Formaldehyde Selectivity in Methanol Partial Oxidation on Silver: Effect of Reactive Oxygen Species, Surface Reconstruction, and Stability of Intermediates

Abstract

Selective oxidation reactions on heterogeneous silver catalysts are essential for the mass production of numerous industrial commodity chemicals. However, the nature of active oxygen species in such reactions is still debated. To shed light on the role of different oxygen species, we studied the methanol oxidation reaction on Ag(111) single-crystal model catalyst surfaces containing two dissimilar types of oxygen (electrophilic, O e and nucleophilic, O n ). X-ray photoelectron spectroscopy and low energy electron diffraction experiments suggested that the atomic structure of the Ag(111) surface remained mostly unchanged after accumulating low O e coverage at 140 K. Temperature-programmed reaction spectroscopic investigation of low coverages of O e on Ag(111) revealed that O e was active for methanol oxidation on Ag(111) with a high selectivity toward formaldehyde (CH 2 O) production. High surface oxygen coverages, on the other hand, triggered a reconstruction of the Ag(111) surface, yielding Ag oxide domains, which catalyzes methanol total oxidation to CO 2 and decreases the formaldehyde selectivity. This important finding indicates a trade-off between CH 2 O selectivity and methanol conversion, where 93% CH 2 O selectivity can be achieved for an oxygen surface coverage of θ O = 0.08 ML (ML = monolayer) with moderate methanol conversion, while methanol conversion could be boosted by a factor of ∼4 for θ O = 0.26 ML with a suppression of CH 2 O selectivity to 50%. Infrared reflection absorption spectroscopy results and density functional theory calculations indicated that Ag oxide contains dissimilar adsorption sites for methoxy intermediates, which are also energetically less stable than that of the unreconstructed Ag(111). The current findings provide important molecular-level insights regarding the surface structure of the oxidized Ag(111) model catalyst directly governing the competition between different reaction pathways in methanol oxidation reaction, ultimately dictating the reactant conversion and product selectivity.

Nature and Reactivity of Oxygen Species on/in Silver Catalysts during Ethylene Oxidation

“…The trend of Ag x –O 2 agrees with the evolution of Raman bands corresponding to Ag 4 –O 2 that vibrate at 691–721 cm –1 (Figure c). The corresponding Ag 3d NAP-XPS spectra presented in Figure c indicate that Ag is predominately metallic during ethylene oxidation . The Ag 3d difference spectra (Figure d) at various temperatures, however, reveal that the model silver catalyst gradually becomes mildly oxidized during ethylene oxidation, forming Ag δ+ (Ag 3d peaks at 367.3 and 373.2 eV).…”

Identification of Adsorbed Species and Surface Chemical State on Ag(111) in the Presence of Ethylene and Oxygen Studied with Infrared and X-ray Spectroscopies

“…The absence of asymmetric O-C-O stretch peak in the 1530-1610 cm −1 range in our spectra indicates bidentate bridging configuration [41,49], which was reported to appear at low surface coverages of formate on Ag(111) [41]. A weaker feature at around 2907 cm −1 is also present; in fact, the peaks at 2819 cm −1 and 2907 cm −1 are a Fermi resonance doublet with the latter one either due to a combination mode of the asymmetric O-C-O stretching and C-H bending or the overtone of C-H bending [41,49]. Peak positions for formate slightly differ in the present study compared to published values in the literature [41,49], as exact positions depend on coverage and presence of co-adsorbed species such as carbonates and oxygen.…”

“…A weaker feature at around 2907 cm −1 is also present; in fact, the peaks at 2819 cm −1 and 2907 cm −1 are a Fermi resonance doublet with the latter one either due to a combination mode of the asymmetric O-C-O stretching and C-H bending or the overtone of C-H bending [41,49]. Peak positions for formate slightly differ in the present study compared to published values in the literature [41,49], as exact positions depend on coverage and presence of co-adsorbed species such as carbonates and oxygen. Such formate peaks appear due to the formic acid in air [50].…”

“…We performed an additional experiment to better understand the reaction pathway: We dosed 10 mbar ambient air into the IR measurement chamber. At RT right after dosing, the surface was covered with formate species that produced the peaks at around 1346 cm −1 and 2819 cm −1 (Figure S8), which can be ascribed to the symmetric O-C-O stretching and C-H stretching, respectively [49]. The absence of asymmetric O-C-O stretch peak in the 1530-1610 cm −1 range in our spectra indicates bidentate bridging configuration [41,49], which was reported to appear at low surface coverages of formate on Ag(111) [41].…”

“…At RT right after dosing, the surface was covered with formate species that produced the peaks at around 1346 cm −1 and 2819 cm −1 (Figure S8), which can be ascribed to the symmetric O-C-O stretching and C-H stretching, respectively [49]. The absence of asymmetric O-C-O stretch peak in the 1530-1610 cm −1 range in our spectra indicates bidentate bridging configuration [41,49], which was reported to appear at low surface coverages of formate on Ag(111) [41]. A weaker feature at around 2907 cm −1 is also present; in fact, the peaks at 2819 cm −1 and 2907 cm −1 are a Fermi resonance doublet with the latter one either due to a combination mode of the asymmetric O-C-O stretching and C-H bending or the overtone of C-H bending [41,49].…”