AmericanChemicalSociety.com

Environmental and political problems created by our dependence on fossil fuels, combined with diminishing petroleum resources are causing our society to search for new renewable sources of fuel.  Biofuels, fuels derived from plant biomass, are the only sustainable source of liquid fuels.  Currently cellulosic biomass resources are significantly cheaper than petroleum (at $15 per barrel of oil energy equivalent) and abundant (having the energy content of 60 % of our domestic crude oil consumption).  While lignocellulosic biomass has a tremendous potential as feedstock for biofuel production, the chief impediment is that economical processes for conversion of lignocellulosic biomass into fuels do not yet exist. 
A large number of processes for converting biomass into biofuels involve hydrogenation of biomass-derived feedstocks in the aqueous phase.  In this study we report the results of a combined experimental and theoretical study for the APH of acetic acid catalyzed by several late transition metals, including Ru, Rh, Pd, Ni, Cu, Ir, and Pt.  High selectivity toward ethanol is obtained with many of these metals, and in particular selectivity approaches 90% with Ru at moderate temperatures.  Density functional theory (DFT) calculations combined with experimental evidence suggest that direct decomposition of acetic acid or acetate to acetyl is likely the first step in activating acetic acid on most of these metals, and that the high selectivity to ethanol is due to either the aqueous phase, the high hydrogen partial pressure employed in the experiment, or both. 

We have calculated the free energies of activation for several different initial C-O bond scission steps on the metals, including the dehydroxylation of 1,1-ethanediol, ethane-1-ol-1-olate, 1,1-dihydroxyethyl, acetic acid, and ethene-1-ol-1-olate (yielding 1-hydroxyethyl, acetaldehyde, 1-hydroxyethylidene, acetyl, and ketene, respectively) and the deoxygenation of acetate (yielding acetyl), using the formulation of Andersson et al.  The lowest-barrier mechanism for initial C-O bond scission is different on different metals: On Rh, Pd, Ir, and Pt it is the dehydroxylation of acetic acid; on Ru and Ni it is the deoxygenation of acetate although acetic acid dehydroxylation is closely competitive; on Cu it is the dehydroxylation of ethane-1-ol-1-olate.  Thus on all the metals except Cu, the hydrogenation of acetic acid should proceed through acetyl at given conditions.  Further hydrogenating acetic acid hinders the initial C-O bond scission. 

Concurrently we have measured the intrinsic activity of Ru/C, Pt/C, Pd/C, Rh/C, Ir/C, Raney Ni, and Raney Cu catalysts for APH of acetic acid at temperatures from 100-260 °C, and at 750 psi total pressure.  The non-precious metal catalysts were Raney type catalysts to avoid sintering reactions that happen under aqueous-phase conditions.  The catalytic activity of all catalysts was typically measured at acetic acid conversion of less than 25% in the absence of diffusion limitations.  The TOF of acetic acid activation decreases as Ru > Rh ~ Pt > Pd ~ Ir > Ni ~ Cu.

In the current experiments, however, the Ru catalyst is not only very active for converting acetic acid, but also offers nearly 90% selectivity toward ethanol below ca. 175 °C .  Besides Ru, all the other metals (except Rh) also offer high selectivity to ethanol in certain ranges of temperature (but never as high as on Ru), in contrast to the previous experiments.  The activity and selectivity observed in the current experiments suggest that hydrogenation is not rate-limiting and that the C-C and 2^nd C-O bond scissions are less competitive against the selective pathway than in the vapour-phase, low-pressure limit. 

The side products, which consist of ethyl acetate, acetaldehyde, methane, ethane, and carbon dioxide, provide additional information for the reaction mechanism on these metals.  The fact that ethyl acetate is a major side product on all the metals suggests the presence of acetyl.¹  Comparison of the selectivities at different temperatures shows that ethanol production is followed closely by ethyl acetate and acetaldehyde production on the Ru, Rh, Pd, Pt, and Ni catalysts.   The synchronized change in the amounts of ethanol, ethyl acetate, and acetaldehyde produced (even though they are not in equilibrium) suggests a common precursor, likely also acetyl, whose formation is rate-controlling.  The small amounts of methane produced at moderate temperatures on Ru suggest that CO poisoning is probably not significant in our experiments.

In conclusion, the activity and selectivity of several transition metal catalysts, including Ru/C, Rh/C, Pd/C, Ir/alumina, Pt/C, Raney Ni, and Raney Cu, toward the selective hydrogenation of acetic acid to ethanol have been shown to be distinctly different.  While non-selective decomposition of acetic acid prevails was found to prevail in previous vapour-phase experiments, high selectivity to ethanol is obtained on most of the metals in the current experiments, with Ru being by far the most active and selective (nearly 90% at mild temperature).  DFT calculations suggest that the initial C-O bond scission occurs in acetic acid or in acetate, rather than in hydrogenated precursors, on all the metals (except Cu), thereby generating acetyl (CH₃CO).  The experimental total activities across the metals show strong correlation with the adsorption energy of atomic C, which is interpreted in light of the DFT results as due to a rate-limiting step involving the generation of either acetyl or 1-hydroxyethylidene.  The observed production of ethyl acetate and acetaldehyde in tandem with ethanol on most of the catalysts also suggests the involvement of a common surface intermediate, likely acetyl.  The combined DFT and experimental evidence thus points to the stability of acetyl as a key parameter that likely controls both the activity and selectivity of this reaction.  Since the adsorption energy of acetyl is strongly correlated with that of atomic C, and since the latter is a readily calculated quantity, it could serve as a simple indicator for the activity of acetic acid conversion on a given metal.  The mechanistic underpinning of the improved selectivity is the subject of continued investigation.