electrode Catalysts in Nonthermal Atmospheric-Pressure Plasma of N2 and H2 Masakazu Iwamoto,*,†,§ Mao Akiyama,† Keigo Aihara,† and Takashi Deguchi† †

Research and Development Initiative, Chuo University, 1-13-27 Kasuga, Bunkyo-ku, Tokyo 112-8551, Japan S Supporting Information *

ABSTRACT: Developing an ammonia synthesis process from N2 and H2 is of interest in the catalysis and hydrogen research community. Wool-like metal electrodes used to produce nonthermal plasma were determined to serve as efficient catalysts for ammonia synthesis under atmospheric pressure without heating. The catalytic activity of Pt, Pd, Ag, Cu, and Ni wools increased as the experiment was repeated, while that of Au, Fe, Mo, Ti, W, and Al was almost constant. The activity change was mainly due to migration of metals from the electrode to the inner wall of a silica reactor or increases in surface areas of metal catalysts. The order of the activity at each initial experiment was Au > Pt > Pd > Ag > Cu > Fe > Mo > Ni > W > Ti > Al. DFT calculations using Gaussian 09 and CASTEP were applied for energy changes in a reaction M3 + 1/2 N2 → M3N and in adsorption of a nitrogen atom on metal surface, in which M3 was a virtual minimum unit of the metal surface. The reactions were assumed to be an essential step in the ammonia production after plasmaactivation of N2. The resulting values correlated with the respective initial catalytic activity, indicating that a more unstable M3N surface intermediate produced higher catalytic activity. Emission spectra in the plasma process using various electrodes were measured and showed that the efficiency of electrodes for plasma activation of nitrogen molecules was almost independent of the metals, while the reactivity of the activated species to form ammonia depended greatly on the metal used. The N2/H2 ratio dependence and formation/decomposition rate constants of ammonia were finally determined on Au and Cu, which were different from those for the conventional Haber−Bosch process. The decomposition of produced ammonia was suggested to proceed in a plasma-irradiated gas phase. KEYWORDS: ammonia synthesis, plasma, nonthermal, catalysis, gold, platinum, palladium, copper

■

catalysis,5−9,13−15 the turnover number of the recently reported catalyst has been confined to approximately 100,9,13 indicating the need for extensive research efforts to make a practical homogeneous catalysis system. Electro-catalysis using Ag−Pd alloys or Ru/perovskite oxides16−19 was also reported to be active for the ammonia formation, although the production rates were far lower than those of the Haber−Bosch process. In the field of plasma-assisted catalysis, a nonthermal atmospheric plasma was applied to synthesize ammonia,20−24 and catalystloaded ceramic membranes or bimetallic oxide systems were used to obtain yields of approximately 2%. However, these have disadvantages, including slow production rates, due to the slow flowing velocity of the reactant gases.21 No experimental result, except for the copper wool catalysis in our previous report,25 could provide a useful and novel strategy or approach for ammonia synthesis that might be superior to the current Haber−Bosch process. Our efforts were devoted to developing the combined process of heterogeneous catalysis with plasma irradiation.

INTRODUCTION A new synthesis method for ammonia is strongly desired from an energy-saving perspective1−9 because the Haber−Bosch process, the main industrial process for producing ammonia, requires extreme reaction conditions, high temperatures and pressures, and it consumes 1−2% of the world’s annual primary supply. However, one of the significant challenges with hydrogen use is the storage and transportation of hydrogen. Ammonia is very suitable for storage and transportation because of its easy liquefaction under slight pressurization, the high level of hydrogen stored in a molecule, and the production of dinitrogen and water after use.10 Active studies have been continued in four fields to improve or displace the current heterogeneous catalytic systems, including synthesis using newly developed catalysts in a heterogeneous phase, metal complex catalysis in a homogeneous phase, electrochemical synthesis in an aqueous solution, and synthesis with plasma irradiation. In the heterogeneous catalysis,8,11−13 Ru/electride catalysts were recently reported to give 0.7% of an ammonia yield at atmospheric pressure and 2.1% at 10 atm.8,12 However, the reaction rate on the Ru/ electride was low due to the low surface area of the newly developed crystalline support. In the field of metal complex © 2017 American Chemical Society

