(582z) Mechanistic and Spectroscopic Evidence for Reactive Intermediate Structures during C-O Bond Rupture in Oxygenates over Metal Phosphide Clusters

Selective hydrogenolysis of C-O bonds in small oxygenates is an important step for the conversion of biomass-pyrolysis oils into fuels and valuable platform chemicals, such as α,ω-diols from furanic species.¹ Metal phosphides (MPs) are promising catalysts for C-O bond rupture as they are inexpensive, chemically and thermally stable, and selectively cleave C‑X bonds (where X=S, N, or O) over C-C bonds.^2,3

Here, we combine kinetic analysis of reaction networks, modulation excitation in situ infrared spectroscopy (MES), and density functional theory (DFT) calculations to describe the mechanism and reactive intermediates for C-O bond rupture over MPs. Rate measurements of C-O bond rupture within 2-methyltetrahydrofuran (2-MTHF; 1-50 kPa 2-MTHF; 0.1-6 MPa H₂; 543 K) and DFT calculations show adsorption and dehydrogenation steps of 2-MTHF are quasi-equilibrated and precede kinetically relevant C-O bond rupture.⁴ The reactive species have lost 1-2 H-atoms from the fully hydrogenated reactant, and ratios of sterically-hindered C-O (³C-O) bond rupture rates to those for unhindered C-O (²C-O) bonds increase with [H₂]^1/2, which suggests H-content of the reactive intermediates for ³C-O and ²C-O rupture differ by the removal of one H‑atom. Sites that bind the reactive intermediates for both ³C-O and ²C-O bond rupture resemble those that bind CO*, NH₃*, and H* on all Ni_xP (where x = 2-2.4) catalysts, which combined with predicted adsorption enthalpies suggest the reactive intermediates are coordinated to Ni ensemble active sites within Ni-rich terminating planes of Ni_xP surfaces. Selectivities for ³C-O bonds are higher than for ²C-O bonds over Ni_xP and much greater than the ³C-O bond selectivities on Ni. Measured and DFT-predicted apparent activation enthalpies (ΔH^‡,473-583 K) for C-O bond rupture indicate that the phosphorus content decreases the ΔH^‡ for ³C-O bond rupture relative to that of ²C-O bond rupture (i.e., ΔH^‡_2CO - ΔH^‡_3CO increases with phosphorus content). Selectivity differences between specific C-O bonds within 2-MTHF reflect differences in the H-content of reactive intermediates, activation enthalpy barriers, and phosphorus content of Ni and Ni_xP clusters.

Direct experimental evidence for the structure of reactive intermediates would distinguish clear connections between the relationship between C-O bond rupture selectivity, surface coordination, and enthalpic barriers over Ni_xP clusters. MES experiments coupled with phase sensitive detection reveal spectral features of reactive intermediates by suppressing those of inactive surface species.⁵ Multivariate curve resolution-alternating least squares algorithm⁶ extract pure-component spectra and surface coverages that indicate the presence of distinct reactive surface intermediates during C-O bond rupture of 2-MTHF on Ni_xP, whose composition and surface coordination are consistent with those implied by kinetic measurements and predicted by DFT calculations. Modulations in H₂ pressure < 2 ks^-1 (i.e., the measured turnover rate for C-O bond rupture in 2-MTHF) allow all intermediates to form and oscillate at the applied frequency; however, experiments performed with modulation frequencies > 2 ks^-1 (e.g., 2.9 ks^-1 – 0.13 s^-1) suppress the creation of reactive intermediates whose barriers for formation correspond to characteristic timescales of formation that are greater than the period of the modulation. Such experiments show that the compositions of the most abundant reactive intermediate (MARI) on Ni and Ni_xP clusters during C-O bond rupture of 2-MTHF are identical; however, the MARI changes from Ni₃(μ³-C₅H₁₀O) to Ni₃(η⁵-C₅H₁₀O) with increasing phosphorus content. The shift in binding configuration with phosphorus content is likely the fundamental cause for increased selectivity towards ³C-O bond rupture. The combination of these kinetic, spectroscopic, and theoretical techniques provides insight into underlying causes of selectivity for reactions of small oxygenates across MP clusters.

References

(1) Chia, M.; Pagan-Torres, Y. J.; Hibbitts, D.; Tan, Q. H.; Pham, H. N.; Datye, A. K.; Neurock, M.; Davis, R. J.; Dumesic, J. A. J. Am. Chem. Soc. 2011, 133, 12675.

(2) Oyama, S. T.; Wang, X.; Lee, Y. K.; Chun, W. J. J. Catal. 2004, 221, 263.

(3) Rensel, D. J.; Rouvimov, S.; Gin, M. E.; Hicks, J. C. J. Catal. 2013, 305, 256.

(4) Witzke, M. E., Almithn, A.; Coonrod, C. L., Hibbitts, D. D.; Flaherty, D. W. J. Catal. to be submitted.

(5) Urakawa, A.; Burgi, T.; Baiker, A. Chem. Eng. Sci. 2008, 63, 4902.

(6) Jaumot, J.; de Juan, A.; Tauler, R. Chemometr. Intell. Lab. 2015, 140, 1.