Surface science letters

The remarkable surface structure sensitivity of the ammonia synthesis over rhenium single crystals☆

This article describes a theory unifying the unusual performance of the undercoordinated adatoms, point defects, terrace edges, surfaces, and nanostructures of various shapes. The ideas of bond order-length-strength correlation and the associated nonbonding electron polarization (BOLS-NEP) feature that bonds between undercoordinated atoms contract spontaneously. Bond contraction raises the local density of charge and energy. Bond strength gain deepens the interatomic potential well to trap the core and bonding electrons deeply. In turn, the locally and densely entrapped electrons polarize those partially occupying the valence band and above pertaining to the lower-coordinated atoms. The BOLS-NEP theory reconciles the unusual behaviors of undercoordinated systems and the size dependency of nanostructures in their lattice oscillating dynamics, mechanical strength, thermal stability, photon emissivity, chemical reactivity, dielectric permeability, associated with generation of polarized Dirac fermions, serving as carriers for extraordinary catalysis, hydrophobicity, fluidity, lubricity, as well as monolayer high-T_C superconductivity and topological insulator conductivity.

A new bis-benzimidazole-based diamide ligand N²,N^2′-bis((1H-benzo[d]imidazol-2-yl)methyl)-[1,1′-biphenyl]-2,2′-dicarboxamide, L and its three Cu(II) metallatriangles of general formula [Cu₃(L)₃X₃]·3X·nH₂O (where X = Cl, Br, NO₃) have been synthesized and one of them is structurally characterized. X-ray diffraction work reveals that the metallatriangle [Cu₃(L)₃Cl₃]·3Cl·15H₂O crystallizes in trigonal R$\overline{3}$ space group with two independent molecules in the asymmetric unit. The asymmetric unit contains only one-third of the molecules and the rest are generated by the crystallographic $\overline{3}$ axis. Each copper(II) ion adopts a highly distorted square pyramidal geometry. The copper(II) metallatriangles are used as catalyst to carry out the oxidation of substituted benzyl alcohols heterogeneously, in the presence of tert-butyl hydroperoxide. Interestingly, the ratio of product profile and the percentage conversion of the products changes with the surface morphology of the metallatriangle employed as a catalyst. A kite type morphology is found to be highly selective to the formation of acid product over the aldehyde, while a hexagonal type morphology results in a mixed acid + aldehyde product. The initial rate of formation of the aldehyde is found to be almost independent of the amount of catalyst employed.

Catalytic methane combustion was studied over the palladium single crystals Pd(1 1 1), Pd(1 0 0) and Pd(1 1 0). Under lean reaction conditions at 600 K (O₂:CH₄ = 10:1), stoichiometric palladium oxide was formed with an increase in surface area by a factor of approximately two. The oxide phase formed a “cauliflower-like” surface structure composed of approximately 4 nm sized semispherical oxide agglomerates. This oxide structure was independent of the original metal single crystal orientation. The turnover rates over the oxide structure starting with metal single crystals were 0.7 s⁻¹ on Pd(1 1 1), 0.9 s⁻¹ on Pd(1 0 0) and 0.9 s⁻¹ on Pd(1 1 0) at 600 K, 160 Torr O₂, 16 Torr CH₄, 1 Torr H₂O and N₂ balance to 800 Torr, suggesting that the methane combustion reaction is independent of the initial structure of the catalyst. Methane combustion on palladium single crystals experienced an activation period in which the initial turnover rates based on the initial Pd surface area were about 1/8–1/4 of the steady-state rates determined based on the oxide surface area. This activation period was caused by the slow oxidation of palladium single crystals and concomitant surface area increase during reaction. The increase in surface area happened mostly in the first 10 min of reaction. Carbon dissolution into the crystal was not found during methane combustion under reaction conditions in excess oxygen.

Alumina-supported rhenium catalysts were prepared and tested for ammonia synthesis. Ammonium perrhenate was used as the precursor of the supported rhenium catalysts and γ-Al₂O₃ (100 m² g⁻¹) was impregnated with this compound in an aqueous solution. NH₄ReO₄/Al₂O₃ was reduced with pure hydrogen at 823 K for 2 h, and then the sample was activated with the reactant gas or pure hydrogen prior to ammonia synthesis. The catalyst, which was activated with the reactant gas, was found to contain not only Re metal, but also Re nitride, which is more active than Re metal. The cesium promoter (Cs:Re = 1:1) was added to the catalyst by the impregnation method; the resulting activity was significantly higher than that of the non-promoted catalyst. H₂ chemisorption and N_ads+H₂ TPSR experiments revealed that the Re area was decreased by the addition of cesium, whereas the onset of NH₃ formation during the TPSR experiment was not affected. XRD and elemental analyses revealed that the alkali addition has some drawbacks: the formation of cesium perrhenate and the decomposition of rhenium nitride. The ammonia synthesis rates over supported rhenium catalysts were found to increase under high-pressure. Kinetic studies based on power-law rate expressions indicated that the unpromoted Re surfaces were covered by H_ads and NH_x,ads. By the addition of cesium, the degree of ammonia poisoning was not changed, but the inhibition by hydrogen was removed. In summary, cesium is identified as a very efficient promoter for supported Re catalysts.

The rhenium nitride was prepared and applied for ammonia synthesis. Under the synthesis conditions (623 K, 0.1 MPa), the nitride was partly decomposed (Re₃N+Re) but the activity was stable and higher (179 μmol h⁻¹ g⁻¹ at 623 K under 0.1 MPa) than the other transition metal nitrides (VN, NbN, Mo₂N, and W₂N), although the BET surface area was immeasurably low. The active phase was considered to be Re₃N.

The effect of transition metal addition to the rhenium catalyst for ammonia synthesis was investigated. Cobalt and iron were effective, but nickel, chromium, and copper were not effective. The best combination was Co-Re (1:4) (492 μmol h⁻¹ g⁻¹ at 623 K under 0.1 MPa, expressed as Co-Re₄ catalyst). NH₃ treatment of Co-Re₄ at 973 K gave a new XRD pattern (probably some kind of rhenium nitride phase), which was suggested to be responsible to the high activity of Co-Re₄ catalyst. The ammonia synthesis rate over Co-Re₄ catalyst was increased under high pressure (2372 μmol h⁻¹ g⁻¹ at 623 K under 3.1 MPa), however, the rate overshot above 1.1 MPa. The kinetic analysis indicated that Re surface was strongly covered either by hydrogen or by nitrogen (ammonia). By the addition of cobalt, the degree of nitrogen (ammonia) poisoning was weakened, but at the same time, the degree of hydrogen poisoning was strengthened.

The causes of surface structure sensitivity and structure insensitivity of catalytic reactions have been explored on the molecular level by modern surface science. The scanning tunnelling microscopy (STM) and sum-frequency generation (SFG), surface vibrational spectroscopy, were used to monitor metal substrate and adsorbate molecular structures during catalytic reactions of high pressures (atmosphere). There are two phenomena that dominate the surface chemistry on the molecular level.

• 1.

  (1) Absorbate induced restructuring of the metal catalyst is readily detectable.

• 2.

  (2) The strongly chemisorbed molecules are stagnant spectators during olefin hydrogenation while the weakly adsorbed π-bonded species that are present on the metal surface only at high pressures are the reactive intermediates that turn over rapidly.