Received: May 18, 2017 Revised: August 25, 2017 Published: August 29, 2017 6924

DOI: 10.1021/acscatal.7b01624 ACS Catal. 2017, 7, 6924−6929

Research Article

ACS Catalysis We found a new plasma reaction to effectively synthesize ammonia at atmospheric pressure without heating.25 When copper wool (a thin-wire assembly) was used as an inner electrode to produce nonthermal plasma, it worked as a very effective synthetic catalyst. The reaction rate and yield greatly increased as the experiments were repeated. The new reaction system gave an ammonia yield of 3.5% at H2/N2 = 3 and a total flow rate of 100 mL min−1 after the activity was stabilized in repeated experiments. There is a possibility that the plasmaassisted catalysis provides new insights into ammonia synthesis; therefore, our study was expanded to the catalysis of various metal wools. The idea demonstrated high catalytic activity of gold and others; furthermore, a DFT calculation indicated the factor determining the catalytic activity for ammonia synthesis. The results will be significant for achieving future progress in an ammonia synthesis and hydrogen society.

■

Figure 1. Change in the catalytic activity of various metal-wool electrodes for ammonia synthesis with repetition of plasma experiments. Reaction conditions: applied voltage, 5 kV; frequency, 50 kHz; electrode length, 150 mm; wool-like metal, 61.3 cm2; total flow rate, 100 mL min−1; and H2/N2 = 1.

EXPERIMENTAL SECTION Various metal thin-wires (wool) were employed as the inner electrode in the plasma experiments, as summarized in Table S1 (in Supporting Information). The quartz tubular reactor and electrodes are shown in Figure S1. The outer diameter and thickness of the quartz tube were 12.7 mm and 1.0 mm, respectively. The outer side of the quartz reactor was surrounded by the outer electrode, which consisted of a copper net. The internal wool-like electrode was connected to a high voltage power supply, and the outer electrode acted as a grounded electrode. All experiments were performed at atmospheric pressure without heating. A mixture of N2 and H2 was flowed into the reactor from the top of the reactor, and the exit gas was delivered to a diluted H2SO4 aqueous solution to gather the produced ammonia. Typical reaction conditions included an applied voltage of 5 kV and a frequency of 50 kHz for the reaction port (Figures S2 and S3), which was 150 mm in length. A mixture of H2/N2 = 1 was flowed into the reactor at a flow rate of 100 mL min−1 unless otherwise stated. Wool-like metallic wires with surface areas of 61.3 cm2 (for example, 2.07 g of Au with a diameter of 0.07 mm) were employed as the inner electrode. It is important to note that the level of ammonia produced in the current reaction system varied with the number of production runs. The production rates and yields of ammonia were measured 5 times, and the results observed in the first experiment were used to compare the catalytic activity of the respective metals because they were the values with no influence on the repetition or little change in the surface states. A detailed experimental description is given in the Supporting Information.

