Low-Temperature Ammonia Synthesis on Iron Catalyst with an Electron Donor

Haber–Bosch process produces ammonia to provide food for over 5 billion people; however, it is currently required to be produced without the use of fossil fuels to reduce global CO2 emissions by 3% or more. It is indispensable to devise heterogeneous catalysts for the synthesis of ammonia below 100–150 °C to minimize the energy consumption of the process. In this paper, we report metallic iron particles with an electron-donating material as a catalyst for ammonia synthesis. Metallic iron particles combined with a mixture of BaO and BaH2 species in an appropriate manner could catalyze ammonia synthesis even at 100 °C. The iron catalyst revealed that iron can exhibit a high turnover frequency (∼12 s–1), which is over an order of magnitude higher than those of other transition metals used in highly active catalysts for ammonia synthesis. This can be attributed to the intrinsic nature of iron to desorb adsorbed hydrogen atoms as hydrogen molecules at low temperatures.

The schematic illustration of the reactor set-up in this study is shown in Figure S1. The stainless steel fixed bed reactor (SUS 304) used for evaluation of the catalytic performance included a duplex tube with a 300 mm SUS 304 outer reaction tube (170 mm (external/internal diameter: 9.53 mm/7.05 mm) + 130 mm (external/internal diameter: 6.35 mm/4.35 mm)) and a SUS 304 inner tube (200 mm (external/internal diameter: 3.18 mm/2.18 mm)) as a thermocouple well. The front end of the inner tube was closed. Stainless steel bellow valves were connected to both ends of the outer tube. The middle of the outer reaction tube that was vertically located was filled with catalyst particles. The height of catalyst bed was 9~11 mm. In this study, the catalytic activities of all catalysts were examined without adding any diluent to catalyst bed. A sheathed K-type (chromel-alumel) thermocouple (1.8 mm in diameter) was inserted into the inner tube and the tip of the sheathed thermocouple was in contact with the closed front end of the inner tube which reached the middle of the catalyst bed (a height of ca. 5 mm from the bottom of the catalyst bed) as shown in Figure  S1; the reaction temperature in this study represents that of the middle of the catalyst bed. The reactor was heated in a 800 W ceramic tube furnace (Asahi Rika ARF1-200 (Kanthal wire)).
In preparation for alkaline earth metal hydride-containing catalysts, the reactor was loaded with a mixture of catalyst precursors in an Ar-filled glovebox, as shown in Figure S1. It was confirmed in a quartz reactor that there is no significant difference in catalyst bed height among the mixture of catalyst precursors, the catalyst immediately after preparation in a flow of H 2 and the catalyst after ammonia synthesis reaction. The reactor loaded with the mixture of catalyst precursors was removed from the Ar-filled glovebox to the atmosphere and connected to a flow reaction system without exposure of the mixture in the reactor to the atmosphere. The reactor was heated in the ceramic tube furnace in a flow of H 2 , which resulted in the tested catalysts. The catalytic activities of the prepared catalysts were examined by flowing N 2 -H 2 into the reactor.

Ammonia synthesis over BaH 2 -BaO/Ru/CaH 2
BaH 2 -BaO/Ru/CaH 2 was prepared from commercial metallic Ru nanoparticles (average particle size: 30 nm) in a similar manner to BaH 2 -BaO/Fe/CaH 2 . However, BaH 2 -BaO/Ru/CaH 2 did not synthesize ammonia under the present reaction conditions. It was confirmed that an increase in the Ru particle size (from 4 nm to >10 nm) in the Ru/BaH 2 -BaO catalyst by increasing the amount of Ru deposition significantly decreased the catalytic activity for ammonia synthesis, which suggests that BaH 2 -BaO loading is not effective for large metallic Ru particles. The details are currently under investigation.

