(495a) Effects of Acid Strength on Solid Brønsted Acid Catalyzed Aldol Condensation Reactions: Experimental and Theoretical Insights

Aldol condensation of carbonyl-containing molecules form intermolecular C-C bonds and remove O-atoms as H₂O providing an effective way to convert biomass-derived molecules – which are high in oxygen content – to valuable fuels and chemicals [1,2]. In this work, condensation reactions of C₃- and C₄-alkanones are investigated experimentally and theoretically using a range of solid acid catalysts, specifically MFI (Al-, Ga-, Fe-, and B-substituted samples), BEA, FAU, and MCM-41. Addition of hydrogenation-dehydrogenation sites on Pt metal catalysts helps elucidate the mechanism for condensation and secondary reactions, and determine the consequences of acid strength and confining environment of microporous zeolite voids.

Solid Brønsted acid catalysts show low selectivity and significant deactivation during aldol condensation due to the formation of reactive α,β-unsaturated oxygenate intermediates that undergo subsequent β-scission or further condensation reactions. The C=C bond of these intermediates can be hydrogenated by co-fed H₂ on Pt, which is present as physical mixtures of supported catalysts (Pt/SiO₂) or ion-exchanged into zeolite pores, leading to selective and stable catalysts useful for the rigorous mechanistic studies of aldol condensation of C₃- and C₄-alkanones.

For the samples investigated, aldol condensation occurs primarily on Brønsted acid sites as determined by in-situ, selective titration of the protons with 2,6-di-tert-butyl-pyridine during reaction, which suppressed rates irreversibly to zero. Infrared spectroscopy at reaction conditions (0.1 to 10 kPa acetone pressure, Al-MFI, 473 K) suggests the H-bonded alkanone is the most abundant surface species as observed by a broadening and shift of the ν(O-H) from 3600 cm^-1to 2440 cm^-1, which is in agreement with density functional theory (DFT) derived frequencies. Kinetic studies show that the condensation rate is first order in oxygenate pressure, consistent with a transition state containing two oxygenate molecules forming a C-C bond when the surface is saturated with the H-bonded oxygenate. This is in agreement with a reaction mechanism where the acid site, here a proton, stabilizes the reactant oxygenate by H-bonding to the oxygen while a basic site, the adjacent framework oxygen, abstracts the α-H to form a reactive enol species. The α-C on this enol species then performs a kinetically-relevant nucleophilic attack on the carbonyl C of a second acetone molecule, creating a C-C bond and forming a ketol that is readily dehydrated. Absence of a kinetic isotope effect (1.0, acetone-D₆) and DFT calculations support that C-C bond formation is the kinetically relevant step. This proposed mechanism defines the measured condensation rate constant as the product of the intrinsic kinetic rate constant to form a C-C bond (k_CC) and the thermodynamic equilibrium constant for the formation of the enol-species from the alkanone (K_enol), as shown in the rate expression below.

r_cond = (k_CCK_enol)P_alkanone

Consequences of acid strength were investigated using a series of isomorphously substituted MFI samples (Al, Fe, Ga, and B). The measured condensation rate constant trends exponentially with DFT-calculated deprotonation energies (DPE) [3] of the solid acid catalysts, suggesting that the relevant precursors for aldol condensation are uncharged prior to forming an ion-pair transition state. This is consistent with the mechanism proposed from the kinetic studies, DFT-calculated barriers, and in-situ IR characterization.

The first-order rate constant for butanone condensation on Al-MFI is four times smaller than that for acetone condensation while maintaining the same mechanism. DFT charge analysis of the transition state structures for the similar kinetically-relevant step of acetone and butanone condensation on catalysts of different acid strength highlight the difference in charge distribution which can be related to the reactivity of the oxygenate species and catalyst.

The authors acknowledge the financial support of BP through the XC² program and computational resources from XSEDE supported by the National Science Foundation.

References

[1] Sad, M., Neurock, M., Iglesia, E. J. Am. Chem. Soc. 133, (2011), 20384.

[2] Taarning, E., et al., Energy Environ. Sci., 4, (2011), 793.

[3] Jones, A., Carr, R., Zones, S., and Iglesia, E., J. Catal. 312, (2014), 58.