activity of the former metals from the first run to the fifth were 36.4 μmol min−1 on Ag, 29.9 on Cu, 22.0 on Pd, 19.7 on Pt, and 17.5 on Ni. The repetition on these metals was accompanied by the formation of metallic spots, lines, and wafers on the inner wall of the silica tubular reactor, which is similar to those observed on a copper-wool electrode.25 The levels of deposited metals, determined by chemical analysis, were 1.5, 21.2, 3.3, 3.5, and 2.1 μmol after the fifth experimental runs for Ag, Cu, Pd, Pt, and Ni, respectively. The order of the amounts of deposited metals on the silica wall was considerably different from the rank of the increments in catalytic activity after the fifth experiment. On the other hand, the levels of metals deposited after the fifth runs of Au, Fe, Mo, Ti, W, or Al electrodes were 0.8 μmol or lower. The findings indicated that the increases in the activity might result from the deposited metals. This possibility was confirmed on copper electrodes. After the fifth experiment of copper wool, the inner electrode was replaced with an 8 mmϕ carbon rod and the catalytic activity was measured. The NH 3 production rate on combination of the carbon rod and the copper film deposited on the silica wall was 41.0 μmol min−1, while that on an 8 mmϕ carbon rod alone was 10.6 μmol min−1. The difference, 30.4 μmol min−1, was almost equal to the increment observed on the copper wool catalyst in Figure 1, 29.9 μmol min−1, indicating that the deposited copper functioned as the catalyst. In a previous report,25 the formation and movement of some copper nitride intermediates was suggested to result in the formation of metal deposition on the inner wall of the silica tube. Therefore, properties of corresponding metal nitrides, including melting points, were investigated with a literature search. The results are summarized in Table S2, although only a few data were available. Unfortunately, we did not find any direct correlation among the properties, the amounts of deposited metals, and the catalytic activities. However, it could be noted that the higher melting points of metal nitrides (for example, AlN, TiN, MoN, and WN) resulted in the lower catalytic activity of metals and little or no changes in catalytic activity with the repetition of experiments. The activity of ruthenium, which is known to be very active for the catalytic formation of ammonia in the Haber−Bosch reaction, could not be determined due to the lack of commercial Ru wires.

■

RESULTS AND DISCUSSION Effect of Electrode Material on Ammonia Production. The ammonia yield in the plasma-assisted reaction substantially increased when some wool-like metal thin wires were used as electrodes. Therefore, the change in the ammonia yield was first studied as a function of the repetition of experimental runs. The results are summarized in Figure 1 where the surface areas of the respective metal wires were adjusted to 61.3 cm2, which is the surface area of 1.10 g of copper wool, by changing their weights. The catalytic activity of Pt, Pd, Ag, Cu, and Ni increased as the experiments were repeated, while that of Au, Fe, Mo, Ti, W, and Al was approximately constant during the experiments, except for changes in the first experiments of Au and Fe, as shown in Figure 1. The increments in the catalytic 6925

DOI: 10.1021/acscatal.7b01624 ACS Catal. 2017, 7, 6924−6929

Research Article

ACS Catalysis

calculation (Gaussian 09 with the Hamiltonian B3LYP) on various metals. An M3 cluster (M = an electrode metal atom) was employed as a minimum model of the catalyst surface. The reaction energy without zero-point energy correction, ΔE, was calculated according to eq 1, in which E represents the selfconsistent field (SCF) energies of the molecules shown in the brackets (a complex, cluster, and substrate).

The order of the initial catalytic activity was Au > Pt > Pd > Ag > Cu > Fe > Mo > Ni > W > Ti > Al, indicating that the gold thin-wire assembly exhibited the highest activity for ammonia synthesis. It is important to note that the activity of the metal catalysts for ammonia synthesis in plasma greatly differed from that of the Haber−Bosch process because of the difference between the activation of N2 in the conventional heterogeneous catalysis and that in the plasma. In a typical Haber−Bosch process, the triple bond between the two nitrogen atoms is activated by electron donation from the catalyst. However, in the plasma process, a nitrogen molecule is activated in advance by the plasma, and the catalyst provides the reaction site for the activated species. It is worth adding that the present results would be the first experimental results showing that gold has the highest catalytic activity of the precious metals even in the lump state. It has been reported in a plasma chemistry that an energy consumption is one of important factors for reaction. The energy consumptions were calculated using an equation ∫ V(t) I(t)dt from the V−I results summarized in Figures S2 and S3 instead of the conventional Lissajous method, because the Lissajous calculations for the current complicated waveforms were not yet decided. The values are compared with the respective ammonia synthesis rates in Figure 2. The measured

ΔEf = E[M3N] − E[M3] − E[1/2N2]

(1)

