A Combined Theoretical and Kinetic Assessment of C–O Bond Rupture Pathways within 2-Methyltetrahydrofuran over Nickel Phosphide Catalyst

The conversion of oxygenates in biomass-derived feedstocks into more valuable fuels and chemicals requires selective C–O bond cleavage while leaving other C–O and C–C bonds intact. Previous studies have shown that metal phosphides (MPs) are promising catalysts for selective C–O, C–S, and C–N bond rupture over C–C bonds [1]. MPs also exhibit remarkably higher selectivity towards breaking sterically-hindered C–O (³C–O) bonds compared to pure transition metals. Our measurements indicate that Ni₂P cleaves ³C–O bonds at a rate three times faster than ²C–O bond cleavage, and has 50 times higher selectivity for ³C–O bond cleavage than pure Ni. The specific mechanism behind C–O bond cleavage on Ni_xP_y, as well as the identity of the reactive intermediate and active site, are not fully understood despite recent studies. Here, we combine kinetic studies and density functional theory (DFT) calculations to investigate the mechanisms of hindered and unhindered C–O bond cleavage in 2-methyltetrahydrofuran (2-MTHF) on Ni_xP_y and Ni surfaces.

Previous studies showed that C–C and C–O hydrogenolysis in alkanes and alkanols occur via unsaturated species formed by sequential quasi-equilibrated dehydrogenation steps [2,3]; the removal of H atoms weakens the C–C and C–O bonds by replacing C–H bonds with C–metal bonds. The measured ratio of products to reactants pressures during the dehydrogenation of 2-methyltetrahydrofuran (2-MTHF) to 2-methylfuran (2-MF) was constant at low conversions (< 1%), suggesting that C–H activation steps are quasi-equilibrated and leaving the C–O bond rupture as the kinetically relevant step. The measured ratio of ³C–O to ²C–O bond cleavage rates (χ) has a half order dependence on hydrogen pressure [H₂]^1/2, indicating that the reactive intermediate for ²C–O bond rupture must lose an additional hydrogen atom compared with the reactive intermediate for ³C–O bond rupture; the identity of those intermediates that undergo C–O bond cleavage can be assessed using DFT calculations.

Periodic plane-wave DFT calculations were performed using the Vienna ab initio Simulation package (VASP) with the RPBE exchange correlation functional and PAW potentials. The (001) and (111) surfaces were chosen to model Ni_xP_y and Ni catalysts, respectively, because they exhibit the lowest surface formation energies. Vibrational frequency calculations were performed to estimate enthalpies and free energies of each state at temperatures relevant to the catalytic reactions. Results show that the effective enthalpy barrier (ΔH҂) for ³C–O bond rupture is 34 kJ mol⁻¹ lower than for ²C–O on Ni₂P(001) and 21 kJ mol⁻¹ on Ni₁₂P₅(001) at 543 K. These DFT-derived enthalpy barriers are consistent with the measured values on 12 nm Ni₂P (23 ± 9 kJ mol⁻¹) and on 19 nm Ni₁₂P₅ (25 ± 5 kJ mol⁻¹). ²C–O bond rupture on Ni(111), however, is more favorable by 73 kJ mol⁻¹. Although C–O bonds activations through partially dehydrogenated intermediates (i.e. ²CH₂–O, ²CH–O, and ³CH–O) have lower enthalpy barriers, the removal of H atoms from the carbon atom vicinal to the C–O bond being broken results in an increase in entropy, rendering ΔG҂ barrier the lowest for the fully dehydrogenated C–O bond. The contribution of each transition states to the measured hydrogenolysis rate depends on their ΔG҂ values, and thus partially dehydrogenated intermediates are unlikely to contribute to the measured rate. These findings confirm the essential role of entropy in addition to enthalpy in determining the rate constants. DFT-predicted ratio of ³C–O to ²C–O bond cleavage rates χ is equal to 2.7 on Ni₂P(001) and 0.53 on Ni₁₂P₅(001) at 543 K and 1 MPa H₂, in agreement with the measured values (~3 on 12 nm Ni₂P and ~0.4 on 19 nm Ni₁₂P₅) at low conversions.

References

[1]   Oyama, S.; Wang, X.; Lee, Y.; Chun, W., Active phase of Ni2P/SiO2 in hydroprocessing reactions. J. Catal. 2004, 221.2, 263−273.

[2]   Flaherty, D.; Hibbitts, D.; Iglesia, E., Theoretical and kinetic assessment of the mechanism of ethane hydrogenolysis on metal surfaces saturated with chemisorbed hydrogen. J. Am. Chem. Soc. 2014, 136, 9664.

[3]   Gürbüz, E.; Hibbitts, D.; Iglesia, E., Kinetic and Mechanistic Assessment of Alkanol/Alkanal Decarbonylation and Deoxygenation Pathways on Metal Catalysts. J. Am. Chem. Soc. 2015, 137(37), 11984−11995.