H 2 desorption from BaH 2 -BaO/Fe/CaH 2 and Fe/CaH 2
Prior to H 2 -TPD experiments, the sample after the ammonia synthesis experiment at 300 °C for over 20 h was kept at 300 °C for 1 h in a flow of Ar and then cooled down to room temperature. The sample was then heated at a rate of 1 °C min -1 in an Ar flow (see the figure caption Figure 2D). NH 3 -and N 2 -TPD measurements confirmed that ammonia and its derivatives are not adsorbed on the sample before H 2 -TPD experiment. Hydrogen adatoms on the metallic iron-alkaline earth metal hydride system are desorbed as hydrogen molecules below about 100 °C, as shown in Figure 5; therefore, it is difficult for the catalytic system to adsorb hydrogen before H 2 -TPD experiment under the present experimental conditions. In addition, the H 2 and H 2 O concentrations in the TPD experimental system were below the detection limit of mass spectrometry. Therefore, it was considered that H 2 desorption observed on Fe/CaH 2 is due to hydride anions in CaH 2 .

D 2 desorption from BaH 2 -BaO/Fe/CaH 2
There are two types of H 2 desorption on transition metal-alkaline earth metal hydride systems; recombination among H adatoms from gas phase H 2 and H 2 derived from Hanions in alkaline earth metal hydrides. The H 2 desorption temperature for the latter is higher than that for the former because it proceeds through a multi-step process, including hydride defect formation, migration of hydrogen to transition metal surfaces and recombination of hydrogen. In the case of BaH 2 -BaO/Fe/CaH 2 , there is a large difference (about 100 °C) between the former and latter desorption temperatures. It was confirmed that H 2 is desorbed from the catalyst at >100 °C after D 2 desorption below 100 °C. HD formation was not observed on BaH 2 -BaO/Fe/CaH 2 , because almost all the D adatoms are desorbed as D 2 before the migration of H in the hydride to the Fe surface.   DFT results for BaH 2 with H − defects. Total energies and structural relaxations of BaH 2 with/without surface Hanion defects were estimated from density functional theory (DFT) computation based on VASP first-principles code. We adopted the Perdew-Burke-Ernzerhof (PBE) exchange-correlation functional in DFT. The convergence criteria of energy and force were, respectively, 0.5 × 10 -4 eV and 1.0 × 10 -1 eV nm -1 for all models. The core electrons were handled with the projector augmented wave (PAW) method. The k-point mesh was created to keep a single k-point per 1/4 (nm -1 ) in the reciprocal space. In BaH 2 , (0 0 1), (0 1 0), (1 0 0), (0 1 1), (1 0 1), (1 1 0), and (1 1 1) surface models were relaxed using DFT, and the (1 0 0) surface was the most stable surface for BaH 2 (0.37 J m -2 ). A notable feature in the computation is that a vacuum region of 2 nm is maintained in the unit cell.DFT calculations revealed that the (100) surface is the most stable in BaH 2 . The work function of BaH 2 with H − defects (Ba 2+ H -(2-1/9) e -1/9 ) that trap electrons was estimated to be 2.6 eV.   Figure S8. FT-IR spectra for 14 N 2 and 15 N 2 adsorbed BaH 2 -BaO/Fe/CaH 2 (10 kPa of 14 N 2 and 15 N 2 at 25 °C). Spectrum " 15 N 2 +73 cm -1 " was obtained by blue-shifting the FT-IR spectrum for 15 N 2 adsorbed BaH 2 -BaO/Fe/CaH 2 (Spectrum " 15 N 2 ") by 73 cm -1 . Spectrum 15 N 2 +73 cm -1 is consistent with the spectrum for 14 N 2 -adsorbed BaH 2 -BaO/Fe/CaH 2 although the peaks observed in these spectra are weak and broad. Taking the isotope effect into account (ca. 2175 cm -1 × (14/15) 1/2 ), the broad peak below 2175 cm -1 can be attributed to the νN 2 band of N 2 adsorbed on BaH 2 -BaO/Fe/CaH 2 . One possible explanation for the broad band is that the electron-donating capability in the iron catalyst has a broad distribution. While the νN 2 band for BaH 2 -BaO/Fe/CaH 2 is broad, the wavelength range is lower that of Ru/C12A7:e -, which suggests that some of the electron-donating sites on BaH 2 -BaO/Fe/CaH 2 have a stronger electron donating capability than those on Ru/C12A7:e -.