The initial bond lengths of metals were the values of the respective crystals, and the initial M-M-M bond angle was 60°. The metal atoms were fully relaxed or frozen in the calculation. The E[M3N] values were first calculated for all potential spin multiplicities. The multiplicity giving the minimum-energy structure was assumed as the most allowable value for each M3N complex and was used for calculating E[M3]. The results are summarized in Table 1 where the metals are given in the order of the atomic number. The table indicated that the nitride species of Al, Ti, Fe, Mo, and W are stable because of their Table 1. DFT Calculations of Energy Changes in Formation of Metal Nitrides M3N (Gaussian 09) or Adsorption of a Nitrogen Atom on Metal Surface (CASTEP)

Figure 2. Catalytic activity for ammonia synthesis (green rods), energy consumptions (sky blue circles), and intensities of photoemission spectra of N2* (337 nm, red circles) and N2+ (392 nm, deep blue rectangles; the intensity was multiplied by 10) on various wool-like metal electrodes. Reaction conditions: applied voltage, 5 kV; frequency, 50 kHz; electrode length, 150 mm; wool-like metal, 61.3 cm2; total flow rate, 100 mL min−1; and N2 = 100% (for measurement of photoemission spectra) or H2/N2 = 1 (for ammonia synthesis).

values considerably varied with the metals used as inner electrodes, but they could roughly be regarded as constant and were not correlated with the production rates of ammonia. Next, photographs of plasma experiments are shown in Figure S4 to investigate the possibility that the plasma might not be uniform in the reactor due to the unevenly packed wool electrodes. In the experiments of Figure S4, ordinary copper wool was used as the inner electrode and a 10 mesh copper net was employed as the outer electrode instead of an ordinary 80 mesh copper net to observe the inside of the quartz reactor. Uniform atmospheric-pressure glow discharge was observed in the reactor as shown in the pictures, indicating little heterogeneity of plasma. Factors Controlling the Catalytic Activity for Ammonia Synthesis in Plasma. To identify the reasons for the high activity of gold and others, energy changes in the surface nitride formation ΔEf were calculated using a density functional theory

Optimized structures, multiplicity, and formation energies ΔEf (kcal mol−1) of metal nitrides, M3N, were calculated using Gaussian 09 with B3LYP as the Hamiltonian and LANL2DZ as the basis. bTo calculate adsorption energies of a nitrogen atom ΔEad (kcal mol−1) on various metal surfaces, a CASTEP program was applied on the plane-wave basis set with GGA-PBE, polarized spin, and a 2 × 2 × 3 super cell with a fixed third layer. cThe M3 clusters were relaxed or frozen in the calculation. The multiplicity indicates the spin multiplicity of a M3N cluster that as optimized by the method described in the main text. A nitrogen atom is shown with a blue ball in each structure. a

6926

DOI: 10.1021/acscatal.7b01624 ACS Catal. 2017, 7, 6924−6929

Research Article

ACS Catalysis

Figure 3. Dependence of the respective initial activity of metal wool catalysts for ammonia synthesis (A) on the formation energy of trinuclear nitride M3N or (B) on the adsorption energy of a nitrogen atom on the metal surface. The former and the latter was determined by a DFT calculation using Gaussian 09 or CASTEP, respectively.

negative ΔEf values; those of Cu, Pd, Ag, Pt, and Au instable; and that of Ni intermediate stability. In Figure 3A, the initial catalytic activity in Figure 1 is plotted as a function of the formation energy of the corresponding nitride (ΔEf) where the results of the relaxed and frozen M3N were both employed. One can recognize a fair linear correlation between the ΔEf and catalytic activity. The more instable M3N had a better ammonia yield. Although calculation for the energy profiles of detailed reaction pathways should be performed, the linear correlation indicated the first step of the current plasma synthesis of ammonia, the reaction of an activated nitrogen molecule with the electrode surface, would be the most important, rate-limiting step for the catalysis. It would be worthwhile to note that the melting points of corresponding metal nitrides listed in Table S2 were roughly correlated with the calculated M3N formation energies: that is, the metal nitrides having the lower melting points showed the positive ΔEf values. Since the formation of surface nitride species was indicated to be the rate-limiting step, adsorption energies of a nitrogen atom on the metal-wool catalysts was estimated using firstprinciples quantum mechanics calculations with supercell models using CASTEP,26 which employs the density functional theory plane-wave pseudopotential method. GGA-PBE functional27 was used with ultrasoft pseudopotentials.28 The energy cutoff and Monkhorst−Pack mesh of k points were 400 eV and (4 × 4 × 1), respectively, for all the models. A vacuum space with a height of 20 Å was placed over the surface of each supercell model. The energy level of a nitrogen atom located on the surface of a (2 × 2 × 3) supercell was determined by eq 2, in which ΔEad, E[N−M], E[M], and E[N2] represent the energy level, the energies of the N-bearing slab, the mother metal, and N2, respectively. Here, calculation of [N−M] and [M] was performed in a spin-polarized mode and on each closest packing strucuture. ΔEad = E[N−M] − E[M] − E[N2]/2

triangular metal atoms in the CASTEP calculation. This adsorption state will be also expressed with M3N in the following discussion. Emission spectra were measured during the plasma reaction using a UV−vis spectrometer JASCO V-670 with an optical fiber unit OBF-832 to observe active species for the NH3 synthesis. A typical spectrum is shown in Figure S5 where N2 alone was flowed onto 1.10 g of Cu wool and the spectrum was observed from the lower end of the quartz reactor. A very intense N2* band (C3Πu → B3Πg, a second positive band from an activated neutral nitrogen molecule) was observed with weak bands for N2* (B3Πg → A3Σu, a first positive band), N2* (C″5Σu → A′5Σg, a Herman infrared band), N2+ (B2Σu+ → X2Πg+, a first negative band), and NO (A2Σ3+ → X2Π, a γ band).24 The appearance of NO γ bands has widely been reported to be due to a trace level of oxygen impurity.24 In the current experiment, the intensity of the N2+ emission bands was far lower than that of N2*. All of these bands, except the NO γ band, increased with the increasing applied voltage and decreased with the introduction of H2. The intensities of N2* and N2+ in the N2 flow are summarized in Figure 2 with the formation rate of NH3 in the presence of H2 (H2/N2 = 1, the result of Figure 1). The N2* emission intensity was regarded as roughly constant, irrespective of the metal species, although they varied widely. This would be because plasma activation of nitrogen molecules roughly depends on the waveform, frequency, voltage, and current of plasma instead of on the metal species of electrodes. Taking the results of Figure 3 into consideration, we could propose the following reaction schemes for ammonia production. Here, a trinuclear active site is an exemplification for the active site. The generation of the N2* species (eq 3) was independent of metal species used as the electrodes. The metal types would strongly affect the reaction of generated N2* species with the metal surface (eq 4). The reactivity of the resultant M3N might be important for progress in subsequent reactions with hydrogen atoms on the surface (eq 5). Decomposition of produced NH3 (eq 6) is possibly catalyzed on the metal surface or proceeds in the gas phase, which will be discussed in the next section.

(2)

The results are summarized in Table 1, and the correlation of ΔEad with the catalytic activity is plotted in Figure 3B. A good correlation similar to that of Figure 3A was found, although there were some dispersion. The surface nitride formation would be almost the same as the adsorption of a nitrogen atom on metal surface. It is noteworthy that all the adsorption of a nitrogen atom on the surface progressed in the dimple of 6927

N2 + e → N2* + e (in a gas phase)

(3)

N2* + 2M3 (a tri‐nuclear surface site) → 2M3N

(4)

M3N + 3H (on metal surface) → → → M3 + NH3

(5)

DOI: 10.1021/acscatal.7b01624 ACS Catal. 2017, 7, 6924−6929

Research Article

ACS Catalysis 2NH3 + metal surface or in a gas phase → → → N2 + 3H 2

The production rate of ammonia was studied as a function of the residence time of the reactant gases. The residence time of the mixture in the reaction port was defined by V/F [min], where V is the volume of the reaction port [mL] and F the total flow rate of the reactants [ml min−1]. The results are summarized in Figure S7. The production rate increased as the residence time increased and then became roughly constant for a longer time. The correlation was analyzed using the following equations.

(6)

Kinetics of Ammonia Synthesis on a Gold Thin-Wire Assembly Electrode. The ammonia yield on a gold thin-wire assembly electrode was examined as a function of the molar ratio of H2 and N2. The ammonia synthesis rates and yields are summarized in Figure 4. The synthesis rate was maximized at

(8)

x = (r0/kd)(1 − exp(−kd t))

(9)

Here we assumed that the formation rate on the electrode was constant (r0 [min−1]) because the changes in the partial pressures of N2 and H2 were small before and after the reaction and both pressures can be treated as approximately constant. Also, the decomposition rate of ammonia was proportional to the rate constant (kd [min−1]) and the partial pressure of produced ammonia (x [−]). eq 8 is thus formulated and converted to eq 9 by integration. We applied a least-square method to determine the parameters on the gold wires. r0(Au) and kd (Au) were 0.212 min−1 and 8.4 min−1, respectively. For comparison, the measurements and analyses were applied for catalysis using Cu. The production rates on the fresh Cu wires are also summarized in Figure S7 and gave the following values: r0i(Cu) = 0.097 min−1 and kdi(Cu) = 8.4 min−1. The formation rate r0i on the fresh Cu was smaller than that on Au, and the decomposition rate constant kdi on the fresh Cu was the same as that on Au, resulting in the low activity of the fresh Cu for ammonia synthesis in the plasma system. We have already reported that the r0s and kds values on the stabilized Cu catalyst were 0.268 and 8.9 min−1, respectively,25 which showed a significant change in the r0 values and almost unchanged kd values, resulting in a significant increment in the ammonia synthesis activity. Another significant finding is that all decomposition constants determined here are approximately coincident. This might suggest progress of the decomposition reaction of ammonia in a plasma-irradiated gas phase, although more detailed experiments on ammonia decomposition should be performed. In conclusion, a wool-like gold electrode was determined to be a novel and very effective catalyst for ammonia synthesis using nonthermal atmospheric-pressure plasma. The initial yield of ammonia from H2 and N2 reached 2.5% without heating. Gold was the most active catalyst among the studied catalysts, Al, Ti, Fe, Ni, Cu, Zn, Mo, Pd, Ag, W, Pt, and Au. This would be due to the appropriate instability of surface nitride on Au to promote the reaction of nitrogen and hydrogen. The reaction mechanism, including the active species in the gas phase and on the surface, is under investigation. However, the current results pave the way for a new approach to ammonia synthesis that may be significant for future hydrogen applications.

Figure 4. Change in the synthesis rate and yield of ammonia as a function of the H2/N2 ratio using a wool-like gold or copper electrode. Reaction conditions: applied voltage, 5 kV; frequency, 50 kHz; electrode length, 150 mm; wool-like electrode, 61.3 cm2; and total flow rate, 100 mL min−1.

H2/N2 = 1 and reached 73 μmol min−1, while the yield increased monotonously as the H2/N2 ratio increased due to a decrease in the concentration of introduced N2. The yield reached 2.5% under the current reaction conditions. The correlation was analyzed using the following kinetic equation, where the reaction rate was expressed using a power equation. rNH3 = kPH2 α PN2 β

dx /dt = r0 − kdx

(7)

A least-square method was applied to determine the α and β values, as shown in the Supporting Information (Figure S6). α(Au) = 1.50 and β(Au) = 1.46 were obtained from the calculations. These values differed from those reported on wellknown solid catalysts (α = 1.5−2.2 and β = 0.9−1.2 on Fe, and α = −0.43 and β = 1.0 on Ru−Cs/MgO)22 and previously reported values of αs(Cu) = 0.77 and βs(Cu) = 1.1625 on a Cu electrode stabilized by the repetition of the plasma experiments (the superscript “s” means “stabilized”). Therefore, similar analysis was again performed on a fresh copper thin-wire assembly electrode for comparison. The H2/N2 ratio dependence was determined on copper electrodes that were newly installed in the respective experiments, and the results are shown in Figure 4. The observed values were αi(Cu) = 1.51 and βi(Cu) = 0.17 (the superscript “i” means “initial”), which were very different from those observed on the stabilized Cu. The changes in the partial pressure dependences with the metal and repetition of experiments were recognized, but detailed discussion would require in-depth understanding of the reaction mechanisms. The positive values on the partial pressure of hydrogen on Au and Cu catalysts are significant for the practical application of the current method because the increase in the hydrogen partial pressure would not result in a decrease in the ammonia formation rate.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acscatal.7b01624. Details of experimental, materials used, physical properties of metal nitrides, details of apparatus, waveform of applied voltage and current, V(t)−I(t) plots, photographs of plasma experiments, emission spectra of 6928

DOI: 10.1021/acscatal.7b01624 ACS Catal. 2017, 7, 6924−6929

Research Article

ACS Catalysis

■

(19) Kosaka, F.; Noda, N.; Nakamura, T.; Otomo, J. J. Mater. Sci. 2017, 52, 2825−2835. (20) Bai, M.; Zhang, Z.; Bai, X.; Bai, M.; Ning, W. IEEE Trans. Plasma Sci. 2003, 31, 1285−1291. (21) Mizushima, T.; Matsumoto, K.; Ohkita, H.; Kakuta, N. Plasma Chem. Plasma Process. 2007, 27, 1−11. (22) Gomez-Ramirez, A.; Cotrino, J.; Lambert, R. M.; GonzalezElipe, A. R. Plasma Sources Sci. Technol. 2015, 24, 065011. (23) Hagen, S.; Barfod, R.; Fehrmann, R.; Jacobsen, C. J. H.; Teunissen, H. T.; Chorkendorff, I. J. Catal. 2003, 214, 327−335. (24) For example: (a) Fan, Z. Y.; Newman, N. Appl. Phys. Lett. 1998, 73, 456−458. (b) Ricard, A.; Decomps, Ph.; Massines, F. Surf. Coat. Technol. 1999, 112, 1−410. (c) Naveed, M. A.; Qayyum, A.; Ali, S.; Zakaullah, M. Phys. Lett. A 2006, 359, 499−503. (d) Rehman, N. U.; Anjum, Z.; Masood, A.; Farooq, M.; Ahmad, I.; Zakaullah, M. Opt. Commun. 2013, 296, 72−78. (25) Aihara, K.; Akiyama, M.; Deguchi, T.; Tanaka, M.; Hagiwara, R.; Iwamoto, M. Chem. Commun. 2016, 52, 13560−13563. (26) Clark, S. J.; Segall, M. D.; Pickard, C. J.; Hasnip, P. J.; Probert, M. J.; Refson, K.; Payne, M. C. Z. Kristallogr. - Cryst. Mater. 2005, 220, 567−570. (27) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996, 77, 3865−3868. (28) Vanderbilt, D. Phys. Rev. B: Condens. Matter Mater. Phys. 1990, 41, 7892−7895.

plasma-activated nitrogen, calculation of ammonia synthesis rates, dependence of the ammonia yield on the residence time (PDF)

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Masakazu Iwamoto: 0000-0001-9141-1873 Present Address §

M.I.: Research Institute for Science and Engineering, Waseda University, 3-4-1 Okubo, Shinjuku-ku, Tokyo 169−8555, Japan. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS A part of the catalytic activity was measured by Dr. Masashi Tanaka. The authors appreciate his cooperation. This work was supported by Grants-in-Aid from the Japan Society for Promotion of Science (JSPS, METI), Japan Science and Technology Agency (JST, METI), and the New Energy and Industrial Technology Development Organization (NEDO, MITI).

■

REFERENCES

(1) Vojvodic, A.; Medford, J.; Studt, F.; Abild-Pedersen, F.; Khan, T. S.; Bligaard, T.; Norskov, J. K. Chem. Phys. Lett. 2014, 598, 108−112. (2) Hoelscher, M.; Leitner, W. Chem. Eur. J. 2017, DOI: 10.1002/ chem.201604612. (3) Connor, G. P.; Holland, P. L. Catal. Today 2017, 286, 21−40. (4) Shipman, M. A.; Symes, M. D. Catal. Today 2017, 286, 57−68. (5) Cherkasov, N.; Ibhadon, A. O.; Fitzpatrick, P. Chem. Eng. Process. 2015, 90, 24−33. (6) Khoenkhoen, N.; de Bruin, B.; Reek, J. N. H.; Dzik, W. I. Eur. J. Inorg. Chem. 2015, 2015, 567−598. (7) Giddey, S.; Badwal, S. P. S.; Kulkarni, A. Int. J. Hydrogen Energy 2013, 38, 14576−14594. (8) Hara, M.; Kitano, M.; Hosono, H. ACS Catal. 2017, 7, 2313− 2324. (9) (a) Tanabe, Y.; Nishibayashi, Y. Coord. Chem. Rev. 2013, 257, 2551−2564. (b) Nishibayashi, Y. Inorg. Chem. 2015, 54, 9234−9247. (c) Tanaka, H.; Nishibayashi, Y.; Yoshizawa, K. Acc. Chem. Res. 2016, 49, 987−995. (10) Dalebrook, A. F.; Gan, W.; Grasemann, M.; Moret, S.; Laurenczy, G. Chem. Commun. 2013, 49, 8735−8751. (11) Aika, K.; Hori, H.; Ozaki, A. J. Catal. 1972, 27, 424−431. (12) (a) Kitano, M.; Inoue, Y.; Yamazaki, Y.; Hayashi, F.; Kanbara, S.; Matsuishi, S.; Yokoyama, T.; Kim, S.; Hara, M.; Hosono, H. Nat. Chem. 2012, 4, 934−940. (b) Kanbara, S.; Kitano, M.; Inoue, Y.; Yokoyama, T.; Hara, M.; Hosono, H. J. Am. Chem. Soc. 2015, 137, 14517−14524. (13) (a) Arashiba, K.; Miyake, Y.; Nishibayashi, Y. Nat. Chem. 2011, 3, 120−125. (b) Arashiba, K.; Kinoshita, E.; Kuriyama, S.; Eizawa, A.; Nakajima, K.; Tanaka, H.; Yoshizawa, K.; Nishibayashi, Y. J. Am. Chem. Soc. 2015, 137, 5666−5669. (c) Kuriyama, S.; Arashiba, K.; Nakajima, K.; Matsuo, Y.; Tanaka, H.; Ishii, K.; Yoshizawa, K.; Nishibayashi, Y. Nat. Commun. 2016, 7, ArticleNo. 12181. (14) Yandulov, D. V.; Schrock, R. R. Science 2003, 301, 76−78. (15) Shima, T.; Hu, S.; Luo, G.; Kang, X.; Luo, Y.; Hou, Z. Science 2013, 340, 1549−1952. (16) Amar, I. A.; Lan, R.; Petit, C. T. G.; Tao, S. J. Solid State Electrochem. 2011, 15, 1845−1860. (17) Saadatjou, N.; Jafari, A.; Sahebdelfar, S. Chem. Eng. Commun. 2015, 202, 420−448. (18) Klinsrisuk, S.; Irvine, J. T. S. Catal. Today 2017, 286, 41−50. 6929

DOI: 10.1021/acscatal.7b01624 ACS Catal. 2017, 7, 6924−6929